article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
in march of 2011 the large hadron collider ( lhc ) had begun , once again , colliding protons at a center - of - mass energy of @xmath3 tev . by june
the atlas detector had recorded more than @xmath4 fb@xmath2 of proton - proton collisions . ] .
these proceedings summarize the search for a @xmath0 boson decaying to a lepton ( where lepton is defined as a muon or electron ) plus a neutrino using @xmath1 fb@xmath2 of data collected by the atlas detector @xcite .
there are many extensions to the standard model ( sm ) of particle physics that predict a new high - mass state similar to the @xmath5 boson in the sm @xcite . the sequential , standard model ( ssm ) boson , i.e. the extended gauge boson of ref .
@xcite , is used to model a @xmath0 boson with the same couplings to leptons as the sm @xmath5 boson but at a higher mass .
no @xmath0 boson signal is observed , and the data is used to set new limits @xcite on @xmath6 ( cross section times branching fraction ) as a function of the @xmath0 boson mass .
table [ tab : xsec ] shows the signal production cross sections used for each mass point .
the search is performed in the individual electron and muon channels where limits on @xmath7 and @xmath8 are set .
the channels are also combined assuming the same branching fraction for both channels .
the limits are determined by examining the transverse mass ( @xmath9 ) distribution for resonant states above @xmath10 gev , where @xmath9 is constructed from events with a single high transverse momentum ( @xmath11 ) lepton plus large missing transverse energy ( @xmath12 ) , @xmath13 here @xmath14 is the angle between the lepton @xmath11 and @xmath12 vectors .- axis and the @xmath15
@xmath16 plane is transverse to the beam direction .
the positive @xmath15-axis is defined as pointing from the interaction point to the center of the lhc ring and the positive @xmath16-axis is defined as pointing upwards .
the azimuthal angle @xmath17 is measured around the beam axis and the polar angle @xmath18 is the angle from the beam axis .
the pseudorapidity is defined as @xmath19 .
the distance @xmath20 in @xmath21
@xmath17 space is defined as @xmath22 . ]
a bayesian method is used to search for an excess of the @xmath9 signal above the expected backgrounds . when no excess is observed ( above @xmath23 ) a limit on the cross section ( @xmath24 significance ) times branching ratio ( @xmath25 ) is set where the ssm @xmath0 boson is used as a benchmark for setting limits .
these proceedings begin with a brief description of the atlas detector in section [ sec : detector ] , followed by an explanation of the event selection in section [ sec : evtsel ] .
the explanation of the modeling of the backgrounds is done in section [ sec : bknds ] .
section [ sec : results ] presents the final results and these proceedings are concluded in section [ sec : conclusion ] .
the atlas detector is a general purpose particle detector that is characterized by three different components , the inner - detector ( i d ) , the calorimeter , and the muon system ( ms ) .
the innermost system is the inner - detector which consists of a silicon pixel detector and a silicon strip detector covering @xmath26 .
a transition radiation tracker is the outer tracker of the i d and covers @xmath27 .
the i d is immersed in a homogenous @xmath28 t field produced from a superconducting solenoid .
the i d allows reconstruction of the vertices and tracks of charged particles originating from the interaction region .
a finely segmented , hermetic calorimeter system surrounds the i d and covers @xmath29 as well as provides three - dimensional reconstruction of particle showers . the electromagnetic ( em ) calorimeter uses liquid argon , while the outer calorimeter , which primarily detects hadronic showers , uses scintillating tiles in the barrel ( @xmath30 ) and liquid argon for higher @xmath31 . the energy resolution for an electron approaches @xmath32 for electrons above @xmath33 gev in @xmath34 .
the ms , situated outside the calorimeters , is immersed in a @xmath35 t toroidal field that is produced by superconducting air - core toroids .
the integral magnetic field for the ms is about @xmath36 tm on average .
the curvature of the muons as they pass through the magnetic field is measured with three layers of precision drift - tube chambers in the region @xmath27 and one layer of cathode - strip chambers followed by two layers of drift - tube chambers for @xmath37 .
resistive - plate and thin - gap chambers provide the triggering for muons and measure the @xmath17 coordinate .
the muon system is able to achieve a momentum resolution of @xmath38 at @xmath4 tev .
the event selection for this analysis first requires that all events were recorded while the detector was operating in an optimal fashion . a well reconstructed primary vertex , with at least three reconstructed tracks associated with the vertex is required to be present in the event ( all objects selected in the event are then required to be associated with this primary vertex ) .
the primary vertex is also required to be reconstructed within @xmath39 mm , along the @xmath40-axis , of the center of the interaction region .
events are then examined to determine if any jets have abnormal shapes .
if an event has a poorly reconstructed jet , then it is thrown out as this type of event could bias the @xmath12 measurement .
only events that fire specific single - lepton triggers are included in the search .
for the electron channel the lowest unprescaled lepton trigger for the dataset has a @xmath41 gev @xmath34 threshold .
the trigger requires an em cluster in the em calorimeter that has an energy deposition profile consistent with that of an electron .
the trigger also requires that the em cluster be matched to a track in the inner detector .
the muon trigger has to be fired by the rpc or the tgc systems , and to pass a @xmath11 threshold of @xmath42 gev .
then the trigger object has to match an inner detector track . in the electron channel
, the reconstructed electron energy must be larger than @xmath43 gev and the electron must have an @xmath44 or @xmath45 . the offline energy cluster must have an energy deposit that is consistent with that of an electron and is matched to an offline track . the offline track is required to have a transverse impact parameter within @xmath46 cm of the primary vertex and a longitudinal distance from the primary vertex within @xmath47 cm .
a hit from the first layer of the pixel detector is required to be associated with the electron track .
finally the energy isolation of the electron , @xmath48 is required to be less than @xmath49 gev .
events with an additional electron are vetoed .
the muon channel makes the following offline requirements .
the offline muon track is first matched to the trigger object .
a @xmath11 cut of @xmath43 gev is then required , where the @xmath11 is the combined transverse momentum in the ms and the i d .
the @xmath31 of the muon must be less than @xmath47 or @xmath50 .
the combined track of the muon is required to have a transverse impact parameter less than @xmath46 cm from the primary vertex and a longitudinal distance from the primary vertex less than @xmath47 cm .
the muon track in the ms is required to register hits in all three stations of the ms , ensuring the best possible momentum measurement . as an additional quality cut , the significance of the difference of the curvature measurements between the i d and the ms is required to be less than @xmath51 .
this requirement ensures that no poorly measured muons pass into the final sample . the track isolation , @xmath52 , is to be less than @xmath53 of the muon s @xmath11 .
events with additional muons are also vetoed .
finally both channels are required to have @xmath12 greater than @xmath43 gev . in order to further reduce
the amount of background from light flavor or heavy flavor decays in the electron channel , the @xmath12 is required to be larger than @xmath54 of the electron @xmath34 .
the final distributions of the lepton @xmath11 and @xmath12 are shown in figs .
[ fig : mu - distributions ] and [ fig : el - distributions ] .
these figures also show several mass point distributions of ssm @xmath0 boson mc signal .
the ssm @xmath0 boson production mc has been generated using pythia and corrected for nnlo qcd corrections .
( left ) and @xmath12(right ) spectra of the muon channel after event selection .
the points represent data and the filled histograms show the stacked backgrounds .
open histograms are @xmath55 signals added to the background with masses in gev indicated in parentheses in the legend .
the qcd backgrounds estimated from data are also shown .
the signal and other background samples are normalized using the integrated luminosity of the data and the nnlo ( approximate - nnlo for @xmath56 ) cross sections listed in table [ tab : xsec ] [ fig : mu - distributions].,title="fig:",width=302](left ) and @xmath12(right ) spectra of the muon channel after event selection .
the points represent data and the filled histograms show the stacked backgrounds .
open histograms are @xmath55 signals added to the background with masses in gev indicated in parentheses in the legend .
the qcd backgrounds estimated from data are also shown .
the signal and other background samples are normalized using the integrated luminosity of the data and the nnlo ( approximate - nnlo for @xmath56 ) cross sections listed in table [ tab : xsec ] [ fig : mu - distributions].,title="fig:",width=302 ] . calculated values of @xmath6 for @xmath57 and the leading backgrounds .
the value for @xmath58 includes all final states with at least one lepton ( @xmath59 , @xmath60 or @xmath61 ) .
the other processes are exclusive and are used for both @xmath62 and @xmath63 .
all calculations are nnlo except @xmath56 which is approximate - nnlo .
[ cols="<,^ , < " , ] the ratio on the lower limit @xmath64 for @xmath0 boson production to the ssm production cross section versus @xmath65 is shown in fig .
[ fig : atlas2011 - 2-emu ] . for @xmath8 ( left ) and @xmath7 ( right ) .
the nnlo cross section for a ssm @xmath0 boson and its uncertainty are also shown.[fig : indiv - limits],title="fig:",width=302 ] for @xmath8 ( left ) and @xmath7 ( right ) . the nnlo cross section for a ssm @xmath0 boson and its uncertainty are also shown.[fig : indiv - limits],title="fig:",width=302 ] for the combined @xmath57 ( @xmath66 channels ) assuming the same branching fraction for both channels . the nnlo cross section and its uncertainty
are also shown .
on the right are the normalized cross section limits ( @xmath67 ) for @xmath57 as a function of mass for this measurement and from cdf @xcite , as well as the previous atlas search @xcite .
the cross section calculations assume the @xmath0 boson has the same couplings as the sm @xmath5 boson .
the region above each curve is excluded at @xmath68 cl.[fig : atlas2011 - 2-emu],title="fig:",width=302 ] for the combined @xmath57 ( @xmath66 channels ) assuming the same branching fraction for both channels .
the nnlo cross section and its uncertainty are also shown .
on the right are the normalized cross section limits ( @xmath67 ) for @xmath57 as a function of mass for this measurement and from cdf @xcite , as well as the previous atlas search @xcite .
the cross section calculations assume the @xmath0 boson has the same couplings as the sm @xmath5 boson .
the region above each curve is excluded at @xmath68 cl.[fig : atlas2011 - 2-emu],title="fig:",width=302 ]
using @xmath1 fb@xmath2 of proton - proton collisions collected at the lhc , atlas has searched for a massive boson decaying to a lepton plus missing transverse energy with @xmath69 gev .
no significant excess has been observed , so lower @xmath68 cl limits have been evaluated on the production cross section of a ssm @xmath0 boson .
this analysis has been submitted to physics review letters b @xcite .
we acknowledge the support of anpcyt , argentina ; yerphi , armenia ; arc , australia ; bmwf , austria ; anas , azerbaijan ; sstc , belarus ; cnpq and fapesp , brazil ; nserc , nrc and cfi , canada ; cern ; conicyt , chile ; cas , most and nsfc , china ; colciencias , colombia ; msmt cr , mpo cr and vsc cr , czech republic ; dnrf , dnsrc and lundbeck foundation , denmark ; artemis , european union ; in2p3-cnrs , cea - dsm / irfu , france ; gnas , georgia ; bmbf , dfg , hgf , mpg and avh foundation , germany ; gsrt , greece ; isf , minerva , gif , dip and benoziyo center , israel ; infn , italy ; mext and jsps , japan ; cnrst , morocco ; fom and nwo , netherlands ; rcn , norway ; mnisw , poland ; grices and fct , portugal ; merys ( mects ) , romania ; mes of russia and rosatom , russian federation ; jinr ; mstd , serbia ; mssr , slovakia ; arrs and mvzt , slovenia ; dst / nrf , south africa ; micinn , spain ; src and wallenberg foundation , sweden ; ser , snsf and cantons of bern and geneva , switzerland ; nsc , taiwan ; taek , turkey ; stfc , the royal society and leverhulme trust , united kingdom ; doe and nsf , united states of america . the crucial computing support from all wlcg partners is acknowledged gratefully , in particular from cern and the atlas tier-1 facilities at triumf ( canada ) , ndgf ( denmark , norway , sweden ) , cc - in2p3 ( france ) , kit / gridka ( germany ) , infn - cnaf ( italy ) , nl - t1 ( netherlands ) , pic ( spain ) , asgc ( taiwan ) , ral ( uk ) and bnl ( usa ) and in the tier-2 facilities worldwide
. 9 atlas collaboration , cern - ph - ep-2011 - 121 ( 2011a ) . k. nakamura et al .
( particle data group ) , j. phys .
g * 37 * , 075021 ( 2010 ) .
g. altarelli , b. mele , and m. ruiz - altaba , z. phys .
c * 45 * , 109 ( 1989 ) .
atlas collaboration , phys .
b * 701 * , 50 ( 2011b ) .
s. moch and p. uwer , phys .
d * 78 * , 034003 ( 2008 ) , 0804.1476 .
u. langenfeld , s. moch , and p. uwer , arxiv : hep - ph/0907.2527 ( 2009 ) , 0907.2527 .
m. aliev et al . , comput .
* 182 * , 1034 ( 2010 ) , 1007.1327 .
atlas collaboration , phys .
d * 83 * , 052005 ( 2011c ) . t. aaltonen et al .
, cdf collaboration , phys .
d * 83 * , 031102 ( 2011 ) . | the atlas detector has been used to search for high - mass states decaying into a single high momentum lepton and missing transverse energy , such as new heavy charged gauge bosons .
the latest search results for a @xmath0 boson decaying to lepton plus neutrino in @xmath1 fb@xmath2 of proton - proton collisions at a center of mass energy of @xmath3 tev produced at the large hadron collider are presented . |
the weakly connected oscillator theory is often used to show conditions under which the oscillators are synchronized @xcite . in this theory ,
small perturbations do not influence the amplitude of the oscillator , however they have a significant effect on its phase .
this allows one for a drastic reduction in the description of the oscillator : instead of operating with the original , possibly high- dimensional set of equations , only one phase variable is used for each oscillator . using the phase description alone , it is possible to find the conditions when oscillators are synchronized , provided the phase reduction is accurate enough , what is the case if the coupling is weak .
one of the assumptions used in this approach is the principle of superposition , which states that the effect of several small perturbations on the period of the oscillation could be considered independently , and then summed . in this paper
, we examined the phase dynamics beyond the superposition principle .
more precisely , we consider the effect of two relatively small perturbations on the phase for various types of oscillators .
our main tool in the description of the phase dynamics is the phase response curve ( prc , @xmath0 ) which is widely used in both theoretical and experimental studies , especially in the field of neuroscience @xcite .
the prc measures the shift of the phase of the oscillator due to an external pulse , in dependence on the phase at which the pulse is applied . or , in terms of the oscillator period
, the prc measures the local change in the period of the oscillator due to a pulse perturbation at various time points within the period .
the oscillation can either advance or delay based on the sign of the prc . to determine the prc ,
one often performs an experiment just according to the definition of the prc ; this can be accomplished for individual biological neurons @xcite and for complex oscillating systems like those responsible for circadian rhythms in the brain @xcite , see refs .
@xcite for other biophysical examples .
furthermore , the prc concept can be applied not only to individual oscillators , but to collective modes as well @xcite .
the shape of the prc curve is shown to be critical for the synchronization properties of the oscillator networks @xcite . however , in the synchronization problem , an oscillator is subject not just to one external pulse , but to a series of pulses from the external force or another oscillator ( or several other oscillators if more then two oscillators are coupled ) .
thus , in order to apply the prc concept in such a situation , one has to know how the oscillator responds to a series of pulses .
if the superposition principle holds , then the sum of two small perturbations will independently influence the period of the oscillator according to the prcs for single inputs , and thus can be linearly added to predict the overall phase shift .
however , generally one expects deviation from this simple superposition if the perturbations are not small . in this paper
we systematically consider the effect of two pulses on the oscillator s phase , and characterize the deviations from the pure superposition as nonlinear effects .
we illustrate these effects also for several realistic models of neuron dynamics .
we start with the simplest case where the oscillator is described just by one variable , the phase @xmath1 ( to be assumed @xmath2-periodic ) that grows uniformly in time @xmath3 suppose that the action of a forcing pulse with strength @xmath4 is described by the standard prc @xmath5 ( here dependence of @xmath6 on @xmath4 accounts for nonlinear terms , so that @xmath7 is the linear prc ) .
consider now the action of two pulses , at times @xmath8 and @xmath9 , having strengths @xmath10 and @xmath11 , respectively .
just after the first pulse @xmath12 just prior to the second pulse the phase is @xmath13 and after the second pulse @xmath14 thus , the overall effect of two pulses @xmath15 \label{eq : genprc0}\ ] ] can be simply calculated via the superposition of two single - pulse prc functions @xmath5 . in the linear approximation , where @xmath16 one obtains just the sum @xmath17 now we consider a general situation where periodic oscillations are described by a limit cycle @xmath18 in a generally @xmath19-dimensional phase space .
the crucial notion simplifying the consideration , is that of isochrons @xcite , which are submanifolds of codimension one foliating the phase space and having the same phase as the corresponding points on the limit cycle .
this allows one to represent the phase space as @xmath20 where @xmath21 is an @xmath22-dimensional `` amplitude '' , and the phase obeys the same equation ( [ eq : phdyn ] ) . without loss of generality , to simplify notations
, we can assume that on the limit cycle the amplitude vanishes @xmath23 . in terms of the phase and the amplitude , a pulse that kicks the system resets the state
@xmath20 as @xmath24 where again @xmath25 and @xmath26 correspond to a linear approximation .
the usual prc is defined for the initial state on the limit cycle ( @xmath23 ) , so @xmath27 .
we now apply as above two pulses at times @xmath8 and @xmath9 , assuming that the system is initially on the limit cycle . then after the first pulse @xmath28 just prior to the second pulse @xmath29 where @xmath30 is the operator describing the evolution of the amplitudes . after the second pulse
the new phase is @xmath31 and the overall phase shift due to two pulses ( two - pulse prc ) is @xmath32 comparing this with expression ( [ eq : genprc0 ] ) we see that now the effect is not a superposition of two prcs , but contains the amplitude - dependent phase reset function @xmath33 .
the difference between expressions ( [ eq : genprc0 ] ) and ( [ eq : genprc1 ] ) gives the nontrivial effect of multiple pulses on the phase of the oscillator as the nonlinear correction term @xmath34 : @xmath35 from this expression one can see that the result essentially depends on the action of the amplitude evolution operator @xmath36 : if @xmath37 , then the correction ( [ eq : genprc2 ] ) vanishes .
thus , the nontrivial effect of the two - pulse excitation of an oscillator depends crucially on the relation between the time interval between the pulses @xmath38 and the relaxation time of the amplitude @xmath39 ( characteristic time scale of the amplitude evolution operator @xmath40 ) , it is mostly pronounced if @xmath41 .
if the amplitude is multidimensional , @xmath39 is the time of most slow decay . in the leading order in the powers of @xmath42
, we can represent the nonlinear correction as @xmath43 where @xmath44 is the linearized evolution operator for the amplitudes , which describes their relaxation to zero @xmath45 $ ] .
this can be generalized to @xmath46 pulses with amplitudes ( @xmath47 ) and different time shifts between them @xmath48 .
then the leading terms will be quadratic ones ( @xmath49 ) , while also higher - order corrections ( e.g. @xmath50 ) will appear . mostly important will be nonlinear terms including neighboring pulses , because , as argued above , the effect decreases with the time interval between the pulses .
another straightforward generalization is the case where two pulses are different and are described by functions @xmath51 .
the stuart - landau oscillator is a two - dimensional model described in polar coordinates as @xmath52 here the frequency of the limit cycle , which is a circle with radius @xmath53 , is normalized to one , parameter @xmath54 describes nonisochronicity of oscillations , while @xmath55 is the relaxation rate of the amplitude .
the phase @xmath1 defined in the whole plane ( except for the origin ) is @xmath56 evolution of the variables @xmath57 can be explicitly solved as @xmath58^{-1/2}\\ \theta(t_0)+t - t_0-\frac{\alpha}{2\mu}\ln(r(t_0)^2 + ( 1-r(t_0)^2)e^{-2\mu(t - t_0)})\end{pmatrix}\;,\ ] ] which defines the operator @xmath40 .
we assume that the pulse is acting in direction @xmath59 , i.e. at the pulse @xmath60 this fully describes the system , and one can find expressions for the prcs @xmath61 and @xmath62 , see appendix [ ap : prcsl ] . according to these formulas we calculated the nonlinear correction term @xmath34 and plot it in fig .
[ fig : slo_1 ] . here
we take @xmath63 and present results for different @xmath55 . as expected , the mostly pronounced effect is for small @xmath55 .
for this equation it is possible to obtain the leading term in order @xmath64 in the expansion of the nonlinear correction term in the pulse strengths analytically ( see appendix [ ap : prcsl ] ) : @xmath65 this expression fits numerics very good for @xmath66 .
vs the phase @xmath67 and the time shift between pulses @xmath38 ( normalized by the cycle period ) , for @xmath68 , and three different values of @xmath55 ( @xmath69 in ( a ) , @xmath70 in ( b ) , and @xmath71 in ( c ) ) . for large @xmath55
the effect is pronounced for very small time intervals between two pulses only . ]
our second example is a modification of the stuart - landau oscillator proposed in @xcite : @xmath72 here large values of parameter @xmath73 produce highly nonuniform growth of angle variable @xmath74 , so that the relation between @xmath1 and @xmath74 is strongly nonlinear . as a result , the isochrons crowd at the region around @xmath75 where the evolution of @xmath74 is slow .
also the nonlinear correction term becomes very large at this region , as illustrated in fig .
[ fig : slo_2 ] but for the modified stuart - landau oscillator eq .
( [ eq : mslo ] ) .
parameter values : ( a ) @xmath76 ; ( b ) : @xmath77 ; ( c ) : @xmath78 . ]
is the original period when no input occurred , @xmath79 is the period when there was an input at time @xmath80 with input phase @xmath81 . c. prc @xmath82 for single input ( black cross ) along with its fit ( solid red line ( @xmath83 ) ) .
d. prc for 2 inputs with isi=10ms .
gray points indicate the change in period against the first input ; prc with respect to the second input is plotted as black cross .
the prediction from superposition principle is plotted as solid blue line ( @xmath84 ) . ]
the prc is commonly used to describe neuron models . in this context
, the prc can characterize the properties of neurons , especially their synchronizability .
in many systems a neuron receives inputs from many other neurons , therefore , it is critical to understand how multiple pulses affect the prc response of neurons .
below we test four different neuron models for the dual pulse effect .
although generally the theory presented in section [ sec : gc ] is applicable to spiking neurons as well , practically one does not follow the continuous phase of the oscillations , but focuses on the spiking events ( these events are readily available in experiment , too ) . therefore , for spiking neurons , the prc and the non - linear correction term have to be measured in terms of the spike times as oppose to phase shifts at arbitrary points as done in previous section .
it is convenient to normalize the correction term should by the peak to trough value of the prc for single input as shown below .
we first illustrate these definitions in fig .
[ prc_measure ] , using the quadratic integrate - and - fire model @xcite .
( note that traditionally in this context the phase @xmath85 and the prc @xmath86 are normalized by one and not by @xmath2 ) .
this one- dimensional model corresponds to the pure phase dynamics as in section [ sec : ppd ] , so it does not demonstrate nonlinear effects of deviations from the superposition .
in general , the models of neurons are classified based on their prc curves as type i and ii .
the type i prc , has only phase advance in response to perturbation , while type ii prc includes both phase delay and advance @xcite .
we further considered both type i and ii neuron models in this study .
we tested wang - buzski model ( type i , based on @xcite ) , the original hodgkin - huxley model ( type ii , based on @xcite ) , and a modified hodgkin - huxley model ( type i , based on @xcite ) .
the equations for all the models are given in the appendix [ app : nm ] .
all of these models are three - dimensional which is required for any deviation from linear superposition . to characterize the deviation from the superposition of two input pulses , we use the following quantity @xmath87 where , @xmath88 is the measured prc for the second pulse , @xmath83 and @xmath89 are the expected prcs of single pulses , and @xmath90 is the amplitude of the single - pulse prc ( this normalization provides a better visualization of the deviation from the superposition of single - pulse prcs ) .
we present the results for the three neuron models in figs .
[ prc_wb_model]-[prc_mhh_model ] , where the main dependence of of @xmath91 on the first input phase and on the interspike interval ( isi ) between two applied pulses @xmath38 is depicted in panel c by a color coding . ) .
b. prc for 2 inputs with isi=10ms .
gray points indicate the prc with respect to the first input and prc with respect to second input is plotted as black cross .
the prediction from superposition principle is plotted as solid blue line . c. the deviation from linear superposition ( @xmath92 ) for different isis . ]
we observed that some neuron models show pronounced nonlinear effects , as the two - pulse response deviates from the expected prc based on superposition principle , while for other models the linear superposition was able to predict the prc for two - pulses accurately .
the wang - buzski model ( [ prc_wb_model ] ) appears to be of the latter type , while both the original hodgkin - huxley system ( [ prc_hh_model ] ) and its modified version ( [ prc_mhh_model ] ) show nonlinear effects in two - pulse prc .
the wang - buzski and modified hodgkin - huxley model had type i prc , the original hodgkin - huxley model had type ii prc , suggesting that there is no relationship between the type of prc and the origin of this deviation .
further , the bifurcation type for spiking from the resting state also did not determine the existence of error , since the wang - buzski model and the modified hodgkin - huxley model possess a saddle - node bifurcation , while the original hodgkin - huxley model demonstrates an andronov - hopf bifurcation .
there were two main findings from the analytical results of section [ sec : pd ] that can be tested with the neuronal models presented in this section .
the first result was that the non - linear correction term was proportional to the square of the perturbation [ eq : sl_error ] .
we tested this in the modified hodgkin - huxley model and observed similar qualitative results ( fig . [ quadratic - error ] ) .
second , the decay time constant of the amplitude term in the oscillator due to perturbation was inversely proportional to the non - linear correction term ( [ eq : genprc2 ] ) . thus
, a slower dynamics of the neuron will lead to lower error .
the modified neuron model also showed reduced error when the dynamics was slowed ( fig [ prc_mhh_model ] ) .
thus , the main features of general limit cycle oscillators , and in particular of stuart - landau oscillator , could be qualitatively extended to some of the neuron models . and
@xmath94 / period .
the division by period was used to compare with the landau - stewart oscillator simulations and does not change the relationship between @xmath95 and @xmath94 , since period was same for all @xmath95 . ]
in this paper we have developed a theory of a response of an autonomous oscillator to two - pulse perturbation .
we found that the action of two pulses generally deviates from the superposition of two one - pulse responses , and this nonlinear effect , which is proportional to the product of the perturbations amplitudes , significantly depends on the relation of the interval between the pulses and the relaxation time of the oscillator .
for fast relaxation and large time interval between the pulses , the nonlinear effect vanishes .
we have demonstrated this property for several models : the standard stuart - landau oscillator , the modified version for this oscillator with a highly non - uniform motion over the cycle , and for three neuron models , including the classical hodgkin - huxley system .
we stress here that in our study , the term `` nonlinearity '' of the prc has been used in two contexts . for a one - pulse prc ,
nonlinearity means , that the phase response can not be represented as an amplitude of the pulse multiplied by a function of the phase ; in particular , the form of the curve may depend on the pulse amplitude . for two pulses , we use term `` nonlinearity '' to describe a deviation from the superposition principle ,
this effect in the leading order is proportional to the product of the pulses amplitudes .
we have shown that the nonlinearity of the single - pulse prc does not necessarily lead to the the nonlinearity for a two - pulse excitation : the purely phase model of section [ sec : ppd ] is a good illustration of this . both nonlinear effects may distort a simple picture of the neuron s dynamics under external forcing . in particular , in @xcite it was suggested that relatively weak noisy current to neuron can be applied to obtain the prc , by solving the equation that relates the infinitesimal ( linear ) prc to the external voltage through optimization methods .
however , this method does not account for the non - linear correction which we showed in this study . in some neurons
, we showed that the non - linear effect could have significant effect on the multi - pulse prc compared to the single pulse prc , and thus the continuous perturbation method may produce erroneous results .
thus , further studies are required to validate the method proposed in @xcite .
we see two main application fields of our approach .
first , it can be used for a diagnostics of oscillators . while the usual prc allows to characterize sensitivity of the phase to an external action , the nonlinear terms in the two - pulse response allow to characterize relaxation processes .
in particular , one - pulse prcs of the wang - buzski model and of the modified hodgkin - huxley model are very similar ( cf . figs .
[ prc_wb_model]a and [ prc_mhh_model]a ) , but their two - pulse prcs are completely different ; this may be useful for designing models to fit experimental data .
two - pulse prc can be estimated from experimental data and this information can be used to design optimally a model that provides a best description of the data .
the second field of application is in the incorporating these effects in the synchronization theory of pulse coupled oscillators .
indeed , in a network an oscillator usually experiences inputs from many other units , and in the absence of synchrony the time intervals between the incoming pulses can be rather small . in this case the nonlinear `` interference '' of the actions is mostly pronounced , and may contribute significantly to the synchronization properties .
20ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ]
+ 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty @noop _
( , , ) @noop _ _ ( , , ) @noop _ _ ( , , ) @noop * * ( ) @noop _ _ , springer , springer series in computational neuroscience , vol . 6 ( , , ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop _ _ ( , , ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( )
we briefly outline the majors steps in obtaining the approximate non - linear correction term for stuart - landau oscillator using an expansion in @xmath4 .
several intermediate steps are not shown to make presentation brief .
section [ sec : slo ] provides the phase equation and the evolution operator for stuart - landau oscillator . at @xmath96 ,
let @xmath97 , @xmath98 .
after @xmath99 pulse at @xmath96 , @xmath100 , @xmath101 , which upon expansion gives , @xmath102 and @xmath103 .
from here we ignore all , @xmath104 , cubic and higher order terms .
+ + the phase shift due to @xmath99 pulse is ,
@xmath105 @xmath107 + \epsilon_0 ^ 2 [ \sin\theta_0 \cos\theta_0 - \frac{\alpha}{2\mu } ( 1 - e^{-2\mu\tau } - 2 \cos^2\theta_0 ( 1 - e^{-2\mu\tau})^2 ) ] \\ & \equiv \tau + \theta_0 - \epsilon_0 e + \epsilon_0 ^ 2 f\\ r(\tau)&= \big [ \frac{\tilde r^2}{\tilde r^2+(1-\tilde r^2)e^{-2\mu\tau } } \big]^{1/2 } \approx \big [ \frac{1 + 2\epsilon_0\cos\theta_0+\epsilon_0 ^ 2}{1 + 2\epsilon_0\cos\theta_0+\epsilon_0 ^ 2+(-2\epsilon_0\cos\theta_0-\epsilon_0 ^ 2 ) e^{-2\mu\tau } } \big]^{1/2}\\ r(\tau)^2 & \approx \frac{1 + 2\epsilon_0\cos\theta_0+\epsilon_0 ^ 2}{1 + 2\epsilon_0\cos\theta_0 ( 1 - e^{-2\mu\tau } ) + \epsilon_0 ^ 2 ( 1 - e^{-2\mu\tau } ) } \\ & \approx 1 + 2 \epsilon_0\cos\theta_0 e^{-2\mu\tau } - 4 \epsilon_0 ^ 2 \cos^2\theta_0 ( 1 - e^{-2\mu\tau } ) + \epsilon_0 ^ 2 e^{-2\mu\tau } + 4 \epsilon_0 ^ 2 \cos^2\theta_0 ( 1 - e^{-2\mu\tau})^2 \\ & \approx 1 + 2 \epsilon_0\cos\theta_0 e^{-2\mu\tau } + \epsilon_0 ^ 2 [ - 4 \cos^2\theta_0 ( 1 - e^{-2\mu\tau})e^{-2\mu\tau } + e^{-2\mu\tau } ] \\ & \equiv 1 + \epsilon_0 c + \epsilon_0 ^ 2 d\end{aligned}\ ] ] after the second pulse ( @xmath108 ) at time @xmath109 , @xmath110 and @xmath111 .
so the final phase shift is , @xmath112 , which needs to be compared to prediction from linear superposition which is given as @xmath113 .
we first expand @xmath114 , @xmath115 , @xmath116 , since it is used in the comparison as , @xmath117 + { \varepsilon}_1 ^ 2 \sin(\theta_0 + \tau ) \cos(\theta_0 + \tau)\end{aligned}\ ] ] @xmath118 \\ & + \frac{1}{2 } { \varepsilon}_1 ^ 2 [ \sin2(\theta_0 + \delta\varphi_0 + \tau ) + \frac{\alpha}{\mu } \cos2(\theta_0 + \delta\varphi_0 + \tau)]\\ \delta \varphi_{0,1 } & = \hat\theta(\tau ) - \theta_0 - \tau - \frac{\alpha}{2\mu}\ln \hat r^2(\tau)\\ & = \epsilon_0 e - { \varepsilon}_1 \sin(\tau + \theta_0 ) + \epsilon_0 ^ 2 f + { \varepsilon}_0{\varepsilon}_1[-\cos(\theta_0 + \tau ) e + \frac{1}{2 } \sin(\theta_0 + \tau ) c ] \\ & + { \varepsilon}_1 ^ 2 \sin(\theta_0 + \tau ) \cos(\theta_0 + \tau ) \\ & - \frac{\alpha}{2\mu}\ln [ 1 + { \varepsilon}_0 c + { \varepsilon}_0 ^ 2 d + 2{\varepsilon}_1(1+\frac{1}{2 } { \varepsilon}_0 c ) ( \cos(\theta_0 + \tau ) + { \varepsilon}_0 e \sin(\tau + \theta_0 ) ) + \epsilon_1 ^ 2]\end{aligned}\ ] ] finally , the difference between the prediction from superposition and actual phase reset ( after substitutions ) is given as , @xmath119 - { \varepsilon}_1 \sin(\tau + \theta_0 ) + \epsilon_0 ^ 2 [ \sin\theta_0 \cos\theta_0 - \frac{\alpha}{2\mu } ( 1 - e^{-2\mu\tau } - 2 \cos^2\theta_0 ( 1 - e^{-2\mu\tau})^2 ) ] \\ & + { \varepsilon}_0{\varepsilon}_1\big[-\cos(\theta_0 + \tau ) [ \sin\theta_0 + \frac{\alpha}{\mu } \cos ( 1 - e^{-2\mu\tau } ) ] + \frac{1}{2 } \sin(\theta_0 + \tau ) 2\cos\theta_0 e^{-2\mu\tau } \big ] + { \varepsilon}_1 ^ 2 \sin(\theta_0 + \tau ) \cos(\theta_0 + \tau ) \\ & - \frac{\alpha}{2\mu}\ln \big [ 1 + { \varepsilon}_0 2\cos\theta_0 e^{-2\mu\tau } + { \varepsilon}_0 ^ 2 [ - 4 \cos^2\theta_0 ( 1 - e^{-2\mu\tau})e^{-2\mu\tau } + e^{-2\mu\tau } ] \\ & \qquad\qquad + 2{\varepsilon}_1(1+\frac{1}{2 } { \varepsilon}_0 2\cos\theta_0 e^{-2\mu\tau } ) ( \cos(\theta_0 + \tau ) + { \varepsilon}_0 [ \sin\theta_0 + \frac{\alpha}{\mu } \cos ( 1 - e^{-2\mu\tau } ) ] \sin(\tau + \theta_0 ) ) + \epsilon_1 ^ 2 \big]\\ & -\big [ -\epsilon_0 ( \sin\theta_0 + \frac{\alpha}{\mu } \cos\theta_0 ) + \frac{1}{2 } \epsilon_0 ^ 2(\sin 2\theta_0 + \frac{\alpha}{\mu } \cos 2\theta_0 ) + { \varepsilon}_1[\sin(\theta_0 + \delta\varphi_0 + \tau ) \\
& + \frac{\alpha}{\mu } \cos(\theta_0+\delta\varphi_0+\tau ) ] + \frac{1}{2 } { \varepsilon}_1 ^ 2 [ \sin2(\theta_0 + \delta\varphi_0 + \tau ) + \frac{\alpha}{\mu } \cos2(\theta_0 + \delta\varphi_0 + \tau ) ] \big]\\\end{aligned}\ ] ] which upon several steps of algebraic reductions and ignoring cubic or higher @xmath120 terms gives , @xmath121 | in this paper , we introduce and systematically study , in terms of phase response curves ( prc ) , the effect of a dual pulse excitation on the dynamics of an autonomous oscillator . specifically , we test the deviations from a linear summation of phase advances from two small perturbations .
we derive analytically a correction term , which generally appears for oscillators whose intrinsic dimensionality is greater than one .
we demonstrate this effect in case of the stuart - landau model , and also in various higher dimensional neuronal model . the non - linear correction is found to be proportional to the square of the perturbation .
this deviation from the superposition principle needs to be taken into account in studies of networks of pulse - coupled oscillators .
further , this deviation could be used for verification of oscillator models via a dual pulse excitation . |
in present paper we study the local time with respect to spatial variable of solution to the following white noise driven heat equation @xmath0 here @xmath1 is a 1-dimensional gaussian space - time white noise in @xmath2 since @xmath1 is a random generalized function on @xmath3 the equation is understood in the sense of generalized functions .
@xmath1 can be considered as a centered gaussian random measure with independent values on the disjoint subsets of @xmath4 some times it is called by brownian sheet @xcite .
it is known @xcite that solution to is given by @xmath5 where the stochastic integral with respect to brownian sheet @xmath1 is defined as an ito integral @xcite by means of the isometry @xmath6 here @xmath7 the interest in the equation arises since it describes the random motion of the string .
for example , m.hairer in @xcite considered the following model .
take @xmath8 particles with positions @xmath9 immersed in a fluid and assume that nearest - neighbours are connected by harmonic springs .
if the external force @xmath10 is acting on the particles , then the equations of motion in the overdamped regime have the following representation @xmath11 where @xmath12 is a constant factor characteristic of the spring : its stiffness .
this constant appears from hooke s law .
the described model can be considered as a primitive model for a polymer chain consisting @xmath8 monomers and without self - interaction .
it does not take into account the effect of the molecules of water that would randomly `` kick '' the particles .
assume that these kicks occur randomly and independently at high rate .
this effect can be modelled by increments of independent wiener processes @xmath13 as follows @xmath14 formally taking the continuum limit ( with the scalings @xmath15 and @xmath16 ) as @xmath17 one can see that this system is well - described by the solution to a stochastic partial differential equation @xmath18 more general @xmath19 white noise driven heat equation was considered by t.funaki in @xcite . for given two functions @xmath20 and @xmath21 let @xmath22 be a diffusion process on @xmath23 determined by the stochastic differential equation @xmath24 where @xmath1 is a @xmath25-dimensional wiener process .
funaki introduced the following @xmath26 , { \mathbb r}^d)$]-valued stochastic differential equation @xmath27 @xmath28.$ ] here @xmath1 is a @xmath25-dimensional gaussian white noise with two parameters in @xmath29\times[0;1]).$ ] the equation describes a string which moves in @xmath23 being interfered by the process @xmath30 the idea behind the derivation of the equation is to take the scaling limit of a sequence of polygonal approximations . more precisely , funaki approximates the string by a polygon and sets on each corner a particle which moves governed by a stochastic differential equation with a suitable scaling . in each step of the approximation ,
the interaction between neighboring two particles was always taken into account .
m.kardar , g.parisi and y .- c.zhang in @xcite proposed a model for the time evolution of the profile of a growing interface .
the interface profile is described by a height @xmath31 for convenience overhangs are ignored .
the langevin equation for a local growth of the profile is given by @xmath32 the mentioned equations , , can be considered as the mathematical models which describe the motion of curve at the random medium . since all these equations constructed by means of white noise ,
then the corresponding solutions would be nonsmooth random processes .
as it was discussed in works of authors @xcite geometric characteristics of random processes are local times and self - intersection local times . at present paper
we want to examine the existence of local time for the solution to the simplest of these equations . to do this
we will show that the solution to the equation belongs to the class of gaussian integrators .
this class of gaussian processes firstly was introduced by a.a.dorogovtsev in the work @xcite .
the original definition is the following .
[ defn1.1 ] @xcite a centered gaussian process
@xmath33,\ x(0)=0 $ ] is said to be an integrator if there exists the constant @xmath34 such that for an arbitrary partition @xmath35 and real numbers @xmath36 the following relation holds @xmath37 this inequality allows to define a stochastic integral for any square integrable function with respect to the integrator .
the following statement describes the structure of all gaussian integrators .
[ lem1.1]@xcite the centered gaussian process @xmath38 $ ] is the integrator iff there exist a white noise @xmath39 in the space @xmath40)$]and a continuous linear operator in the same space such that @xmath41 } , \xi),\ t\in[0;1].\ ] ] it follows from lemma [ lem1.1 ] that all properties of the gaussian integrator can be characterized in terms of the operator @xmath42 since we are going to study the local time for the solution to the equation , let us recall the general definition of local time for a 1-dimensional random process @xmath43.$ ] [ defn1.2 ] for any @xmath44 @xmath45 is said to be a local time of the process @xmath46 at the point @xmath47 whenever the limit exists .
put @xmath48 this paper is organized as follows . in the beginning we prove that solution to a stochastic heat equation at the fixed moment of time is the integrator .
then we check that for a sufficiently small interval this gaussian integrator is generated by a continuously invertible linear operator . using that
, we establish the existence of local time with respect to spatial variable on any finite interval .
in this section we will prove that the solution @xmath22 to the equation considered at the fixed moment of the time has a local time with respect to spatial variable .
our proof of existence of the local time for @xmath22 has the following outline .
first of all we prove that @xmath22 belongs to the class of gaussian integrators .
the local times of integrators were discussed in the works @xcite .
it was proved that if the integrator is generated by a continuously invertible operator , then it has a local time .
we will check that on the small intervals the process @xmath22 satisfies this condition .
then we discuss the existence of the local time for a linear combination of gaussian process and some jointly gaussian random variable and finally prove the main statement for an arbitrary finite interval .
denote by @xmath54 using the properties of stochastic integral and isometry property of fourier transform one can get the following relations @xmath55 @xmath56 @xmath57 @xmath58 @xmath59 @xmath60 \big|^2d\lambda ds=\ ] ] @xmath61 where @xmath62}(y).\ ] ] by integrating over @xmath63 one can obtain that equals @xmath64 @xmath65 which proves the theorem . since @xmath66 is an integrator ,
then there exist a continuous linear operator @xmath67 in the space @xmath68)$ ] and a white noise @xmath39 in the same space such that the process @xmath69 $ ] admits the representation @xmath70 } , \xi\big).$ ] suppose that @xmath76 let us check that there exists a constant @xmath77 such that for any @xmath78)$ ] @xmath79 really , for @xmath80 with @xmath81 $ ] @xmath82 @xmath83 note that @xmath84 @xmath85 @xmath86 applying the cauchy inequality one can obtain the following relation @xmath87 @xmath88 it follows from that @xmath89 in the @xcite we prove that the gaussian integrator generated by a continuously invertible operator has a local time at any real point which is jointly continuous in time and space variables . consequently , the gaussian integrator @xmath90 $ ] with @xmath91 has a jointly continuous local time at any real point .
now we will consider the interval of an arbitrary length .
consider the gaussian integrator @xmath92 $ ] and the partition @xmath93 such that for any @xmath94 @xmath95 every gaussian integrator @xmath96 $ ] has a jointly continuous local time at any real point .
does the same hold for the gaussian integrator @xmath97?$ ] the answer is given by the next theorem .
consider the partition @xmath93 such that for any @xmath94 @xmath95 then one can written the following relation @xmath99 denote by @xmath100 then to prove the theorem it suffices to check that @xmath101 since @xmath102 then follows from the existence of finite limit @xmath103 as @xmath104 let us check it .
one can see that @xmath105 @xmath106 @xmath107 since @xmath108 $ ] is the gaussian integrator , then there exist a continuous linear operator @xmath109 in the space @xmath110)$ ] and a white noise @xmath111 in @xmath110)$ ] such that @xmath112 } , \widetilde{\xi}).\ ] ] note that the random variable @xmath113 is jointly gaussian with the white noise @xmath114 to check this let us recall the construction of the white noise @xmath114 denote by @xmath115 the closure of the linear span of values of @xmath116 with respect to square mean norm .
@xmath115 is a separable hilbert space
. there exists the subspace @xmath117 of @xmath110)$ ] which is isomorphic to @xmath118 denote by @xmath119 where @xmath120 is isomorphism .
then @xmath110)$ ] can be represented as a direct sum @xmath121 suppose that @xmath122 is a gaussian white noise in @xmath123 which is independent of @xmath30 put @xmath124 and @xmath125 notice that @xmath126 since @xmath127 $ ] is the gaussian process , then the random variables @xmath128 and @xmath113 are jointly gaussian .
the random variables @xmath128 and @xmath129 are independent .
it implies that the random variables @xmath130 and @xmath113 are jointly gaussian .
hence , the random variable @xmath113 and @xmath111 are jointly gaussian .
to continue the proof of the theorem we need the following statement which can be easily proved .
[ lem2.2 ] let @xmath39 be a white noise in the space @xmath40).$ ] for a jointly gaussian with @xmath39 random variable @xmath131 there exist @xmath132)$ ] and the random variable @xmath133 that is independent of the white noise @xmath39 , such that @xmath131 admits the representation @xmath134 applying lemma 2.2 to the white noise @xmath111 and the random variable @xmath113 one can see that can be represented as follows @xmath135 } , \widetilde{\xi})-(h,\widetilde{\xi})-\zeta)\cdot\ ] ] @xmath136 } , \widetilde{\xi})-(h,\widetilde{\xi})-\zeta)d\vec{v},\ ] ] where @xmath137 denote by @xmath138 the projection onto linear span generated by @xmath139 then possesses the representation @xmath140 } , \widetilde{\xi } ) -(h,\widetilde{\xi})-\zeta\big)\cdot\ ] ] @xmath141 } , \widetilde{\xi } ) -(h,\widetilde{\xi } ) -\zeta)/ ( p_h\widetilde{a}{1\!\!\,{\rm i}}_{[u_k ; v_i ] } , \widetilde{\xi}),\ i=\overline{1 , 2p},\ ( h,\widetilde{\xi}),\ \zeta\big\ } d\vec{v}.\ ] ] denote by @xmath142 let @xmath143 be the gramian matrix constructed by elements @xmath144 @xmath145 put @xmath146 @xmath147},\ldots,(i - p_h)\widetilde{a}{1\!\!\,{\rm i}}_{[u_k ; v_{2p}]})+i(\varepsilon_1 , \varepsilon_2),\ ] ] @xmath148}-(h , \widetilde{\xi})-\zeta\\ \vdots\\ p_h\widetilde{a}{1\!\!\,{\rm i}}_{[u_k ; v_{2p}]}-(h , \widetilde{\xi})-\zeta \end{pmatrix}\ ] ] then equals @xmath149 it follows from that to end the proof of the theorem one must check that the integral @xmath150 } , \ldots , ( i - p_h)\widetilde{a}{1\!\!\,{\rm i}}_{[u_k ; v_{2p}]})}}\ ] ] converges .
really , for any @xmath151 @xmath152 @xmath153 } , \ldots , ( i - p_h))\widetilde{a}{1\!\!\,{\rm i}}_{[u_k ; v_{2p}]}}}\cdot\ ] ] @xmath154 } , \ldots , ( i - p_h))\widetilde{a}{1\!\!\,{\rm i}}_{[u_k ; v_{2p}]})\vec{\theta},\vec{\theta})}\ ] ] as @xmath104 since for any @xmath155 less or equal to @xmath153 } , \ldots , ( i - p_h))\widetilde{a}{1\!\!\,{\rm i}}_{[u_k ; v_{2p}]}}},\ ] ] then using and applying the lebesgue dominated convergence theorem one can obtain the statement of the theorem . therefore , let us check . to do this we need the following statements which are related to the properties of the gram determinant in the integrand .
those statements were proved in the works @xcite .
let @xmath156 be the finite - dimensional subspace of the space @xmath40).$ ] suppose that @xmath157 is an orthonormal basis for @xmath158 denote by @xmath159 the projection onto @xmath158 one can prove the following statement .
it follows from lemma [ lem2.2 ] that @xmath162 } , \ldots,(i - p_h)\widetilde{a}{1\!\!\,{\rm i}}_{[u_k;v_p ] } ) = g\big(\widetilde{a}{1\!\!\,{\rm i}}_{[u_k;v_1 ] } , \ldots,\widetilde{a}{1\!\!\,{\rm i}}_{[u_k;v_p ] } , \frac{h}{\|h\| } \big).\ ] ] since @xmath109 is a continuously invertible operator in the space @xmath110),$ ] then ( 16 ) can be represented as follows @xmath163 } , \ldots,\widetilde{a}{1\!\!\,{\rm i}}_{[u_k;v_p ] } , \widetilde{a}\widetilde{a}^{-1}\frac{h}{\|h\| } \big).\ ] ] one can check that the following theorem holds .
it follows from theorem [ thm2.3 ] and properties of the determinant that greater or equal to @xmath167 } , \ldots , { 1\!\!\,{\rm i}}_{[u_k , v_p ] } , \widetilde{a}^{-1}h).\ ] ] denote by @xmath168 the subspace of all step functions in @xmath158 let @xmath169 be an orthonormal basis for @xmath168 and @xmath170 be an orthonormal basis for the orthogonal complement of @xmath168 in @xmath158 note that @xmath171 is an orthonormal basis for @xmath158 one can see that the following estimates take place .
[ lem2.3]@xcite there exists a constant @xmath34 such that the following relation holds @xmath172 } , \ldots , { 1\!\!\,{\rm i}}_{[0 , t_k ] } , f_1 , \ldots , f_s , e_1 , \ldots , e_m)\geq c\ g({1\!\!\,{\rm i}}_{[0;t_1 ] } , \ldots , { 1\!\!\,{\rm i}}_{[0 , t_k ] } , f_1 , \ldots , f_s).\ ] ] @xcite [ lem2.4 ] let @xmath173 be the points of jumps of functions @xmath174 then there exists a constant @xmath175 which depends on @xmath176 such that the following relation holds @xmath172 } , \ldots , { 1\!\!\,{\rm i}}_{[0 , t_k ] } , f_1 , \ldots , f_s)\geq c_{\vec{s}}\ g({1\!\!\,{\rm i}}_{[0;t_1 ] } , \ldots , { 1\!\!\,{\rm i}}_{[0 , t_k ] } , { 1\!\!\,{\rm i}}_{[0;s_1 ] } , \ldots , { 1\!\!\,{\rm i}}_{[0 , s_n]}).\ ] ] if @xmath177 is not a step function , then it follows from lemma [ lem2.3 ] that @xmath178 } , \ldots,{1\!\!\,{\rm i}}_{[u_k ; v_p ] } , \widetilde{a}^{-1}h)\geq c\ g({1\!\!\,{\rm i}}_{[u_k ; v_1 ] } , \ldots,{1\!\!\,{\rm i}}_{[u_k ; v_p]}).\ ] ] the relation implies that less or equal to @xmath179},\ldots,{1\!\!\,{\rm i}}_{[u_k ; v_{2p } ] } ) } } = \ ] ] @xmath180 @xmath181 here @xmath182 is positive constant which depends on @xmath183 and @xmath184 if @xmath177 is a step function , then it follows from lemma [ lem2.4 ] that @xmath185 } , \ldots,{1\!\!\,{\rm i}}_{[u_k ; v_{2p } ] } , \widetilde{a}^{-1}h)\geq c_{\vec{s}}\ g({1\!\!\,{\rm i}}_{[u_k ; v_1 ] } , \ldots,{1\!\!\,{\rm i}}_{[u_k ; v_{2p}]},{1\!\!\,{\rm i}}_{[u_k ; s_1 ] } , \ldots,{1\!\!\,{\rm i}}_{[u_k ; s_n]}),\ ] ] where @xmath186 are the points of jumps of the function @xmath187 the relation implies that less or equal to @xmath188 } , \ldots,{1\!\!\,{\rm i}}_{[u_k ; v_{2p}]},{1\!\!\,{\rm i}}_{[u_k ; s_1 ] } , \ldots,{1\!\!\,{\rm i}}_{[u_k ; s_n]})}}.\ ] ] consider the following partition of the domain @xmath189 @xmath190 where @xmath191 @xmath192 @xmath193 note that @xmath194 it follows from that to finish the proof it suffices to check that @xmath195 the convergence of the integral we proved in @xcite . for convenience of the reader ,
let us now briefly recall essential points of the proof .
the change of variables in the integral implies that @xmath196 @xmath197 @xmath198 } , \ldots,{1\!\!\,{\rm i}}_{[0;v_k]},{1\!\!\,{\rm i}}_{[0;1]})}}=\ ] ] @xmath199 } , \ldots,(i-\widetilde{p}){1\!\!\,{\rm i}}_{[0;v_k]},{1\!\!\,{\rm i}}_{[0;1]})}},\ ] ] where @xmath200 is a positive constant which depends on @xmath201 and @xmath202 is a projection onto linear span generated by @xmath203}.$ ] one can check that equals @xmath204 where @xmath205 $ ] is the brownian bridge .
to finish the proof it suffices to check that @xmath206 really , note that @xmath207 it is known that joint probability density of the random variables @xmath208 and @xmath209 has the following representation @xcite @xmath210 then @xmath211 @xmath212 where @xmath213 is the gamma function .
99 d. khoshnevisan , _ five lectures on brownian sheet : summer intership program university of wisconsin - madison _ , 2001 , 61 p. j.b.walsh , _ an introduction to stochastic partial differential equations _ , lecture notes in math . 1180 ( 1986 ) , 265 - 439 .
r.cairoli , j.b.walsh , _ stochastic integrals in the plane _ , acta mathematica 134 ( 1975 ) , 111 - 183 .
m.hairer , _ an introduction to stochastic _ , pdes ,
the university of warwick , 2009 , 78 p. t.funaki , _ random motion of strings and related stochastic evolution equations _ , nagoya math.j .
89 ( 1983 ) , 129 - 193 .
m.kardar , g.parisi , y .- c.zhang , _ dynamic scaling of growing interfaces _ , physical reviev letters 56 ( 1986 ) , no . 9 , 889 - 892 .
a.a.dorogovtsev , o.l.izyumtseva , _ self - intersection local time _ , ukrainian math .
journal 68 ( 2016 ) , no.3 , 291 - 341 .
a.a.dorogovtsev , o.l.izyumtseva , _ self - intersection local times for gaussian processes _ , germany : lap lambert academic publishing , 2011 , 152 p. a.a.dorogovtsev , _ stochastic integration and one class of gaussian random processes _ ,
ukr.math.journal 50 ( 1998 ) , no.4 , 495 - 505 .
o.l.izyumtseva , _ on the local times for gaussian integrators _ , theory of stochastic processes 19 ( 35 ) ( 2014 ) , no.1 , 11 - 25 .
a.a.dorogovtsev , o.l.izyumtseva , _ properties of gaussian local times _ , lithuanian mathematical journal 55 ( 4 ) ( 2015 ) , 489 - 505 .
a.a.dorogovtsev , o.l.izyumtseva , _ on self - intersection local times for generalized brownian bridges and the distance between step functions _ ,
theory of stochastic processes 20 ( 36 ) ( 2015 ) , no.1 , 1 - 13 .
o.l.izyumtseva , _ moments estimates for local times of a class of gaussian processes _ , communications on stochastic analysis 10 ( 2016 ) , no.1 , 97 - 116 .
a.n.borodin , _ brownian local time _ , russian math .
surveys 44 ( 1989 ) , no.2 , 1 - 51 . | in this article we discuss the existence of local time for a class of gaussian processes which appears as the solutions to some stochastic evolution equations .
we show that on small intervals such processes are gaussian integrators generated by a continuously invertible operators .
this allows us to conclude that the considered processes have a local time on any finite interval with respect to spatial variable . |
one of the most fundamental results on the symmetric functions is the determinantal expression of the schur function @xmath7 for any skew shape @xmath0 ; see @xcite . the jacobi - trudi determinant @xcite and its dual @xcite , the giambelli determinant @xcite as well as lascoux and pragacz s rim ribbon determinant @xcite are all of this kind .
hamel and goulden @xcite remarkably found that all above mentioned determinants for the schur function @xmath7 can be unified through the concept of _ outside decompositions _ of the skew shape @xmath0 . in what follows all definitions will be postponed until subsection [ ss:2 ] and we first present hamel and goulden s theorem ( theorem [ t : hg ] ) . [ t : hg ] if the skew diagram of @xmath0 is edgewise connected .
then , for any outside decomposition @xmath11 of the skew shape @xmath0 , it holds that @xmath12_{i , j=1}^{g},\end{aligned}\ ] ] where @xmath13 and @xmath14 if @xmath15 is undefined .
their proof is based on a lattice path construction and the lindstrm - gessel - viennot methodology @xcite . in this paper
we generalize the concept of outside decompositions even further .
we introduce the concept of _ outside nested decompositions _ of the skew shape @xmath0 and our first main result is a generalization of theorem [ t : hg ] with respect to any outside nested decomposition @xmath1 of the skew shape @xmath0 . for any such decomposition @xmath1 of the skew shape @xmath0 where @xmath2 is a thickened strip for every @xmath3 , if @xmath4 is the number of boxes that are contained in two distinct thickened strips of @xmath5 .
then , our main theorem provides a determinantal formula of the function @xmath6 with the schur functions of thickened strips as entries .
the precise statement is the following .
[ t : thick ] if the skew diagram of @xmath0 is edgewise connected .
then , for any outside nested decomposition @xmath1 of the skew shape @xmath0 , we set @xmath16 that is the number of common special corners of @xmath5 and we have @xmath17_{i , j=1}^{g}\quad \mbox { where } \ , p_{1^{r}}(x)=(\sum_{i=1}^{\infty}x_i)^r,\end{aligned}\ ] ] @xmath13 and @xmath18 if @xmath19 is undefined .
the function @xmath8 is the power sum symmetric function index by the partition @xmath9 and @xmath20 if @xmath21 . when @xmath21 and all thickened strips @xmath2 are strips , we retrieve hamel and goulden s theorem on the _ outside decompositions _ of the skew shape @xmath0 .
with the help of theorem [ t : thick ] , it suffices to find an outside nested decomposition with minimal number of thickened strips in order to reduce the order of the determinantal expression of the schur function @xmath7 .
let @xmath22 and @xmath23 denote the number of boxes contained in the skew shape @xmath0 and the number of standard young tableaux of shape @xmath0 with the entries from @xmath24 to @xmath22 ( similarly for @xmath25 and @xmath26 ) .
then , by applying the exponential specialization on both sides of ( [ e : hg ] ) , one immediately gets [ c : thick2 ] if the skew diagram of @xmath0 is edgewise connected .
then , for any outside nested decomposition @xmath1 of the skew shape @xmath0 , we have @xmath27_{i , j=1}^{g } \quad \mbox { where } \ , a_{i , j}=\vert\theta_i\#\theta_j\vert,\end{aligned}\ ] ] @xmath28 and @xmath29 if @xmath19 is undefined .
it should be noted that the parameter @xmath4 vanishes in ( [ e : hg1 ] ) .
our second main result is an enumeration of @xmath10-strip tableaux by applying corollary [ c : thick2 ] , which provides another proof of baryshnikov and romik s results in @xcite .
baryshnikov and romik @xcite counted @xmath10-strip tableaux via extending the transfer operator approach due to elkies @xcite . in subsection
[ ss:1 ] and [ ss:2 ] we introduce all necessary notations and definitions . in section [ s : proof1 ] we prove theorem [ t : thick ] and corollary [ c : thick2 ] . in section [ e : app ] we introduce the notion of @xmath10-strip tableaux and count the number of @xmath10-strip tableaux . * a _ partition _ @xmath30 of @xmath31 , denoted by @xmath32 , is a sequence @xmath33 of non - negative integers such that @xmath34 and their sum is @xmath31 .
the non - zero @xmath35 are called the _ parts _ of @xmath30 and the number of parts is the length of @xmath30 , denoted by @xmath36 .
let @xmath37 denote the number of parts of @xmath30 that equal @xmath3 , we simply write @xmath38 . * given a partition @xmath33 , the _ standard diagram _ of @xmath30 is a left - justified array of @xmath39 boxes with @xmath40 in the first row , @xmath41 in the second row , and so on . * a _ skew diagram _ of @xmath0 ( also called _ a skew shape @xmath0 _ ) is the difference of two skew diagrams where @xmath42 .
note that the standard shape @xmath30 is just the skew shape @xmath0 when @xmath43 . *
the _ content _ of a box @xmath44 in a skew shape @xmath0 equals @xmath45 if the box @xmath44 is in column @xmath46 from the left and row @xmath47 from the top of the skew shape @xmath0 .
we refer to box @xmath44 as box @xmath48 and @xmath48 is called its _
coordinate_. a _ diagonal of content @xmath49 _ in a skew diagram is a set of boxes with content @xmath49 in a skew diagram . * a skew diagram
` starts ' at a box ( called _ the starting box _ ) if that box is the bottommost and leftmost box in the skew diagram , and a skew diagram ` ends ' at a box ( called _ the ending box _ ) if that box is the topmost and rightmost box in the skew diagram . * a _ semistandard young tableau _ ( resp . _
standard young tableau _ ) of skew shape @xmath0 is a filling of the boxes of the skew diagram of @xmath0 with positive integers such that the entries strictly increase down each column and weakly ( resp .
strictly ) increase left to right across each row . in a semistandard young tableau
@xmath50 we use @xmath51 to represent the positive integer in the box @xmath44 of @xmath50 .
the _ schur function _ , @xmath7 , in the variables @xmath52 , is given by @xmath53 where the summation is over all semistandard young tableaux @xmath50 of shape @xmath0 and @xmath54 means that @xmath44 ranges over all boxes in the skew diagram of @xmath0 .
in particular , @xmath13 . the _ complete symmetric functions _
@xmath55 are defined by @xmath56 the jacobi - trudi identity is a determinantal expression of schur function @xmath7 in terms of complete symmetric functions @xmath55 ; see @xcite .
[ t : jacobi - trudi ] let @xmath0 be a skew shape partition , let @xmath57 and @xmath58 have at most @xmath59 parts .
then @xmath60_{i , j=1}^{k}.\end{aligned}\ ] ] the classical aitken formula for the number of standard young tableaux of skew shape can be directly obtained by applying the exponential specialization on the jacobi - trudi identity ; see chapter 7 of @xcite .
we denote by @xmath22 the number of boxes contained in the skew diagram of @xmath0 and denote by @xmath23 the number of standard young tableaux of shape @xmath0 with the entries from @xmath24 to @xmath22 .
let @xmath0 be a skew shape partition , let @xmath57 and @xmath58 have at most @xmath59 parts
. then @xmath61_{i , j=1}^{k}.\end{aligned}\ ] ] it is clear that the order of the determinant in the jacobi - trudi identity and in the aitken formula equals the number @xmath36 of parts in @xmath30 . using ( [ e : aitken ] ) to compute @xmath23 becomes difficult when the partitions @xmath30 and @xmath62 are large , even when their difference @xmath0 is small .
we start with the strips and outside decompositions .
hamel and goulden described the notion of an _ outside decomposition _ of the skew shape @xmath0 , which generalizes lascoux and pragacz s rim ribbon decomposition @xcite . with the help of hamel and goulden s theorem @xcite , for any skew shape @xmath0
, one can reduce the order of the determinant in the jacobi - trudi identity to the number of strips contained in any outside decomposition of skew shape @xmath0 .
two boxes are said to be edgewise connected if they share a common edge . a skew diagram @xmath63
is said to be _ edgewise connected _ if @xmath63 is an edgewise connected set of boxes .
[ d : ribbon ] a skew diagram @xmath63 is a _ strip _
if @xmath63 is edgewise connected and it contains no @xmath64 blocks of boxes .
the strips in definition [ d : ribbon ] are called ` border strips ' by macdonald @xcite and are called ` ribbons ' by lascoux and pragacz @xcite .
we adopt the name ` strips ' from @xcite .
[ d : podec ] suppose that @xmath65 are strips of a skew diagram of @xmath0 and every strip has a starting box on the left or bottom perimeter of the diagram and an ending box on the right or top perimeter of the diagram .
then we say the totally ordered set @xmath11 is an _ outside decomposition _ of @xmath0 if the union of these strips is the skew diagram of @xmath0 and every two strips @xmath66 in @xmath67 are disjoint ,
that is , @xmath68 and @xmath69 have no boxes in common .
the rim ribbon decomposition of @xmath0 introduced by lascoux and pragacz @xcite is an outside decomposition with minimal number of strips ; see @xcite and @xcite .
see figure [ f : pic-3 ] for an outside decomposition and two non - outside decompositions where all boxes are marked by black dots .
the first two decompositions in figure [ f : pic-3 ] are not outside decompositions since the strip @xmath70 of the left one has a starting box neither on the left nor on the bottom perimeter of the skew diagram and the strip @xmath71 of the middle one has an ending box neither on the right nor on the top perimeter of the skew diagram . .
[ f : pic-3 ] ] we next introduce the notion of _ thickened strips _ and we will decompose the skew diagram of @xmath0 into a sequence of thickened strips , in order to extend hamel and goulden s theorem @xcite on the determinantal expression of the schur function @xmath7 . our extension is motivated by the enumeration of @xmath72-strip tableaux where any outside decomposition of @xmath72-strip diagram with @xmath31 columns consists of at least @xmath31 strips ( see subsection [ ss : odd ] ) . so the order of the determinantal expression of @xmath7 can not be further reduced by applying hamel and goulden s theorem ( theorem [ t : hg ] ) .
[ d : to ] a skew diagram @xmath73 is a thickened strip if @xmath73 is edgewise connected and it neither contains a @xmath74 block of boxes nor a @xmath75 block of boxes . by definition
the only difference between strips and thickened strips is that thickened strips could have some @xmath64 blocks of boxes ; see figure [ f : pic-9 ] .
] we next define the _ corners _ and the _ special corners _ of a thickened strip @xmath2 because in contrast to the outside decompositions , we allow two thickened strips in an outside nested decomposition to have special corners in common . in
what follows , note that the box @xmath48 always refers to the box with coordinate @xmath48 in the skew diagram of @xmath0 .
( corner , special corner)[d : sc ] when a thickened strip @xmath2 has more than one box , we define that _ a corner _
@xmath48 of a thickened strip @xmath2 is an upper corner or a lower corner , where _ an upper corner _ @xmath48 of @xmath2 is a box @xmath48 such that neither the box @xmath76 nor the box @xmath77 is contained in @xmath2 .
likewise , _ a lower corner _
@xmath48 of @xmath2 is a box @xmath48 such that neither the box @xmath78 nor the box @xmath79 is contained in @xmath2 .
we say that a corner @xmath48 of a thickened strip @xmath2 is _ special _ if the corner @xmath48 satisfies one of the following conditions : 1 .
the corner @xmath48 is the starting box or the ending box of @xmath2 ; 2 .
the corner @xmath48 is contained in a @xmath64 block of boxes of @xmath2 .
consider the thickened strip in figure [ f : pic-9 ] ( the left one ) , the only corner that is not special in this thickened strip is the box @xmath80 .
now we are ready to present the outside thickened strip decomposition .
[ d : todec ] suppose that @xmath81 are thickened strips in the skew diagram of @xmath0 and every thickened strip has a starting box on the left or bottom perimeter of the diagram and an ending box on the right or top perimeter of the diagram .
then we say the totally ordered set @xmath1 is an _ outside thickened strip decomposition _ of the skew diagram of @xmath0 if the union of the thickened strips @xmath2 of @xmath5 is the skew diagram of @xmath0 , and for all @xmath82 , one of the following statements is true : 1 . two thickened strips @xmath2 and @xmath83 are disjoint , that is , @xmath2 and @xmath83 have no boxes in common ; 2 .
one thickened strips @xmath83 is on the right side or the bottom side of the other thickened strip @xmath2 and they have some special corners in common , where each common special corner @xmath48 is a lower corner of @xmath2 and an upper corner of @xmath83 .
every special corner of a thickened strip in @xmath5 is called a _ special corner of @xmath5 _ and every common special corner of any two distinct thickened strips of @xmath5 is called a _ common special corner of @xmath5_.
if @xmath2 has only one box @xmath48 and box @xmath48 is also a special corner of @xmath83 .
then the outside thickened strip decomposition @xmath5 is essentially the same to the one without @xmath2 .
so we exclude this scenario .
figure [ f : pic2 - 01 ] ( middle , right ) shows an outside thickened strip decomposition @xmath84 of the skew diagram of @xmath85 where the boxes @xmath86 and @xmath87 are the common special corners of @xmath88 and @xmath89 .
the box @xmath90 is the only common special corner of @xmath91 and @xmath88 . in figure
[ f : pic2 - 01 ] and figure [ f : pic-12 ] every common special corner of @xmath5 is marked by a black square , while other boxes are marked by black dots .
( middle , right ) of the skew diagram @xmath85 ( left).[f : pic2 - 01 ] ] figure [ f : pic-12 ] ( middle , right ) gives an outside thickened strip decomposition @xmath92 of the skew diagram of @xmath93 where all the common special corners of @xmath5 are boxes @xmath94 . and
@xmath95 are drawn separately ( middle , right ) .
the dashed line with integer @xmath96 represents the diagonal of content @xmath96.[f : pic-12 ] ] we observe that , unlike the strips in any outside decomposition , the thickened strips in any outside thickened strip decomposition @xmath5 are _ not necessarily nested _ ; see definition [ d : nesttodec ] . however , the _ nested property _ of thickened strips in an outside thickened strip decomposition is of central importance in the proof of theorem [ t : thick ] . in view of this
, we need to introduce the _ enriched diagrams _ and the _ directions _ of all boxes in the skew shape @xmath0 to describe the nested property of thickened strips .
[ d : enrich ] suppose that @xmath1 is an outside thickened strip decomposition of the skew shape @xmath0 , for every @xmath3 such that @xmath97 , and for box @xmath48 that is the starting box or the ending box of @xmath2 , we shall add new boxes to @xmath2 according to the following rules : 1 .
if box @xmath48 is a lower corner of @xmath2 and an upper corner of some other thickened strip in @xmath5 , we add boxes @xmath98 that are not contained in @xmath2 to @xmath2 ; 2 .
if box @xmath48 is an upper corner of @xmath2 and a lower corner of some other thickened strip in @xmath5 , we add boxes @xmath99 that are not contained in @xmath2 to @xmath2 . where all the coordinates of new boxes are _ relative _ to the coordinates of the boxes in the skew diagram of @xmath0 .
we denote by @xmath100 the diagram after adding the new boxes to @xmath2 and we call @xmath100 _ an enriched thickened strip_. if neither the starting box nor the ending box of @xmath2 satisfies @xmath101 or @xmath102 , then @xmath103 . _ an enriched diagram _
@xmath104 is the union of all enriched thickened strips @xmath100 for every @xmath2 of @xmath5 . in figure
[ f : pic2 - 01 ] the box @xmath86 contained in @xmath89 and @xmath88 is the only box that satisfies conditions @xmath101 and @xmath102 of definition [ d : enrich ] .
so we add the boxes @xmath105 to @xmath89 and add the boxes @xmath106 to @xmath88 ; see figure [ f : pic2 - 1 ] where all newly added boxes are colored grey .
( left ) and the enriched thickened strips ( middle , right ) where @xmath84 is given in figure [ f : pic2 - 01].[f : pic2 - 1 ] ] the enriched diagram @xmath104 may not be a skew diagram ; see figure [ f : pic2 - 1 ] and [ f : pic-13 ] . with enriched
thickened strips @xmath100 where the outside thickened strip decomposition @xmath5 is given in figure [ f : pic-12 ] and the dashed line with integer @xmath96 represents the diagonal of content @xmath96 .
[ f : pic-13 ] ] with the help of enriched diagram @xmath104 , one can define the directions of all boxes other than the special corners of @xmath5 in the skew diagram of @xmath0 . for every box @xmath48 of the skew diagram of @xmath0 , if box @xmath48 is _ not _ a special corner of @xmath5 , then box @xmath48 is contained in only one thickened strip @xmath2 of @xmath5 .
we may define the _ direction _ of box @xmath48 in the enriched diagram @xmath104 according to the following rules : 1 .
if both boxes @xmath76 and @xmath79 are contained in the enriched thickened strip @xmath100 of @xmath104 , then we say the box @xmath48 _ goes right and up _ ; 2 .
if not both boxes @xmath76 and @xmath79 are contained in @xmath100 , then we say that the box @xmath48 _ goes right _ or
_ goes up _ if @xmath79 or @xmath76 is contained in @xmath100 ; 3 .
if neither box @xmath79 nor box @xmath76 is contained in @xmath100 , then box @xmath48 must be the ending box of @xmath2 , thus it must be on the top or right perimeter of the skew diagram of @xmath0 , and we say that box @xmath48 _ goes up _ if it is on the top perimeter of @xmath0 and that box @xmath48 _ goes right _ if it is on the right perimeter but not on the top perimeter of the skew diagram of @xmath0 .
[ d : nesttodec ] an outside thickened strip decomposition @xmath5 is an outside nested decomposition if @xmath1 is _ nested _ , that is , for all @xmath49 , one of the following statements is true : 1 .
all boxes of content @xmath49 all go right or all go up ; 2 .
all boxes of content @xmath49 or all boxes of content @xmath107 are all special corners of @xmath5 .
it should be noted that all boxes of content @xmath107 are special corners of @xmath5 if and only if all boxes of content @xmath49 all go right and up .
definition [ d : nesttodec ] is analogous to the nested property of the strips in any outside decomposition where all boxes on the same diagonal of the skew shape @xmath0 all go right or all go up ; see @xcite . by definition [ d : nesttodec ]
the outside thickened strip decomposition in figure [ f : pic-12 ] is not an outside nested decomposition because two boxes on the diagonal of content @xmath96 are special corners , but one box goes right , while the outside thickened strip decomposition in figure [ f : pic2 - 01 ] is an outside nested decomposition because all boxes on the diagonal of content @xmath108 are all special corners , all boxes on the diagonal of content @xmath109 all go up , and all boxes on the diagonal of content @xmath96 all go right .
hamel and goulden @xcite defined a non - commutative operation @xmath110 for every two strips of an outside decomposition @xmath11 of the skew shape @xmath0 , also when the skew shape @xmath0 is edgewise disconnected .
subsequently , chen , yan and yang @xcite came up with the notion of cutting strips so as to derive a transformation theorem for hamel and goulden s determinantal formula , in which one of the key ingredients is a bijection between the outside decompositions of a given skew diagram and the cutting strips .
based on these previous work , we will extend the non - commutative operation @xmath110 for every two thickened strips of an outside nested decomposition @xmath1 of the skew shape @xmath0 . in order to provide a simple definition of @xmath19
, we need to introduce the _ thickened cutting strips _ , which are called ` cutting strips ' for any outside decomposition in @xcite .
[ d : tcs ] the _ thickened cutting strip _
@xmath111 with respect to an outside nested decomposition @xmath1 is a thickened strip obtained by successively superimposing the enriched thickened strips @xmath112 of @xmath104 along the diagonals .
we say that _ a box @xmath44 of the thickened cutting strip @xmath111 has content @xmath49 _ if box @xmath44 is on the diagonal of content @xmath49 in the skew diagram of @xmath0 and we represent each box of the thickened cutting strip @xmath111 as follows : 1 .
box @xmath113 $ ] denotes the unique box of @xmath111 with content @xmath49 ; 2 .
box @xmath114 $ ] and box @xmath115 $ ] denote the upper and the lower corner of @xmath111 with content @xmath49 if they are contained in a @xmath64 block of boxes in @xmath111 .
because of the nested property in definition [ d : nesttodec ] , the thickened cutting strip @xmath111 with respect to any outside nested decomposition @xmath5 is a thickened strip .
consider the outside nested decomposition @xmath84 in figure [ f : pic2 - 01 ] , the thickened cutting strip with respect to @xmath5 is constructed in figure [ f : pic-15 ] , where the dashed lines represent the diagonals of content @xmath108 respectively .
( right ) with respect to the outside nested decomposition @xmath116 in figure [ f : pic2 - 01 ] and the enriched thickened strips @xmath117 are given in figure [ f : pic2 - 1].[f : pic-15 ] ] [ d : operator1 ] if the skew diagram of @xmath0 is edgewise connected , let @xmath118 @xmath119 be an outside nested decomposition of skew shape @xmath0 , and let @xmath111 be the thickened cutting strip with respect to @xmath5 .
for each thickened strip @xmath2 in @xmath5 , if @xmath120 is the content of the starting box of @xmath2 , the starting box @xmath121 of @xmath2 is given as below : 1 .
@xmath122 $ ] if the starting box is not a special corner of @xmath5 ; 2 . @xmath123 $ ] if the starting box is a special corner of @xmath5 and an upper corner of @xmath2 ; 3 .
@xmath124 $ ] if the starting box is a special corner of @xmath5 and a lower corner of @xmath2 . likewise , we denote the ending box of @xmath2 by @xmath125 if we replace @xmath126 by @xmath125 and replace the starting box by the ending box from the above notations
. then @xmath2 forms a segment of the thickened cutting strip @xmath111 starting with the box @xmath126 and ending with the box @xmath125 , which is denote by @xmath127 $ ] .
we may extend the notion to @xmath128 $ ] in the following way : 1 .
if @xmath129 or @xmath130 , then @xmath128 $ ] is a segment of @xmath111 starting with the box @xmath131 and ending with the box @xmath125 ; 2 .
if @xmath131 and @xmath125 are in the same diagonal of @xmath111 , or @xmath132 , then @xmath128=\varnothing$ ] ; 3 . if @xmath133 , then @xmath128 $ ] is undefined .
for any two thickened strips @xmath2 and @xmath83 of @xmath5 , the thickened strip @xmath19 is defined as @xmath128 $ ] .
we only need to deal with the outside nested thickened strip decompositions of an edgewise connected skew diagram because the schur function of any edgewise disconnected diagram is a product of schur functions of edgewise connected components .
since @xmath5 is an outside nested decomposition , we can identify every thickened strip @xmath2 as a segment of @xmath111 starting with the box @xmath126 and ending with the box @xmath125 .
[ e : op1 ] consider the outside nested decomposition @xmath84 in figure [ f : pic2 - 01 ] , one has @xmath134=[[-3,+],[5]]=(5,5,5,4,4)/(4,3,3,2)$ ] , that is , a segment of the thickened cutting strip @xmath111 in figure [ f : pic-15 ] starting with box @xmath135 $ ] and ending with box with content @xmath136 $ ] .
similarly , the thickened strips obtained by the operation @xmath110 are given below : @xmath137 if @xmath5 is an outside decomposition where all thickened strips @xmath2 are strips , then the starting box of any strip @xmath2 and the ending box of any strip @xmath83 are @xmath122 $ ] and @xmath138 $ ] .
in @xcite hamel and goulden noticed that any order of the strips in an outside decomposition play the same role .
chen , yan and yang @xcite also showed hamel and goulden s theorem in terms of the canonical order of strips and our extension ( theorem [ t : thick ] ) also works for any total order of the thickened strips in an outside nested decomposition .
since it is convenient to construct an involution in the context of lattice paths , we choose to represent semistandard young tableaux of thickened strip shape in the language of lattice paths .
our proof of theorem [ t : thick ] consists of three main steps . in the first step
we build a one - to - one correspondence between semistandard young tableaux of thickened strip shape to _ double lattice paths _ , which is based on a bijection between semistandard young tableaux of strip shape and lattice paths in @xcite . in the second step
we introduce the _ separable _ @xmath139-tuples of double lattice paths and show that the generating function of all separable @xmath139-tuples of double lattice paths is @xmath140 . in the last step we will construct a sign - reversing and weight - preserving involution @xmath141 on all non - separable @xmath139-tuples of double lattice paths , so that only the separable ones contribute to the determinant @xmath142_{i , j=1}^g$ ] in theorem [ t : thick ]
. we will prove corollary [ c : thick2 ] by using the exponential specializations of the schur functions and power sum symmetric functions .
first we recall that @xmath111 is the thickened cutting strip which corresponds to @xmath5 ( see definition [ d : tcs ] ) and @xmath19 is given in definition [ d : operator1 ] . for any @xmath82
, we will introduce the double lattice path @xmath143 in definition [ d : plattice ] .
[ d : plattice ] under the assumption of theorem [ t : thick ] , for every @xmath82 , the _
double lattice paths _ with respect to @xmath5 , are pairs @xmath144 of lattice paths
where @xmath145 and @xmath146 start at @xmath147 and end at @xmath148 .
the starting point @xmath147 and the ending point @xmath148 are fixed as below : 1 . if the starting box @xmath48 of @xmath83 is a common special corner of @xmath5 , and * if box @xmath48 is a lower corner of @xmath83 , then @xmath149 ; * otherwise if box @xmath48 is an upper corner of @xmath83 , then @xmath150 ; 2 .
if the starting box @xmath48 of @xmath83 is not a common special corner of @xmath5 , and * if box @xmath48 is on the left perimeter of the skew shape @xmath0 , then @xmath149 ; * otherwise if box @xmath48 is only on the bottom perimeter of the skew shape @xmath0 , then @xmath150 ; 3 . if the ending box @xmath151 of @xmath2 is a common special corner of @xmath5 , and * if box @xmath151 is a lower corner of @xmath2 , then @xmath152 ; * otherwise if box @xmath151 is an upper corner of @xmath2 , then @xmath153 ; 4
if the ending box @xmath151 of @xmath2 is not a common special corner of @xmath5 , and * if box @xmath151 is on the right perimeter of the skew shape @xmath0 , then @xmath153 ; * otherwise if box @xmath151 is only on the top perimeter of the skew shape @xmath0 , then @xmath152 .
furthermore , the lattice paths @xmath145 and @xmath146 consist of four types of steps : an up - vertical step @xmath154 , a down - vertical step @xmath155 , a horizontal step @xmath156 and a diagonal step @xmath157 , which satisfy the conditions 1 .
a down - vertical step @xmath158 must not precede an up - vertical step @xmath159 and must not precede a horizontal step @xmath160 ; 2 .
an up - vertical step @xmath159 must not precede a down - vertical step @xmath158 and must not precede a diagonal step @xmath161 .
moreover , there is a horizontal step @xmath160 of the lattice path @xmath145 or @xmath146 between lines @xmath162 and @xmath163 if one of @xmath164 holds : 1 .
a box of content @xmath165 is to the left of a box of content @xmath49 in @xmath19 ; 2 .
the starting box of @xmath83 has content @xmath49 and @xmath166 .
there is a diagonal step @xmath161 of the lattice path @xmath145 or @xmath146 between lines @xmath162 and @xmath163 if one of @xmath167 holds : 1 .
a box of content @xmath165 is right below a box of content @xmath49 in @xmath19 ; 2 .
the starting box of @xmath83 has content @xmath49 and @xmath168 .
we connect all these non - vertical steps by up - vertical and down - vertical steps so that every non - vertical step of @xmath145 is either above or the same as the one of @xmath146 between any lines @xmath162 and @xmath163 .
when @xmath5 is an outside decomposition , for all @xmath3 and @xmath169 , the double lattice path @xmath170 with respect to @xmath5 is a lattice path , that is , @xmath171 where all steps between any lines @xmath162 and @xmath163 are all horizontal or all diagonal ; see @xcite . because @xmath5 is an outside nested decomposition , by @xmath101-@xmath102 in definition [ d : plattice ] , all starting points and all ending points are all different .
once the starting point @xmath147 and the ending point @xmath148 are chosen , the shape of any double lattice path @xmath143 is fixed , that is , whether any non - vertical step of @xmath143 is horizontal or diagonal , is determined by @xmath19 .
this allows us to identify the schur function @xmath172 as the generating function of all weighted double lattice paths from @xmath147 to @xmath148 in subsection [ ss : invo ] . for every @xmath3 and @xmath169 , let @xmath173 represent the set of all double lattice paths from @xmath147 to @xmath148 , and let @xmath174 represent the set of all semistandard young tableaux of thickened strip shape @xmath19 .
we next establish that [ l : bij3 ] there is a bijection @xmath175 between the set @xmath174 and the set @xmath173 .
if @xmath176 , according to @xmath177-@xmath178 of definition [ d : plattice ] , @xmath173 contains only one double lattice path @xmath143 that has no non - vertical steps from @xmath147 to @xmath148 , which corresponds to the empty tableau from @xmath179 . if @xmath19 is undefined , the starting point @xmath147 is on the right hand side of the ending point @xmath148 , so by definition [ d : plattice ] there exist no double lattice paths from @xmath147 to @xmath148 , that is , the set @xmath173 is undefined , which corresponds to the undefined set @xmath174 .
otherwise , given a semistandard young tableau @xmath180 of thickened strip shape @xmath19 , we build the corresponding double lattice path @xmath181 starting with @xmath147 and ending at @xmath148 . for every box @xmath44 in @xmath180 , suppose that the box @xmath44 of content @xmath49 has entry @xmath182 in @xmath180
. then we put a horizontal step from @xmath183 to @xmath184 if one of @xmath185 is true : 1 .
a box of content @xmath165 is to the left of @xmath44 in @xmath19 ; 2 .
@xmath44 is the starting box of @xmath83 and @xmath166 .
we put a diagonal step from @xmath186 to @xmath184 if one of @xmath187 is true : 1 .
a box of content @xmath165 is right below @xmath44 in @xmath19 ; 2 .
@xmath44 is the starting box of @xmath83 and @xmath168 .
we connect all these non - vertical steps by up - vertical and down - vertical steps . in this way
we get a pair @xmath170 of lattice paths where every non - vertical step of @xmath145 is either above or the same as the one of @xmath146 . by construction in the lattice path @xmath145 or @xmath146 , there is no down - vertical step preceding an up - vertical step and there is no up - vertical step preceding a down - vertical step . since @xmath180 is a semistandard young tableau , there is no down - vertical step preceding a horizontal step because otherwise , the entries along each row of @xmath180 is not weakly increasing from left to right .
similarly there is no up - vertical step preceding a diagonal step because otherwise , the entries along each column of @xmath180 is not strictly decreasing from bottom to top .
so by definition [ d : plattice ] the path @xmath170 is a double lattice path .
the map @xmath188 is a bijection because the above process is reversible . in lemma
[ l : bij3 ] we observe that the point @xmath184 is the ending point of some non - vertical step of @xmath143 if and only if a box of content @xmath49 has entry @xmath182 in @xmath180 where @xmath189 . for @xmath190 ,
consider the thickened strip @xmath2 of the skew shape @xmath85 in figure [ f : pic2 - 01 ] , the corresponding double lattice path @xmath191 of the thickened strip tableau @xmath192 is given in figure [ f : pic2 - 7 ] where all integers represent the @xmath193-th coordinates of all ending points from the non - vertical steps in @xmath194 .
we have discussed the shape of @xmath195 in example [ e : op1 ] .
since the starting box @xmath196 $ ] of @xmath195 is an upper corner of @xmath88 , according to @xmath101 in definition [ d : plattice ] , the starting point @xmath197 is @xmath198 and we put a diagonal step from @xmath199 to @xmath200 in figure [ f : pic2 - 6 ] because of @xmath178 in definition [ d : plattice ] .
similarly , since the ending box of @xmath195 is @xmath201 $ ] , the ending point @xmath202 is @xmath203 .
in addition , the corresponding double lattice path @xmath204 of the empty thickened strip tableau @xmath205 consists of only vertical steps from @xmath206 to @xmath207 .
( left ) , the corresponding thickened strip tableau @xmath208 ( middle ) and the thickened cutting strip @xmath111 where the starting box @xmath196 $ ] and the ending box @xmath201 $ ] of @xmath195 are marked with empty squares ( right ) .
[ f : pic2 - 7 ] ] with the help of lemma [ l : bij3 ] we can establish the relation between semistandard young tableaux of skew shape @xmath0 and @xmath139-tuples of double lattice paths . under the assumption of theorem [ t : thick ] , for any @xmath209 , we write @xmath210 and consider any @xmath139-tuple ( [ e : good ] ) of double lattice paths , where the starting points and the ending points of all steps are called _ points _ of ( [ e : good ] ) .
[ d : noncr ] consider a @xmath139-tuple @xmath211 of double lattice paths where @xmath212 .
then ( [ e : good ] ) is _ non - crossing _ if for any @xmath3 and @xmath169 , @xmath213 and @xmath214 are non - crossing .
this holds if and only if 1 .
@xmath213 and @xmath214 are non - intersecting , that is , have no points in common ; 2 .
@xmath214 is on the top side of @xmath213 and they have some points in common , where each common point @xmath184 occurs only when one diagonal step of @xmath214 and one horizontal step of @xmath213 end at the same point @xmath184 . otherwise @xmath213 and @xmath214 are crossing and ( [ e : good ] ) is crossing .
if ( [ e : good ] ) is non - crossing , we call every common point of any two double lattice paths in ( [ e : good ] ) a _ touchpoint _ of ( [ e : good ] ) . when @xmath215 , two double lattice paths @xmath216 and @xmath217 are non - crossing if and only if two semistandard young tableaux @xmath218 and @xmath219 are disjoint or have the same entry in every common special corner of @xmath2 and @xmath83 , where @xmath220 for @xmath221 .
the triple @xmath222 of double lattice paths given in figure [ f : pic2 - 6 ] where the @xmath193-coordinates of @xmath223 are all infinity , is non - crossing and all touchpoints have coordinates @xmath224 .
[ l : compatible ] if a @xmath139-tuple ( [ e : good ] ) of double lattice path is non - crossing , then @xmath225 must be the identity permutation , that is , @xmath215 .
we shall prove the equivalent statement , namely , if @xmath209 and @xmath226 , then any @xmath139-tuple ( [ e : good ] ) of double lattice path is crossing .
first we consider a total order @xmath227 of all starting points @xmath228 and a total order @xmath227 of all ending points @xmath229 of the double lattice paths .
for every @xmath3 , let @xmath230 and @xmath231 denote the @xmath232-th coordinate and @xmath193-th coordinate of point @xmath233 , similarly for @xmath234 and @xmath235 .
we recall that @xmath236 according to definition [ d : plattice ] .
we define @xmath237 if and only if one of the following conditions is true : 1 .
@xmath238 ; 2 . @xmath239 and @xmath240 ; 3 .
@xmath241 and @xmath242 .
we define @xmath243 if and only if one of the following conditions is true : 1 .
@xmath244 ; 2 . @xmath245 and @xmath246 ; 3 .
@xmath247 and @xmath248 .
we claim that for any @xmath3 and @xmath47 , @xmath237 holds if and only if @xmath243 holds . the essential reason for this
is the fact that @xmath5 is an outside thickened strip decomposition ( definition [ d : todec ] ) , so when we read the boxes on the bottom perimeter and the left perimeter of the skew shape @xmath0 in the right - to - left and bottom - to - top order , the starting box of @xmath249 comes earlier than the starting box of @xmath2 if and only if @xmath237 holds .
since one thickened strip is on the right side or the bottom side of the other thickened strip ; see definition [ d : todec ] , when we read the boxes on the right perimeter and the top perimeter of the skew shape @xmath0 in the bottom - to - top and right - to left order , the ending box of @xmath249 comes earlier than the ending box of @xmath2 if and only if @xmath243 holds .
this implies that for any @xmath3 and @xmath47 , @xmath237 holds if and only if @xmath243 holds .
second , for any @xmath225 such that @xmath250 , there exist two integers @xmath47 and @xmath46 such that @xmath251 and @xmath252 because otherwise it contradicts the assumption @xmath253 .
we wish to show that @xmath254 and @xmath255 are crossing , which can be proved by discussing all cases when one of the previous conditions @xmath101-@xmath256 for @xmath251 is true , and one of the previous conditions @xmath257-@xmath258 for @xmath259 is true .
so we conclude that if @xmath209 and @xmath253 , then ( [ e : good ] ) is crossing . here
note that we need the condition that @xmath5 is an outside thickened strip decomposition .
lemma [ l : compatible ] actually verifies the condition of stembridge s theorem on the non - intersecting lattice paths @xcite .
though stembridge considered only the non - intersecting lattice paths , his theorem is still applicable to the non - crossing double lattice paths .
[ l : bij5 ] under the assumption of theorem [ t : thick ] , there is a bijection between semistandard young tableaux of skew shape @xmath0 and non - crossing @xmath139-tuples of double lattice paths with @xmath4 touchpoints . in view of lemma
[ l : compatible ] , we shall establish a bijection between semistandard young tableaux of skew shape @xmath0 and non - crossing @xmath139-tuples @xmath260 of double lattice paths with @xmath4 touchpoints where @xmath261 for every @xmath3 . for a semistandard young tableau @xmath50 of the skew shape @xmath0
, we can express @xmath50 as a @xmath139-tuple @xmath262 of thickened strip tableaux where @xmath218 is @xmath50 that is restricted to the thickened strip shape @xmath2 . combining the bijection @xmath175 in lemma [ l : bij3 ] ,
one gets the @xmath139-tuple ( [ e : noncr ] ) of double lattice paths where @xmath263 is a double lattice path from @xmath233 to @xmath148 and the fact that ( [ e : noncr ] ) is non - crossing follows from the fact that all entries of boxes on the same diagonal of @xmath50 are strictly increasing from the top - left side to the bottom - right side .
the map @xmath264 is a bijection because , for any @xmath3 and @xmath169 , two double lattice paths @xmath265 and @xmath266 are non - intersecting if and only if two thickened strip tableaux @xmath218 and @xmath219 are disjoint .
furthermore , @xmath267 is on the top side of @xmath265 such that the diagonal step of @xmath267 and the horizontal step of @xmath265 end at the same point @xmath184 if and only if the box of content @xmath49 and with entry @xmath182 in @xmath50 , is an upper corner of @xmath83 and a lower corner of @xmath2 . since there are @xmath4 common special corners of @xmath5 , there are @xmath4 touchpoints of ( [ e : noncr ] ) .
consider the semistandard young tableau @xmath268 of skew shape @xmath85 in figure [ f : pic2 - 9 ] , the corresponding triple of double lattice paths @xmath263 is displayed in figure [ f : pic2 - 6 ] where the @xmath193-coordinates of @xmath223 are all infinity .
in addition , from the non - crossing triple @xmath222 of double lattice paths in figure [ f : pic2 - 6 ] , one has @xmath269 and @xmath270 .
so for instance , any double lattice paths @xmath271 and @xmath272 are crossing because @xmath273 but @xmath274 . which is equivalent to a triple @xmath275 of thickened strip tableaux and @xmath84
is given in figure [ f : pic2 - 01 ] .
[ f : pic2 - 9 ] ] uniquely corresponds to the semistandard young tableau @xmath218 in figure [ f : pic2 - 9 ] .
[ f : pic2 - 6 ] ] for a @xmath139-tuple @xmath276 of double lattice paths where @xmath277 for every @xmath3 , we will describe a separable @xmath139-tuple of double lattice paths and our main task is to establish the bijection in proposition [ p : nonspr ] , from which it follows that the generating function of all weighted separable @xmath139-tuples of double lattice paths is @xmath140 ; see subsection [ ss : invo ] . [
d : spe1 ] for all @xmath49 such that neither @xmath49 nor @xmath165 is the content of some special corner of @xmath5 , we say that two double lattice paths @xmath278 and @xmath279 are _ non - separable at the point @xmath280 _ if and only if @xmath278 intersects @xmath279 at the point @xmath280 , that is , @xmath278 and @xmath279 have the point @xmath280 in common . for a @xmath139-tuple @xmath281 ( see ( [ e : good3 ] ) ) of double lattice paths , we say that @xmath281 is non - separable at a single point if there exist two double lattice paths in @xmath281 such that they are non - separable at a single point .
otherwise we say that @xmath281 is separable at all single points .
the point @xmath280 in definition [ d : spe1 ] is not a touchpoint because if the point @xmath280 is a touchpoint , then @xmath165 must be the content of some special corner of @xmath5 , which is impossible according to the assumption in definition [ d : spe1 ] .
when the outside nested decomposition @xmath5 is an outside decomposition , there is no special corners of @xmath5 and any double lattice path is a lattice path . so in this case any two double lattice paths are non - separable at the point @xmath280 if and only if two lattice paths are intersecting at the point @xmath280 .
[ d : spe2 ] for all @xmath49 such that @xmath49 is the content of some special corner of @xmath5 , and for all @xmath3 , if @xmath278 has a point on line @xmath162 , we consider the unique _ @xmath49-point _ of @xmath278 , which is the ending point of the non - vertical step of @xmath278 between lines @xmath282 and @xmath162 , or the starting point of @xmath278 on line @xmath162 . if @xmath283 , the @xmath49-point @xmath284 of @xmath279 is above the one @xmath285 of @xmath278 , that is , @xmath286 , and there is no other @xmath49-points between @xmath285 and @xmath284 .
then we call @xmath287,c)$ ] a _ @xmath288-pair_. by definition [ d : spe2 ] it is clear that the number of @xmath288-pairs and the number of touchpoints of a non - crossing @xmath139-tuple of double lattice paths are the same , which are both equal to the number of common special corners of @xmath5 . for a triple @xmath289 of double lattice paths in figure
[ f : pic3 - 4 ] , and for @xmath290 , the @xmath291-points of @xmath292 and @xmath293 are @xmath198 and @xmath294 .
so @xmath295,-3)$ ] is a @xmath288-pair .
similarly , the @xmath296-points of @xmath292 and @xmath293 are @xmath297 and @xmath298 , as well as the @xmath256-points of @xmath299 and @xmath292 are @xmath300 and @xmath87 .
consequently the triple of double lattice paths @xmath301 contains three @xmath288-pairs , which are @xmath295,-3)$ ] , @xmath302,0)$ ] and @xmath302,3)$ ] . .
[ f : pic3 - 4 ] ] we define that @xmath303 $ ] is constructed from @xmath304 by the following steps : * remove the vertical steps on lines @xmath162 and @xmath163 ; * shift the non - vertical step ending at @xmath305 to the non - vertical step ending at @xmath306 ; * add vertical steps on lines @xmath162 and @xmath163 to connect with the new non - vertical step .
it should be noted that @xmath303 $ ] may not be a double lattice path .
[ d : rep ] for all @xmath288-pairs @xmath287,c)$ ] of a @xmath139-tuple @xmath281 ( see ( [ e : good3 ] ) ) of double lattice paths , suppose that @xmath307 are the @xmath49-points of @xmath279 and @xmath278 , the diagonal step of @xmath279 ends at @xmath306 and the horizontal step of @xmath278 ends at @xmath305 .
then we say that @xmath278 and @xmath279 are _ non - separable at a @xmath288-pair @xmath287,c)$ ] _ if and only if neither @xmath308 $ ] nor @xmath309 $ ] is a double lattice path . for a @xmath139-tuple @xmath281 ( see ( [ e : good3 ] ) ) of double lattice paths , we say that @xmath281 is non - separable at a @xmath288-pair if there exist two double lattice paths in @xmath281 such that they are non - separable at a @xmath288-pair .
otherwise we say that @xmath281 is separable at all @xmath288-pairs .
[ d : sdlp ] for a @xmath139-tuple @xmath281 ( see ( [ e : good3 ] ) ) of double lattice paths , we say that @xmath281 is _ separable _ if and only if @xmath281 is neither non - separable at any single point nor non - separable at any @xmath288-pair . the triple @xmath310 of double lattice paths in figure
[ f : pic3 - 4 ] is separable . for all @xmath49 such that
@xmath311 , any two double lattice paths from @xmath281 are not intersecting on line @xmath162 . for the @xmath288-pair @xmath295,-3)$ ] , we find that between lines @xmath312 and @xmath313 , the diagonal step of @xmath292 ends at @xmath314 , the horizontal step of @xmath293 ends at @xmath315 , and @xmath316 $ ] is a double lattice path . similarly , @xmath317 $ ] and @xmath318 $ ] are double lattice paths .
[ p : nonspr ] under the assumption of theorem [ t : thick ] , given any fixed total order of all points in the @xmath319-dimensional @xmath320 grid , there is a bijection between all separable @xmath139-tuples of double lattice paths with @xmath4 distinct @xmath288-pairs and all pairs @xmath321 where @xmath322 is a sequence of @xmath4 positive integers and @xmath323 is a non - crossing @xmath139-tuple of double lattice paths with @xmath4 distinct touchpoints . from lemma [ l : compatible ] we know that if @xmath323 is a non - crossing @xmath139-tuple ( [ e : good ] ) of double lattice paths , then one gets @xmath324 where @xmath325 for every @xmath3 .
first we establish that all pairs @xmath321 where @xmath322 is a sequence of @xmath4 positive integers and @xmath323 is a non - crossing @xmath139-tuple ( [ e : good2 ] ) of double lattice paths with @xmath4 distinct touchpoints , are in bijection with all separable @xmath139-tuples @xmath326 of double lattice paths with @xmath4 distinct @xmath288-pairs where @xmath327 for every @xmath3 .
that is , to prove the map @xmath328 is a bijection .
second , we prove that for any separable @xmath139-tuple ( [ e : good3 ] ) of double lattice paths , @xmath225 must be the identity permutation .
given such a pair @xmath321 , by assumption all double lattice paths @xmath216 have @xmath4 distinct touchpoints , suppose that @xmath329 is the coordinate of the @xmath3-th touchpoint with respect to any total order of all points in the @xmath319-dimensional @xmath320 grid .
then for all @xmath3 such that @xmath330 , assume that the diagonal step of @xmath331 and the horizontal step of @xmath332 intersect at the point @xmath329 , we shall insert @xmath333 to the double lattice paths according to the following steps : 1 .
if @xmath334 $ ] is a double lattice path , then we replace the non - vertical steps between lines @xmath335 and @xmath336 , together with the vertical steps on lines @xmath335 and @xmath336 of @xmath331 by the ones from @xmath334 $ ] ; 2 .
otherwise , @xmath337 and we replace the non - vertical steps between lines @xmath335 and @xmath336 , together with the vertical steps on lines @xmath335 and @xmath336 of @xmath332 by the ones from @xmath338 $ ] .
we choose @xmath301 to be the double lattice path @xmath216 after inserting all integers @xmath339 to the @xmath139-tuple @xmath323 ( see ( [ e : good2 ] ) ) of double lattice paths .
so it suffices to prove the @xmath139-tuple @xmath281 ( see ( [ e : good4 ] ) ) of double lattice paths is separable .
we observe that for all @xmath49 such that neither @xmath49 nor @xmath165 is the content of some special corner of @xmath5 , all non - vertical steps of @xmath281 between lines @xmath162 and @xmath163 are the same as the ones of @xmath323 , so any two double lattice paths from @xmath281 are separable at any single point since @xmath323 is non - intersecting between lines @xmath162 and @xmath163 . in addition , we notice that all @xmath288-pairs of @xmath281 and all @xmath288-pairs of @xmath323 are the same .
so we claim that for all @xmath3 , @xmath340 and @xmath341 are separable at any @xmath288-pair @xmath287,b_i)$ ] . if not , by definition [ d : rep ] it would contradict the facts that the point @xmath329 is a touchpoint of @xmath331 and @xmath332 and @xmath323 is non - crossing .
conversely , for a separable @xmath139-tuple @xmath281 of double lattice paths and for all @xmath288-pairs @xmath342,b_i)$ ] of @xmath281 , if the horizontal step of @xmath341 ends at point @xmath343 and the diagonal step of @xmath340 ends at point @xmath344 .
since @xmath281 is separable , according to definition [ d : sdlp ] , one of @xmath345 $ ] and @xmath346 $ ] must be a double lattice path . 1 .
if @xmath346 $ ] is a double lattice path , then we replace all non - vertical steps between lines @xmath335 and @xmath336 , together with the vertical steps on lines @xmath335 and @xmath336 of @xmath340 by the ones of @xmath346 $ ] , and we set @xmath347 .
2 . otherwise ,
@xmath345 $ ] is a double lattice path , then we replace all non - vertical steps between lines @xmath335 and @xmath336 , together with the vertical steps on lines @xmath335 and @xmath336 of @xmath341 by the ones of @xmath345 $ ] , and we set @xmath348 . in this way we retrieve the non - crossing @xmath139-tuple @xmath323 of double lattice paths as well as a sequence @xmath322 of positive integers , so that the @xmath3-th touchpoint @xmath329 of @xmath323 corresponds to the @xmath288-pair @xmath287,b_i)$ ] of @xmath281 for every @xmath3 .
in fact , @xmath256-@xmath257 is the inverse process of @xmath101-@xmath102 .
we notice that for any @xmath209 , any separable @xmath139-tuple ( [ e : good3 ] ) of double lattice paths , after the above process , yields a non - crossing @xmath139-tuple of double lattice paths .
because of lemma [ l : compatible ] , @xmath225 in ( [ e : good3 ] ) must be the identity permutation and the proof is complete .
consider the outside nested decomposition @xmath84 in figure [ f : pic2 - 01 ] where @xmath5 has three common special corners @xmath349 .
given a pair @xmath350 where @xmath351 and @xmath323 is a non - crossing triple of double lattice paths given in figure [ f : pic2 - 6 ] .
the corresponding separable triple @xmath281 of double lattice paths is shown in figure [ f : pic3 - 4 ] .
for instance , when we insert @xmath352 to the triple @xmath323 of double lattice paths in figure [ f : pic2 - 6 ] . since the @xmath291-point of @xmath272 is above the one of @xmath353 , and @xmath354 $ ] is not a double lattice path because a down - vertical step on line @xmath313 precedes a horizontal step ; see condition @xmath256 of definition [ d : plattice ] and figure [ f : f2 ] .
so between and on lines @xmath312 and @xmath313 , @xmath293 has the same steps as in @xmath355 $ ] .
$ ] is not a double lattice path , while @xmath355 $ ] is a double lattice path .
[ f : f2 ] ] for any permutation @xmath356 , the _ inversion _ of @xmath225 is @xmath357 and we may interpret the determinant in theorem [ t : thick ] as @xmath358_{i=1}^g=\sum_{\pi\in s_g}(-1)^{\scriptsize{\mbox{inv}}(\pi)}\prod_{i=1}^g s_{\theta_i\#\theta_{\pi_i}}(x).\end{aligned}\ ] ] by lemma [ l : bij3 ] we know that all semistandard young tableaux from @xmath359 are in bijection with all double lattice paths in @xmath360 .
it follows that all @xmath139-tuples @xmath361 where @xmath362 is a semistandard young tableau of thickened shape @xmath363 , are in bijection with all @xmath139-tuples ( [ e : good3 ] ) of double lattice paths where @xmath364 . for every double lattice path @xmath278 in ( [ e : good3 ] ) , if two non - vertical steps end at the same point @xmath365 , we assign these two steps with a single weight @xmath366 . for every other horizontal step or diagonal step ,
we assign each step with a weight @xmath366 if the step ends at @xmath365 . for every vertical step ,
we assign it with weight @xmath24 .
furthermore , the weight of every double lattice path @xmath278 , denoted by @xmath367 , is the product of all weights on the steps of @xmath278 and we use @xmath368(x)$ ] to denote the generating function of all double lattice paths in @xmath369 , that is , @xmath368(x)$ ] is the sum of all weighted double lattice paths from @xmath369 .
the relation between these notations is @xmath370(x)\end{aligned}\ ] ] where the first sum runs over all @xmath139-tuples ( [ e : good3 ] ) of double lattice paths and the second sum runs over all double lattice paths @xmath278 from the set @xmath369 .
we recall that lemma [ l : bij3 ] implies @xmath368(x)=s_{\theta_{i}\#\theta_{\pi_i}}(x)$ ] .
if @xmath371 , then @xmath372(x)=1 $ ] since the only double lattice path from @xmath373 to @xmath148 has no non - vertical steps . if @xmath363 is undefined , then @xmath374(x)=0 $ ] since the set @xmath369 is undefined . together with ( [ e : gen ] ) and ( [ e : re ] ) , one obtains @xmath375_{i=1}^g&=
\sum_{\pi\in s_g}\sum_{\tilde{p}}(-1)^{\scriptsize{\mbox{inv}}(\pi)}\prod_{i=1}^g \cmcal{w}(\tilde{p}(u_{\pi_i},v_i))\end{aligned}\ ] ] which can be viewed as a generating function for all pairs @xmath376 where @xmath209 and @xmath281 is any @xmath139-tuple ( [ e : good3 ] ) of double lattice paths . from proposition
[ l : bij5 ] and proposition [ p : nonspr ] , it follows that the generating function for all pairs @xmath377 when @xmath281 is separable , equals the generating function for all pairs @xmath321 where @xmath323 is a non - crossing @xmath139-tuple ( [ e : good2 ] ) of double lattice paths , that is , @xmath378 so in order to prove theorem [ t : thick ] , it remains to find an involution on all pairs @xmath376 when @xmath281 is non - separable . from definition
[ d : sdlp ] it is clear that a @xmath139-tuple @xmath281 ( see ( [ e : good3 ] ) ) of double lattice paths is non - separable if and only if @xmath281 is non - separable at a point or at a @xmath288-pair . it should be noted that there is _ no _ common integer @xmath49 such that @xmath281 is non - separable at point @xmath280 and at a @xmath288-pair @xmath287,c)$ ] according to definition [ d : spe1 ] and [ d : rep ] .
so we consider the minimal integer @xmath379 such that * @xmath281 is non - separable at a point on line @xmath380 or is non - separable at a @xmath288-pair @xmath287,c^*)$ ] for some @xmath381 ; * @xmath281 is neither non - separable at any point @xmath280 nor at any @xmath288-pair @xmath287,\tilde{c})$ ] when @xmath382 .
we choose a _ minimum _ of any non - separable @xmath139-tuple @xmath281 ( see ( [ e : good3 ] ) ) of double lattice paths to be 1 .
the point @xmath383 if it is the first point on line @xmath380 from top to bottom such that @xmath281 is non - separable at the point @xmath383 ; 2 . the @xmath288-pair @xmath287,c^*)$ ] if @xmath384 are the first two @xmath379-points on line @xmath380 from top to bottom such that @xmath281 is non - separable at the @xmath288-pair @xmath287,c^*)$ ] .
we are now ready to construct the involution @xmath141 on all non - separable @xmath139-tuples @xmath281 ( see ( [ e : good3 ] ) ) of double lattice paths by distinguishing the cases when the minimum of @xmath281 is a single point @xmath383 or a @xmath288-pair @xmath287,c^*)$ ] . for each case
, we will express the involution @xmath141 as @xmath385 where @xmath386 and @xmath387 are two non - separable @xmath139-tuples of double lattice paths with @xmath388 for each case below , the involution @xmath141 has the following properties : 1 .
@xmath141 is weight - preserving , that is , @xmath389 2 .
@xmath141 is sign - reversing , that is , @xmath390 ; 3 .
@xmath141 is closed , that is , @xmath281 and @xmath391 belong to the same case . _
case @xmath24 _ : if the minimum of @xmath281 is the point @xmath383 , and among all double lattice paths that are passing the point @xmath383 , assume that @xmath278 and @xmath279 of @xmath281 are two double lattice paths whose indices @xmath3 and @xmath169 are the smallest and the second smallest . since neither @xmath379 nor @xmath392 is the content of some common special corner of @xmath5 , all steps of @xmath281 between lines @xmath380 and @xmath393 are all horizontal steps or all diagonal steps . by our choice of @xmath379 ,
all steps of @xmath278 between lines @xmath394 and @xmath380 are disjoint with the ones of @xmath279 . using the notations @xmath395 and @xmath396 to denote the segments of the double lattice path @xmath278 from @xmath373 to the point @xmath397 and from the point @xmath397 to @xmath148 ( similarly for @xmath279 )
, we may define the pair @xmath398 where @xmath399 as follows . for @xmath400 , @xmath401 , we set @xmath402 and @xmath403 we will show that @xmath404 is a double lattice path from @xmath405 to @xmath148 and @xmath406 is a double lattice path from @xmath407 to @xmath408 by discussing the ending points of the non - vertical steps between lines @xmath394 and @xmath393 . here , without loss of generality , we assume that the steps between lines @xmath394 and @xmath380 are horizontal , while the steps between lines @xmath380 and @xmath393 are diagonal .
suppose that the ending points of non - vertical steps from @xmath278 are the points @xmath409 and @xmath410 where @xmath411 , and the ones from @xmath279 are the points @xmath383 and @xmath412 where @xmath413 ; see figure [ f : c1 ] .
since @xmath278 and @xmath279 are intersecting at the point @xmath383 , one has @xmath414 , which implies @xmath415 .
so there is no single up - vertical step on line @xmath380 that is preceding the diagonal step in @xmath404 or @xmath406 .
this indicates that @xmath404 and @xmath406 are double lattice paths according to ( 4 ) in definition [ d : plattice ] .
furthermore , @xmath141 is closed within all non - separable @xmath139-tuples of double lattice paths that belong to case @xmath24 , because by construction @xmath281 and @xmath391 are non - separable at the same points and the same @xmath288-pairs . in particular , the minimum of @xmath391 is also the point @xmath383 .
see figure [ f : c1 ] . for case @xmath24
when the minimum of @xmath281 is a point @xmath383 ( marked by a black square ) and all integers represent the @xmath193-th coordinates of all ending points from the non - vertical steps .
[ f : c1 ] ] _ case @xmath319 _ : if the minimum of @xmath281 is the @xmath288-pair @xmath287,c^*)$ ] , we assume that @xmath82 are the smallest indices of @xmath416 satisfying @xmath417 are non - separable at the @xmath288-pair @xmath287,c^*)$ ] . by our choice of @xmath379 ,
all steps of @xmath278 between lines @xmath394 and @xmath380 are disjoint with the ones of @xmath279 .
consequently , suppose that between lines @xmath380 and @xmath418 , the diagonal step and the horizontal step of @xmath279 end at @xmath419 and the horizontal step and the diagonal step of @xmath278 end at @xmath420 furthermore , if there is a horizontal step and a diagonal step of @xmath279 ending on line @xmath393 and on line @xmath418 , we assume they end at @xmath421 it should be mentioned that such horizontal step and diagonal step are not contained in @xmath279 if the starting box or the ending box of @xmath83 is the common special corner of @xmath2 and @xmath83 . but for this situation the discussion on the involution @xmath141 follows analogously , so we focus on the case when the horizontal step ending at @xmath422 and the diagonal step @xmath423 are contained in @xmath279 .
likewise , if there is a diagonal step and a horizontal step of @xmath213 ending on line @xmath393 and on line @xmath418 , we assume that they end at @xmath424 see figure [ f : c0 ] where all integers represent the @xmath193-th coordinates of all ending points from the non - vertical steps . and
@xmath279 between lines @xmath394 and @xmath418 if the minimum of @xmath281 is the @xmath288-pair @xmath287,c^*)$ ] . [ f : c0 ] ] since @xmath278 and @xmath279 are double lattice paths , from definition [ d : plattice ] we find that @xmath425 , @xmath426 , @xmath427 and @xmath428 .
since neither @xmath429 $ ] nor @xmath430 $ ] is a double lattice path , then the integer @xmath431 satisfies @xmath432 or @xmath433 and the integer @xmath434 satisfies @xmath435 or @xmath436 .
so under the assumption @xmath286 , we shall consider the following disjoint sub - cases : * case @xmath437 : @xmath438 ; * case @xmath439 : @xmath432 and @xmath440 ; * case @xmath441 : @xmath442 and @xmath435 ; * case @xmath443 : @xmath442 and @xmath440 .
_ case @xmath437 _ : if @xmath438 , we use the notations @xmath444 and @xmath445 to denote the segments of the double lattice path @xmath278 from @xmath373 to all non - vertical steps ending on line @xmath380 and from all non - vertical steps of @xmath278 starting on line @xmath380 to @xmath148 ( similarly for @xmath279 ) . furthermore , @xmath446 is obtained by connecting two segments @xmath444 and @xmath447 with new vertical steps on line @xmath380 . here
we may define the pair @xmath398 as follows . for @xmath400 , @xmath401
, we set @xmath402 and @xmath448 where @xmath406 is a double lattice path from @xmath449 to @xmath408 and @xmath404 is a double lattice path from @xmath450 to @xmath148 .
this is guaranteed by the relations @xmath451 and @xmath452 .
furthermore , @xmath141 is closed within all non - separable @xmath139-tuples of double lattice paths that belong to case @xmath437 , because @xmath391 also belongs to case @xmath437 since @xmath453 , and by construction @xmath387 are non - separable at the same points and at the same @xmath288-pairs . in particular
, the minimum of @xmath391 is also the @xmath288-pair @xmath287,c^*)$ ] .
see figure [ f : c2 ] for an example of case @xmath437 . for case @xmath437
when the minimal of @xmath281 is the @xmath288-pair @xmath287,c^*)$ ] and @xmath438 .
[ f : c2 ] ] _ case @xmath439 _ : if @xmath432 and @xmath440 , we use the notation @xmath454 to denote the segment of the double lattice path @xmath278 from @xmath373 to its @xmath379-point , as well as the step from the @xmath379-point to the ending point of the horizontal step between lines @xmath380 and @xmath393 .
we use the notation @xmath455 to denote the segment of the double lattice path @xmath278 that is complement to the segment @xmath454 except the down - vertical steps on line @xmath380 and @xmath393 ; similarly for @xmath279 .
furthermore , @xmath456 is obtained by connecting two segments @xmath454 and @xmath457 by new down - vertical steps on lines @xmath380 and @xmath393 . here
we may define the pair @xmath398 where @xmath458 as follows . for @xmath400 , @xmath401
, we set @xmath402 and @xmath459 where @xmath406 is a double lattice path from @xmath449 to @xmath408 and @xmath404 is a double lattice path from @xmath450 to @xmath148 .
this is guaranteed by the assumption @xmath432 and @xmath440 . to be precise
, @xmath460 holds because @xmath461 and @xmath286 ; @xmath462 holds because @xmath463 and @xmath464 ; @xmath432 and @xmath436 hold because of the assumption .
furthermore , @xmath141 is closed within all non - separable @xmath139-tuples of double lattice paths that belong to case @xmath439 , that is , @xmath391 also belongs to case @xmath439 since @xmath461 and @xmath465 , and by construction @xmath387 are non - separable at the same points and the same @xmath288-pairs .
in particular , the minimum of @xmath391 is also the @xmath288-pair @xmath287,c^*)$ ] .
see figure [ f : c3 ] for an example of case @xmath439 . for case @xmath439 when the minimal of @xmath281 is the @xmath288-pair @xmath287,c^*)$ ] such that @xmath432 and @xmath440 .
[ f : c3 ] ] _ case @xmath441 _ : if @xmath442 and @xmath435 , we use the notation @xmath466 to denote the segment of the double lattice path @xmath278 from @xmath373 to its @xmath379-point , together with the step from the @xmath379-point to the ending point of the diagonal step between lines @xmath380 and @xmath393 .
we use the notation @xmath467 to denote the segment of the double lattice path @xmath278 that is complement to the segment @xmath466 except the up - vertical steps on lines @xmath380 and @xmath393 ; similarly for @xmath279 .
furthermore , @xmath468 is obtained by connecting two segments @xmath466 and @xmath469 by new up - vertical steps on lines @xmath380 and @xmath393 . here
we may define the pair @xmath398 where @xmath458 as follows . for @xmath400 , @xmath401
, we set @xmath402 and @xmath470 where @xmath406 is a double lattice path from @xmath449 to @xmath408 and @xmath404 is a double lattice path from @xmath450 to @xmath148 .
this is guaranteed by the assumption @xmath442 and @xmath435 . to be precise
, @xmath471 holds because @xmath472 and @xmath473 ; @xmath474 holds because @xmath475 ; @xmath435 and @xmath433 hold because of the assumption .
see figure [ f : c4 ] for an example of case @xmath441 . for case @xmath441 when the minimal of @xmath281 is the @xmath288-pair @xmath287,c^*)$ ] such that @xmath442 and @xmath435 .
[ f : c4 ] ] furthermore , @xmath141 is closed within all non - separable @xmath139-tuples of double lattice paths that belong to case @xmath441 , because @xmath391 also belongs to case @xmath441 since @xmath476 and @xmath477 , and by construction @xmath387 are non - separable at the same points and the same @xmath288-pairs . in particular , the minimum of @xmath391 is also the @xmath288-pair @xmath287,c^*)$ ] .
see figure [ f : c4 ] for an example of case @xmath441 .
_ case @xmath443 _ : if @xmath442 and @xmath440 , then we may define the pair @xmath398 where @xmath458 as follows . for @xmath400 , @xmath401 , we set @xmath402 and @xmath478 where @xmath406 is a double lattice path from @xmath449 to @xmath408 and @xmath404 is a double lattice path from @xmath450 to @xmath148
. this is guaranteed by the assumption @xmath442 and @xmath440 . to be precise , @xmath479 holds because @xmath480 ; and @xmath481 holds because @xmath482 .
furthermore , @xmath141 is closed within all non - separable @xmath139-tuples of double lattice paths that belong to case @xmath443 , because @xmath391 also belongs to case @xmath443 since @xmath476 and @xmath465 , and by construction @xmath387 are non - separable at the same points and the same @xmath288-pairs . in particular , the minimum of @xmath391 is also the @xmath288-pair @xmath287,c^*)$ ] .
see figure [ f : c5 ] for an example of case @xmath443 . for case @xmath443 when the minimal of @xmath281 is the @xmath288-pair @xmath287,c^*)$ ] such that @xmath442 and @xmath440 .
[ f : c5 ] ] for each case ( case @xmath24 or case @xmath437-@xmath443 ) , it is clear that @xmath483 is an involution which preserves the weight of the double lattice path and changes the inversion of the permutation by @xmath24 .
_ proof of theorem [ t : thick]_. from ( [ e : gen1 ] ) and the involution @xmath484 in subsection [ ss : invo ] , we find that only the generating function for all pairs @xmath377 where @xmath281 is any separable @xmath139-tuple of double lattice paths , is remained on the right hand side of ( [ e : gen1 ] )
. in combination of ( [ e : nonspe ] ) , ( [ e : hg ] ) follows immediately .
_ proof of corollary [ c : thick2]_. we refer the readers to chapter @xmath485 of @xcite for a full description of the exponential specialization .
let @xmath486f$ ] denote the coefficient of @xmath487 in @xmath175 , the exponential specialization ex of the symmetric function @xmath175 is defined as @xmath488f\frac{t^n}{n!}\end{aligned}\ ] ] and @xmath489 .
let @xmath490 and @xmath491 , then one has @xmath492 consequently ( [ e : hg1 ] ) follows directly after we apply @xmath493 on both sides of ( [ e : hg ] ) .
we will count the number of @xmath10-strip tableaux by applying corollary [ c : thick2 ] .
it should be pointed out that the enumeration of @xmath494-strip tableaux is a direct consequence of theorem [ t : hg ] ; see @xcite . in @xcite ,
morales , pak and panova also found that the enumeration of @xmath494-strip tableaux can be simplified by applying lascoux - pragacz s theorem @xcite , or more generally , hamel and goulden s theorem ( theorem [ t : hg ] ) .
baryshnikov and romik @xcite counted the number of @xmath10-strip tableaux as a generalization of the classical formula from d. andr @xcite on the number of up - down permutations . an _ @xmath10-strip diagram _
@xmath495 contains three parts : head @xmath496 , tail @xmath497 and body .
the body of an @xmath10-strip diagram consists of an elongated hexagonal shape with @xmath31 columns , where the numbers of boxes in the @xmath31 columns are @xmath498 the first ( resp .
last ) @xmath499 columns forms a standard diagram and the columns where each contains @xmath10 boxes forms a skew diagram of shape @xmath500 the _ head _ @xmath496 and _ tail _ @xmath497 are standard diagrams of length at most @xmath499 that are rotated and connected to the body by leaning against the sides of the body .
the empty partition @xmath296 is always denoted by @xmath501 and _ an @xmath10-strip tableau _ is a standard young tableau of the @xmath10-strip shape .
our definition of @xmath10-strip diagrams is slightly different to the one in @xcite because @xcite contains a minor typo on the number of boxes in the leftmost and the rightmost columns of any @xmath10-strip diagram .
our notation @xmath495 is the notation @xmath502 in @xcite and we find it more convenient to use @xmath495 to represent some @xmath10-strip diagrams for small @xmath10
. see figure [ f:1 ] for an example of @xmath503-strip diagram with head partition @xmath496 and tail partition @xmath497 from @xcite .
the standard diagrams of @xmath504 are rotated and attached to the body of the @xmath503-strip diagram .
-strip diagram @xmath505 , a @xmath319-strip diagram @xmath506 and a @xmath507-strip diagram @xmath508 ( left , middle , right ) .
[ f:1 ] ] to avoid confusion , we adopt the definitions and notations of euler numbers and tangent numbers from @xcite .
a permutation @xmath509 is called an _ up - down permutation _ if @xmath510 .
it is well - known that the exponential generating function of the numbers @xmath511 of up - down permutations of @xmath512=\{1,2,\ldots , n\}$ ] is @xmath513 this is also called andr s theorem @xcite , which connects the numbers @xmath511 with the _ euler numbers _ @xmath514 and _ tangent numbers _ @xmath515 by the taylor expansions of @xmath516 and @xmath517 , that is , @xmath518 this implies that @xmath519 it should be mentioned that euler numbers are defined differently in some literature @xcite .
it is clear that an up - down permutation of @xmath520 $ ] can be identified as a @xmath319-strip tableau of shape @xmath521 . by thickening the @xmath319-strip diagram , baryshnikov and
romik @xcite introduced the @xmath10-strip diagram and enumerated the @xmath10-strip tableaux via transfer operators , which proved that the determinant to count @xmath10-strip tableaux has order @xmath499 .
this is certainly to their advantage that baryshnikov and romik s determinant for @xmath72-strip diagrams is much simpler than the one directly from hamel and goulden s theorem ( theorem [ t : hg ] ) .
we next recall the baryshnikov and romik s determinant for the @xmath10-strip tableaux .
we define the numbers @xmath522 and denote the head young diagram by @xmath523 and the tail young diagram by @xmath524 where @xmath525 . for any non - negative integers @xmath526 ,
we denote by @xmath527 and @xmath528 the number of @xmath319-strip tableaux of shape @xmath529 and the number of @xmath507-strip tableaux of shape @xmath530 where the empty partition @xmath296 is denoted by @xmath501 .
in other words , @xmath531 in particular , @xmath532 and some values of @xmath528 are given in theorem [ c:345 ] .
note that @xmath533 holds for any non - negative integers @xmath534 and @xmath182 .
this is true because for any standard young tableau @xmath50 of shape @xmath529 , if we replace every entry @xmath535 of @xmath50 by @xmath536 and flip the diagram @xmath529 upside - down and reverse it left - to - right , we obtain a standard young tableau of shape @xmath537 . similarly @xmath538 holds for any non - negative integers @xmath534 and @xmath182 .
furthermore , we define the numbers @xmath539 and @xmath540 as below : @xmath541 our notation @xmath527 is @xmath542 in @xcite and we need the parameters @xmath526 to describe the thickened strips later .
for the readers convenience , we should mention that the left @xmath319-strip diagram in figure @xmath543 of @xcite should be the middle one in figure [ f:1 ] .
baryshnikov and romik proved that [ t:1 ] let @xmath544 and @xmath545 for @xmath546 .
then the number of standard young tableaux of shape @xmath495 is given by @xmath547_{i , j=1}^k \,\mbox { if } \
, m=2k \,\mbox { or by}\\ \label{e : odd}f^{\cmcal{d}_m(\tilde{\lambda};\tilde{\mu})}&=(-1)^{\binom{k}{2}}\vert \cmcal{d}_m(\tilde{\lambda};\tilde{\mu})\vert!\det[y_{2n - m+1}(l_i , m_j)]_{i , j=1}^k \,\mbox { if } \ , m=2k+1.\end{aligned}\ ] ] theorem [ t:1 ] is a combination of theorem 4 and 5 in @xcite .
here we use the combinatorial interpretations of @xmath548 to introduce the numbers @xmath549 , whose expressions in terms of the numbers @xmath550 , @xmath551 and @xmath552 can be derived by the recursions of @xmath527 and @xmath528 .
here we omit the computational details .
baryshnikov and romik @xcite also presented some explicit formulas for small @xmath10
. we will establish theorem [ c:345 ] by decomposing @xmath507-strip tableaux directly and by choosing two different outside nested decompositions respectively for @xmath553-strip tableaux .
[ c:345 ] some numbers of @xmath507-strip tableaux are @xmath554 some numbers of @xmath543-strip tableaux are @xmath555,\\ \label{e:4strip2}f^{\cmcal{d}_4((1);(1))}&=\binom{4n}{2n}e_{2n}^2 -\binom{4n}{2n-2}e_{2n-2}e_{2n+2}=(4n)!\det\left[\begin{array}{ll } \bar{a}_{2n } & \bar{a}_{2n+2}\\ \bar{a}_{2n-2 } & \bar{a}_{2n } \end{array}\right],\end{aligned}\ ] ] and the number of @xmath556-strip tableaux without head and tail is @xmath557.\end{aligned}\ ] ] we count the number @xmath558 of @xmath494-strip tableaux by choosing an outside decomposition @xmath559 of the @xmath494-strip diagram @xmath560 , which is a special outside nested decomposition without common special corners . given a @xmath494-strip diagram @xmath560 , we can peel this diagram off into successive maximal outer strips @xmath561 beginning from the outside ; see the left one in figure [ f : pic-6 ] . that we choose for the @xmath503-strip diagram @xmath562 ( left ) and the outside
nested decomposition that we will not choose for the @xmath485-strip diagram @xmath563 ( right ) .
[ f : pic-6 ] ] we recall the numbers @xmath544 and @xmath545 , for @xmath564 . in the outside decomposition @xmath67 , every strip @xmath68 is a @xmath319-strip of @xmath565 columns , with head partition @xmath566 and tail partition @xmath567 .
the number of such tableaux are denoted by @xmath568 , that is , @xmath569 . by definition [ d : operator1 ] , we see that the thickened cutting strip @xmath570 is a @xmath319-strip of @xmath565 columns , with head partition @xmath571 and tail partition @xmath572 .
so it follows that @xmath15 is a @xmath319-strip diagram with @xmath565 columns , with head partition @xmath566 and with tail partition @xmath573 . consequently , @xmath574 . by corollary
[ c : thick2 ] we know that the number @xmath558 of standard young tableaux of @xmath494-strip shape with @xmath31 columns , is expressed as a determinant where the @xmath575-th entry is @xmath576 .
that is to say , @xmath577_{i , j=1}^{k}\\ & = ( -1)^{\binom{k}{2}}\vert\cmcal{d}_{2k}(\tilde{\lambda};\tilde{\mu } ) \vert!\det[x_{2n-2k+1}(l_i , m_j)]_{i , j=1}^{k},\end{aligned}\ ] ] which is ( [ e : even ] ) .
we observe that any outside decomposition of @xmath72-strip diagram will not reduce the order of the jacobi - trudi determinant in theorem [ t : jacobi - trudi ] because the minimal number of strips contained in any outside decomposition is exactly the number of columns in any @xmath72-strip diagram @xmath578 ; see the outside decomposition of the @xmath485-strip diagram @xmath563 in figure [ f : pic-6 ] . given a @xmath72-strip diagram @xmath578
, we can peel this diagram off into successive maximal outer thickened strips @xmath579 beginning from the outside ; see figure [ f : pic-81 ] . of the @xmath485-strip diagram @xmath563 where each common special corner is marked by a black square in the thickened strip .
[ f : pic-81 ] ] consider the outside nested decomposition @xmath580 , every thickened strip @xmath2 is a @xmath507-strip of @xmath565 columns , with head partition @xmath566 and tail partition @xmath567 . the number of such tableaux are denoted by @xmath581 , that is , @xmath582 . by definition
[ d : operator1 ] , we see that the thickened cutting strip @xmath111 is a @xmath507-strip of @xmath565 columns , with head partition @xmath571 and tail partition @xmath572 .
so it follows that @xmath19 is a @xmath507-strip diagram with @xmath565 columns , with head partition @xmath566 and with tail partition @xmath573 . consequently , @xmath583 . by corollary
[ c : thick2 ] we know that the number @xmath584 of standard young tableaux of @xmath72-strip shape with @xmath31 columns , is expressed as a determinant where the @xmath575-th entry is @xmath585 .
that is to say , @xmath586_{i , j=1}^{k}\\ & = ( -1)^{\binom{k}{2}}\vert\cmcal{d}_{2k+1}(\tilde{\lambda};\tilde{\mu } ) \vert!\det[y_{2n-2k}(l_i , m_j)]_{i , j=1}^{k},\end{aligned}\ ] ] which is ( [ e : odd ] ) .
here we need the parameter @xmath31 to describe the number of columns when we decompose the @xmath507-strip diagrams .
so we set @xmath587 and let @xmath588 denote a @xmath507-strip diagram which is obtained by adding a new box to the right of the topmost and rightmost box of @xmath589 ; see figure [ f:3strip ] .
first we have two simple observations .
-strip diagrams [ f:3strip ] ] [ l : funlem ] the numbers @xmath590 , @xmath591 and @xmath592 satisfy @xmath593 let @xmath594 denote the set of all standard young tableaux of shape @xmath595 .
then , in order to prove ( [ e : div1 ] ) , we will establish the bijection @xmath596\times \cmcal{t}_{\cmcal{d}_{3n-2}}\rightarrow \cmcal{t}_{\cmcal{d}_{3n-1}}\dot{\cup } \cmcal{t}_{\cmcal{d}_{3n-1}^*}.\end{aligned}\ ] ] given a pair @xmath597 where @xmath598 $ ] and @xmath50 is a standard young tableau of shape @xmath599 with entries from the set @xmath600-\{i\}$ ] .
suppose that the rightmost and topmost box @xmath44 of @xmath50 has entry @xmath182 . if @xmath601 , then we put a box with entry @xmath3 on the top of box @xmath44 , which gives us a standard young tableau of shape @xmath602 with entries from @xmath24 to @xmath603 .
otherwise we put a box with entry @xmath3 to the right of box @xmath44 , which , after transposing the rows into columns , is a standard young tableau of shape @xmath604 with entries from @xmath24 to @xmath603 .
it is clear that this procedure is reversible , so the bijection ( [ e : in1 ] ) follows .
furthermore , it holds that @xmath605 since for any standard young tableau of shape @xmath602 , if we replace every entry @xmath182 by @xmath606 and flip the diagram @xmath602 upside - down and reverse it left - to - right , we obtain a standard young tableau of shape @xmath604 . in combination of ( [ e : in1 ] )
, it follows that ( [ e : div1 ] ) is true . in order to prove ( [ e : retf ] ) , we next establish the bijection @xmath607\times \cmcal{t}_{\cmcal{d}_{3n-1}^*}\rightarrow \cmcal{t}_{\cmcal{d}_{3n}}\dot{\cup } \cmcal{t}_{\cmcal{c}_{3n}}\end{aligned}\ ] ] which is analogous to ( [ e : in1 ] ) . given a pair @xmath597 where @xmath608 $ ] and
@xmath50 is a standard young tableau of shape @xmath604 with entries from the set @xmath609-\{i\}$ ] .
suppose that the rightmost and topmost box @xmath44 of @xmath50 has entry @xmath182 . if @xmath601 , then we put a box with entry @xmath3 on the top of box @xmath44 , which gives us a standard young tableau of shape @xmath610 with entries from @xmath24 to @xmath611 .
otherwise we put a box with entry @xmath3 to the right of box @xmath44 , which is a standard young tableau of shape @xmath588 with entries from @xmath24 to @xmath611 .
this implies that ( [ e : in2 ] ) is a bijection , thus in view of @xmath612 , ( [ e : retf ] ) holds . by lemma [ l : funlem ]
it suffices to count the numbers @xmath590 and @xmath613 .
consider the boxes @xmath614 of the @xmath507-strip diagram @xmath599 , one of these boxes has the minimal entry @xmath24 for any standard young tableau from @xmath615 .
let @xmath616 be the @xmath507-strip diagram @xmath599 after removing the box @xmath617 .
then we have [ l:31 ] for @xmath618 , the numbers @xmath619 satisfy @xmath620 let @xmath621 denote the set of all @xmath622-subsets of @xmath623 $ ] , we aim to construct the bijection @xmath624\times \cmcal{t}_{\cmcal{d}_{3n-2,i}}\rightarrow \cmcal{t}_{\cmcal{d}_{3n-2}}\dot{\cup}(\cmcal{s } \times\cmcal{t}_{\cmcal{d}_{3i-1}^*}\times \cmcal{t}_{\cmcal{d}_{3n-3i-1}^*}),\end{aligned}\ ] ] from which ( [ e:31 ] ) follows immediately .
given a pair @xmath625 where @xmath626 $ ] and @xmath50 is a standard young tableau of shape @xmath616 with entries from the set @xmath623-\{r\}$ ] .
suppose that the entries of box @xmath627 and box @xmath628 are @xmath629 and @xmath630 in @xmath50 , we set @xmath631 . if @xmath632 , then we add a box @xmath617 with entry @xmath4 to @xmath50 , which is a standard young tableau of shape @xmath599 . if @xmath633 , then we consider a segment of @xmath50 from the starting box of @xmath616 to box @xmath627 and we add a box with entry @xmath4 to the right of box @xmath627 , which , after transposing the rows into columns , leads to a standard young tableau of shape @xmath634 with entries coming from a @xmath635-subset @xmath636 of @xmath623 $ ] .
moreover , the segment of @xmath50 from box @xmath637 to the ending box of @xmath616 , is a standard young tableau of shape @xmath638 with entries coming from the complement set @xmath639 of @xmath636 with respect to @xmath623 $ ] .
if @xmath640 , then we consider a segment of @xmath50 from box @xmath628 to the ending box of @xmath616 and we add a box with entry @xmath4 right below the box @xmath628 , which leads to a standard young tableau of shape @xmath638 with entries coming from a @xmath622-subset @xmath641 of @xmath623 $ ] . moreover , the segment of @xmath50 from the starting box of @xmath616 to box @xmath637 , which , after transposing the rows into columns , is a standard young tableau of shape @xmath634 with entries coming from the complement set @xmath642 of @xmath641 with respect to @xmath623 $ ] . conversely , given a standard young tableau @xmath643 of shape @xmath599 , we set @xmath4 to be the entry of box @xmath617 in @xmath643 and after we remove box @xmath617 from @xmath643 , we obtain a standard young tableau of shape @xmath616 . given a triple @xmath644 where @xmath645 is a standard young tableau of shape @xmath638 with entries from @xmath646 , and @xmath647 is a standard young tableau of shape @xmath634 with entries from the complement set @xmath648 .
suppose that the entry of box @xmath649 in @xmath645 is @xmath650 and the entry of box @xmath651 in @xmath647 is @xmath652 , if @xmath653 , we remove the box @xmath654 of @xmath647 , then transpose it from columns into rows and put the box with entry @xmath652 to the left of box @xmath655 of @xmath645 .
this gives us a standard young tableau of shape @xmath616 such that the entry of box @xmath628 is larger than the one of box @xmath627 and we choose @xmath4 to be the entry of box @xmath654 of @xmath647 , so that @xmath656 . if @xmath657 , we transpose @xmath647 from columns into rows , then put its rightmost and topmost box right below the box with entry @xmath650 after we remove the box @xmath655 of @xmath645 .
this gives us a standard young tableau of shape @xmath616 such that the entry of box @xmath628 is smaller than the one of box @xmath627 and we choose @xmath4 to be the entry of box @xmath655 of @xmath645 , so that @xmath658 .
since all cases are disjoint and cover all possible scenarios , ( [ e : b1 ] ) is a bijection . in view of @xmath605 , ( [ e:31 ] ) follows . for @xmath659
, we consider the pairs @xmath660 and @xmath661 .
since @xmath662 , we put a box with entry @xmath543 to @xmath645 . since @xmath663
, we separate @xmath647 of shape @xmath664 into two standard young tableaux of shapes @xmath665 and @xmath666 . ) in lemma [ l:31 ] .
[ f:3strip ] ] let @xmath667 be the @xmath507-strip diagram @xmath588 after removing the box @xmath617 , we can decompose the skew diagram @xmath588 in exactly the same way .
so we omit the proof of lemma [ l:32 ] .
[ l:32 ] for @xmath618 , the numbers @xmath668 satisfy @xmath669 with the help of recursions ( [ e:31 ] ) and ( [ e:32 ] ) , we could use the generating function approach to finally derive the numbers of @xmath507-strip tableaux .
_ proof of ( [ e:3strip])-([e:3strip2])_. summing ( [ e:31 ] ) and ( [ e:32 ] ) over all @xmath3 gives us @xmath670 we can translate the recursions ( [ e : rec1 ] ) and ( [ e : rec2 ] ) into two identities of exponential generating functions .
we define that @xmath671 from ( [ e : div1 ] ) we have @xmath672 .
furthermore , ( [ e : rec1 ] ) is equivalent to @xmath673 where @xmath674 .
this leads to a unique solution , @xmath675 .
together with the exponential generating function for @xmath676 ; see ( [ e : andre ] ) , we can prove ( [ e:3strip ] ) and ( [ e:3strip1 ] ) . similarly , ( [ e : rec2 ] ) is equivalent to @xmath677 where @xmath678 .
this yields a unique solution @xmath679 from which we can derive the numbers @xmath613 by expanding @xmath680 , i.e. , @xmath681 together with ( [ e : retf ] ) and ( [ e:3strip1 ] ) , ( [ e:3strip2 ] ) is proved . for the @xmath543-strip diagram @xmath682
, we choose another outside decomposition @xmath683 , which is slightly different to the one @xmath559 for the @xmath494-strip diagram @xmath560 . the benefit to make such
a slight change is that the determinant in ( [ e : even ] ) is further simplified , which only has the numbers @xmath552 as entries .
we call a @xmath319-strip _ a zig - zag strip _ if the corresponding standard young tableaux are in bijection with up - down permutations . for instance
, the @xmath319-strip diagram @xmath521 is a zig - zag strip and all strips in figure [ f : pic-82 ] are zig - zag strips . of the @xmath543-strip diagrams @xmath682 ( left ) and @xmath684 ( right ) .
[ f : pic-82 ] ] for the @xmath543-strip diagrams @xmath682 , we can peel each diagram off into successive maximal zig - zag outer strips @xmath685 beginning from the outside .
see the left one in figure [ f : pic-82 ] .
it is clear that the numbers of any zig - zag strip @xmath686 or @xmath687 are euler numbers or tangent numbers . by definition [ d : operator1 ]
we find that @xmath688 by corollary [ c : thick2 ] , we can prove ( [ e:4strip1 ] ) and ( [ e:4strip2 ] ) follows in the same way . for the @xmath556-strip diagram @xmath689
, we choose another outside nested decomposition @xmath690 , which is slightly different to the one @xmath580 for the @xmath72-strip diagram @xmath578 .
the benefit to make such a slight change is that the determinant in ( [ e : odd ] ) is further simplified , which only has the numbers @xmath550 and @xmath691 as entries .
we call a @xmath507-strip _ a zig - zag thickened strip _ if the number of such @xmath507-strip tableaux is one of the numbers @xmath590 , @xmath591 , @xmath592 and @xmath613 . for instance
, two thickened strips in figure [ f : pic-83 ] are zig - zag thickened strips .
it is clear that the numbers of the zig - zag thickened strips @xmath692 or @xmath693 are @xmath694 . by definition [ d : operator1 ]
we find that @xmath695 by corollary [ c : thick2 ] , we can prove that @xmath696 combining ( [ e:3strip2 ] ) and ( [ e : c ] ) , we can conclude that ( [ e:5strip ] ) is true . of the @xmath556-strip diagram @xmath689 .
[ f : pic-83 ] ]
the author would like to thank angle hamel , igor pak , dan romik and john stembridge for their very helpful suggestions and encouragements and would like to give special thanks to the joint seminar arbeitsgemeinschaft diskrete mathematik for their valuable feedback .
this work was partially done during my stay in the ag algorithm and complexity , technische universitt kaiserslautern , germany .
the author also thanks markus nebel , sebastian wild and raphael reitzig for their kind help and support . , de functionibus alternantibus earumque divisione per production e differentiis elementorum conflatum , j. reine angew .
, 22 , 360 - 371 , 1841 .
also in mathematische werke , vol .
3 , chelsea , new york , 439 - 452 , 1969 . | in this paper we describe the _ thickened strips _ and the _ outside nested decompositions _ of any skew shape @xmath0 . for any such decomposition @xmath1 of the skew shape @xmath0 where @xmath2 is a thickened strip for every @xmath3 , if @xmath4 is the number of boxes that are contained in any two distinct thickened strips of @xmath5 , we establish a determinantal formula of the function @xmath6 with the schur functions of thickened strips as entries , where @xmath7 is the schur function of the skew shape @xmath0 and @xmath8 is the power sum symmetric function index by the partition @xmath9
. this generalizes hamel and goulden s theorem on the outside decompositions of the skew shape @xmath0 ( _ planar decompositions of tableaux and schur function determinants _ , europ .
j. combinatorics , 16 , 461 - 477 , 1995 ) . as an application of our theorem
, we derive the number of @xmath10-strip tableaux which was first counted by baryshnikov and romik ( _ enumeration formulas for young tableaux in a diagonal strip _ , israel j. math , 178 , 157 - 186 , 2010 ) via extending the transfer operator approach due to elkies . |
this work is dedicated to the memory of professor sergio del campo .
cc acknowledges partial support by conacyt grant cb-2012 - 177519-f .
cc also acknowledges conacyt grant i0010 - 2014 - 02 estancias internacionales-233618-c .
this work was partially supported by sni ( mxico ) .
vhc acknowledges partial support by grant diuv 50/2013 .
one of the authors ( vhc ) would like to thank the warm hospitality of the facultad de fsica at the universidad veracruzana in xalapa , where part of this work was carried out .
rh was supported by the comisin nacional de ciencias y tecnologa of chile through fondecyt grant n@xmath66 1130628 and di - pucv n@xmath66 123.724 . | we present a new cosmological model that mimics the lambda cold dark matter by using a stealth field .
this kind of field is characterized as not coupling directly to gravity ; however , it is connected to the underlying matter content of the universe model .
as is known , stealth fields do not back - react on the space - time ; however , their mimicry skills show how this field and its self - interaction potential determines the cosmic evolution .
we show the study of the simplest model that can be developed with the stealth field . _
i. introduction_. precise astronomical measurements of the universe indicate that nearly @xmath0 of its content is in the form of dark matter ( dm ) , the key ingredient necessary to explain large scale structure formation , and @xmath1 dark energy ( de ) , the unknown component driving the recent cosmic acceleration . in this context , the best model to describe almost all
the observational data is a mixture of elements from the standard cosmological model plus a cosmological constant , the so - called `` concordance '' @xmath2cdm model .
although successful in fitting the observational data , from a theoretical point of view the model seems too arbitrary .
first , there is no clue about where this cosmological constant came from , and with it , why its value is so close to the critical energy density , and second , why we live in a very special epoch where the contributions from dm and de are of the same order of magnitude , the well - known `` cosmic coincidence problem '' ( ccp ) .
physicists have proposed different ways to overcome this dilemma .
the first was to adopt a dynamical cosmological constant , trying to adjust the dynamics of it to alleviate the ccp .
this is the idea behind quintessence @xcite , where a scalar field is responsible for driving the current cosmic acceleration .
the second was to modify the left - hand side of einstein s equations , trying to explain the presence of a cosmological constant as a non - standard geometric effect @xcite .
the third was to violate the copernican principle , i.e. , by assuming we live in an inhomogeneous universe @xcite .
although some successes have been obtained in each one of these alternatives scenarios , there is so far no clear evidence of a preference compared to the @xmath2cdm model .
since einstein s field equations link the geometric properties of the universe with its total content , there is a well - known degeneracy between these two components ; dm and de .
this is in fact one good reason to consider unified dark models .
of course the _ simplicity _ of considering a single component acting as both dm and de is also a good reason . in this letter
, we show that a new class of scalar field model that exhibits a non - trivial response to geometry , dubbed _ stealth _ , which serves as a unified model of the dark sector .
the idea of considering unified scalar field models ( see @xcite for a review ) to describe dm and de emerges as a natural way to alleviate the so - called `` coincidence problem '' , namely , to explain why the energy densities of these two dark components are of the same order of magnitude today .
models of this type have been proposed in the past , assuming the stress energy tensor of the scalar field back - react to the geometry according to einstein s equations . among them
we can mention the model of @xcite ; where a potential @xmath3 is considered , the chaplygin gas @xcite , the generalized chaplygin gas model @xcite , and models with a non - canonical kinetic term called k - essence models @xcite . on the other hand ,
it is well - known that in the general theory of relativity gravity is understood as a manifestation of the curvature of space - time and the latter is caused by the presence of matter .
this fundamental principle is codified in the equations proposed by einstein .
so the slightest presence of matter on the right hand side of einstein s equations is sufficient to alter the geometry of space - time .
however , the stealth is a kind of matter that remains present in the space - time without altering or changing its geometry .
the stealth appears only for a scalar field non - minimally coupled to gravity , and its origins date back to the improved energy - momentum tensor considered first in @xcite , where the authors showed the possibility of this new tensor becoming the source of the gravitational field ; meanwhile , the dynamic of the scalar field is dictated by the klein - gordon equation .
the original proposal of the stealth was reported for a three - dimensional btz black hole in @xcite , in higher dimensions in minkowski space - time @xcite and ( anti-)de sitter [ ( a)ds ] space @xcite . also in lifshitz space - time in @xcite , for a four dimensional black hole @xcite , for an ads black hole in lovelock gravity @xcite , in einstein - gauss - bonnet gravity for topological black hole @xcite , for a rotating ads black hole in new massive gravity @xcite , and finally for a btz rotating black hole present two solutions in @xcite . lately , as was shown in @xcite , there are stealth fields during the cosmological evolution and some cosmological solutions are given in order to have a lcdm cosmology , in particular those with polynomials and power - law evolution are analyzed .
also , the general solutions for de sitter cosmologies and inhomogeneous stealths have been studied , concluding that only for de sitter backgrounds allow a full dependence on the space - time coordinates . in this letter
, we examine the case of a cosmological model coming from a non - minimal coupling with a stealth scalar field as a unified component describing both dm and de , and thus mimicking the @xmath2cdm model . _
ii .
stealths as a unified dark model_. in the present work we study a cosmological model coming from a non - minimal coupling with a stealth scalar field , described by the action : @xmath4.\nonumber\\\end{aligned}\ ] ] here , @xmath5 is the lagrangian matter .
clearly , for @xmath6 the scalar field stress tensor reduces to the usual case of a minimally coupled field . by varying the action ( [ action ] ) ,
the field equations are written as @xmath7 where @xmath8 is the stress energy tensor of matter , and @xmath9 is the stress energy tensor of the stealth field @xmath10 given by @xmath11 it is worth noting that for @xmath12 , the variation on @xmath13 produces the stealth stress tensor @xmath9 to get a contribution from the einstein tensor .
the stealth configuration emerges once we set both sides of eq.([feq ] ) to zero : the left - hand side is the einstein s equation for a universe with a matter content described by @xmath8 , and the right - hand side is the stealth equations @xmath14 .
once a solution to the right side is found , the stealth obeys the dynamics dictated by the space - time and at the same time it is invisible to it . while on the one hand the existence of gravitational stealth is a fact , and its feature of not having back - reaction on the gravitational field is of interest , the gravitational field equations say very little about their interaction with matter .
there are a few works on this topic ; some remarkable results in that direction are given in @xcite , where the interaction between ordinary matter and stealth is shown and @xcite , where a relation with the axionic field is shown .
furthermore , it is possible to show the ability of the stealth fields to mimic any kind of matter , which is another surprising characteristic of these fields @xcite . as was demonstrated in @xcite ,
there is a stealth solution in the context of a friedmann - lemaitre - robertson - walker ( flrw ) space - time .
now we obtain our cosmological model by coupling @xmath15,\ ] ] to a perfect fluid with zero pressure , i.e. , the dust case , plus a cosmological constant , and when the scalar field depend only on time . on the left - hand side of ( [ feq ] )
we use the stress - energy tensor for a perfect fluid as the dm contribution , leading to the usual @xmath2cdm model , where the dm density @xmath16 and the cosmological constant @xmath2 determine the cosmic evolution @xmath17 . at the same time , from the right - hand side of ( [ feq ] ) , the stealth field @xmath10 and its self interacting potential @xmath18 determines completely the cosmic evolution @xmath17 . as a consequence of this the evolution must be the _ same _ as that of the @xmath2cdm model the stealth here works as a unifying field simultaneously describing the effects of the action of both dm and de .
we get : @xmath19 and @xmath20 _ iii .
specific stealth realizations _ as is well - known , the @xmath2cdm model is so far the best fit model for a large set of astronomical observations , such as , type ia supernovae ( snia ) , baryon acoustic oscillations ( bao ) , cosmic microwave background radiation ( cmbr ) , growth of structure , etc . @xcite . in this setup
the cosmological constant @xmath2 drives the current accelerated expansion of the universe , detected for the first time using type ia supernovae @xcite , @xcite . as we mentioned in the introduction , although the stealth field does not back - react to the geometry , the existence of a non - zero coupling @xmath21 enables the stealth to appear dynamically coupled to the matter content . in this section
we characterize the stealth field associated with this cosmological model . in order to give a complete description of the model we are presenting
, we display the features of the @xmath2cdm model .
the friedmann equation is @xmath22 and the stress - energy conservation equation implies @xmath23 where we have assumed explicitly an equation of state @xmath24 for the matter content ( cold dark matter ) . from equation ( [ 8 ] )
we obtain the energy density , which evolves as @xmath25 .
this simple model fits several observational probes quite well .
the best fit parameters so far , assuming a curved frw metric , are those from the planck collaboration @xcite : @xmath26 , @xmath27 , and @xmath28 , where @xmath29 , @xmath30 , and @xmath31 . on the other hand , from the vanishing of the stealth stress - energy tensor , eqs.([eq1 ] , [ eq2 ] )
, we can read the equivalence relations between the set @xmath32 $ ] for the @xmath2cdm model , and the set @xmath33 $ ] for the stealths .
it is easy to show that an equivalence can be met by proposing the following relation : @xmath34 this means that the self interacting potential is related only algebraically to the cosmological constant .
using this relation and after some manipulations , the equivalence is complete after we impose @xmath35 = \kappa \rho,\ ] ] as well as @xmath36 where @xmath37 , and @xmath38 is an integration constant .
what we have obtained here is a stealth field @xmath10 given by the solution of ( [ key ] ) with self - interaction ( [ potfix ] ) that ( by construction ) _ generates _ an evolution
@xmath17 totally indistinguishable from that obtained from the @xmath2cdm model .
the equivalence enables us to use cosmological observations to fix the values ( and the uncertainties ) of the model parameters . using ( [ key ] ) in the equation for @xmath16
we get for the hubble function : @xmath39 before testing the model , we have to write it in terms of the redshift @xmath40 . recalling that @xmath41 the relation ( [ key ] ) can be written as @xmath42 where @xmath43 , @xmath44 , @xmath45 and @xmath46 . on the other hand , eq.([heq ] ) can be written as @xmath47 the free parameters to constrain are clearly @xmath21 and @xmath48 .
there is no way to constrain @xmath49 based on sets ( [ phiz ] ) and ( [ heq2 ] ) .
however , if we test the model using @xmath50 measurements , we can get a number for @xmath49 just by minimizing the residuals of @xmath51.\ ] ] in practice we solve differential equation ( [ phiz ] ) numerically with the initial condition @xmath52 , and by making use of ( [ heq2 ] ) to get @xmath53 .
then , we compute the residuals . in
what follows , we use observational measurements of @xmath50 extracted from @xcite consisting of 30 data points to constrain the free parameters in the model .
in addition to the parameters @xmath21 , @xmath48 and @xmath49 , we must also consider the parameters associated with the stealth potential @xmath18 . given our choice of ( [ potfix ] )
, the potential can be described by just one parameter that is fixed by relation ( [ potfix ] ) .
in fact , by writing @xmath54 , from ( [ potfix ] ) we find that @xmath55 . in this way
, it is not necessary to fit it along with the other three , because it depends on the best fit value of @xmath21 , @xmath49 and should be consistent with the known value of @xmath56 . along these lines , it is clear that our stealth have more free parameters than the original @xmath2cdm model ( the former has four and the latter three ) .
however , as we just mentioned , the only free parameters that we can fix using ( [ phiz ] ) and ( [ heq2 ] ) are @xmath21 , @xmath48 and @xmath49 . after the fit ( assuming the plus sign in @xmath57 ) , we get @xmath58 , @xmath59 and @xmath60 . in fig.([pdfz ] ) we show the @xmath61 and @xmath62 c.l . among the free parameters . in fig .
( [ fith ] ) we display the data points used to constrain the model along with the best theoretical curve . [ cols="^ " , ] the reduced hubble parameter ( @xmath63 ) is the most sensitive parameter in the fit . as we mentioned in the last paragraph
, this parameter essentially controls the amplitude of the theoretical curve displayed in fig.([fith ] ) .
by contrast parameters @xmath48 and @xmath21 are not very sensitive to changes , so it was very difficult to find a best fit set based on the @xmath50 data .
a close study of the system of eqs.([phiz ] ) and ( [ heq2 ] ) enables us to understand this behavior .
in fact , the use of observational data for each value of @xmath53 instead of using an analytical expression for it renders these two parameters highly correlated .
measurements obtained from @xcite . , width=302 ] _ iv . conclusions _ in this paper we have shown an example of how the stealth can operate during the cosmological evolution describing the @xmath2cdm model .
this example make use of an explicit quadratic potential for the stealth , and opens the possibility of extends this finding using other forms of @xmath18 .
we have also re - write the stealth equations in a way to find explicitly the equivalence between the @xmath2cdm content non - relativistic matter and a cosmological constant and the stealth field and its potential .
it was through this philosophy working with the stealth equivalent of the @xmath2cdm that we have found the example we studied . to put our model to the test , considering the special features of the current cosmological model
, we opt not to use an ansatz for the scale factor @xmath17 ; instead , we make use of observational data directly to constrain the free parameters of the stealth .
what we have shown here is the _ observationally _ induced stealth that best describes the @xmath2cdm model . here
, the stealth field with its self - interacting potential enables us to describe both the dark matter and the cosmological constant contributions , thus being a unified scalar field model . the best fit curve
i.e. , the function @xmath53 extracted from ( [ heq2 ] ) together with the data is shown in fig .
( [ fith ] ) , showing the capacity of the stealth to describe the observational data directly .
although by construction the stealth mimics the evolution of the @xmath2cdm model , we have performed a direct test of the model against observational data , and we have found that in this case the best fit modifies the value for @xmath48 ( instead of the typical @xmath64 , by our value @xmath65 ) at the expense of fixing an extra parameter , @xmath21 , which is absent in the @xmath2cdm model . finally , we would like to emphasize the importance of understanding the potential role the stealth may play in cosmic evolution . as we have shown ,
although stealth does not back - react to the space - time , it can describe both dark contributions at once , accounting for almost 95% of the matter content of the universe .
there is no doubt that we must continue to explore the consequences of the stealth in the recent cosmic evolution as well as in the early ages of the universe . |
in earlier publications @xcite most of the then available data on strong interaction effects in antiprotonic atoms have been used in ` global ' analyses in terms of a @xmath0-nucleus optical potential .
the simplest form of the potential is @xmath4~~,\ ] ] where @xmath5 is the @xmath0-nucleus reduced mass , @xmath6 and @xmath7 are the neutron and proton density distributions normalized to the number of neutrons @xmath8 and number of protons @xmath9 , respectively , @xmath10 and @xmath11 is the mass of the nucleon . in the zero - range ` @xmath12 '
approach the parameters @xmath13 and @xmath14 are minus the @xmath0-nucleon isoscalar and isovector scattering lengths , respectively , otherwise these parameters may be regarded as ` effective ' .
the fits of ref .
@xcite were based on 48 data points which included only two pairs of isotopes . as a typical result we quote re@xmath15 fm , im@xmath16 fm , leading to @xmath17/n , the @xmath17 per point , of 1.5 , which represents a very good fit to the data .
additional terms had also been cosidered in @xcite such as a @xmath3-wave term , which led to @xmath17/n in the range of 1.2 to 1.4 .
such a reduction in @xmath17 is , however , not statistically significant in these circumstances .
consequently only the above zero - range @xmath13 version has since been used in several analyses of experimental results . the recent publication of results for strong interaction effects in antiprotonic atoms by the ps209 collaboration @xcite changed the experimental situation dramatically by providing high quality data for several sequences of isotopes along the periodic table .
such data hold promise to pin - point isospin effects in the potential and/or explicit surface effects such as the existence of a @xmath3-wave term in the potential .
this is the scope of the present work which is based on data for five isotopes of ca , four isotopes each of fe and of ni , two isotopes each of zr and of cd , four isotopes of sn , five isotopes of te and the isotope @xmath18pb , all from results of the ps209 collaboration @xcite .
note that the results for cd , sn and te isotopes have been revised very recently to include effects of e2 resonance @xcite . in order to expand the range of the data further
we have included also earlier results for @xmath19o ( see @xcite for references to the original experiments . ) in section [ sect : dens ] we discuss the choice of nuclear density distributions @xmath20 and @xmath21 , which form an essential ingredient of the optical potential eq .
( [ equ : potl ] ) .
section [ sect : res ] summarizes the results and presents a ` recommended ' set of parameters which best describes the data .
in choosing the nuclear density distributions for use in the analysis of strong interaction effects in antiprotonic atoms , one is faced with the fact that due to the strength of the @xmath0-nucleus interaction and the dominant role played by the imaginary potential , strong interaction effects reflect the interaction with the extreme outer regions of the nucleus where the density is well below 10% of the central density .
it is therefore not very reliable to infer from these data quantities which relate to the bulk of the density such as root mean square ( rms ) radii . in what follows we therefore rely also on information from pionic atoms @xcite when choosing the densities , bacause pionic atoms are sensitive to the region where the density is 50% or more of the central density .
the density distribution of the protons is usually considered known as it is obtained from the nuclear charge distribution @xcite by unfolding the finite size of the charge of the proton .
the density distribution of the neutrons is usually not known to sufficient accuracy . in a recent paper @xcite trzciska et al .
have suggested , based on analyses of strong interaction effects in antiprotonic atoms and of radiochemical data following @xmath0 absorption on nuclei , that the difference between neutron and proton rms radii in nuclei depend on the asymmetry parameter @xmath22 as follows : @xmath23 addressing the question of the _ shape _ of the neutron density , they have concluded that a ` halo ' rather than a ` skin ' shape is preferred .
these shapes are defined as follows : assuming a two - parameter fermi distribution both for the proton ( unfolded from the charge distribution ) and for the neutron density distributions @xmath24 then for the ` skin ' form the same diffuseness parameter for the protons and the neutrons @xmath25 is used and the @xmath26 parameter is determined from the rms radius @xmath27 . in the ` halo '
form the same radius parameter @xmath28 is assumed for @xmath20 and for @xmath21 , and @xmath29 is determined from @xmath27 .
we also use an ` average ' option where the diffuseness parameter is set to be the average of the above two diffuseness parameters , namely , @xmath30 and the radius parameter @xmath26 is determined from the rms radius @xmath27 . to allow for different values of @xmath31
we have used eq .
( [ equ : radii ] ) but with variable slope parameter , where , e.g. a slope of 1.5 fm correspons to the average behaviour of results of relativistic mean field ( rmf ) calculations @xcite .
figure [ fig : dens ] shows examples of densities used in the present work .
the single - particle ( sp ) densities are described in @xcite .
the various neutron densities have the same value of @xmath27 , as given by rmf calculations .
note that where the densities are well below 10% of their central values the neutron densities are significantly larger than the proton density .
fits were made to 90 data points using a variety of shapes of nuclear densities and values of @xmath27 as described in the previous section .
finite range was also introduced as an option , where the nuclear densities have been folded with a gaussian , separately for the real and for the imaginary potentials , such that the densities in eq .
( [ equ : potl ] ) are replaced by ` folded ' densities @xmath32 table [ tab : res ] summarizes the main features of the results .
the columns marked ` rmf ' were obtained with values of @xmath31 taken from results of rmf calculation @xcite whereas the columns marked ` @xmath0 ' are for @xmath31 due to ref .
@xcite . in both cases
the 2-parameter fermi distribution was used for the proton and for the neutron densities .
the column marked with sp is for both densities obtained by filling in levels in single particle potentials @xcite with @xmath31 values from the rmf model .
the shapes of the neutron distributions ( ` ave . ' , ` halo ' , ` skin ' ) are as explained in the previous section .
the top part of the table is for zero range interaction and the bottom part of the table is for finite range interaction with gaussian range parameters @xmath33 and @xmath34 for the real and for the imaginary parts , respectively . from the values of @xmath17/n for the zero range
fits it is seen that reasonable fits to the data can be achieved for all options except for the ` rmf - halo ' variety .
the increased values of @xmath17/n compared to results of analyses of the previous data @xcite reflect the improved accuracy and the additional information contained in the new data .
the latter is observed in the lower part of the table where the need for a finite - range version of the @xmath0-nucleus optical potential is self - evident .
the values of @xmath17 and of @xmath17/n mean significant improvements in the fits compared to the zero - range results .
the best values for @xmath17/n are close to what has been achieved in global fits to pionic atom data @xcite .
.global fits to 90 antiprotonic atom points .
` rmf ' and ` @xmath0 ' stand for values of @xmath31 used .
top part is for zero - range potentials and bottom part for finite - range potentials with range parameters @xmath35 . [
cols="^,^,^,^,^,^,^,^ " , ] next we turn to the question of whether the data require the presence of an isovector term @xmath14 in the potential eq .
( [ equ : potl ] ) .
fits to the full data set of 90 points failed to produce any reduction in @xmath17 when such a term was introduced . in a separate calculation
we have been able to obtain a non - zero value for the parameter @xmath14 when fits were made to only the data for the sn isotopes , leading to @xmath17/n of 1.0 for these isotopes .
however , applying that @xmath14 to the full data set resulted in a major deterioration of the fit , and a further global fit restored the results of table [ tab : res ] with @xmath14=0 .
this is not surprizing since antiprotonic atoms are sensitive to the nuclear density in the extreme outer regions of the nucleus where the neutron density is dominant .
explicit isospin effects are , therefore , hard to observe .
another attempt to improve the agreement between zero - range calculations and experiment was made by introducing a @xmath3-wave term into the optical potential @xcite , a term which is active mostly at the surface region .
starting with the zero - range version of the potential , the values of @xmath17 could be reduced from those of the upper part of table [ tab : res ] to those of the lower part , but with @xmath3-wave parameters which depended critically on the choice of shape for the neutron distribution . including in addition
finite - range folding failed to further improve the fits . in conclusion , global analyses of most of the data from ps209 reveal a need to improve on the previously successful zero - range @xmath1-wave isoscalar @xmath0-nucleus optical potential . including an isovector term is not required by the data nor is a @xmath3-wave term recommended at our present state of knowledge , although it undoubtedly improves the fits to the data when the zero - range form is used .
the use of a finite - range interaction in the optical potential is definitely required by the data .
assuming that the values of the differences between rms radii @xmath31 are , on the average , between those advocated in ref .
@xcite and those predicted by rmf calculations @xcite , and assuming the shape of the neutron distributions is between those described by the ` skin ' and by the ` average ' prescriptions , we then recommend the values re@xmath13
= [email protected] fm , im@xmath13 = [email protected] fm , with @xmath37 = @xmath38 = 0.85 fm within the folding expression eq .
( [ equ : fold ] ) . | the experimental results for strong interaction effects in antiprotonic atoms by the ps209 collaboration consist of high quality data for several sequences of isotopes along the periodic table .
global analysis of these data in terms of a @xmath0-nucleus optical potential achieves good description of the data using a @xmath1-wave finite - range @xmath2 interaction .
equally good fits are also obtained with a poorly - defined zero - range potential containing a @xmath3-wave term .
, antiproton - nucleus optical potentials , global fits 13.75.cs , 14.20.dh , 21.10.gv , 36.10.-k |
in the dual superconductor mechanism of confinement a dual meissner effect @xcite confines color electric flux to narrow tubes connecting a quark - antiquark pair . the dual abelian higgs model , which couples dual potentials @xmath4 to a monopole higgs field @xmath5 , provides an example of this mechanism
. @xcite the action @xmath6 $ ] of the theory is @xmath7 = \int d^4x \left [ -\frac{\left(\partial_\mu c_\nu - \partial_\nu c_\mu\right)^2}{4 } - \frac{\left|\left(\partial_\mu - ig c_\mu\right)\phi \right|^2}{2 } - \frac{\lambda}{4 } \left ( |\phi|^2 - \phi_0 ^ 2 \right)^2 \right ] \ , .
\label{field action}\ ] ] the dual coupling constant is @xmath8 , where @xmath9 is the yang mills coupling constant . in the confined phase the monopole fields @xmath5
condense to the value @xmath10 .
the dual potentials @xmath4 acquire a mass @xmath11 via a dual higgs mechanism .
quarks couple to dual potentials via a dirac string connecting the quark - antiquark pair along a line l , the ends of which are sources and sinks of color electric flux .
the color field of the pair destroys the dual meissner effect near l so that @xmath5 vanishes on l. at distances greater than @xmath12 the monopole field returns to its bulk value @xmath10 so that the color field is confined to a tube of radius @xmath13 surrounding the line l , which is the location of the vortex .
@xcite this model provides an effective theory of long distance qcd , i.e. for distances greater than @xmath14 .
here we want to use this theory to obtain the energy levels of mesons having large angular momentum . under such circumstances
the distance between quarks is much larger than the flux tube radius and we must take into account the contribution of flux tube fluctuations to the interaction between the quark and antiquark , the fluctuating vortex line l sweeps out a world sheet @xmath15 whose boundary is the loop @xmath16 formed from the world lines of the moving quark - antiquark pair .
their interaction is determined by the wilson loop @xmath17 $ ] : @xmath18 = \int \scrd c_\mu \scrd\phi \scrd\phi^ * e^{is[c_\mu , \phi ] } \ , .
\label{wilson loop def}\ ] ] the path integral ( [ wilson loop def ] ) goes over all field configurations for which the monopole field @xmath19 vanishes on some sheet @xmath15 bounded by the loop @xmath16 .
the wilson loop @xmath17 $ ] describes the quantum fluctuations of a field theory having classical vortex solutions . we want to express the functional integration over fields as a path integral over the vortex sheets @xmath15 to obtain an effective string theory of these vortices . to do this
we carry out the functional integration ( [ wilson loop def ] ) in two stages : 1 .
we integrate over all field configurations in which the vortex is located on a particular surface @xmath15 , where @xmath20 .
this integration determines the action @xmath21 $ ] of the effective string theory .
2 . we integrate over all vortex sheets @xmath22 , @xmath23 , @xmath24 .
this integration gives @xmath17 $ ] the form of an effective string theory of vortices .
the action @xmath21 $ ] is invariant under reparameterizations @xmath25 of the world sheet @xmath22 of the vortex .
we choose a particular parameterization of @xmath15 in terms of the amplitudes @xmath26 and @xmath27 , of the two transverse fluctuations of the vortex .
@xmath28 using the parameterization ( [ x param ] ) , we can write the integration over vortex sheets as a path integral over the transverse vortex fluctuations @xmath26 and @xmath27 .
the path integral ( [ wilson loop def ] ) over fields then takes the form : @xcite @xmath18 = \int \scrd f^1 \scrd f^2 \delta_{fp } e^{is_{{\rm eff}}[\tilde x^\mu ] } \ , , \label{wilson loop eff}\ ] ] where @xmath29\ ] ] is a faddeev - popov determinant , and @xmath30 is the square root of the determinant of the induced metric @xmath31 , @xmath32 the path integral ( [ wilson loop eff ] ) goes over fluctuations in the shape of the vortex sheet with wave lengths greater than the radius @xmath12 of the flux tube .
all the long distance fluctuations in @xmath17 $ ] are contained in this path integral over string fluctuations .
the presence of the determinant @xmath33 makes the path integral ( [ wilson loop eff ] ) invariant under reparameterizations @xmath34 of the vortex worldsheet .
specifying the parameterization ( [ x param ] ) is analogous to fixing a gauge in a gauge theory , and gives rise to the path integral ( [ wilson loop eff ] ) over @xmath26 and @xmath27 .the factor @xmath33 arises from writing the original field theory path integral ( [ wilson loop def ] ) as a ratio of path integrals of two string theories @xcite , and reflects the field theory origin of the effective string theory .
this ratio is anomaly free .
anomalies @xcite present in string theory are not present in field theory .
the parameterization invariant measure in the path integral ( [ wilson loop eff ] ) is universal and is independent of the explicit form of the underlying field theory . on the other hand , the action @xmath21
$ ] of the effective string theory is not universal .
it is determined by the integration over all field configurations which contain a vortex on the surface @xmath15 . however , for wavelengths @xmath35 of the string fluctuations greater than the flux tube radius @xmath14 , which are those included in ( [ wilson loop eff ] ) , the action @xmath21 $ ] can be expanded in powers of the extrinsic curvature tensor @xmath36 of the sheet @xmath15 , @xmath37 = - \int d^4x \sqrt{-g } \left [ \sigma + \beta \left(\scrk^a_{ab}\right)^2 + ... \right ] \ , .
\label{curvature expansion}\ ] ] the extrinsic curvature tensor is @xmath38 where @xmath39 , @xmath40 are vectors normal to the worldsheet at the point @xmath22 .
the string tension @xmath41 and the rigidity @xmath42 are determined by the parameters of the underlying effective field theory , whose long distance fluctuations are described by the effective string theory .
the expansion parameter in ( [ curvature expansion ] ) is the ratio @xmath43 of the square of the flux tube radius to the square of the wave length of the string fluctuations . in mesons of angular momentum
@xmath44 this parameter is of order @xmath45 . for mesons of large angular momentum
the effective action ( [ curvature expansion ] ) can therefore be approximated by the nambu
goto action , @xmath46 and the path integral for @xmath17 $ ] becomes @xmath18 = \int \scrd f^1 \scrd f^2 \delta_{fp } e^{i\sigma \int d^4x \sqrt{-g } } \ , .
\label{wilson nambu goto}\ ] ] in the next section we describe the results @xcite of a semiclassical expansion of this effective string theory .
we calculate @xmath17 $ ] in the leading semiclassical approximation in the background of a worldsheet generated by a straight string attached to quarks rotating with uniform angular velocity @xmath47 ( see fig . [
rot string fig ] ) .
the quarks have masses @xmath48 and @xmath49 , move with velocities @xmath50 and @xmath51 , and are separated by a fixed distance @xmath52 .
the parameters @xmath53 are the time @xmath54 and the coordinate @xmath55 , which runs along the string from @xmath56 to @xmath57 , so that the transverse velocity of the straight string is zero when @xmath58 .
the amplitudes @xmath26 and @xmath27 of the transverse fluctuations are the spherical coordinates @xmath59 and @xmath60 of a point on the string .
these angles are defined in an unconventional manner so that @xmath61 is a straight string rotating in the @xmath62 plane .
the angle @xmath59 is the fluctuation perpendicular to the plane of rotation , and the angle @xmath60 is the fluctuation lying in the plane of the rotating string .
the ends of the string are fixed to their classical trajectories , @xmath63 the fluctuating worldsheet @xmath15 then has the parameterization @xmath64 given by @xmath65 \
, , \label{x mu param}\end{aligned}\ ] ] where @xmath66 , @xmath67 are unit vectors along the four fixed spacetime axes , @xmath68 .
the parameterization ( [ x mu param ] ) is a concrete example of ( [ x param ] ) . the classical rotating straight string @xmath69 has the parameterization , @xmath70 \,.\end{aligned}\ ] ] the corresponding metric @xmath71 $ ] and classical action @xmath72 $ ] are independent of the time @xmath54 , so that @xmath17 $ ] has the form @xmath18 = e^{itl^{{\rm string}}(r_1 , r_2 , \omega ) } \ , , \label{wilson loop integrated}\ ] ] where @xmath73 is the elapsed time . for massless quarks ,
the ends of the string move with the velocity of light , and singularities appear in @xmath74 . to regulate these singularities
, we retain the quark mass as a cutoff and take the massless limit at the end when evaluating physical quantities .
the lagrangian @xmath74 is the sum of a classical part @xmath75 and a fluctuating part @xmath76 , @xmath77 where @xmath78 we will calculate @xmath76 in @xmath0 spacetime dimensions.the fluctuation @xmath59 is replaced by @xmath79 fluctuations perpendicular to the plane of rotation and there is still just 1 fluctuation in the plane of rotation . the effective lagrangian for the quark
antiquark pair is obtained by adding quark mass terms to @xmath74 : @xmath80 the effective lagrangian is the sum of a classical part and a fluctuating part , @xmath81 where @xmath82 \ , , \label{l_cl def formal}\end{aligned}\ ] ] with @xmath83 the expression for @xmath76 is obtained from ( [ wilson loop integrated ] ) and the semiclassical calculation of @xmath17 $ ] .
it contains terms which are quadratically , linearly , and logarithmically divergent in the cutoff @xmath84 .
the quadratically divergent term is a renormalization of the string tension , the linearly divergent term is a renormalization of the quark mass , and the logarithmically divergent term is proportional to the integral of the scalar curvature over the whole worldsheet @xcite . after absorbing the quadratically and linearly divergent terms into renormalizations ,
we obtain @xcite @xmath85 + \frac{\omega}{2 } + \omega f(v_1 , v_2 ) \ , , \label{l fluc eval}\ ] ] where @xmath86 is the proper length of the string , @xmath87 and @xmath88 \ , .
\label{f v_1 v_2 def}\ ] ] the function @xmath89 vanishes when @xmath90 and @xmath91 approach unity , so that the last term in ( [
l fluc eval ] ) is small for relativistic quarks .
the contribution to @xmath76 of the additional @xmath92 fluctuations perpendicular to the plane of rotation , coming from the extension to d dimensional spacetime , is contained in the first term in ( [ l fluc eval ] ) . in the limit @xmath93 , @xmath94 , and @xmath76
reduces to the result of lscher @xcite for the correction to the static quark
antiquark potential due to string fluctuations , @xmath95 for @xmath96 , @xmath76 contains a logarithmically divergent part .
we simplify this term using the classical equation of motion , @xmath97 to express @xmath98 in terms of @xmath47 .
( [ canonical classical eqn ] ) gives the relation @xmath99 where @xmath100 is equal to @xmath101 evaluated at @xmath102 .
the solution of ( [ classical eqn ] ) for @xmath98 as a function of @xmath47 is @xmath103 using the relation ( [ classical eqn ] ) in ( [ l fluc eval ] ) gives @xmath104 + \frac{\omega}{2 } + \omega f(v_1 , v_2 ) \ , .
\label{l fluc pre renorm}\ ] ] use of the classical equations of motion has eliminated the logarithmic dependence of @xmath76 on the dynamical parameter @xmath47 . the logarithmically divergent term in ( [
l fluc pre renorm ] ) is proportional to the quantity @xmath105 which diverges when @xmath106 .
this term can be absorbed into a renormalization of a term in the boundary action called the geodesic curvature @xcite so that the theory will be finite in the @xmath106 limit .
for a straight string rotating with angular velocity @xmath47 , the geodesic curvature is equal to @xmath107 . the logarithmic divergence in ( [ l fluc pre renorm ] )
can then be removed by adding to the quark mass term in the boundary lagrangian a counterterm containing the geodesic curvature .
this gives @xmath108 \ , , \label{l boundary expand}\end{aligned}\ ] ] where the second term in ( [ l boundary expand ] ) is the geodesic curvature term .
the logarithmic divergence in ( [ l fluc pre renorm ] ) can then be regarded as a renormalization of the coefficient @xmath109 of the geodesic curvature in ( [ l boundary expand ] ) . since the quantity @xmath105 diverges when @xmath106 , the requirement that the theory is finite in the @xmath106 limit forces the renormalized value of @xmath109 to be zero .
( we take the @xmath106 limit with the cutoff @xmath84 fixed , because we have an effective theory ) . removing the terms in ( [ l fluc pre renorm ] ) proportional to @xmath110
gives an expression for @xmath76 which is applicable in the massless quark limit , @xmath111 in the case of two light quarks , @xmath112 ( @xmath113 ) , eqs .
( [ proper length ] ) and ( [ f v_1 v_2 def ] ) give @xmath114 and @xmath115 , so that ( [ l fluc renorm ] ) becomes @xmath116 the first term in ( [
l fluc ll ] ) is the negative of lscher potential with the length @xmath117 of the string replaced by its proper length @xmath114 .
it is the contribution of @xmath118 tranverse fluctuations in the background of a flat metric .
the @xmath119 term accounts for the curvature of the classical background metric @xmath71 $ ] generated by the rotating string . in the case of one heavy and one light quark , @xmath120 ( @xmath121 ) and @xmath122 ( @xmath123 ) , @xmath124 and @xmath125 so @xmath126 in this
talk we use ( [ l fluc ll ] ) to determine regge trajectories of mesons composed of two light quarks .
in the previous discussion , the quark antiquark pair moved in a fixed classical trajectory in the @xmath62 plane ( see fig . [
rot string fig ] and eq s ( [ classical boundary angles ] ) and ( [ solve for r_i ] ) ) . to take into account the fluctuations of the positions @xmath127 and @xmath128 of the quarks at the ends of the rotating string , we extend the functional integral ( [ wilson nambu goto ] ) to include a path integral over @xmath127 and @xmath129 , and add the action of the quarks to the string action ( [ nambu goto action ] ) .
this extension replaces @xmath17 $ ] by the partition function @xmath130 , @xmath131 where @xmath132 is the action of two free ( scalar ) quarks , @xmath133 and where @xmath134 is the corresponding quark partition function , @xmath135 dividing by @xmath134 removes the vacuum energy of the quarks .
we use the methods of dashen , hasslacher , neveu @xcite to carry out a semiclassical calculation of the partition function @xmath130 around periodic classical solutions and determine the energies of the physical meson states .
we find : 1 . for massless quarks , the contribution to @xmath136 arising from the fluctuations in the motion of the ends of the string along the vector @xmath137 shown in fig .
[ rot string fig ] vanishes .
for massless quarks the longitudinal mode is unphysical and the meson energy comes from the fluctuations of the interior of the string , i.e. @xmath138 .
+ ( for massive quarks the fluctuations of the boundary also contribute to the meson energy . )
the usual wkb quantization condition for angular momentum .
the angular momentum @xmath44 , @xmath139 takes on the values , @xmath140 the energy @xmath141 is given by the corresponding hamiltonian , @xmath142 eq s ( [ l_cl def formal ] ) and ( [ l fluc ll ] ) give @xmath143 so that for large @xmath44 , @xmath144 can be treated as a perturbation .
the meson energy is then @xmath145 where @xmath146 and where @xmath47 is given as a function of @xmath44 by the classical relation @xmath147 eqs .
( [ j class ] ) and ( [ e of j ] ) determine meson regge trajectories in the leading semiclassical approximation .
we first evaluate @xmath148 to obtain classical regge trajectories . for massless quarks eqs .
( [ l_cl def formal ] ) , ( [ e class ] ) , and ( [ j class ] ) give @xmath149 using eqs .
( [ e of j ] ) , ( [ l fluc ll ] ) , and ( [ classical regge eqn ll ] ) to include the correction to the energy due to the zero point energy of the string fluctuations gives @xmath150 the energies @xmath151 of the excited states of the rotating string ( light hybrid mesons ) are obtained by adding to ( [ e of omega ll ] ) the energies @xmath152 of the excited vibrational modes .
@xmath153 since there are many combinations of string normal modes which give the same @xmath154 ( e.g. , a doubly excited @xmath155 mode and a singly excited @xmath156 mode each give @xmath157 ) , the spectrum is highly degenerate .
there are @xmath118 trajectories having @xmath158 , each corresponding to a single excitation of one of the @xmath155 normal modes .
higher values of @xmath154 have higher degeneracies . the value of @xmath47 is given as a function of @xmath44 through the classical relation @xmath159 .
squaring both sides of ( [ hybrid energy ll ] ) yields @xmath160 where we have used the wkb quantization condition ( [ j quant ] ) . setting @xmath2 in ( [ e^2 l ] ) and solving for @xmath161 we obtain the the regge trajectories ,
@xmath162 the excited states , @xmath163 , of the rotating string give rise to daughter regge trajetories determining the spectrum of hybrid mesons composed of zero mass quarks .
the first semiclassical correction to the leading regge trajectory , @xmath164 , adds the constant @xmath3 to the classical regge formula .
the small size of this correction could explain why regge trajectories are linear at values of @xmath161 of order 1 .
for @xmath1 , eq . ( [ e^2 l ] ) yields the spectrum @xmath165 the spectrum of energies ( [ 26 dim spectrum ] ) coincides with the spectrum of open strings in classical bosonic string theory . however , ( [ 26 dim spectrum ] ) is valid only in the leading semiclassical approximation , so that it can not be used for @xmath166 , where it would yield the scalar tachyon of the open bosonic string
1 . we have seen how the dual superconducting model of confinement leads to an effective string theory of long distance qcd .
we have calculated , in the semiclassical large angular momentum domain , where this theory is applicable , the effect of string fluctuations on regge trajectories for mesons containing light quarks .
similar calculations @xcite give regge trajectories for mesons containing one heavy and one light quark .
the spectrum of the energies of the excited states formally coincides with the spectrum of the open string of bosonic string theory in its critical dimension @xmath167 . here
, we obtained the spectrum for any @xmath0 , but only for large @xmath161 , from the semiclassical expansion of an effective string theory .
the presence of functional determinant @xmath33 in the path integral ( [ wilson nambu goto ] ) was essential in obtaining this result .
we treated the light quarks as massless scalar particles to determinine the energies of the high angular momentum excited states of mesons .
however , the effect of chiral symmetry breaking , generating a constituent quark mass , must play a dominant role in determining the masses of mesons which are ground states of quark antiquark systems .
4 . the derivation of the effective string theory made no use of the details of the effective field theory from which it was obtained .
furthermore , arguments based on the work of t hooft @xcite indicate that the confined phase of a non abelian gauge theory is characterized by a dual order parameter , which vanishes in regions of space where `` dual supercondutivity '' is destroyed
. this generic description of the confined phase of qcd leads to the effective string theory of long distance qcd described here , which provides a concrete picture of the qcd string .
we would like to thank s.olejnik and the other organizers for providing the opportunity to participate in this stimulating workshop . | we show how an effective field theory of long distance qcd , describing a dual superconductor , can be expressed as an effective string theory of superconducting vortices .
we evaluate the semiclassical expansion of this effective string theory about a classical rotating string solution in any spacetime dimension @xmath0 .
we show that , after renormalization , the zero point energy of the string fluctuations remains finite when the masses of the quarks on the ends of the string approach zero . for @xmath1 the semiclassical energy spectrum of the rotating string
formally coincides with that of the open string in classical bosonic string theory .
however , its physical origin is different .
it is a semiclassical spectrum of an effective string theory valid only for large values of the angular momentum .
for @xmath2 the first semiclassical correction adds the constant @xmath3 to the classical regge formula for the angular momentum of mesons on the leading regge trajectory .
the excited vibrational modes of the rotating string give rise to daughter regge trajectories determining the spectrum of hybrid mesons . |
topological insulators ( tis ) are an emerging quantum phase in condensed matter physics @xcite .
the key feature of tis is that the electronic band structure resembles an ordinary insulator in the bulk , while gapless edge or surface states exist within the bulk energy gap @xcite .
these edge states are protected by time - reversal symmetry .
ti materials have been theoretically predicted and experimentally observed in various systems such as hgte / cdte and inas / gasb quantum well , and bi@xmath0sb@xmath1 , bi@xmath2te@xmath3 , bi@xmath4se@xmath5 , bi@xmath4te@xmath4se @xcite . a novel quantized topological magnetoelectric effect ( tme ) is predicted in tis that an applied electric field could induce parallel magnetization and an applied magnetic field could induce parallel electric polarization @xcite .
an additional term in the lagrangian : @xmath6 , which gives rise to the axion electrodynamics " , contains a complete description of the electromagnetic(em ) responses of tis @xcite . here
@xmath7 is the fine structure constant , @xmath8 is the axion angle with @xmath9 being an integer , @xmath10 is electric field and @xmath11 is magnetic field . owing to this additional lagrangian , the constitutive relations for tis should be modified as : @xmath12 , where @xmath13 and @xmath14 are respectively the electric displacement and magnetic field strength , @xmath15 and @xmath16 are respectively the permittivity and permeability of tis , @xmath17 is proportional to the fine structure constant .
these modified constitutive relations provide us a compact form to deal with the tme effect as the doping level is in the gap .
many unusual phenomena due to tme effect have been revealed @xcite .
the study of em waves scattering by small particles and its application is an enormous field @xcite .
em waves scattering by scatterers of tis exhibits lots of interesting properties due to the tme effect .
for instance , parity - violating light scattering under oblique incidence , and strong perturbation of dipole radiation near the ti spheres are demonstrated @xcite .
broadband strong scattering in backward direction and interesting antiresonances are found for ti cylinders @xcite .
the antiresonance effect reveals that the tme effect is a surface effect essentially in the language of em scattering .
moreover , the quantization of tme effect of ti cylinders can even be determined by measuring the electric - field components of scattered waves in the far fields at one or two scattering angles in the rayleigh scattering limit @xcite . in this article , we study light scattering by spheres of tis .
based on mie theory , we derive the scattering coefficients and scattering matrix for ti spheres analytically .
exotic strong backward scattering due to tme effect are also found in both rayleigh and mie scattering regimes , similar to the case for ti cylinders . at certain frequencies , antiresonances of cross - polarized scattering coefficients of ti spheres
are revealed , wherein the cross - polarized fields induced by tme effect are trapped inside ti spheres . in the rayleigh limit ,
we propose a simple way to determine the quantized tme effect of tis by measuring the electric - field components of scattered waves in the far field .
direction and polarized along @xmath18 direction.,width=113 ] the system under study is shown in fig .
we consider a ti sphere which is illuminated by a time - harmonic em wave with frequency @xmath19 .
the radius of the sphere is @xmath20 .
the dielectric permittivity and magnetic permeability for the ti sphere are ( @xmath21 ) , and those for the background medium are ( @xmath22 ) .
this scattering problem can be solved analytically by the standard mie theory . in the spherical coordinate system , ( @xmath23 , @xmath24 , @xmath25 ) ,
the em fields can be expanded by the vector spherical harmonics @xcite : @xmath26 , @xmath27 , @xmath28 and @xmath29 , where subscripts @xmath30 and @xmath31 denote respectively the even and odd modes with respect to the @xmath18 axis , @xmath32 is the corresponding wavevector ( @xmath33 , @xmath34 are the wavevectors in the ti sphere and background respectively ) , @xmath35 is a integer , and @xmath36 stands for which kind of spherical bessel(hankel ) functions we use .
we assume that the incident em wave is @xmath37 , where @xmath38 is the amplitude .
the scattered and internal electric fields can be expanded in terms of the vector spherical harmonics @xcite : @xmath39 where @xmath40 , \{@xmath41 } are the scattering coefficients that have one - to - one correspondence to the ones for the conventional dielectrics . however , two new scattering coefficients , \{@xmath42 , @xmath43 } , or the so called _ cross - polarized _ scattering coefficients arise from the tme effect .
the superscripts 1 and 3 stand for the spherical bessel function of the first kind and spherical hankel function of the first kind , respectively . by matching the standard boundary conditions at @xmath44 , and applying the modified constitutive relations @xmath45 inside the ti sphere ,
the scattering coefficients @xmath46 , @xmath47 , and the internal coefficients @xmath48 , @xmath49 are found to be : @xmath50^{^{\prime } } \beta _ { 2}-\mu _ { 1}j_{n}(x)[mxj_{n}(mx)]^{^{\prime } } } { \mu _ { b}m^{2}j_{n}(mx)[xh_{n}(x)]^{^{\prime } } \beta _ { 2}-\mu _ { 1}h_{n}(x)[mxj_{n}(mx)]^{^{\prime } } } \\ b_{n } & = & \frac{\mu _ { 1}j_{n}(mx)[xj_{n}(x)]^{^{\prime } } -\mu _ { b}j_{n}(x)[mxj_{n}(mx)]^{^{\prime } } \beta _ { 1}}{\mu _ { 1}j_{n}(mx)[xh_{n}(x)]^{^{\prime } } -\mu _ { b}h_{n}(x)[mxj_{n}(mx)]^{^{\prime } } \beta _ { 1 } } \\ c_{n } & = & \frac{\mu _ { 1}j_{n}(x)[xh_{n}(x)]^{^{\prime } } -\mu _ { 1}h_{n}(x)[xj_{n}(x)]^{^{\prime } } } { \mu _ { 1}j_{n}(mx)[xh_{n}(x)]^{^{\prime } } -\mu _ { b}h_{n}(x)[mxj_{n}(mx)]^{^{\prime } } \beta _ { 1 } } \\ d_{n } & = & \frac{\mu
_ { 1}mj_{n}(x)[xh_{n}(x)]^{^{\prime } } -\mu _ { 1}mh_{n}(x)[xj_{n}(x)]^{^{\prime } } } { \mu _ { b}m^{2}j_{n}(mx)[xh_{n}(x)]^{^{\prime } } \beta _ { 2}-\mu _ { 1}h_{n}(x)[mxj_{n}(mx)]^{^{\prime } } } \end{aligned}\ ] ] where @xmath51 and @xmath52 are respectively the spherical bessel and hankel functions of the first kind , @xmath53 is the size parameter and @xmath54 is the relative refractive index ; @xmath55 and @xmath56 are auxilliary functions related to the axion angle , given by @xmath57 and @xmath58 , where @xmath59 and @xmath60^{^{\prime } } } { \mu _ { b}m^{2}j_{n}(mx)[xh_{n}(x)]^{^{\prime } } -\mu _ { 1}h_{n}(x)[mxj_{n}(mx)]^{^{\prime } } } \\ \chi_{2 } & = & \frac{\mu _ { b}h_{n}(x)[mxj_{n}(mx)]^{^{\prime } } } { \mu _ { 1}j_{n}(mx)[xh_{n}(x)]^{^{\prime } } -\mu _ { b}h_{n}(x)[mxj_{n}(mx)]^{^{\prime } } } \end{aligned}\ ] ]
we assume that @xmath76 in this article since the materials considered here are non - magnetic . in the rayleigh scattering limit ( @xmath77 and @xmath78 )
, it can be shown that only the following scattering coefficients are in the order of @xmath79 : @xmath80 all other scattering coefficients are in the order of @xmath81 or higher , and can be hence neglected in the rayleigh scattering limit .
the scattering coefficients @xmath82 and @xmath83 correspond to the electric and magnetic dipolar terms , whereas @xmath84 and @xmath85 correspond to the cross - polarized electric dipolar term and cross - polarized magnetic dipolar term respectively .
since the wavevector of the incident wave is along the @xmath86 direction , the scattering direction @xmath87 and the forward direction @xmath88 define a scattering plane .
the incident electric fields and the scattered electric fields in the far field ( @xmath89 ) can be related by a scattering matrix @xcite : @xmath90where @xmath91 and @xmath92 are respectively the components of scattered electric fields parallel and perpendicular to the scattering plane .
the corresponding unit vectors parallel and perpendicular to the scattering plane are defined as : @xmath93=@xmath94 and @xmath95=-@xmath96 .
the components of incident electric fields parallel and perpendicular to the scattering plane are respectively @xmath97 and @xmath98 , which depend on azimuth angle .
the elements of amplitude scattering matrix @xmath99 ( @xmath100=1 , 2 , 3 , 4 ) for a single ti sphere are given by the following relations : @xmath101 , @xmath102 , @xmath103 , @xmath104 , where @xmath105 and @xmath106 @xcite .
@xmath107 are associated legendre functions . for conventional dielectric spheres ,
the condition @xmath108 leads to a diagonal scattering matrix , i.e. , @xmath109 . for ti spheres , however , scattering matrix is not a diagonal matrix in general since @xmath110 and @xmath43 may not vanish owing to the tme effect .
the intensities of scattered waves , defined as energy flux per unit area , characterize scattering properties of the system .
based on the scattering matrix shown above , we ll discuss the role that the dielectric background plays in the em scattering . in the rayleigh limit ,
only electric and magnetic dipolar terms dominate the em scattering .
the angle - dependent functions , @xmath111 and @xmath112 for @xmath113 , are given by : @xmath114 .
two different kinds of limit can be estimated in the following : ( i ) , @xmath115 .
therefore @xmath116 , in other words , the contribution of the conventional electric dipole is much larger than those of the cross - polarized electric and magnetic dipoles induced by tme effect . the intensity of scattered waves is mainly determined by the diagonal terms of scattering matrix .
thus , the scattering patterns of ti spheres are similar to those of conventional dielectric spheres .
( ii ) , @xmath117 .
the refractive index of the background is very close to that of the ti sphere . in this case
we have @xmath118 , namely , the conventional electric and magnetic dipoles are negligible compared to cross - polarized ones .
we thus have : @xmath119 where we assume that the incident plane wave has an unit intensity , @xmath120 and @xmath121 are intensities correspond to the polarized components with scattered electric - field vectors parallel and perpendicular to the scattering plane , respectively .
the scattered intensity is contributed mainly by off - diagonal terms of scattering matrix . clearly , in the forward direction ( @xmath122 ) ,
the scattered intensity reduce to zero for both @xmath120 and @xmath123 since @xmath124 . in the backward direction ( @xmath125 )
, however , the scattered intensity shows a strong backscattering stemming from the constructive interference of the cross - polarized electric and magnetic dipole radiation , i.e. the terms proportional to @xmath84 and @xmath85 . for ti spheres .
the dielectric permittivity of the background medium is the same as that of ti spheres .
( a ) for rayleigh scattering ( @xmath18=0.01 ) and ( b ) for mie scattering ( @xmath18=1 ) .
the boundary of the sphere in ( b ) is indicated by a white circle .
the incident wave propagates along @xmath86 direction.,width=264 ] to show these unusual scattering patterns , we plot the distributions of scattered electric fields in fig . [ fig2 ] .
the dielectric permittivity of ti sphere and background medium are chosen to be @xmath126 .
the resulting scattering patterns are completely different from those for the conventional media , showing a strong backward scattering for both rayleigh and mie scattering .
it is noted that such enhanced or even complete backward scattering has also been found in other systems such as magnetic@xcite , metallic @xcite , or semiconducting @xcite particles .
however , unlike metallic spheres @xcite or semiconducting spheres with high refractive index @xcite , this enhanced backward scattering is a broadband effect for ti spheres since the tme effect is non - resonant intrinsically . with @xmath113 .
the resonant peaks in ( a ) and ( c ) mainly come from the bulk scattering of ti spheres .
the corresponding size parameters of antiresonances for cross - polarized coefficients are shown by dotted lines in ( b ) and ( d ) .
the dielectric permittivity of ti spheres is @xmath127 and the axion angle is chosen to be @xmath128 .
the background is vacuum.,width=264 ]
in fig . [ fig3 ] , the scattering and internal coefficients corresponding to dipolar modes as functions of size parameter @xmath18 are shown . for scattering coefficients @xmath82 and @xmath83 , there exist resonances and antiresonances .
for instance , the resonances at @xmath129 for @xmath130 and @xmath131 for @xmath132 correspond to the excitation of electric and magnetic dipolar modes respectively . meanwhile , the internal coefficients @xmath133 and @xmath134 have similar resonant behaviors as @xmath82 and @xmath83 respectively , which indicates that the em fields inside the sphere have also been strengthened at the corresponding frequencies . in contrast , at the antiresonances ( @xmath135 for @xmath130 or @xmath136 for @xmath132 ) no such electric or magnetic dipolar radiation can be detected outside the sphere since the corresponding scattering coefficients vanish at these particular @xmath18 . however , the electric and magnetic dipolar modes still exist inside the sphere , since the internal coefficients @xmath133 and @xmath134 are actually finite instead of zero as shown in fig . [ fig3](c ) .
the scattering coefficients @xmath137 and @xmath138 and internal coefficients @xmath68 and @xmath69 for a ti sphere should be similar to those for a dielectric sphere with the same refractive index , since the fine structure constant is in the order of @xmath139 . as a result , tme effect could not be manifested in the scattering coefficients @xmath137 , @xmath138 and internal coefficients @xmath75 .
however , it plays a key role in the cross - polarized coefficients @xmath140 and @xmath141 since these four coefficients should vanish in conventional dielectric spheres .
interestingly , there also exist resonances and _ antiresonances _ for @xmath61 and @xmath62 .
the cross - polarized scattering coefficients vanish at the antiresonances . in order to guarantee @xmath142 , the condition @xmath143^{^{\prime } } = 0 $ ] should be satisfied . in other words ,
the antiresonances correspond to the roots of @xmath144 or @xmath145^{^{\prime } } = 0 $ ] exactly . in fig .
[ fig3](b ) , we take the dipolar mode as an example .
the first antiresonance at @xmath146 corresponds to @xmath147^{^{\prime } } = 0 $ ] , whereas the second antiresonance at @xmath148 corresponds to @xmath149 .
although the corresponding scattering fields are zero outside the sphere at the antiresonances , the fields inside the sphere are not necessarily zero . for instance , by checking the fields inside the ti sphere at @xmath146 , it is found that cross - polarized electric dipolar mode could be exited since @xmath150 whereas cross - polarized magnetic dipolar mode does not exist since @xmath151 . on the contrary , for @xmath148 we have @xmath152 and @xmath153 as shown in fig .
[ fig3](d ) , which implies the existence of cross - polarized magnetic dipolar mode and the lack of cross - polarized electric dipolar mode . at antiresonances @xmath146 . in ( a ) and
( b ) , the field components @xmath154 and @xmath155 of total fields contributed by the superposition of electric and magnetic dipolar modes are shown respectively . in ( c ) and
( d ) , the components @xmath156 and @xmath157 are trapped inside the sphere , which are contributed from the cross - polarized electric dipolar mode due to the tme effect .
( e ) and ( f ) show the field components @xmath154 and @xmath155 of the electric dipolar mode arised from bulk scattering ; while in ( g ) and ( h ) , @xmath154 and @xmath155 of the magnetic dipolar mode from bulk scattering are plotted . the field components @xmath158 and @xmath159 vanish at the boundary of the ti sphere .
the physical parameters are the same as those in fig .
the boundary of the ti sphere is shown by the dashed lines.,width=377 ] from the definition of the vector spherical harmonics @xcite , the tangential components ( i.e. , @xmath160 and @xmath161 ) of @xmath162 and @xmath163 are proportional to @xmath145^{^{\prime } } $ ] and @xmath164 respectively .
thus , the underlying physics for the existence of anti - resonances of cross - polarized scattering coefficients is that the tangential field components corresponding to @xmath162 or @xmath163 vanish at the boundary of sphere .
note that the electric(magnetic ) fields of electric multipoles are contributed from @xmath162(@xmath163 ) , while those of magnetic multipoles are proportional to @xmath163(@xmath162 ) .
in other words , the corresponding tangential field components of the electric and magnetic multipoles should vanish at the boundary . in most cases ,
the electric and magnetic multipolar fields yield from the bulk scattering are strong and the cross - polarized ones induced by tme effect can not manifest themselves . however , tme effect can be clearly observed at two particular scattering planes , i.e. , planes with @xmath165 and @xmath166 . in fig .
[ fig4 ] , we show the field distributions at scattering plane @xmath165 for the first antiresonance at @xmath146 which corresponds to a root of @xmath147^{^{\prime } } = 0 $ ] .
the upper panel of fig .
[ fig4 ] shows the total fields distribution , the dominant field components are @xmath167 and @xmath168 , which are the superposition of the fields from electric dipole ( polarized in @xmath18 direction ) and magnetic dipole ( polarized in @xmath169 direction ) modes due to the bulk scattering .
the field patterns are almost the same with those for conventional dielectric spheres .
interestingly , the field components @xmath170 and @xmath171 induced by cross - polarized electric dipole(polarized along @xmath169 direction ) are nonzero and trapped inside the ti sphere due to tme effect , whereas for conventional dielectric spheres these two components should be zero . as shown in fig . [
fig4](c ) and ( d ) , the cross - polarized fields induced by tme effect are completely trapped inside the sphere .
observation of the trapping effect can be made clearly since the incident and scattered waves do nt have these two components(@xmath170 and @xmath171 ) at the scattering plane @xmath165 . for conventional spheres @xcite ,
the scattered waves are reduced to zero at antiresonances and thus the spheres become transparent , however , the corresponding field components of incident waves still exist inevitably .
in contrast , at the antiresonances of ti cylinders , the electric multipolar modes ( corresponding to @xmath137 ) and magnetic multiple modes ( corresponding to @xmath138 ) due to the bulk scattering can be excited separately via waves with different polarization , i.e. , te waves and tm waves respectively . as is well known , for ti spheres electric multipolar modes ( @xmath137 ) and magnetic multipolar modes ( @xmath138 )
will always be excited simultaneously .
therefore , at the antiresonances , the total fields of bulking scattering at the boundary can not be reduced to zero in general
. however , the tangential components of the electric and magnetic fields for respectively the electric and magnetic multipolar modes , are zero as shown in fig .
[ fig4](e ) and [ fig4](h ) for the dipolar case with @xmath147^{^{\prime } } = 0
$ ] satisfied .
the vanishment of the tangential field components at antiresonances @xcite for electric and magnetic dipoles is an exotic phenomenon and also a strong evidence that the tme effect is intrinsically a surface effect . for the second antiresonance at @xmath148 similar conclusions
can be drawn .
the field trapping for the cross - polarized magnetic dipole ( note that @xmath172 here ) inside the sphere can also be observed .
a quantitative determination of the tme effect can also be found in ti spheres similar to the case in ti cylinder @xcite .
we can conduct a rayleigh - scattering experiment with incident plane waves as shown in fig .
[ fig1 ] schematically . in the far field ,
both electric - field components of scattered waves @xmath173 and @xmath174 are measurable quantities and we can thus define a quantity @xmath175 . in the rayleigh limit , the scattering is dominantly contributed from dipoles .
assuming that @xmath176 we have : @xmath177 where @xmath178 is basically a bulk parameter . to extract the fine structure constant , we firstly focus on the scattering angles that re in the first quadrant , i.e. @xmath179 .
obviously , @xmath180 is a small term compared to @xmath181 , @xmath182 and hence could be neglected except in the vicinity @xmath183 . with these assumptions
we have : @xmath184}\]]this offers a simple way to determine @xmath185 if @xmath186 is measured and the parameter @xmath187 is known .
clearly , the parameter @xmath187 is material - dependent and so does this measurement .
however , if we set a special observation anlge @xmath188 , from eq .
( [ alpha ] ) we have : @xmath189owing to the complete suppression of bulk scattering in this particular angle , the to - be measured @xmath185 now can be solely determined by the value of @xmath190 which is material - independent . from the obtained @xmath185 , it offers a way to measure the fine structure constant @xmath191 since the axion angle is quantized .
we ve studied em wave scattering by ti spheres systematically .
the scattering coefficients for ti spheres are derived analytically .
we ve found that the scattering patterns of ti spheres showing strong backward scattering with impedance - matched background @xmath192 .
the enhanced backscattering or the cancellation of the forward scattering is non - trivial since it is a strong manifestation of the tme effect where the em scattering due to the surface of tis play a key role .
the strong backward scattering has a broadband feature and holds for both mie and rayleigh scattering regimes due to the non - resonant nature of tme effect .
the anti - resonances in the mie scattering regime are another kind of novel phenomenon of ti spheres , wherein the cross - polarized fields induced by tme effect are trapped inside the ti spheres .
finally , we proposed a simple way to determine the quantized tme effect of tis in the rayleigh limit by measuring the electric - field components of scattered waves in the far field .
this work is supported by the 973program ( grant nos .
2013cb632701 and 2011cb922004 ) . the research of h.d.z .
is further supported by the national natural science foundation of china ( grant no .
11304038 ) and the fundamental research funds for the central universities ( grant no .
cqdxwl-2014-z005 ) .
b. s. lukyanchuk , m. i. tribelsky , z. b. wang , m. h. h. y. zhou , l. p. shi , t. c. chong , extraordinary scattering diagram for nanoparticles near plasmon resonance frequencies , appl .
phys . a 89 ( 2007 ) 259264 .
j. m. geffrin , b. garca - cmara , r. gmez - medina , et al . ,
magnetic and electric coherence in forward- and back - scattered electromagnetic waves by a single dielectric subwavelength sphere , nat .
commun . 3 ( 2012 ) 11711178 . | the electromagnetic scattering properties of topological insulator ( ti ) spheres are systematically studied in this paper .
unconventional backward scattering caused by the topological magneto - electric ( tme ) effect of tis are found in both rayleigh and mie scattering regimes .
this enhanced backward scattering can be achieved by introducing an impedance - matched background which can suppress the bulk scattering . for the cross - polarized scattering coefficients ,
interesting antiresonances are found in the mie scattering regime , wherein the cross - polarized electromagnetic fields induced by the tme effect are trapped inside ti spheres . in the rayleigh limit
, the quantized tme effect of tis can be determined by measuring the electric - field components of scattered waves in the far field . |
particles of high - spins @xcite continue being among the most enigmatic challenges in contemporary theoretical physics .
the difficulties in their descriptions , both at the classical- , and the quantum - field theoretical levels , are well known and take their origin from the circumstance that such particles are most naturally described by differential equations of orders twice their respective spins @xcite-@xcite .
higher - order theories are difficult to tackle and various strategies have been elaborated over the years to lower the order of the corresponding differential equations , the linear ones by rarita - schwinger @xcite being the most popular so far .
however , the latter framework is plagued by various inconsistencies , the acausal propagation within an electromagnetic environment @xcite , the violation of unitarity in compton scattering in the ultraviolet in schemes with minimal gauge couplings @xcite , and the violation of lorentz - symmetry upon quantization , being the most serious ones . in parallel , also second order spin-@xmath8 @xcite , @xcite , @xcite , @xcite and spin-@xmath9 @xcite fermion theories have been developed by different authors and shown to provide a reasonable compromise between the rigorous linear and the natural higher - order descriptions in so far as they were able to circumvent , among others , the acausality problem and the related violation of unitarity in compton scattering @xcite .
however , for spins higher than @xmath9 no second order theory has been developed so far .
it is the goal of the present work to fill this gap .
the interest in such a study is motivated by the observation that distinct representation spaces of the lorentz algebra @xmath10 describe particles of different physical properties .
for example , due to the representation dependence of the boost operator , the electromagnetic quadrupole and octupole moments of fundamental particles with spin-@xmath9 transforming in the four - vector spinor come out different from those of particles transforming as @xmath11 @xcite .
same holds valid regarding spin-@xmath12 in the four - vector , @xmath13 , versus the anti - symmetric tensor , @xmath14 @xcite . in view of the expected production of new particles in the experiments run by the large hadron collider it is important to have at ones disposal a reliable and comfortable to deal with universal calculation scheme for high spins transforming in carrier spaces of the lorentz algebra different from the totally symmetric tensors of common use , the weinberg - joos states being the prime candidates .
the present study is devoted to the elaboration of such a scheme .
+ the path we take is to embed single spin-@xmath0 weinberg - joos states @xcite-@xcite , @xmath7 , into direct sums of properly selected irreducible @xmath10 representation spaces which join to reducible representation spaces that are large enough as to allow to be equipped by lorentz , and if needed , separate dirac indexes
. then we pin down the state of our interest in a two step procedure .
first we pin down the @xmath1 irreducible representation space by means of a momentum independent ( static ) projector designed on the basis of one of the casimir invariants of the lorentz algebra and then impose on it the klein - gordon equation . in this fashion , a second order formalism for any single - spin valued weinberg- joos state is furnished .
the scheme also allows for an extension to spin-@xmath0 as the highest spins in the two - spin valued representation spaces , @xmath15 .
+ the paper is organized as follows . in the next section
we formulate the concepts underlying our suggested method . in section iii
we design the calculation algorithm and write down a second order master equation for any single spin-@xmath0 transforming as @xmath7 . in same section
we apply the scheme elaborated to the description of the particular case of spin-@xmath9 . in section
iv we present the spin-@xmath9 lagrangian and couple it to the electromagnetic field , find the electromagnetic current , and calculate the associated electromagnetic multipole moments .
we show that the observables obtained in this fashion reproduce the predictions reported in @xcite where the pure spin-@xmath9 states have been considered in the standard way as eight - dimensional spinors .
also there we show that the pure spin-@xmath9 sector of the antisymmetric tensor spinor describes a particle that propagates causally within an electromagnetic environment .
finally , we calculate the process of compton scattering off @xmath16 and report on a finite differential cross section in forward direction .
the paper closes with brief conclusions and has one appendix .
the method for high - spin description advocated in this work is based upon representation spaces of the lorentz algebra @xmath10 , which are as a rule different from those of common use
. while the representation spaces underlying the weinberg - joos formalism are of non - tensorial nature , those underlying the rarita - schwinger framework are totally symmetric tensors .
our method permits to consider besides totally symmetric also totally antisymmetric lorentz tensors , and also such of mixed symmetries .
our idea is to embed @xmath7 carrier spaces of @xmath10 into finite direct sums of properly chosen dummy irreducible representation spaces with the aim to end up with reducible representation spaces of minimal dimensionality which are nonetheless large enough as to allow to be equipped by lorentz ( and if needed , separate dirac ) indexes , i.e. @xmath17 \oplus \sigma _ { ( k , l ) } \ , n_{(kl ) } \left [ \left(j_k , j_l\right)\oplus \left(j_l , j_k\right)\right].\nonumber\\ \label{chudo_1}\end{aligned}\ ] ] for example , pure spin-@xmath9 can be embedded into the totally antisymmetric tensor of second rank , @xmath18}$ ] , with dirac spinor components , @xmath19 , a representation space that is reducible according to , @xmath20 } & \simeq&b_{\left[\mu\nu \right]}\otimes \psi\nonumber\\ & \simeq & \left [ ( 1,0)\oplus ( 0,1)\right]\otimes \left[\left(\frac{1}{2},0\right)\oplus \left(0 , \frac{1}{2 } \right ) \right]\nonumber\\ & \longrightarrow & \left[\left ( \mathbf { \frac{3}{2 } } , { \mathbf 0}\right)\oplus \left({\mathbf 0 } , \mathbf{\frac{3}{2 } } \right)\right]\oplus \left[\left(\frac{1}{2},0\right ) \oplus \left(0 , \frac{1}{2 } \right)\right]\nonumber\\ & \oplus & \left[\left ( 1,\frac{1}{2}\right)\oplus \left(\frac{1}{2 } , 1 \right ) \right ] \nonumber\\ & \longrightarrow & \left[\left ( \mathbf { \frac{3}{2 } } , { \mathbf 0}\right)\oplus \left({\mathbf 0 } , \mathbf{\frac{3}{2 } } \right)\right ] \oplus \psi_\mu,\nonumber\\ \psi_\mu & \simeq & \left(\frac{1}{2},\frac{1}{2}\right ) \otimes\left [ \left(\frac{1}{2},0 \right ) \oplus \left(0,\frac{1}{2 } \right)\right ] , \label{tensor_spinor}\end{aligned}\ ] ] where we use @xmath21 $ ] to denote an antisymmetric pair of indexes .
spin-@xmath22 is part of the antisymmetric tensor with anti - symmetric tensor components , @xmath23\left[\eta \gamma \right ] } & \simeq&b_{\left[\mu\nu \right]}\otimes b_{\left[\eta\gamma \right]}\nonumber\\ & = & \left [ ( 1,0)\oplus ( 0,1)\right]\otimes \left[\left(1,0\right)\oplus \left(0,1\right)\right ] \nonumber\\ & \longrightarrow & \left[\left({\mathbf 2},0 \right)\oplus \left(0,{\mathbf 2 } \right)\right]\oplus 2\left(0,0 \right)\oplus 2\left(1,1 \right)\nonumber\\ & \oplus & \left [ \left(1,0 \right)\oplus \left(0,1 \right)\right ] , \label{tensor_vector}\end{aligned}\ ] ] where the numbers in front of the irreps indicate their multiplicity upon reduction .
+ similarly , spin-@xmath24 can be viewed as a resident of the direct product of the antisymmetric tensor - tensor from ( [ tensor_vector ] ) with a dirac spinor , giving rise to @xmath25\left[\eta\gamma\right]}$ ] , a representation space reducible according to @xmath26\left [ \eta\gamma \right ] } & \otimes & \psi \simeq \psi_{\left [ \mu \nu\right]\left [ \eta \gamma\right ] } \nonumber\\ & \simeq & \left [ ( 1,0)\oplus ( 0,1)\right]\otimes \left [ ( 1,0)\oplus(0,1 ) \right]\nonumber\\ & \otimes & \left[\left(\frac{1}{2},0\right)\oplus \left(0 , \frac{1}{2 } \right ) \right ] \nonumber\\ & \longrightarrow & \left[\left(\mathbf{\frac{5}{2}},0\right)\oplus \left(0 , \mathbf{\frac{5}{2 } } \right ) \right]\oplus 3\left[\left ( 1,\frac{1}{2}\right)\oplus \left(\frac{1}{2 } , 1 \right ) \right]\nonumber\\ & \oplus & 2\left[\left(1,\frac{3}{2}\right)\oplus \left(\frac{3}{2 } , 1 \right ) \right ] \oplus 3\left[\left ( \frac{1}{2},0\right)\oplus \left(0,\frac{1}{2 } \right ) \right]\nonumber\\ & \oplus & 2\left[\left ( \frac{3}{2},0\right)\oplus \left(0,\frac{3}{2 }
\right ) \right ] \oplus \left[\left(2,\frac{1}{2}\right)\oplus \left(\frac{1}{2 } , 2\right ) \right].\nonumber\\ \label{tensor_4vspinor}\end{aligned}\ ] ] in order to pin down the @xmath27 sector of our interest we have to remove the dummy irreducible sectors in the above reducible tensorial representation spaces . for this purpose
we develop an algorithm on the basis of static projectors constructed from one of the casimir invariants of the lorentz algebra . such projectors have the property to unambiguously identify and exclude anyone of the irreducible representation spaces , no matter whether single- or multiple - spin valued @xcite .
below we outline this representation reduction algorithm .
the lorentz algebra has a casimir operator , denoted by @xmath28 and given in @xcite in terms of the lorentz - group generators , @xmath29 , as @xmath30_{ab}&=&\frac{1}{4}[m^{{\mu}{\nu}}]_a{}^c [ m_{{\mu}{\nu}}]_{c b } , \label{fg_c}\end{aligned}\ ] ] with the capital latin letters @xmath31,@xmath32 , @xmath33 , ... standing for the generic indexes characterizing the representation space of interest .
its eigenvalue problem for any irreducible representation spaces of the type @xmath34 , with the generic representation functions denoted by , @xmath35 , reads , @xmath36 where @xmath37 on the basis of @xmath28 we design the following momentum independent lorentz projector , @xmath38 , @xmath39w^{(j_k , j_l ) } = w^{(j_1,j_2 ) } , \label{feq12}\end{aligned}\ ] ] where @xmath40 denotes the operation of successive multiplication , @xmath41 is the @xmath28 eigenvalue of the searched sector , @xmath42 , while @xmath43 are the @xmath28 eigenvalues of the dummy sectors , @xmath44 , of the hosting tensor , which need all to be excluded .
the mayor advantage of such projectors is that they are of zeroth order in the momenta , and do not contribute at all to the order of the wave equations . in what follows
we shall mainly consider only such reducible lorentz tensors ( or , tensors - spinors ) which allow the spin-@xmath0 of our interest to reside within one single - spin valued irreducible subspace , @xmath1 , i.e. in @xmath45 however , in the next section we show that the algorithm suggested allows for an extension also toward double - spin valued spaces such as @xmath46 with @xmath47 , i.e. @xmath48 the dynamics into the spin-@xmath0 sectors of interest from above is introduced by implementing the mass - shell condition , @xmath49 for the case of single spins transforming as @xmath1 . for double - spin valued representation spaces , the lorentz projector only tracks down the @xmath50 sector as a whole , but does not distinguish between its @xmath51 , and @xmath52 residents . in order to single out , say , the spin-@xmath0 , a different projector , here denoted by @xmath53 , and based on the squared pauli - lubanski vector , has to be employed .
the mass projector , @xmath54 , and the spin-@xmath0 projector , @xmath53 , will be occasionally referred to as `` poincar projectors '' @xcite in reference to the fact that they express in terms of the casimir invariants of the poincar algebra , the squared four momentum , @xmath55 , and the squared pauli - lubanski vector , @xmath56 , as @xmath57 here , @xmath58 denotes the pauli - lubanski ( pseudo)vector , defined as @xmath59 where @xmath60 are the generators of the lorentz algebra in the representation space of interest , while @xmath31 , and @xmath32 are again as already mentioned above the sets of indexes that characterize the dimensionality of that very representation space .
furthermore , @xmath61 , and @xmath62 , are the respective eigenvalues of the eigenstates of the operator @xmath56 , corresponding to the spins-@xmath0 , and @xmath63 , residing within @xmath50 , with the common mass-@xmath64 satisfying @xmath65 .
+ combining now the equation ( [ feq12 ] ) either with ( [ p2j0 ] ) , or with ( [ w2eq120])([gen_pl ] ) , the following master equations emerge , @xmath66_{b}{}^{a}\ , w^{(j,0)}_{a}=w^{(j,0)}_{b } , \label{masterequations_1}\end{aligned}\ ] ] and @xmath67_{d}{}{}^{c}\ , w^{\left(j_1,\frac{1}{2}\right);j}_{c } = w^{\left(j_1,\frac{1}{2}\right);j}_{d},\nonumber\\ \label{masterequations_2}\end{aligned}\ ] ] respectively . here , the functions , @xmath68 , and @xmath69 , have the property each to simultaneously diagonalize both the lorentz and the poincar projectors .
the indexes @xmath31 for bosons are given by , @xmath70 , while those for fermions carry in addition a dirac index , denoted by small latin letters , @xmath71 , @xmath72 , etc .
according to @xmath73 .
along this path , one necessarily encounters lagrangians that are second order in the momenta .
second order fermion approaches have traditions in field theory @xcite,@xcite , and are of growing popularity in qed as well as in qcd @xcite , @xcite , @xcite .
the present work focuses on the description of the pure spin-@xmath9 weinberg - joos state , @xmath74 as part of the antisymmetric tensor - spinor of second rank in ( [ tensor_spinor ] ) .
+ fermionic representation spaces of the types ( [ tensor_spinor ] ) , ( [ tensor_4vspinor ] ) have been earlier employed by niederle and nikitin in @xcite in a linear framework of the rarita - schwinger type , with the special emphasis on spin-@xmath9 .
however , the focus of @xcite has been in first place the separation of parities , while the question on the precise assignment of spin-@xmath9 to the irreducible @xmath11 sector of the anti - symmetric tensor spinor of second rank has been left aside .
moreover , differently from the present work , no physical observables have been calculated in @xcite for the sake of a comparison to predictions by other formalisms .
finally , the scheme of @xcite confines to fractional spins alone , while the one elaborated here applies to both fermions and bosons .
the antisymmetric lorentz tensor - spinor of second rank has been defined in eq .
( [ tensor_spinor ] ) and is the ordinary anti - symmetric lorentz - tensor of second rank , @xmath76 , with dirac spinor , @xmath77 , components .
the @xmath78 sector is well known and has been frequently elaborated in the literature , listed among others in @xcite,@xcite .
we here present it in the momentum space , and denote it by , @xmath79}({\mathbf p},\lambda^\prime ) $ ] with @xmath80 standing for the three momentum , and @xmath81 denoting the polarization quantum number , taking the values @xmath82 .
the spin-@xmath12 tensor @xmath79}({\mathbf p},\lambda^\prime ) $ ] allows for a representation in terms of the three spin-@xmath12 basis states , @xmath83 , and the one spin-@xmath84 state , @xmath85 , spanning the four - vector space , @xmath86 .
these states have been constructed for example in @xcite , @xcite and are summarized in the equations ( [ eta1pl])([eta0 ] ) in the appendix below . in terms of the aforementioned @xmath86 basis vectors , the tensor under discussion expresses as , @xmath87}({\mathbf p},\lambda^\prime ) = \eta ^\alpha ( { \mathbf p } , 0,0)\eta^\beta ( { \mathbf p},1,\lambda^\prime ) - \eta^\alpha ( { \mathbf p},1,\lambda^\prime ) \eta ^\beta ( { \mathbf p } , 0,0).\ ] ] now the tensor - spinor of interest here , to be denoted by @xmath88}({\mathbf p},\lambda^\prime , \lambda^\prime{}^\prime ) $ ] , its dual being @xmath89}({\mathbf p},\lambda^\prime , \lambda^\prime{}^\prime ) $ ] , reads , @xmath90}_\pm ( { \mathbf p},\lambda^\prime , \lambda^\prime{}^\prime ) & = & { \mathcal b}^{\left [ \alpha\beta\right]}({\mathbf p},\lambda^\prime ) \otimes u_\pm ( { \mathbf p},\lambda^\prime{}^\prime ) , \label{ts_msp}\end{aligned}\ ] ] where @xmath91 and @xmath92 denote in their turn the @xmath93 and the @xmath94 dirac spinors of positive / negative parities , and @xmath95 .
this tensor is reducible according to ( [ tensor_spinor ] ) and its irreducible sectors can be identified by means of projectors constructed from one of the casimir invariants of the lorentz algebra , an issue on which we shed light in subsection 3.3 below . before that , in the subsequent section we present the generators of the lorentz algebra in the anti - symmetric tensor - spinor of second rank .
the generators within the anti - symmetric tensor - spinor ( ats ) of second rank are @xmath96_{\left [ { \alpha}{\beta}\right]\left[{\gamma}{\delta}\right]}&=&[m^{at}_{{\mu}{\nu}}]_{\left[{\alpha}{\beta}\right]\left[{\gamma}{\delta}\right]}\mathbf{1}^s+\mathbf{1}_{\left[{\alpha}{\beta}\right]\left[{\gamma}{\delta}\right]}\,\ , \left[m^s_{{\mu}{\nu}}\right ] , \label{def : gensats } \\
\mathbf{1}_{\left[{\alpha}{\beta}\right]\left[{\gamma}{\delta}\right]}&=&\frac{1}{2}(g_{{\alpha}{\gamma}}g_{{\beta}{\delta}}-g_{{\alpha}{\delta}}g_{{\beta}{\gamma}}),\\ m^{s}_{{\mu}{\nu}}&=&\frac{1}{2}{\sigma}_{{\mu}{\nu}}=\frac{i}{4}[{\gamma}_{\mu},{\gamma}_{{\nu}}].\label{genss } \label{identityat}\end{aligned}\ ] ] here , @xmath97_{\left[{\alpha}{\beta}\right]\left[{\gamma}{\delta}\right]}$ ] are the generators within the anti - symmetric tensor ( at ) space , @xmath98\left[{\gamma}{\delta}\right]}$ ] stands for the identity in this space , while @xmath99 and @xmath100 are the unit operator and the generators within the dirac space ( [ genss ] ) , where @xmath101 are the standard dirac matrices . in what follows
we shall suppress the dirac indexes for the sake of avoiding cumbersome notations and will keep only the lorentz indexes .
we always will mark anti - symmetric pairs of lorentz indexes by @xmath102 $ ] any times when more than one pair is involved .
the @xmath97_{\left[{\alpha}{\beta}\right]\left[{\gamma}{\delta}\right]}$ ] generators express in terms of the generators in the four - vector , @xmath103_{{\alpha}{\beta}}&=&i(g_{{\alpha}{\mu}}g_{{\beta}{\nu}}-g_{{\alpha}{\nu}}g_{{\beta}{\mu}}),\label{gensv}\\end{aligned}\ ] ] as , @xmath104_{\left[{\alpha}{\beta}\right]\left[{\gamma}{\delta}\right]}&=&\frac{1}{2 } { \big(}\left[m^v_{{\mu}{\nu}}\right]_{{\alpha}{\gamma}}g_{{\beta}{\delta}}+g_{{\alpha}{\gamma } } \left[m^v_{{\mu}{\nu}}\right]_{{\beta}{\delta } } -\left[m_{{\mu}{\nu}}^v\right]_{{\alpha}{\delta}}g_{{\beta}{\gamma}}- g_{{\alpha}{\delta } } \left[m_{{\mu}{\nu}}^v\right]_{{\beta}{\gamma}}{\big)}\nonumber\\ & = & -2\,\,\mathbf{1}_{\left[{\alpha}{\beta}\right]}{}^{\left[{\kappa}\sigma \right]}\left[m^v_{{\mu}{\nu}}\right]_{\sigma } { } ^{\rho}\mathbf{1}_{\left[{\rho}{\kappa}\right]\left[{\gamma}{\delta}\right]}.\end{aligned}\ ] ] then , the explicit expression for the casimir invariant @xmath28 in ( [ fcasm ] ) takes the form , @xmath30_{\left[{\alpha}{\beta}\right]\left[{\gamma}{\delta}\right]}&= & -\frac{1}{8}{\big(}{\sigma}_{{\alpha}{\beta}}{\sigma}_{{\gamma}{\delta}}- { \sigma}_{{\gamma}{\delta}}{\sigma}_{{\alpha}{\beta}}-22\ , \mathbf{1}_{\left[{\alpha}{\beta}\right]\left[{\gamma}{\delta}\right]}{\big)}.\nonumber\\ \label{f_antes}\end{aligned}\ ] ] in the antisymmetric tensor - spinor space under investigation there are three lorentz sectors of the type @xmath106 , corresponding to @xmath107 = @xmath108 , @xmath109 , and @xmath110 .
the associated representation functions , @xmath111 , are characterized by their @xmath41 eigenvalues with respect to the @xmath28 invariant according to the equations ( [ fcasm ] ) , and ( [ fev ] ) , @xmath112 with @xmath113 all three eigenvalues are different and given by , @xmath114 this allows us to define the three independent operators , @xmath115 , and @xmath116 as @xmath117 which have the properties to remove from the antisymmetric tensor - spinor in ( [ tensor_spinor ] ) the dummy @xmath118 , and @xmath119 , companions to @xmath75 .
specifically , the lorentz projector for the irreducible sector @xmath120 is cast into the following form , @xmath121 with @xmath122 , @xmath123 , @xmath124 from eq .
( [ tr_org ] ) . in this way we find the following lorentz projector on @xmath120 , @xmath125_{\left[{\alpha}{\beta}\right]\left[{\gamma}{\delta}\right]}&=&\frac{1 } { 8 } ( { \sigma}_{{\alpha}{\beta}}{\sigma}_{{\gamma}{\delta}}+{\sigma}_{{\gamma}{\delta}}{\sigma}_{{\alpha}{\beta}})-\frac{1}{12}{\sigma}_{{\alpha}{\beta}}{\sigma}_{{\gamma}{\delta}},\nonumber\\ \label{cs_3}\end{aligned}\ ] ] which satisfies the conditions , @xmath126_{\left [ { \alpha}{\beta}\right]\left [ { \mu}{\nu}\right ] } & = & 0,\\ \left[\mathcal{p}^{\left(\frac{3}{2},0\right)}_f\right]_{\left [ { \alpha}{\beta}\right]\left[{\mu}{\nu}\right ] } { \gamma}^{\mu}&= & 0.\end{aligned}\ ] ] we now consider the action of the lorentz projector in ( [ eq30 ] ) on the tensor - spinor and its dual and generate in this way the set of twenty four momentum space wave functions , @xmath127^{\left [ \alpha\beta \right]}$ ] , and @xmath128^{\left [ \alpha\beta \right]}$ ] , as , @xmath129^{\left [ { \alpha}{\beta}\right]}&= & \left[\mathcal{p}^{\left(\frac{3}{2},0\right)}_f\right]^{\left [ \alpha \beta\right]}{}{}_{\left [ { \gamma}{\delta}\right ] } { \mathcal t}^{\left [ { \gamma}{\delta}\right]}_\pm ( { \mathbf p},\lambda^\prime , \lambda^\prime{}^\prime ) , \nonumber\\ \left[{\widetilde \phi}^{\left(\frac{3}{2},0\right)}_\pm \left({\mathbf p},\frac{3}{2},\lambda^\prime , \lambda^\prime{}^\prime \right)\right]^{\left [ { \alpha}{\beta}\right]}&= & \left[\mathcal{p}^{\left(\frac{3}{2},0\right)}_f\right]^{\left [ \alpha \beta\right]}{}{}_{\left [ { \gamma}{\delta}\right ] } { \widetilde { \mathcal t}}^{\left [ { \gamma}{\delta}\right]}_\pm ( { \mathbf p},\lambda^\prime , \lambda^\prime{}^\prime ) , \nonumber\\ \lambda^\prime&=&\pm 1,0 , \,\ , \lambda^\prime{}^\prime=\pm\frac{1}{2 } , \nonumber\\ \lambda&=&\pm\frac{1}{2},\pm \frac{3}{2}. \label{drct}\end{aligned}\ ] ] after some algebraic manipulations
, it can be verified that these are all @xmath130 eigenfunctions , meaning that they all reside in @xmath131 , although only eight of them are linearly independent , as it should be . out of them
, a set of pure spin-@xmath9 tensor spinors of positive and negative parities , henceforth denoted by @xmath132^{\left [ { \alpha}{\beta}\right]}$ ] , can be constructed and cast into the form of the lorentz projector @xmath133^{\left[{\alpha}{\beta}\right]}{}{}_{\left[{\mu}{\nu}\right]}$ ] applied to the anti - symmetric combination , @xmath134^{\left[{\mu}{\nu}\right]}$ ] , of the four - momentum , @xmath135 , with a four - vector spinor , @xmath136 , as explained in eq .
( [ ss32ats ] ) in the appendix . in other words , one can write , @xmath137^{\left [ { \alpha}{\beta}\right]}&= & 2 \left[\mathcal{p}^{\left(\frac{3}{2},0\right)}_f\right]^{\left[{\alpha}{\beta}\right]}{}{}_{\left[{\mu}{\nu}\right]}\left[u_\pm{\left(}\mathbf{p},\frac{3}{2},{\lambda}{\right)}\right]^{\left[{\mu}{\nu}\right]}\nonumber\\ & = & \frac{2}{m } \left[\mathcal{p}^{\left(\frac{3}{2},0\right)}_f\right]^{\left[{\alpha}{\beta}\right]}{}{}_{\left[{\mu}{\nu}\right ] } p^{\mu}{\mathcal u}^{\nu}_\pm\left(\mathbf{p},\frac{3}{2},{\lambda}\right ) , \label{spin32ats}\end{aligned}\ ] ] with @xmath138 defined in eqs .
( [ fvs324 ] ) in the appendix .
the factor 2 ensures the normalization of these states to @xmath139 .
to the amount @xmath140^{\left [ { \alpha}{\beta}\right]}$ ] are considered on their mass shell , they satisfy the klein - gordon equation , as explained in ( [ feq12 ] ) , ( [ masterequations_1 ] ) , and ( [ finv ] ) . in effect , the @xmath141 degrees of freedom are found to solve the following second order equation , @xmath142^{\left[{\alpha}{\beta}\right]}{}{}_{\left[{\gamma}{\delta}\right ] } p^2-m^2 \mathbf{1}^{\left[{\alpha}{\beta}\right]}{}{}_{\left[{\gamma}{\delta}\right]}{\big ) } \left[w^{\left(\frac{3}{2},0\right)}_\pm \left({\mathbf p},\frac{3}{2},\lambda\right)\right]^{\left[{\gamma}{\delta}\right ] } & = & 0.\,\,\ , \label{eqats32}\end{aligned}\ ] ] in this fashion , the pure spin-@xmath9 degrees of freedom are manifestly generated in the shape of tensor - spinors .
compared to the eight - component weinberg - joos spin-@xmath9 `` vectors '' , listed among others in @xcite , the eight tensor - spinors , @xmath143^{\left [ { \alpha}{\beta}\right]}$ ] , open an avenue towards efficient tensor calculations of scattering cross - sections off @xmath120 by the help of the symbolic software feyncalc , a reason for which we consider the equation ( [ drct ] ) as the first achievement of the present study , worth reporting .
the aforementioned tensor spinors do not satisfy the dirac equation , due to the nonzero commutator , @xmath144\not=0 $ ] , between the lorentz - projector and the feynman slash , @xmath145 .
this is a crucial circumstance for the neat separation of the rarita - schwinger from the weinberg - joos sector in the anti - symmetric lorentz tensor - spinor of second rank .
the tensor - spinors are conditioned through , @xmath146_{\left [ { \alpha}{\beta}\right]}=0 .
\label{ts_gcnd}\ ] ] we now introduce the short - hand @xmath147^{{\alpha}{\beta}{\mu}}$ ] for convenience as @xmath148^{{\alpha}{\beta}{\mu}}=\frac{2}{m } \left[\mathcal{p}_f^{(j_1,j_2)}\right]^{\left[{\alpha}{\beta}\right]\left[{\gamma}{\mu}\right]}p_{\gamma } , \label{f_tensors}\end{aligned}\ ] ] with @xmath149^{{\alpha}{\beta}}{}_{{\mu } } \left[f^{\left(\frac{3}{2},0\right)}(p)\right]_{{\alpha}{\beta}{\nu}}&=&\frac{1 } { m^2 } { \big(}g_{{\alpha}{\beta}}g_{{\mu}{\nu } } -\frac{1}{3}{\sigma}_{{\alpha}{\mu}}{\sigma}_{{\beta}{\nu}}-g_{{\alpha}{\mu}}g_{{\beta}{\nu}}{\big)}p^{\alpha}p^{\beta}. \nonumber\\ \label{eq : fintern320}\end{aligned}\ ] ] this allows us to write the tensor - spinors in ( [ spin32ats ] ) as @xmath150^{\left[{\alpha}{\beta}\right]}=\left[f^{(\frac{3}{2},0 ) } ( \mathbf{p})\right]^{{\alpha}{\beta}}{}{}_{\mu}\left[{\mathcal u}_\pm\left(\mathbf{p},\frac{3}{2},{\lambda}\right)\right]^{{\mu}}. \label{eq : atsvs}\ ] ] the conjugate tensor - spinors are then introduced as @xmath151_{\left[{\alpha}{\beta}\right]}= \overline{{\mathcal u}}_\pm^{\mu}\left(\mathbf{p},\frac{3}{2},{\lambda}\right ) \left[\overline{f}^{\left(\frac{3}{2},0\right ) } ( \mathbf{p})\right]_{{\alpha}{\beta}}{}{}^{\mu } , \label{eq : atsvsd}\ ] ] were @xmath152_{{\alpha}{\beta}{\mu}}={\gamma}^0{\left(}\left[f^{\left(\frac{3}{2},0\right ) } ( \mathbf{p})\right]_{{\alpha}{\beta}{\mu } } { \right)}^\dagger{\gamma}^0 $ ] . + the above spin-@xmath9 tensor - spinors are normalized as @xmath153_{\left[{\alpha}{\beta}\right ] } \left[w^{\left(\frac{3}{2},0\right)}_{\pm}\left(\mathbf{p},\frac{3}{2},{\lambda}\right)\right]^{\left[{\alpha}{\beta}\right]}=\pm 1 .
\label{spin_norm}\ ] ] we now calculate their propagator as , @xmath154_{\left[\alpha \beta\right ] \left[\gamma \delta\right ] } = \frac{\left[\delta^{\left(\frac{3}{2 } , 0\right)}(p)\right]_{\left[\alpha \beta\right ] \left[\gamma \delta\right ] } } { p^2- m^2+i{\epsilon}},\end{aligned}\ ] ] with @xmath155_{\left[\alpha \beta\right]\left [ \gamma \delta \right]}&= & \frac{p^2}{m^2}\left[\mathcal{p}^{\left(\frac{3}{2},0\right)}_f\right]_{\left[{\alpha}{\beta}\right]\left[{\gamma}{\delta}\right]}+\frac{{\left(}m^2-p^2{\right)}}{m^2}\mathbf{1}_{\left[{\alpha}{\beta}\right ] \left[{\gamma}{\delta}\right]}.\end{aligned}\ ] ] ll @xmath7 & @xmath156^{\lbrack \mu\nu \rbrack } $ ] .
+ & + tensor - spinor & @xmath157^{\lbrack \mu\nu\rbrack } $ ] , eq .
( [ ts_msp ] ) + and its dual : & @xmath158^{\lbrack \mu\nu\rbrack } $ ] + lorentz projector : & @xmath159 + & - @xmath160 , eq .
( [ cs_3 ] ) + & + primordial spin-@xmath9 & @xmath161_{\left [ \alpha \beta \right]}=\left [ { \mathcal p}_f^{\left(\frac{3}{2},0\right))}\right]_{\left [ \alpha\beta\right]\left [ { \gamma}{\delta}\right]}$ ] + & + tensor - spinors , & @xmath162^{\left[{\gamma}{\delta}\right ] } $ ] , eqs .
( [ drct ] ) + & + and dual tensor - spinors : & @xmath163_{\left [ \alpha \beta \right]}=\left [ { \mathcal p}_f^{\left(\frac{3}{2},0\right))}\right]_{\left [ \alpha\beta\right]\left [ { \gamma}{\delta}\right]}$ ] + & @xmath164^{\left[{\gamma}{\delta}\right]}$ ] + & + parity & @xmath140^{\left [ { \alpha}{\beta}\right ] } = \frac{2}{m } \left[\mathcal{p}^{\left(\frac{3}{2},0\right)}_f\right]^{\left[{\alpha}{\beta}\right]}{}{}_{\left[{\mu}{\nu}\right]}$ ] + representation : & + & @xmath165 , eq . ( [ spin32ats]),([ss32ats ] ) + & + & + wave equation : & @xmath166= @xmath167 + & @xmath168^{\lbrack \alpha\beta \rbrack}$ ] eq . ( [ eqats32 ] ) + & + electromagnetic & @xmath169 +
current : & @xmath170 + & @xmath171 , + & + & eq .
( [ elmcr ] ) + & + second order theories present the notorious problem that the propagators are of unspecified spatial parities . for bosons
this circumstance does not present an obstacle in so far as bosons and anti - bosons are of equal parities .
such is due to the commutativity of the parity and the charge conjugation operators .
a discussion on this issue can be found around the equation ( 2.34 ) in the reference @xcite . for massive charged fermions , however , for which particles and anti - particles are of opposite parities
, the problem is more serious and can affect the quantization procedure .
however , we expect that the method would allow for quantization of majorana fermions ( if they were to exist ) and massless particles . moreover , at the level of scattering processes , the problem of parity distinction of charged massive particles still can be attended by constructing the amplitudes for states of a fixed parity and then inserting into the squared amplitudes the relevant parity projector , a path that we take in the evaluation of the process of compton scattering as presented in subsection 5 of the next section .
the method is highlighted in table 1 .
as a next step after having designed the appropriate free equations of motion in ( [ masterequations_1 ] ) , ( [ eqats32 ] ) , we introduce the electromagnetic interaction .
we here confine to minimal gauging , i.e. the coupling is found by replacing ordinary by covariant derivatives according to , @xmath173 where @xmath174 is the electric charge of the particle . in order to obtain the gauged equations
, we first pass the states from momentum to position space using the standard quantization prescription , @xmath175^{\left [ { \alpha}{\beta}\right]}e^{\mp i\,x\cdot p}$ ] @xmath176 @xmath177^{\left [ { \alpha}{\beta}\right]}$ ] , and then write the momenta in operator form yielding , @xmath178_{[{\alpha}{\beta}]}{}{}^{[\gamma\delta]}\partial^2 [ { \psi}^{\left(\frac{3}{2},0\right)}(x,{\lambda})]^{\lbrack { \alpha}{\beta}\rbrack } = -m^2 [ { \psi}^{\left(\frac{3}{2},0\right ) } ( x,{\lambda})]^{\lbrack { \gamma}{\delta}\rbrack}. \label{fdeqs32_prl}\ ] ] we now define at the free particle level a new tensor , @xmath179^ { \left[\gamma \delta\right ] } { } { } _ { \left [ { \alpha}{\beta}\right]}$ ] , as @xmath180_{[{\alpha}{\beta}][{\gamma}{\delta}]}\partial ^{\mu}\partial ^{\nu}= [ \mathcal{p}^{\left(\frac{3}{2},0\right)}_f]_{[{\alpha}{\beta}][\gamma\delta]}\partial^2 , \label{g_3}\end{aligned}\ ] ] and then cast the gauged eq .
( [ fdeqs32_prl ] ) as , @xmath181_{[{\alpha}{\beta}][{\gamma}{\delta}]}d^{\mu}d^{\nu}[\psi^{\left(\frac{3}{2},0\right ) } ( x)]^{\lbrack { \alpha}{\beta}\rbrack } = -m^2 [ \psi^{\left(\frac{3}{2},0\right ) } ( x ) ] ^{\lbrack { \gamma}{\delta}\rbrack } , \label{fdeqs32}\ ] ] where we denoted by @xmath182^{\left [ { \alpha}{\beta}\right]}$ ] the new gauged solutions . in order to guarantee that the gauged solutions continue being eigenstates of the lorentz projectors and thereby to ensure equality of the number of the degrees of freedom before and after gaugung
, the @xmath183_{[{\alpha}{\beta}][{\gamma}{\delta}]}$ ] tensor has to satisfy , @xmath184_{[{\alpha}{\beta}]}{}{}^{\left[\sigma
\rho\right]}\left [ \gamma^{\left(\frac{3}{2},0\right)}_{\mu\nu}\right]_{\left[\sigma \rho\right][\gamma\delta]}= \left [ \gamma^{\left(\frac{3}{2},0\right)}_{\mu\nu}\right]_{[\alpha \beta][\gamma\delta]}.\label{g32tensor}\ ] ] once the validity of the equation ( [ g32tensor ] ) has been ensured , the representations space and therefore the spin after gauging continues being same as before .
it should be noticed that the @xmath185 tensor in ( [ g_3 ] ) is determined at the free particle level modulo additive terms leading to vanishing @xmath186 $ ] commutators prior gauging , which upon gauging give rise to the electromagnetic field strength tensor , @xmath187 and thereby to non - vanishing contributions to the gauged equations . as first discussed in @xcite ,
exploiting this freedom could be of help in achieving causality of particle propagation , and/or unitarity of scattering amplitudes in the ultraviolet . using the aforementioned freedom we chose @xmath189^ { \left[{\alpha}{\beta}\right ] } { } { } _ { \left[{\gamma}{\delta}\right]}$ ] in such a way that ,
@xmath190^{\left [ { \alpha}{\beta}\right ] } { } { } _ { \left[{\gamma}{\delta}\right]}=4\left[\mathcal{p}^{\left(\frac{3}{2},0\right)}_f\right]^{\left[{\alpha}{\beta}\right ] \left [ { \sigma } { \mu}\right]}\ \left[\mathcal{p}^{\left(\frac{3}{2},0\right)}_f\right]_{\left[\sigma { \nu}\right]\left[{\gamma}{\delta}\right]}&&\nonumber\\ = \frac{1}{2}{\big ( } { { \sigma}}^{\alpha\beta } { { \sigma}}^{\sigma\mu}{\mathbf { \sigma}}_{\sigma\nu}{{\sigma}}_{\gamma\delta}+ { { \sigma}}^{\sigma\mu}{{\sigma}}^{\alpha\beta}{{\sigma}}_{\gamma\delta}{{\sigma}}_{\sigma\nu}&&\nonumber\\ -3 { { \sigma}}^{\alpha\beta}{{\sigma}}^{\sigma\mu } { { \sigma}}_{\gamma\delta } { { \sigma}}_{\sigma\nu}-3{{\sigma}}^{\sigma\mu } { { \sigma}}^{\alpha\beta}{{\sigma}}_{\sigma\nu } { { \sigma}}_{\gamma\delta } { \big)}&&\nonumber\\ + 4{{\sigma}}^{\sigma\mu } { { \sigma}}^{\alpha\beta}{{\sigma}}_{\sigma\nu } { { \sigma}}_{\gamma\delta},&&\nonumber\\ \label{gamma - s32}\end{aligned}\ ] ] is satisfied , thus ensuring validity of ( [ g_3 ] ) and ( [ g32tensor ] ) .
this is the precise tensor that enters the gauged equation for the @xmath188 sector in ( [ fdeqs32 ] ) .
+ by virtue of ( [ g32tensor ] ) , on expects the solutions to to continue behaving as @xmath191 states and depending on only eight degrees of freedom . in order to see this , it is convenient to become aware of the fact that that upon accounting for the dirac label , the equation ( [ fdeqs32 ] ) in reality stands for a @xmath192 dimensional matrix equation
. it is straightforward to cross - check that the matrix in ( [ fdeqs32 ] ) in combination with ( [ gamma - s32 ] ) can be block - diagonalized with one of the blocks being precisely the expected @xmath193 dimensional one .
the corresponding eight - dimensional gauged solutions , @xmath194 , can be now written as linear combinations within any basis spanning the @xmath75 representation space .
in particular , if we choose the orthogonal set of the rest - frame states given by ( [ eq : atsvs ] ) , @xmath195 can be decomposed according to , @xmath196 as a reminder , the reduction to eight degrees of freedom is possible by virtue of the condition in ( [ g32tensor ] ) imposed by the lorentz projector on the @xmath197_{[\alpha \beta][\gamma\delta]}$ ] tensor which amounted to ( [ gamma - s32 ] ) . below
we show that the expansion of the gauged solutions in ( [ sollc ] ) significantly facilitates the proof of their causal propagation within an electromagnetic environment .
+ the causality and hyperbolicity of the equations of motion of order @xmath198 in the derivatives are tested using the courant - hilbert criterion , which requires us to calculate the so called `` characteristic determinant '' @xcite of the gauged equation .
the latter is found by replacing the highest order derivatives by the components of the vectors @xmath199 , interpreted as the normals to the characteristic surfaces , and which characterize the propagation of the ( classical ) wave fronts of the gauged equation . if the vanishing of the characteristic determinant demands to have a real - valued time - like component @xmath200 , then the equation is hyperbolic .
if this determinant nullifies for @xmath201 , then the equation is in addition causal @xcite .
the @xmath202 coefficients are calculated upon substitution of ( [ sollc ] ) in ( [ fdeqs32 ] ) , and making use of the normalization of the basis tensors in ( [ spin_norm ] ) as , @xmath203_{\left[{\alpha}{\beta}\right ] } \left[{\gamma}^{\left(\frac{3}{2 } , 0\right)}_{{\mu}\nu}\right]^ { \left[{\alpha}{\beta}\right ] } { } { } _ { \left [ { \gamma}{\delta}\right]}d^{\mu}d^{\nu}\left[\psi^{\left(\frac{3}{2},0\right)}(x ) \right]^{\left[{\gamma}{\delta}\right]}= - m^2 a_\pm(x ) _ { { \lambda}^\prime } , \ ] ] meaning that they satisfy the equation , @xmath204_{\left[{\alpha}{\beta}\right ] } & & \left[{\gamma}^{\left(\frac{3}{2 } , 0\right)}_{{\mu}\nu}\right]^ { \left[{\alpha}{\beta}\right ] } { } { } _ { \left [ { \gamma}{\delta}\right]}d^{\mu}d^{\nu}\sum_{{\lambda}^\prime { } ^\prime } a_\pm ( x ) _ { { \lambda}^\prime{}^\prime } \left[w^{\left(\frac{3}{2},0\right)}_{\pm}\left({\mathbf 0},\frac{3}{2},{\lambda}^\prime{}^\prime \right)\right]^{\left[{\gamma}{\delta}\right]}\nonumber\\ & = & - m^2 a_\pm(x ) _ { { \lambda}^\prime } , \label{eqfora}\\ d^\mu d^\nu&=&\partial^\mu\partial^\nu + ie(\partial^{\mu}a^{\nu } ) + iea^\nu\partial^\mu + iea^\mu\partial^\nu -e^2a^\mu a^\nu.\label{gd_prdct}\end{aligned}\ ] ] the equation represents a system of eight partial differential equation for the eight coefficients @xmath205 , because the polarization label @xmath206 runs over the four allowed values , @xmath207 , and is counted twice because of the two parities ( the lower case @xmath208 indexes ) , of the basis tensor - spinors , @xmath209 .
the elements of the characteristic matrix are proportional to @xmath210 , as can been seen from replacing the principal part , @xmath211 of @xmath212 in ( [ gd_prdct ] ) by @xmath213 yielding , @xmath204_{\left[{\alpha}{\beta}\right ] } \left[{\gamma}^{\left(\frac{3}{2 } , 0\right)}_{{\mu}\nu}\right]^ { \left[{\alpha}{\beta}\right ] } { } { } _ { \left [ { \gamma}{\delta}\right]}n^{\mu}n^{\nu}&&\nonumber\\ \times \eta_t\left[w^{\left(\frac{3}{2},0\right)}_{\pm } \left({\mathbf 0 } , \frac{3}{2},{\lambda}^\prime{}^\prime \right)\right]^{\left[{\gamma}{\delta}\right]}= \left[\overline{w}^{\left(\frac{3}{2},0\right)}_{\pm}\left(\mathbf{0},\frac{3}{2},{\lambda}^\prime \right)\right]_{\left[{\alpha}{\beta}\right]}\nonumber\\ \times n^2\left[\mathcal{p}_f^{\left(\frac{3}{2 } , 0\right)}\right]^ { \left[{\alpha}{\beta}\right ] } { } { } _ { \left [ { \gamma}{\delta}\right]}\eta_t \left[w^{\left(\frac{3}{2},0\right)}_{\pm}\left({\mathbf 0},\frac{3}{2},{\lambda}^\prime{}^\prime \right)\right]^{\left[{\gamma}{\delta}\right]}&&\nonumber\\ = n^2 \left[\overline{w}^{\left(\frac{3}{2},0\right)}_\pm \left({\mathbf 0},\frac{3}{2},{\lambda}^\prime\right)\right]_{\left[{\alpha}{\beta}\right ] } \eta_t\left[w^{\left(\frac{3}{2},0\right)}_{\pm}\left({\mathbf 0},\frac{3}{2},{\lambda}^\prime{}^\prime \right)\right]^{\left[{\alpha}{\beta}\right]}&&\nonumber\\ = n^2 \delta _ { { \lambda}^\prime { \lambda}^\prime { } ^\prime } \delta_{\pm \pm},&\quad & \nonumber\\ t=\pm , \ , \eta_+=-\eta_-=1.&&\nonumber\\\end{aligned}\ ] ] the characteristic matrix in question is therefore diagonal and each element equals @xmath210 , correspondingly , its determinant is @xmath214 nullifying the latter amounts to the condition @xmath215 , which guarantees causal propagation within the electromagnetic environment .
the free equations of motion ( [ fdeqs32 ] ) can now be derived from the following lagrangian : @xmath216^a{\big ) } [ { \gamma}_{{\mu}{\nu}}^{{\left(}\frac{3}{2},0{\right)}}]_{ab}{\partial}^{\nu}[{\psi}^{{\left(}\frac{3}{2},0{\right)}}]^b - m^2[\overline{{\psi}}^{{\left(}\frac{3}{2},0{\right)}}]^a[{\psi}^{{\left(}\frac{3}{2},0{\right)}}]_a,\nonumber\\ a&= & \left[\mu\nu\right],\quad b=\left [ { \gamma}{\delta}\right ] , \label{lagr}\end{aligned}\ ] ] where we suppressed the arguments for the sake of simplifying notations .
the gauged lagrangian then emerges as , @xmath217^a\right ) [ { \gamma}_{{\mu}{\nu}}^{{\left(}\frac{3}{2},0{\right)}}]_{ab}d^{\nu}[\psi^{{\left(}\frac{3}{2},0{\right)}}]^b -m^2[\overline{\psi}^{{\left(}\frac{3}{2},0{\right)}}]^a[\psi^{{\left(}\frac{3}{2},0{\right)}}]_a,\nonumber\\\end{aligned}\ ] ] and its decomposition into free and interacting parts reads , @xmath218 back to momentum space , we extract the following vector and tensor bi - linear forms , @xmath219^a [ \mathcal{v}^{\left(\frac{3}{2},0\right)}_{\mu}(p',p)]_{ab } \left[w_\pm^{\left(\frac{3}{2},0\right)}\left(\mathbf{p},\frac{3}{2},{\lambda}\right)\right]^b,\nonumber\\ \label{ecrrnt}\\ k_{{\mu}{\nu}}^{\left(\frac{3}{2},0\right)}({\mathbf p},{\lambda } , { \mathbf p}^\prime , { \lambda}^\prime)&=&e\left[\overline{w}_\pm ^{\left(\frac{3}{2},0\right)}\left(\mathbf{p}',\frac{3}{2},{\lambda}'\right)\right]^a [ \mathcal{c}^{{\left(}\frac{3}{2},0{\right)}}_{{\mu}{\nu}}]_{ab}\left[w_\pm^{{\left(}\frac{3}{2},0{\right)}}\left(\mathbf{p},\frac{3}{2},{\lambda}\right)\right]^b , \label{ktnsr}\end{aligned}\ ] ] the first standing for the electromagnetic current .
the vertexes are given as , @xmath220_{ab}&= & [ { \gamma}^{{\left(}\frac{3}{2},0{\right)}}_{{\nu}{\mu}}]_{ab}p'^{\nu}+ [ { \gamma}^{{\left(}\frac{3}{2},0{\right)}}_{{\mu}{\nu}}]_{ab}p^{\nu},\label{vert1}\\ ~[\mathcal{c}^{{\left(}\frac{3}{2},0{\right)}}_{{\mu}{\nu}}]_{ab}&=&\frac{1}{2}\ ( [ { \gamma}^{{\left(}\frac{3}{2},0{\right)}}_{{\mu}{\nu}}]_{ab}+[{\gamma}^{{\left(}\frac{3}{2},0{\right)}}_{{\nu}{\mu}}]_{ab } ) \label{vert2}.\end{aligned}\ ] ] the feynman rules following from this lagrangian are depicted in figs . [ propfig ] , [ regla1fig ] , [ regla2fig ] .
feynman rule for the two - photon contact vertex in ( [ vert2 ] ) . ]
feynman rule for the two - photon contact vertex in ( [ vert2 ] ) . ]
feynman rule for the two - photon contact vertex in ( [ vert2 ] ) . ]
it is not difficult to verify that the one - photon vertex obeys @xmath221_{ab}=[s^{{\left(}\frac{3}{2},0{\right ) } } ( p')]_{ab}-[s^{{\left(}\frac{3}{2},0{\right)}}(p)]_{ab } , \label{wt_id}\ ] ] which is the ward - takahashi identity , as it should be .
we now calculate the electromagnetic multipole moments of a particle transforming in the single spin- @xmath9 irreducible weinberg - joos sector of the antisymmetric tensor - spinor of interest from the current in ( [ ecrrnt ] ) , taken between positive parities .
this current can be further simplified in taking advantage of the equations ( [ spin32ats ] ) from above , of eqs . ( [ ss32atsprl])-([ss32ats ] ) in the appendix , and of the mass shell condition , yielding , @xmath222 the procedure of extracting the multipole moments from known currents is well established and will not be highlighted here ( see for example @xcite,@xcite and references therein for technical details ) , and amounts to @xmath223&=&e,\label{qe0}\\ ~[q_m^1({\lambda})]&=&\frac{2}{3}{\left(}\frac{e}{2m}{\right)}\langle m_{12}\rangle,\label{qm1}\\ ~[q_e^2({\lambda})]&=&-\frac{1}{3}{\left(}\frac{e}{m^2}{\right)}\langle { \mathcal a}\rangle,\label{qe2}\\ ~[q_m^3({\lambda})]&=&-2{\left(}\frac{e}{2m^3}{\right)}\langle { \mathcal b}\rangle,\label{qm3}\end{aligned}\ ] ] with @xmath224 the expressions in ( [ qe0])([qm3 ] ) fully coincide in form with the corresponding predictions by the weinberg - joos formalism reported in @xcite , where the calculation has been carried out while treating the states under consideration as eight - component spinors .
the only difference concerns the value of the gyromagnetic ratio which in the present work comes out fixed to the inverse of the spin , @xmath225 , and in accord with belinfante s conjecture , while in @xcite , a method exclusively based on the poincar covariant spin - projector alone , @xmath226 had remained unspecified according to , @xmath223_{ts}&=&e,\\ ~[q_m^1({\lambda})]_{ts}&=&g{\left(}\frac{e}{2m}{\right)}\langle m_{12}\rangle,\\ ~[q_e^2({\lambda})]_{ts}&=&-(1-g){\left(}\frac{e}{m^2}{\right)}\langle { \mathcal a}\rangle,\\ ~[q_m^3({\lambda})]_{ts}&=&-3 g{\left(}\frac{e}{2m^3}{\right)}\langle { \mathcal b}\rangle.\end{aligned}\ ] ] we conclude that the anti - symmetric tensor - spinor provides a description of particles of spin-@xmath9 transforming as @xmath141 that is equivalent to the weinberg - joos formalism . the great advantage of the tensor basis employed here that it allows to comfortably carry out more complicated calculations such a compton scattering , the subject of the next section .
the tree - level compton scattering amplitude contains contributions from the three different channels displayed in the figs .
[ mafig ] , [ mbfig ] , [ mcfig ] . in denoting by @xmath227 and @xmath228 in turn the four - momenta of the incident and scattered single spin-@xmath9 target particles , while by @xmath229 and @xmath230 the four - momenta of the incident and scattered photons , respectively , the amplitude takes the following form @xcite , @xmath231 in the following we shall evaluate the process off positive parity states , in which case one has , @xmath232^{a } & & \left [ { \mathcal u}_{{\mu}{\nu}}(p^\prime , q , p ) \right]_{ab}\nonumber\\ \left [ w_+^{\left(\frac{3}{2},0\right)}\left({\mathbf p},\frac{3}{2},{\lambda}\right ) \right]^{b } & & \left[{\epsilon}^{\mu}({\mathbf q}^\prime , \ell^\prime ) \right]^ * { \epsilon}^{\nu}({\mathbf q},\ell),\nonumber\\ \label{maj1}\\ i{\mathcal m}_2 = e^2\left [ { \overline w}_+^{\left(\frac{3}{2},0\right ) } \left({\mathbf p}^\prime , \frac{3}{2},{\lambda}'\right ) \right]^{a } & & \left[{\mathcal u}_{{\nu}{\mu}}(p',r , p)\right]_{ab}\nonumber\\ \left [ w_+^{\left(\frac{3}{2},0\right)}\left({\mathbf p},\frac{3 } { 2},{\lambda}\right)\right]^{b}&&[{\epsilon}^{\mu}(\mathbf{q } ' , \ell^\prime ) ] ^ * { \epsilon}^{\nu}({\mathbf q},\ell),\nonumber\\ \label{mbj1}\\ -i{\mathcal m}_3=e^2\left [ { \overline w}_+^{\left(\frac{3}{2},0\right)}\left({\mathbf p}^\prime , \frac{3 } { 2},{\lambda}'\right)\right]^{a}&&\left[{\mathcal x}_{{\mu}{\nu}}\right]_{ab}\nonumber\\ \left [ w_+^{\left(\frac{3}{2},0\right)}\left({\mathbf p},\frac{3}{2},{\lambda}\right)\right ] ^{b}&&\left[{\epsilon}^{\mu}({\mathbf q}^\prime , \ell^\prime ) \right]^*{\epsilon}^{\nu}({\mathbf q},\ell)\nonumber\\ \label{mcj1},\end{aligned}\ ] ] with @xmath233 and @xmath234 .
furthermore , we define the short - hands , @xmath235_{ab}&=&\left[\mathcal{v}^{\left(\frac{3 } { 2},0\right)}_{{\mu } } ( p',q)\right]_{ac } ~\left[s^{\left(\frac{3}{2},0\right ) } ( q)\right]^{cd}\nonumber\\ & & \left[\mathcal{v}^{\left(\frac{3}{2},0\right)}_{\nu}(q , p)\right]_{db},\\ ~[\mathcal{x}_{{\mu}{\nu}}]_{ab}&=&\left[\mathcal{c}_{{\mu}{\nu}}^{\left(\frac{3 } { 2},0\right)}+\mathcal{c}_{{\nu}{\mu}}^{\left ( \frac{3 } { 2},0\right)}\right]_{ab},\end{aligned}\ ] ] with @xmath236 and @xmath237 defined in ( [ vert1 ] ) and ( [ vert2 ] ) , respectively , and the indexes @xmath31,@xmath32 being introduced in ( [ lagr ] )
. the gauge invariance of this amplitude is ensured by the ward - takahashi identity in ( [ wt_id ] ) .
the averaged squared amplitude then reads , @xmath238^\dagger\\ & = & \frac{1}{8}tr\left[[\mathcal{m}_{{\mu}{\nu}}(p',q , r , p)]_{ab } [ { \mathcal m}^{{\nu}{\mu}}(p , r , q , p')]^{ab}\right ] , \label{eq : asa}\nonumber\\\end{aligned}\ ] ] and the corresponding expression contains the projector on positive parity states , @xmath239 , as @xmath240_{ab}&= & e^2[{\mathbf p}^{\left(\frac{3}{2},0\right)}_+ ( { \mathbf p}')]_{a}{}^c[\mathbf{u}_{{\mu}{\nu}}(p',q , r , p)]_{cb},\nonumber\\
~[{\mathbf u}_{{\mu}{\nu}}(p',q , r , p)]_{cb}&=&{\big(}{\mathcal u}_{{\mu}{\nu}}(p',q , p)+{\mathcal u}_{{\nu}{\mu}}(p',r , p)-{{\mathcal x}}_{{\mu}{\nu}}{\big)}_{cb}.\end{aligned}\ ] ] the projector over spin-@xmath9 states of positive parity has same structure as the one given by @xcite , @xmath241_{ab}= \sum_{{\lambda}}\left [ w_+^{\left(\frac{3}{2},0\right)}\left({\mathbf p},\frac{3 } { 2},{\lambda}\right)\right]_{a } \left[\overline{w}_+^{\left(\frac{3}{2},0\right ) } \left({\mathbf p},\frac{3}{2},{\lambda}\right)\right]_{b}. \label{phpr}\end{aligned}\ ] ] diagram for the point - scattering contribution ( [ mcj1 ] ) to the compton scattering amplitude ( [ camp ] ) . ]
diagram for the point - scattering contribution ( [ mcj1 ] ) to the compton scattering amplitude ( [ camp ] ) . ]
diagram for the point - scattering contribution ( [ mcj1 ] ) to the compton scattering amplitude ( [ camp ] ) . ]
we furthermore used @xmath242^{*}&=&- g^{{\mu}{\nu}}.\end{aligned}\ ] ] the pure spin-@xmath9 projector , @xmath243_{ab}$ ] can equivalently be rewritten to @xmath244_{ab } & = & \left[f^{\left(\frac{3}{2},0\right)}(\mathbf{p})\right]_{a}{}^{{\mu}}{\left(}\frac{- { p}\!\!\!/+m}{2m}{\right)}\left[\overline{f}^{\left(\frac{3}{2},0\right)}(\mathbf{p})\right]_{b{\mu}},\end{aligned}\ ] ] with @xmath245 from ( [ f_tensors ] ) .
the contractions indicated in ( [ eq : asa ] ) are comfortably evaluated with the aid of the feyncalc symbolic software package amounting to : @xmath246 where @xmath247 are the standard mandelstam variables and we are using the notations @xmath248 the differential cross section in the laboratory frame is now calculated from the standard formulas , @xmath249 where @xmath250 and @xmath251 are the energies of the incident and scattered photons respectively , while @xmath252 is the scattering angle in the laboratory frame . furthermore , with @xmath253 and after some algebraic manipulations , the final result is cast in the form of an expansion in powers of @xmath254 ( with @xmath255 ) according to , @xmath256 here , @xmath257 , @xmath258 , and the expansion coefficients are , @xmath259 in the low energy limit , we recover as expected the correct expression for the differential cross section in the thompson limit as , @xmath260 while in forward direction , the differential cross section takes an energy independent value , @xmath261 and in accord with unitarity . in all the other directions however , the differential cross section increases with energy . in the figure [ dsfig ] we present a plot of the quantity @xmath262 as a function of the @xmath263 variable , at energies of @xmath264 ( solid curve ) , @xmath265 ( long dashed curve ) and @xmath266 ( short dashed curve ) ,
here we see how the differential cross section approaches the classical limit at low energy ( symmetric curve ) and raises as the energy grows except in the forward direction .
the differential cross section , @xmath267 , as a function of @xmath258 .
the solid curve represents the classical limit at @xmath268 , the long dashed line corresponds to an energy comparable to the mass of the particle , @xmath269 , while the short dashed curve corresponds to @xmath266 .
this cross section increases with energy , except in the forward direction , @xmath270 , where it approaches @xmath271 . ] integrating over the solid angle we find the total cross section as : @xmath272 being @xmath273 the thompson cross section and @xmath274 the total cross section ( [ tcs ] ) has the following limits , @xmath275 this behavior is show in figure [ sfig ] , where we make a plot of @xmath276 here we see the decreasing behavior of the cross section at low energies as well as its growing behavior at high energies .
the total cross section @xmath277 as a function of @xmath278
. in the low energy limit the thompson limit , @xmath279 , is recovered , otherwise the cross section grows with the energy increase . ]
the first achievement of this work is the representation in the equations ( [ drct])-([spin32ats ] ) of the spin-@xmath9 degrees of freedom spanning the @xmath141 carrier space of the lorentz algebra as anti - symmetric second rank lorentz tensor - spinors .
this representation is equipped by separate lorentz and dirac indexes and provides a comfortable tool in calculations of scattering cross sections by means of the symbolic software package feyncalc .
a similar experience has been made in @xcite regrading spin-@xmath12 transforming as @xmath76 , where the calculation of compton scattering could not be tackled in terms of six - dimensional spinors and was instead easily executed in the anti - symmetric tensor - basis .
we here specifically worked out the compton scattering off @xmath141 and reported in eq .
( [ unt_fd ] ) on finite differential cross section in forward direction in the ultra relativistic regime and in accord with unitarity . before we had calculated in the eqs .
( [ qe0])([qm3 ] ) the multipole moments of a particle transforming according to the representation space under investigation and found that they reproduced results earlier predicted by the weinberg - joos theory , the gyromagnetic ratio coming out fixed to the inverse of the spin , i.e. @xmath280 and in accord with belinfante s conjecture .
it is this very same value that gives rise to unitarity in the process of compton scattering in forward direction and at variance with the @xmath281 value required by spin-@xmath9 transforming in the four - vector spinor .
our observation suggests that spin-@xmath9 particles transforming within the two - spin valued four - vector spinor , on the one side , and as the single spin valued @xmath141 , on the other side , behave as distinct species . within the framework of the second order approach elaborated here
, we provided arguments in favor of the causal propagation of a pure spin-@xmath9 interacting with an electromagnetic field .
we furthermore explained how our scheme allows for an extension toward any spin , be it bosonic or fermionic , in retaining the quadratic wave equations and associated lagrangians .
our conclusion is that @xmath7 representation spaces preserve their individuality upon embeddings into reducible lorentz - tensors .
we have checked that also all the other irreducible sectors of the antisymmetric tensor - spinor , namely , the single spin-@xmath8 dirac sector , as well as the double spin valued @xmath282 rarita - schwinger , are characterized before and after embeddings by same sets of multiple moments and compton scattering cross sections .
the embeddings under discussion bring the advantage of separate lorentz and dirac indexes which significantly speeds up the tensor calculus relatively to the matrix calculus .
as long as according to ( [ tensor_spinor ] ) the anti - symmetric tensor spinor represents itself as the direct sum of @xmath188 ( @xmath283 degrees of freedom ) with a dirac spinor , @xmath77 ( 4 degrees of freedom ) , and @xmath284 ( 12 degrees of freedom ) , the number of degrees of freedom is this space is @xmath285 .
the set of the first @xmath286 four - vector spinor degrees of freedom has been constructed in @xcite , with those corresponding to spin-@xmath9 in @xmath287 being here @xmath289 are negative and positive energy dirac spinors , coupled to the following spin-@xmath12 four - vectors , @xmath290 spanning @xmath291 , given ( modulo notational differences ) by @xcite @xmath292 there is one more four - vector residing within the @xmath293 representation space which is of spin-@xmath84 and reads @xcite , @xmath294 the coupling of the latter vector to @xmath295 in the equations ( [ fvs324 ] ) from above provides a new independent orthogonal complete set of spin-@xmath9 states according to , @xmath296^{\left[{\alpha}{\beta}\right]}&= & \frac{1}{2}{\big ( } \eta^{\alpha}(\mathbf{p},0,0){\mathcal u}_\pm^{\beta}\left(\mathbf{p},\frac{3}{2},{\lambda}\right ) - \eta^{\beta}(\mathbf{p},0,0){\mathcal u}^{\alpha}_\pm\left(\mathbf{p},\frac{3}{2},{\lambda}\right ) { \big)}\nonumber\\ \label{ss32atsprl}\\ & = & \frac{1}{2}{\big ( } \frac{p^{\alpha}}{m}{\mathcal u}^{\beta}_\pm\left(\mathbf{p},\frac{3}{2},{\lambda}\right ) - \frac{p^{\beta}}{m}{\mathcal u}_\pm^{\alpha}\left(\mathbf{p},\frac{3}{2},{\lambda}\right){\big)}. \label{ss32ats}\end{aligned}\ ] ] these are the tensor - spinors which , according to ( [ spin32ats ] ) , find themselves at the root of the degrees of freedom of the @xmath297 representation space .
finally , the conjugate states to @xmath298 , and @xmath299^{\left[{\alpha}{\beta}\right]}$ ] are defined as @xmath300^{\dagger{\alpha } } , \label{dualdef1 } \\
\left [ \overline{u}_\pm \left({\mathbf p},\frac{3}{2},{\lambda}\right)\right]^{\left[{\alpha}{\beta}\right ] } & = & \left[\gamma_0 u_\pm \left({\mathbf p},\frac{3}{2},{\lambda}\right)\right]^{\dagger { \left[{\alpha}{\beta}\right]}}. \label{dualdef2}\end{aligned}\ ] ] | we propose a general method for the description of arbitrary single spin-@xmath0 states transforming according to @xmath1 carrier spaces of the lorentz algebra in terms of lorentz - tensors for bosons , and tensor - spinors for fermions , and by means of second order lagrangians .
the method allows to avoid the cumbersome matrix calculus and higher @xmath2 order wave equations inherent to the weinberg - joos approach .
we start with reducible lorentz - tensor ( tensor - spinor ) representation spaces hosting one sole @xmath1 irreducible sector and design there a representation reduction algorithm based on one of the casimir invariants of the lorentz algebra .
this algorithm allows us to separate neatly the pure spin-@xmath0 sector of interest from the rest , while preserving the separate lorentz and dirac indexes .
however , the lorentz invariants are momentum independent and do not provide wave equations .
genuine wave equations are obtained by conditioning the lorentz - tensors under consideration to satisfy the klein - gordon equation . in so doing
, one always ends up with wave equations and associated lagrangians that are second order in the momenta .
specifically , a spin-3/2 particle transforming as @xmath3 is comfortably described by a second order lagrangian in the basis of the totally antisymmetric lorentz tensor - spinor of second rank , @xmath4}$ ] . moreover ,
the particle is shown to propagate causally within an electromagnetic background . in our study of @xmath5 as part of @xmath6 we reproduce the electromagnetic multipole moments known from the weinberg - joos theory .
we also find a compton differential cross section that satisfies unitarity in forward direction .
the suggested tensor calculus presents itself very computer friendly with respect to the symbolic software feyncalc .
* bosonic and fermionic weinberg - joos @xmath7 states of arbitrary spins as lorentz - tensors or tensor - spinors and second order theory * e. g. delgado acosta , v.m .
banda guzmn , and m. kirchbach instituto de fisica , uaslp , av .
manuel nava 6 , zona universitaria , + san luis potosi , slp 78290 , mxico pacs - key : 11.30.cp ( lorentz and poincar invariance ) , 03.65.pm ( relativistic wave equations ) 00h |
in this supplementary material we first discuss the fabrication and characterization techniques ( section i ) . in section
ii , we show that two synchronously bending waveguides are equivalent to two straight waveguides with some effective coupling constant . in section
iii the techniques used to obtain the topological phase diagram of fig .
[ fig2 ] ( main text ) are presented .
we discuss wavelength tuning and the effect of disorder in section iv before finally concluding with a study of how this disorder affects the topology in section v. the photonic lattice with two driving periods was fabricated inside a 70-mm - long glass ( corning eagle@xmath69 ) substrate using the ultrafast laser inscription technique , where the refractive index profile of each waveguide was controlled using the slit - beam shaping " method @xcite .
the glass substrate , mounted on @xmath70-@xmath71-@xmath16 translation stages , was translated at 8 mm / s through the focus of a 500 khz train of 1030 nm femtosecond laser pulses to fabricate each waveguide .
the laser inscription parameters were optimized to produce waveguides that were single - mode and well confined in the measurement wavelength range of 700 - 830 nm . to study the response of the edge modes in the presence of a defect at the edge , the waveguide at the ( 8 , 4 ) lattice site was not fabricated .
it should also be highlighted that the waveguide paths are designed such that all waveguides exhibit identical bend radii at a given @xmath16 , although the direction of this bending is site dependent .
this ensures that there is minimal site - dependent losses . to measure the coupling constants , we fabricated five sets of each bond separately inside the same substrate .
these bonds ( or couplers ) were then characterised in the wavelength range 705 - 795 nm , to obtain the variation of the mean and standard deviation of coupling strength with wavelength ( see supplementary for details ) . to excite the lattices with different wavelengths ,
a photonic crystal fiber was pumped by sub - picosecond laser pulses of 1064 nm wavelength to generate a broadband supercontinuum .
a tunable monochromator placed after the supercontinuum source was used to select narrow band ( @xmath72 nm ) light , which was coupled into an optical fibre ( smf-600 ) .
the fiber was then coupled to the lattice sites .
the output intensity distribution was observed using a cmos camera .
a polarizer passing only vertically polarized light is placed in front of the camera to ensure that the measurements are not affected by polarization - dependent coupling . in the experimental setup ,
the lattice constant was chosen such that , in the absence of any bending , the waveguides comprising the lattice were practically uncoupled . in order to turn on the coupling between two neighbouring sites
[ see fig .
[ fig1 ] ( b ) in the main text ] , the two waveguides were bent together and then remained straight before moving apart again [ see fig .
[ fig1 ] ( d ) in the main text ] .
the coupling between two such synchronously bending waveguides was demonstrated in ref .
@xcite to be equivalent to an effective coupling between two straight waveguides .
this mapping is vital in order for the experiment to emulate our theoretical model , which assumes that the lattice is comprised of straight waveguides but with controllable couplings to neighbouring waveguides [ see fig .
[ fig1 ] ( b ) in the main text ] . in light of the importance of this result in the present work ,
a derivation is provided here for completeness .
the dynamics of how light propagates through an array of weakly - guiding waveguides is governed by the scalar helmholtz equation , @xmath73 here @xmath74 is the electric field amplitude , @xmath75 , where @xmath76 is the wavelength of the light in free space , and @xmath77 is the refractive index profile in the domain of interest . if the wave propagation is primarily along the @xmath16-axis , then the field @xmath78 can be represented as a slowly varying complex field amplitude and a fast oscillating wave , @xmath79 .
substituting @xmath80 into the helmholtz equation and making the slowly varying envelope approximation for @xmath81 gives @xmath82 the refractive index profile can be written as @xmath83 .
however , in laser - written waveguide arrays , the refractive index change caused by the laser is small and so @xmath84 $ ] .
consider now an array of moving waveguides where @xmath85 $ ] is given by @xmath86=\sum_k v_0^{k } = \sum_{k } v_0[x - x_{k}^0-x_{k}(z),y - y_{k}^0 ] .
\notag\ ] ] here @xmath87 is the refractive index profile of a single isolated waveguide , the index @xmath88 labels the different waveguides and @xmath89 is the @xmath16-dependent position of the @xmath88th waveguide .
+ in the regime where the waveguides are far apart , compared to the width of @xmath90 , the electric field can be well approximated by a superposition of the fundamental modes of the different waveguides , @xmath91e^{-i\beta z } , \notag\ ] ] where @xmath92 $ ] is the fundamental mode of an isolated waveguide and satisfies the equation @xmath93=-\frac{\lambdabar^2}{2n_s}(\frac{\partial^2}{\partial x^2}+\frac{\partial^2}{\partial y^2})u[x , y]+v_0[x , y]u[x , y ] .
\label{eigenmode}\ ] ] inserting the electric field expansion into eq .
gives , after some rearrangement , @xmath94u_k+ \sum_{n \neq k } v_0^{n } \sum_k c_k u_k,\end{aligned}\ ] ] where we have introduced the notation @xmath95 $ ] .
the expression in the square brackets is equal to zero as a result of eq . .
+ multiplying by @xmath96 and integrating transversely gives @xmath97 where @xmath98 a number of pertinent statements can be made about these integrals : @xmath99 these statements have the corollary that @xmath100 as previously mentioned , in the absence of any bending there is negligible coupling between any of the waveguides .
the consequence of this is that in the presence of bending , only pairs of waveguides are coupled at any given moment . the coupled mode equations for this system
can then be written as block diagonal matrices with each block comprising a @xmath101 matrix .
this simplifies the following to a study of a two waveguide coupler as the analysis will straightforwardly generalise to the full lattice .
therefore , if there are two waveguides , labelled @xmath102 and @xmath103 , that are synchronously bending , such that @xmath104 , then it can be readily seen that @xmath105 .
the coupled - mode equations for this @xmath106 block can therefore be written in matrix notation as @xmath107 where the matrices @xmath108 and @xmath109 have the form @xmath110 here @xmath111 and @xmath112 . to proceed
further we introduce a new set of variables by @xmath113 , where @xmath114 is chosen such that @xmath115 .
this transformation is always possible as both @xmath116 and @xmath108 are both hermitian , positive definite matrices and so can be connected via a similarity transformation by a non - singular matrix @xcite .
this change of variable matrix can be written as @xmath117 using this change of variables and multiplying from the left by @xmath118 allows eq . to be rewritten as @xmath119 with @xmath120 taking the diagonal form @xmath121 the equation for @xmath122
can be straightforwardly solved to yield @xmath123 where @xmath109 has the form @xmath124 in the second equality , the cross - power term , @xmath125 , has been neglected , which is a valid approximation as long as the waveguides remain far apart throughout the bending motion .
+ a return to the original @xmath126 variables can be made by inverting the non - singular matrix @xmath114 , @xmath127 where @xmath128 are the initial conditions .
the waveguides in the experimental lattice are coming together and then moving apart .
consequently , at the end of the bending section , @xmath129 , the waveguides are at the same separation as at @xmath130 and so @xmath131 .
this allows the matrix product @xmath132 to take the form @xmath133 with @xmath134 for comparison , the evolution operator for two straight waveguides of length @xmath135 is given by @xmath136 where @xmath137 is the coupling between the waveguides .
as can be readily observed , this has the same form as eq . if we write @xmath138 as @xmath139 where @xmath140 is an effective coupling .
this analysis has been conducted for two waveguides , but can be readily extended to the full lattice discussed in the main text .
the experimental lattice is written such that only pairs of waveguides are coupled at any given moment .
the @xmath108 and @xmath114 matrices therefore have block - diagonal matrix form with the blocks all having the form of eq . .
in this section the methodology used to obtain the topological phase diagram in fig .
@xmath141 ( main text ) is detailed ; this method is based on ref .
@xcite , and we refer the reader to this reference for more details on concepts related to the topological characterisation of floquet systems . the calculation of the different phases is assisted by utilizing the property that the topology of the system can only change when there is a gap closing between the two bulk bands .
the position of these gap - closing events can be found analytically by diagonalizing the evolution operator at the end of the driving period .
it is found that for @xmath142 and @xmath143 the system is gapless at quasienergy zero , whilst for @xmath144 and @xmath145 the system is gapless through the fundamental zone edge .
the position of these gap closings thereby divides the phase space into the eight different sectors shown in fig .
the topology within these different sectors can then be defined by calculation of the winding numbers . as discussed in eq .
, at any time @xmath22 the evolution operator may be diagonalized to yield the instantaneous bloch bands of the driven system .
the eigenstates associated with these bands can be used to calculate an instantaneous chern number for that band . in accordance with eq .
the winding number can then be calculated by tracking changes in the chern number of the lowest band that occur through the zone edge throughout the driving period .
this procedure is illustrated in fig .
@xmath146 ( a ) for the parameters @xmath147 , @xmath148 , which shows how the instantaneous chern number of the lowest band changes throughout the driving period .
it can be readily observed that that the chern number changes twice within a driving period and computing the spectra at these times shows that the second of these changes occurs through the zone edge , see fig .
@xmath146 ( b - c ) .
this latter degeneracy causes the chern number of the lowest band to decrease by one .
this topological transition through the zone edge when combined with the chern numbers of all the bands being zero at @xmath149 implies , in accordance with eq .
, that both @xmath47 and @xmath48 are equal to one . over one driving period
; the system parameters are @xmath150 and @xmath151 .
the chern number changes twice within a driving period ( at times @xmath152 and @xmath153 ) .
the first of these topological transitions occurs when the gap closes at @xmath154 , ( b ) , whilst the second occurs at @xmath155 , ( c ) .
this changing of the instantaneous chern number that occurs through the zone edge is responsible for the winding numbers @xmath47 and @xmath48 being non - zero , and the chern number of the floquet bands to be zero .
this non - trivial value for the winding number allows for the presence of edge modes in the spectrum at @xmath149 , ( d ) , even though the chern numbers of all the floquet bands ( evaluated at @xmath156 ) are zero.,width=321 ] in the main text it was discussed how the experimental setup provides the ability to modify the wavelength of light used as an input .
this ability allows the effective coupling between waveguides to be altered without modifying the lattice . in order to understand the behaviour of the bonds @xmath157 , see main text , as a function of wavelength ,
five sets of isolated bonds were written and the percentage of light transferred from the launch waveguide as a function of wavelength was measured , fig .
[ coupling ] .
this bond characterisation data illustrates how at 785 nm the bonds @xmath157 have almost equal transfer which is close to 100@xmath158 .
this wavelength therefore closely matches the desired parameters discussed in the main text of having three bonds equal and perfectly transferring . as a function of wavelength . *
the effective coupling can be extracted from this characterisation data using the relation transfer @xmath159 , where @xmath135 is the physical length of the sample .
the dotted line indicates that the transfer is very close to 100@xmath158 for all the three bonds at 785 nm wavelength.,width=321 ] the capability of this wavelength tuning technique in our experimental setup allows for the transfer of the bonds to be changed substantially from @xmath160 right down to @xmath161 .
this provides the ability to compare experimental and theoretical predictions for a wide range of parameters . in order to perform such a comparison ,
the coupling strengths extracted from the bond characterisation data were used to calculate a theoretical centre of mass drift , @xmath162 , that could be compared to experimental results .
the centre of mass drift is defined as @xmath163 where @xmath164 is the intensity in the @xmath165-th site at the output , @xmath166 is the coordinates of the @xmath165-th site and @xmath167 are the coordinates of the launch site .
experimental data was taken at four different physical locations in the lattice .
these different positions should , in the absence of disorder , produce identical results for @xmath162 , up to a sign change .
however , the non - zero standard deviations measured in the bond characterisation data indicates the presence of bond strength disorder within the lattice ( such a disorder will be referred to as off - diagonal disorder " ) .
the effects of this disorder can be directly observed in the experimental lattice with the four different launch sites producing slightly different results for @xmath162 , see the error bars in fig .
[ cm](a ) . in order to theoretically model this disorder ,
the strengths of the bonds @xmath157 were randomly selected to lie within the range measured in the bond characterisation data .
the comparison between the experimental data and the theoretical prediction shows good agreement around @xmath168 nm , but deviations are observed at smaller wavelengths , fig .
[ cm](a ) .
the failure of the theoretical model in this low transfer region is further illustrated in fig .
[ cm](c@xmath169 ) , which compares the averaged experimental output image to the theoretical image averaged over 1000 disorder realisations .
the match between theory and experimental results can be improved by including a small on - site disorder term , @xmath170 , in addition to the coupling disorder of the previous model ( this form of disorder will be referred to as diagonal disorder " ) .
the @xmath170 for the @xmath171-th site is a random number drawn from within a uniform distribution covering the interval @xmath172 $ ] .
the parameter @xmath13 is the disorder strength which is chosen as @xmath173 , as this particular value results in the best fit to the experimental data over the whole wavelength range investigated . the corresponding center - of - mass prediction , fig .
[ cm](b ) , and output facet image , fig .
[ cm](c@xmath174 ) , produced by this model are in closer agreement to the experimental results which points towards a disordering of this type being present in the lattice . and
@xmath71 directions whilst the black error bars indicate the range of values measured experimentally .
the shaded regions in ( a ) and ( b ) indicate the theoretical disorder averaged value plus / minus one standard deviation .
( c@xmath175 ) is the experimental output image averaged over the 4 launch sites for 705 nm input whilst ( c@xmath176 ) and ( c@xmath177 ) are the disordered averaged output facet images coming from models 1 and 2 respectively.,width=325 ] the analysis conducted in the previous section illustrated the presence of off - diagonal disorder as well as possibly indicating the presence of diagonal disorder terms . in the absence of any disorder ,
the experimental lattice should reproduce a driven - lattice model that exhibits anomalous topological edge modes ( i.e. chiral edge modes that are associated with non - zero winding numbers , while the chern number of the floquet bands is trivial ; see main text ) .
the inclusion of disorder in the system can allow bandgap closings to occur which can modify the topology of the system .
in particular , the disorder could potentially change the chern numbers of the floquet bands to a non - trivial value such that the experimentally observed edge modes would no longer be anomalous " .
the elimination of this possibility thereby requires the calculation of the chern number in a disordered system .
the presence of disorder means that quasimomentum is no longer a good quantum number which in turn precludes the use of the usual momentum space techniques in calculating the chern number . in order to overcome this limitation we use a real - space chern invariant , @xmath178 , as introduced by bianco and resta in ref .
this invariant allows for a generalisation of the ( _ static _ ) bulk - edge correspondence to disordered and quasi - periodic lattices @xcite .
in particular , when the reference energy , @xmath179 , is placed in a mobility gap of the spectrum the chern invariant will measure the number of chiral edge modes which cross this energy , as predicted by the chern numbers of the bands below this energy .
this real - space chern invariant is based around an operator called the chern marker , @xmath180 , which is defined by @xmath181.\ ] ] here the operators @xmath182 and @xmath183 are expressed in terms of the position operator @xmath184 and the projection operators @xmath185 in the absence of disorder a real space chern number can be defined by averaging the chern marker operator over a unit cell , @xmath186 anywhere within the bulk of the lattice . this local chern marker
is narrowly peaked around zero which implies that the static bulk - edge correspondence would predict no robust chiral edge modes present in the system.,width=325 ] in the presence of disorder this real - space chern number fluctuates depending upon the position of the unit cell . these fluctuations can be accounted for by replacing the unit - cell average by an average over an area @xmath187 that is located within the bulk and is large compared to the fluctuation length @xcite .
these ideas thereby provide a mechanism to investigate whether the disorder terms discussed in the previous section can alter the chern numbers of the floquet bands to non - trivial values .
in particular we calculate the local chern marker , in the presence of these disorder terms , for a reference energy of zero , @xmath188 .
the real - space average of this local chern marker , the chern invariant , can be viewed as the prediction coming from the static bulk - edge correspondence for the number of chiral edge modes that are traversing the mobility gap between the two bands .
the results , however , of this analysis reveal that even in the presence of disorder the local chern marker , @xmath189 , is still heavily peaked around zero with this local quantity fluctuating only slightly within the bulk , fig .
the chern invariant is therefore equal to zero independently of the size of the averaging area that is chosen .
consequently , the static bulk - edge correspondence predicts the experimental lattice should not feature any chiral edge modes . | the discovery of the quantised hall effect , and its subsequent topological explanation , demonstrated the important role topology can play in determining the properties of quantum systems @xcite .
this realisation led to the development of topological band theory , where , in addition to band index and quasimomentum , bloch bands are also characterised by a set of topological invariants .
this topological theory can be readily extended to periodically - driven systems . in the limit of fast driving
, the topology of the system can still be captured by the topological invariants used to describe static systems @xcite . in the limit of slow driving , however , situations can arise where standard topological invariants are zero , but yet , topologically protected edge modes are still observed @xcite .
these anomalous " topological edge modes have no static analogue , and are associated with a distinct topological invariant , which takes into account the full time - evolution over a driving period . here
we demonstrate the first experimental observation of such anomalous topological edge modes in an ultrafast - laser - inscribed photonic lattice .
this inscription technique allows one to address each bond of a lattice independently and dynamically , generating a rich band structure with robust anomalous chiral edge modes and the potential for perfectly localised bulk states .
subjecting a system to time - periodic modulations constitutes a powerful method to engineer band structures with non - trivial topological properties @xcite , as recently demonstrated in cold - atom experiments @xcite and photonics @xcite . in this
_ floquet - engineering _ approach , a static system , described by a hamiltonian @xmath0 , is driven periodically in time by a modulation @xmath1 , whose frequency will be denoted @xmath2 . the time - evolution operator over an arbitrary long duration @xmath3 then takes the general form @xcite @xmath4 here , the time - independent ( effective ) hamiltonian @xmath5 describes the time - averaged dynamics over the duration @xmath6 , which can be isolated by probing the system stroboscopically at discrete times @xmath7 .
this effective hamiltonian stems from a rich interplay between the static system @xmath0 and the time - modulation @xmath8 @xcite ; its band structure ( the quasienergies or floquet spectrum " ) can feature topological properties , such as non - zero chern numbers and chiral edge modes @xcite , in direct analogy with non - driven systems .
the time - evolution operator [ eq . ] also includes the micro - motion operator @xmath9 , which describes the dynamics that takes place within each period of the driving , i.e. @xmath10 . in the high - frequency regime of the driving ( @xmath11 ) , observables are generally only slightly affected by the micro - motion , fig .
[ fig1](a ) . in this regime , the topological properties of the driven system
are then entirely captured by the effective hamiltonian @xmath5 .
the observation of topological edge modes can be directly related to the chern numbers associated with the effective band structure @xcite .
such edge modes were directly visualized in photonic realizations of floquet band structures @xcite ( see also refs .
@xcite for other photonic implementations of standard topological edge modes ) .
importantly , this is no longer the case in the low - frequency regime , where @xmath12 becomes comparable to the typical energy scales of the system ( e.g. the band - width of the effective band structure @xcite ) . in this regime ,
the micro - motion strongly affects the time - evolution of observables [ fig .
[ fig1](a ) ] , and the topological properties of the system can no longer be simply related to the effective hamiltonian only . in particular , the presence of topological edge modes is now entirely ruled by a distinct topological invariant , a winding number @xmath13 that takes into account the micro - motion @xcite . in the following , topological edge modes that emerge in the low - frequency regime , due to the micro - motion , will be referred to as anomalous " topological edge modes , as proposed in ref . @xcite . since these topological states are captured by the winding numbers @xmath13 , and not by the chern numbers of the effective hamiltonian , they constitute genuine floquet " topological modes .
this intriguing regime of periodically - driven systems has recently been investigated in a diverse range of experimental platforms : topologically protected bound states associated with non - trivial @xmath13 were first demonstrated in a one - dimensional photonic setup realizing a discrete time quantum walk @xcite ; more recently ref .
@xcite reported on the observation of anomalous edge modes in a designer surface plasmon platform .
finally , ref .
@xcite considered a thouless - pump approach to measure @xmath13 in a one - dimensional microwave network . .
in the low - frequency regime , the micro - motion can significantly affect the dynamics , and the time - evolution is no longer well described by the time - averaged hamiltonian only .
( b ) the four different bonds present in the lattice ( with coupling constants @xmath14 ) and the cyclic driving protocol employed .
( c ) in the simplest case where all bond strengths are equal , @xmath15 , chiral edge modes arise whilst the evolution operator associated with the bulk is identity .
( d ) sketch illustrating how , using laser inscription , different pairs of waveguides can be moved together to turn on a coupling and then apart to switch it off .
this flexibility allows for a realisation of the driving protocol shown in ( b ) . in waveguide arrays the propagation direction @xmath16 plays a role analogous to time @xcite .
, width=325 ] the theory works in ref .
@xcite introduced conceptually simple models that exhibited such intriguing floquet topological structures . the model in ref .
@xcite is a square lattice with nearest - neighbour couplings which are engineered so that the couplings between a lattice site and its four nearest neighbours are independently controllable .
these four couplings , denoted @xmath17 through @xmath18 , are then varied in a spatially homogeneous and time - periodic manner so that any lattice site is , at any given moment , coupled to only one of its nearest neighbours , see fig .
[ fig1 ] ( b ) .
the simplest demonstration of this model is when the driving period @xmath19 is split into four equal steps , and @xmath19 is selected , such that a particle that is initially localised on a certain site will be completely transferred to the neighbouring site after a time @xmath20 .
therefore , considering for now a system without edges , this means that after one complete period , any initial state is exactly reproduced , i.e. the propagator , @xmath21 , is the identity matrix [ fig .
[ fig1 ] ( c ) ] . as a corollary , the effective hamiltonian in eq .
is the zero matrix and the floquet spectrum consists of two degenerate flat bands at zero energy .
the bulk is completely localised , and the chern numbers associated with the effective hamiltonian are necessarily trivial . in a geometry with edges , however , it is found that there are chiral propagating edge modes that are localised along the edge .
these occur because a particle launched at an edge can not get back to its initial position and ends up moving one unit cell along the edge [ fig .
[ fig1 ] ( c ) ] , in direct analogy with the skipped cyclotron orbits of quantum hall systems .
as previously mentioned , the topological properties of slowly - driven systems are well captured by a winding number that takes the full time evolution into account , including the micro - motion @xcite .
for the two - band model introduced above , the time evolution operator at time @xmath22 , where @xmath23 , may be written as @xmath24 where @xmath25 and @xmath26 are the projectors and eigenvalues of @xmath27 , respectively .
the instantaneous energies " , @xmath28 , are defined modulo @xmath29 , and we define the corresponding floquet - brillouin zone " to be in the range @xmath30 $ ] . in a driven system , there are two types of degeneracy that can occur within a driving period : inter - zone degeneracy , where @xmath31 , and degeneracy through the zone edge where @xmath32 and @xmath33 .
it is the existence of this latter type of degeneracy that profoundly alters the topological picture of driven systems .
zone - edge singularities ( zes ) allow the chern number of the bands to change without the inter - zone gap closing .
consequently , these zone - edge degeneracies can lead to violations of the static bulk - edge correspondence @xcite , so that the chern number of both floquet bands are zero but yet protected edge modes are still present @xcite .
the adequate topological characterisation of driven systems is captured by a winding number which includes the changes in chern numbers that occur through the zone edge , within a period of the driving . in an two - band driven system
, there can be at most two bandgaps , and for each of these bandgaps , there is a winding number @xmath34 , which has the form @xcite @xmath35 where @xmath36 is the chern number of the @xmath37-th band at time @xmath19 , and where @xmath38 corresponds to the change in chern number of the lowest band that occurs in the @xmath39-th zone edge degeneracy .
this winding number directly gives the number , @xmath40 , of topologically protected chiral edge modes present in the @xmath41-th gap .
this driven - system " bulk - edge correspondence can be shown to have the form @xcite @xmath42 the first term in the second equality of eq .
is the term that is found in static systems , whilst the second applies only to driven systems .
it is this term that is the source of the anomalous edge modes as it allows the number of edge modes to be non - zero even if the standard topological invariants , i.e. the chern numbers @xmath36 , are zero for all of the bands . , which allows the bond @xmath17 to be of a different strength to the bonds @xmath43 . here
@xmath44 and @xmath45 .
each phase is described by three topological invariants : the chern number of the lowest floquet band @xmath46 , the winding number @xmath47 of the gap centred around zero , and the winding number @xmath48 of the gap centred around @xmath49 .
the white regions indicate where anomalous edge modes can be observed ( @xmath50 ) .
( b ) quasienergy spectrum for the parameters indicated by the black dot in ( a ) .
red curves indicate topological - edge - modes dispersions , while black curves correspond to the bulk bands .
( c ) quasienergy spectrum for the experimentally achieved parameters .
the first floquet - brillouin zone is taken to be defined in the range @xmath30 $ ] . ] in this work , we experimentally implement a variant of the model introduced in ref .
@xcite . in our model
the coupling of the first bond ( @xmath17 ) is different than the following three bonds ( @xmath51 ) , which are chosen so as to satisfy @xmath52 and @xmath53 .
this gives two controllable experimental parameters , @xmath54 and @xmath55 , that can be tuned independently .
we find that the interplay between these two parameters produces a rich phase diagram , shown in fig .
[ fig2 ] ( a ) , which demonstrates how this simple model can be used to explore many different topological regimes ( see sup
. mat . ) .
each topological phase of this driven two - band system is accurately labelled by the winding numbers associated with the two bandgaps ; in the phase diagram shown in fig .
[ fig2 ] ( a ) , we have chosen @xmath47 to be the winding number for the bandgap centred around quasienergy zero and @xmath48 for the bandgap centred around @xmath49 .
in addition , we also provide the chern number of the lowest floquet band @xmath46 , so as to highlight the anomalous regimes where @xmath56 and @xmath57 . for the sake of experimental practicability , we aim to focus on the anomalous edge modes that are predicted for @xmath58 and @xmath59 , as indicated by the black dot in fig .
[ fig2 ] ( a ) . in this configuration , the blue bonds in fig .
[ fig1 ] ( b ) are never turned on , while nearly complete transfer of light can occur via the other three bonds .
these parameters are desirable as not only is the resultant system located well within the anomalous regime [ fig .
[ fig2 ] ( a ) ] but also the associated spectrum consists of two gapped flat bands [ fig . [ fig2 ] ( b ) ] . additionally , the floquet states have a simple analytical form , and there are robust chiral edges states that can be excited with unit efficiency , through the simple experimental technique of single - site excitation .
the realisation of this model requires a high degree of control over the couplings present in the lattice .
ultrafast - laser - inscribed arrays of optical waveguides with three - dimensional geometry offer this control , as they allow the coupling between lattice sites to be individually controlled . in the experiment presented here , a lattice constant of 40 @xmath60 was chosen , such that the coupling between lattice sites was insignificant over the maximum observable propagation distance . to turn on the coupling between any two waveguides ,
the inter - waveguide separation is reduced such that the two waveguides propagate together for a 4.5 mm straight section , with a centre - to - centre waveguide separation of 11 @xmath60 .
after this interaction region , the waveguides again separate in a reverse manner , see the sketch in fig .
[ fig1 ] ( d ) . for a particular wavelength ,
the coupling that occurs between two synchronously bending waveguides can be shown to be equivalent to that of two straight waveguides with some effective coupling constant ( sup . mat . ) .
this effective coupling can be controlled by changing the interaction length and/or the separation of the waveguides ; these parameters are fixed upon writing the sample and so do not allow fine - tuning of the coupling in - situ . however , the wavelength of the light used to excite the lattice is a tunable parameter that can be used to control the effective coupling strength ( sup . mat . ) . to demonstrate the existence of anomalous edge modes ,
a lattice of size @xmath61 , consisting of two driving cycles , was fabricated .
this lattice also contains a single defect , namely , a missing lattice site on the edge , allowing for the detection of the robustness of edge modes . to characterise the bonds present in the lattice , and hence validate our theoretical model , we fabricated five sets of each bond in isolation , inside the same substrate , and measured their behaviour as a function of the wavelength of light ( sup . mat . ) .
when characterised with 785 nm light , the mean and standard deviation of @xmath62 ( @xmath63 ) were found to be @xmath64 , @xmath65 , @xmath66 .
for the bond @xmath17 , all the light remains in the waveguide excited at the input , which indicates no transfer of light ( @xmath67 ) .
the close proximity of these measured couplings to the aforementioned desired values suggest that anomalous edge modes should be detected for an input wavelength of 785 nm . indeed , the quasienergy spectrum corresponding to these experimental values [ fig .
[ fig2](c ) ] is similar to the ideal case [ fig .
[ fig2](b ) ] .
while a finite dispersion of the bulk now becomes apparent , we note that the velocity of the edge modes remains significantly larger compared to the dispersion of the bulk .
it should be noted that the non - zero standard deviations indicate that there will be bond - strength disorder within the lattice , whilst a detailed numerical studies of the experimental results also suggest the additional presence of a phase disorder effect ; the latter can be phenomenologically modelled by a small random on - site term .
however , we verified that the topological properties of the lattice are unaffected by these sources of disorder ( sup . mat . ) . in order to experimentally demonstrate the presence of anomalous topological edge modes , 785 nm light was launched at multiple locations around the edge of the lattice ; the white circles in fig .
[ exp ] indicate the launch site . if the light is launched at the middle of an edge , as shown in fig .
[ exp ] ( a ) , then at the output it is observed that the light has moved along the edge with minimal penetration into the bulk . moreover , if the input position of the beam is moved further down the edge , fig .
[ exp ] ( b ) , the close proximity of the launch site to the left edge means that the light will encounter the corner of the lattice during its evolution .
this corner , as can be observed in the figure , does not cause backscattering but instead the light simply turns the corner and continues to propagate .
this robustness is also observed in fig .
[ exp ] ( c - d ) , which demonstrates the light moving around a missing lattice site without backscattering or penetrating into the bulk .
these observations provide evidence for the existence of protected chiral edge modes that can be almost exclusively excited by single - site excitation .
these experimental observations are well complemented by a theoretical analysis , which was obtained using the couplings extracted from the bond characterisation data . the right - hand image in each panel of fig .
[ exp ] shows the quasienergy spectrum for the lattice and the overlap , as a function of this quasienergy , of the different initial launch states with the numerically calculated floquet states .
fig .
[ exp ] ( a - e ) demonstrate how launching in a single edge site leads to almost ideal excitation of the edge modes with little probability for bulk modes being excited , even when the launch site is close to a defect or corner .
the absence of this bulk coupling accounts for the strong confinement of the light to the lattice edge .
the central result of these experimental images is that they show the existence of edge modes propagating in a chiral ( unidirectional ) manner , and which are not scattered by corners nor defects .
the absence of scattering provides strong evidence that there is topological protection for these edge modes . from the bond - strength measurements , combined with our theoretical analysis of the corresponding model , we demonstrate that these propagating states correspond to anomalous topological edge modes .
the single - site excitation that is used in the experimental setup can also be exploited to probe the different edge modes that are present in the lattice . moving the input site onto the bottom left edge , fig .
[ exp ] ( e ) , shows the light propagating with twice the group velocity to what is observed in fig .
[ exp ] ( a ) .
this experimentally demonstrates that , when the bond @xmath17 is different to the other three bonds , edge modes with different group velocities can be observed .
the exact dispersion relations of the edge modes can be altered by modifying the strength of @xmath17 .
the ability to tune the group velocity could present a valuable experimental tool for the future , which could allow , for instance , the investigation of how the non - scattering behaviour of edge modes is modified by the interplay between group velocity and non - linearity . the behaviour observed when light is launched in the bulk of the lattice , fig .
[ exp ] ( f ) , is markedly different to what occurs on the edge with the input state being almost exactly reformed at the output facet .
this refocusing is due to the beating between the two bulk bands which are separated by a bandgap of approximately @xmath49 . in an ideal lattice , @xmath68
, the bulk bands would be dispersionless [ fig .
[ fig2](b ) ] and so the initial state would continue to reform every two driving cycles .
in the experimental lattice , however , deviations of the bonds strengths from the ideal case cause the bands to become dispersive [ fig .
[ fig2](c ) ] , which indicates that the initial state would eventually disperse in lattices of very long propagation lengths .
these deviations could be addressed in future work by careful optimisation of the experimental setup and the waveguide bending profile , so as to allow a closer realisation of the ideal model .
the ability to easily excite edge modes , with almost unit efficiency , and the possible co - existence of chiral edge modes with a dispersionless bulk , are two particularly interesting features of this slowly - driven photonic system .
these properties make this lattice a promising platform for investigating topological transport properties in response to perturbations , such as external ( engineered ) fields , disorder , and particle - particle interaction ( as generated by optical non - linearities ) .
_ acknowledgements _ s. m. thanks heriot - watt university for a james watt phd scholarship .
a. s. acknowledges support from the epsrc cm - cdt .
n. g. thanks d. t. tran for discussions , and also the frs - fnrs belgium and the bspo under the pai project p7/18 dygest for support .
m. v. and p. . acknowledge support from epsrc ep / m024636/1 .
r. r. t. gratefully acknowledges funding from the uk science and technology facilities council ( stfc ) in the form of an stfc advanced fellowship ( st / h005595/1 ) . 36 natexlab#1#1bibnamefont # 1#1bibfnamefont # 1#1citenamefont # 1#1url # 1`#1`urlprefix[2]#2 [ 2][]#2 , , ,
, * * , ( ) . ,
* * , ( ) . , * * , ( ) . , , , , * * , ( ) . , , , * * , ( ) . , , , , * * , ( ) . , * * , ( ) . , ( ) . , * * , ( ) . , , , * * , ( ) . , , , , , , , , , , , * * , ( ) . , * * , ( ) . , , , , * * , ( ) . , , , , , , , * * , ( ) . , , , , , , , , , * * , ( ) . , , , , , , ,
, , * * , ( ) . , , , * * , ( ) .
, , , , * * , ( ) . , , , , , * * , ( ) . , , , , , , * * , ( ) . , , , , , , * * , ( ) . , * * , ( ) . , , , * * , ( ) . , , , , , , , , , * * , ( ) . , , , , , , , , , , , ( ) . , , , , , , * * , ( ) . , , , , * * , ( ) . , * * , ( ) . , * * , ( ) . , , , , , , , * * , ( ) . , , , , * * , ( ) . , , , ,
* * , ( ) . ,
* * , ( ) . , _ _ ( , ) . , * * , ( ) . , , , , * * , ( ) . |
steady - state flows of incompressible viscous newtonian fluids are modeled by the navier - stokes equations @xmath1 for the velocity @xmath2 and the pressure @xmath3 . for this system
, one can consider different boundary conditions .
for example on solid walls , we have the dirichlet condition @xmath4 . on other parts of the boundary ( an artificial boundary such as the exit of a channel , or a free surface ) a no - friction condition ( neumann condition ) @xmath5
may be useful . here
@xmath6 denotes the matrix with the components @xmath7 , and @xmath8 is the outward normal .
note that the neumann problem naturally appears in the theory of hydrodynamic potentials ( see @xcite ) .
it is also of interest to consider boundary conditions containing components of the velocity and of the friction .
frequently used combinations are the normal component of the velocity and the tangential component of the friction ( slip condition for uncovered fluid surfaces ) or the tangential component of the velocity and the normal component of the friction ( condition for in / out - stream surfaces ) . in the present paper
, we consider mixed boundary value problems for the system ( [ navier - s ] ) in a three - dimensional domain @xmath9 of polyhedral type , where components of the velocity and/or the friction are given on the boundary . to be more precise , we have one of the following boundary conditions on each face @xmath10 : * @xmath11 , * @xmath12 , * @xmath13 , * @xmath14 , where @xmath15 denotes the normal and @xmath16 the tangential component of @xmath2 , @xmath17 is the vector @xmath18 , @xmath19 is the normal component and @xmath20 the tangential component of @xmath17 .
weak solutions , i.e. variational solutions @xmath21 , always exist if the data are sufficiently small . in the case when the boundary conditions ( ii ) and ( iv ) disappear , such solutions exist for arbitrary @xmath22 ( see the books by ladyzhenskaya @xcite , temam @xcite , girault and raviart @xcite ) . our goal is to prove regularity assertions for weak solutions . as is well - known , the local regularity result @xmath23 is valid outside an arbitrarily small neighborhood of the edges and vertices if the data are sufficiently smooth . here
@xmath24 denotes the sobolev space of functions which belong to @xmath25 together with all derivatives up to order @xmath26 .
the same result holds for the hlder space @xmath27 .
since solutions of elliptic boundary value problems in general have singularities near singular boundary points , the result can not be globally true in @xmath9 without any restrictions on @xmath26 and @xmath28 . here
we give a few particular regularity results which are consequences of more general theorems proved in the present paper .
suppose that the data belong to corresponding sobolev or hlder spaces and satisfy certain compatibility conditions on the edges .
then the following smoothness of the weak solution is guaranteed and is the best possible . *
if @xmath29 is a solution of the dirichlet problem in an arbitrary polyhedron or a solution of the neumann problem in an arbitrary lipschitz graph polyhedron , then @xmath30 here @xmath31 is a positive number depending on the domain @xmath9 . *
if @xmath29 is a solution of the dirichlet problem in a convex polyhedron , then @xmath32 * if @xmath29 is a solution of the mixed problem in an arbitrary polyhedron with the dirichlet and neumann boundary condition prescribed arbitrarily on different faces , then @xmath33 * let @xmath29 be a solution of the mixed boundary value problem with slip condition ( iii ) on one face @xmath34 and dirichlet condition on the other faces .
then @xmath35 where @xmath36 is the maximal angle between @xmath34 and the adjoining faces .
general facts of such a kind imply various precise regularity statements for special domains .
let us consider for example the flow outside a regular polyhedron @xmath37 .
on the boundary of @xmath37 , the dirichlet condition is prescribed .
then we obtain @xmath38 on every bounded subdomain of the complement of @xmath37 , where the best possible value for @xmath28 is ( with all digits shown correct ) + in the last decades , numerous papers appeared which treat boundary value problems for elliptic equations and systems in piecewise smooth domains . for the stationary linear stokes system
see e.g. the references in the book @xcite .
the properties of solutions of the dirichlet problem to the nonlinear navier - stokes system in 2-dimensional polygonal domains were studied in papers by kondratev @xcite , kellogg and osborn @xcite , kalex @xcite , orlt and sndig @xcite .
in particular , kellogg and osborn proved that the solution of the dirichlet problem belongs to @xmath39 if @xmath40 and the polygon @xmath9 is convex .
kalex , orlt and sndig considered solutions of mixed boundary value problems in polygonal domains .
mixed boundary value problems with boundary conditions ( i ) and ( iii ) in 3-dimensional domains with smooth non - intersecting edges were handled by solonnikov @xcite , mazya , plamenevski and stupyalis @xcite .
they proved in particular the solvability in weighted sobolev and hlder spaces . for the case of the dirichlet problem and a polyhedral domain , solvability and regularity results in weighted sobolev and
hlder spaces were proved by mazya and plamenevski @xcite .
concerning regularity results in @xmath41 sobolev spaces , we refer also to the papers of nicaise @xcite ( dirichlet problem ) , ebmeyer and frehse @xcite ( mixed problem with boundary conditions ( i ) and ( iii ) ) .
ebmeyer and frehse proved that @xmath42 with arbitrary real @xmath43 if the angle at the edge , where the boundary conditions change , is less than @xmath44 .
finally , we mention the papers by deuring and von wahl @xcite , dindos and mitrea @xcite dealing with the navier - stokes system in lipschitz domains .
the present paper consists of four sections .
section 1 concerns the existence of weak solutions in @xmath45 . in section 2 ,
we introduce and study weighted sobolev and hlder space . here
the weights are powers of the distances @xmath46 and @xmath47 to the vertices and edges of the domain @xmath9 , respectively .
in particular , we establish imbedding theorems for these spaces .
in contrast to the papers @xcite , we use weighted spaces with `` nonhomogeneous '' norms .
the weighted sobolev space @xmath48 in our paper is defined as the set of all functions @xmath2 in @xmath9 such that @xmath49 for all @xmath50 , where @xmath51 .
the norm in the weighted hlder space @xmath52 has a similar structure .
the use of weighted spaces with nonhomogeneous norms has several advantages .
first , these spaces are applicable to a wider class of boundary value problems . for @xmath53 and @xmath54
they are closely related to the nonweighted spaces . furthermore in some cases
( e.g. the dirichlet problem when the edge angles are less than @xmath44 ) , it is possible to obtain higher regularity results when considering solutions in weighted spaces with nonhomogeneous norms .
so we can partially improve the results in @xcite .
the main results of the paper are contained in section 3 .
for the proofs , we use results of our previous papers @xcite , where we studied mixed boundary value problems for the linear stokes system .
we show in particular that the weak solution @xmath29 belongs to the weighted space @xmath55 if the data are from corresponding spaces , satisfy certain compatibility conditions , and the numbers @xmath56 and @xmath57 are positive and sufficiently small .
the precise conditions on @xmath58 and @xmath59 are given in terms of eigenvalues of certain operator pencils .
the general results in section 3 together with estimates for the eigenvalue of these pencils ( see @xcite ) allow us in particular to deduce regularity assertions in nonweighted sobolev and hlder spaces .
a number of examples is given at the end of section 3 .
the last section concerns the solvability in the space @xmath60 , where @xmath28 may be less than 2 . here
, we assume that the dirichlet condition is given on at least one of the adjoining faces of every edge @xmath61 .
one of our results is the following .
let @xmath9 be an arbitrary polyhedron .
then the problem ( [ navier - s ] ) with @xmath62 and dirichlet condition @xmath4 on the boundary has a weak solution @xmath63 for arbitrary @xmath64 , @xmath65 , provided the norm of @xmath22 is sufficiently small .
the same result holds for the mixed problem with boundary condition ( i)(iii ) if we suppose that the angles at the edges where the boundary conditions change are less or equal to @xmath66 .
in the following , let @xmath67 be the dihedron @xmath68 where @xmath69 are the polar coordinates in the @xmath70-plane , @xmath71 , @xmath72 .
furthermore , let @xmath73 be a polyhedral cone with plane faces @xmath74 and edges @xmath75 .
the bounded domain @xmath76 is said to be a _ domain of polyhedral type _
if * the boundary @xmath77 consists of smooth ( of class @xmath78 ) open two - dimensional manifolds @xmath10 ( the faces of @xmath9 ) , @xmath79 , smooth curves @xmath61 ( the edges ) , @xmath80 , and vertices @xmath81 , * for every @xmath82 there exist a neighborhood @xmath83 and a diffeomorphism ( a @xmath78 mapping ) @xmath84 which maps @xmath85 onto @xmath86 , where @xmath87 is a dihedron of the form ( [ dih ] ) and @xmath88 is the unit ball , * for every vertex @xmath89 there exist a neighborhood @xmath90 and a diffeomorphism @xmath91 mapping @xmath92 onto @xmath93 , where @xmath94 is a polyhedral cone with vertex at the origin .
the set @xmath95 of the singular boundary points is denoted by @xmath96 . for every face @xmath10 , @xmath79 , let a number @xmath97 be given .
we consider the boundary value problem @xmath98 where @xmath99 by a weak solution of the problem ( [ ns1 ] ) , ( [ ns2 ] ) , we mean a vector function @xmath100 satisfying @xmath101 where @xmath102 , @xmath103 note that for arbitrary @xmath104 , the functional @xmath105 is continuous on @xmath106 .
this follows from the inequality @xmath107 and the continuity of the imbedding @xmath108 .
we consider the weak solution of the boundary value problem for the stokes system @xmath109 i.e. a vector function @xmath100 satisfying @xmath110 where @xmath111 is given by ( [ a4 ] ) .
for the proof of the following theorem , we refer to ( * ? ? ?
* th.5.1 ) .
[ stokest1 ] let @xmath112 and @xmath113 be such that there exists a vector function @xmath114 , @xmath115 on @xmath10 , @xmath79 . in the case when @xmath116 for all @xmath117 , we assume in addition that @xmath118 furthermore , let the functional @xmath119 satisfy the condition @xmath120 where @xmath121 denotes the set of all @xmath122 such that @xmath123 for @xmath124 .
then there exists a solution @xmath125 of the problem _ ( [ stokes1 ] ) , ( [ stokes2])_. here @xmath3 is uniquely determined if @xmath126 for at least one @xmath117 and unique up to constants if @xmath127 for all @xmath117 .
the vector function @xmath2 is unique up to elements from @xmath121 .
note that @xmath121 contains only functions of the form @xmath128 , where @xmath129 is a constant vector and @xmath130 is a constant matrix , @xmath131 ( rigid body motions ) .
in particular , @xmath132 for @xmath133 . in most cases
( e.g. if the dirichlet condition is given on at least one face @xmath10 ) , the set @xmath121 contains only the function @xmath134 .
let the operator @xmath135 be defined by @xmath136 obviously , @xmath135 realizes a mapping @xmath137 .
furthermore , there exist constants @xmath138 such that @xmath139 using the last two estimates together with theorem [ stokest1 ] , we can prove the following statement .
[ stokest1a ] let @xmath140 and @xmath141 be as in theorem _
[ stokest1]_. furthermore , we suppose that @xmath142 and @xmath143 is sufficiently small .
then there exists a solution @xmath144 of the problem _ ( [ a1 ] ) , ( [ a2])_. here , @xmath2 is unique on the set of all functions with norm less than a certain positive @xmath31 , @xmath3 is unique if @xmath126 for at least one @xmath117 , otherwise @xmath3 is unique up to a constant .
p r o o f. let @xmath145 be the solution of the linear problem ( [ stokes1 ] ) , ( [ stokes2 ] ) . by our assumptions on @xmath111 , @xmath140 and @xmath141 , we may assume that @xmath146 where @xmath147 is a small positive number .
let @xmath148 denote the set of all @xmath122 such that @xmath132 .
we put @xmath149 and @xmath150 .
then @xmath29 is a solution of the problem ( [ a1 ] ) , ( [ a2 ] ) if and only if @xmath151 and @xmath152 by theorem [ stokest1 ] , there is a linear and continuous mapping @xmath153 defined by @xmath154 we write ( [ 1stokest1a ] ) as @xmath155 due to ( [ q2 ] ) , the operator @xmath156 is contractive on the set of all @xmath157 with norm @xmath158 if @xmath147 and @xmath159 are sufficiently small . hence there exist @xmath160 and @xmath161 satisfying ( [ 1stokest1a ] ) .
the result follows .
_ if @xmath163 for all @xmath117 , then @xmath164 ( see ( * ? ?
* le.iv.2.2 ) ) .
thus , analogously to ( * ? ?
* th.iv.2.3 ) ( see also @xcite ) , the problem ( [ a1 ] ) , ( [ a2 ] ) hast at least one solution for arbitrary @xmath119 , @xmath62 , @xmath113 satisfying ( [ 1stokest1 ] ) . _
here , we introduce weighted sobolev and hlder spaces in polyhedral domains and prove imbeddings for these spaces which will be used in the next section .
we start with the case of a polyhedral cone .
let @xmath165 be a polyhedral cone in @xmath166 whose boundary consists of plane faces @xmath10 and edges @xmath61 , @xmath167 .
we denote by @xmath168 the distance of @xmath169 to the vertex of the cone , by @xmath170 the distance to the edge @xmath61 , and by @xmath171 the distance to the set @xmath172 .
note that there exist positive constants @xmath173 , @xmath174 such that @xmath175 let @xmath26 be a nonnegative integer , @xmath176 , @xmath177 , and @xmath178 .
we define @xmath179 as the closure of the set @xmath180 with respect to the norm @xmath181 the weighted sobolev space @xmath182 , where @xmath183 for @xmath184 , is defined as the closure of the set @xmath185 with respect to the norm @xmath186 if @xmath59 is a real number , then by @xmath179 and @xmath182 , we mean the above introduced spaces with @xmath187 . for the proof of the following lemma we refer to ( * ? ? ?
* le.1 ) .
[ al2 ] let @xmath188 , @xmath189 , @xmath190 , and @xmath191 for @xmath184 .
then @xmath179 is continuously imbedded into @xmath192 .
in particular , we have @xmath193 for @xmath194 , @xmath195 and @xmath196 , @xmath184 . if in addition @xmath197 and @xmath198 for @xmath184 , then also @xmath199 the spaces @xmath179 and @xmath182 coincide if @xmath200 for all @xmath201 ( see @xcite ) .
[ al3 ] let @xmath202 and @xmath203 .
then @xmath204 and @xmath205 for all @xmath206 with a constant @xmath129 independent of @xmath2 .
furthermore , for arbitrary @xmath207 , @xmath208 , the following inequality is valid .
@xmath209 p r o o f. let @xmath210 , and let @xmath211 be infinitely differentiable functions with support in @xmath212 such that @xmath213 we define @xmath214 and @xmath215 .
then @xmath216 vanishes for @xmath217 and @xmath218 .
therefore for @xmath219 , we have @xmath220 ( see ( * ? ? ?
* th.3 ) ) .
this inequality together with the continuity of the imbedding @xmath221 implies @xmath222 where @xmath129 is independent of @xmath2 and @xmath201 . using the equalities
@xmath223 we obtain @xmath224 consequently , @xmath225 this proves ( [ 1al3 ] ) . the proof of ( [ 2al3 ] ) proceeds analogously .
@xmath162 [ al4 ] let @xmath206 , and let @xmath117 be an integer , @xmath226 . then there exist functions @xmath227 and @xmath228 such that @xmath229 and @xmath230 p r o o f.
let @xmath211 be the same functions as in the proof of lemma [ al3 ] , and let @xmath2 be an arbitrary function from @xmath182
. obviously , the function @xmath231 defined by @xmath232 belongs also to @xmath182 and vanishes for @xmath217 and @xmath218 .
consequently by ( * ? ? ?
* le.1.3 ) ( for integer @xmath233 see ( * ? ? ?
* th.5 ) ) , there exist functions @xmath234 and @xmath235 with supports in @xmath236 such that @xmath237 and @xmath238 where @xmath129 is independent of @xmath2 and @xmath201 .
let @xmath239 and @xmath240 .
then the supports of @xmath241 and @xmath242 are contained in @xmath243 , and we have @xmath244 for all @xmath201 .
moreover , @xmath245 let @xmath246 and @xmath247 . then @xmath229 .
since @xmath241 and @xmath248 have disjoint supports for @xmath249 , we have @xmath227 and @xmath250 analogously , the norm of @xmath251 in @xmath252 can be estimated by the norm of @xmath2 in @xmath182
. the lemma is proved .
let @xmath26 be a nonnegative integer , @xmath176 , @xmath177 , and @xmath253 .
we define the weighted hlder space @xmath254 as the set of all @xmath26 times continuously differentiable functions on @xmath255 with finite norm @xmath256 furthermore , the space @xmath257 is defined for nonnnegative @xmath258 , @xmath184 , as the set of all @xmath26 times continuously differentiable functions on @xmath255 with finite norm @xmath259 where @xmath260 , @xmath261 + 1 $ ] , @xmath262 $ ] denotes the greatest integer less or equal to @xmath28 .
the trace spaces for @xmath254 and @xmath257 on @xmath10 are denoted by @xmath263 and @xmath264 , respectively .
obviously , @xmath265 is a subset of @xmath257 . if @xmath266 for @xmath184 , then both spaces coincide
furthermore , the following imbedding holds ( and is continuous , see @xcite ) .
@xmath267 if in addition @xmath258 and @xmath268 are nonnegative , then @xmath257 is continuously imbedded into @xmath269 .
next , we prove a relation between the spaces @xmath270 and @xmath271 .
[ al5 ] suppose that @xmath272 , @xmath273 and @xmath274 for @xmath184 . then @xmath179 is continuously imbedded into @xmath275 .
p r o o f. it suffices to prove the lemma for @xmath276 , @xmath184 .
let @xmath277 .
for an arbitrary point @xmath278 , we denote by @xmath279 the set @xmath280 first , let @xmath281 .
from the continuity of the imbedding @xmath282 it follows that there exists a constant @xmath129 independent of @xmath2 and @xmath169 such that @xmath283 due to ( [ rx ] ) and ( [ 1al5 ] ) , this implies @xmath284 and analogously @xmath285 now let @xmath169 be an arbitrary point in @xmath286 and @xmath287 .
we put @xmath288 , @xmath289
. then @xmath290 and @xmath291 .
consequently , the function @xmath292 satisfies the inequalities @xmath293 for @xmath294 and @xmath295 for @xmath296 . here
, the constant @xmath129 is independent of @xmath2 , @xmath169 and @xmath297 . using ( [ rx ] )
, we obtain the inequalities @xmath298 the result follows .
@xmath162 [ ac2 ] let @xmath206 , @xmath299 .
then @xmath300 where @xmath301 for @xmath302 , @xmath303 for @xmath304 , and @xmath305 for @xmath200 ( @xmath31 is an arbitrarily small positive number ) .
p r o o f. let @xmath306 be a smooth function on the unit sphere @xmath307 such that @xmath308 in a neighborhood of the points @xmath309 for @xmath310 .
we extend @xmath306 to @xmath311 by @xmath312
. then @xmath313 . obviously , it suffices to prove ( [ 1ac2 ] ) for the function @xmath314 . if @xmath302 , then by lemma [ al3 ] , @xmath315 . if @xmath200 , then @xmath316 . by lemma [ al5 ]
, the last space is imbedded into @xmath317 for arbitrary @xmath318 .
therefore in particular , @xmath319 for @xmath320 , the assertion follows from the imbedding @xmath321 .
@xmath162 + let @xmath9 be a domain of polyhedral type ( see section 1.1 ) with faces @xmath74 , edges @xmath322 and vertices @xmath81 .
we denote the distance of @xmath169 to the edge @xmath61 by @xmath170 , the distance to the vertex @xmath89 by @xmath323 , the distance to @xmath96 ( the set of all edge points and vertices ) by @xmath171 , and the distance to the set @xmath324 by @xmath325 . for arbitrary integer @xmath326 , real @xmath327 and real tuples
@xmath328 , @xmath329 , we define @xmath330 and @xmath48 as the weighted sobolev spaces with the norms @xmath331 respectively . in the case of the space @xmath48 , we suppose that @xmath183 for @xmath80 .
the corresponding trace spaces on the faces @xmath10 are denoted by @xmath332 and @xmath333 , respectively .
furthermore , let @xmath334 denote the dual space of @xmath335 , @xmath336 , with respect to the @xmath41 scalar product .
[ al7 ] let @xmath337 , @xmath338 for @xmath339 , and @xmath340 for @xmath184 .
then @xmath341 and @xmath342 .
these imbeddings are continuous .
p r o o f. let @xmath343 . by hlder s inequality , @xmath344 where @xmath345 if @xmath346 and @xmath347 .
this proves the imbedding @xmath342 .
analogously , the imbedding @xmath348 holds .
@xmath162 + the following result can be directly deduced from lemma [ al2 ] .
[ al8 ] let @xmath188 , @xmath189 , @xmath349 for @xmath339 , and @xmath350 for @xmath80 .
then @xmath330 is continuously imbedded into @xmath351 .
[ ac1 ] let @xmath188 , @xmath352 , @xmath353 for @xmath339 , and @xmath354 for @xmath184 .
then @xmath355 is continuously imbedded into @xmath356 .
p r o o f. according to lemma [ al8 ] , we have @xmath357 , where @xmath336 , @xmath358 .
the result follows .
@xmath162 + let us further note that ( as in the case of a cone ) @xmath359 for @xmath339 , @xmath80 . if @xmath200 for @xmath80 , then @xmath360 .
+ we introduce the following weighted hlder spaces in the domain @xmath9 .
the space @xmath361 is defined as the set all @xmath26 times continuously differentiable functions on @xmath362 with finite norm @xmath363 suppose that @xmath364 for @xmath80 .
then @xmath365 , is defined as the set of all @xmath26 times continuously differentiable functions on @xmath366 with finite norm @xmath367 where @xmath368 and @xmath261 + 1 $ ] .
the trace space on @xmath10 for @xmath365 is denoted by @xmath264 .
analogously to the case when the domain is a cone , we have @xmath369 for @xmath339 , @xmath80 .
if in addition @xmath258 and @xmath268 are nonnegative , then @xmath370 .
furthermore , it follows from lemma [ al5 ] that @xmath371 for @xmath339 , @xmath80 .
we introduce the following notation . if @xmath372 and @xmath373 , then by @xmath374 and @xmath375 , we mean the spaces @xmath376 and @xmath377 with @xmath378 , @xmath379 .
analogous notation will be used for the weighted sobolev spaces @xmath270 and @xmath380 .
the next lemma follows immediately from the definition of the space @xmath381 .
[ al6 ] if @xmath382 and @xmath383 , then @xmath384 .
finally , we define @xmath385 as the space of all distributions of the form @xmath386 note that every @xmath387 , i.e. every distribution of the form @xmath388 belongs to @xmath389 .
indeed , let @xmath390 be infinitely differentiable cut - off functions equal to one near @xmath391 and to zero near the vertices @xmath392 , @xmath393 .
then the distribution ( [ cmin ] ) can be written as @xmath394 @xmath395 . here , @xmath396 , and from @xmath397
, @xmath398 it follows that @xmath399 for @xmath395 .
in this section , we establish regularity results for weak solutions in weighted sobolev and hlder spaces .
the regularity assertions are formulated in terms of eigenvalues of operator pencils generated by the boundary value problem at the edge points and vertices of the domain .
we introduce the operator pencils generated by the problem ( [ ns1 ] ) , ( [ ns2 ] ) for the edge points and vertices of the domain @xmath9 .
\1 ) let @xmath400 be a point on an edge @xmath61 , and let @xmath401 be the faces of @xmath9 adjacent to @xmath400 .
then by @xmath87 we denote the dihedron which is bounded by the half - planes @xmath402 tangent to @xmath403 at @xmath400 and the edge @xmath404 .
the angle between the half - planes @xmath402 is denoted by @xmath405 .
furthermore , let @xmath69 be polar coordinates in the plane perpendicular to @xmath406 such that @xmath407 then we define the operator @xmath408 as follows : @xmath409 where @xmath410 , @xmath411 , @xmath412 .
the operator @xmath408 depends quadratically on the parameter @xmath413 and realizes a continuous mapping @xmath414 for every @xmath415 , where @xmath416 denotes the interval @xmath417 .
the spectrum of the pencil @xmath408 consists of eigenvalues with finite geometric and algebraic multiplicities .
these eigenvalues are zeros of certain transcendental functions ( see @xcite ) .
for example , in the cases @xmath418 ( dirichlet conditions on @xmath403 ) and @xmath419 ( neumann conditions on @xmath403 ) , the spectrum of @xmath408 consists of the solutions of the equation @xmath420 @xmath421 for @xmath418 . let @xmath422 be the eigenvalue with smallest positive real part of this pencil , and let @xmath423 be the eigenvalue with smallest real part greater than 1 . we define @xmath424 where @xmath425 if @xmath426 , @xmath427 if @xmath428 . finally , let @xmath429 note that in the case of even @xmath430 , the number @xmath431 belongs always to the spectrum of the pencil @xmath408 .
+ 2 ) let @xmath89 be a vertex of @xmath9 and let @xmath432 be the set of all indices @xmath201 such that @xmath433 . by our assumptions , there exist a neighborhood @xmath434 of @xmath89 and a diffeomorphism @xmath435 mapping @xmath436 onto @xmath437 and @xmath438 onto @xmath439 for @xmath440 , where @xmath441 is a polyhedral cone with vertex @xmath442 and @xmath443 are the faces of this cone . without loss of generality , we may assume that the jacobian matrix @xmath444 is equal to the identity matrix @xmath445 at the point @xmath89 .
we introduce spherical coordinates @xmath446 , @xmath447 in @xmath94 and define @xmath448 on the space @xmath449 , we define the bilinear form @xmath450 as @xmath451 where @xmath452 , @xmath453 , @xmath454 , @xmath455 , @xmath456 , @xmath457 , and @xmath458 .
this bilinear form generates the linear and continuous operator @xmath459 by @xmath460 the operator @xmath461 depends quadratically on the complex parameter @xmath413 .
the spectrum of the pencil @xmath461 consists of isolated points , eigenvalues with finite geometric and algebraic multiplicities .
the following two theorems are proved in @xcite .
[ stokest2a ]
let @xmath462 be a solution of the problem _
( [ stokes1 ] ) , ( [ stokes2])_. suppose that the following conditions are satisfied .
* @xmath463 , @xmath464 , @xmath465 , * there are no eigenvalues of the pencils @xmath461 , @xmath339 in the closed strip between the lines @xmath466 and @xmath467 , * the components of @xmath59 satisfy the inequalities @xmath468
. then @xmath469 and @xmath470 .
[ stokest2b ] let @xmath462 be a solution of the problem _
( [ stokes1 ] ) , ( [ stokes2])_. suppose that * @xmath471 , @xmath472 , and @xmath119 has the representation _ ( [ a4 ] ) _ with @xmath473 , @xmath474 , * there are no eigenvalues of the pencils @xmath461 , @xmath339 , in the closed strip between the lines @xmath475 and @xmath476 , * the components of @xmath59 satisfy the inequalities @xmath477 , * @xmath140 , @xmath141 and @xmath478 are such that there exist @xmath479 and @xmath480 satisfying @xmath481 then @xmath482 and @xmath483 . for the proof of the following two regularity assertions in weighted hlder spaces , we refer to @xcite .
[ stokest3a ] let @xmath462 be a weak solution of the problem _
( [ stokes ] ) , ( [ ns2])_. suppose that * @xmath484 , @xmath485 , @xmath486 , @xmath487 , * @xmath488 for @xmath339 , and the strip @xmath489 is free of eigenvalues of the pencils @xmath461 , @xmath339 , * the components of @xmath59 are nonnegative and satisfy the inequalities @xmath490 , @xmath491 , * @xmath140 , @xmath141 and @xmath478 are such that there exist @xmath492 and @xmath493 satisfying @xmath494 then @xmath495 .
note that under the conditions of theorem [ stokest3a ] on @xmath58 and @xmath59 , there are the imbeddings @xmath496 , @xmath497 , and @xmath498 . [ stokest3 ]
let @xmath462 be a weak solution of the problem _
( [ stokes ] ) , ( [ ns2])_. suppose that * @xmath499 , @xmath500 , @xmath501 , @xmath502 , * @xmath503 for @xmath339 , the strip @xmath504 is free of eigenvalues the pencil @xmath461 , @xmath339 , * the components of @xmath59 are nonnegative , satisfy the inequalities @xmath505 , @xmath491 , and @xmath506 , * @xmath140 , @xmath141 and @xmath478 are such that there exist @xmath507 and @xmath493 satisfying @xmath508 then @xmath509 .
_ for the validity of condition ( iv ) in theorems [ stokest2b][stokest3 ] it is necessary and sufficient that the functions @xmath141 and their derivatives , @xmath478 and @xmath140 satisfy certain compatibility conditions on the edges of the domain @xmath9 ( see @xcite ) . _
our goal is to extend the results of theorems [ stokest2a][stokest3 ] to the nonlinear problem ( [ ns1 ] ) , ( [ ns2 ] ) .
we start with regularity results in weighted sobolev spaces .
[ dl1 ] let @xmath510 , @xmath511 , @xmath512 for @xmath339 , and @xmath513 .
then @xmath514 for @xmath515 .
p r o o f. let @xmath516 . by hlders inequality , @xmath517 furthermore by corollary [ ac1 ] , the space @xmath518 is continuously imbedded into @xmath519 if @xmath520 and @xmath521 . the result follows .
@xmath162 [ dt1 ] let @xmath462 be a solution of the problem _ ( [ a1 ] ) , ( [ a2])_. we suppose that @xmath522 and that the conditions _
( i)-(iii ) _ of theorem _ [ stokest2a ] _ are satisfied . then @xmath523 and @xmath470 .
p r o o f. 1 ) first , let @xmath524 . from @xmath525 and @xmath526
it follows that @xmath527 .
this together with corollary [ ac1 ] implies @xmath528 .
hence , @xmath29 is a solution of the problem @xmath529 where @xmath530 from theorem [ stokest2a ] we conclude that @xmath531
. then by lemma [ dl1 ] , we have @xmath532 and therefore , @xmath533 consequently , theorem [ stokest2a ] implies @xmath534 . repeating the last consideration ,
we obtain @xmath535 .
\2 ) next , let @xmath536 .
then by lemma [ al7 ] , @xmath537 , @xmath538 , and @xmath539 , where @xmath540 , @xmath541 , @xmath31 is an arbitrarily small positive number . for sufficiently small @xmath31
, we conclude from part 1 ) that @xmath542 . by hlders inequality , @xmath543 due to corollary [ ac1 ] , @xmath544 if @xmath545 .
therefore , @xmath29 is a solution of the problem ( [ 1dt1 ] ) , ( [ 2dt1 ] ) , where @xmath546 . using theorem [ stokest2a ] ,
we obtain @xmath547 .
\3 ) finally , let @xmath548 .
then again by lemma [ al7 ] , @xmath549 , @xmath550 , and @xmath551 , where @xmath552 , @xmath553 , @xmath31 is an arbitrarily small positive number . for sufficiently small @xmath31 , we conclude from part 2 ) that @xmath554 .
since @xmath555 , it follows that @xmath556 .
the last space is embedded to @xmath557 if @xmath558 ( see corollary [ ac1 ] ) .
therefore , @xmath29 is a solution of the problem ( [ 1dt1 ] ) , ( [ 2dt1 ] ) , where @xmath559 . applying theorem [ stokest2a ]
, we obtain @xmath560 .
@xmath162 + for the proof of the analogous @xmath561 regularity result , we need the following lemma .
[ dl2 ] let @xmath562 , @xmath563 , @xmath564 , @xmath565 .
then @xmath566 for every @xmath567 , @xmath521 , @xmath568 .
1 ) suppose that @xmath569 for @xmath80 .
then by lemma [ al8 ] , @xmath570 , @xmath571 . from this , from the assumption @xmath572 and from hlder
s inequality it follows that @xmath573 , @xmath574 2 ) we consider the case when @xmath575 for all @xmath201 . then @xmath2 admits the decomposition @xmath576 from lemma [ al3 ] it follows that @xmath577 since @xmath578 , we have in particular @xmath579 and therefore also @xmath580 .
we estimate the norms of @xmath581 and @xmath582 in @xmath583 .
let @xmath571 .
using hlder s inequality and lemma [ al8 ] , we obtain @xmath584 by lemma [ al8 ] , the space @xmath585 is continuously imbedded into @xmath586 . consequently , @xmath587 thus , we have @xmath588 . furthermore ,
using the continuity of the imbedding @xmath589 , we obtain @xmath590 this proves the lemma for the case @xmath591 , @xmath80 .
\3 ) the case when @xmath591 for some but not all @xmath201 can be reduced to cases 1 ) and 2 ) using suitable cut - off functions .
@xmath162 [ dt2 ]
let @xmath462 be a solution of the problem _
( [ a1 ] ) , ( [ a2])_. we assume that the conditions _
( i)(iv ) _ of theorem _ [ stokest2b ] _ are satisfied and that @xmath592 for @xmath339
. then @xmath482 and @xmath593 .
p r o o f. 1 ) let first @xmath594 .
we put @xmath595 if @xmath43 , @xmath596 if @xmath597 . since @xmath598
we obtain @xmath599 if @xmath600 and @xmath601 or , what is the same , if @xmath602 for @xmath79 , @xmath603 for @xmath80 .
let @xmath604 , @xmath605 , where @xmath31 is a sufficiently small positive number .
then @xmath29 is a solution of the problem @xmath606 where @xmath607 , @xmath608 , @xmath609 and @xmath610 .
consequently by theorem [ stokest2b ] , we have @xmath611 . applying lemma [ dl2 ] ,
we obtain @xmath612 , where @xmath613 and @xmath614 .
hence , theorem [ stokest2b ] implies @xmath615 . repeating this procedure ,
we obtain @xmath616 .
\2 ) next , we consider the case @xmath617 .
let @xmath31 be a positive number less than 1/2 such that @xmath618 for all @xmath201 .
then by lemma [ al7 ] , @xmath619 , and from part 1 ) it follows that @xmath620 if @xmath31 is sufficiently small .
in particular , @xmath621 ( see lemma [ al8 ] ) .
let @xmath622 for @xmath80 .
then @xmath623 where @xmath624 consequently , @xmath625 , and theorem [ stokest2b ] implies @xmath626 . using lemma [ dl2 ] , we obtain @xmath627 analogously to the first part of the proof .
\3 ) let @xmath628 , and let @xmath31 be a positive number less than 1/2 such that @xmath629 for all @xmath201 .
then by lemma [ al7 ] , @xmath630 .
therefore by part 2 ) , we have @xmath631 provided @xmath31 is sufficiently small .
consequently , @xmath632 let @xmath622 for @xmath80 .
then @xmath633 where @xmath634 thus , we have @xmath625 and @xmath635 ( by theorem [ stokest2b ] ) . using again lemma [ dl2 ] , we obtain @xmath636 .
\4 ) finally , let @xmath637 .
we define @xmath638 , where @xmath31 is a sufficiently small positive number .
then we have @xmath639 furthermore , the functional ( [ a4 ] ) belongs to @xmath640 . since @xmath641 it follows from theorem [ dt1 ] that @xmath642 .
then by corollary [ ac2 ] , @xmath643 where @xmath301 for @xmath644 , @xmath303 for @xmath645 . by hlder s inequality
, we have @xmath646 for the last inequality , we used the fact that @xmath647 and @xmath648 . hence , @xmath29 is a solution of the problem ( [ 1dt1 ] ) , ( [ 2dt1 ] ) , where @xmath649 is a functional of the form ( [ a4 ] ) with @xmath650 , @xmath651 . applying theorem [ stokest2b ] , we obtain @xmath652 and @xmath593 . @xmath162 in order to extend the results of theorems [ stokest3a ] and [ stokest3 ] to problem ( [ ns1 ] ) , ( [ ns2 ] ) , we consider first the nonlinear term in the navier - stokes system
. [ el1 ] let @xmath653 , where @xmath654 for @xmath339 , @xmath655 , @xmath656 is not an integer for @xmath80 .
then @xmath657 for every @xmath567 such that @xmath658 , @xmath80 .
p r o o f. we have to show that there exists a constant @xmath659 such that @xmath660 and @xmath661 here @xmath662 .
inequality ( [ 1el1 ] ) follows immediately from the estimates @xmath663 and the inequalities @xmath664 , @xmath665 .
furthermore for @xmath666 , we have @xmath667 and analogously @xmath668 hence , @xmath669 for @xmath666 . from this estimate and from the inequalities
@xmath670 @xmath671 , @xmath672 , and @xmath664 , we obtain ( [ 2el1 ] ) .
analogously , we obtain the estimate @xmath673 since @xmath674 , there exists a constant @xmath659 such that @xmath675 this together with ( [ 4el1 ] ) implies @xmath676 thus , estimate ( [ 3el1 ] ) holds .
the proof is complete .
@xmath162 [ et1 ]
let @xmath462 be a weak solution of the problem _
( [ ns1 ] ) , ( [ ns2 ] ) _ , and let the conditions _ ( i)(iv ) _ of theorem _ [ stokest3 ] _ are satisfied
. then @xmath677 and @xmath678 .
p r o o f. suppose first that @xmath679 for @xmath80 .
let @xmath31 be an arbitrarily small positive number and @xmath28 an arbitrary real number greater than 1 .
we put @xmath680 for @xmath79 and @xmath681 . from our assumptions on @xmath682 and
@xmath478 it follows that @xmath683 using theorem [ dt2 ] , we obtain @xmath684 .
consequently , @xmath685 for @xmath124 , where @xmath686 for @xmath80 .
this together with hlder s inequality implies @xmath687 therefore , @xmath688 .
the numbers @xmath31 and @xmath28 can be chosen such that @xmath689 for @xmath339 , @xmath690 for @xmath80 , @xmath691 , and @xmath692 .
then @xmath693 , @xmath694 , and @xmath695 .
consequently , @xmath696 ( see lemma [ al5 ] ) . hence , @xmath29 is a solution of the problem ( [ 1dt2 ] ) , ( [ 2dt2 ] ) , where @xmath697 .
applying theorem [ stokest3 ] , we obtain @xmath698 .
suppose now that @xmath699 for at least one @xmath201 . by the first part of the proof , we obtain @xmath700 , where @xmath701
, @xmath31 is an arbitrarily small positive real number .
then lemma [ el1 ] implies @xmath697 , and from theorem [ stokest3 ] it follows that @xmath698 .
the proof is complete .
@xmath162 + finally , we prove the analogous @xmath702-regularity result .
[ et2 ] let @xmath462 be a weak solution of the problem _
( [ ns1 ] ) , ( [ ns2])_. suppose that conditions _ ( i)(iv ) _ of theorem _ [ stokest3a ] _ are satisfied
. then @xmath703 and @xmath704 .
p r o o f. 1 ) let first @xmath679 for @xmath80 .
then @xmath705 and @xmath706 , where @xmath707 , @xmath708 , @xmath31 is an arbitrarily mall positive number , and @xmath327 .
furthermore , the functional ( [ a4 ] ) belongs to @xmath709 . using theorem [ dt1 ]
, we obtain @xmath710 .
we consider the term @xmath711 from the inclusions @xmath712 , @xmath713 it follows that @xmath714 if @xmath31 is sufficiently small and @xmath28 is sufficiently large ( see lemma [ al5 ] ) .
furthermore , from @xmath715 , @xmath716 and lemma [ al6 ] it follows that @xmath717 consequently , we have @xmath718 , and theorem [ stokest3a ] implies @xmath719 .
\2 ) suppose that @xmath720 for some @xmath201 . by the first part of the proof
, we have @xmath721 , where @xmath701 for @xmath80 , @xmath31 is an arbitrarily small positive number . in particular , @xmath722 , @xmath723 , and therefore ( by lemma [ al6 ] ) @xmath724
the last space is contained in @xmath725 for sufficiently small @xmath31 . applying theorem [ stokest3a ]
, we obtain @xmath726 .
@xmath162 let @xmath727 be the eigenvalue of the pencil @xmath728 with smallest real part @xmath729 .
we show that the inequalities @xmath730 in theorem [ dt2 ] can not be weakened .
we assume first that @xmath731 for some @xmath117 and that @xmath59 satisfies the second condition of ( [ nec1 ] ) . by our assumptions on the domain
, there exist a neighborhood @xmath90 of @xmath89 and a diffeomorphism @xmath435 mapping @xmath92 onto the intersection of a cone @xmath94 with the unit ball such that @xmath732 . in the new coordinates @xmath733 , the navier - stokes system takes the form @xmath734 where @xmath735 . here by @xmath736 we mean the matrix @xmath737 .
we consider the functions latexmath:[\[u= \zeta(y ) \ , |y|^{\lambda_j}\ , \phi(y/|y|),\quad p= \zeta(y ) \
, an eigenvector of the pencil @xmath728 corresponding to the eigenvalue @xmath727 and @xmath739 is a smooth cut - off function equal to one near the origin .
the eigenvector @xmath740 belongs to the space @xmath741 with arbitrary @xmath742 and @xmath743 satisfying @xmath744 . here , @xmath745 is the closure of the set @xmath746 with respect to the norm @xmath747 where @xmath2 is extended by @xmath748 to @xmath94 . in particular ,
@xmath749 and @xmath750 for @xmath751 .
since @xmath740 is an eigenvector , the vector function @xmath752 is a solution of the linear stokes system with zero right - hand sides . from this and from the equalities @xmath732 and @xmath735
, it follows that @xmath753 and @xmath754 if @xmath755 , @xmath756 .
analogously , the corresponding boundary data are from @xmath757 and @xmath757 , respectively .
however , @xmath758 for @xmath759 .
this example shows that the inequality @xmath760 can not be weakened .
now we show that the inequality @xmath761 can not be weakened .
we assume for the sake of simplicity that the edge @xmath61 is a part of the @xmath762-axis and that the adjacent faces @xmath763 and @xmath764 are plane .
let @xmath765 , and let @xmath766 be an eigenvalue of the pencil @xmath767 with the real part @xmath768 .
then we consider the functions @xmath769 where @xmath770 are the polar coordinates in the @xmath70-plane , @xmath740 is an eigenvector of the pencil @xmath767 corresponding to the eigenvalue @xmath766 , and @xmath739 is a smooth cut - off function equal to one in a neighborhood of a certain point @xmath771 and equal to zero near the other edges .
since the vector function @xmath772 is a solution of the linear stokes system with zero right - hand sides , it follows that @xmath773 if @xmath774 .
however , @xmath775 for @xmath776 .
this means , the result of theorem [ dt2 ] fails if @xmath777 .
analogously , it can be shown that the inequalities for @xmath778 and @xmath258 in theorems [ dt1 ] , [ et1 ] and [ et2 ] can not be weakened . here , we establish some regularity results for weak solutions in the class of the nonweighted spaces @xmath779 and @xmath780 .
we assume that @xmath9 is a polyhedron with faces @xmath10 , @xmath781 , and edges @xmath61 , @xmath80 .
the angle at the edge @xmath61 is denoted by @xmath782 . for the sake of simplicity ,
we restrict ourselves to the case @xmath62 and to homogeneous boundary conditions @xmath783 analogous results are valid for inhomogeneous boundary conditions provided the boundary data satisfy certain compatibility conditions on the edges .
note that there are the following equalities @xmath784 _ the dirichlet problem_. for arbitrary @xmath785 , there exists a solution @xmath786 of the dirichlet problem @xmath787 ( see e.g. ( * ? ? ?
* th.iv.2.1 ) ) .
this solution is unique for sufficiently small @xmath22 .
the regularity results established below are based on the following properties of the operator pencils @xmath728 ( see @xcite or ( * ? ? ?
* ch.5 ) ) . *
the strip @xmath788 is free of eigenvalues of the pencils @xmath461 . *
if the cone @xmath94 is contained in a half - space , then the strip @xmath789 contains only the eigenvalue @xmath431 of the pencil @xmath461 .
this eigenvalue has only the eigenvector @xmath790 , @xmath791 , and no generalized eigenvectors . *
the eigenvalues of the pencil @xmath461 in the strip @xmath789 are real and monotonous with respect to the cone @xmath94 . moreover , the eigenvalues for a circular cone are solutions of a certain transcendental equation ( see @xcite or ( * ? ? ?
* section 5.6 ) ) . the numbers @xmath768 can be easily calculated . in the case
@xmath792 , we have @xmath793 .
if @xmath794 , then @xmath768 is the smallest positive solution of the equation @xmath795 note that @xmath796 for every @xmath797 , @xmath798 if @xmath799 , @xmath800 if @xmath801 , and @xmath802 if @xmath803 . using these facts together with theorems [ dt1 ] and [ dt2 ] , we obtain the following assertions . *
if @xmath804 , @xmath805 , @xmath336 , then @xmath806 . if the polyhedron @xmath9 is convex , then this assertion is true for all @xmath807 . * if @xmath808 , @xmath809 , then @xmath810 . if @xmath811 for @xmath80 , then this result is true for @xmath594 . if @xmath9 is convex , then this result is valid for @xmath812 .
if , moreover , the angles at the edges are less than @xmath813 , then the result holds even for @xmath814 .
furthermore , the following assertion is valid . * if @xmath9 is convex , @xmath387 , and @xmath815 is sufficiently small ( such that @xmath816 and there are no eigenvalues of the pencils @xmath817 in the strip @xmath818 ) , then @xmath819 .
we prove the last result .
let @xmath31 be a positive number , @xmath820 .
since @xmath821 , it follows from theorem [ et2 ] that @xmath822 if @xmath823 .
in particular , we have @xmath824 .
this implies @xmath825 .
since @xmath826 , it follows that @xmath827 .
thus , @xmath29 satisfies the stokes system @xmath828 where @xmath829 .
hence by ( * ? ? ?
* th.4.3 ) , the solution @xmath29 admits the decomposition @xmath830 in a neighborhood of the vertex @xmath391 , where @xmath831 .
this is true for every vertex @xmath391 , @xmath832 .
consequently , @xmath833 .
+ for special domains , it is possible to obtain precise regularity results .
let for example @xmath9 have the form of steps as in the first two pictures below with angles @xmath834 or @xmath66 at every edge or the form of two beams , where one lies on the other as in the third picture .
note that the third polyhedron is not lipschitz .
the greatest edge angle is @xmath66 , and we obtain @xmath835 .
moreover for every vertex , there exists a circular cone with the same vertex and aperture @xmath66 which contains the polyhedron . the left polyhedron is even contained in a half - space bounded by a plane through an arbitrary of the vertices .
consequently , the smallest positive eigenvalue of the pencils @xmath461 does not exceed the first eigenvalue for a circular cone with vertex @xmath66 .
a numerical calculation shows that this eigenvalue is greater than @xmath836 .
this means that for @xmath53 and @xmath54 , the condition ( iii ) in theorems [ dt1 ] and [ dt2 ] is stronger than the condition ( ii ) in the same theorems .
thus , we obtain * @xmath837 if @xmath838 , @xmath839 , * @xmath840 if @xmath841 , @xmath842 . here the condition on @xmath28 is sharp .
we give some comments concerning the examples in the introduction ( the flow outside a regular polyhedron @xmath37 ) . by theorem [ dt2 ] , the regularity result @xmath843 in an arbitrary bounded subdomain of the complement of @xmath37 holds if @xmath844 and there are no eigenvalues of the pencils @xmath461 in the strip @xmath845 .
here , @xmath768 is the smallest positive solution of the equation ( [ eqmu ] ) , where @xmath846 is the edge angle in the exterior of @xmath37 , @xmath847 is equal to @xmath848 if @xmath37 is a regular tetrahedron or octahedron , @xmath849 if @xmath37 is a cube , @xmath850 if @xmath37 is a regular dodecahedron , and @xmath851 if @xmath37 is a regular icosahedron .
the smallest positive solutions of ( [ eqmu ] ) are @xmath852 for the regular tetrahedron , @xmath853 for a cube , @xmath854 for the regular octahedron , @xmath855 for the regular dodecahedron , and @xmath856 for the regular icosahedron . in the case of a regular tetrahedron , cube , octahedron or dodecahedron
, the inequality @xmath844 implies @xmath857 .
then the absence of eigenvalues of the pencil @xmath461 in the strip @xmath845 follows from ( * ? ? ?
* th.5.5.6 ) .
the exterior of a regular icosahedron is contained in a right circular cone with aperture less than @xmath858 .
numerical results for right circular cones together with the monotonicity of the eigenvalues of the pencils @xmath461 in the interval @xmath859 show that also for this polyhedron , the strip @xmath860 is free of eigenvalues of the pencil @xmath461 .
thus , the above mentioned regularity result holds for all @xmath844 . this inequality can not be weakened .
+ _ the neumann problem for the navier - stokes system_. we consider a weak solution @xmath861 of the neumann problem @xmath862 for this problem it is known that the strip @xmath863 contains only the eigenvalues @xmath864 and @xmath431 of the operator pencils @xmath461 ( see ( * ? ? ? * th.6.3.2 ) ) if @xmath9 is a lipschitz polyhedron .
the numbers @xmath768 are the same as for the dirichlet problem .
therefore , the following assertions are valid . * if @xmath804 , @xmath336 , @xmath865 , then @xmath866 . * if @xmath867 , @xmath809 , then @xmath868.if the angles @xmath782 are less than @xmath869 , then this result is true for @xmath870 . _ the mixed problem with dirichlet and neumann boundary conditions_. we assume that on each face @xmath10 either the dirichlet condition @xmath4 or the neumann condition @xmath871 is given .
if on both adjoining faces of the edge @xmath61 the same boundary conditions are given , then @xmath872 .
if on one of the adjoining faces the dirichlet condition and on the other face the neumann condition is given , then @xmath873 .
this implies the following result . *
if @xmath867 , @xmath874 , then every weak solution @xmath29 belongs to @xmath875 .
_ the mixed problem with boundary conditions _ ( i)(iii ) .
let @xmath876 be a weak solution of the problem ( [ ns1 ] ) , ( [ ns2 ] ) , where @xmath62 , @xmath877 , @xmath878 for @xmath79 , and @xmath879 for all @xmath201 ( i.e. , the neumann condition does not appear in the boundary conditions ) .
we assume that the dirichlet condition is given on at least one of the adjoining faces of every edge .
then , by ( * ? ? ?
* th.6.1.5 ) , the strip @xmath880 is free of eigenvalues of the pencils @xmath461 .
furthermore , we have @xmath881 if the dirichlet condition is given on both adjoining faces of the edge @xmath61 .
for the other indices @xmath201 , we have @xmath882 and @xmath883 if @xmath884 . * if @xmath804 , @xmath885 , then @xmath886 .
suppose that @xmath887 if the boundary condition ( ii ) or ( iii ) is given on one of the adjoining faces of the edge @xmath61 . then this result is even true for @xmath888 .
* if @xmath867 , @xmath874 , then @xmath868 .
suppose that @xmath889 if the dirichlet condition is given on both adjoining faces of @xmath61 , @xmath890 if the boundary condition ( ii ) is given on one of the adjoining faces of @xmath61 , and @xmath891 if the boundary condition ( iii ) is given on one of the adjoining faces of @xmath61 . then the last result is true for @xmath892 .
note that in the last case , we have @xmath893 for @xmath80 . finally , we assume that the homogeneous dirichlet condition @xmath4 is given on the faces @xmath894 , while the homogeneous boundary condition ( iii ) is given on @xmath895 .
let @xmath445 be the set of all @xmath201 such that @xmath896 and @xmath897 we suppose that the polyhedron @xmath9 is convex and @xmath898 for @xmath899 .
then @xmath800 for all @xmath201 , and the strip @xmath900 contains only the simple eigenvalue @xmath431 of the pencils @xmath461 ( see ( * ? ? ?
* th.6.2.7 ) ) . if @xmath901 for @xmath902 and @xmath903 for @xmath904 , then even @xmath905 .
this implies the following result .
* let @xmath867 , @xmath812
. then any weak solution belongs to @xmath875 . if @xmath901 for @xmath902 and @xmath903 for @xmath904 , then the result holds even for @xmath814 .
furthermore analogously to the dirichlet problem , the following assertion holds .
* let @xmath906 .
then for sufficiently small @xmath815 , we have @xmath907 .
in section 1 we proved the existence of weak solutions of the boundary value problem in @xmath910 . using the regularity result of theorem [ dt1 ] ,
we obtain also the existence of weak solutions in @xmath911 for sufficiently small @xmath807 . in this section
, we will prove that weak solutions exist also in the space @xmath908 with @xmath909 provided the norms of the right - hand sides of ( [ a1 ] ) , ( [ a2 ] ) in the corresponding sobolev spaces are sufficiently small . throughout this section ,
we suppose that the dirichlet condition is given on at least one of the adjoining faces of every edge @xmath61 .
let @xmath286 be the same polyhedral cone as in section 2.1 .
we consider weak solutions @xmath912 of the linear stokes system in @xmath286 with boundary conditions ( i)(iv ) on the faces @xmath10 .
this means that @xmath29 satisfies the equations @xmath913 here @xmath914 denotes the bilinear form ( [ a3 ] ) , where @xmath9 has to be replaced by @xmath286 .
we define the space @xmath915 as the set of all linear and continuous functionals on the space @xmath916 , where @xmath336 .
furthermore , the pencils @xmath767 for the edges @xmath61 and @xmath917 for the vertex of the cone @xmath286 are defined as in section 3.1 .
if the dirichlet condition ( i ) is given on at least one of the adjoining faces of the edge @xmath61 , then @xmath864 is not an eigenvalue of the pencil @xmath767 .
the following lemma is proved in @xcite under the condition @xmath918 . using the sharper estimates of green s matrix given in @xcite for the case
when @xmath864 is not an eigenvalue of the pencils @xmath767 , this theorem can be proved in the same way if @xmath919 for @xmath920 [ zl1 ] suppose that @xmath921 , @xmath922 , @xmath923 , there are no eigenvalues of the pencil @xmath917 on the line @xmath924 , and the components of @xmath59 satisfy the inequalities _ ( [ z1])_. then there exists a unique solution @xmath925 of problem _
( [ stokesk1 ] ) , ( [ stokesk2])_. [ zl2 ] _ 1 ) _ suppose that in addition to the assumptions of lemma _ [ zl1 ] _ , we have @xmath926 , @xmath927 , @xmath928 , where @xmath929 for @xmath184 and @xmath930 is such that there are no eigenvalues of the pencil @xmath917 in the closed strip between the lines @xmath931 and @xmath932 . then the solution @xmath925 belongs to @xmath933 .
_ 2 ) _ suppose that in addition to the assumptions of lemma _
[ zl1 ] _ , @xmath934 , @xmath935 , and the functional @xmath111 has the form @xmath936 with vector function @xmath937 , @xmath938 , where @xmath930 is such that there are no eigenvalues of the pencil @xmath917 in the closed strip between the lines @xmath939 and @xmath940 , and the components of @xmath567 satisfy the inequalities @xmath941 .
then the solution @xmath942 belongs to @xmath943 .
we consider the operator @xmath944 of problem ( [ stokes1 ] ) , ( [ stokes2 ] ) and denote this operator by @xmath945 . here
again @xmath946 is defined as the dual space of @xmath947 .
we show that this operator is fredholm if there are no eigenvalues of the pencil @xmath461 on the line @xmath948 , @xmath339 , and the components of @xmath59 satisfy the inequalities @xmath949 for @xmath80 . for this end
, we construct a left and right regularizer for the operator @xmath945 .
[ zl3 ] let @xmath434 be a sufficiently small open subset of @xmath9 and let @xmath950 be a smooth function with support in @xmath951 .
suppose that there are no eigenvalues of the pencil @xmath461 on the line @xmath952 , @xmath339 , and the components of @xmath59 satisfy _ ( [ z3])_. then there exists an operator @xmath953 continuously mapping the space of all @xmath954 with support in @xmath951 onto @xmath955 such that @xmath956 for all @xmath957 with support in @xmath958 and @xmath959 for all @xmath29 with support in @xmath958 .
p r o o f. suppose first that @xmath958 contains the vertex @xmath960 of @xmath9 .
then there exists a diffeomorphism @xmath435 mapping @xmath434 onto a subset @xmath961 of a polyhedral cone @xmath286 with vertex at the origin such that @xmath962 and the jacobian matrix @xmath736 coincides with the identity matrix @xmath445 at @xmath960 .
we assume that the supports of @xmath2 and @xmath3 are contained in @xmath958
. then the coordinate change @xmath733 transforms ( [ stokes1 ] ) , ( [ stokes2 ] ) into @xmath963 where @xmath964 , @xmath965 , @xmath966 are the faces of @xmath286 , @xmath967 is a first order differential operator of the form @xmath968 and @xmath969 is a bilinear form having the representation @xmath970 here @xmath971 and @xmath972 are quadratic matrices such that @xmath973 and @xmath974 let @xmath739 be an infinitely differentiable cut - off function on @xmath975 equal to 1 in @xmath976 and to zero in @xmath977 . for arbitrary positive @xmath978 , we define @xmath979 , moreover , we put @xmath980 and @xmath981 for @xmath982 .
we consider the operator @xmath983 defined by @xmath984 where @xmath985
we denote the operator ( [ z4 ] ) by @xmath986 . according to lemma [ zl1 ]
, the operator @xmath987 is an isomorhism .
since the norm of @xmath988 is small for small @xmath978 , the operator @xmath986 is an isomorphism if @xmath989 and @xmath990 is sufficiently small .
we may assume that @xmath991 on @xmath961 for @xmath992 .
then problem ( [ st1 ] ) , ( [ st2 ] ) can be written as @xmath993 if the supports of @xmath994 and @xmath995 are contained in @xmath996 .
let @xmath997 outside @xmath434 , let @xmath29 be continuously extended to a vector function from @xmath955 .
the so defined mapping @xmath998 is denoted by @xmath953 and has the desired properties .
suppose now that @xmath958 contains an edge point @xmath999 but no points of other edges and no vertices of @xmath9
. then again there exists a diffeomorphism mapping @xmath434 onto a subset of a cone @xmath286 .
we assume that the point @xmath1000 lies on the edge @xmath1001 of @xmath286 and coincides with the origin ( in contrast to the first part of the proof , the vertex of the cone is not the origin ) .
let @xmath1002 be the same operator as above .
then there exist a number @xmath1003 and a tuple @xmath567 , @xmath1004 , such that @xmath987 and for sufficiently small @xmath978 also @xmath986 are isomorphisms @xmath1005 since @xmath958 does not contain points of the edges @xmath61 , @xmath1006 , the vector function ( [ z5 ] ) can be continuously extended to a vector function from @xmath1007 .
the so defined mapping @xmath1008 defines the desired operator @xmath1009 .
analogously , the lemma can be proved for the case when @xmath958 contains no edge points of @xmath9 .
@xmath162 [ zr1 ] _ suppose that @xmath1010 and there are no eigenvalues of the pencil @xmath728 in the closed strip between the lines @xmath948 and @xmath1011 .
then it follows from lemma [ zl2 ] that the operator @xmath953 constructed in the proof of lemma [ zl3 ] continuously maps the subspace of all @xmath957 , where @xmath1012 and the functional @xmath1013 has the form @xmath1014 into @xmath1015 .
_ [ zl4 ] suppose that there are no eigenvalues of the pencil @xmath728 on the line @xmath467 , @xmath339 , and the components of @xmath59 satisfy _ ( [ z3])_. then there exists a continuous operator @xmath1016 such that @xmath1017 and @xmath1018 are compact operators in @xmath1019 and @xmath1020 , respectively .
p r o o f. for the sake of brevity , we write @xmath1021 instead of @xmath945 .
let @xmath1022 be a sufficiently fine open covering of @xmath9 , and let @xmath1023 , @xmath1024 be infinitely differentiable functions such that @xmath1025 for every @xmath117 , there exists an operator @xmath1026 having the properties of lemma [ zl2 ] for @xmath1027 .
we consider the operator @xmath953 defined by @xmath1028 obviously , @xmath1029\ , ( u , p)\big ) = ( u , p ) - \sum_j \varphi_j\ , { \cal r}_j\ , [ { \cal a},\psi_j]\ , ( u , p),\ ] ] where @xmath1030 $ ] is the commutator of @xmath1021 and @xmath1024 . here , the mapping @xmath1031\ , ( u , p)$ ] is continuous from @xmath1019 into the set of all @xmath957 , where @xmath1032 , @xmath1033 , and @xmath111 is a functional of the form @xmath1034 with arbitrary @xmath1035 , @xmath1036 .
we can choose @xmath930 and @xmath567 such that @xmath1037 for @xmath339 , @xmath1038 for @xmath80 , and @xmath1039 satisfy the conditions of remark [ zr1 ] with @xmath1040 .
then the mapping @xmath1041\ , ( u , p)$ ] is continuous from @xmath1019 into @xmath1042 . since the last space is compactly imbedded into @xmath1019 for @xmath1043 , @xmath1044 ( cf .
* le.6.2.1 ) ) , it follows that @xmath1045 is compact .
analogously , the compactness of @xmath1046 holds .
@xmath162 + from lemma [ zl4 ]
it follows that the operator @xmath1047 is fredholm under the conditions of this lemma . using theorem [ stokest1 ] and lemma [ zl2 ] , we can prove the following theorem on the existence of unique solutions .
[ zt1 ] let @xmath1013 , @xmath1048 , and @xmath1049 .
we suppose that the components of @xmath59 satisfy condition _ ( [ z1 ] ) _ and that there are no eigenvalues of the pencils @xmath1050 in the closed strip between the lines @xmath1051 and @xmath1052 . in the case when @xmath127 for all @xmath117 , we assume in addition that @xmath140 and @xmath141 satisfy condition _ ( [ 1stokest1])_. then there exists a solution @xmath1053 of problem _
( [ stokes1 ] ) , ( [ stokes2])_. ( here @xmath1054 has to be replaced by the set of all @xmath1055 satisfying @xmath1056 on @xmath10 . )
the vector function @xmath2 is unique , @xmath3 is unique if @xmath1057 for at least one @xmath117 , @xmath3 is unique up to a constant if @xmath1058 for all @xmath117 .
p r o o f. first note that in the case when @xmath1058 for all @xmath117 , the spectra both of @xmath461 and @xmath767 contain the eigenvalue @xmath431 .
an eigenvector corresponding to this eigenvalue is @xmath1059 .
furthermore , the spectra of the pencils @xmath461 contain the eigenvalue @xmath1060 .
therefore from the conditions of the lemma on @xmath58 and @xmath59 it follows in particular that @xmath1061 and @xmath1062
. then @xmath1063 . in particular
, any constant belongs to the space @xmath355 , and the integrals in ( [ 1stokest1 ] ) exist for @xmath1064 , @xmath1065 .
we prove the uniqueness of @xmath2 and @xmath3 .
let @xmath1066 be a solution of problem ( [ stokes1 ] ) , ( [ stokes2 ] ) with @xmath1067 , @xmath62 , @xmath877 .
then according to lemma [ zl2 ] , we have @xmath1068 and @xmath1069 . from theorem [ stokest1 ]
it follows that @xmath4 and @xmath3 is constant , @xmath1070 if @xmath1071 for at least one @xmath117 .
we prove the existence of solutions for the case when @xmath1072 for all @xmath117 .
due to lemma [ zl4 ] and the uniqueness of the solution , every solution @xmath1066 of problem ( [ stokes1 ] ) , ( [ stokes2 ] ) , @xmath1073 , satisfies the inequality @xmath1074 with a constant @xmath129 independent of @xmath2 and @xmath3 .
let @xmath1075 , @xmath1048 and @xmath1049 satisfying ( [ 1stokest1 ] ) be given .
then there exist sequences @xmath1076 , @xmath1077 and @xmath1078 converging to @xmath111 , @xmath140 , and @xmath141 , respectively . from ( [ 1stokest1 ] )
it follows that the sequences of the numbers @xmath1079 converges to zero .
therefore , the sequence of the functions @xmath1080 converges also to @xmath140 .
moreover , @xmath1081 and @xmath1082 satisfy condition ( [ 1stokest1 ] ) . by theorem [ stokest1 ] ,
there exist solutions @xmath1083 of problem ( [ stokes1 ] ) , ( [ stokes2 ] ) with right - hand sides @xmath1084 , @xmath1081 and @xmath1082 , @xmath1085 . from lemma [ zl2 ] and from the imbedding @xmath1086 with arbitrary positive @xmath31
it follows that @xmath1087 . due to ( [ 1zt1 ] )
, the vector functions @xmath1088 form a cauchy sequence in @xmath1019 .
the limit of this sequence solves problem ( [ stokes1 ] ) , ( [ stokes2 ] ) .
the proof of the existence of solutions in the case when @xmath1089 for at least one @xmath117 proceeds analogously .
@xmath162 we consider the nonlinear problem ( [ ns1 ] ) , ( [ ns2 ] ) . for the proof of the existence of solutions in @xmath1019 , we have to have to show inequalities analogous to ( [ q1 ] ) , ( [ q2 ] ) for the operator @xmath1090
p r o o f. let @xmath742 be a real number such that @xmath1096 for example , we can put @xmath1097 . then according to lemma [ al8 ] , the space @xmath1098 is continuously imbedded into @xmath1099 , where @xmath1100 , @xmath1101 .
let @xmath1102 . from the conditions on @xmath742 it follows that @xmath1103 . by hlders s inequality
, we have @xmath1104 for arbitrary @xmath1095 . using the continuity of the imbedding @xmath1105 which follows from corollary [ ac1 ] , we obtain the desired inequality . @xmath162 + as a consequence of lemma [ zl5 ] , the following inequalities hold for arbitrary @xmath1106 : @xmath1107 using these estimates , the following theorem can be proved analogously to theorem [ stokest1a ] .
[ zt2 ] let the conditions of theorem _ [ zt1 ] _ on @xmath111 , @xmath140 @xmath141 , @xmath58 and @xmath59 be satisfied .
in addition , we assume that @xmath1091 , @xmath1108 , @xmath1093 and that @xmath1109 is sufficiently small
. then there exists a solution @xmath1110 of problem _ ( [ a1 ] ) , ( [ a2 ] ) _ , where @xmath1054 has to be replaced by the set of all @xmath1111 such that @xmath1112 on @xmath10 , @xmath79 .
the function @xmath2 is unique on the set of all functions with norm less than a certain positive @xmath31 , @xmath3 is unique if @xmath126 for at least one @xmath117 , otherwise @xmath3 is unique up to a constant .
p r o o f. let @xmath1113 . by theorem [ zt1 ] , the norm of @xmath1114 can be assumed to be small .
obviously , @xmath29 is a solution of problem ( [ a1 ] ) , ( [ a2 ] ) if and only if @xmath1115 is a solution of the problem @xmath1116 this problem can be written as @xmath1117 where @xmath130 is the inverse to the operator @xmath1118 considered in the preceding subsection . due to ( [ q4 ] ) ,
the operator @xmath1119 is contractive on the set of all @xmath1120 with norm @xmath1121 if @xmath31 and the norm of @xmath1114 are sufficiently small .
hence this operator has a fixed point .
this proves the theorem .
@xmath162 + finally , we give a result in nonweighted sobolev spaces which follows immediately from the last theorem . *
_ let @xmath9 be a polyhedron .
we assume that on every face @xmath10 , one of the boundary conditions _ ( i)(iii ) _ is given , that the dirichlet condition is given on at least one of the adjoining faces of every edge @xmath61 , and that @xmath1122 if the boundary conditions _ ( ii ) _ or _ ( iii ) _ are given on one of the adjoining faces of @xmath61 . here , @xmath782 denotes again the angle at the edge @xmath61 .
then the problem @xmath1123 with @xmath1124 , @xmath1125 , has a solution @xmath1126 if the norm of @xmath111 is sufficiently small . _ for the proof it suffices to note that the strip @xmath1127 does not contain eigenvalues of the pencils @xmath461 and that @xmath1128 for @xmath80 . therefore , the conditions of theorem [ zt2 ] are satisfied for @xmath53 , @xmath54 , @xmath65 . , _ the first boundary value problem for classical equations of mathematical physics in domains with piecewise smooth boundaries _ , part 1 : zeitschr
* 2 * ( 4 ) 1983 , 335 - 359 , part 2 : zeitschr .
* 2 * ( 6 ) 1983 , 523 - 551 .
orlt , m. , sndig , a .- m . , _ regularity of viscous navier - stokes flows in nonsmooth domains _ , boundary value problems and integral equations in nonsmooth domains , lecture notes in pure and appl
* 167 * ( 1993 ) 185 - 201 .
solonnikov , v. a. , _ solvability of a three - dimensional boundary value problem with a free surface for the stationary navier - stokes system _ , partial differential equations , banach center publications , vol . 10 , warsaw 1983 , 361 - 403 . | mixed boundary value problems for the navier - stokes system in a polyhedral domain are considered .
different boundary conditions ( in particular , dirichlet , neumann , slip conditions ) are prescribed on the faces of the polyhedron .
the authors obtain regularity results for weak solutions in weighted ( and non - weighted ) @xmath0 sobolev and hlder spaces .
+ keywords : navier - stokes system , nonsmooth domains + msc ( 1991 ) : 35j25 , 35j55 , 35q30 |
shor s algorithms for factoring and discrete logarithms @xcite , as well as grover s search algorithm @xcite promise substantial quantum speed - up for certain important problems .
another exciting application area for quantum computers is simulation of physical systems , including molecules , strongly correlated spin systems , nuclei , and qcd .
the majority of these problems admit a natural description using continuous variables , such as particle coordinates or field modes .
( `` mode '' may refer to the field strength at a grid point or to a fourier component , depending on which representation is chosen .
) however , we want to perform the simulation using a _ digital _ quantum computer , i.e. , one composed of qubits . in this paper , we address some general issues pertaining to the discretization of continuous variables and round - off errors in the quantum case .
it is worth noting that continuous variables play an important role in algebraic number theory ( for example , consider the use of logarithmic coordinates in the proof of dirichlet s unit theorem @xcite . )
thus , we anticipate that discretization techniques might also be helpful in constructing new quantum algorithms for number - theoretic problems .
to see why the discretization is not entirely trivial , consider a quantum particle characterized by a single coordinate @xmath0 . to represent it in a quantum computer , we approximate @xmath0 by @xmath1 , where @xmath2 is a ( binary - encoded ) integer and @xmath3 is some small constant ( the discretization grid spacing ) . to simplify the notation , we may rescale @xmath0 so that @xmath4 ; thus , we approximate real numbers by integers .
now , suppose we want to change this standard grid to a different one , which is equivalent to stretching or squeezing the wavefunction and putting it back on the same grid .
we call this transformation _ resampling_. if @xmath0 is mapped to @xmath5 , a continuous wavefunction would transform as follows : @xmath6 in the simplest case , @xmath5 = @xmath7 , so that @xmath8 and @xmath9 . naively , this appears to be implementable as a permutation of basis vectors .
however , the discrete version of @xmath10 , @xmath11 is not a permutation ! for @xmath12 the transformation is a stretching ; it maps integers to a subset of integers , leaving some gaps .
for @xmath13 it is a squeezing , therefore two distinct integers may be mapped to the same integer .
fortunately , transformation ( [ discrete_map ] ) is to be applied to a special class of wavefunction , namely , ones that vary at distances much larger than the grid spacing . in this case , a few lowest bits of @xmath0 are almost in the uniform superposition and thus can be easily removed or added again while maintaining quantum coherence . but adding or removing @xmath14 bits corresponds to stretching ( resp .
squeezing ) by factor @xmath15 .
it is easy to generalize this trick to any scaling factor @xmath16 , or to an arbitrary function @xmath10 .
multidimensional resampling is somewhat more challenging though .
our solution for this case uses multidimensional gaussian states .
of course , the preparation of such states begins with the one - dimensional case .
the transition to higher dimensions requires some care .
one may be tempted to say that a general gaussian is just a product of several one - dimensional gaussians .
however , the implementation of this approach involves the diagonalization of the covariance matrix and a linear change of variables , which in turn requires resampling . to avoid this problem
, we use a special linear transformation , namely shearing , and construct its discrete version explicitly . in the rest of the paper , we simply expand this outline .
the states we wish to prepare are discrete approximations to continuous gaussian functions . in this section
we examine these functions and the approximate forms of interest . in one dimension ,
the gaussian is just the normal distribution familiar from statistics : @xmath17 ( up to a normalization factor ) .
our algorithm is similar to the uniform superposition preparation technique @xcite and its generalization to efficiently integrable distributions @xcite .
however , the latter procedures begin with dissecting the distribution at @xmath18 and setting the highest bit of @xmath0 according to @xmath19 . in comparison
, we first set the _
lowest _ bit and proceed with generating a discrete gaussian on either even integers or odd integers .
this method is slightly more efficient since it does not involve the gauss error function , but rather , the jacobi theta function , which is easier to calculate . if we treat the gaussian wavefunction as continuous , it has form @xmath20 so that @xmath21 and @xmath22 if we consider it to be a function of an integer value , we have @xmath23 where the function @xmath10 is related to the third jacobi theta function ( @xmath24 ) and is defined as @xmath25 so that @xmath26 if @xmath27 , then @xmath28 is approximately equal to @xmath29 .
we now define the function @xmath30 note that this function is periodic with period @xmath15 .
this allows us to define the quantum state @xmath31 for @xmath32 , this state is very close to a gaussian with mean @xmath33 and standard deviation @xmath34 . a recursive description of the state , @xmath35 where @xmath36 leads almost immediately to an algorithm for preparing it .
note that the sum represented by @xmath28 converges , so we can approximate it with a finite number of terms and keep some number of digits of the answer . however , the convergence is poor when @xmath34 is large .
when @xmath34 is small ( which is the case at a late stage of our recursive algorithm ) , a few terms of the sum will dominate , and only these need be calculated and stored . for large @xmath34 we use @xmath28 = @xmath37 to avoid having to sum an exponentially large number of terms .
the algorithm recursively calls the following subroutine , using the input data @xmath38 and some number of supplementary qubits .
note that the @xmath39 qubits are indexed 0 to @xmath40 , where the 0-th qubit corresponds to the rightmost qubit in a binary representation . 1 .
* calculate and store @xmath41 ( see eqn .
@xmath42 ) and @xmath43 in a k - bit binary representation on supplementary qubits . *
apply the rotation operator r(@xmath41 ) to the 0-th qubit @xmath44\ ] ] using the bits of @xmath43 as described below . *
uncompute @xmath41 and @xmath43 .
2 . compute @xmath45 , where @xmath46=@xmath47 and @xmath48 = @xmath49 if the previously rotated qubit is @xmath50 and @xmath48 = @xmath51 if it is @xmath52 .
note that all pairs @xmath53 are stored in supplementary memory until a cleanup phase , see below .
3 . on the remaining @xmath40 qubits
prepare state @xmath54 .
note that for the last ( ( n-1)th ) qubit , we proceed only through step 1 ( c ) , as after this qubit is rotated we do not need another pair of parameters .
after this last qubit is rotated , right before termination , a last series of steps is performed in which all stored pairs @xmath53 are cleared . to do this ,
the machine erases @xmath55 by recalculating it using @xmath56 and the last output qubit , performing a second controlled xor to the region where @xmath55 is stored , thus clearing @xmath55 from memory .
the process is repeated on @xmath56 , @xmath57 and so on until the supplementary memory is cleared .
since quantum recursion can be counterintuitive and difficult to describe verbally , this algorithm is perhaps easier to understand when presented in circuit form .
the figure shows the gist of the algorithm as well as some of the workings of the supplementary qubits .
note that the double line across the top denotes the classical part of the computer , which in this scheme functions only to transfer the specifications of the gaussian to the quantum part at the end of the process to clear memory .
the single lines with forward slashes ( / ) overlying their left ends represent multiple qubits in the supplementary memory . note that each such line represents the same array at different stages of the computation .
we deviate slightly from conventional notation by introducing the downward arrow conditional , which represents a generalized form of controlled operation .
the downward arrow terminating on the xor symbol on the supplementary array directly above the @xmath58th qubit denotes that the previously used pair @xmath59 and the @xmath60th qubit ( which was just rotated ) are used to control the calculation of @xmath61 and @xmath62 .
note that classically , @xmath62 is a rotation angle equal to @xmath63 if the @xmath60-th qubit is @xmath50 and @xmath64 if the @xmath60-th qubit is @xmath52 .
however , since the previous qubit is always in a superposition state , @xmath62 is always entangled with the previous qubits .
the downward arrow terminating on the i - th rotation operator box indicates that the rotation angle is specified by the controlling qubits , specifically , the qubits in the supplementary containing @xmath62 ( which is shown directly to the right of the downward arrow ) .
this rotation is actually a series of standardized rotations , each controlled by a single qubit of the register storing @xmath62 , as described below .
are stored until the end of computation .
the multi - qubit wires feeding into the black box g are the working " supplementary memory that is inverted and cleared in each step.,title="fig:",width=529 ] are stored until the end of computation .
the multi - qubit wires feeding into the black box g are the working " supplementary memory that is inverted and cleared in each step.,title="fig:",width=529 ] due to finite computational resources , the rotation operator must be implemented approximately .
note that since @xmath41 is a rotation angle between 0 and @xmath65 , so @xmath66 and we can approximate in @xmath14-digits of binary notation that @xmath67 so that @xmath68 .
note that the @xmath69 are either 1 or 0 , as this is a binary representation .
the rotation operator is realized as a sequence of at most @xmath14 standard rotations as @xmath70 .
the precision of this operation on the entire array ( assuming the standard rotations are exact ) is thus @xmath71 and the size of the circuit required to execute the algorithm is polynomial in @xmath72 .
in @xmath73 dimensions , the general gaussian function is @xmath74 , where @xmath75 is a constant , @xmath76 is a real symmetric positive definite matrix ( @xmath77=@xmath76 , and @xmath78 @xmath79 ) , and @xmath80 is vector with s components , @xmath81 where @xmath82 denotes position in @xmath73 space and @xmath83 is a vector denoting the mean of the distribution .
we assume that the register ( call it @xmath84 ) contains @xmath85 qubits , where @xmath14 is some integer .
thus , we can break the array into @xmath73 subarrays ( index them @xmath86 for @xmath87 ) each containing @xmath14 qubits , and write @xmath84=@xmath88 , with multiplication representing array concatenation .
the state space corresponding to @xmath84 is of dimension @xmath89 .
the wavefunction we wish to prepare has form @xmath90 in this expressions @xmath91 = @xmath92 and @xmath93 , @xmath94 = @xmath95 $ ] , where negative integers are represented using two s complement . in the quantum computer , @xmath96 is the basis vector formed by concatenating all binary representations of the entries of @xmath97 .
for example , if @xmath98 , in binary representation , @xmath99 , so if we assume @xmath14=2 , @xmath97 = @xmath100 = @xmath101 . for
@xmath86 , the basis state @xmath102 corresponds to the integer @xmath103 along the @xmath58th coordinate axis in @xmath73 space .
thus , each subarray generates one coordinate axis of the space in question .
note that this space is not a state space in the system , rather , each integer point in the cube described corresponds to one dimension / basis vector of the system s state space .
the reason for the distinction will be clear momentarily .
our method of preparing the general @xmath73-dimensional gaussian relies on a special case of the problem .
note that if the matrix @xmath76 in @xmath104 is diagonal ( in which case , we call it @xmath105 ) , then @xmath106 factors into 1-dimensional gaussians as @xmath107 where @xmath108 is the @xmath58th diagonal element of @xmath105 and @xmath109 the @xmath58-th entry of @xmath80 .
thus in this special case , the state to prepare becomes ( with @xmath97=@xmath110 , where the @xmath111 are @xmath14-digit binary integers ) @xmath112 this state is just a product of 1-dimensional gaussians , which we have already shown how to prepare .
we can prepare this state by preparing @xmath39 simple gaussians on @xmath14 qubit subarrays .
the overall normalization constant ( here denoted by @xmath75 ) is accounted for by the normalization constants of the gaussians on subarrays .
some simple properties of linear algebra and a change of variables allow us to transform this easily prepared state into any arbitrary gaussian .
any real symmetric positive definite matrix @xmath76 can be decomposed as @xmath76 = @xmath113 for some @xmath105 and @xmath114 , where @xmath105 is diagonal and @xmath114 is upper triangular with 1 s on the diagonal ( in other words , a multidimensional shearing matrix ) .
thus , if we apply transformation @xmath115 , then @xmath116 the matrix @xmath114 can be decomposed as the product of shearing matrices with exactly one non - zero off - diagonal entry each .
note that @xmath114 acts numerically on the basis vectors in a classically reversible fashion ( provided relevant parameters are stored ) .
it is not a unitary operator and is not applied directly to the state space of the system .
rather , it maps a vector @xmath97 to a vector @xmath80 ( and the corresponding basis vector @xmath96 to @xmath117 ) as in the following example .
let @xmath114 be a square , 2 by 2 upper triangular matrix with 1 s the diagonal and a real number @xmath41 in the upper right entry and let @xmath97 = @xmath118 .
then , we see that @xmath119 @xmath120 thus , @xmath114 can act on basis vectors numerically by matrix multiplication with rounding , which , provided the specifications for @xmath114 are stored , is an efficiently reversible classical operation that can thus be carried out on a quantum computer . applying @xmath114 to a superposition of states results in a changes of variable of the kind described above , which turns our set of independent gaussians into the general gaussian , @xmath121
thus , we have an algorithm for preparing the general gaussian function in @xmath39 dimensions .
one application of the gaussian states described above is multidimensional resampling .
this is more clear after a description of resampling in one dimension .
the goal is to stretch or contract a wavefunction with respect to some function @xmath10 , see eqn .
( [ wfmap ] ) .
( we assume that @xmath10 is smooth , monotone , and its derivative is bounded from above and from below . )
such deformations are easily accomplished , provided that @xmath122 is a slowly varying function ( @xmath123 and @xmath124 are very close for all @xmath0 in the range of interest ) .
assume we have two qubit ensembles , @xmath76 and @xmath94 , each containing @xmath39 qubits .
the basis states of ensemble @xmath76 correspond to the @xmath0 axis , those of ensemble @xmath94 to the @xmath125 axis .
the desired original wavefunction @xmath122 is prepared on ensemble @xmath76 , while ensemble @xmath94 is prepared in state @xmath126 .
the state of the combined system is thus @xmath127 if we want to perform a resampling with respect to @xmath128 , then we prepare the uniform superposition state @xmath129 on ensemble b using an algorithm described in problem 9.4 of @xcite .
we describe how parameter @xmath130 is selected shortly . after this operation
, the whole ensemble is in state @xmath131 a simple reversible computation can efficiently transform this state into @xmath132 recall that we wish to prepare the state @xmath133 on an @xmath39-qubit ensemble .
note that if we started with that wavefunction prepared on ensemble @xmath94 with ensemble @xmath76 initialized to @xmath126 , we could carry out a transformation analogous to that above to yield the state @xmath134 if we think of @xmath76 and @xmath94 as generating coordinate axes @xmath0 and @xmath125 , then wavefunctions can be visualized as sets of points in the @xmath135 plane , specifically , as points occupying the vertices of the infinite grid with integer spacing . in this scheme ,
@xmath136 and @xmath137 appear as bands of width @xmath138 about the line @xmath139 .
the amplitudes of basis vectors in @xmath140 and @xmath141 agree on that line ; the only difference is that the amplitudes in @xmath140 are constant on vertical strips whereas the amplitudes in @xmath141 are constant on horizontal strips . if @xmath130 is much larger than the grid spacing but small enough relative to the grid length at which @xmath122 and @xmath142 vary , the amplitudes in intersecting horizontal and vertical strips are close .
thus we have the approximate equality @xmath143 .
this means that by choosing @xmath39 and @xmath130 appropriately and preparing the state @xmath136 , it is virtually equivalent to having prepared the state @xmath137 .
if we then apply ( to the combined ensemble ) the inverse of the computation we would have used to prepare @xmath137 , it is as if we prepared and then uncomputed @xmath137 , leaving the combined ensemble in ( or rather , arbitrarily close to ) the state such that the wavefunction on @xmath94 is @xmath144 and the wavefunction on @xmath76 is @xmath126 .
thus if we can prepare @xmath123 satisfying the necessary smoothness conditions , we can use resampling with supplementary qubits to transform the state to one arbitrarily close to @xmath144 . and @xmath136 . the grey region ( and the dotted region in the inset ) represents the basis states of the combined ensemble ab that have non - zero amplitudes . in state
@xmath136 the amplitudes on vertical strips are equal , in state @xmath137 amplitudes are constant on horizontal strips .
if n is much larger than the spacing between basis vectors but much smaller than the size of the grid , the amplitudes of basis vectors in a vertical strip will be roughly equal to those of the basis vectors of any horizontal strip that intersects it .
when this is the case , @xmath145 and resampling will work.,width=529 ] this particular transformation takes advantage of the fact that uniform superpositions on an interval are invariant under stretching or contracting transformation .
multidimensional resampling requires a family of states that is similarly invariant under a general linear transformation .
multidimensional gaussian states satisfy this condition and thus may be used for multidimensional resampling .
we illustrate with a simple example .
note that in the procedure above , if we prepare on ensemble @xmath94 the state @xmath146 with parameters @xmath34 and @xmath33 such that @xmath147 with @xmath148 and @xmath149 instead of the uniform superposition state @xmath150 , the procedure still works .
intuitively , we prepare a series of vertical , parallel gaussian - weighted strips centered somewhere above the @xmath0 axis and shift them up along the line @xmath125 = @xmath7 so that each strip is centered at the line .
the result is that the entire grid has non - zero amplitudes , but all amplitudes except those immediately about @xmath125 = @xmath7 are very , very small , so the result looks very similar to figure 2 and is amenable to resampling as described .
this research was funded by caltech s summer undergraduate research fellowship ( surf ) program .
w.w . would like to thank ning bao , erik winfree , and david renshaw for helpful comments and discussion .
plain p. w. shor .
algorithms for quantum computation : discrete logarithms and factoring . in proceedings ,
35th annual symposium on foundations of computer science , ieee press , los alamitos , ca , 1994 . l. k. grover .
quantum mechanics helps in searching for a needle in a haystack .
lett.,79(2):325 , 1997 .
k. ireland , m. rosen . a classical introduction to modern number theory . graduate texts in mathematics , v. 84 .
springer , 1998 ( 2nd edition ) .
a. yu kitaev , a. h. shen , m. n. vyalyi .
classical and quantum computation .
graduate studies in mathematics , v. 47 .
american mathematical society , 2002 .
l. k. grover , t. rudolph . creating superpositions that correspond to efficiently integrable probability distributions .
quant - ph/0208112 . | we present an algorithm that prepares multidimensional gaussian wavefunctions on qubit arrays and an application of such wavefunctions to multidimensional resampling , a technique useful in quantum digital simulation . |
the forthcoming ten petawatt laser systems , vulcan-10pw in the united kingdom ( 300 j , 30 fs ) and apollon in france ( 150 j , 15 fs ) , will pave the way to the exawatt eli project aiming to reach intensities in excess of @xmath1 .
this will lead to unexplored laser - matter scenarii , characterized by coupled plasma and quantum electrodynamics ( qed ) processes . in order to enhance these novel effects
, a favourable configuration relies upon the interaction of the laser pulse with counter - propagating , multi - gev electrons . at laser intensities
up to @xmath2 , the electron dynamics enters the so - called radiation - dominated regime characterized by strong production of high - energy photons via nonlinear compton scattering . interacting back with the laser field
, these photons can lead to efficient production of @xmath3 pairs ( via the breit - wheeler mechanism ) . the resulting dense pair plasma , beyond its fundamental interest ,
can potentially modify the overall laser - plasma interaction @xcite .
the importance of qed effects on a relativistic electron of mass @xmath4 , charge @xmath5 and energy @xmath6 , and on a photon of energy @xmath7 , both interacting with a monochromatic plane wave @xmath8 , depends on the respective lorentz - invariant quantum parameters : @xmath9^{1/2 } , & & \chi_{\gamma } = \frac{\gamma _ \gamma}{e_s } \left [ \left(\mathbf{e_{\perp } } + \mathbf{c}\times \mathbf{b}\right)^2 \right]^{1/2 } , \label{eq : etachi}\end{aligned}\ ] ] where @xmath10 is the schwinger field beyond which the electromagnetic field can no longer be treated classically .
the energy distribution of the production rate of photons is given by @xcite @xmath11 , \label{eq : dnsync_dtdgammae}\end{aligned}\ ] ] where @xmath12 is the fine structure constant , @xmath13 the compton length and @xmath14 $ ] .
for @xmath15 ( @xmath16 , @xmath17 ) , large numbers of photons are emitted , each taking away a small fraction of the electron energy .
the emission process can be therefore treated classically and modelled as a continuous damping reaction . by constrast , in the quantum regime , for @xmath18 ( @xmath19 )
, fewer photons are emitted , but their energies can be comparable to that of the electron , leading to strong recoils and discontinuous trajectories .
finally , for high - energy photons ( @xmath20 ) interacting with the laser field , the energy distribution of the probability to decay into @xmath0 pair is given by @xcite @xmath21 } .
\label{eq : dnpm_dtdgamma}\end{aligned}\ ] ] here , @xmath22 refers to the energy of the created electron .
particle - in - cell ( pic ) codes are efficient tools for simulating the kinetic evolution of plasmas subjected to intense electromagnetic waves @xcite .
the pic code calder has been enriched with physical models describing photon and pair generation . in the pic description ,
particles are represented by super - particles with given density , shape and momenta .
radiative and qed models are applied on the super - particles and we similarly define the notion of super - photons generated during the emission process . in the classical radiative regime ,
the model considered is based on sokolov s scheme for radiation damping @xcite : @xmath23 where @xmath24 is the lorentz force vector and @xmath25 .
the radiated power is calculated using the quantum equation ( [ eq : dnsync_dtdgammae ] ) : @xmath26 in the quantum radiative regime , the monte carlo ( mc ) description is well - suited to treating the stochastic character of the photon generation @xcite .
this is valid provided that the field is uniform and quasi - static during the emission process .
propagating electrons are assigned an optical depth @xmath27 , evolving with time according to the field and the electron energy , @xmath28 in the mc algorithm , a discrete photon is emitted when @xmath29 , where the total optical depth @xmath30 is sampled from @xmath31 with @xmath32 0,1]$ ] a uniform random number .
the photon quantum paramater @xmath33 is calculated by inverting @xmath34 where @xmath35 is the the photon cumulative distribution function and @xmath36 $ ] is a new random parameter .
the photon is emitted to the direction parallel to the electron velocity with wave vector @xmath37 ( since the angular spread is @xmath38 ) .
the electron momentum @xmath39 therefore becomes @xmath40 a similar mc scheme is applied to the super - photons to model pair creation using the probability function ( [ eq : dnpm_dtdgamma ] ) .
each photon is assigned an optical depth @xmath41 and a randomly sampled final one @xmath42.when @xmath43 , the super - photon is deleted and a @xmath3 pair is created at the current position . solving @xmath44 gives the newly - created electron quantum parameter @xmath45 , @xmath46 being randomly calculated in @xmath470,1\right]$ ] and @xmath48 the cumulative distribution function associated to the pair generation .
the positron quantum parameter is @xmath49 . as the mc radiative model leads to the generation of many low - energy photons unable to decay into pairs and prejudicial for the numerical efficiency ( most notably in configurations prone to pair cascading )
, we usually restrict the creation of super - photons to energies above @xmath50 .
moreover , we have worked out a coalescence scheme ensuring energy and momentum conservation on average .
as a test simulation , we consider the photon and pair generation by electrons of initial energy @xmath51 in a constant magnetic field @xmath52 .
the initial quantum parameter is equal to @xmath53 .
the initial synchrotron frequency and radius are respectively equal to @xmath54 and @xmath55 .
we define @xmath56 the energy distributions of the positrons , electrons and photons , respectively .
the time evolution of the latter can be described by the following set of equations @xcite @xmath57 ) and ( [ eq : spc_th2 ] ) ( top).,scaledwidth=85.0% ] ) and ( [ eq : spc_th2 ] ) ( top).,scaledwidth=87.0% ] the time evolution of the total energy , density and averaged quantum parameter is plotted in fig .
[ fig : time_diag_group ] for each species .
figure [ fig : spc_group ] displays the corresponding energy distributions . very good agreement is obtained between the numerical solutions of eqs . ( [ eq : spc_th1 ] ) and ( [ eq : spc_th2 ] ) and the mc - pic simulations .
the interaction starts in the quantum regime .
the firstly created photons have an energy close to initial electron energy , @xmath58 .
consequently , the radiated energy significantly grows during the first rotation at the expense of the electron energy ( fig .
[ fig : time_diag_group]a ) .
the average photon quantum parameter is close to @xmath59 ( fig [ fig : time_diag_group]c ) , which is favourable for pair production . in turn , newly generated particles radiate their energy away , hence leading to pair cascading ( fig .
[ fig : time_diag_group]b ) .
when entering the semi - classical regime , the production rate of positrons tends to stabilize since the average photon energy drops below the pair creation threshold ( @xmath60 ) .
radiation losses , originally responsible for pair creation , now only act as a cooling mechanism . note that in the quantum regime , the stochasticity , intrinsic to the qed processes stretches the electron energy spectrum , whereas in the semi - classical regime , radiation losses tend to shrink it back .
we have presented the implementation in a pic code of the photon and pair generation via the compton scattering and breit wheeler processes .
these physical models will be used in forthcoming papers to study a number of laser - plasma scenarii at ultra - relativistic intensities .
the authors acknowledge support by the french agence nationale de la recherche ( labex palm - anr-10-labx-39 ) and thanks r duclous , a debayle for interesting discussions .
the pic simulations were performed using hpc resources at tgcc / ccrt ( grant no . 2013 - 052707 ) .
99 g. a. mourou _ et al .
_ 2006 _ rev . mod .
phys . _ * 78 * 309 ; a. macchi _ et al . _ 2013 _ rev
phys . _ * 85 * 751 a. r. bell and j. g. kirk 2008 _ phys .
lett . _ * 101 * 200403 ; s. v. bulanov _ et al .
_ 2011 _ nimpra _ * 660 * 31 - 34 ; a. di piazza _ et al .
_ 2012 _ rev .
phys . _ * 84 * 1177 c. k. birdsall and a. b. langdon 1985 _ plasma physics via computer simulation _
( new york : mcgraw - hill ) ; e. lefebvre _ et al .
_ 2003 _ nucl . fusion _ * 43 * 629 - 633 a. i. nikishov and v. i. ritus 1964 _ jetp _ * 19 * 529 ; t. erber 1966 _ rev . mod .
phys . _ * 38 * 626 i. v. sokolov _ et al .
_ 2009 _ phys .
plasmas _ * 16 * 093115 r. duclous _ et al .
_ 2011 _ plas .
control fusion _ * 53 * 015009 ; n.v .
_ 2011 _ phys . rev .
. beams _ * 14 * 054401 i. v. sokolov 2010 _ phys
* 105 * 195005 ; j. g. kirk _ et al .
_ 2009 _ plas .
control fusion _ * 51 * 085008 ; n. neitz _ et al .
_ 2013 _ phys .
lett . _ * 111 * 054802 | next generation of ultra - intense laser facilities will lead to novel physical conditions ruled by collective and quantum electrodynamics effects , such as synchrotron - like emission of high - energy photons and @xmath0 pair generation . in view of the future experiments performed in this regime ,
the latter processes have been implemented into the particle - in - cell code calder . |
in this section we recall the reduction , due to bruce westbury @xcite , of the study of finite dimensional simple representations of the three string braid group @xmath47 to those of a particular quiver @xmath48 .
we will then identify the luna slices at specific @xmath48-representations to representation spaces of corresponding local quivers as introduced and studied in @xcite .
recall that the center of @xmath7 is infinite cyclic with generator @xmath49 and hence that the corresponding quotient group ( taking @xmath50 and @xmath51 ) @xmath52 is the free product of cyclic groups of order two and three and therefore isomorphic to the modular group @xmath2 . by schur
s lemma , @xmath53 acts via scalar multiplication with @xmath54 on any finite dimensional irreducible @xmath7-representation , hence it suffices to study the irreducible representations of the modular group @xmath37 . bruce westbury @xcite established the following connection between irreducible representations of @xmath37 and specific stable representations of the directed quiver @xmath48 : @xmath55[f-]{\scriptscriptstyle } } \\ { * + [ o][f-]{\scriptscriptstyle } } \ar[rrrru ] \ar[rrrrd ] \ar[rrrrddd ] & & & & \\ & & & & { * + [ o][f-]{\scriptscriptstyle } } \\ { * + [ o][f-]{\scriptscriptstyle } } \ar[rrrru ] \ar[rrrruuu ] \ar[rrrrd ] & & & & \\ & & & & { * + [ o][f-]{\scriptscriptstyle } } } \ ] ] for @xmath15 an @xmath1-dimensional representation of @xmath37 , decompose @xmath15 into eigenspaces with respect to the actions of @xmath56 and @xmath57 @xmath58 ( @xmath42 a primitive 3rd root of unity ) .
denote the dimensions of these eigenspaces by @xmath59 resp . @xmath60 and @xmath61 , then clearly @xmath62 . choose a vector - space basis for @xmath15 compatible with the decomposition @xmath63 and another basis of @xmath15 compatible with the decomposition @xmath64
, then the associated base - change matrix @xmath65 determines a @xmath48-representation @xmath66 of dimension vector @xmath67 @xmath68[f-]{\scriptscriptstyle x } } \\ { * + [ o][f-]{\scriptscriptstyle a } } \ar[rrrru]^(.3){b_{11 } } \ar[rrrrd]^(.3){b_{21 } } \ar[rrrrddd]_(.2){b_{31 } } & & & & \\ & & & & { * + [ o][f-]{\scriptscriptstyle y } } \\ { * + [ o][f-]{\scriptscriptstyle b } } \ar[rrrruuu]_(.7){b_{12 } } \ar[rrrru]_(.7){b_{22 } } \ar[rrrrd]_(.7){b_{23 } } & & & & \\ & & & & { * + [ o][f-]{\scriptscriptstyle z}}}~\qquad \text{with } \qquad b= \begin{bmatrix } b_{11 } & b_{12 } \\
b_{21 } & b_{22 } \\
b_{31 } & b_{32 } \end{bmatrix}\end{aligned}\ ] ] a @xmath48-representation @xmath26 of dimension vector @xmath6 is said to be @xmath69-stable , resp .
@xmath69-semi - stable if for every proper sub - representations @xmath70 , with dimension vector @xmath71 , we have that @xmath72 , resp
. @xmath73 .
@xmath15 is an @xmath1-dimensional simple @xmath37-representation if and only if the corresponding @xmath48-representation @xmath66 is @xmath69-stable .
moreover , @xmath74 as @xmath37-representations if and only if corresponding @xmath48-representations @xmath66 and @xmath75 are isomorphic as quiver - representations .
the affine git - quotient @xmath76 classifying isomorphism classes of @xmath1-dimensional semi - simple @xmath37-representations decomposes into a disjoint union of irreducible components @xmath77 one component for every dimension vector @xmath67 satisfying @xmath62 . if @xmath78 satisfies @xmath79 , then the component @xmath80 contains an open subset of simple representations if and only if @xmath81 . in this case
, the dimension of @xmath80 is equal to @xmath82 .
the remaining simple @xmath37-representations ( that is , those of dimension vector @xmath67 with @xmath83 ) are of dimension one or two
. there are @xmath18 one - dimensional simples ( of the abelianization @xmath84 ) corresponding to the @xmath48-representations @xmath85[f-]{\scriptscriptstyle 1 } } \\ { * + [ o][f-]{\scriptscriptstyle 1 } } \ar[rrrru]^1 & & & & \\ & & & & { * + [ o][f-]{\scriptscriptstyle 0 } } \\ { * + [ o][f-]{\scriptscriptstyle 0 } } & & & & \\ & & & & { * + [ o][f-]{\scriptscriptstyle 0 } } } \qquad s_2 = \xymatrix@=.1 cm { & & & & { * + [ o][f-]{\scriptscriptstyle 0 } } \\ { * + [ o][f-]{\scriptscriptstyle 0 } } & & & & \\ & & & & { * + [ o][f-]{\scriptscriptstyle 1 } } \\ { * + [ o][f-]{\scriptscriptstyle 1 } } \ar[rrrru]^1 & & & & \\ & & & & { * + [ o][f-]{\scriptscriptstyle 0 } } } \qquad s_3 = \xymatrix@=.1 cm { & & & & { * + [ o][f-]{\scriptscriptstyle 0 } } \\ { * + [ o][f-]{\scriptscriptstyle 1 } } \ar[rrrrddd]^1 & & & & \\ & & & & { * + [ o][f-]{\scriptscriptstyle 0 } } \\ { * + [ o][f-]{\scriptscriptstyle 0 } } & & & & \\ & & & & { * + [ o][f-]{\scriptscriptstyle 1}}}\ ] ] @xmath86[f-]{\scriptscriptstyle 1 } } \\ { * + [ o][f-]{\scriptscriptstyle 0 } } & & & & \\ & & & & { * + [ o][f-]{\scriptscriptstyle 0 } } \\ { * + [ o][f-]{\scriptscriptstyle 1 } } \ar[rrrruuu]^1 & & & & \\ & & & & { * + [ o][f-]{\scriptscriptstyle 0 } } } \qquad s_5 = \xymatrix@=.1 cm { & & & & { * + [ o][f-]{\scriptscriptstyle 0 } } \\ { * + [ o][f-]{\scriptscriptstyle 1 } } \ar[rrrrd]^1 & & & & \\ & & & & { * + [ o][f-]{\scriptscriptstyle 1 } } \\ { * + [ o][f-]{\scriptscriptstyle 0 } } & & & & \\ & & & & { * + [ o][f-]{\scriptscriptstyle 0 } } } \qquad s_6 = \xymatrix@=.1 cm { & & & & { * + [ o][f-]{\scriptscriptstyle 0 } } \\ { * + [ o][f-]{\scriptscriptstyle 0 } } & & & & \\ & & & & { * + [ o][f-]{\scriptscriptstyle 0 } } \\ { * + [ o][f-]{\scriptscriptstyle 1 } } \ar[rrrrd]^1 & & & & \\ & & & & { * + [ o][f-]{\scriptscriptstyle 1}}}\ ] ] and three one - parameter families of two - dimensional simple @xmath37-representations corresponding to the @xmath48-representations @xmath87[f-]{\scriptscriptstyle 0 } } \\ { * + [ o][f-]{\scriptscriptstyle 1 } } \ar[rrrrd]|(0.6){\lambda } \ar[rrrrddd]|(0.7){1 } & & & & \\ & & & & { * + [ o][f-]{\scriptscriptstyle 1 } } \\ { * + [ o][f-]{\scriptscriptstyle 1 } } \ar[rrrru]|(0.6){1 } \ar[rrrrd]|(0.6){1 } & & & & \\ & & & & { * + [ o][f-]{\scriptscriptstyle 1 } } } \quad t_2(\lambda ) = \xymatrix@=.1 cm { & & & & { * + [ o][f-]{\scriptscriptstyle 1 } } \\ { * + [ o][f-]{\scriptscriptstyle 1 } } \ar[rrrru]|(0.6){\lambda } \ar[rrrrddd]|(0.6){1 } & & & & \\ & & & & { * + [ o][f-]{\scriptscriptstyle 0 } } \\ { * + [ o][f-]{\scriptscriptstyle 1 } } \ar[rrrruuu]|(0.6){1 } \ar[rrrrd]|(0.6){1 } & & & & \\ & & & & { * + [ o][f-]{\scriptscriptstyle 1 } } } \quad t_3(\lambda ) = \xymatrix@=.1 cm { & & & & { * + [ o][f-]{\scriptscriptstyle 1 } } \\ { * + [ o][f-]{\scriptscriptstyle 1 } } \ar[rrrru]|(0.6){\lambda } \ar[rrrrd]|(0.6){1 } & & & & \\ & & & & { * + [ o][f-]{\scriptscriptstyle 1 } } \\ { * + [ o][f-]{\scriptscriptstyle 1 } } \ar[rrrru]|(0.6){1 } \ar[rrrruuu]|(0.7){1 } & & & & \\ & & & & { * + [ o][f-]{\scriptscriptstyle 0 } } } \ ] ] satisfying @xmath88 . with @xmath89
we will denote the set @xmath90 of all one - dimensional @xmath37-representations .
consider a finite set @xmath91 of simple @xmath37-representations and identify @xmath92 with the corresponding @xmath48-representation of dimension vector @xmath93 .
consider the semi -
simple @xmath37-representation @xmath94 the theory of luna slices allows us to describe the tale local structure of the component @xmath0 , where @xmath95 , in a neighborhood of the point corresponding to @xmath96 .
we will assume throughout that @xmath97 is the dimension vector of a @xmath69-stable representation , that is , that @xmath81 .
let @xmath98 be the @xmath99-orbit of @xmath96 in the representation space @xmath100 , then the normal space to the orbit @xmath101 is the extension space , see for example ( * ? ? ?
* ii.2.7 ) .
because @xmath102 we can identify the vectorspace @xmath103 to the representation space @xmath104 of the quiver @xmath105 on @xmath106 vertices @xmath107 ( vertex @xmath108 corresponding to the simple @xmath37-representation @xmath92 ) such that the number of directed arrows from vertex @xmath108 to vertex @xmath109 is equal to @xmath110[f-]{\scriptscriptstyle i } } \ar[r ] & { * + [ o][f-]{\scriptscriptstyle j } } } \ } = dim_{{\mathbb{c}}}~ext^1_{q_0}(v_i , v_j)\ ] ] and the dimension vector @xmath111 of @xmath105 is given by the multiplicities of the simple factors in @xmath96 , that is , @xmath112 . observe that the stabilizer subgroup of @xmath96 is equal to @xmath113 . the quiver @xmath105 is called the local quiver of @xmath96 , see for example @xcite or @xcite , and can be determined from the euler form of @xmath48 which is the bilinear map @xmath114 determined by the matrix @xmath115 for @xmath15 and @xmath26 @xmath48-representation of dimension vector @xmath6 and @xmath116 , we have @xmath117 recall that @xmath118 whenever @xmath15 and @xmath26 are @xmath69-stable quiver representations . from the luna slice theorem we obtain ,
see for example @xcite .
[ general ] let @xmath119 be a finite set of simple @xmath37-representations with corresponding @xmath48-dimension vectors @xmath93 . consider the semi - simple @xmath37-representation @xmath94 with @xmath48-dimension vector @xmath95 .
let @xmath105 be the local quiver described above and let @xmath120 be the @xmath105-dimension vector determined by the multiplicities .
then , the action map @xmath121 sending the class of @xmath122 in the associated fibre bundle to the representation @xmath123 where @xmath124 is the representation in the normal space to the orbit @xmath98 corresponding to the @xmath105-representation @xmath125 , is a @xmath126-equivariant tale map with a zariski dense image .
taking @xmath126 quotients on both sides , we obtain an tale action map @xmath127 with a zariski dense image .
in order to apply theorem [ general ] we need to consider a family @xmath128 of simple @xmath37-representations generating a semi - simple representation @xmath96 in every component @xmath0 , and , we need to make the action map explicit , that is , we need to identify representations of the quiver @xmath105 with representations in the normal space to the orbit @xmath98 .
we consider the set @xmath129 of all one - dimensional @xmath37-representations , using the notations as before .
consider the semi -
simple @xmath37-representation of dimension @xmath130 @xmath131 then @xmath96 corresponds to a @xmath69-semistable @xmath48-representation of dimension vector @xmath132 under the identification ( [ identification ] ) the @xmath133 matrix @xmath134 determined by @xmath96 consists of the following block - matrices @xmath135 containing themselves blocks of sizes @xmath136 for the appropriate @xmath137 and @xmath138 @xmath139 @xmath140 [ matrices ] with @xmath141 , the tale action map @xmath142 is induced by sending a representation in @xmath143 defined by the matrices @xmath144[f-]{\scriptscriptstyle a_1 } } \ar@/^/[ld]^{c_{16 } } \ar@/^/[rd]^{c_{12 } } & \\ { * + [ o][f-]{\scriptscriptstyle a_6 } } \ar@/^/[ru]^{c_{61 } } \ar@/^/[d]^{c_{65 } } & & { * + [ o][f-]{\scriptscriptstyle a_2 } } \ar@/^/[lu]^{c_{21 } } \ar@/^/[d]^{c_{23 } } \\ { * + [ o][f-]{\scriptscriptstyle a_5 } } \ar@/^/[u]^{c_{56 } } \ar@/^/[rd]^{c_{54 } } & & { * + [ o][f-]{\scriptscriptstyle a_3 } } \ar@/^/[u]^{c_{32 } } \ar@/^/[ld]^{c_{34 } } \\ & { * + [ o][f-]{\scriptscriptstyle a_4 } } \ar@/^/[lu]^{c_{45 } } \ar@/^/[ru]^{c_{43 } } & } \ ] ] to the representation in @xmath145 corresponding to the @xmath133 matrix ( via identification ( [ identification ] ) ) @xmath146 under this map , simple @xmath147-representations with invertible matrix @xmath148 are mapped to irreducible @xmath1-dimensional @xmath37-representations .
hence , if the coefficients in the matrices @xmath149 give a parametrization of ( an open set of ) the quotient variety @xmath150 , the @xmath1-dimensional representations of the three string braid group @xmath7 given by @xmath151 contain a zariski dense set of the simple @xmath7-representations in @xmath152 .
using the euler - form of the quiver @xmath48 and the @xmath48-dimension vectors of the representations @xmath153 one verifies that the quiver @xmath147 is the one given in the statement of the theorem . to compute the components of the tangent space in @xmath96 to the orbit ,
take @xmath154 as the set of matrices ( in block - matrices of sizes @xmath136 ) @xmath155 and hence the tangent space to the orbit is computed using the action of @xmath156 on the quiver - representations , giving for example for the @xmath157-arrow @xmath158 which is equal to @xmath159 and , similarly , the @xmath160-components of @xmath161 resp .
@xmath162 are calculated to be @xmath163 @xmath164 @xmath165 here the zero blocks correspond precisely to the matrices @xmath149 describing a representation in @xmath143 which can therefore be identified with the normal space in @xmath96 to the orbit @xmath98 . here
we use the inproduct on @xmath166 defined for all @xmath167 and @xmath168 @xmath169 hence , the representation @xmath170 is determined by the matrices @xmath171 @xmath172 @xmath173 to any @xmath48-representation @xmath66 of dimension vector @xmath78 , with invertible @xmath133 matrix @xmath148 corresponds , via the identifications given in the previous section , the @xmath1-dimensional representation of the modular group @xmath37 defined by @xmath174 as @xmath175 and @xmath176 it follows that @xmath177 and @xmath178 . therefore , when @xmath66 is @xmath69-stable it determines a @xmath179-family of @xmath1-dimensional representations of the three string braid group @xmath7 given by @xmath180 using the dimension vector @xmath181 and the matrices @xmath135 found before , we obtain the required @xmath7-representations .
the final statements follow from theorem [ general ] .
the method of proof indicates how one can make the action maps explicit for any given finite family of irreducible @xmath37-representations .
note that the condition for @xmath181 to be the dimension vector of a @xmath69-stable representation is equivalent to the condition on @xmath111 @xmath182 which is the condition for @xmath141 to be the dimension vector of a simple @xmath147-representation , by @xcite .
[ rationality ] for any @xmath48-dimension vector @xmath97 admitting a @xmath69-stable representation , that is satisfying @xmath81 there exist semi - simple representations @xmath183 @xmath184 such that for @xmath141 the quotient variety @xmath150 is rational . as a consequence ,
the tale action map @xmath185 given by theorem [ matrices ] determines a rational parametrization of a zariski dense subset of @xmath186 .
the condition on @xmath116 ensures that there are simple @xmath111-dimensional representations of @xmath147 , and hence that @xmath111 is a schur root . by a result of aidan schofield @xcite
this implies that the quotient variety @xmath150 is birational to the quotient variety of @xmath187-tuples of @xmath188 matrices under simultaneous conjugation , where @xmath189 by procesi s result ( * ? ? ?
iv.6.4 ) and the known rationality results ( see for example @xcite ) the result follows when we can find such an @xmath111 satisfying @xmath190 . for small dimensions one verifies this by hand , and , for larger dimensions having found a representation @xmath96 with @xmath191 , one can modify the multiplicities of the simple components by @xmath192 ( or a cyclic permutation of these ) to obtain another semi - simple representation @xmath193 lying in the same component , but having @xmath194 .
in this section we will determine for every irreducible component @xmath0 , with @xmath196 such that @xmath197 satisfying @xmath81 , a semi - simple @xmath37-representation @xmath184 contained in @xmath0 .
then , we will explicitly determine a rational family of representations in @xmath143 for @xmath198 .
as we are only interested in the corresponding @xmath7-representations we may assume that @xmath196 is such that @xmath199 and @xmath200 .
observe that in going from @xmath37-representations to @xmath7-representations we multiply with @xmath54 .
hence , there is a @xmath201-action on the dimension vectors of @xmath48-representations giving the same component of @xmath7-representations .
that is , we have to consider only one dimension vector from the @xmath201-orbit @xmath202 hence we may assume that @xmath199 and that @xmath203 . if we find a module @xmath204 and corresponding rational parametrization for @xmath196 by representations in @xmath143 , then we can mirror that description to obtain similar data for @xmath205 via @xmath144[f-]{\scriptscriptstyle a_1 } } \ar@/^/[ld]^{c_{16 } } \ar@/^/[rd]^{c_{12 } } & \\ { * + [ o][f-]{\scriptscriptstyle a_6 } } \ar@/^/[ru]^{c_{61 } } \ar@/^/[d]^{c_{65 } } & & { * + [ o][f-]{\scriptscriptstyle a_2 } } \ar@/^/[lu]^{c_{21 } } \ar@/^/[d]^{c_{23 } } \\ { * + [ o][f-]{\scriptscriptstyle a_5 } } \ar@/^/[u]^{c_{56 } } \ar@/^/[rd]^{c_{54 } } & & { * + [ o][f-]{\scriptscriptstyle a_3 } } \ar@/^/[u]^{c_{32 } } \ar@/^/[ld]^{c_{34 } } \\ & { * + [ o][f-]{\scriptscriptstyle a_4 } } \ar@/^/[lu]^{c_{45 } } \ar@/^/[ru]^{c_{43 } } & } ~\quad \leftrightarrow \quad \xymatrix@=1.1 cm { & { * + [ o][f-]{\scriptscriptstyle a_1 } } \ar@/^/[ld]^{c_{12 } } \ar@/^/[rd]^{c_{16 } } & \\ { * + [ o][f-]{\scriptscriptstyle a_2 } } \ar@/^/[ru]^{c_{21 } } \ar@/^/[d]^{c_{23 } } & & { * + [ o][f-]{\scriptscriptstyle a_6 } } \ar@/^/[lu]^{c_{61 } } \ar@/^/[d]^{c_{65 } } \\ { * + [ o][f-]{\scriptscriptstyle a_3 } } \ar@/^/[u]^{c_{32 } } \ar@/^/[rd]^{c_{34 } } & & { * + [ o][f-]{\scriptscriptstyle a_5 } } \ar@/^/[u]^{c_{56 } } \ar@/^/[ld]^{c_{54 } } \\ & { * + [ o][f-]{\scriptscriptstyle a_4 } } \ar@/^/[lu]^{c_{43 } } \ar@/^/[ru]^{c_{45 } } & } \ ] ] proving that we may indeed restrict to @xmath196 with @xmath199 and @xmath200 . [ lem1 ] the following representations provide a rational family determining a dense subset of the quotient varieties @xmath206 for these ( minimalistic ) dimension vectors @xmath6 .
@xmath207[f-]{\scriptscriptstyle 1 } } \ar@/^/[rr]^1 & & { * + [ o][f-]{\scriptscriptstyle 1 } } \ar@/^/[ll]^a}$ & ( b ) : ~$\xymatrix@=.5cm{{*+[o][f-]{\scriptscriptstyle 1 } } \ar@/^/[rr]^{\begin{bmatrix } 0 \\ 1 \end{bmatrix } } & & { * + [ o][f-]{\scriptscriptstyle 2 } } \ar@/^/[ll]^{\begin{bmatrix } 1 & a \end{bmatrix } } \ar@/^/[rr]^{\begin{bmatrix } b & c \end{bmatrix } } & & { * + [ o][f-]{\scriptscriptstyle 1 } } \ar@/^/[ll]^{\begin{bmatrix } 1 \\ 0 \end{bmatrix}}}$ \\ ( c ) : ~$\xymatrix@=.5cm{{*+[o][f-]{\scriptscriptstyle 1 } } \ar@/^/[rr]^{\begin{bmatrix } 0 \\ 1 \end{bmatrix } } & & { * + [ o][f-]{\scriptscriptstyle 2 } } \ar@/^/[ll]^{\begin{bmatrix } 1 & a \end{bmatrix } } \ar@/^/[rr]^{\begin{bmatrix } 1 & 0 \\ 0 & 1 \end{bmatrix } } & & { * + [ o][f-]{\scriptscriptstyle 2 } } \ar@/^/[ll]^{\begin{bmatrix } b & c \\ d & 1 \end{bmatrix}}}$ & ( d ) : ~$\xymatrix@=.5cm{{*+[o][f-]{\scriptscriptstyle 2 } } \ar@/^/[rr]^{\begin{bmatrix } a & 0 \\ 0 & b \end{bmatrix } } & & { * + [ o][f-]{\scriptscriptstyle 2 } } \ar@/^/[ll]^{\begin{bmatrix } 1 & 0 \\ 0 & 1 \end{bmatrix } } \ar@/^/[rr]^{\begin{bmatrix } 1 & 0 \\ 0 & 1 \end{bmatrix } } & & { * + [ o][f-]{\scriptscriptstyle 2 } } \ar@/^/[ll]^{\begin{bmatrix } c & 1 \\ d & e \end{bmatrix}}}$ \\ ( e ) : ~$\xymatrix@=.5cm{{*+[o][f-]{\scriptscriptstyle 1 } } \ar@/^/[rr]^{\begin{bmatrix } 1 \\ 0 \\ 0 \end{bmatrix } } & & { * + [ o][f-]{\scriptscriptstyle 3 } } \ar@/^/[ll]^{\begin{bmatrix } a & 1 & b \end{bmatrix } } \ar@/^/[rr]^{\begin{bmatrix } 1 & c & d \\ 0 & e & 1 \end{bmatrix } } & & { * + [ o][f-]{\scriptscriptstyle 2 } } \ar@/^/[ll]^{\begin{bmatrix } 0 & 0 \\ 1 & 0 \\ 0 & 1 \end{bmatrix}}}$ & \\ \end{array}\ ] ] recall from @xcite that the rings of polynomial quiver invariants are generated by traces along oriented cycles in the quiver .
it also follows from @xcite that these dimension vectors are the dimension vectors of simple representations and that their quotient varieties are rational of transcendence degrees : 1(a ) , 3(b ) , 4(c ) , 5(d ) and 5(e ) .
this proves ( a ) .
cases ( b),(c ) and ( d ) are easily seen ( focus on the middle vertex ) to be equivalent to the problem of classifying couples of @xmath208 matrices @xmath209 up to simultaneous conjugation in case ( d ) . for ( c )
the matrix @xmath210 needs to have rank one and for ( b ) both @xmath210 and @xmath148 have rank one .
it is classical that the corresponding rings of polynomial invariants are : ( b ) @xmath211 $ ] , ( c ) @xmath212 $ ] and ( d ) @xmath213 $ ] . in case
( e ) , as @xmath214 we can invoke the first fundamental theorem for @xmath21-invariants ( see ( * ? ? ?
ii.4.1 ) ) to eliminate the @xmath215-action by composing arrows through the @xmath216-vertex .
this reduces the study to the quiver - representations below , of which the indicated representations form a rational family by case ( d ) @xmath217[f-]{\scriptscriptstyle 1 } } \ar@(ul , dl)_a \ar@/^/[rr]^{\begin{bmatrix } 1 \\ 0 \end{bmatrix } } & & { * + [ o][f-]{\scriptscriptstyle 2 } } \ar@(ur , dr)^{\begin{bmatrix } c & d \\ e & 1 \end{bmatrix } } \ar@/^/[ll]^{\begin{bmatrix } 1 & b \end{bmatrix}}}\ ] ] and one verifies that the indicated family of ( e ) representations does lead to these matrices .
[ lem2 ] adding one vertex at a time ( until one has at most @xmath218 vertices ) to either end of the full subquivers of @xmath147 of lemma [ lem1 ] one can extend the dimension vector and the rational dense family of representations as follows in these allowed cases * @xmath219[f-]{\scriptscriptstyle 1 } } \ar@/^/[rr]^1 & & { * + [ o][f-]{\scriptscriptstyle 1 } } \ar@/^/[ll]^u \ar@/^/[rr ] & & \hdots \ar@/^/[ll]}$ ] * @xmath219[f-]{\scriptscriptstyle 1 } } \ar@/^/[rr]^{\begin{bmatrix } 1 \\ u \end{bmatrix } } & & { * + [ o][f-]{\scriptscriptstyle 2 } } \ar@/^/[ll]^{\begin{bmatrix } v & w \end{bmatrix } } \ar@/^/[rr ] & & \hdots \ar@/^/[ll]}$ ] * @xmath219[f-]{\scriptscriptstyle 2 } } \ar@/^/[rr]^{\begin{bmatrix } 1 & 0 \\ 0 & 1 \end{bmatrix } } & & { * + [ o][f-]{\scriptscriptstyle 2 } } \ar@/^/[ll]^{\begin{bmatrix } u & v \\ w & x \end{bmatrix } } \ar@/^/[rr ] & & \hdots \ar@/^/[ll]}$ ] before we add an extra vertex , the family of representations is simple and hence the stabilizer subgroup at the end - vertex is reduced to @xmath179 . adding the vertex and arrows , we can quotient out the base - change action at the new end - vertex by composing the two new arrows ( the first fundamental theorem for @xmath21-invariant theory , see ( * ? ? ?
ii.4.1 ) ) . as a consequence
, we have to classifying resp . a @xmath220 matrix , a rank one @xmath208 matrix or a @xmath208 matrix with trivial action .
the given representations do this . assuming the total number of vertices is @xmath17 we can glue two such subquivers and families of representations at a common end - vertex having dimension @xmath221 .
indeed , the ring of polynomial invariants of the glued quivers is the tensor product of those of the two subquivers as any oriented cycle passing through the glue - vertex can be decomposed into the product of two oriented cycles , one belonging to each component .
when the two subquivers @xmath5 and @xmath222 are glued along a @xmath221-vertex located at vertex @xmath223 we will denote the new subquiver @xmath224 .
remains only the problem of closing - the - circle by adding the @xmath18-th vertex to one of these families of representations on the remaining five vertices .
[ lem3 ] assume we have one of these dimension vectors on the full subquiver of @xmath147 on five vertices and the corresponding family of generically simple representations .
then , we can extend the dimension vector and the rational family of representations to the full quiver @xmath147 * @xmath225 & & { * + [ o][f-]{\scriptscriptstyle 1 } } \ar@/^/[ll ] \ar@/^/[rr]^{1 } & & { * + [ o][f-]{\scriptscriptstyle 1 } } \ar@/^/[ll]^u \ar@/^/[rr]^v & & { * + [ o][f-]{\scriptscriptstyle 1 } } \ar@/^/[ll]^w \ar@/^/[rr ] & & \hdots \ar@/^/[ll]}$ ] * @xmath225 & & { * + [ o][f-]{\scriptscriptstyle 1 } } \ar@/^/[ll ] \ar@/^/[rr]^{1 } & & { * + [ o][f-]{\scriptscriptstyle 1 } } \ar@/^/[ll]^u \ar@/^/[rr]^{\begin{bmatrix } v \\ w \end{bmatrix } } & & { * + [ o][f-]{\scriptscriptstyle 2 } } \ar@/^/[ll]^{\begin{bmatrix } x & y \end{bmatrix } } \ar@/^/[rr ] & & \hdots \ar@/^/[ll]}$ ] * @xmath225 & & { * + [ o][f-]{\scriptscriptstyle 2 } } \ar@/^/[ll ] \ar@/^/[rr]^{\begin{bmatrix } v & w \end{bmatrix } } & & { * + [ o][f-]{\scriptscriptstyle 1 } } \ar@/^/[ll]^{\begin{bmatrix } 1 \\ u \end{bmatrix } } \ar@/^/[rr]^{\begin{bmatrix } x \\ y \end{bmatrix } } & & { * + [ o][f-]{\scriptscriptstyle 2 } } \ar@/^/[ll]^{\begin{bmatrix } p & q \end{bmatrix } } \ar@/^/[rr ] & & \hdots \ar@/^/[ll]}$ ] forgetting the right - hand arrows , the left - hand representations are added as in the previous lemma .
then , the action on the right - hand arrows is trivial . for any of the obtained @xmath147-dimension vectors
, we can now describe a rational family of generically simple representations via a code containing the following ingredients * @xmath210,@xmath148,@xmath226,@xmath227,@xmath228 will be representations as in lemma [ lem1 ] , @xmath229 and @xmath230 will denote the mirror images of @xmath226 and @xmath228 * we add @xmath221 s or @xmath231 s when we add these dimensions to the appropriate side as in lemma [ lem2 ] * we add @xmath232 when we glue along a @xmath221-vertex placed at spot @xmath223 , and * we add @xmath233 if we close - up the family at a @xmath221-vertex placed at spot @xmath234 as in lemma [ lem3 ] in figure 1 we list rational families for all irreducible components of @xmath23 for dimensions @xmath195 .
a @xmath235-sign in column two indicates there are two such components , mirroring each other , @xmath181 indicates the @xmath48-dimension vector and @xmath111 gives the multiplicities of the @xmath89-simples of a semi - simple @xmath37-representation in the component .
the last column gives the dimension of this @xmath23-component .
[ ratfams ]
@xmath236 we are looking for @xmath48-dimension vectors @xmath196 satisfying @xmath237 , @xmath81 and @xmath199 , @xmath200 ( if @xmath238 there is a mirror component ) .
one verifies that one only obtains the given 5-tuples and that the indicated @xmath147-dimension vectors @xmath111 are compatible with @xmath116 and that the code is allowed by the previous lemmas . only in the final entry , type @xmath239
we are forced to close - up in a 2-vertex , but by the argument given in lemma [ lem3 ] it follows that in this case the following representations @xmath240 & & { * + [ o][f-]{\scriptscriptstyle 2 } } \ar@/^/[ll ] \ar@/^/[rr]^{\begin{bmatrix } 1 & 0 \\ 0 & 1 \end{bmatrix } } & & { * + [ o][f-]{\scriptscriptstyle 2 } } \ar@/^/[ll]^{\begin{bmatrix } u & v \\ w & x \end{bmatrix } } \ar@/^/[rr]^{\begin{bmatrix } y & z \end{bmatrix } } & & { * + [ o][f-]{\scriptscriptstyle 1 } } \ar@/^/[ll]^{\begin{bmatrix } p \\
q \end{bmatrix } } \ar@/^/[rr ] & & \hdots \ar@/^/[ll]}\ ] ] will extend the family to a rational dense family in @xmath150 . for example
, the code @xmath241 of component @xmath242 will denote the following rational family of @xmath147-representations @xmath243[f-]{\scriptscriptstyle 2 } } \ar@/^/[ld]|{\begin{bmatrix } 1 & a \end{bmatrix } } \ar@/^/[rd]^{\begin{bmatrix } 1 & 0 \\ 0 & 1 \end{bmatrix } } & \\ { * + [ o][f-]{\scriptscriptstyle 1 } } \ar@/^/[ru]^{\begin{bmatrix } 0 \\ 1 \end{bmatrix } } \ar@/^/[d]^{\begin{bmatrix } 0 \\ 1 \end{bmatrix } } & & { * + [ o][f-]{\scriptscriptstyle 2 } } \ar@/^/[lu]^{\begin{bmatrix } b & c \\ d & 1 \end{bmatrix } } \ar@/^/[d]^{\begin{bmatrix } i & j \end{bmatrix } } \\ { * + [ o][f-]{\scriptscriptstyle 2 } } \ar@/^/[u]^{\begin{bmatrix } 1 & e \end{bmatrix } } \ar@/^/[rd]^{\begin{bmatrix } 1 & 0 \\ 0 & 1 \end{bmatrix } } & & { * + [ o][f-]{\scriptscriptstyle 1 } } \ar@/^/[u]^{\begin{bmatrix } 1 \\ k \end{bmatrix } } \ar@/^/[ld]^{\begin{bmatrix } l \\
m \end{bmatrix } } \\ & { * + [ o][f-]{\scriptscriptstyle 2 } } \ar@/^/[lu]^{\begin{bmatrix } f & g \\ h & 1 \end{bmatrix } } \ar@/^/[ru]|{\begin{bmatrix } n & p \end{bmatrix } } & } \ ] ] corresponding to these representations are the @xmath148-matrices of theorem [ matrices ] @xmath244 which give us a rational @xmath245-dimensional family of @xmath7-representations , zariski dense in their component .
the rational zariski dense families of @xmath7-representations can be used to separate conjugacy classes of three string braids by taking traces .
recall that a _ flype _ is a braid of the form @xmath246 where @xmath247 and @xmath248 .
a flype is said to be non - degenerate if @xmath12 and the reversed braid @xmath249 are in distinct conjugacy classes .
an example of a non - degenerate flype of minimal length is @xmath250 @xmath251 hence we can ask whether a @xmath18-dimensional @xmath7-representation can detect that @xmath12 lies in another conjugacy class than its reversed braid @xmath252 @xmath253 note that from the classification of all simple @xmath7-representation of dimension @xmath17 by tuba and wenzl @xcite no such representation can separate @xmath12 from @xmath45 .
let us consider @xmath18-dimensional @xmath7-representations . from the previous section we retain that there are @xmath254 irreducible components in @xmath39 ( two being mirror images of each other ) and that rational parametrizations for the corresponding @xmath37-components are given by the following representations of type 6a,6b and 6c @xmath144[f-]{\scriptscriptstyle 1 } } \ar@/^/[rd]^{1 } & \\ & & { * + [ o][f-]{\scriptscriptstyle 1 } } \ar@/^/[lu]^{a } \ar@/^/[d]^{b } \\ { * + [ o][f-]{\scriptscriptstyle 1 } } \ar@/^/[rd]|{{\tiny \begin{bmatrix } 0 \\ 1 \end{bmatrix } } } & & { * + [ o][f-]{\scriptscriptstyle 1 } } \ar@/^/[u]^{1 } \ar@/^/[ld]^{{\tiny \begin{bmatrix } 1 \\ 0 \end{bmatrix } } } \\ & { * + [ o][f-]{\scriptscriptstyle 2 } } \ar@/^/[lu]^{{\tiny \begin{bmatrix } 1 & e \end{bmatrix } } } \ar@/^/[ru]|{{\tiny \begin{bmatrix } c & d \end{bmatrix } } } & } ~\qquad~\xymatrix@=1.1 cm { & { * + [ o][f-]{\scriptscriptstyle 1 } } \ar@/^/[ld]^{g } \ar@/^/[rd]^{1 } & \\ { * + [ o][f-]{\scriptscriptstyle 1 } } \ar@/^/[ru]^{f } \ar@/^/[d]^{e } & & { * + [ o][f-]{\scriptscriptstyle 1 } } \ar@/^/[lu]^{a } \ar@/^/[d]^{b } \\ { * + [ o][f-]{\scriptscriptstyle 1 } } \ar@/^/[u]^{1 } \ar@/^/[rd]^{1 } & & { * + [ o][f-]{\scriptscriptstyle 1 } } \ar@/^/[u]^{1 } \ar@/^/[ld]^{1 } \\ & { * + [ o][f-]{\scriptscriptstyle 1 } } \ar@/^/[lu]^{d } \ar@/^/[ru]^{c } & } ~\qquad~ \xymatrix@=1.1 cm { & { * + [ o][f-]{\scriptscriptstyle 1 } } \ar@/^/[rd]^{1 } & \\ & & { * + [ o][f-]{\scriptscriptstyle 1 } } \ar@/^/[lu]^{a } \ar@/^/[d]^{{\tiny \begin{bmatrix } 1 \\ 0 \end{bmatrix } } } \\ { * + [ o][f-]{\scriptscriptstyle 1 } } \ar@/^/[rd]^{1 } & & { * + [ o][f-]{\scriptscriptstyle 2 } } \ar@/^/[u]^{{\tiny \begin{bmatrix } c & d \end{bmatrix } } } \ar@/^/[ld]^{{\tiny \begin{bmatrix } 1 & e \end{bmatrix } } } \\ & { * + [ o][f-]{\scriptscriptstyle 1 } } \ar@/^/[lu]^{b } \ar@/^/[ru]|{{\tiny \begin{bmatrix } 0 \\ 1 \end{bmatrix } } } & } \ ] ] as a consequence we obtain a rational zariski dense subset of @xmath39 from theorem [ matrices ] using the following matrices @xmath148 for the three components @xmath255 using a computer algebra system , for example sage @xcite , one verifies that @xmath256 for all @xmath18-dimensional @xmath7-representations belonging to components @xmath257 and @xmath258 .
however , @xmath259 for an open subset of representations in component @xmath260 and hence @xmath18-dimensional @xmath7-representations can detect non - degeneracy of flypes . the specific representation given in the introduction ( obtained by specializing @xmath261 and @xmath262
gives @xmath263 whereas @xmath264 . in order to use the family of @xmath18-dimensional @xmath7-representations as an efficient test to separate 3-braids from their reversed braids
it is best to specialize the variables to random integers in @xmath265 $ ] to obtain an @xmath266-parameter family of @xmath7-representations over @xmath267 . in sage
@xcite one can do this as follows k.<l>=numberfield(x^2+x+1 ) ; a = randint(1,1000)*l+randint(1,1000 ) ; b = randint(1,1000)*l+randint(1,1000 ) ; c = randint(1,1000)*l+randint(1,1000 ) ; d = randint(1,1000)*l+randint(1,1000 ) ; e = randint(1,1000)*l+randint(1,1000 ) ; f = randint(1,1000)*l+randint(1,1000 ) ; g = randint(1,1000)*l+randint(1,1000 ) ; h = randint(1,1000)*l+randint(1,1000 ) ; b = matrix(k,[[1,0,0,a,0,f],[0,1,1,0,1,0],[1,1,0,1,0,0],[0,0,1,0,d , e ] , [ 0,1,0,b , c,0],[g,0,1,0,0,1 ] ] ) ; binv = b.inverse ( ) ; mat2=matrix(k,[[1,0,0,0,0,0],[0,1,0,0,0,0],[0,0,1,0,0,0],[0,0,0,-1,0,0 ] , [ 0,0,0,0,-1,0],[0,0,0,0,0,-1 ] ] ) ; mat3=matrix(k,[[1,0,0,0,0,0],[0,1,0,0,0,0],[0,0,l^2,0,0,0 ] , [ 0,0,0,l^2,0,0],[0,0,0,0,l,0],[0,0,0,0,0,l ] ] ) ; s1=h*binv*mat3*b*mat2 ; s2=h*mat2*binv*mat3*b ; s1inv = s1.inverse ( ) ; s2inv = s2.inverse ( ) ; .... one can then test the ability of this family to separate braids from their reversed braid on the list of all knots having at most @xmath266 crossings and being closures of three string braids , as provided by the knot atlas @xcite .
it turns out that the braids of the following knots can be separated from their reversed braids : @xmath268 , @xmath269 , @xmath270 , @xmath271 , @xmath272 ( all flypes ) , as well as the smallest non - invertible knot @xmath9 as mentioned in the introduction . | every irreducible component @xmath0 of semi - simple @xmath1-dimensional representations of the modular group @xmath2 has a zariski dense subset contained in the image of an tale map @xmath3 from the quotient variety @xmath4 of representations of a fixed quiver @xmath5 and a dimension vector @xmath6 such that @xmath4 is a rational variety . as an application
we will prove that there is a unique component of 6-dimensional simple representations of the three string braid group @xmath7 detecting braid - reversion .
further , we give explicit rational parametrizations of dense families of simple @xmath7-representations of all dimensions @xmath8 .
a knot is said to be invertible if it can be deformed continuously to itself , but with the orientation reversed
. there exist non - invertible knots , the unique one with a minimal number of crossings is knot @xmath9 which is the closure of the three string braid @xmath10 @xmath11{8_17.pdf}~\qquad~\qquad~\qquad~b~\includegraphics{braid817.pdf}\ ] ] proving knot - invertibility of @xmath9 essentially comes down to separating the conjugacy class of the braid @xmath12 from that of its reversed braid @xmath13 .
recall that knot invariants derived from quantum groups can not detect non - invertibility , see @xcite . on the other hand , @xmath14 for a sufficiently large @xmath7-representation @xmath15 .
in fact , bruce westbury discovered such @xmath16-dimensional representations , and asked for the minimal dimension of a @xmath7-representation able to detect a braid from its reversed braid , @xcite .
imre tuba and hans wenzl have given a complete classification of all simple @xmath7-representations in dimension @xmath17 , @xcite . by inspection ,
one verifies that none of these representation can detect invertibility , whence the minimal dimension must be @xmath18 .
unfortunately , no complete classification is known of simple @xmath7-representations of dimension @xmath19 . in this note
, we propose a general method to solve this and similar separation problems for three string braids .
let @xmath20 be the affine variety of all @xmath1-dimensional representations of the three string braid group @xmath7 .
there is a base change action of @xmath21 on this variety having as its orbits the isomorphism classes of @xmath1-dimensional representations .
the git - quotient of this action , that is , the variety classifying closed orbits @xmath22 is the affine variety @xmath23 whose points correspond to the isomorphism classes of semi - simple @xmath1-dimensional @xmath7-representations . in general , @xmath23 will have several irreducible components @xmath24 if we can prove that @xmath25 for all representations @xmath26 in a zariski - dense subset @xmath27 , then no representation @xmath15 in that component will be able to separate @xmath28 from @xmath29 . in order to facilitate the calculations , we would like to parametrize the dense family @xmath30 by a minimal number of free parameters .
that is , if @xmath31 we would like to construct explicitly a morphism @xmath32 from a rational affine variety of dimension @xmath33 , having a zariski dense image in @xmath34 .
the theory of luna - slices in geometric invariant theory , see @xcite and @xcite , will provide us with a supply of affine varieties @xmath35 and specific tale action-maps @xmath32 .
rationality results on quiver representations , see @xcite and @xcite , will then allow us to prove rationality of some specific of these varieties @xmath35 .
as the modular group @xmath36 is a central quotient of @xmath7 it suffices to obtain these results for @xmath37 . in the first two sections we will prove the affine variety classifying @xmath1-dimensional semi - simple representations of the modular group decomposes into a disjoint union of irreducible components @xmath38 there exists a fixed quiver @xmath5 having the following property . for every component @xmath0 containing a simple representation
, there is a @xmath5-dimension vector @xmath6 with rational quotient variety @xmath4 and an tale action map @xmath3 having a zariski dense image in @xmath0 .
we apply this general method to solve westbury s separation problem . of the four irreducible components of @xmath39
the three of dimension 6 can not detect invertibility , whereas the component of dimension 8 can .
a specific representation in that component is given by the matrices @xmath40 @xmath41 where @xmath42 is a primitive third root of unity .
one verifies that @xmath43 for the braid @xmath12 describing knot @xmath9 , whereas @xmath44 for the reversed braid @xmath45 . in order to facilitate the application of his method to other separation problems of three string braids , we provide in section three explicit rational parametrizations of dense families of simple @xmath1-dimensional @xmath7-representations for all components and all @xmath46 .
this can be viewed as a first step towards extending the tuba - wenzl classification @xcite . |
[ see eq . [ eq : symenergy ] ] and the solid line to the full band structure with @xmath112 , @xmath113 .
black lines are for @xmath114 and red lines for @xmath115 .
( c ) the bottom of the conduction band for @xmath114 showing the absence of sombrero structure for finite @xmath116 .
( d ) the bottom of the conduction band for @xmath115 showing the restoration of the sombrero structure for finite @xmath116 . ( e ) @xmath0 for clean bilayer graphene .
the lines correspond to the same as in ( a ) .
[ fig : cleanbs ] ] the band structure associated with the hamiltonian presented in eq .
is shown in fig .
[ fig : cleanbs ] . in the symmetric case with @xmath117 ,
the low energy bands have the dispersion @xmath118 where @xmath119 . writing this as a function of density and differentiating yields the following analytical expression for @xmath0 in this limit in the conduction band : @xmath120 where @xmath121 and @xmath122 .
in contrast , when @xmath8 and @xmath9 are included in the analysis , analytical solutions of the eigenvalue problem are not possible .
the energy spectrum is found as the roots of the following quartic polynomial equation for @xmath123 : @xmath124 e^2 \\
+ \left [ \frac{\delta u^2}{2 } + 2\kappa^2\left ( \delta + \chi^2\delta + 2 \chi \gamma_1 \right ) \right ] e
\\ + \frac{u^2}{4}\left ( \gamma_1 ^ 2 - \delta^2 \right ) + \left [ \frac{u^2}{4 } - \kappa^2\left ( 1 - \chi^2 \right ) \right]^2 = 0 \label{eq : g4poly}\end{gathered}\ ] ] where @xmath125 .
the fermi surface may be ring - shaped for finite values of @xmath10 so that the wave vectors corresponding to the inner and outer fermi surfaces @xmath126 are given by the equation @xmath127 with the constraint that @xmath128 and @xmath129 must give the same energy for the appropriate band when substituted into eq . .
when the fermi energy is above the sombrero region , @xmath130 .
the extracted value of @xmath128 can be substituted into eq . to find @xmath32 .
the differentiation to obtain @xmath3 can the be carried out numerically .
evaluations of these equations are shown in fig .
[ fig : cleanbs ] for the band structure and @xmath3 for two different values of the band gap , and with and without the tight - binding parameters @xmath8 and @xmath9 .
the step feature in @xmath3 is a result of the change in topology of the fermi surface from a ring to a disc as the fermi energy leaves the sombrero region . for small values of @xmath10 ,
the conduction band does not show the sombrero shape when @xmath116 are finite and this step is absent .
the spike at the origin is a @xmath131-function resulting from the discontinuity of the fermi energy at zero density as it jumps from the valence band to the conduction band .
r. nandkishore and l. levitov , phys .
104 * , 156803 ( 2010 ) .
j. jung , f. zhang , and a. h. macdonald , phys .
b * 83 * , 115408 ( 2011 )
. f. zhang , j. jung , g. a. fiete , q. niu , and a. h. macdonald , phys .
lett . * 106 * , 156801 ( 2011 ) . a typical disorder - induced charge imbalance between the two blg layers as small as @xmath132 , ( which is of the same order as the disorder - induced in - plane density fluctuations @xmath27 in the system , although they are independent physical quantities ) , is sufficient to produce the band gap necessary for interpreting the observed peak in @xmath3 in the experiment @xcite .
although this is suggestive that the experimental peak is arising entirely from extrinsic effects , we can not , however , rule out a small many - body spontaneous gap underlying the disorder effects . | we present an interpretation of recent experimental measurements of @xmath0 in suspended bilayer graphene samples .
we demonstrate that the data may be quantitatively described by assuming a spatially varying inter - layer potential asymmetry ( which generates a band gap ) induced by local electric fields resulting from charged impurity disorder in the graphene environment .
we demonstrate that the fluctuations in the inter - layer potential asymmetry and density vary between different samples , and that the potential asymmetry fluctuations increase in magnitude as the density fluctuations increase .
this indicates that the mechanism causing this effect is likely to be extrinsic .
we also provide predictions for the optical conductivity and mobility of suspended bilayer graphene samples with small band gaps in the presence of disorder .
recently , the fabrication of suspended bilayer graphene ( blg ) devices @xcite has allowed the direct local measurement of the thermodynamic quantity @xmath1 using scanning single electron transistor spectroscopy .
this is an important quantity to study since it is linked directly to thermodynamic observables such as the thermodynamic density of states , electronic compressiblility and the quantum capacitance of the interacting electron liquid in the blg . in samples mounted on a sio@xmath2 substrate ,
the level of disorder is generally too high for the intrinsic physics near the charge neutrality point to be seen because of the existence of strong inhomogeneity ( electron - hole puddles ) induced by the disorder@xcite .
however , suspended samples provide a setting where the effective strength of the impurities is reduced as reflected , for example , in the enhanced mobility@xcite and it is likely that the low - density physics is more readily accessible . in the experiments@xcite , a region of decreased compressiblilty ( or equivalently a peak in @xmath3 )
is seen near the charge - neutrality point in zero magnetic field indicating the possible presence of a small gap in the low energy band structure .
it is claimed @xcite that the size of the gap which generates this peak is approximately @xmath4 , and that the magnitude of charge inhomogeneity is of the order of @xmath5 .
the authors of ref .
state that castro _ et al .
_ @xcite show that this level of in - plane charge inhomogeneity by itself would create a gap which is a factor of 10 smaller than that observed , ruling out disorder - induced fluctuations as a mechanism for the generation of this gap .
instead , they claim@xcite that this peak is evidence for a correlated ground state of the interacting electron liquid , such as one of those proposed in the recent literature @xcite
. some of these proposed states may induce a spontaneous many - body charge transfer instability in the bilayer system which could open a small band gap at zero density , and the region of decreased compressibility seen in the experiments may be evidence for the formation of one of these charge - transfer - instability states @xcite .
however , we believe that this interpretation of the experimental results is problematic because the out - of - plane charge imbalance which would drive a disorder - induced gap is not the same quantity as the inplane charge inhomogeneity which martin _ et al . _ use@xcite to estimate the gap size .
therefore it is important to look more closely at the role of disorder in the suspended bilayer graphene system to investigate in depth the issue of the possible existence ( or absence ) of a many - body quantum phase transition creating a spontaneous band gap .
the goal of the current work is to carry out such a phenomenological investigation to see whether random charge impurities in the bilayer environment could produce a signature in the compressibility consistent with the observations of refs . .
transport measurements suggest that there are some residual impurities in the environment of the suspended graphene , each of which will have an electric field associated with it , and it is well established that the electrons in the blg order themselves into puddles of electrons and holes in order to screen these fields @xcite . additionally , it was shown recently @xcite that these electric fields also cause a spatial variation in the interlayer potential asymmetry which is highly correlated with the disorder profile .
this is a different effect from the inplane reorganization of charge , and the two effects may in principle be of quite different magnitudes , and both are likely to be present in disordered bilayer graphene .
although this interlayer effect was reported for bilayer graphene supported on a substrate@xcite , the general principle applies to suspended samples as well . in this article
, we revisit the issue of density fluctuations induced by charged impurity disorder , and show that our calculation of @xmath3 by averaging over the density of the electrons in the puddles@xcite can be applied in two different phenomenological ways to this situation and gives results for the gap fluctuations , density fluctuations , and required charge imbalance which are self - consistent . by fitting this theory to measurements of @xmath3 from two different devices@xcite
, we show that the gap and charge inhomogeneity are device - dependent , which is , of course , not unexpected since disorder in different devices is likely to be different leading to different interlayer potential asymmetry in each sample .
we also demonstrate that the transport characteristics associated with our phenomenological disorder fits are reasonable , and predict that in an optical spectroscopy experiment , the gap will still be obscured by disorder in these samples@xcite .
since our results are internally consistent , we claim that the mechanism of local density fluctuations induced by charged impurity disorder can not be ruled out as a cause of the observed@xcite peak in @xmath3 at low carrier density .
our work by no means eliminates the possible occurence of a many - body blg instability , it just establishes the qualitative and quantitative importance of disorder in thinking about the quantitative aspects of the observed effects , and points out that there is a possible single - particle extrinsic cause for the observed effective gap , namely , the inter - layer potential asymmetry induced by the random charged impurities in the blg environment . in reality
, both interaction and disorder may be present in a non - perturbative manner , making a microscopic theoretical analysis quite challenging , particularly since the graphene landscape becomes inhomogeneous due to the formation of electron - hole puddles in the presence of the charged impurity disorder . in order to model the clean blg system
, we consider the continuum limit of the tight - binding hamiltonian . in this approach ,
the wave functions are described by bloch functions modulated by a spinor , the elements of which correspond to the four sites in the blg unit cell@xcite .
we make this assumption despite the fact that the inhomogeneity in the disorder potential breaks translational symmetry because we assume that the changes in the potential landscape are spatially slowly varying enough that the physics associated with p
n junctions or quantum dots do not manifest .
we also point out that this theory has been very successful in describing capacitance - based compressibility experiments in samples which are much dirtier than those we currently discuss@xcite .
the band structure is determined by the relative strength of electron hopping between the four lattice sites . to simulate the experimental situation as accurately as possible ,
we go beyond the standard nearest - neighbor tight binding formalism ( where @xmath6 is the fermi velocity of monolayer graphene which comes from the intra - layer hopping between adjacent atoms and @xmath7 is the energy associated with the inter - layer dimer bond which generates the finite effective mass at low energy ) by including the next - nearest neighbor inter - layer hops with energy @xmath8 and the onsite energies given by the lattice site asymmetry @xmath9 and a potential asymmetry @xmath10 between the two layers .
this potential asymmetry directly generates a band gap at the k point of the band structure , and we shall therefore use the terms ` inter - layer potential asymmetry ' and ` gap ' synonymously in this paper .
the hamiltonian for one valley is then @xmath11 where @xmath12 with @xmath13 the lattice constant , and @xmath14 and @xmath15 are respectively the radial and angular part of the wave vector measured from the k point .
the spectrum is valley - symmetric and isotropic .
the effect of the @xmath8 and @xmath9 terms is to introduce an asymmetry between the low - energy conduction and valence bands , and has previously been used to explain an observed asymmetry in capacitance measurements @xcite . using the results collated in ref .
we choose representative parameters @xmath16 , @xmath17 , @xmath18 , @xmath19 and @xmath20 .
we describe the band structure and @xmath3 in the absence of disorder in the appendix .
the existence of puddles of carriers in graphene is a well - established phenomenon both in experiment and in theory@xcite . in a previous paper@xcite
, we introduced the macroscopic averaging over density fluctuations and applied this theory to measurements of the capacitance of blg devices finding good agreement with the experimental data .
the idea of the theory is that the potential fluctuations associated with charged impurity disorder generates a spatially - varying on - site term in the tight - binding hamiltonian .
this in turn leads to a local fluctuation in the fermi energy and hence in the carrier density .
the statistics of this fluctuation can be encapsulated within a distribution function @xmath21 .
the coupling between the tip of the set and the electron liquid in the bilayer graphene is capacitative , so we assume that each region with a given carrier density has a local capacitance associated with it .
then , the total capacative coupling of the set to the sample is the parallel coupling of all of the areas , which equates to the geometrical average over the region sampled by the set .
therefore , the areal average can be replaced by an average over the charge density distribution @xmath21 .
we note that the dimension of the region sampled by the set in the experiments of ref .
is @xmath22 whereas the puddle size seen in ref .
is @xmath23 so that the averaging should be reasonable in this case .
we stress that this philosophy deliberately retains the inhomogeneity of the system and treats it as the primary manifestation of disorder .
this is a fundamentally different approach from , for example , a diagrammatic expansion of the electron impurity interaction which contains within it an average over realizations of disorder which explicitely restores the translational symmetry .
we believe that the important disorder physics to be included in the low - density regime of graphene near the charge neutrality point is the formation of electron - hole puddles in the system around the discrete quenched charged impurity centers leading to strong spatial inhomogeneity modelled by the distribution function @xmath21 mentioned above which has been calculated in the literature in a few instances @xcite and measured in experiment for a bare sio@xmath2 substrate@xcite . for density fluctuations associated with puddles of carriers in graphene systems we write the local density as @xmath24 where @xmath25 is the average density induced by external gates and the fluctuations are encapsulated within @xmath26 .
we parameterize these fluctuations by their standard deviation @xmath27 , and computing the average of @xmath3 over the probability distribution for the density results in the following expression for the average inverse compressibility @xmath28 the distribution function @xmath21 is talken as the following gaussian @xmath29 , \label{eq : pdef}\ ] ] since there is compelling theoretical@xcite and experimental@xcite evidence that this is approximately the correct form . in this theory ,
the layer asymmetry @xmath30 is constant in space ( hence the subscript ` @xmath31 ' ) and is generated by either external electric fields ( such as those from gates ) or by an asymmetry of charge between the two layers of the blg generated by the intrinsic electron - electron interactions .
analytical evaluation of this procedure is not possible since , for finite @xmath8 and @xmath9 , the eigenvalues of the hamiltonian are the roots of a quartic polynomial .
therefore , the expression for @xmath32 as a function of @xmath33 , its derivative , and the integration in eq . must be calculated numerically . from eq . to the experimental data .
note that this theory includes a spatially uniform band gap @xmath34 .
the black lines are for published data @xcite , and the red lines for the unpublished data@xcite .
( b ) fits of the double averaged @xmath35 from eq . with @xmath36 to the experimental data .
note that in this theory , the band gap @xmath37 fluctuates in space , and the distribution of these fluctuations is parameterized by @xmath38 .
the table shows a comparison of the fitting parameters for the two samples and two procedures . [
fig : averagefits],title="fig : " ] from eq . to the experimental data .
note that this theory includes a spatially uniform band gap @xmath34 .
the black lines are for published data @xcite , and the red lines for the unpublished data@xcite .
( b ) fits of the double averaged @xmath35 from eq . with @xmath36 to the experimental data .
note that in this theory , the band gap @xmath37 fluctuates in space , and the distribution of these fluctuations is parameterized by @xmath38 .
the table shows a comparison of the fitting parameters for the two samples and two procedures .
[ fig : averagefits],title="fig : " ] [ cols="^,^,^,^,^ " , ] figure [ fig : averagefits](a ) shows the results of finding the best fit between eq . and the experimental data@xcite .
we find that the published data@xcite ( which we shall denote by the label prl ) is quantitatively described by a constant gap @xmath39 with density fluctuations characterized by @xmath40 , which is very close to the values claimed by martin _
et al .
_ in their paper@xcite .
the unpublished data@xcite ( denoted by pc , as in ` private communication ' ) is described by @xmath41 and @xmath42 .
therefore , a different degree of in - plane charge homogeneity and different size of gap is required to describe the experimental data in each sample which is perfectly reasonable since we expect the disorder to have sample - to - sample variations . also , the band gap required to fit the data is larger for the more strongly disordered sample
. the opposite would be true if the gap was the result of an intrinsic interaction - driven effect since in this case , the gap due to interactions would be universal , but disorder would work to reduce its size@xcite .
therefore , these results , particularly the fact that the more strongly disordered sample exhibits the larger effective experimental band gap , indicate that there is some role being played by extrinsic effects in these experiments .
we now introduce another step in this theory , motivated by recent observations of spatial fluctuations in the interlayer potential asymmetry in bilayer graphene @xcite .
the interlayer asymmetry was measured in electron puddles and in hole puddles of a bilayer graphene flake which was mounted on an sio@xmath2 substrate .
the difference in the asymmetry of approximately @xmath43 was seen between the two puddles at zero back gate voltage , and an associated band gap was seen in the spectrum of landau levels measured via stm .
although we expect that the magnitude of this effect will be smaller in suspended samples , the presence of disorder - induced puddles ( as demonstrated above by our fitting to the appropriate theory ) means that it is sensible to assume that the same disorder may also induce ( albeit smaller in magnitude ) local fluctuations in the potential asymmetry just as it does in graphene on substrates in ref . .
in this case , the potential asymmetry becomes a function of position such that @xmath44 .
the fluctuations are contained within @xmath45 , and we assume that they can be described by the distribution @xmath21 with standard deviation @xmath38 ( just as the density fluctuations can ) because the same disorder is generating both types of fluctuation . note that it is the in - plane component of the electric field associated with disorder which causes the formation of puddles , while the spatially varying inter - layer potential asymmetry is driven by the out - of - plane component of the electric field .
therefore , while we expect the two types of fluctuation to be correlated , it is not obvious exactly what relationship should exist between them .
however , on a phenomenological level , we can treat @xmath27 and the potential asymmetry fluctuations parameterized by @xmath38 as independent parameters .
we can therefore perform an average over the gap fluctuations in the same way we previously did for the density .
the spatially varying part of the band gap @xmath45 is described by @xmath38 , and the gaussian is centered around @xmath46 .
the spatial average then corresponds to the following average over the distribution function : @xmath47 we call this procedure ` double averaging ' . in order to assess
if the observed gap might be entirely due to disorder - induced fluctuations , we fit this expression with @xmath36 to the experimental data .
therefore the fitting parameters are now @xmath48 and @xmath27 , so we have two parameters as was the case for the density averaging procedure .
the results are displayed in fig .
[ fig : averagefits](b ) .
we find that the prl data is fitted by @xmath49 and @xmath50 and the pc data by @xmath51 and @xmath52 .
the fit of the double - averaged theory is actually slightly better than that which is averaged only over density ( especially in the region of the peak ) , but both averaging procedures give results which are in reasonable agreement with the experiment and therefore it is not possible to distinguish between a spatially fluctuating disorder - induced gap and a uniform intrinsic gap in the current data@xcite .
therefore , our work convincingly demonstrates they key importance of disorder , particularly the density and potential fluctuations associated with the inhomogeneous puddles , in determining the experimental compressibility measurements of refs . . and
@xmath53 for ( a ) @xmath54 , ( b ) @xmath55 , and ( c ) @xmath56 . in all cases , we have @xmath57 . [
fig : gaussians ] ] using the parameters from these fits , we can make the following observation . in principle , if there was truly no disorder then both @xmath27 and @xmath38 should be zero . however , if we take the two pairs extracted from the double averaging fitting procedure and assume a linear extrapolation to @xmath58 , we find a residual @xmath59 . this may indicate the presence of an intrinsic ( possibly , many - body ) gap not generated by the fluctuations , so we examine that issue now .
the experimental measurements of @xmath3 in the prl data suggest that the size of the uniform gap , if it exists , is approximately @xmath4 . since this is of the same order as the gap predicted by extrapolating to the clean limit , we have run fits of our double - averaged theory with a constant gap of this size to the experimental data .
we find that there is no improvement in the fitting for any value of @xmath27 or @xmath38 .
to explain this , we need to analyze the distribution function for the gap fluctuations . we know that @xmath3 is an even function of the gap size @xmath10 , and
therefore the absolute value @xmath60 is the key for determining the distribution of @xmath3 .
therefore , although we average over the gaussian @xmath21 in eq . , the effective distribution of @xmath3 is non - gaussian , and given by @xmath61 where @xmath62 and @xmath63 is defined in eq . .
figure [ fig : gaussians ] shows the distribution @xmath53 for the cases @xmath64 , @xmath55 , and @xmath56 .
we see that in all but the last case , the distribution is dominated by small values of @xmath65 because of the significant contribution from the @xmath66 term , and it is only when @xmath67 that the gaussian shape is revealed in @xmath53 . in the two sets of data which we have studied ,
we see that @xmath68 found from the single averaging procedure and the @xmath38 from the double averaging obey the relation @xmath69 .
this suggests that the disorder is too strong to determine if a spontaneous gap exists in the absence of disorder . in order to measure the spontaneous gap , samples with levels of disorder such that @xmath70 are needed
this necessitates the development of samples with much higher quality ( i.e. less disorder ) , perhaps mobilities which are a factor of five or so larger , to decisively establish the existence of a spontaneous many - body blg gap . in short
, the disorder - induced gap needs to be a factor of 2 or more smaller than the many - body gap for an unambiguous conclusion on this issue . and ( b ) @xmath71 for various degrees of disorder broadening .
the color of the lines is the same in both panels .
dashed lines indicate the clean case .
[ fig : optcond ] ] the gap size may also be extracted from measurements of the optical conductivity via absorption or reflection experiments @xcite . to demonstrate that this may be an unreliable process in the case of the small band gaps ( of the order of a few mevs ) we are discussing in this context , we plot in fig .
[ fig : optcond ] the optical conductivity derived from the kubo formula with @xmath72 in panel ( a ) and with @xmath73 in panel ( b ) . by way of comparison , in the linear ( high density ) part of the hyperbolic spectrum of bilayer graphene , the density fluctuations @xmath27 correspond to fluctuations in the fermi level of approximately @xmath74 , and with @xmath75 , this gives an energy scale of approximately @xmath76 .
when the fermi energy is in the sombrero ( low density ) region with @xmath72 , the flatness of the bands means that this density fluctuation corresponds to a fluctuation of less than 1@xmath77 .
we include the disorder through a constant broadening of the bands parameterized by the energy @xmath78 which appears in the green s function as @xmath79 the optical conductivity is calculated from this green s function from the following expression@xcite : @xmath80 where @xmath81 and @xmath82 are the spin and valley degeneracies , and @xmath83 figure [ fig : optcond ] shows that for @xmath84 , the peak due to the onset of the intraband conductivity is already significantly blurred . by the time @xmath85 , the peak is completely smeared , as shown by the @xmath86 line .
we mention in this context that the same conclusion can also be reached by considering an inhomogeneous broadening effect rather than a homogeneous broadening as used in eq .
for example , we can introduce a density fluctuation leading to an uncertainty in the fermi energy by 1mev , leading to exactly the same conclusion that the optical gap measurement would be unable to discern the small band gap in the presence of this inhomogeneous broadening effect .
for a charged impurity density , @xmath87 , and a fixed short range potential @xmath88 ( ev)@xmath89 .
different curves correspond to the band gap @xmath90 , 0.15 , 0.3 , 0.33 ev ( from top to bottom ) .
inset shows the calculated conductivities as a function of carrier density for two different charged impurity densities @xmath91 ( dashed line ) and @xmath92 ( solid line ) with zero band gap .
[ fig : sig ] ] we also predict the transport properties of suspended bilayer graphene with disorder so that independent characterization of the samples in which @xmath3 was measured can be done and compared with our theory incorporating inhomogeneous puddle effects of density fluctuations .
we use the highly successful boltzmann - rpa formalism , which has been used to study the transport in monolayer graphene and blg systems @xcite .
this theory incorporates both long - range coulomb impurities and short - range scatterers . in fig .
[ fig : sig](a ) we show the mobility of suspended blg as a function of charged impurity density @xmath93 , calculated at a carrier density @xmath94 for different short - ranged disorder potentials , @xmath95 , where @xmath96 is the density of short - ranged impurities and @xmath97 is the strength of the potential .
if we assume a charged impurity density @xmath98 , then we have @xmath99 cm@xmath89/vs with a fixed short range potential @xmath88 ( ev)@xmath89 .
however , we note that even though the carrier density fluctuation is related to the impurity density , the theoretical relation between two densities is not yet known . in fig . [
fig : sig](b ) we show the conductivity calculated within effective medium theory@xcite as a function of gate voltage @xmath101 in the presence of an energy band gap .
we assume that the energy dispersion of carriers is given by @xmath102 , where @xmath103 indicates the electron ( @xmath104 ) or hole ( @xmath105 ) band , @xmath106 is the effective mass , and @xmath107 is the band gap . in this figure a charged impurity density @xmath108 and a fixed short range potential @xmath88 ( ev)@xmath89 are used . our calculation of fig .
[ fig : sig](b ) shows that due to the electron hole puddles the transport gap ( i.e. the region of zero conductivity ) does not occur even though there is a band gap ( top three curves in the figure ) .
only large enough energy gaps ( the lowest curve ) reflect the transport gap .
however , our calculation also shows that for small density fluctuations ( i.e. shallow electron - hole puddles ) the transport gap can be induced by the smaller energy gap . in the inset to fig .
[ fig : sig](b ) the conductivity of suspended blg calculated within the effective medium theory @xcite is shown for two different charged impurity densities , @xmath109 and 4@xmath110 for a fixed short range potential @xmath88 ( ev)@xmath89 .
our calculation shows the minimum conductivity at the charge neutral point is higher for a low mobility ( more strongly disordered ) sample . in conclusion ,
we have demonstrated that even though suspended graphene samples are known to be significantly less disordered than those mounted on substrates , the effect of disorder in recent measurements of @xmath0 can not be ruled out .
using phenomenological averaging procedures , which specifically incorporate the effects of strong inhomogeneous broadening arising from electron - hole puddle induced spatial density and potential fluctuations , we have demonstrated that current experimental data can not distinguish between an intrinsic , spatially uniform band gap generated by electron - electron interactions , and a spatially fluctuating band gap induced by disorder . specifically , these disorder - induced fluctuations can account for the observed data without invoking the presence of an intrinsic gap@xcite . in order to be confident about the origin of the band gap , higher quality samples where @xmath111 must be fabricated .
one possibility is to carry out measurements on blg samples fabricated on bn substrates which typically have much lower charged impurity disorder @xcite .
the theory we develop involves exactly one uncontrolled approximation , assuming the density fluctuations and the fluctuations in @xmath10 to be uncorrelated .
going beyond this approximation would require knowledge about the details of the local electric fields producing the fluctuations in @xmath33 and @xmath10 which is currently unavailable experimentally .
it would be straightforward to include the correlation in our theory if microscopic information about the correlator between fluctuations in @xmath33 and @xmath10 is available from experiments or some other microscopic calculations . in the absence of such information we assumed them to be independent .
we also emphasize that the theory presented in this article is phenomenological and is a first step toward the understanding of the role of disorder in the experiments of refs . .
what we have achieved in this paper is a demonstration that spatial density and potential fluctuations are capable of producing a spatially fluctuating gap consistent with the experimental data .
many questions of details could be raised with respect to our phenomenological procedure , for example , the averaging procedure we employ is certainly not unique , but in the presence of real spatial fluctuations which explicitly break the translational invariance in graphene near the dirac point , our procedure is physically motivated as it averages precisely over the quantities which fluctuate according to some distribution function which has been calculated in the literature @xcite . a better ( purely numerical ) method to tackle
the problem would be to use the techniques developed in refs .
to explicitly and self - consistently calculate the real - space electronic structure in the presence of the electron - hole puddles generated by the random charge impurities , and then to use this real - space electronic structure to directly numerically calculate the compressibility .
such a calculation is obviously numerically prohibitively difficult , and our work is a short - cut ( albeit a rather simple one ) in carrying out such a numerical procedure . since the gaussian probability distribution function we use is numerically well - verified in refs . through self - consistent calculations
, we believe that our results , although phenomenological , should be qualitatively and quantitatively valid . the eventual theory involving a microscopic calculation of the compressibility in the presence of long - range coulomb disorder and electron - electron interaction including
the realistic band structure of bilayer graphene is very far in the future since no one knows how to approach such a non - perturbative problem even at a formal level .
we acknowledge support from us - onr and nri - swan and thank amir yacoby for sharing unpublished data with us . |
the shadowing property of dynamical systems ( diffeomorphisms or flows ) is now well - studied ( see , for example , the monographs [ 1 , 2 ] and the recent survey [ 3 ] ) .
this property means that , near approximate trajectories ( so - called pseudotrajectories ) , there exist exact trajectories of the system . in this paper , we are interested in shadowing near fixed points . mostly , standard methods allow one to show that a diffeomorphism has the shadowing property near a hyperbolic fixed point , and this property is lipschitz , see [ 1 ] .
one can mention several papers which contain methods of proving the shadowing property for systems with nonhyperbolic behavior ( see , for example , [ 4 , 5 ] ) .
our approach is based on the method of lyapunov type functions .
first , lyapunov type functions were used in the study of shadowing and topological stability by lewowicz in [ 6 ] .
we apply results on lyapunov type functions obtained in our paper [ 7 ] .
let us state the problem of shadowing in general .
let @xmath0 be a homeomorphism of a metric space @xmath1 . in this paper
, we define a finite @xmath2-pseudotrajectory of @xmath0 as a set of points @xmath3 such that @xmath4 in the study of shadowing in noninvariant sets ( such as neighborhoods of fixed points ) , the concept of finite shadowing is natural .
we say that @xmath0 has the finite shadowing property in a set @xmath5 if , for any @xmath6 , there exists a @xmath7 such that if @xmath8 is a @xmath2-pseudotrajectory of @xmath0 , then there exists a point @xmath9 such that @xmath10 let us emphasize that in the above definition , @xmath2 depends on @xmath11 and @xmath12 but not on @xmath13 .
the structure of the paper is as follows . in sec .
[ sec2 ] , we treat the one - dimensional case .
we prove a simple general statement ( theorem [ res1 ] ) and show that , in some cases , the dependence of @xmath2 on @xmath12 can be clarified .
section [ sec3 ] is devoted to the method of lyapunov type functions developed in [ 7 ] . in sec .
[ sec4 ] , we give general conditions of finite shadowing in the two - dimensional case and then treat in detail an important example of a diffeomorphism of the form @xmath14 where @xmath15 are natural numbers and @xmath16 are smooth functions that vanish at the origin together with their jacobi matrices . finally , sec .
[ sec5 ] is devoted to shadowing near a nonisolated fixed point .
we study a simple ( but nontrivial ) example of the diffeomorphism @xmath17 for which the origin is a nonisolated fixed point ( any point @xmath18 is a fixed one ) .
of course , such a system does not have the usual shadowing property . in this case
, we work with a new concept of shadowing in which we control the size of one - step errors .
our methods can be applied to dynamical systems with phase space of arbitrary dimension ; in this paper , we restrict the consideration to one - dimensional and two - dimensional systems to clarify the presentation of the main ideas .
first we consider the problem of shadowing near a nonhyperbolic fixed point of a homeomorphism in the simplest , one - dimensional , case .
let @xmath0 be a homeomorphism of a neighborhood @xmath19 of a fixed point @xmath20 to its image .
we consider in detail the case where @xmath0 is nonhyperbolically expanding in a neighborhood of a fixed point ; the case of nonhyperbolic contraction near a fixed point is treated similarly .
we impose simplest possible conditions on @xmath0 ; in a sense , precisely this topological form of conditions is generalized by our conditions in the two - dimensional case
. * condition 1 .
* there exist numbers @xmath21 such that if @xmath22 and @xmath23 , then @xmath24 denote by @xmath25 the closed @xmath9-neighborhood of a set @xmath26 .
[ res1 ] if condition _
1 _ is satisfied , then @xmath0 has the finite shadowing property in the set @xmath27 fix @xmath6 ; we assume that @xmath28 .
condition 1 implies that if @xmath29 , then @xmath30 the compactness of the set @xmath31 and the continuity of @xmath0 imply that there exists a @xmath7 such that @xmath32 for @xmath29 .
let @xmath33 be a @xmath2-pseudotrajectory of @xmath0 ; denote @xmath34 .
we claim that @xmath35 indeed , let @xmath36 $ ] and let @xmath37 .
inclusion ( [ 2 ] ) with @xmath38 implies that @xmath39 now it follows from the inequalities @xmath40 and @xmath41 that @xmath42 ; hence , @xmath43 similarly one shows that @xmath44 , which proves ( [ 3 ] ) .
it follows from inclusions ( [ 3 ] ) that the set @xmath45 is not empty .
clearly , for any point @xmath46 , inequalities ( [ 0 ] ) hold . in theorem [ res1 ]
, we can say nothing about the dependence of @xmath2 on @xmath12 . for a particular example of a diffeomorphism with a nonhyperbolic fixed point considered below ,
such a dependence can be clarified .
* example 1 .
* let @xmath47 , where @xmath48 is a natural number .
take @xmath6 and let @xmath49 then @xmath50 is a polynomial in @xmath51 of even degree with positive leading coefficient . since the derivative of @xmath50 in @xmath51 has a unique zero @xmath52 , the inequality @xmath53 holds .
thus , the form @xmath54 of degree @xmath55 is positive definite , and there exists a positive number @xmath56 such that @xmath57 assume that @xmath58 then there exist @xmath59 such that if @xmath22 and @xmath60 , then @xmath61 if @xmath62 is a @xmath2-pseudotrajectory with @xmath63 , then it follows from ( [ 4 ] ) and ( [ 6 ] ) that @xmath64 @xmath65 this relation and a similar relation for @xmath66 mean that an analog of inclusion ( [ 2 ] ) holds for any @xmath67 . repeating the proof of theorem 1 ,
we conclude that @xmath0 has the finite shadowing property in the set @xmath68 .
note that , for example , condition ( [ 5 ] ) is valid if @xmath69 .
our reasoning also shows that if @xmath70 , then @xmath0 has the finite shadowing property in the whole line @xmath71 with the same dependence of @xmath2 on @xmath12 given by ( [ 7 ] ) . in [ 8 ] ,
s. tikhomirov used a different approach to show that the diffeomorphism @xmath72 has the shadowing property with @xmath73 .
as was mentioned in the introduction , we consider in detail the problem of finite shadowing for a homeomorphism @xmath0 of the plane @xmath74 in two cases : case i ( the origin is an isolated nonhyperbolic fixed point ) and case ni ( the @xmath75-axis consists of fixed points ) .
we apply the approach based on pairs of lyapunov type functions developed in the paper [ 7 ] .
let us formulate the sufficient conditions of finite shadowing obtained in [ 7 ] in a form modified to fit our purposes in this paper .
let @xmath76 in case i , and let @xmath77 in case ni .
we assume that there exist two continuous nonnegative functions @xmath78 and @xmath79 defined on @xmath80 such that @xmath81 for any @xmath82 and the conditions ( c1)-(c9 ) stated below are satisfied .
we formulate our conditions not directly in terms of the functions @xmath79 and @xmath78 but in terms of some geometric objects defined via these functions .
fix a positive number @xmath83 and a point @xmath82 and let @xmath84 @xmath85 and @xmath86 set @xmath87 @xmath88 and @xmath89 let @xmath11 be a subset of @xmath90 ( in our basic examples , @xmath11 is a small closed neigborhood of the origin in case i and @xmath11 is a neighborhood of the origin in @xmath90 in case ni ) . conditions
( c1 ) ( c4 ) contain our assumptions on the geometry of the sets introduced above . in conditions ( c2 ) ( c4 ) , @xmath82 , and @xmath91 are arbitrary positive numbers such that @xmath92 .
( c1 ) for any @xmath6 there exists a @xmath93 such that @xmath94 for @xmath95 .
( c2 ) @xmath96 is not a retract of @xmath97 ; ( c3 ) @xmath96 is a retract of @xmath98 ; ( c4 ) there exists a retraction @xmath99 such that if @xmath100 , then @xmath101 . in the next group of conditions
, we state our assumptions on the behavior of the introduced objects and their images under the homeomorphism @xmath0 .
let @xmath102 and let @xmath103 .
we say that condition @xmath104 is satisfied if ( c5 ) @xmath105 ; ( c6 ) @xmath106 ; ( c7 ) @xmath107 ; ( c8 ) @xmath108 ; ( c9 ) @xmath109 .
the same reasoning as in [ 7 ] proves the following statement .
[ prop ] assume that conditions _ ( c2 ) ( c4 ) _ hold .
let @xmath110 .
if @xmath103 and condition @xmath111 is satisfied for any @xmath112 , then there is a point @xmath113 such that @xmath114 thus , to show that @xmath0 has the finite shadowing property in a neighborhood @xmath11 of the origin , it is enough to find functions @xmath78 and @xmath79 that satisfy conditions ( c1 ) ( c4 ) and to show that for any @xmath115 there exists a @xmath116 with the following property : there exists a @xmath7 such that if @xmath102 and @xmath117 , then condition @xmath104 is satisfied . indeed , take any @xmath6 , find a corresponding @xmath118 ( see condition ( c1 ) ) , then take suitable @xmath119 and @xmath83 , and finally find a @xmath7 having the above property .
then , if @xmath110 is a @xmath2-pseudotrajectory of @xmath0 , this pseudotrajectory is @xmath12-shadowed by any point @xmath9 that satisfies inclusions ( [ 2.1 ] ) .
we realize this scheme in the next section considering case i. in case ni , @xmath0 does not have the usual shadowing property , and we have to modify the concept of shadowing ( see sec . [ sec5 ] ) .
now we consider a two - dimensional homeomorphism @xmath120 having a fixed point at the origin and assume that @xmath0 is contracting in the direction of variable @xmath51 and expanding in the direction of variable @xmath75 ( and both the contraction and expansion are not assumed to be hyperbolic ) .
let @xmath121 be the coordinate representation of a point @xmath122 .
in the case considered , we introduce two functions , @xmath123 for such functions , conditions ( c1 ) ( c4 ) are obviously satisfied for any @xmath122 and any @xmath103 .
we first formulate general conditions of finite shadowing for an arbitrary compact subset @xmath11 of the plane and then apply them to our first basic example .
* condition 2 . * for any @xmath124 there exist @xmath125 such that @xmath126 and if @xmath95 , then condition ( c5 ) with @xmath127 is satisfied , @xmath128 latexmath:[\[\label{3.2 } and latexmath:[\[\label{3.3 } _ if @xmath11 is a compact subset of the plane and condition _ 2 _ is satisfied , then @xmath0 has the finite shadowing property in the set _
@xmath11 first we show that condition 2 implies that condition @xmath130 is satisfied for any @xmath95 . since @xmath131 conditions ( [ 3.1 ] ) with @xmath132 and ( [ 3.2 ] )
imply condition ( c6 ) .
similarly , condition ( [ 3.2 ] ) implies condition ( c7 ) . since condition ( [ 3.3 ] ) holds , @xmath133 inequalities ( [ 3.1 ] ) for @xmath134 combined with ( [ 3.18 ] ) show that the image of the boundary of the set @xmath97 under the homeomorphism @xmath0 does not intersect the set @xmath135 .
set @xmath136 it is obviously that @xmath137 . from @xmath138 , @xmath139 and @xmath140 it follows that @xmath141 . because of connectedness of @xmath142 it follows also that @xmath143 and condition ( c8 )
is proved .
finally , condition ( c9 ) follows from condition ( [ 3.3 ] ) .
thus , we have shown that conditions ( [ 3.1 ] ) ( [ 3.3 ] ) imply that condition @xmath130 is satisfied for any @xmath95 .
finally , we note that since @xmath11 is compact and @xmath0 , @xmath78 , and @xmath79 are continuous , the form of conditions ( c5 ) ( c9 ) implies that there exists a @xmath7 depending only on @xmath83 and @xmath144 such that if @xmath95 and @xmath145 , then condition @xmath104 is satisfied .
now theorem 4.1 is a corollary of the proposition stated in the previous section .
* example 2 . *
consider a diffeomorphism ( [ 3.0 ] ) in which @xmath15 are natural numbers and @xmath16 are smooth functions that vanish at the origin together with their jacobi matrices .
first we fix a small closed neighborhood @xmath11 of the origin ( in what follows , we make it as small as our future conditions require ) .
let us assume that if @xmath95 , then @xmath146 and @xmath147 we take the same functions @xmath78 and @xmath79 as above in this section .
it follows from the form of @xmath0 that for any @xmath148 we can find a neighborhood @xmath11 such that @xmath149 thus , if @xmath148 is given , then , for @xmath83 small enough ( and @xmath11 properly chosen ) , condition ( c5 ) is satisfied with @xmath150 .
we take @xmath151 ( in which case we may take @xmath152 in condition ( c5 ) ) and assume that @xmath11 is so small that @xmath153 if @xmath95 and @xmath154 , then @xmath155 thus , condition ( [ 3.2 ] ) is satisfied . now let us check condition ( [ 3.3 ] ) .
if @xmath156 and @xmath157 , then @xmath158 for some @xmath159 . assume , for definiteness , that @xmath160 and estimate , using condition ( [ 3.7 ] ) : @xmath161 @xmath162dt=\ ] ] @xmath163 @xmath164dt > w\geq{\delta},\ ] ] where we take into account that @xmath165 is not identically zero .
this proves condition ( [ 3.3 ] ) ( the case @xmath166 is treated similarly ) . to prove condition ( [ 3.1 ] ) , we consider the case where @xmath167 ; the case where @xmath132 is treated similarly ( note that in both cases , @xmath168 with @xmath159 ) . we assume , for definiteness , that @xmath169 , represent @xmath168 with @xmath159 , and estimate , using condition ( [ 3.5 ] ) : @xmath170 @xmath171 @xmath172dt<{\delta}.\ ] ]
the case @xmath173 is treated similarly .
let us consider as a test perturbation " a monomial @xmath174 in ( [ 3.0 ] ) , where @xmath175 , @xmath176 , and @xmath177 .
let @xmath178 be a point in a neighborhood @xmath11 of the origin with @xmath179 .
taking @xmath180 in ( [ 3.7 ] ) and dividing the result by @xmath181 , we get the following necessary condition : @xmath182 this condition is obviously satisfied in a small @xmath11 if @xmath183 ( and if @xmath184 is small in the case where @xmath185 and @xmath186 ) . if @xmath187 , then the necessary condition looks as follows : @xmath188 , @xmath189 is even , and @xmath190 is odd .
let us write condition ( [ 3.7 ] ) in @xmath11 in the form @xmath191 if @xmath183 ( and if @xmath184 is small in the case where @xmath185 and @xmath186 ) , condition ( [ 3.7 ] ) is obviously satisfied in a small neighborhood of the origin .
in the case where @xmath192 , we get one more necessary condition : @xmath193 . indeed , since @xmath188 , @xmath189 is even , and @xmath190 is odd , it is enough to consider ( [ 3.7 ] ) for @xmath194 .
now let us write ( [ 3.7 ] ) in the form @xmath195 if @xmath196 , set @xmath197 and @xmath198 with small @xmath199 .
we get an inequality of the form @xmath200 which can not be satisfied for all small @xmath201 .
elementary calculations show that if @xmath192 , @xmath188 , @xmath189 is even , @xmath190 is odd , and @xmath193 , then condition ( [ 3.10 ] ) is satisfied in a small neighborhood of the origin .
we can obtain similar conditions if @xmath202 in ( [ 3.0 ] ) is also a monomial .
our methods allow us to estimate the dependence of @xmath2 on @xmath12 in the finite shadowing property for the considered case .
for example , if @xmath203 and @xmath204 in ( [ 3.0 ] ) with @xmath205 and @xmath206 , then the same reasoning as above shows that there exists a neighborhood @xmath11 of the origin and a small number @xmath207 such that if a @xmath208 is given and @xmath62 is a finite set of points in @xmath11 with @xmath209 where @xmath210 , then conditions @xmath211 are satisfied .
this means that @xmath0 has in @xmath11 the finite shadowing property with the following dependence of @xmath2 on @xmath12 : @xmath212 .
in this section , we consider a model example of the diffeomorphism ( [ 4.0 ] ) for which the origin is a nonisolated fixed point ( any point @xmath18 is a fixed one ) .
nevertheless , we show that @xmath0 has an analog of the finite shadowing property if we consider pseudotrajectories @xmath62 with @xmath213 and allow the errors " @xmath214 to depend on @xmath215 .
the errors must be smaller for smaller values of @xmath216 .
such an approach ( in the case of a nontransverse homoclinic point ) has been suggested by s. tikhomirov .
we restrict our consideration to the case of a diffeomorphism @xmath0 of a very simple form ( [ 4.0 ] ) to simplify presentation ( as the reader will see , even this case is not completely trivial ) ; of course , our reasoning can be applied in more general situations .
[ theor5 ] there exists a neighborhood @xmath11 of the origin and a number @xmath207 such that , for any @xmath6 and any pseudotrajectory @xmath218 in @xmath221 that satisfies conditions _ ( [ 4.1 ] ) _ with @xmath222 there exists a point @xmath9 for which inequalities _ ( [ 0 ] ) _ are satisfied . as above , in our proof
we use the approach based on lyapunov functions , but now one of the functions is modified .
we take @xmath223 clearly , these functions are nonnegative and continuous on @xmath80 and they vanish if their arguments coincide . our main goal is to check condition @xmath111 for consecutive points @xmath230 of the pseudotrajectory considered for properly chosen @xmath83 and @xmath144 ; after that , we apply proposition [ prop ] stated in sec .
[ sec3 ] .
thus , to prove our claim , we have to show that there exists a number @xmath226 such that if @xmath238 and @xmath239 , then @xmath241 and @xmath242 in addition , we have to show that if @xmath243 and @xmath239 , then @xmath244 and @xmath245 ( see the formula ( [ 4.01 ] ) for @xmath246 ) .
we note that @xmath256 thus , the projection of @xmath257 to the @xmath51 axis is the segment @xmath258 $ ] , where @xmath259 at the same time , if @xmath260 , then the projection of @xmath240 to the @xmath51 axis is the segment @xmath261 $ ] , where @xmath262 @xmath263 and @xmath264 ( in the above formulas , we note that if @xmath2 and @xmath51 are small , then @xmath265 ) .
let us represent @xmath274 since @xmath275 , we have the estimates @xmath276 if the neighborhood @xmath11 is small ( so that @xmath277 and @xmath278 are small enough ) , we conclude from the inequality @xmath279 ( with @xmath229 fixed above ) that @xmath280 for small @xmath83 , i.e. , @xmath281 does not intersect the upper " component of @xmath271 .
a similar reasoning is applicable to the lower " component .
this proves condition ( c7 ) ( and , combined with inclusion ( [ 4.2 ] ) , condition ( c6 ) ) .
it remains to check conditions ( c8 ) and ( c9 ) .
let us start with condition ( c9 ) .
consider a point @xmath282 of the upper " component of @xmath96 . in this case ,
the @xmath75 component of the point @xmath284 is @xmath285 if @xmath270 , then the projection of the set @xmath240 to the @xmath75 axis is the segment @xmath286 $ ] .
let us represent @xmath287 @xmath288 the estimates @xmath289 and @xmath290 ( which is valid if @xmath278 and @xmath83 do not exceed a small value independent of @xmath51 ) imply that there exists a constant @xmath229 ( inependent of @xmath51 , @xmath75 , and @xmath83 ) such that if @xmath291 , then the projection of the point @xmath284 to the @xmath75 axis does not belong to the segment @xmath292 . | we use lyapunov type functions to find conditions of finite shadowing in a neighborhood of a nonhyperbolic fixed point of a one - dimensional or two - dimensional homeomorphism or diffeomorphism .
a new concept of shadowing in which we control the size of one - step errors is introduced in the case of a nonisolated fixed point .
alexey a. petrov sergei yu . pilyugin ( communicated by lan wen ) |
the bayesian approach to parametric model selection requires the specification of a prior probability distribution over the parameter space .
the jeffreys prior , which is proportional to the square root of the determinant of the fisher information computed in the parameter space , has been shown to be the uniform prior over all _ distributions _ indexed by the parameters in a parametric family @xcite .
geometrically , its integral over a region of the parameter space computes a volume that essentially measures the fraction of statistically distinguishable probability distributions within that region @xcite . in this interpretation , the jeffreys prior distribution @xmath0 where @xmath1 simply measures the fractional volume of the small element @xmath2 relative to total volume of the parametric manifold @xmath3 . here
@xmath4 is the fisher information on the parameter space @xmath5 and @xmath2 is the standard riemannian volume element on @xmath6 .
the volume @xmath7 also appears in the minimum description length ( mdl ) approach to model selection @xcite , conceptually because it effectively measures how many different distributions are describable by different parameter choices .
an important difficulty in applying the mdl approach to model selection occurs when the parameter space is noncompact and the volume @xmath7 diverges . in this case , from the bayesian perspective , a uniform prior on the parameter space does not exist , while from the mdl perspective the number of models that might be describable diverges , leading to problems with the definition of the description length .
of course the parameter space can be cut off by hand , but unless the choice of cut - off is well founded , it can lead to artifacts in the comparison of different model families @xcite . unfortunately in many practical problems the parameter space _ is _ noncompact and @xmath7 diverges .
for example , in astrophysics , the detection of exoplanets depends on a model of the light coming from the occluded star .
this model will contain a non - compact direction representing the orbital period of the planet
see , e.g. , @xcite . for examples from psychophysics see , e.g. , @xcite . in this note
we argue that merely specifying the experimental set - up before the measurement of any actual data influences the prior distribution on the parameter space .
this occurs because , given the finite number of measurements in any experiment , many of the probability distributions indexed by a parametric manifold will be statistically indistinguishable . in cases where the parameter space is noncompact ,
the uniform prior conditioned on the experimental setup can thus become well - defined . in the geometric language of @xcite , the volume that measures the number of probability distributions in the parametric family that are statistically distinguishable given a _
finite _ number of measurements can be finite even if the parameter space is non - compact . in effect , specifying the experimental set - up can render the parameter space compact .
our results illustrate how the choice of experimental set - up influences the measure on the parameter space of a model , thereby affecting which model is regarded as most ` correct ' . in section [ general - argument ]
we briefly review the computation of posterior probabilities , and consider the effect of conditioning on the experimental set - up on the parameter space measure . in section [ sec :
example ] we apply these considerations to a physical problem : the analysis of light - curves of stars with orbiting planets . in this example
we see that the volume of the parameter space is rendered effectively finite after the experimental set - up is specified .
[ general - argument ] suppose one is interested in some physical phenomenon , and has made @xmath8 relevant measurements : @xmath9 .
further suppose that there are two different parametric models , @xmath10 and @xmath11 , that aim to describe the phenomenon in question .
the basic question to be answered is which of the two models is the better one , considering the experimental data @xmath12 .
the probability - theoretic answer to this question is to compute the posterior probabilities @xmath13 and @xmath14 , which we can write using the bayes rule as @xmath15 where @xmath16 is the vector of variables parametrising @xmath10 , and @xmath17 is the volume form associated to the measure on the parameter space , which we will define shortly .
a corresponding expression can also be written for @xmath14 .
since we wish to compare @xmath13 and @xmath14 , we can ignore the common factor @xmath18 , and we will assume @xmath19 and drop this factor as well .
thus the only remaining ingredient to be defined is the volume form @xmath17 ; we simply quote the result from @xcite : the volume form that gives equal weight to all statistically distinguishable distributions in the parametric family is @xmath20 where @xmath21 is the _ fisher information matrix _ , defined as the second derivative of the kullback
leibler distance @xmath22 : @xmath23 where @xmath24 is the integration measure over the sample space @xmath25 , and @xmath26 is the distribution function associated to the values of the parameters @xmath27 .
now we have defined everything needed to compute the posterior probabilities , and we illustrate the formalism by applying it to the analysis of light - curves . using this , we can compute the fisher information matrix by computing the kullback
leibler distance between two nearby points and taylor expanding : @xmath28 on the third line , the terms linear in @xmath29 vanish , as exchanging the order of integration and derivation , the integral of @xmath29 will yield a constant 1 , which then differentiates to zero . the measure ( [ measure ] ) is independent of the experimental data @xmath12 and is constructed under the assumption that the entire sample space can be measured by the observer .
however , in real experiments , instrumental and design limitations only allow observation of some subset @xmath30 of the sample space .
thus an observation either results in no detected outcome , or in a measurement @xmath31 .
thus the effective predicted distribution of measured outcomes is not the @xmath32 , but rather @xmath33 where @xmath34 .
we will argue that if the models in the asymptotic regions of a noncompact parameter space differ in their predictions mostly outside the observable region @xmath30 , the fisher information for the effective distributions ( [ effdist ] ) can decay sufficiently quickly to render the volume @xmath35 finite . in this section
we will give one set of sufficient conditions for this to happen and in sec .
3 we will give a detailed example . consider a model , specified by parameters @xmath36 , and a distribution @xmath37 , with @xmath38 .
we will slightly simplify notation simply referring to the distribution as @xmath32 and understanding the implicit parameter dependence .
let us use spherical coordinates in the parameter space @xmath6 with @xmath39 being the radial coordinate , i.e. @xmath40 . also consider an experimental set - up that can only make measurements inside some compact region @xmath41 .
thus , the probability of no measurement being registered by this experiment is @xmath42 .
our first assumption is a smoothness condition , so that inside the region @xmath30 the distribution does not fluctuate too much as one approaches the asymptotics of parameter space : @xmath43 where @xmath44 goes to zero as @xmath39 goes to infinity ; we will later specify the exact scaling needed .
intuitively , this condition says that as the parameter @xmath45 , the models do not differ too much inside the observable part of the sample space @xmath30 .
this allows us to estimate @xmath46 where vol@xmath47 denotes the volume of the compact region @xmath30 .
secondly we assume that inside @xmath30 , the distributions @xmath32 do not decay too quickly as @xmath45 .
intuitively , since any experiment will only measure a finite amount of data ( say @xmath8 points ) , if the probability of a single measurement lying inside @xmath30 is significantly less than @xmath48 , then the experimental set - up will not detect anything .
thus we will require @xmath49 where again we will later specify the scaling of @xmath50 with @xmath39 .
decays too quickly as @xmath45 , then the models in the asymptotic region of the parameter space make no measurable predictions for experiments designed with a finite number of measurements .
the example in the sec .
3 will illustrate such a scenario . ] using these assumptions , we can establish an upper bound for the fisher information ( [ eq : fishertheor ] ) : @xmath51 thus the determinant of the fisher information scales as @xmath52 and for the integral @xmath7 to be finite one must have suppression stronger than @xmath53 .
thus the integral converges if @xmath54 is suppressed more strongly than @xmath55 from the experimental set - up one can estimate how @xmath50 scales with @xmath39 , which then determines how @xmath44 needs to scale for the integral to converge .
this is thus a sufficient condition for rendering the parameter space effectively finite .
it is worth stressing that , following the above analysis , any method of deciding the validity of a model is impacted by the choice of the experiment in a completely computable way , and this should be taken into account when designing experiments .
[ sec : example ] consider a star orbited by a planet so that the planet periodically passes between the star and earth .
the light output ( light - curve ) of such a star is a constant line , with a small periodic dip when the planet is eclipsing part of the star .
one model for such a light - curve was proposed in @xcite as @xmath56 , \label{eq : curve}\ ] ] where @xmath57 an example light - curve is shown in figure [ fig : curve ] ; @xmath58 is the period of the planet ; @xmath59 is the duration of the transit , i.e. how long the planet eclipses the star ; @xmath60 is the depth of the dip in the curve ; @xmath61 is the total observed brightness of the star ; and @xmath62 is a phase parameter specifying when the planets transit occurs . finally , @xmath63 is a constant parameter specifying the sharpness of the edges of the light - curve , expected to be fairly large as the transition between transit / no - transit is relatively quick .
the assumption @xmath64 greatly simplifies our analysis , and is not physically very restrictive .
the parameter space for this model is clearly non - compact as @xmath58 can range to infinity .
however , we will argue that the space is effectively rendered compact after the experimental set - up is specified . to be precise
, the parameter space is to be a constant , not a parameter . ] : @xmath65 , \quad \tau \in [ 0,t ] , \quad \eta \in [ 0,\delta t ] , \quad b \in [ 0,b_{max}],\ ] ] where @xmath54 is a small number that we will estimate , and the maximal brightness @xmath66 is naturally given by the brightness of sirius , the brightest star visible from earth . assuming a circular orbit as in figure [ fig : exo ] , the ratio of the transit time to the period of the planet is given by @xmath67 for the currently known transiting exo - planets this ratio is around @xmath68 @xcite , although for a typical system one expects it to be smaller as large planets orbiting close to the star are easier to observe , which favors largest values of the ratio . for an elliptical orbit
, the answer will differ by an @xmath69 factor , but will have the same dependence on @xmath70 .
thus , @xmath59 will always be a small fraction of @xmath58 . with an average distance @xmath71 .
] now we can write down the probability density for measuring values @xmath72 for the light - curve at times @xmath73 with the light - curve specified by parameters as @xmath74 where we have assumed that the uncertainty in each measurement is gaussian , and further we have chosen the standard deviation to be equal for all measurements for simplicity . using ( [ gauss ] ) in the formula ( [ eq : fishertheor ] ) , we see that the the integrals in the fisher information are gaussian in @xmath75 ; thus we can compute them analytically to get @xmath76 this is our key formula , and we shall spend the next subsection analysing its properties .
[ finiteness ] we now wish to apply the general arguments of section [ general - argument ] to the exo - planet system .
consider an experimental set - up that can barely measure two periods , and then consider shortening the experiment slightly so that only one dip is detected ; this is depicted in figure [ fig : curve ] . to be precise , the shorter set - up measures the beginning and end of a transit at @xmath77 and @xmath78 , @xmath79 points in between , and @xmath80 points after the transit .
the longer set - up makes measurements at the same times , and additionally at times @xmath81 and @xmath82 , detecting the second transit . in the next subsections
we will show that @xmath83 , indicating that detecting the second dip is of fundamental importance to experimental design ; without the second dip the experimental set - up ca nt differentiate models with large enough @xmath58 .
this renders the parameter space effectively finite , as an experiment can not differentiate between models that have period @xmath58 larger than the duration of the experiment .
[ sec : estimate ] in this subsection we will give an estimate for the magnitude of the determinant of the fisher information , and show how it is affected by the inclusion of the second transit in the data .
in subsequent subsections we will exactly compute the determinant for a few specific experimental set - ups . from ( [ eq : fisher ] ) and the definition of a determinant
, we see that in each term of the determinant each parameter @xmath85 appears exactly twice in the derivatives , i.e. each term is of the form @xmath86 as a rough estimate of the size the determinant , we investigate how large terms of this type can be .
the derivatives are @xmath87 from ( [ eq : f ] ) we see that @xmath88 only when @xmath89 , again assuming large @xmath63 .
this tells us that the measurements that contribute most to the fisher information are the ones on the edges of the dips ; for our current purposes it is sufficiently accurate .
] , i.e. at times @xmath90 and @xmath82 in figure [ fig : curve ] .
we write the condition @xmath91 as @xmath92 and note that the ratio of transit time to period is very small , @xmath93 .
this gives us the solutions @xmath94 where @xmath79 is an integer indexing the number of the dip , with @xmath95 denoting the solitary dip if only one is present in the data .
we wish to estimate the ratio of the determinants of the fisher information by an order of magnitude estimate @xmath96 where both the numerator and the denominator are of the form ( [ generic ] ) , and according to the argument above the maximal contributions come from the edge measurements . from ( [ dd]-[deta ] )
we see that the derivatives with respect to @xmath97 and @xmath61 are all periodic at the edges : @xmath98 for @xmath99 , and thus will cancel in the ratio ( [ ratio ] ) .
it is crucial that @xmath100 , however , is not periodic due to the second term in ( [ dt ] ) . at the first dip , @xmath101 , we expand ( [ dt ] ) to find @xmath102 while at the second dip , @xmath103 , the contribution is @xmath104 ignoring signs that are irrelevant for this estimate .
thus we see that the fisher information increases strongly as the second dip is included : @xmath105 where we ignored order one coefficients .
this is an explicit example of how our arguments from section [ general - argument ] work for a realistic model : when an experimental set - up does not have the capability to detect two dips , it becomes impossible to determine the period , and consequently the fisher information is very small ( or vanishing ) compared to an experiment that is able to detect two dips and determine the period more accurately . for any given experiment of finite duration @xmath106
, the fisher information will decline with @xmath58 when @xmath107 effectively rendering the parameter space compact .
to verify our claim that the parameter space is really rendered compact we need to show that det @xmath109 strongly enough as @xmath58 is taken to infinity .
it is easy enough to find the @xmath58-scaling of the derivatives ( [ dt]-[deta ] ) ; @xmath100 scales as @xmath110 , while the others stay finite in the large @xmath58 limit .
thus , as seen from ( [ generic ] ) , the determinant will scale as @xmath111 which shows that that the parameter space measure vanishes fast enough for large @xmath58 to render the parameter space volume finite .
[ sec : dets ] while the order of magnitude estimate of the previous subsection offers an intuitive reason as to why the fisher information decreases sharply when the number of peaks detected falls below two , it is still instructive to explicitly compute the determinant in a few experimental set - ups .
[ [ detecting - two - dips ] ] detecting two dips : + + + + + + + + + + + + + + + + + + + let us first consider the case @xmath113 from section [ finiteness ] , i.e. measurements at times indicated in figure [ fig : curve ] . using the derivatives ( [ dt]-[deta ] ) one can write down the fisher information matrix ( [ eq : fisher ] ) as @xmath114 where for brevity we defined @xmath115 in computing this matrix we used that @xmath116 for @xmath117 , which is true up to corrections of order @xmath118 , as seen from ( [ eq : f ] )
; for this reason one does not need to specify the exact times of the @xmath79 measurements during the dip , or the @xmath80 measurements outside the dip , as up to @xmath118 corrections they all contribute equally .
the determinant of the fisher information is simple , @xmath119 this result explains the subtlety referred to earlier : although measurements at the edges contribute the most to the fisher information , if one only has measurements at the edges ( @xmath120 ) the fisher information actually vanishes .
physically this is easy to interpret , as only measuring the edges @xmath121 will yield four points lying on a line , and thus they can not be used to determine any information about the curve ; other data points are needed to ` anchor ' the data . [ [ detecting - only - one - dip ] ] detecting only one dip : + + + + + + + + + + + + + + + + + + + + + + + similarly one can compute the fisher information in the ` short ' experimental set - up , where measurements are made at the same times as before , except not at @xmath81 and @xmath82 .
this yields @xmath122 and perhaps surprisingly the determinant vanishes : det@xmath123 , up to tiny @xmath118 corrections .
this indicates that the estimate in section [ sec : estimate ] was an overestimate : terms in the determinant of @xmath124 are of the magnitude estimated , but the determinant is arranged in such a way that the terms cancel to a high accuracy , and the compactness of the parameter space is strengthened .
our analysis has shown how the specification of an experimental design affects the measure on model parameter spaces in mdl model selection ( or equivalently the prior probability distribution on parameters in the bayesian approach ) .
interestingly , the finite number of measurements within a bounded sample space in any practical experiment can effectively render a non - compact parameter space compact thereby leading to a well - defined prior distribution ( [ measure ] ) .
our analysis could be turned around to design experiments to discriminate well between models in some chosen region of the parameter space by ensuring that the fisher information ( [ eq : fisher ] ) is large in the desired region .
it would also be useful to determine general conditions under which experimental design effectively makes model parameter spaces compact , perhaps following the arguments of sec . 2 .
this paper was written in honor of jorma rissanen s 75th birthday and his many seminal achievements in statistics and information theory .
vb and kl were partially supported by the doe under grant de - fg02 - 95er40893 , and kl was also partly supported by a fellowship from the academy of finland .
vb was also partly supported as the helen and martin chooljian member at the institute for advanced study .
v. balasubramanian , `` statistical inference , occam s razor and statistical mechanics on the space of probability distributions , '' [ arxiv : cond - mat/9601030 ] , v. balasubramanian , `` a geometric formulation of occam s razor for inference of parametric distributions , '' [ arxiv : adap - org/9601001 ] .
myung , v. balasubramanian and m.a .
pitt , `` counting probability distributions : differential geometry and model selection '' , proceedings of the national academy of science , 97(21 ) 1117011175 , 2000 . | we apply the minimum description length model selection approach to the detection of extra - solar planets , and use this example to show how specification of the experimental design affects the prior distribution on the model parameter space and hence the posterior likelihood which , in turn , determines which model is regarded as most ` correct ' .
our analysis shows how conditioning on the experimental design can render a non - compact parameter space effectively compact , so that the mdl model selection problem becomes well - defined . |
with the observation of a standard model - like higgs boson @xcite at the large hadron collider ( lhc ) , our understanding of electroweak symmetry breaking has been significantly enhanced .
the attention of experiments at the lhc now turns to establishing the properties of this new resonance .
the possibility that the resonance is actually a composite bound state remains open , and in this work we study a top partner search in the context of composite higgs scenarios @xcite .
the discovery of such top partners would spectacularly elucidate the mechanism by which the scalar resonance mass is stabilized at the electroweak scale .
although the higgs boson can be a generic composite bound state , we focus on the realization of the higgs boson as a pseudo nambu - goldstone boson ( pngb ) of the coset @xmath1 , @xmath2 is an approximate global symmetry and @xmath3 is an unbroken subgroup .
we assume partial compositeness @xcite to evade flavour physics constraints @xcite . here , the standard model ( sm ) fermions obtain masses via mixing with composite bound states from the strongly coupled sector .
since sm fermions arise as admixtures of the elementary and the corresponding composite bound states , there are necessarily accompanying heavy excitations with the same sm quantum numbers , so - called top partners .
they belong to a representation of the unbroken subgroup @xmath3 . following ref .
@xcite we adopt a minimal setup for @xmath4 and @xmath5 .
we assume @xmath6 to be a completely composite chiral state and only @xmath7 to be partially composite . in this short article
we focus on evaluating the lhc sensitivity for discovering or excluding a top partner belonging to the singlet @xmath8 of the unbroken so(4 ) .
the operators @xmath9 induced by the interactions of the fermions in the strong sector are assumed to transform in the representation @xmath10 of the strong - sector fermions transform in the representation @xmath11 . ] the interaction lagrangian for @xmath12 is given by @xmath13\end{aligned}\ ] ] and the mass eigenvalue of the top partner @xmath12 is approximately given by @xmath14,\ ] ] where @xmath15 and @xmath16 . as a result ,
the phenomenology of @xmath12 is determined by four parameters : @xmath17 .
@xmath18 is the mass of the top partner , @xmath19 controls the mixing between the composite and elementary states , @xmath20 is an @xmath21 parameter associated with the interactions of @xmath6 and @xmath22 is the symmetry breaking scale of the strong sector . by requiring the four parameters to conspire to give the observed top quark mass , one degree of freedom in this parameter space
can be removed , i.e. we choose @xmath23 . in the following quantitative analysis we also choose @xmath24 to ensure compatibility with electroweak precision tests @xcite .
the cross section also depends on the parameter @xmath20 , as described in ref @xcite .
here we choose @xmath25 for a @xmath12 mass of 700gev .
pair production of fermionic top partners has been thoroughly searched for at the lhc @xcite . in composite higgs models ,
light top partners are well motivated @xcite but so far masses less than 687 782gev have been excluded depending on the branching ratio to different final states . in this work ,
we focus instead on single production : @xmath26 .
the cross section for single production is smaller than that for pair production at low masses of the top partner .
however , depending on the weak coupling to the top partner , single production can begin to have a larger cross section in the mass - range of 6001000gev @xcite .
thus , well - designed searches for single production can potentially extend the mass - reach of the lhc experiments . singlet top partners decay to @xmath0,@xmath27 or @xmath28 in the approximate ratio 2:1:1 respectively , and the significant branching fractions to @xmath27 and @xmath28 have attracted previous attention@xcite . in particular , multi - leptonic final states arising from these decays make for searches with low background , albeit at a relatively low overall signal efficiency . here , we have chosen to focus on the most abundant expected decay : @xmath29 , with subsequent decay of the @xmath30 boson to leptons . trading a clean final state for
a larger signal efficiency can be beneficial for large top partner masses @xcite .
the present experimental limits are weakest for large values of the @xmath31 branching ratio @xcite . we demonstrate that , by requiring a forward jet and exploiting jet substructure to suppress backgrounds from top - quark production , this relatively neglected channel can become a promising one for discovery of single top - partner production at the lhc .
candidate mass used in the statistical analysis is shown for 20fb@xmath32 integrated luminosity at the 8tev lhc .
the shape of the sm background from @xmath30+jets , @xmath33 and single top production are compared to the shape of the expected signal.,scaledwidth=43.0% ] signal events were simulated using 5 @xcite interfaced with 8.185 @xcite for parton - shower and fragmentation .
for a top partner with mass 700gev , the leading - order cross sections from are 0.163pb ( 0.957pb ) for @xmath34tev ( 14tev ) for our parameter choices is 0.55 ] .
background samples of and @xmath30+jets were generated with up to two and four additional partons using @xcite 2.1.0 .
single top quark production in the @xmath35-channel was modelled , in a four - flavor scheme @xcite , using @xcite showered with 8.185 .
the production cross section of the sample was scaled to the next - to - next - to - leading - order prediction including resummation of next - to - next - to - leading logarithmic soft gluon terms @xcite calculated using the nnpdf2.3 pdf set @xcite .
approximate corrections to next - to - leading order for the signal cross - section were calculated using the single - top cross section from , with the top - mass set to the mass of the top - partner compared to the equivalent leading order cross - section calculated from .
these corrections to the signal were approximately 14(7)% for 8(14)tev .
analysis of the samples of simulated events is performed on the stable final - state particles using rivet 2.1.1 @xcite .
the event selection was designed to select singly produced @xmath12 quarks with subsequent @xmath36 decay .
for leptonically decaying @xmath30-bosons , the signal event topology is characterized by a charged lepton , missing transverse momentum , a high- @xmath37-tagged central jet and a forward jet .
final - state electrons are corrected for energy loss due to photon emission by combining them with all final - state photons within @xmath38 , where @xmath39 is the distance in the @xmath40@xmath41 plane between the electron and the photon .
charged leptons are required to be isolated from other particles ) , @xmath42 , with @xmath43 should be less than 10% of the electron .
similarly for muons , @xmath42 for all particles with @xmath44 is required to be less than 6% of the muon , but in addition , @xmath42 is required to be less than @xmath45 @xmath46 . ]
. the typical fiducial acceptance of the atlas detector @xcite is mimicked by only using electrons with @xmath47 or @xmath48 , and muons with @xmath49 .
small - radius jets are clustered from all final - state particles with @xmath50 , except muons and neutrinos , using the anti-@xmath51 algorithm @xcite with a radius parameter of 0.4 .
only jets with @xmath52 25gev for @xmath53 or @xmath52 35gev for @xmath54 are used . in order to account for electrons misidentified as small - radius jets , any small - radius jet ( @xmath55 ) with @xmath56
are discarded .
large - radius jets are clustered in a similar manner , but with a radius parameter of 1.0 .
these large - radius jets were required to have @xmath52 100 gev and @xmath571.2 .
jets are referred to as @xmath37-jets if their constituents contain a bottom flavoured hadron or the decay products of at least one such hadron .
the typical performance of experimental jet @xmath37-tagging was mimicked by ascribing a 0.7 probability for tagging @xmath37-jets and a 0.01 mis - tag rate for @xmath37-tagging jets from other sources @xcite .
the missing transverse momentum , , is the magnitude of the total transverse momentum calculated from all final - state particles with @xmath58 except neutrinos .
the missing transverse momentum direction is defined as being directly opposite to this same total transverse momentum .
events are preselected by requiring exactly one electron or muon with @xmath52 25gev , @xmath52 20gev and plus transverse mass , where @xmath59 is the azimuthal angle between the charged lepton and the missing transverse momentum direction .
] of 60gev .
events are also required to contain at least two small - radius jets and at least one large - radius jet with @xmath60gev .
after this selection , the dominant background processes are @xmath30+jet production and @xmath33 production .
further cuts were applied with the goal of optimizing the statistical significance @xmath61 , defined as the number of 700gev top - partner signal ( @xmath62 ) events divided by the square root of the number of background ( @xmath63 ) events .
the @xmath64 jet background is suppressed by requiring that a @xmath37-tagged small - radius jet is matched to the large - radius jet@xmath41 distance of 0.8 . ] .
the @xmath33 background is also suppressed by requiring that the large - radius jet have @xmath65gev and mass @xmath66gev .
the latter requirement exploits the fact that high @xmath67 large - radius jets in @xmath33 tend to have a genuinely large mass since they contain at least two decay products from @xmath68 , as illustrated in figure [ fig : illustration ] . to ensure that the lepton is not within the active area of the large - radius jet , it is required that @xmath69 .
the expected isolation of the b - tagged jet from the @xmath12 decay can be exploited further by vetoing events with extra central jets ( those with @xmath70 ) above a certain threshold .
theoretical uncertainties when vetoing extra jets decrease for higher thresholds , i.e. the resummation of contributions @xmath71 becomes increasingly important for lower cut - off scales @xcite .
hence , this cut was optimized by looking for both the largest @xmath72 and the largest jet threshold .
the optimal threshold was found to be 75gev .
finally , events must have at least one jet with @xmath52 35gev and @xmath54 , consistent with the single top partner production mechanism . in table
[ tab : cutflow ] , the number of expected events for 20fb@xmath32 integrated luminosity at the 8tev and 14tev lhc is shown at the preselection stage , after requiring the central jet - veto , and after the final selection .
the selection cuts lead to a @xmath73 of 0.22 ( 0.20 ) and @xmath72 of 2.5 ( 5.65 ) for a 700gev @xmath12 at 8tev(14tev ) .
the addition of the forward jet requirement leads to modest improvements in @xmath72 but more than doubles @xmath73 , it can thus be very useful in the case of large experimental uncertainties on the background normalization . to search for @xmath12 production ,
a candidate @xmath12-mass constructed from the charged lepton and the @xmath37-jet and neutrino candidates is used .
the @xmath37-tagged small - radius jet , that is matched to the large radius jet , is used as the @xmath37-jet candidate .
the neutrino candidate is constructed from the and charged lepton by imposing a @xmath30-mass constraint on the lepton+ system @xcite .
the resulting @xmath12-candidate mass distribution is shown for signal and backgrounds in figure [ fig : tmass ] for 20fb@xmath32 integrated luminosity at the 8tev lhc .
the signal distributions peak close to the @xmath12 mass while background distributions turn over at lower masses . [ cols="^,^,^,^,^,^,^ " , ] a comparison was made with the strategy presented in ref .
@xcite , where the authors address the @xmath0 final state by using a high lepton , large summed transverse momentum ( h@xmath74 ) and the masses of combinations of final state particles . after running both selections on two signal samples with masses of 700 and 800 gev ,
it was found that although the @xmath75 for both analyses is comparable , our approach accepts between 50% and 70% more events .
this will either allow for additional cuts to purify the final state or in an improved reach of the resonance mass or smaller production cross section .
in order to estimate whether a singly - produced top partner can be excluded with data collected at the lhc , we use a binned log - likelihood hypothesis test as described in ref .
the results of this procedure , as a function of integrated luminosity for 8tev and 14tev lhc data , are presented in figure [ fig : cls700at8 ] , which shows the confidence level at which single production of a 700gev @xmath12 quark can be excluded in the absence of signal .
these results are shown both before and after the forward - jet requirement and for different assumed levels of systematic uncertainty on the backgrounds : 0% , 5% and 10% .
it can be seen that the sensitivity is improved by the forward - jet requirement and that the full selection is sufficient to exclude the single - top partner production using 8tev data even for the most pessimistic assumed systematic uncertainty .
although this systematic treatment is rather simplistic we expect the search strategy to retain sufficient sensitivity to exclude the 700gev @xmath12 if deployed in a real experimental analysis . for 14tev data ,
the exclusion of the full 8tev data set is already surpassed at an integrated luminosity of 5@xmath76 .
this is the first study to demonstrate a simple and feasible strategy for discovering single production of a top partner in the @xmath29 channel .
the study shows that use of a forward jet and the requirement of a low - mass central large - radius jet and central jet veto are powerful tools for rejecting the sm background processes whilst retaining acceptable signal efficiency .
this analysis is indicative that single production of 700gev top partners could be excluded already at the 8tev lhc .
the mass reach of searches at higher - energy lhc running is likely to be significantly extended .
this study considers @xmath12 production within a specific composite higgs scenario .
however , the analysis is generic and its applicability to top partner searches in other models can be inferred from the production cross sections in our chosen model and from the exclusion limits shown in figure . [
fig : cls700at8 ] .
we reserve study of @xmath77 and @xmath78 for a future work , noting that in most searches the extra efficiency in a dominant decay channel outweigh signal purity in the selected sample .
_ acknowledgements .
_ we thank andy buckley and karl nordstrom for assistance with rivet , frank siegert for advice on configuration , and andrew pickford for computing support .
we thank oleskii matsedonskyi , andrea de simone and andrea wulzer for providing valid parameter points for the composite higgs scenarios under consideration .
we thank tony doyle for useful discussion .
the work of jf and dk is supported by the uk science and technology facilities council ( stfc ) .
the work of ng is supported by the scottish universities physics alliance ( supa ) .
g. aad _ et al . _ [ atlas collaboration ] , phys .
b * 716 * , 1 ( 2012 ) [ arxiv:1207.7214 [ hep - ex ] ] .
s. chatrchyan _ et al . _
[ cms collaboration ] , phys .
b * 716 * , 30 ( 2012 ) [ arxiv:1207.7235 [ hep - ex ] ] .
h. georgi and d. b. kaplan , phys .
b * 145 * , 216 ( 1984 ) .
d. b. kaplan and h. georgi , phys .
b * 136 * , 183 ( 1984 ) .
d. b. kaplan , h. georgi and s. dimopoulos , phys .
b * 136 * , 187 ( 1984 ) .
m. j. dugan , h. georgi and d. b. kaplan , nucl .
b * 254 * , 299 ( 1985 ) .
r. contino , y. nomura and a. pomarol , nucl .
b * 671 * , 148 ( 2003 ) [ hep - ph/0306259 ] .
k. agashe , r. contino and a. pomarol , nucl .
b * 719 * , 165 ( 2005 ) [ hep - ph/0412089 ] .
r. contino , t. kramer , m. son and r. sundrum , jhep * 0705 * , 074 ( 2007 ) [ hep - ph/0612180 ] .
d. b. kaplan , nucl .
b * 365 * , 259 ( 1991 ) .
r. barbieri , b. bellazzini , v. s. rychkov and a. varagnolo , phys .
d * 76 * , 115008 ( 2007 ) [ arxiv:0706.0432 [ hep - ph ] ] . g. f. giudice , c. grojean , a. pomarol and r. rattazzi , jhep * 0706 * , 045 ( 2007 ) [ hep - ph/0703164 ] . c. grojean , o. matsedonskyi and g. panico , jhep * 1310 * , 160 ( 2013 ) [ arxiv:1306.4655 [ hep - ph ] ] . g. aad _ et al . _ [ atlas collaboration ] , phys.lett .
b * 718 * ( 2013 ) 1284 - 1302 s. chatrchyan _ et al . _
[ cms collaboration ] , phys.lett .
b * 729 * ( 2014 ) 149 - 171 o. matsedonskyi , g. panico and a. wulzer , jhep * 1301 * , 164 ( 2013 ) [ arxiv:1204.6333 [ hep - ph ] ] . j.a .
aguilar - saavedra _ et al .
_ phys.rev .
d * 88 * ( 2013 ) 094010 j. mrazek and a. wulzer phys.rev .
d * 81 * ( 2010 ) 075006 .
t. gleisberg , s. hoeche , f. krauss , m. schonherr , s. schumann , f. siegert and j. winter , jhep * 0902 * ( 2009 ) 007 . j. m. campbell , r. frederix , f. maltoni and f. tramontano , phys . rev .
lett . * 102 * , ( 2009 ) 182003 ; r. frederix , e. re and p. torrielli , jhep * 1209 * ( 2012 ) 130 m. cacciari et al . ,
* 710 * ( 2012 ) 612 - 622 , [ arxiv:1111.5869 [ hep - ph ] ] ; p. brnreuther et al .
* 109 * ( 2012 ) 132001 , [ arxiv:1204.5201 [ hep - ph ] ] ; m. czakon and a. mitov . , jhep * 1212 * ( 2012 ) 054 , [ arxiv:1207.0236 [ hep - ph ] ] ; m. czakon , a. mitov .
, jhep * 1301 * ( 2013 ) 080 , [ arxiv:1210.6832 [ hep - ph ] ] ; m. czakon , p. fiedler , a. mitov .
* 110 * ( 2013 ) 252004 [ arxiv:1303.6254 [ hep - ph ] ] ; m. czakon and a. mitov , arxiv:1112.5675 [ hep - ph ] c. f. berger , c. marcantonini , i. w. stewart , f. j. tackmann and w. j. waalewijn , jhep * 1104 * , 092 ( 2011 ) [ arxiv:1012.4480 [ hep - ph ] ] ; a. banfi , p. f. monni , g. p. salam and g. zanderighi , phys .
* 109 * , 202001 ( 2012 ) [ arxiv:1206.4998 [ hep - ph ] ] . | fermionic top partners are a feature of many models of physics beyond the standard model .
we propose a search strategy for single production of top partners focussing specifically on the dominant decay to @xmath0 .
the enormous background can be reduced by exploiting jet substructure to suppress top - pair production and by requiring a forward jet .
this simple strategy is shown to produce a sensitive search for single top - partner production , in the context of composite higgs models , that has competitive mass reach with existing experimental searches for top - partner - pair production at the 8tev lhc . |
a major impediment that prevents pressurized water reactors ( pwrs ) from operating at higher duty and longer cycles is the accumulation of boron within metal oxide scales that deposit on the upper spans of fuel assemblies .
boron , in the form of boric acid ( h@xmath2bo@xmath2 ) , is added to pwr coolant to control the neutron flux while lithium hydroxide ( lioh ) is also dosed to control the acidity of the coolant and to reduce corrosion . because of the large neutron absorption cross section of @xmath3b , a small amount of accumulated b is sufficient to cause an abnormal decrease in the neutron flux , which shifts the power output toward the bottom half of the reactor core .
this phenomenon , known as axial offset anomaly ( aoa ) , has been observed in high - duty reactors that run long fuel - cycles . in extreme form , aoa can decrease the reactor shutdown margin sufficiently to force major power reduction leading to substantial economic losses @xcite .
aoa modeling efforts traditionally assume that boron deposition within the metal oxide scales ( which are commonly referred to as crud , an acronym for chalk river unidentified deposits ) occurs predominantly through precipitation of lithium borate compounds such as libo@xmath0 , li@xmath0bo@xmath4 , li@xmath0b@xmath1o@xmath4 @xcite .
although the retrograde solubility of these borates could explain the `` lithium return '' experienced during plant shutdown , they have never been observed experimentally in pwr crud .
using mssbauer spectroscopy together with xrd on crud scrapes recovered from high duty pwrs , sawicki identified the precipitation of ni@xmath0febo@xmath5 as a possible mechanism for b deposition @xcite .
mesoscale crud models developed by short _
_ , assume that supersaturation of boric acid leads to precipitation of boron trioxide ( b@xmath0o@xmath2 ) within the crud @xcite . in recent work
, we combined _ ab initio _ calculations with experimental formation enthalpies to investigate the incorporation of b into the structure of nickel ferrite ( nife@xmath0o@xmath1 , nfo ) as a potential new mechanism for b deposition within crud @xcite . assuming solid - solid ( and solid - gas ) equilibrium between nickel ferrite and elemental reservoirs of fe , ni ,
b ( and o@xmath0 gas ) we found that it is thermodynamically favorable for b to form secondary phases with fe , ni , and o ( _ e . g. _
b@xmath0o@xmath2 , fe@xmath2bo@xmath5 , and ni@xmath2b@xmath0o@xmath6 ) instead of entering the nfo structure as a point defect .
building on previous works @xcite , the present study attempts to deal with the same question , however , here the defect formation energies are evaluated assuming solid - liquid equilibrium between nfo and the surrounding aqueous solution of ni , fe and dissolved boric acid ( h@xmath2bo@xmath2 ) . to set up solid - liquid equilibrium ,
the chemical potentials of individual aqueous species are defined as a function of temperature , pressure , and concentration and are linked to the chemical potentials of the ionic species in solid .
this new scheme allows for the evaluation of defect formation energies under conditions that are specific to operating nuclear pwrs .
the approach is quite general and applicable to a large variety of solids in equilibrium with aqueous solutions .
the first - principles calculations required for the present study have been carried out within the density functional theory ( dft ) using the same computational parameters and crystal models that are specified in ref . .
the formation of a defect in a crystalline solid can be regarded as an exchange of particles ( atoms and electrons ) between the host material and chemical reservoirs .
the formation energy of a defect _
d _ in charge state _
q _ can be written as @xcite @xmath7 in eq .
, @xmath8 and @xmath9 are the total energies of the defect - containing and defect free solids , calculated within the dft .
the third term on the right side of eq . represents the change in energy due to the exchange of atoms between the host compound and the chemical reservoir , where @xmath10 is the atomic chemical potential of the constituent _ i _ ( _ i _ = ni , fe , or b ) .
the quantity @xmath11 represents the number of atoms added to @xmath12 or removed from @xmath13 the supercell .
the quantity @xmath14 is the fermi energy referenced to the energy of the valence band maximum ( vbm ) of the defective supercell , @xmath15 .
this value is calculated as the vbm energy of the pure nfo , corrected by aligning the core potential of atoms far away from the defect in the defect - containing supercell with that in the defect free supercell @xcite .
the quantity @xmath16 represents the charge state of the defect , _
i. e. _ the number of electrons exchanged with the electron reservoir with chemical potential @xmath14 . under solid - liquid equilibrium conditions , the chemical potentials of the ionic species in the solid,@xmath17 , are equal to the chemical potential of the aqueous species in the saturated solution , @xmath18 .
to derive an expression for defect formation energy , eq .
has to be written in terms of ionic chemical potentials , @xmath17 , instead of atomic chemical potentials , @xmath10 .
this can be accomplished by adding and subtracting the term @xmath19 from eq . .
this term can be interpreted as the energy necessary to exchange electrons between the electron reservoir and the atomic species in the pure nfo . if we combine this energy with the atomic chemical potential ( third term in eq . )
we obtain the ionic chemical potential of species in nfo : @xmath20 in eq .
@xmath21 represents the ionic charge , _ i. e_. the number of electrons exchanged with the electron reservoir to create the ionic species in nfo , and @xmath22 is the energy of the vbm in the pure nfo to which the energy of the electron reservoir @xmath23 is referenced . using eq .
together with the solid - liquid equilibrium condition , @xmath24 , the defect formation energy becomes : @xmath25 where @xmath26 is the energy difference between the defect containing and defect free supercells .
therefore , to calculate the defect formation energy , the chemical potentials of the aqueous species have to be evaluated .
the scheme described above has the advantage that it decouples the ionic charge from the charge state of the defect ; charge neutrality is achieved through exchange of electrons with the electron reservoir with energy equal to the fermi level .
the chemical potential of the aqueous ions , @xmath27 , can be written as the sum of the standard chemical potential , @xmath28 , and a temperature dependent term in eq . , @xmath30 is the universal gas constant and @xmath31 is the activity of the ionic species @xmath32 in the aqueous solution . in the present case , because nfo is weakly soluble in water , the activity of the ionic species can be approximated by the concentration of ions in the solution .
the gibbs energies of formation required for eq . are obtained from that supcrt database @xcite which uses the revised helgeson - kirkham - flowers equation of state to predict the thermodynamic behavior of aqueous species at high temperature and pressure @xcite .
the chemical potentials of the solid phases @xmath37 are usually approximated by the total energy per atom of the elemental solid calculated within the dft framework .
however , as pointed out in earlier work , this approach suffers from incomplete error cancellation when total dft energies of physically and chemically dissimilar systems are compared @xcite . therefore , to compute the elemental - phase chemical potentials of the fe , ni , b , and o , we extend the database of 50 elemental energies published by stevanovic _
et al . _
@xcite to include b. to do this we add 26 b - containing binaries to the large fitting set of 252 compounds that have been used by stevanovic _
et al . _ and solve the overdetermined system of 278 equations for 51 elements using a least - square approach , as described in refs . .
the calculated dft energies and experimental formation enthalpies of the 26 b - containing binaries are listed in table [ table1 ] , while the 51 elemental - phase chemical potentials are given in table [ table2 ] in the appendix .
.[table1 ] dft energies and experimental enthalpies of formation of 26 b - containing binaries that have been added to the fitting set in ref . , to calculate the elemental - phase chemical potentials .
theoretical enthalpies of formation are also listed . [ cols="<,^,^,^",options="header " , ] _ pwr axial offset anomaly ( aoa ) guidelines , revision 1 _ , epri , palo alto , ca , 1008102 ( 2004 ) _ rootcause investigation of axial power offset anomaly _ ,
epri , palo alto , ca , tr-108320 ( 1997 ) . _ modeling pwr fuel corrosion product deposition and growth processes _ , epri , palo alto , ca , 1009734 ( 2004 ) . _ modeling pwr fuel corrosion product deposition and growth process : final report _ ,
epri , palo alto , ca , 1011743 ( 2005 ) .
_ axial offset anomaly ( aoa ) mechanism verification in simulated pwr environments _ , epri , palo alto , ca , 1013423 ( 2006 ) .
_ pressurized water reactor ( pwr ) axial offset anomaly mechanism verification in simulated pwr environments _ , epri , palo alto , ca , 1021038 ( 2010 ) .
s. uchida , y. asakura , and h. suzuki , nucl .
eng . des . *
241 * , 2398 ( 2011 ) .
j. a. sawicki , j. nucl . mater . * 374 * , 248 ( 2008 )
. j. a. sawicki , j. nucl . mater . * 402 * , 124 ( 2010 ) .
m. p. short , d. hussey , b. k. kendrick , t. m. besmann , c. r. stanek , and s. yip , j. nucl .
mater . * 443 * , 579 ( 2013 ) .
rk , c. j. obrien , and d. w. brenner , j. nucl .
mater . * 452 * , 446 ( 2014 ) . c. j. obrien , zs .
rk , and d. w. brenner , j phys - condens mat * 25 * , 445008 ( 2013 ) . c. j. obrien , zs .
rk , and d. w. brenner , j. phys .
c * 118 * , 5414 ( 2014 ) .
k. matsunaga , phys .
b. * 77 * , 104106 ( 2008 ) . s. b. zhang and j. e. northrup , phys .
lett . * 67 * , 2339 ( 1991 ) . s. b. zhang ,
j phys - condens mat * 14 * , r881 ( 2002 ) .
t. windman , t. windman , and e. shock , geochim .
72 * , a1027 ( 2008 ) .
j. w. johnson , e. h. oelkers , and h. c. helgeson , computat .
* 18 * , 899 ( 1992 ) . j. c. tanger and h. c. helgeson , am
* 288 * , 19 ( 1988 ) .
r. o. jones and o. gunnarsson , rev .
* 61 * , 689 ( 1989 ) .
v. stevanovic , s. lany , x. w. zhang , and a. zunger , phys .
b * 85 * , 115104 ( 2012 ) .
s. lany , phys .
b * 78 * , 245207 ( 2008 ) .
a. jain , g. hautier , s. p. ong , c. j. moore , c. c. fischer , k. a. persson , and g. ceder , phys .
b * 84 * , 045115 ( 2011 ) .
m. todorova and j. neugebauer , phys .
rev . applied * 1 * , 014001 ( 2014 ) .
k. fujiwara and m. domae , in proceedings of 14th international conference on the properties of water and steam , edited by m. nakahara et al .
( maruzen co. , ltd , kyoto , 2004 ) , p. 581 . a. v. bandura and s. n. lvov , j phys chem ref data * 35 * , 15 ( 2006 ) .
a. e. paladino , j. am .
. soc . * 42 * , 168 ( 1959 ) .
h. m. obryan , f. r. monforte , and r. blair , j. am .
. soc . * 48 * , 577 ( 1965 ) .
a. t. nelson , j. t. white , d. a. andersson , j. a. aguiar , k. j. mcclellan , d. d. byler , m. p. short , and c. r. stanek , j. am .
. soc . * 97 * , 1559 ( 2014 ) . | a serious concern in the safety and economy of a pressurized water nuclear reactor is related to the accumulation of boron inside the metal oxide ( mostly nife@xmath0o@xmath1 spinel ) deposits on the upper regions of the fuel rods .
boron , being a potent neutron absorber , can alter the neutron flux causing anomalous shifts and fluctuations in the power output of the reactor core .
this phenomenon reduces the operational flexibility of the plant and may force the down - rating of the reactor . in this work
an innovative approach is used to combine first - principles calculations with thermodynamic data to evaluate the possibility of b incorporation into the crystal structure of nife@xmath0o@xmath1 , under conditions typical to operating nuclear pressurized water nuclear reactors .
analyses of temperature and ph dependence of the defect formation energies indicate that b can accumulate in nife@xmath0o@xmath1 as an interstitial impurity and may therefore be a major contributor to the anomalous axial power shift observed in nuclear reactors .
this computational approach is quite general and applicable to a large variety of solids in equilibrium with aqueous solutions . |
during the past decade , galactic stars from the sloan digital sky survey ( sdss ) have been used not only to discover substructure , but also to trace the density of the smooth components of the galaxy : the disks and spheroid . because the sample of stars is large and the colors of stars are well measured , it has been possible to use the statistical distributions of star magnitudes to infer the distances to milky way substructure , without the need to know the absolute magnitude of every star . in an early application of this technique @xcite , an estimate of the most common absolute magnitude of very blue f turnoff stars in the sagittarius dwarf spheroidal galaxy
was compared with the most common absolute magnitude of similar stars in the tidal debris stream to determine the distance to the tidal stream .
even though there is a two magnitude spread in the absolute magnitudes of stars near the turnoff , these stars were used as distance indicators to discover stellar substructure and to estimate the distance to it . later studies such as @xcite used photometric parallax of a larger color selection of the stars in the sample , assuming they were overwhelmingly main sequence stars , to measure the shape of the major components of the milky way or to discover substructures .
some searches for stellar halo substructure @xcite have used a matched filter technique to discover stars that have been tidally stripped from dwarf galaxies and globular clusters . since the @xcite paper ,
newberg & yanny have continued to use the apparent magnitude of the turnoff stars in a discovered milky way substructure to estimate the distances to discovered overdensities of stars @xcite .
sometimes the assumed absolute magnitude for f turnoff stars is the same as was calculated for the sagittarius dwarf galaxy , and sometimes it is adjusted slightly based on available information about the stellar population of the substructure being studied . in general , the distance to the structure under study is estimated from the peak of the apparent magnitude distribution of the bluest main sequence stars .
this technique was improved in @xcite , in which a maximum likelihood technique for measuring the spatial distribution of tidal debris from the sagittarius dwarf galaxy was described . in this paper ,
the distribution of magnitudes for f turnoff stars with @xmath4 was represented by a gaussian with center @xmath5 and dispersion @xmath6 .
qsos were eliminated from the sample with @xmath7 .
this distribution is approximately the same as the distribution of similarly selected f turnoff stars for three globular clusters measured by @xcite , except that @xcite fit a slightly larger tail on the faint side of the peak than the bright side of the peak . with the maximum likelihood technique , the star number counts expected for a particular stream and smooth halo model ( assuming a single absolute magnitude value for all turnoff stars )
were broadened in the @xmath8 direction by the absolute magnitude distribution for the turnoff stars , before the number counts were compared with the data . using this method , one could estimate the depth of the tidal debris stream along the line of sight , and could make a more accurate fit to the smooth distribution of stars in the stellar halo .
implicit in all of these studies has been the assumption that the absolute magnitude distribution of f turnoff stars is well represented by a single gaussian distribution with a fixed peak and width .
one might expect that the actual distribution might be a function of age and metallicity of the stellar component .
in fact , the distribution might be different as a function of distance along a tidal debris stream , and different across the stellar halo and disks . to effectively use f turnoff stars as standard candles , one should study the variation in the distribution of f turnoff star absolute magnitudes as a function of stellar population . in this paper
, we analyze eleven globular clusters in the galactic halo , and find that the peak of the distribution of absolute magnitudes of f turnoff stars is typically @xmath9 4.18 , and that the asymmetric distribution can be approximated by a half gaussian on the bright side with a width of @xmath2 0.36 , and a half gaussian on the faint side with a width of @xmath2 0.76 .
this distribution is surprisingly similar for all of the globular clusters studied , which range in age from 9.5 to 13.5 gyr and from [ fe / h]@xmath10 to [ fe / h]@xmath11 .
we explain this surprising result by showing that it is consistent with the age - metallicity relationship for galactic stars .
older clusters should have fainter , redder turnoffs .
however , older clusters also contain fewer metals , which pushes the turnoff brighter and bluer .
although the color of the turnoff varies slightly from cluster to cluster , the absolute magnitude of the turnoff only shifts about 0.1 magnitudes from the mean .
although we find the turnoff magnitude to be similar for the clusters studied , observational effects may considerably change the properties of turnoff star distributions .
as one samples stars to the limit of the survey , the photometric errors increase .
although these photometric errors are small compared to the uncertainty with which we know the absolute magnitude of a single f turnoff star , they can be large compared to the width of our color selection box . for bright magnitudes ,
the actual colors of most of the stars selected are actually within the color range selected . for magnitudes near the survey limit ,
some stars that should be selected are randomly measured with colors that are too red or too blue , and a larger number of stars that are too red or too blue are randomly scattered into our selection range .
the largest flux into and out of the color selection range is on the red side of the range , and primarily broadens the measured width on the faint side of the peak absolute magnitude as one approaches the survey limit .
the results of this paper will make it possible to make more accurate measurements of the intrinsic density of f turnoff stars in the milky way s stellar halo .
we present a function that can be used to estimate the observed width on the faint side of the absolute magnitude distribution .
we also provide numbers for the fractional increase in the number counts as stars leak into the color box at distances where g main sequence stars are near the survey limit ; and the fractional decrease in the number counts as f turnoff stars reach the survey limit and the large color errors scatter them out of the color selection range .
our analysis used photometric data from eleven globular clusters , taken from the sdss database s seventh data release ( dr7 ; abazajian et al .
there are a total of seventeen globular clusters within the sdss dr7 footprint , but five clusters ( ngc 2419 , ngc 7006 , pal3 , pal 4 , and pal 14 ) were eliminated because they are too distant , and one ( pal 1 ) was eliminated because there were too few stars to obtain accurate measures of the f turnoff star distribution using sdss data . a list of all clusters studied in this investigation can be found in table 1 . to determine the limits in right ascension and declination for selecting stars for each field
, we used the sdss skyserver s navigate tool .
the image field of view was expanded so that the cluster was clearly visible , then expanded further until a sizable background distribution was also contained within this view . in general , the entire field of view spanned a rectangle with sides of lengths between 6 and 8 times the apparent radius of the cluster .
the bounds of this rectangle were then used as the right ascension and declination limits in the casjobs data query . within these limits , we selected all objects classified as star , and that had @xmath12 . the latter requirement was designed to avoid possible contamination from quasars @xcite . by selecting from the database of
star"s , we ensured that we obtained only one instance of each object .
we extracted the extinction - corrected ( denoted by the subscript ` 0 ' ) point - spread - function ( psf ) apparent magnitudes with errors .
we determined @xmath13 , the radius within which the majority of the stars belong to the globular cluster , and @xmath14 , the radius outside of which there is little contamination from cluster stars , by visual inspection of the data .
stars with @xmath15 were removed from the data set .
an example is shown in figure 1 .
the @xmath13 and @xmath14 values used for each cluster are included in table 1 .
the following clusters are listed in table 1 , but were excluded in some of our analyses : ngc 5053 has a very low zinn & west metallicity ( @xmath16_{zw84}$ ] @xmath17 - 2.58 ) which falls outside the range of the @xcite conversion scale , and is also below the minimum metallicity value for which padova isochrones can be generated . in the newer work of @xcite , an @xmath16 $ ] of -2.30
is found , which is within the range of the padova isochrones @xcite , and therefore we use this metallicity value for our analysis . to indicate status as a potential outlier
, results for ngc 5053 are plotted using a red - dotted series in figures showing turnoff stars properties .
m15 ( ngc 7078 ) is a ` core - collapsed ' cluster , and so has a very compact core .
it may also have different dynamics and stellar distributions than standard globular clusters @xcite .
the sdss photometric pipeline does not attempt to deblend crowded star fields , and so information on m15 in sdss is highly biased towards stars found on the edges of the cluster .
we do not expect this cluster to necessarily be consistent with the other clusters , but we include it in analysis anyway .
we indicate m15 s status as potential outlier in figures showing turnoff star properties by plotting it with a red dotted series .
ngc 4147 contains only slightly more stars than m92 , ( 583 versus 334 ) but they are more concentrated around the turnoff due to a lack of stars below @xmath18 .
this lack of stars is due to sdss crowded - field photometry detection efficiency problems at fainter magnitudes ( see section 4 ) .
we attempt to analyze ngc 4147 with other clusters later in our analysis , but we expect errors to be large due to the low number of data points available . in figures showing turnoff star properties ,
a blue dotted series is used for ngc 4147 to indicate it s low star counts and high expected errors .
cluster m92 ( ngc 6341 ) fell on the edge of the sdss dr7 footprint , and relatively few stars were observed in the cluster .
the number of cluster stars was large enough to produce a fiducial fit with large bins in magnitude ( 0.5 magnitudes ) .
therefore we were able to fit a modified padova isochrone to this cluster .
however , as there are so few m92 stars in our data , especially close to the turnoff , that m92 was omitted from the turnoff analysis .
in order to convert our observed apparent magnitudes to absolute magnitudes , we rely on the measured distances to each globular cluster . since we would like to study the effects of age and metallicity on the absolute magnitudes of the turnoff stars , we would like measurements of these quantities that are as accurate and as uniform as possible for our sample of globular clusters . in this section ,
we assemble the spectroscopically determined metallicities ( @xmath16 $ ] ) from a single group of authors ; measurements were obtained from zinn & west ( 1984 ) and then converted to more modern carretta & gratton scale using the conversion provided in @xcite . using these metallicities , we then fit ages and distances to the clusters in a consistent fashion using padova isochrones .
the padova theoretical isochrones were fit to fiducial sequences determined from data for eleven clusters found in sdss .
stars in each cluster were separated into @xmath8 bins , then the mean and standard deviation ( @xmath19 ) of the @xmath20 distribution in each bin was found .
any stars in the @xmath8 bin with a @xmath20 value beyond 2@xmath19 from the mean were rejected , and then the @xmath20 mean and @xmath19 of the remaining population was found . the @xmath20 mean and average @xmath0 value were accepted as a point on the fiducial sequence once the entire bin was within 2@xmath19 of the current mean .
isochrones were then fit to the fiducial sequences using distance and age as free parameters , while metallicity was held constant at the spectroscopically determined value . in our initial attempts to use this technique to determine cluster properties
, we found that padova isochrones that are good fits to both the main sequence and the subgiant branch require unreasonably high ages ( @xmath2115 gyr ) for most clusters .
the lack of agreement between theoretical isochrones and cluster data has been explored by previous authors . using eclipsing binary stars in the hyades open cluster
, @xcite showed that theoretical isochrones do not match true star populations ; there are discrepancies in mass , luminosity , temperature , and radius . in the second paper in the series , @xcite uses _ hipparcos _ parallax data for the hyades cluster to further calibrate theoretical isochrones , finding that offsets in color indexes are sufficient to bring a theoretical isochrone in line with real main sequence data . in the fourth and final paper of the series
@xcite fits isochrones to galactic open clusters using color corrections in @xmath22 as a function of @xmath23 , while using cepheid variables as calibration points . in @xcite ,
updated yale rotating evolutionary code with marcs model atmospheres were used to produce _ ugriz _ isochrones which were fit to main sequences of five globular clusters , producing ages and distances to these clusters . using similar techniques ,
we seek to calibrate padova _
isochrones to the @xcite results . comparing a padova isochrone generated from the @xcite derived age and distance for globular cluster ngc 6205 with our derived fiducial fit , we find that the difference between the theoretical isochrone and the data is very nearly linear along the main sequence ( figure 2 ) .
therefore , we apply a linear color correction function in @xmath20 , holding @xmath0 as the independent variable , to the padova isochrone to bring it into agreement with our fiducial fit .
we fit only to the main sequence and subgiant branch , and ignored the giant branch ( @xmath24 ) , as the linear trend was no longer valid brighter than the subgiant branch .
we also did not fit to the lower main sequence , as our data does not extend to fainter absolute magnitudes .
our color - correction functional fit is given by the following equation : @xmath25 this function was applied to all of the colors in the model isochrones and then these models were fit to our fiducial sequences derived from the data .
for each cluster , we first chose a padova isochrone that , modified by equation 1 , appeared to fit the fiducial sequence well .
we then chose additional isochrones , identical except for age , which was varied about our by eye " best fit in increments of 0.2 gyrs .
a gaussian function was fit to the residuals of these isochrones fits , and the mean of this gaussian was taken as the best fit age .
errors in our fit ages were determined through a hessian matrix method by comparing the residuals in the isochrone fit . in the limit of a single fit parameter ( age ) , the hessian error method reduces to : @xmath26 where @xmath27 is the residual of the isochrone fit at age @xmath28 , and @xmath29 is the step size in the age determination method .
using equation 2 , we were able to determine the age fit errors for each cluster through the use of three isochrones : the isochrone of best fit age , and two isochrones generated at the best fit age @xmath30 gyr .
padova isochrones fit using our correction function produce a consistent set of metallicities , ages , and distances to our globular cluster sample , as presented in table 1 .
cluster color - magnitude diagrams , fiducial fits , and modified isochrone fits are shown in figure 3 .
we compare our distance fits to distances in three other sources ( @xcite , @xcite and @xcite ) , and compare our ages to other isochrone - derived ages ( @xcite , @xcite , and @xcite ) , in figure 4 .
our distances appear to be in excellent agreement with other sources .
our ages agree to within the formal errors for each cluster , but appear to have a small linear systematic offset .
the ages are a very close match around 13 gyr , but are a gyr or two higher for ages of 10 gyr , so our age scale is slightly more compressed than the ages in the comparison sources .
it is clear from figure 3 that the cluster data is incomplete at fainter absolute magnitudes , especially amongst the farther clusters : pal 5 , and ngc 4147 , ngc 5024 and ngc 5053 .
this incompleteness begins at a brighter magnitude than is expected from the sdss detection efficiency for stars @xcite .
the poorer detection efficiency in globular clusters is due to difficulty in detecting faint sources in highly crowded star fields , and in particular the poor performance of the sdss photometric pipeline in this regime @xcite . for sufficiently crowded fields
, the code can not deblend and resolve faint stars , since they are washed out by much brighter stars . to quantify the cluster detection efficiency , we examined nearby clusters with relatively complete cmds and compared them to the farther , incomplete cmds under the assumption that there are similar absolute magnitude distributions for the stars in nearby and distant clusters .
we select stars from the cmds in a rectangular box with bounds @xmath31 , @xmath32 , and a color cut of @xmath33 , to coincide with areas of completeness in the near clusters ( specifically @xmath34 for ngc 6205 ) that are incomplete for the more distant clusters .
we choose three nearby clusters , ngc 6205 , ngc 5904 , and ngc 5272 , and found that their normalized @xmath35 histograms in our box are , in fact , similar .
therefore , we average them to create a reference absolute magnitude distribution for cluster stars .
we assume that this histogram represents the true magnitude distribution of globular clusters , then compare the normalized histograms of the four most distant clusters ( ngc 4147 , ngc 5024 , ngc 5053 , and pal 5 ) to this reference . to make the normalizations comparable across the incompleteness of the farther clusters
, we took the average difference of the first five bins ( which showed a similar rising trend in all of our clusters ) and scaled the entire histogram by this value , thereby scaling the clusters by matching initial trends . to quantify the detection efficiency in our globular clusters , we fit the ratios between the reference histogram and the distant cluster histograms using a parabolic function .
we expect these globular clusters to have the same intrinsic absolute magnitude distribution as the bright clusters , but are missing faint stars that were lost to the crowded - field photometry .
the histogram residuals and functional fit can be seen in figure 5 , and the resulting function is given empirically as : @xmath36 we can use this function to reconstruct incomplete cluster star distributions , and to model the effect of sdss crowded - field photometry at farther distances .
errors in the parameter fits in equation 3 are 0.14 @xmath37 0.001 , 20.92 @xmath37 0.007 .
we are now prepared to characterize the absolute magnitude distribution of f turnoff stars in old stellar populations .
we select stars in the f turnoff region ( @xmath38 ) , which is the color range used in the photometric f turnoff density searches in @xcite , and the range originally chosen in @xcite to include stars redder than most blue horizontal branch stars , but bluer than the turnoff of milky way disk stars .
we then build a histogram in @xmath0 from the cluster data , using a bin size of 0.2 magnitudes over the range @xmath39 , which minimizes potential contamination from non - cluster stars .
we then divide each bin by the cluster detection efficiency for the @xmath8 that corresponds to the bin s @xmath0 , thereby rebuilding the component of each cluster lost due to the detection efficiency .
we subtract from this histogram the expected number of field stars in each magnitude bin , as determined from the sky area around each cluster , scaled to the sky area of cluster data .
if the cluster lies on the edge of the sdss footprint , or where sections of the clusters were unresolved by sdss , the areas with no data were excluded from the area used in scaling the background bins .
the magnitude histogram for the background was divided by the field stellar detection efficiency @xcite as a function of apparent magnitude .
the top panel in figure 6 shows the distribution of turnoff star @xmath0 absolute magnitudes in globular cluster ngc 5272 .
we fit each sdss cluster histogram with a ` double - sided ' gaussian distribution , where the standard deviation is different on each side of the mean .
this choice provides a good fit to the data using a simple , well - known function .
we have no theoretical motivation behind our choice of fitting function ; however , this function appears to effectively match the form of our data without over - determining the system .
the form of our fit function is given by : @xmath40 where : @xmath41 when normalized : @xmath42 the fit parameters @xmath43 , @xmath44 , @xmath45 , and @xmath46 are the amplitude , mean , left - side standard deviation , and right - side standard deviations , respectively .
we count all bins outside of the range @xmath39 to be zero .
we assumed poisson counting errors when fitting this function to our data histograms .
typically , the approximation of ` square - root n ' is used to simulate true poisson errors , but we found that this over - emphasized low - count bins at the expense of the overall fit .
we instead modeled our poisson errors in the fashion of equation 10 from @xcite . at the standard 1-@xmath3 confidence level
, this equation simplifies to @xmath47 .
a two stage process was used in fitting the parameters of the double - sided gaussian function to the @xmath0 histograms .
first a markov - chain monte carlo technique was used to sweep the parameter space , then the best fit point was fed into a gradient descent algorithm .
we found that this two - stage fit allowed us to avoid local minima and find the true best fit .
a sample histogram and functional fit are provided for m3 ( ngc 5272 ) on the top plot of figure 6 .
results from the fits to ten clusters can be found in table 2 .
a hessian matrix , calculated at the best fit parameters , was used to determine the model errors in the parameter fit values .
this matrix was multiplied by 2 to make it equivalent to the curvature matrix , then inverted , so that the diagonal elements are equivalent to the squared parameter variances . to determine the turnoff color , @xmath48 , for a cluster , we fit a gaussian profile to the @xmath20 histogram of all stars within 0.5 magnitudes of the @xmath44 parameter determined by the turnoff fits .
the mean value from this gaussian is taken as the turnoff value of the cluster .
this definition of turnoff color is then consistent with the turnoff color of theoretical isochrones ( the bluest point of the isochrone turnoff ) .
note that we do not want to choose the ` bluest ' ( minimum @xmath20 value ) point of our data , since that ignores the intrinsic color spread of the data , and could result in a bluer measurement of the turnoff with distance as color errors become larger .
we now examine our f turnoff distribution fit parameters as a function of the ages , metallicities , and distances to the clusters in our sample .
keep in mind that throughout this section , we are describing the turnoff magnitude in the sdss @xmath49 filter and turnoff color in @xmath20 . in figure 7
, we find no significant correlation with cluster age and our fit parameters @xmath44 , @xmath45 , @xmath46 , and @xmath48 .
intuitively , as a cluster ages @xmath44 should move to higher magnitudes ( fainter ) and the turnoff color should move to smaller ( redder ) values of @xmath20 .
the invariance of these parameters with age is evidence for a fairly uniform globular cluster sample and for the the age - metallicity conspiracy discussed in section 7 .
when these fit parameters are plotted vs. metallicity , possible relationships emerge ( figure 8) . in general , when the metallicity of a theoretical isochrone is increased while holding other variables constant , the turnoff becomes more pinched " and dimmer - that is , the main sequence and subgiant branch move closer together in magnitude while the turnoff point becomes fainter .
the decrease of the @xmath45 and @xmath46 parameters with increasing @xmath16 $ ] indicates that pinching " is occurring , however , our @xmath44 fits _ decrease _ , showing the opposite behavior from what is expected .
this is evidence that metallicity is not the dominant factor in determining turnoff brightness , and will be further explored in section 7 .
a reddening ( increasing ) of turnoff color is expected with increasing @xmath16 $ ] , and is observed in our fits .
therefore , metallicity is dominant over age in determining the turnoff color .
it is important to note that , for our cluster sample , the absolute magnitude distribution has little dependence on metallicity .
having explored the f turnoff distribution as a function of age and metallicity , we now turn our attention to the distance ( figure 9 ) .
as we do not expect position in the galactic halo to significantly change the structure of a cluster , we expect all of our turnoff fit parameters to be intrinsically independent of the observed distance for similar clusters .
any parameter dependence on distance is therefore actually a result of observational errors .
we see from figure 9 that @xmath44 and @xmath45 stay roughly constant with distance , and @xmath46 rises first , then decreases .
the turnoff color ( @xmath48 ) appears to be invariant with distance . in the next section
, we study the observational errors contained within the sdss database , and show that these errors explain the observed changes in the distribution of f turnoff stars as a function of distance .
photometric errors widen the observed magnitude and color distributions of turnoff stars in globular clusters as plotted in color - magnitude diagrams , by an increasing amount with fainter magnitudes .
we now examine the photometric errors in the sdss database and extrapolate the observational biases that result from these errors .
sdss forgoes the conventional pogson " logarithmic definition of magnitude in favor of the asinh magnitude " system described in @xcite .
these systems are virtually identical at high signal - to - noise , but in the low signal - to - noise regime asinh magnitudes are well - behaved and have non - infinite errors .
the two magnitude systems do not differ noticeably in magnitude or magnitude error until fainter than 24th magnitude ( signal - to - noise @xmath172.0 ) , so our analysis will be valid in both systems .
figure 10 shows the relationship between apparent magnitude and magnitude error in the @xmath50 , @xmath8 and @xmath51 passbands for stars in the palomar 5 selection field .
we fit an exponential function with an argument that is linear in apparent magnitude to the error vs. magnitude data , and find : @xmath52 @xmath53 @xmath54 it is the negative constant term in the exponential that determines the magnitude at which the errors start to become significant .
it is evident from the functional fits and figure 10 that , of the three studied passbands , errors in @xmath50 grow most quickly , while errors in @xmath8 are the smallest throughout .
most of our good data is brighter than 23rd magnitude , where @xmath8 and @xmath51 magnitude errors are less than 0.3 .
this will have little effect on the overall appearance of the h - r diagrams , however , magnitude errors of this degree will have a noticeable affect on the observed colors , where the faint magnitude errors are on the order of the observed @xmath20 values .
since f turnoff stars are selected through color cuts , we expect sdss color errors to cause the turnoff to be contaminated by misidentified non - turnoff stars .
we took cluster ngc 6205 as a reference data set , as the relatively low distance ( 7.7 kpc ) implies minimal photometric magnitude errors near the turnoff .
cluster ngc 6205 color errors do not become noticeable until @xmath55 ( @xmath56 ) ; at our absolute magnitude limit of @xmath57 ( @xmath58 ) , the magnitude errors are close to 0.1 in @xmath8 and 0.15 in @xmath51 , resulting in a maximum color error at @xmath57 equal to 0.18 . from figure 3 , we see that ngc 6205 stars between @xmath59 are sufficiently red that even for the maximum color error , these stars are statistically unlikely to to be detected in the turnoff color cut .
we therefore can assume that ngc 6205 will be illustrative as an example of a cluster with an uncontaminated turnoff . in order to understand how the errors affect a cluster with increasing distance
, we must first define a process that will allow us to view a cluster as though it were observed by sdss at a farther distance , taking into account the increasing magnitude errors as apparent magnitudes increase .
we first choose a new ` effective distance ' ( @xmath60 ) to observe the cluster at , and then perform a ` distance shift ' on each star for each of @xmath50 , @xmath8 and @xmath51 ; the magnitude of each star is increased by an amount equivalent to observing the cluster at @xmath60 instead of the original distance ( @xmath61 ) . the magnitude error associated with the new magnitude
is then derived from the appropriate error equation ( equations 6 , 7 and 8) .
the original magnitude error is then subtracted in quadrature from the new magnitude error to produce the relative increase in error .
the new magnitude value is then modified by the relative increase in error : we select a random value from a gaussian distribution with a mean equivalent to the new magnitude , and with a standard deviation equal to the relative increase in error .
this random value produces the new shifted magnitude , equivalent to observing the star at @xmath60 , including the effects of sdss magnitude errors .
having defined the distance shift process , we then separate the ngc 6205 cluster data into three color bins and examine bin cross - contamination as the distance increases .
the three color bins are : the primary ` yellow ' turnoff bin @xmath62 , the ` red ' star bin ( @xmath63 ) and the ` blue ' star bin ( @xmath64 ) .
each of these bins were treated as separate data sets .
we then performed distance shifts on each bin at 1.0 kpc steps , up to a maximum @xmath60 of 80.0 kpc . at each step
we reinforced the @xmath65 color cut , then counted the number of stars that remained in their origin bin , and the number that had leaked into other color bins due to color errors .
we performed this process 100 times with the ngc 6205 cluster data , and averaged the results to smooth out potential random errors .
we then repeated this process , but included the field stellar detection efficiency in our calculations in order to represent the actual observed turnoff population .
figure 11 illustrates the composition of the selected turnoff stars ( @xmath38 ) as a fraction of the original ` true ' turnoff count , and as a function of distance .
this calculation assumed 100% detection efficiency for the shifted stars .
the results of applying the field stellar detection efficiency is shown by the dotted lines .
the trends in this figure are color - coded by the bin of origin ; the black trend is a sum of all lower curves , and represents the total number of stars detected as turnoff stars for a given distance . readily apparent from figure 11
is the quick influx of ` red ' stars into the turnoff between distances of ( 10.0 - 20.0 kpc ) , and the constant loss of true turnoff stars with distance .
red g stars that lie just below the turnoff are on the densely populated main sequence , and have @xmath20 values just above our turnoff color cut of 0.3 .
even slight color error perturbations will tend to shift some these stars into the turnoff color cut , resulting in significant ` red ' star contamination at relatively low halo distances .
it is clear from figure 11 that this is a rapid effect that occurs at fairly low distances ( 10.0 - 20.0 kpc ) , but as the errors continue to increase it becomes just as likely for a red g star to jump over the turnoff selection box as it is to land in it , thereby causing the red star contamination to stop increasing around 20.0 kpc .
it is interesting to note that the @xmath65 color cut causes some suppression of the red contamination effect . since red stars are fainter and farther down the main sequence , they have higher measured magnitudes and magnitude errors than true turnoff stars . as the @xmath50 passband has the highest associated magnitude errors , faint stars with large errors
are perturbed more in @xmath65 than in @xmath20 .
therefore , faint red stars with large color errors may be perturbed beyond the @xmath65 cut and subsequently removed from the data set , even if errors would place that star in the @xmath20 cut for the turnoff .
the @xmath65 cut then serves to remove some of the red contamination stars from the turnoff selection .
as the errors increase , true f turnoff stars can only leak out of the turnoff selection box .
since there are very few bluer ( @xmath64 ) a - type stars or redder subgiants in globular clusters , and these in any event are bright , few a stars leak into the f turnoff star selection due to errors in color . around the distance that the red star contamination stops ( @xmath1725.0 kpc ) ,
the number of stars in the turnoff selection box is comprised of roughly 60% true f turnoff stars and 40% redder g star contamination . as distances increase , the fraction of true turnoff stars in the color selection range decreases to 50% , and the total number of stars selected as turnoff stars decreases .
this is a significant effect that has never been accounted for in previous research papers .
so that future authors may compensate for these effects , we provide analytical functions for the f turnoff dissipation and the red star contamination .
these fits are to the 100% detection efficiency case ( solid curves in figure 11 ) , representing the effect caused by color errors only .
we represent the f turnoff dissipation with a 4th - order polynomial function in @xmath60 , and fit the red contamination with a similar function of 7th - order , with the coefficients given by @xmath66 and @xmath67 ( where @xmath68 , subscripts corresponding to order of term ) , respectively : @xmath69 @xmath70 these functions are valid in the range of @xmath71 , that is , the range of the distance shifts used in the above analysis . in order to study the effects of sdss errors on our measured fit parameters
, we performed distance shifts ( see previous section ) on the @xmath50 , @xmath8 and @xmath51 magnitudes of each cluster at varying @xmath60 steps , to a maximum of 44.0 kpc . at every @xmath60
, we applied the @xmath33 color cut to remove stars that would have been removed as if we had selected this data from the sdss database .
at each distance shift we took into account the background subtraction and detection losses .
since distance shifts must be performed on a dataset of individual stars , while the correction functions must be applied to histogrammed data , we performed the distance shift before background subtraction and detection efficiency correction could occur . in order to keep the background subtraction consistent with a new , effectively more distant cluster data set
, we applied an equivalent distance shift to the background prior to binning and subtraction .
this reproduces the effect of subtracting the background prior to the shift .
we then divided by the cluster detection efficiency , but with the parabola center shifted with the cluster to the new magnitude .
if we did not shift the detection efficiency function before dividing , it would be applied to the wrong portion of the cluster histogram , since the cluster has been shifted to higher magnitudes . before performing the functional fit to our shifted data ,
we apply one of three observational biases to our @xmath0 histogram : the cluster detection efficiency detailed in section 4 , the stellar detection efficiency described in @xcite , and 100% detection efficiency ( in which no correction is applied ) .
the first bias system reveals the evolution with increasing distance of globular cluster turnoff distributions as observed in sdss data .
the second bias system will produce the evolution of non - cluster turnoff distributions as a function of distance .
the final system reveals the evolution of turnoff distributions if no detection bias is applied , that is , if all of the stars originally detected in a cluster continue to be detected as the distances increase .
all ten clusters fit in section 5 were distance shifted to increasingly greater @xmath60 values , using the methods outlined above . at each @xmath60 ,
a new set of fit parameters for the observed absolute magnitude distribution were evaluated using the methods described above .
an example of the histogram and fit of cluster m3 ( ngc 5272 ) , shifted to a @xmath60 of 44.0 kpc , is presented in the lower plot of figure 6 .
we present the results of the distance shifted fits for the parameters @xmath44 , @xmath45 , and @xmath46 in figures 12 , 13 , and 14 respectively .
ngc 5053 and m15 ( ngc 7078 ) are plotted with a red dotted series to indicate their status as expected outliers , while ngc 4147 is plotted with a blue dotted series to indicate that it contains few stars and therefore the fits contain large errors .
it is important to note that all of the clusters studied , including suspected outliers , have similar @xmath44 and @xmath45 values despite differences in distance , age and metallicity .
although we see differences in @xmath46 , we have shown these are due to photometric errors , and not differences in the absolute magnitude distribution of turnoff stars in globular clusters .
this implies that the halo cluster population is intrinsically similar throughout . in the next section
we will show that this similarity is related to the age - metallicity relationship . in figure 15
we show a series of fits to the turnoff star magnitude distribution for nearby cluster ngc 6205 at increasing effective distances , including the effects of the cluster detection efficiency . from this plot ,
the most obvious effect is the loss of stars with distance ; however , one can see that these losses balance to cause @xmath45 to stay constant throughout , and @xmath44 shifts slightly to the left ( brighter magnitudes ) as the detection efficiency cuts in with distance .
we find that @xmath44 is approximately constant with distance , regardless of the applied detection efficiency bias . from the plot of @xmath44 vs @xmath60
( figure 12 ) , we see that @xmath44 values decrease slowly with increasing distance ; however , @xmath44 fit errors increase quickly with distance due to the loss of turnoff stars . within the fit errors , @xmath44
is adequately described by a constant value .
the error - weighted constant - value fit to @xmath44 gives @xmath1 4.18(@xmath37 0.008 ) , with a cluster dispersion of 0.073 .
we also provide a linear fit to @xmath44 with distance , and find that a slope of zero ( which would imply a constant value ) is within the one-@xmath3 errors : @xmath72 we find that @xmath45 values also stay constant with distance , regardless of the applied detection efficiency , to within calculated errors ( figure 13 ) .
it is also apparent that the two expected outlier clusters ngc 5053 and ngc 7078 have @xmath45 behaviors that differ from the other clusters .
an error - weighted average to the @xmath45 values , excluding the two outliers , give @xmath73 0.36(@xmath37 0.006 ) , with a cluster dispersion of 0.18 .
the values of @xmath46 do not stay constant with the distance shifts for any of our three detection efficiency cases .
because of increasing color errors with distance , all but the nearest turnoff star populations are contaminated by redder main - sequence stars , as described in section 6.2 .
these redder stars enter the turnoff histograms on the fainter ( @xmath74 ) side , thereby widening the overall distribution and increasing @xmath46 while leaving the other two parameters unchanged . the nearest clusters ( ngc 5272 , ngc 5904 , ngc 6205 , ngc 7089 ; excluding the core - collapsed ngc 7078 ) are near enough that they do not exhibit significant red main - sequence contamination , and are consistent with each other . we report that these nearby , uncontaminated clusters are representative of the true " globular cluster distribution , with a @xmath46 fit of 0.76(@xmath37 0.04 ) , equivalent to the fit value of @xmath75 , below .
when we apply the sdss detection efficiency for globular clusters , we find that all of our clusters show the same behavior as a function of distance ( figure 14 , left ) .
the initial , quick rise in @xmath46 with distance is due to a large influx of red main - sequence stars due to color errors .
if the cluster is observed at even farther distances , @xmath46 is reduced due to the cluster detection efficiency removing increasingly larger portions of the faint edge of the turnoff .
this consistent series in @xmath46 is evidence that the observed spread in initial cluster @xmath46 values is not a real feature of the clusters , but is instead an observational bias due to the incompleteness of sdss crowded - field photometry at faint magnitudes . in figure 16 , we show a fit to the variation of @xmath46 with distance . in order to see what happens to the distribution of of turnoff stars as a function of distance for sdss field stars , we also apply the sdss stellar detection efficiency when distance shifting the cluster . we find that the nearby clusters ( ngc 5272 , ngc 5904 , ngc 6205 , ngc 7089 , and ngc 5466 ) follow a similar pattern as in the cluster detection efficiency system , while the initially more distant clusters ( ngc 4147 , ngc 5024 , ngc 5053 , pal 5 ) maintain a relatively constant value for @xmath46 .
the initially more distant clusters are already incomplete due to the cluster detection efficiency , which has modified their apparent @xmath46 fit .
for the nearby clusters , the initial rapid rise in @xmath46 with distance peaks at a higher value of @xmath46 , and the subsequent gradual decline is less severe .
this is due to the wider and fainter drop off for the stellar detection efficiency as compared to the cluster detection efficiency ; that is , the stellar detection efficiency starts removing fainter stars at greater distances , and to a less severe degree .
as the stellar field detection efficiency is the dominant observational bias in the sdss , we provide a 4th order polynomial fit to @xmath46 versus distance , with the coefficients given by @xmath76 : @xmath77 if we assume 100% detection efficiency at all magnitudes , we find that nearby clusters see a quick rise in @xmath46 as in the previous two cases , but then level out at a constant value ( figure 14 , right ; figure 16 ) , while farther clusters remain constant ( as in the previous case ) .
as discussed in section 6.2 , the ratio of true " turnoff stars to red contaminants remains constant after the initial influx of red stars into the turnoff color cut .
since this ratio remains constant , there is no appreciable change to the observed turnoff distribution after the initial rise .
figure 16 shows the difference in @xmath46 evolution for different detection efficiency cases .
the sudden inflow of red turnoff contaminants between 10.0 and 15.0 kpc is responsible for the rapid rise in @xmath46 fits for all three detection efficiency systems .
the nearby clusters ( ngc 5272 , ngc 5904 , ngc 6205 , and ngc 7089 ) are close enough ( @xmath78 kpc ) to not contain significant turnoff contamination ; they are therefore representative of the intrinsic @xmath46 value for globular clusters .
when we assume 100% detection efficiency at all magnitudes , the deviation of @xmath46 fits from this intrinsic value is then purely an effect of the color errors due to magnitude , and are not influenced by pre - existing red star contamination or detection efficiency losses .
these @xmath46 trends then serve as a basis for understanding the effects of color errors on the observed distribution of turnoff stars , and so we provide a sigmoid functional fit to the 100% detection efficiency case : @xmath79 where @xmath60 is the effective distance of the shifted cluster .
we find best fit values for the sigmoid functional fit to the nearby clusters are @xmath80 0.52(@xmath37 0.04 ) , @xmath81 12.0(@xmath37 0.31 ) , and @xmath82 0.76(@xmath37 0.04 ) .
the three different trends for @xmath46 with distance , produced by the three separate detection efficiency systems , are compared in figure 16 .
we have shown that the intrinsic turnoff distribution , as quantified by the fit parameters @xmath44 , @xmath45 , and @xmath46 , is similar for all observed clusters , but that the measured value of @xmath46 depends on the distance to the cluster due to photometric errors that increase if the cluster is farther away .
stars that affect @xmath45 are bright , and therefore have smaller color errors than stars that affect @xmath46 , and will therefore not be shifted into or out of the turnoff color cut as easily .
also , there are few stars adjacent to the brighter region of the turnoff , so potential contamination is low .
finally , detection efficiency biases must effect the fainter turnoff stars before they can affect the brighter stars .
therefore , bright turnoff stars are well insulated from contamination and detection bias . the dominant process affecting @xmath45 , then , is the loss of turnoff stars due to color errors .
turnoff star losses also affect @xmath44 , but at large distances @xmath44 will also be modified by detection biases .
the age - metallicity relationship ( amr ) describes the observed relationship between a cluster s age and its average metallicity .
recent observational work @xcite indicates that there are two such amrs present in the milky way : one in which metallicity increases with age , attributed to clusters and dwarf galaxies that were that were gradually captured by the milky way over time ( considered by these authors to be outer halo " clusters ) ; and a set of clusters with age 13 gyr that spans a very large range of metallicities , believed to be due to old clusters that formed rapidly alongside the milky way during the initial formation event ( considered to be inner halo " clusters ) . in figure 17
, we show that the clusters in this study follow the trend of metallicity decreasing with age , which we will henceforth refer to as _ the _ amr , as presented in @xcite .
muratov & gnedin provide a galaxy - independent model of galaxy formation history using semi - analytic models that take into account cosmological simulations .
note that the only two clusters in our sample that are not good matches to muratov & gnedin have already been identified as outliers in previous discussions . since all of our clusters are located at high galactic altitudes , and are likely members of the galactic halo ( and though to have been accreted during galaxy assembly ) , it is not surprising that our clusters are similar to the @xcite relations .
figure 17 shows that the muratov & gnedin amr model does not reproduce the constant ( old , rapid co - forming clusters ) metallicity trend described by @xcite , @xcite , @xcite , but that it is consistent with the amr ( young , late accretion ) cluster values from this work and other papers . using the muratov & gnedin amr and the modified padova isochrones ,
we show that they predict the absolute magnitudes of turnoff stars will be similar for old stellar populations that follow the amr . taking the approximate mean metallicity at range of ages over 8 gyr from figure 8 of @xcite ( plotted in figure 17 as the solid line )
, we produce a set of metallicity versus age points that are representative of the amr .
we show the padova isochrones ( modified by the linear correction function in equation 1 ) in figure 18 . the blue turnoff point can be found between @xmath83 , the turnoff magnitude lies between @xmath84 , and the subgiant branch is constrained between magnitudes of @xmath85 . also shown in figure 18
are two isochrones generated for age - metallicity combinations far from the amr .
these illustrate that the tight grouping of the series is a property of clusters on amr , and is not true for arbitrary combinations of age and metallicity .
the turnoff parameter fits to the unmodified cluster data in table 2 are consistent with the trends seen in figure 18 . along the amr ,
as age decreases and metallicity increases , we see the turnoff move to slightly brighter magnitudes ( lower @xmath44 ) and slightly redder colors ( higher @xmath20 ) .
we find that the amr produces clusters with similar isochrones , but the age of a cluster slightly dominates over the metallicity in determining the @xmath8 turnoff brightness , while the metallicity slightly dominates over the age in determining the @xmath20 turnoff color .
this finding implies that age and metallicity values along the amr conspire " to produce a similar population distribution for all older ( @xmath86 gyr ) clusters . as a cluster ages ,
it s turnoff becomes redder since it has had more time for stars to evolve off of the main - sequence .
however , an increase in metal content also corresponds to a redder value for the entire isochrone .
the plot of isochrones along the theoretical amr in figure 18 implies that these two effects almost exactly cancel each other out , and therefore old stars formed in accordance with the amr should fall in a very narrow range on a color - magnitude diagram .
provided one does not select a cluster from the old ( inner halo " ) population , this finding simplifies distance measurements for the predominantly old stellar population in the galactic halo , but complicates the age determination process , which depends on the uniqueness of isochrones . in this paper , we have removed this complication by using only metallicities determined from spectra ; only age and distance were used free parameters in our isochrone fits .
in this paper we have described the sdss photometric errors and the resulting effects on observed f turnoff distributions ( @xmath38 ) . in this section
we will provide a method by which sdss observations can be corrected for these effects .
if we sum the polynomials whose coefficients are presented in equation 9 and equation 10 , we will produce a function that describes the ratio of stars selected as turnoff stars to the number of actual turnoff stars present , as a function of distance @xmath87 where @xmath88 is the distance to the stellar population . note that this equation assumes a @xmath65 cut to the data ; if this cut is not performed , then additional stars will be included in the data set and this equation will not be valid . instead , the following equation should be used : @xmath89 @xmath90 @xmath91 for simple density searches that ignore f turnoff distribution statistics , dividing the observed density at a given distance by @xmath92 is sufficient to correct data for missing turnoff stars due to magnitude errors .
for example , if one measures the number of background - corrected f turnoff stars in a section of the sagittarius tidal debris stream at a distance of 45 kpc , that number should be corrected for completeness ( if necessary ) and then divided by @xmath93 kpc@xmath94 in order to produce the true number of turnoff stars in that field . for statistical models that seek to quantify f turnoff star densities in the galactic halo ,
a more complicated correction method is required . in @xcite
, the authors sought to map the sagittarius tidal debris stream by performing a maximum likelihood fit of a statistical model to f turnoff stars taken from the sdss data .
they include the effect of the turnoff star distribution by convolving a normalized gaussian , representing the turnoff star distribution ( mean=4.2 , @xmath20.6 ) , with a halo density model in which all of the turnoff stars have the same absolute magnitude ( @xmath94.2 ) .
a correction for the stellar detection efficiency is subsequently applied .
we seek to modify the methods of @xcite by incorporating our results into their analysis . in order to include the turnoff star distribution presented in this paper , the turnoff distribution in the density convolution
must be changed from a symmetrical gaussian ( @xmath95 in that paper ) to the double - sided gaussian , @xmath96 presented in equations 4 and 5 , with a @xmath46 given by equation 13 .
note that we use the 100% detection efficiency function for @xmath46 , as the stellar detection efficiency is applied subsequent to the turnoff convolution .
the normalization for @xmath97 must be adjusted for the fact that turnoff stars are lost ( and a smaller number of g stars are gained ) at farther distances . to account for the turnoff dissipation and contamination ,
the distribution @xmath97 should be multiplied by @xmath98 . putting this together
, we produce a convolution recipe which provides a statistical description of the galactic halo turnoff population , including the effects of the double - sided gaussian magnitude distribution and the effects of dissipation and contamination with distance .
equation 14 in @xcite is replaced by : @xmath99 where @xmath100 represents a density function of magnitude and solid angle ( @xmath101 ) , @xmath102 is the model turnoff density function , in which all stars are assumed to be at absolute magnitude @xmath44 ( 4.18 ) , @xmath103 is equation 14 , and @xmath97 is equation 4 with equation 13 as the input for @xmath46 .
@xmath103 and @xmath46 are both functions of @xmath88 , which is equal to the distance ( @xmath104 ) . using these examples as templates , and with the information contained in the rest of this paper
, future authors should be able to adapt our results to correct their models for observational biases due to turnoff magnitude errors .
future authors must be prepared to compensate for three mechanisms that cause turnoff star incompleteness and misidentification : the combined effects of dissipation and contamination , given by equation 14 ; the sdss stellar field detection efficiency , given in @xcite ; and the intrinsic magnitude distribution of turnoff stars within the galactic halo , as discussed in section 6 and this section .
recent observational studies @xcite ( see also reviews by gratton et al . 2004 and mcwilliam 1997 ) indicate that both outer halo globular clusters and halo field stars have similar alpha - element abundances , consistent with an average [ @xmath105/fe ] value of 0.3 .
this is strong evidence that halo clusters and field stars share a similar formation history , and it is therefore reasonable to assume that they share similar population distributions and photometric properties .
these studies also find that halo clusters and stars are chemically distinct from dwarf spheroidal galaxies @xcite and the galactic disk , including stars in the thick disk .
while thick disk stars can be removed from sdss photometric data through magnitude and color cuts , dwarf spheroidal stars may be present in the halo due to tidal disruptions of infalling dwarfs . if significantly different [ @xmath105/fe ] affect the photometric properties of a population , disrupted dwarf spheroidal galaxies may pose a problem for our technique .
however , the success of @xcite in mapping segments the sagittarius tidal stream with a similar technique is evidence that that dwarf galaxy turnoff star distributions are not significantly different from those in the globular clusters in this study .
the set of globular clusters in this study are limited to old , metal poor galactic halo clusters , with a maximum metallicity of [ fe / h]@xmath10 and a minimum age of 9.5 gyr ( ngc 5904 ) .
figure 18 indicates that our results can be extended to an age of at least 8.0 gyr . from figure 17
we see that our entire cluster sample falls within the metal poor " ( blue ) region of the muratov & gnedin amr .
while our results are well - suited to our goal of describing the old , metal - poor halo of the milky way , we have not tested how far it can be extrapolated to stellar populations less than 8 gyr old or more metal - rich than ngc 5904 .
the results of this paper may not apply to globular clusters with age and metallicity values that fall far from the milky way amr .
our results may not apply to stellar populations in other galaxies , which have differing assembly histories and potentially different age - metallicity relationships .
indeed , the two unusual globular clusters that fall outside of the amr in figure 17 , are the two clusters that differ from each other in turnoff magnitude ( figure 11 ) .
recently , multiple star formation periods have been detected in globular clusters @xcite ( see review by @xcite ) .
we do not expect multiple stellar populations to significantly affect our results , however , certain clusters with several separate , strongly visible main sequences ( such as m54 ) may be poorly described by our results . above average helium enrichment has been proposed as a potential mechanism for producing separations in multiple - population cluster main sequences @xcite . due to the difficulty of measuring he enrichment , the complete effects of enrichment are still under investigation ( see preliminary work by valcarce at al .
2011 , and milone & merlo 2008 ) , but it is thought that it will result in slightly brighter subgiant branches and slightly bluer turnoffs , since he is less opaque than h. if the he abundances have only a small effects on photometry , they will not significantly influence our results .
binary stars present in globular clusters have the potential to bias color - magnitude diagrams towards brighter , redder values , as detailed in @xcite .
the maximum increase in brightness occurs when the binaries are of the same spectroscopic type , @xmath106 magnitude , and the maximum increase in reddening , " which serves to widen the color - magnitude diagram towards redder values , can shift colors by as much as 0.05 magnitudes . depending on the binary fraction ( @xmath107 ) , this effect could introduce significant biases when determining population statistics of a globular cluster .
the binary fraction in old globular clusters is generally low , around a few percent to 20% , with the binaries being concentrated near the cluster s center @xcite , although a select few globular clusters have @xmath107 values as high as 50% .
the primary result of monte - carlo simulations in @xcite is that true " binary fractions are likely to remain constant with age , implying that we can also rule out age - dependent biases on cluster @xmath107 values .
@xcite studied the binary fractions of metal - poor ( [ fe / h ] @xmath108 -1.4 ) field red giants and dwarfs , and found them to be 16%@xmath374% and 17%@xmath372% , respectively , which is on the high end of a typical globular cluster @xmath107 .
globular clusters also feature apparent binaries , which are stars that appear as binaries due to crowding , which become more likely as one looks closer to the densely - packed cluster cores . due to the poor performance of the sdss photometric pipeline in crowded fields ,
the centers of globular clusters are absent from the database , and therefore are not present in this study . if binaries are concentrated towards a cluster s center as in @xcite , then this disadvantage " serves to reduce the potential bias that binaries would have on our results , and will also greatly reduce the number of apparent binaries in our data .
the effect of a binary population is to shift the observed turnoff to brighter magnitudes , and to increase the spread in magnitudes .
we ran an experiment in which we generated 1000 stars with magnitudes given by a double - sided gaussian with @xmath109 , @xmath110 and @xmath111 , and with colors given by a standard gaussian with a @xmath112 and @xmath113 .
we then randomly selected 20% of these stars and shifted them brighter by 0.75 magnitudes and redder by 0.05 magnitudes , which should produce a larger effect than we would expect from the binary fraction of a typical halo globular cluster .
we found that the 20% binary population had a new magnitude fit of @xmath114 , @xmath115 and @xmath116 , and a color fit of @xmath117 and @xmath118 .
the variations in these values are all within the errors of the fitting process , except for @xmath45 in the magnitude fit .
it is possible that the quantity @xmath45 is sensitive to to the binary fractions typically found in halo globular clusters , in the sense that a large binary fraction could produce a larger @xmath45 .
the spread in @xmath45 values in our accepted sample is low ( figure 13 ) , so we do not expect that the effect of binaries has changed our results significantly .
a future study could potentially use a more rigorous study of binary fraction effects to see if the binary fraction is directly correlated to the value of @xmath45 .
we expect the results of this paper to provide a close approximation of the f turnoff absolute magnitude distribution in old , metal poor stars in the milky way halo , as observed by sdss .
these results should also be a reasonable approximation of small dwarf galaxy turnoff distributions .
we do not expect the results to be applicable to young , relatively metal - rich stellar populations , to populations that are outliers on the amr , or to populations that are outside of the milky way .
in this paper we analyze sdss data for 11 globular clusters in the halo of the milky way , and draw conclusions about the intrinsic and observed absolute magnitude distributions of f turnoff stars .
the major conclusions are : \(1 ) the completeness as a function of magnitude in sdss stellar data is different in crowded fields such as globular clusters than in typical star fields .
this incompleteness begins at brighter magnitudes and falls off more steeply than the standard stellar detection efficiency reported in @xcite .
we calculated an approximate parabolic function to describe this crowded - field incompleteness , as presented in equation 3 .
\(2 ) we calculated a linear correction to the turnoff region of padova isochrones in sdss colors in order to place them on an age scale that is consistent with determinations from other authors .
the findings of @xcite were used as a basis for this function . without this correction , padova _ ugriz _
isochrones imply ages greater than the current measured age of the universe . we find the appropriate correction to be @xmath119 , valid for @xmath120 .
\(3 ) using our padova isochrone correction , a uniformly determined set of metallicity ( @xmath16 $ ] ) , distance ( kpc ) , and age ( gyr ) measurements were calculated for 11 globular clusters observed in the sloan digital sky survey .
these results are collected in table 1 .
\(4 ) as color errors increase with distance , the fraction of f turnoff stars detected in a narrow color range decreases rapidly , while redder star contamination becomes significant . using equations 9 and 10 , future authors will be able compensate for both effects .
\(5 ) across a range of old stellar populations ( @xmath121 < -1.17 $ ] , dex ; @xmath122 age @xmath123 , gyr ) , the distribution of sdss @xmath49 filter absolute magnitudes of the turnoff stars ( @xmath38 ) is approximately constant : @xmath124 4.18@xmath37 0.008(dispersion=0.073 ) , and if the distribution is fit with two half gaussians the brighter half gaussian has @xmath125 0.36@xmath37 0.006 ( dispersion=0.18 ) , and the fainter half gaussian , when not affected by observational errors , has @xmath126 0.76@xmath37 0.04 .
\(6 ) although the @xmath0 turnoff absolute magnitude distribution is intrinsically the same for old stellar populations in the milky way , observational errors produce biases which introduce a significant difference in the absolute magnitude distribution for stars selected in a narrow color range . while the average and bright - side half gaussian fit parameters stay relatively constant to within errors , the observed fainter - side half gaussian width ( @xmath46 ) changes dramatically with increasing distance ( and therefore increasing errors ) and depends strongly upon observational biases .
we found that equation 13 provides a good description of the change in @xmath46 with distance , as influenced by color errors only .
this must be accounted for if f turnoff color cuts are to be used to trace stellar structure in the milky way .
\(7 ) the turnoff properties for old stellar populations is consistent with the known age - metallicity relationship for the milky way .
older globular clusters should have fainter , redder turnoffs , but because they are more metal - poor ( which tends to produce brighter , bluer turnoffs ) the turnoff is similar for all older clusters .
for stellar populations at least 8.0 gyr old , the turnoff color ranges from @xmath127 , the turnoff magnitudes range between @xmath84 , and the subgiant branch is constrained between the magnitudes of @xmath85 .
\(8 ) we provide a method for correcting sdss data for the effects of magnitude errors .
we discuss two cases : ( i ) if the stellar population under study is at a single distance , then the observed f turnoff star number counts should be corrected for completeness , and then divided by @xmath92 , as given by equation 14 or 15 ; ( ii ) if the stellar population is distributed in distance , one can convolve the proposed density model with a function that accounts for observational errors before comparing with data .
f turnoff stars have proven useful as tracers of galactic structures , particularly in the halo . before completing this work ,
we were concerned that the assumption of a single absolute magnitude distribution for turnoff stars in the halo would be problematic .
instead , our analysis shows that while a single turnoff is a good assumption , photometric errors significantly affect the stellar population as a function of apparent magnitude . since we have found that all old stellar populations are intrinsically similar , it is possible to model the selected populations of stars ( as selected in a narrow color range ) as a function of magnitude and correct measurements of galactic structure for this effect .
funding for this research was provided by the national science foundation , under grant ast 10 - 09670 .
matthew newby was partially supported by the nasa / ny space grant .
heidi newberg was partially supported by nsfc grants 10973015 and 11061120454 .
jacob simones was supported by an nsf reu supplement to iis 06 - 12213 .
matthew monaco was supported by nsf reu grant dmr 08 - 50934 .
we would like to thank our anonymous referee for providing helpful feedback that improved the results of this paper , and made our conclusions more robust .
we would also like to acknowledge brian yanny for providing guidance with navigating the sdss database , and to benjamin willett , david horne and jeffery carlin for providing advice and discussion that helped move this project forward .
funding for the sdss and sdss - ii has been provided by the alfred p. sloan foundation , the participating institutions , the national science foundation , the u.s .
department of energy , the national aeronautics and space administration , the japanese monbukagakusho , the max planck society , and the higher education funding council for england .
the sdss web site is http://www.sdss.org/. the sdss is managed by the astrophysical research consortium for the participating institutions .
the participating institutions are the american museum of natural history , astrophysical institute potsdam , university of basel , university of cambridge , case western reserve university , university of chicago , drexel university , fermilab , the institute for advanced study , the japan participation group , johns hopkins university , the joint institute for nuclear astrophysics , the kavli institute for particle astrophysics and cosmology , the korean scientist group , the chinese academy of sciences ( lamost ) , los alamos national laboratory , the max - planck - institute for astronomy ( mpia ) , the max - planck - institute for astrophysics ( mpa ) , new mexico state university , ohio state university , university of pittsburgh , university of portsmouth , princeton university , the united states naval observatory , and the university of washington .
abazajian , k. n. , et al .
2009 , , 182 , 543 adelman - mccarthy , j. k. , et al .
2008 , , 175 , 297 an , d. , terndrup , d. m. , & pinsonneault , m. h. 2007 , , 671 , 1640 an , d. , et al .
2009 , , 700 , 523 belokurov , v. , et al .
2006 , , 642 , l137 bedin , l. r. , piotto , g. , anderson , j. , et al . 2004 , , 605 , l125 carney , b. w. , latham , d. w. , stefanik , r. p. , laird , j. b. , & morse , j. a. 2003 , , 125 , 293 carretta , e. , & gratton , r. g. 1997 , , 121 , 95 carretta , e. , bragaglia , a. , gratton , r. , dorazi , v. , & lucatello , s. 2009 , , 508 , 695 cole , n. , et al .
2008 , , 683 , 750 de angeli , f. , piotto , g. , cassisi , s. , et al .
2005 , , 130 , 116 dotter , a. , sarajedini , a. , anderson , j. , et al .
2010 , , 708 , 698 dotter , a. , sarajedini , a. , & anderson , j. 2011 , , 738 , 74 freese , k. , gondolo , p. , newberg , h. j. , & lewis , m. 2004 , physical review letters , 92 , 111301 fregeau , j. m. , ivanova , n. , & rasio , f. a. 2009 , , 707 , 1533 gehrels , n. 1986 , , 303 , 336 girardi , l. , bressan , a. , bertelli , g. , & chiosi , c. 2000 , vizier online data catalog , 414 , 10371 gratton , r. , sneden , c. , & carretta , e. 2004 , , 42 , 385 grillmair , c. j. 2006 , , 645 , l37 grillmair , c. j. , & dionatos , o. 2006 , , 641 , l37 grillmair , c. j. , & dionatos , o. 2006 , , 643 , l17 grillmair , c. j. , & johnson , r. 2006 , , 639 , l17 grillmair , c. j. 2009 , , 693 , 1118 gunn , j. e. , et al .
1998 , , 116 , 3040 harris , w. e. et al .
1997 , , 114 , 1030 haurberg , n. c. , lubell , g. m. g. , cohn , h. n. , lugger , p. m. , anderson , j. , cool , a. m. , & serenelli , a. m. 2010 , , 722 , 158 juri , m. , et al .
2008 , , 673 , 864 lupton , r. h. , gunn , j. e. , & szalay , a. s. 1999 , , 118 , 1406 marigo , p. , girardi , l. , bressan , a. , groenewegen , m. a. t. , silva , l. , & granato , g. l. 2008 , , 482 , 883 marn - franch , a. , aparicio , a. , piotto , g. , et al .
2009 , , 694 , 1498 mcwilliam , a. 1997 , , 35 , 503 milone , a. p. , bedin , l. r. , piotto , g. , et al .
2008 , , 673 , 241 milone , l. a. , & merlo , d. c. 2008 , boletin de la asociacion argentina de astronomia la plata argentina , 51 , 121 muratov , a. l. , & gnedin , o. y. 2010 , , 718 , 1266 newberg , h. j. , et al .
2002 , , 569 , 245 newberg , h. j. , & yanny , b. 2006 , journal of physics conference series , 47 , 195 newberg , h. j. , yanny , b. , cole , n. , beers , t. c. , re fiorentin , p. , schneider , d. p. , & wilhelm , r. 2007 , , 668 , 221 newberg , h. j. , willett , b. a. , yanny , b. , & xu , y. 2010 , , 711 , 32 pasquini , l. , mauas , p. , kufl , h. u. , & cacciari , c. 2011 , , 531 , a35 pinsonneault , m. h. , terndrup , d. m. , hanson , r. b. , & stauffer , j. r. 2003 , , 598 , 588 pinsonneault , m. h. , terndrup , d. m. , hanson , r. b. , & stauffer , j. r. 2004 , , 600 , 946 piotto , g. , bedin , l. r. , anderson , j. , et al .
2007 , , 661 , l53 piotto , g. 2009 , arxiv:0902.1422 pritzl , b. j. , venn , k. a. , & irwin , m. 2005 , , 130 , 2140 rockosi , c. m. , et al .
2002 , , 124 , 349 roederer , i. u. 2009 , , 137 , 272 romani , r. w. , & weinberg , m. d. 1991 , , 372 , 487 sollima , a. , beccari , g. , ferraro , f. r. , fusi pecci , f. , & sarajedini , a. 2007 , , 380 , 781 valcarce , a. a. r. , catelan , m. , & sweigart , a. v. 2011 , revista mexicana de astronomia y astrofisica conference series , 40 , 257 venn , k. a. , irwin , m. , shetrone , m. d. , et al .
2004 , , 128 , 1177 willett , b. a. , newberg , h. j. , zhang , h. , yanny , b. , & beers , t. c. 2009 , , 697 , 207 yanny , b. , et al .
2000 , , 540 , 825 yanny , b. , et al .
2003 , , 588 , 824 yanny , b. , et al .
2009 , , 700 , 1282 zinn , r. , & west , m. j. 1984 , , 55 , 45 lccccccccr ngc 4147 & & 252.85 & 77.19 & -1.58 & 19.3 & 11.2 @xmath37 0.8 & 0.07 & 0.08 & 583 + ngc 5024 & m53 & 332.96 & 79.76 & -1.89 & 18.7 & 12.5 @xmath37 1.3 & 0.25 & 0.30 & 5871 + ngc 5053 & & 335.70 & 78.95 & -2.30 & 18.5 & 11.5 @xmath37 1.1 & 0.14 & 0.17 & 2280 + ngc 5272 & m3 & 42.22 & 78.71 & -1.43 & 10.4 & 12.2 @xmath37 1.1 & 0.30 & 0.40 & 4395 + ngc 5466 & & 42.15 & 73.59 & -2.14 & 15.6 & 13.4 @xmath37 0.9 & 0.18 & 0.25 & 2029 + ngc 5904 & m5 & 3.86 & 46.80 & -1.17 & 8.0 & 9.4 @xmath37 0.1 & 0.21 & 0.28 & 3488 + ngc 6205 & m13 & 59.01 & 40.91 & -1.42 & 7.7 & 12.7 @xmath37 0.2 & 0.24 & 0.33 & 2192 + ngc 6341 & m92 & 68.34 & 34.86 & -2.17 & 8.7 & 13.4 @xmath37 0.9 & 0.135 & 0.20 & 334 + ngc 7078 & m15 & 65.01 & -27.31 & -2.04 & 11.0 & 11.2 @xmath37 0.9 & 0.19 & 0.25 & 2939 + ngc 7089 & m2 & 53.37 & -35.77 & -1.38 & 11.5 & 12.6 @xmath37 0.3 & 0.14 & 0.18 & 1451 + pal 5 & & 0.85 & 45.86 & -1.24 & 21.0 & 11.5 @xmath37 0.7 & 0.087 & 0.12 & 1478 + lcccccccccccc ngc 4147 & & 1.70 & 4.42 & 0.33 & 0.50 & 0.22 & 0.98 & 0.29 & 26.6 & 3.34 & 0.223 & 0.022 + ngc 5024 & m53 & 1.69 & 4.49 & 0.02 & 0.50 & 0.02 & 0.94 & 0.03 & 271.8 & 3.53 & 0.240 & 0.010 + ngc 5053 & & 1.92 & 4.68 & 0.07 & 0.70 & 0.06 & 0.93 & 0.07 & 105.3 & 3.00 & 0.219 & 0.011 + ngc 5272 & m3 & 1.37 & 4.33 & 0.05 & 0.39 & 0.04 & 0.77 & 0.04 & 154.9 & 3.71 & 0.251 & 0.009 + ngc 5466 & & 1.92 & 4.28 & 0.15 & 0.40 & 0.09 & 1.22 & 0.13 & 78.7 & 3.17 & 0.217 & 0.010 + ngc 5904 & m5 & 1.23 & 4.13 & 0.28 & 0.32 & 0.18 & 0.72 & 0.19 & 134.7 & 4.03 & 0.253 & 0.006 + ngc 6205 & m13 & 1.30 & 4.31 & 0.13 & 0.31 & 0.09 & 0.79 & 0.10 & 69.0 & 4.07 & 0.251 &
0.008 + ngc 7078 & m15 & 2.18 & 4.48 & 0.11 & 0.62 & 0.08 & 1.23 & 0.11 & 88.7 & 3.26 & 0.219 & 0.011 + ngc 7089 & m2 & 1.27 & 4.33 & 0.10 & 0.31 & 0.09 & 0.76 & 0.12 & 46.0 & 4.07 & 0.258 & 0.010 + pal 5 & & 1.56 & 4.21 & 0.14 & 0.33 & 0.10 & 0.99 & 0.12 & 54.1 & 3.55 & 0.284 & 0.011 + average & & 1.613 & 4.367 & 0.138 & 0.440 & 0.097 & 0.930 & 0.122 & 103.0 & 3.573 & 0.242 & 0.011 + standard & deviation & 0.309 & 0.149 & 0.092 & 0.129 & 0.059 & 0.173 & 0.073 & 67.5 & 0.367 & 0.021 & 0.003 + | f turnoff stars are important tools for studying galactic halo substructure because they are plentiful , luminous , and can be easily selected by their photometric colors from large surveys such as the sloan digital sky survey ( sdss ) .
we describe the absolute magnitude distribution of color - selected f turnoff stars , as measured from sdss data , for eleven globular clusters in the milky way halo .
we find that the @xmath0 distribution of turnoff stars is intrinsically the same for all clusters studied , and is well fit by two half gaussian functions , centered at @xmath1 4.18 , with a bright - side @xmath2 0.36 , and with a faint - side @xmath2 0.76 .
however , the color errors and detection efficiencies cause the _ observed _ @xmath3 of the faint - side gaussian to change with magnitude due to contamination from redder main sequence stars ( 40% at 21st magnitude ) .
we present a function that will correct for this magnitude - dependent change in selected stellar populations , when calculating stellar density from color - selected turnoff stars . we also present a consistent set of distances , ages and metallicities for eleven clusters in the sdss data release 7 .
we calculate a linear correction function to padova isochrones so that they are consistent with sdss globular cluster data from previous papers .
we show that our cluster population falls along the milky way age - metallicity relationship ( amr ) , and further find that isochrones for stellar populations on the amr have very similar turnoffs ; increasing metallicity and decreasing age conspire to produce similar turnoff magnitudes and colors for all old clusters that lie on the amr . |
networks provide a powerful representation of interaction patterns in complex systems@xcite . the structure of social relations among individuals
, interactions between proteins , food webs , and many other situations can be represented using networks . until recently , the vast majority of studies focused on networks that consist of a single type of entity , with different entities connected to each other via a single type of connection .
such networks are now called _ single - layer _ ( or _ monolayer _ ) networks .
the idea of incorporating additional information such as multiple types of interactions , subsystems , and time - dependence has long been pointed out in various fields , such as sociology , anthropology , and engineering , but an effective unified framework for the mathematical treatment of such multidimensional structures , which are usually called _ multilayer networks _ , was developed only recently@xcite .
multilayer networks can be used to model many complex systems .
for example , relationships between humans include different types of interactions such as relationships between family members , friends , and coworkers that constitute different _ layers _ of a social system .
different layers of connectivity also arise naturally in natural and human - made systems in transportation@xcite , ecology@xcite , neuroscience@xcite , and numerous other areas .
the potential of multilayer networks for representing complex systems more accurately than was previously possible has led to an explosion of work on the physics of multilayer networks .
a key question concerns the implications of multilayer structures on the dynamics of complex systems , and several papers about interdependent networks a special type of multilayer network revealed that such structures can change the qualitative behaviors in a significant way .
for example , several studies have provided insights on percolation properties and catastrophic cascades of failures in multilayer networks@xcite .
these findings helped highlight an important challenge : how does one account for multiple layers of connectivity in a consistent mathematical way ?
an explosion of recent papers has developed the field of multilayer networks into its modern form , and there is now a suitable mathematical framework@xcite , novel structural descriptors@xcite , and tools from fields ( such as statistical physics @xcite ) for studying these systems .
many studies have also started to highlight the importance of analyzing multilayer networks , instead of relying on their monolayer counterparts , to gain new insights about empirical systems ( see , e.g. , @xcite ) .
it has now been recognized that the study of multilayer networks is fundamental for enhancing understanding of dynamical processes on networked systems .
an important example are spreading processes , such as flows ( and congestion ) in transportation networks@xcite , and information and disease spreading in social networks@xcite .
for instance , when two spreading process are coupled in a multilayer network , the onset of one disease - spreading process can depend on the onset of the other one , and in some scenarios there is a curve of critical points in the phase diagram of the parameters that govern a system s spreading dynamics@xcite .
such a curve reveals the existence of two distinct regimes , such that the criticality of the two dynamics is interdependent in one regime but not in the other .
similarly , cooperative behavior can be enhanced by multilayer structures , providing a novel way for cooperation to survive in structured populations@xcite . for additional examples ,
see various reviews and surveys@xcite on multilayer networks and specific topics within them .
a multilayer framework allows a natural representation of coupled structures and coupled dynamical processes . in this article , after we give a brief overview on representing multilayer networks , we will focus on spreading processes in which multilayer analysis has revealed new physical behavior .
specifically , we will discuss two cases : ( i ) a single dynamical process , such as continuous or discrete diffusion , running on top of a multilayer network ; and ( ii ) different dynamical processes , in which each one runs on top of a given layer , but they are coupled by a multilayer structure .
one can represent a monolayer network mathematically by using an adjacency matrix , which encodes information about ( possibly directed and/or weighted ) relationships among the entities in a network . because multilayer networks include multiple dimensions of connectivity , called _ aspects _ , that have to be considered simultaneously
, their structure is much richer than that of ordinary networks .
possible aspects include different types of interactions or communication channels , different subsystems , different spatial locations , different points in time , and more .
one can use tensors to encode the connectivity of multilayer networks as ( multi)linear - algebraic objects@xcite .
multilayer networks include three types of edges : intra - layer edges ( connecting nodes within the same layer ) , inter - layer edges between replica nodes ( i.e. , copies of the same entity ) in different layers , and inter - layer edges between nodes that represent distinct entities .
distinguishing disparate types of edges has deep consequences both mathematically and physically .
mathematically , this yields banded structures in multilinear - algebraic objects that depend on a systems physical constraints , and such structures impact features such as a network s spectral properties .
these , in turn , have a significant impact on dynamical systems ( e.g. , of spreading processes or coupled oscillators ) that are coupled through multilayer networks .
moreover , intra - layer edges and inter - layer edges encode relationships in fundamentally different ways , and they thereby represent different types of physical functionality . for example , in a metropolitan transportation system @xcite , intra - layer edges account for connections between the same type of node ( e.g. , between two different subway stations ) , whereas inter - layer edges connect different types of nodes ( e.g. , between a certain subway station and an associated bus station ) . in some cases , inter - layer edges and intra - layer edges may even be measured using different physical units . for instance
, an intra - layer edge in a multilayer social network could represent a friendship between two individuals on facebook , whereas an inter - layer edge in the same network could represent the transition probability of an individual switching from using facebook to use twitter .
the rich variety of connections in a typical multilayer network can be mathematically represented by the components @xmath0 of a 4th - order tensor @xmath1 , called multilayer adjacency tensor@xcite , encoding the relationship between any node @xmath2 in layer @xmath3 and any node @xmath4 in layer @xmath5 in the system ( where @xmath6 and @xmath7 , @xmath8 denotes the number of nodes in the network and @xmath9 denotes the number of layers ) .
once the connectivity of the nodes and layers are encoded in a tensor , one can define novel measures to characterize the multilayer structure .
however , this is a delicate process , as naively generalizing existing concepts from monolayer networks can lead to qualitatively incorrect or nonsensical results@xcite .
an alternative way of generalizing concepts from monolayer networks to multilayer networks is to use sets of adjacency matrices rather than tensors .
this alternative approach has the advantage of familiarity , and indeed it is also convenient to `` flatten '' adjacency tensors into matrices ( called `` supra - adjacency matrices '' ) for computations@xcite .
however , the compact representation of multilayer networks in terms of tensors allows greater abstraction , which has been very insightful , and it will facilitate further development of the mathematics of complex systems .
studies of structural properties of multilayer networks include descriptors to identify the most `` central '' nodes according to various notions of importance@xcite and quantify triadic relations such as clustering and transitivity@xcite .
significant advances have been achieved to reduce the structural complexity of multilayer networks@xcite , to unveil mesoscale structures ( e.g. , communities of densely - connected nodes ) @xcite , and to quantify intra - layer and inter - layer correlations@xcite in empirical networked systems .
the structural properties of multilayer networks depend crucially on how layers are coupled together to form a multilayer structure .
inter - layer edges provide the coupling and help encode structural and dynamical features of a system , and their presence ( or absence ) produces fascinating structural and dynamical effects . for example , in multimodal transportation systems , in which layers represent different transportation modes , the weight of inter - layer connections might encode an economic or temporal cost to switching between two modes @xcite . in multilayer social networks ,
inter - layer connections allow models to tune , in a natural way , an individual s self - reinforcement in opinion dynamics@xcite .
depending on the relative importances of intra - layer and inter - layer connections , a multilayer network can act either as a system of independent entities , in which layers are structurally decoupled , or as a single - layer system , in which layers are indiscernible in practice . in some multilayer networks , one can even derive a sharp transition between these two regimes@xcite .
there are two different categories of dynamical processes on multilayer networks : ( i ) a single dynamical process on top of the coupled structure of a multilayer network ( see fig .
[ fig : dynamics_types]a ) ; and ( ii ) `` mixed '' or `` coupled '' dynamics , in which two or more dynamical processes are defined on each layer separately and are coupled together by the presence of inter - layer connections between nodes ( see fig .
[ fig : dynamics_types]b ) . * _ single dynamics .
* in this section , we analyze physical phenomena that arise from a single dynamical process on top of a multilayer structure .
the behavior of such a process depends both on intra - layer structure ( i.e. , the usual considerations in networks ) and on inter - layer structure ( i.e. , the presence and strength of interactions between nodes on different layers ) .
one of the simplest types of dynamics is a diffusion process ( either continuous or discrete ) .
the physics of diffusion , which has been analyzed thoroughly in multiplex networks@xcite reveals an intriguing and unexpected phenomenon : diffusion can be faster in a multiplex network than in any of the layers considered independently@xcite .
one can understand diffusion in multiplex networks in terms of the spectral properties of a laplacian tensor ( in particular , we consider the type of laplacian that is known in graph theory as the `` combinatorial laplacian '' @xcite ) , obtained from the adjacency tensor of a multilayer network , that governs the diffusive dynamics .
one first `` flattens''@xcite without loss of information , provided one keeps the layer labels the laplacian tensor@xcite into a special lower - order tensor called `` supra - laplacian matrix '' .
the supra - laplacian matrix has a block - diagonal structure , where diagonal blocks encode the associated laplacian matrices corresponding to each layer separately and off - diagonal blocks encode inter - layer connections .
the supra - laplacian matrix was initially presented in the literature as a matrix for a multilayer network that includes both intra - layer edges and inter - layer edges@xcite .
the time scale of diffusion is controlled by the smallest positive eigenvalue @xmath10 of the supra - laplacian matrix . in fig .
[ fig : diffusion ] , we show a representative result that conveys the existence of two distinct regimes in multiplex networks as a function of the inter - layer coupling strength .
the regimes illustrate how multilayer structure can influence the outcome of a physical process . for small values of the inter - layer coupling
, the multilayer structure slows down the diffusion ; for large values , the diffusion speed converges to the mean diffusion speed of the superposition of layers . in many cases ,
the diffusion in the superposition is faster than that in any of the separate layers .
these findings are a direct consequence of the emergence of more paths between every pair of nodes due to the multilayer structure .
the transition between the two regimes is a structural transition@xcite , a characteristic of multilayer networks that can also arise in other contexts@xcite .
the above phenomenology can also occur in discrete processes .
perhaps the most canonical examples of discrete dynamics are random walks , which are used to model markovian dynamics on monolayer networks and which have yielded numerous insights over the last several decades@xcite . in a random walk , a discretized form of diffusion
, a walker jumps between nodes through available connections . in a multilayer network ,
the available connections include layer switching via an inter - layer edge , a transition that has no counterpart in monolayer networks and which enriches random - walk dynamics@xcite .
an important physical insight of the interplay between multilayer structure and the dynamics of random walkers is `` navigability''@xcite , which we take to be the mean fraction of nodes that are visited by a random walker in a finite time , which ( similar to the case of continuous diffusion ) can be larger than the navigability of an aggregated network of layers . in terms of navigability ,
multilayer networks are more resilient to uniformly random failures than their individual layers , and such resilience arises directly from the interplay between the multilayer structure and the dynamical process .
another physical phenomenon that arises in multilayer networks is related to congestion , which arises from a balance between flow over network structures and the capacity of such structures to support flow .
congestion in networks was analyzed many years ago in the physics literature @xcite , but it has been studied only recently in multilayer networks @xcite , which can be used to model multimodal transportation systems .
it is now known that the multilayer structure of a multiplex network can induce congestion even when a system would remain decongested in each layer independently@xcite .
coupled dynamical processes are a second archetypical family of dynamics in which multilayer structure plays a crucial role . thus far , the most thoroughly studied examples are coupled spreading processes , which are crucial for understanding phenomena such as the spreading dynamics of two concurrent diseases in two - layer multiplex networks@xcite and spread of disease coupled with the spread of information or behavior@xcite .
we illustrate two basic effects : ( i ) two spreading processes can enhance each other ( e.g. , one disease facilitates infection by the other@xcite ) , and ( ii ) one process can inhibit the spread of the other ( e.g. , a disease can inhibit infection by another disease@xcite or the spreading of awareness about a disease can inhibit the spread of the disease@xcite ) .
interacting spreading processes also exhibit other fascinating dynamics , and multilayer networks provide a natural means to explore them @xcite .
the above phenomenology is characterized by the existence of a curve of critical points that separate endemic and non - endemic phases of a disease .
this curve exhibits a crossover between two different regimes : ( i ) a regime in which the critical properties of one spreading process are independent of the other , and ( ii ) a regime in which the critical properties of one spreading process do depend on those of the other .
the point at which this crossover occurs is called a `` metacritical '' point . in fig .
[ fig : spreading ] , we show ( left ) a phase diagram of disease incidence in one layer of two reciprocally enhanced disease spreading processes ; and ( right ) a phase diagram of the incidence in one layer of an inhibitory disease spreading process affecting another disease .
the metacritical point delineates the transition between independence ( dashed line ) and dependence ( solid curve ) of the critical properties of the two processes .
in most natural and engineered systems , entities interact with each other in complicated patterns that include multiple types of relationships and/or multiple subsystems , change in time , and incorporate other complications .
the theory of multilayer networks seeks to take such features into account to improve our understanding of such complex systems . in the last few years
, there have been intense efforts to generalize traditional network theory by developing and validating a framework to study multilayer systems in a comprehensive fashion .
the implications of multilayer network structure and dynamics are now being explored in fields as diverse as neuroscience@xcite , transportation@xcite , ecology@xcite , granular materials@xcite , evolutionary game theory @xcite , and many others .
for instance , in ecological networks , different layers might encode different types of interaction e.g. trophic and non - trophic or different spatial patches ( or different temporal snapshots ) , where the same interaction may or may not appear @xcite . in human brain networks ,
different layers might encode functional connectivity corresponding to specific frequency bands , with inter - layer connections encoding cross - frequency interactions @xcite . in gene interaction networks
, layers might correspond to different genetic interactions ( e.g. , suppressive , additive , or based on physical or chemical associations)@xcite . in financial networks , layers might represent different interdependent networks of entities@xcite e.g. , banking networks and commercial firms or different trade relationships among legal entities , ranging from individuals to countries . despite considerable progress in the last few years@xcite , much remains to be done to obtain a deep understanding of the new physics of multilayer network structure and multilayer network dynamics ( both dynamics of and dynamics on such networks ) . in seeking such a deep understanding ,
it is crucial to underscore the inextricable interdependence of the structure and dynamics of networks .
recent efforts have revealed fundamental new physics in multilayer networks .
the richer types of spreading and random - walk dynamics can lead to enhanced navigability , induced congestion , and the emergence of new critical properties .
such new phenomena also have a major impact on practical goals such as coarse - graining networks to examine mesoscale features and evaluating the importance of nodes two goals that date to the beginning of investigations of networks@xcite . for multilayer networks to achieve their vast potential , there remain crucial problems to address . for example , from a structural point of view , it is much easier to measure edge weights reliably for intra - layer edges than for inter - layer edges .
moreover , inter - layer edges not only play a different role from intra - layer ones , but they also play different roles in different applications , and the research community is only scratching the surface of the implications of their presence and the new phenomena to which they lead .
for example , how to infer or impose inter - layer edges ( and their associated meaning ) is a major challenge in many applications of social networks , where an inter - layer connection could exploit the fact of changing from a social platform to another in time as the probability of switching
. this can be even more complicated in many types of biological networks ( e.g. , when considering protein and genetic interactions ) .
we know that different layers are not independent of each other , but it is much more difficult to quantify and measure the weights of the dependencies in a meaningful way .
another major challenge is to understand the propagation of dynamical correlations , due to network structure , across different layers , which affects not only spreading processes but dynamical systems more generally .
although our manuscript only addresses physical phenomena related to spreading processes , other dynamical processes also pose extremely fascinating questions@xcite .
one important example is synchronization , although there are many others ( e.g. , opinion models , games , and more ) .
a few studies with particular setups have made good progress on multilayer synchronization ( see , e.g. , @xcite ) , but there phenomenology is very rich , and it will require the development of solid theoretical grounding to study synchronization manifolds , stability analysis , transient dynamics , and more . additionally , one can build on diffusion dynamics to study reaction
diffusion systems in multilayer networks @xcite .
a particularly promising approach in network theory that will have a major impact on future studies of multilayer networks is the analysis of network structure that arises from latent geometrical spaces@xcite .
observed connectivity in networks often depends on space@xcite either through explicit constraints or by influencing the existence probability and weights of edges and thus on the distance in that space .
either or both of the latent space ( e.g. , people with connections on more layers can lead to a higher probability of observing an edge between them@xcite ) and an associated observed network connections can have a multilayer structure .
such explicit use of geometry also allows the possibility of incorporating more continuum types of analyses to accompany the traditional discrete approaches to studying networks .
we thus assert that techniques from both geometry and statistics will be crucial for scrutinizing dynamical processes on multilayer networks .
the study of multilayer networks is in its infancy , and new emergent physical phenomena that arise from the interaction of such networks and the dynamical processes on top of them are waiting to be discovered .
all of the authors wrote the paper and contributed equally to the production of the manuscript .
the authors declare no competing financial interests .
all authors were funded by fet - proactive project plexmath ( fp7-ict-2011 - 8 ; grant # 317614 ) funded by the european commission .
mdd acknowledges financial support from the spanish program juan de la cierva ( ijci-2014 - 20225 ) .
cg acknowledges financial support from a james s. mcdonnell foundation postdoctoral fellowship .
aa acknowledges financial support from the icrea academia , the james s. mcdonnell foundation , and fis2015 - 38266 .
map acknowledges a grant ( ep / j001759/1 ) from the epsrc .
the authors acknowledge help from serafina agnello on the creative design of figures .
* multilayer networks*. ( * a * ) an edge - colored multigraph , in which nodes can be connected by different types ( i.e. , colors ) of interactions . in this example
, there are no inter - layer edges .
( * b * ) a multiplex network , which consists of an edge - colored multigraph along with inter - layer edges that connect entities with their replicas on other layers .
( * c * ) an interdependent network , in which each layer contains nodes of a different type ( circles , squares , and triangles ) and includes inter - layer edges to nodes in other layers ; in this case , inter - layer edges can occur either between entities and their replicas or between different entities . ] * dynamical processes on multilayer networks*. ( * a * ) schematic of a single type of dynamical process running on all layers of a multiplex network .
( arcs of the same color represent the same dynamical process . )
( * b * ) schematic of two dynamical processes , each of which is running on a different layer , that are coupled by the interconnected structure of a multilayer network .
, scaledwidth=70.0% ] * single dynamics on a multilayer network .
* the speed of diffusion dynamics in a multilayer network is characterized by the second smallest eigenvalue @xmath10 of a laplacian tensor .
we consider a pair of coupled erds
rnyi ( er ) networks in which we independently vary the probabilities @xmath11 $ ] to connect two nodes within the same layer between . the condition@xcite to observe faster diffusion in the multilayer network than diffusion in each layer separately is @xmath12 . in the central panel ,
we see that the condition is satisfied when the two layers have similar edge - connection probabilities ( i.e. , @xmath13 .
we set the inter - layer connections between each node and its replica in the other layer to a weight @xmath14 $ ] . in the left panel ,
we show the behavior of @xmath10 as a function of the inter - layer coupling weight @xmath15 .
a sharp change in the value of the @xmath10 separates two different regimes that correspond to different structural properties of the multilayer network . ] * coupled dynamics on multilayer networks . * two ( left )
reciprocally - enhanced and ( right ) reciprocally - inhibited disease - spreading processes of susceptible infected susceptible ( sis ) type .
we compute these diagrams for multiplex networks formed by two layers of 5000-node erds rnyi graphs of 5000 with mean intra - layer degree @xmath16 .
the colors in the figure represent the prevalence levels of the diseases at a steady state of monte - carlo simulations .
note the emergence of a curve of critical points ( at a `` metacritical point '' ) in which the spreading in one layer depends on the spreading in the other . ] | the study of networks plays a crucial role in investigating the structure , dynamics , and function of a wide variety of complex systems in myriad disciplines . despite the success of traditional network analysis ,
standard networks provide a limited representation of complex systems , which often include different types of relationships ( i.e. , `` multiplexity '' ) among their constituent components and/or multiple interacting subsystems .
such structural complexity has a significant effect on both dynamics and function .
throwing away or aggregating available structural information can generate misleading results and be a major obstacle towards attempts to understand complex systems . the recent `` multilayer '' approach for modeling networked systems
explicitly allows the incorporation of multiplexity and other features of realistic systems . on one hand
, it allows one to couple different structural relationships by encoding them in a convenient mathematical object . on the other hand
, it also allows one to couple different dynamical processes on top of such interconnected structures .
the resulting framework plays a crucial role in helping achieve a thorough , accurate understanding of complex systems .
the study of multilayer networks has also revealed new physical phenomena that remain hidden when using ordinary graphs , the traditional network representation .
here we survey progress towards attaining a deeper understanding of spreading processes on multilayer networks , and we highlight some of the physical phenomena related to spreading processes that emerge from multilayer structure .
departament denginyeria informtica
i matemtiques , universitat rovira i virgili , 43007 tarragona , spain carolina center for interdisciplinary applied mathematics , department of mathematics , university of north carolina , chapel hill , nc 27599 - 3250 , usa oxford centre for industrial and applied mathematics , mathematical institute , university of oxford , ox2 6gg , uk ; cabdyn complexity centre , university of oxford , oxford ox1 1hp , uk ; and department of mathematics , university of california , los angeles , california 90095 , usa corresponding authors : [email protected] , [email protected] |
a famous result from lamperti @xcite asserts that any continuous state branching process can be represented as a spectrally positive lvy process , time changed by the inverse of some integral functional .
this transformation is invertible and defines a bijection between the set of spectrally positive lvy processes and this of continuous state branching processes .
lamperti s result is the source of an extensive mathematical literature in which it is mainly used as a tool in branching theory .
however , recently lamperti s transformation itself has been the focus of some research papers . in @xcite
several proofs of this result are displayed and in @xcite an extension of the transformation to the case of branching processes with immigration is obtained . in this work
we show an extension of lamperti s transformation to continuous time , integer valued , multitype branching processes .
more specifically , let @xmath2 be such a process issued from @xmath3 , then we shall prove that @xmath4 where @xmath5 are @xmath0 independent @xmath6-valued compound poisson processes .
as in the one dimensional case , absorption of @xmath7 at 0 means that in this transformation , the process @xmath8 must be stopped at some random time which , in the multitype case , will be defined as the ` first passage time ' at level @xmath9 by the multidimensional random field @xmath10\right)= x^{(1)}_{t_1}+\dots+x^{(d)}_{t_d}\,.\ ] ] multitype lamperti transformation is not invertible as in the one dimensional case .
however , by considering the whole branching structure behind the branching process , it is possible , to obtain a one - to - one relationship between @xmath11 and some decomposition of the @xmath0 dimensional process @xmath7 into @xmath1 processes , see theorem [ 2379 ] .
the proofs of these results pass through a special coding of multitype plane forests which leads to a lamperti type representation of discrete time , multitype branching processes .
results in discrete time are displayed and proved in section [ deterministic ] whereas the next section is devoted to the statements of our results in continuous time .
the latter will be proved in section [ proofs ] .
in all this work , we use the notation @xmath12 , @xmath13 and for any positive integer @xmath0 , we set @xmath14=\{1,\dots , d\}$ ] .
we also define the following partial order on @xmath15 by setting @xmath16 , if @xmath17 , for all @xmath18 $ ]
. moreover , we write @xmath19 if @xmath20 and if there exists @xmath21 $ ] such that @xmath22
. we will denote by @xmath23 the @xmath24-th unit vector of @xmath6 . + fix @xmath25 , let @xmath26 , where @xmath27 is any probability measure on @xmath6 such that @xmath28 and let @xmath2 be a @xmath0-type continuous time and @xmath6-valued branching process with progeny distribution @xmath26 and such that type @xmath21 $ ] has reproduction rate @xmath29 . for @xmath30 $ ] , the quantity @xmath31 corresponds to the mean number of children of type @xmath32 , given by an individual of type @xmath24 .
let @xmath33}$ ] be the mean matrix of @xmath7 . recall that if @xmath34 ( or equivalently @xmath35 ) is irreducible , then according to perron - frobenius theorem , it admits a unique eigenvalue @xmath36 which is simple , positive and with maximal modulus .
if moreover , @xmath35 is non degenerate , then extinction holds if and only if @xmath37 , see @xcite , @xcite and chapter v of @xcite . if @xmath38 , we say that @xmath7 is critical and if @xmath39 , we say that @xmath7 is subcritical .
+ we now define the underlying compound poisson process in the lamperti representation of @xmath7 that will be presented in theorems [ 2379 ] and [ 2415 ] .
let @xmath40 , where @xmath5 , @xmath21 $ ] are @xmath0 independent @xmath41-valued compound poisson processes .
we assume that @xmath42 and that @xmath5 has rate @xmath43 and jump distribution @xmath44 where @xmath45 in particular , with the notation @xmath46 , for all @xmath47 , the process @xmath48 is a @xmath49-valued , downward skip free , compound poisson process , i.e. @xmath50 , @xmath51 , with @xmath52 and for all @xmath53 , the process @xmath54 is a standard poisson process .
we emphasize that in this definition , some of the processes @xmath54 , @xmath55 $ ] can be identically equal to 0 .
+ we first present a result on passage times of the multidimensional random field @xmath56\right)= x^{(1)}_{t_1}+\dots+x^{(d)}_{t_d}\,,\ ] ] which is a particular case of additive lvy processes , see @xcite and the references therein .
henceforth , a process such as @xmath57 will be called an _ additive @xmath58downward skip free@xmath59 compound poisson process_. [ 6803 ] let @xmath60 .
then there exists a @xmath58unique@xmath59 random time @xmath61 such that almost surely , @xmath62 and if @xmath63 is any random time satisfying @xmath64 , then @xmath65 .
the time @xmath66 will be called the first passage time of the additive compound poisson process @xmath8 at level @xmath9 .
the process @xmath67 is increasing and additive , that is , for all @xmath68 , @xmath69 where @xmath70 is an independent copy of @xmath71 .
the law of @xmath66 on @xmath72 is given by @xmath73 where the support of this measure is included in @xmath74 .
note that from the additivity property ( [ 6290 ] ) of @xmath75 , we derive that the law of this process is characterised by the law of the variables @xmath76 for @xmath21 $ ] .
+ as the above statement suggests , some coordinates of the time @xmath66 may be infinite .
more specifically , we have : [ 1269 ] assume that @xmath35 irreducible and non degenerate . *
if @xmath35 is @xmath58sub@xmath59critical , then almost surely , for all @xmath3 and for all @xmath21 $ ] , @xmath77 . *
if @xmath35 is super critical , then for all @xmath3 , with some probability @xmath78 , @xmath79 , for all @xmath21 $ ] and with probability @xmath80 , @xmath77 , for all @xmath21 $ ] .
there are instances of reducible distributions @xmath35 such that for some @xmath3 , with positive probability , @xmath77 , for all @xmath81 and @xmath79 , for all @xmath82 , @xmath83 being a non trivial partition of @xmath14 $ ] .
it is the case for instance when @xmath84 , for @xmath85 , @xmath86 , @xmath87 and @xmath88 .
+ then we define @xmath0-type branching forests with edge lengths as finite sets of independent branching trees with edge lengths .
we say that such a forest is issued from @xmath60 ( at time @xmath89 ) , if it contains @xmath90 trees whose root is of type @xmath24 .
the discrete skeleton of a branching forest with edge lengths is a discrete branching forest with progeny distribution @xmath35 .
the edges issued from vertices of type @xmath24 are exponential random variables with parameter @xmath43 .
these random variables are independent between themselves and are independent of the discrete skeleton .
a realisation of such a forest is represented in figure [ fig1 ] .
then to each @xmath0-type forest with edge lengths @xmath91 is associated the branching process @xmath2 , where @xmath92 is the number of individuals in @xmath91 , alive at time @xmath93 . [ 2593 ] for @xmath53 , we denote by @xmath94 the total number of individuals of type @xmath32 whose parent has type @xmath24 and who were born before time @xmath93 .
the definition of @xmath95 for @xmath21 $ ] is the same , except that we add the number of roots of type @xmath24 and we subtract the number of individuals of type @xmath24 who died before time @xmath93 .
then we readily check that the branching process @xmath96 which is associated to this forest can be expressed in terms of the processes @xmath97 , as follows : @xmath98\,.\ ] ] the next theorem asserts that from a given @xmath0-type forest with edge lengths @xmath91 , issued from @xmath99 , we can construct a @xmath0-dimensional additive compound poisson process @xmath100,\,(t_1,\dots , t_d)\in\mathbb{r}_+^d)$ ] stopped at its first passage time of @xmath9 , such that the branching process @xmath7 associated to @xmath91 can be represented as a time change of @xmath57 .
this extends the lamperti representation to multitype branching processes . [
2379 ] let @xmath60 and let @xmath91 be a @xmath0-type branching forest with edge lengths , issued from @xmath101 , with progeny distribution @xmath35 and life time rates @xmath43 .
then the processes @xmath97 , @xmath55 $ ] introduced in definition @xmath102 admit the following representation : @xmath103 where @xmath104
are independent @xmath6-valued compound poisson processes with respective rates @xmath43 and jump distributions @xmath105 defined in @xmath106 , stopped at the first hitting time @xmath66 of @xmath9 by the additive compound poisson process , @xmath107,\,(t_1,\dots , t_d)\in\mathbb{r}_+^d\right)$ ] , i.e. @xmath108 in particular the multitype branching process @xmath7 , issued from @xmath109 admits the following representation , @xmath110 moreover , the transformation @xmath111 is invertible , so that the processes @xmath97 can be recovered from the processes @xmath5 .
conversely , the following theorem asserts that an additive compound poisson process being given , we can construct a multitype branching forest whose branching process is the unique solution of equation ( [ 2415 ] ) . [ 7462 ] let @xmath109 and @xmath104 be independent @xmath6 valued compound poisson processes with respective rates @xmath29 and jump distributions @xmath105 , stopped at the first hitting time @xmath66 of @xmath9 by the additive compound poisson process @xmath112\right)$ ] .
then there is a branching forest with edge lengths , with progeny distribution @xmath35 and rates @xmath29 such that the processes @xmath97 of definition @xmath102 satisfy relation @xmath111 .
moreover , the branching process @xmath2 associated to this forest is the unique solution of the equation , @xmath113 we emphasize that theorems [ 2379 ] and [ 7462 ] do not define a bijection between the set of branching forests with edge lengths and this of additive compound poisson processes , as in the discrete time case , see section [ deterministic ] .
indeed , in the continuous time case , when constructing the processes @xmath97 as in definition [ 2593 ] , at each birth time , we loose the information of the specific individual who gives birth .
in particular , the forest which is constructed in theorem [ 7462 ] is not unique .
this lost information is preserved in discrete time and the breadth first search coding that is defined in subsection [ codingforests ] allows us to define a bijection between the set of discrete forests and this of coding sequences .
we will denote by @xmath114 the set of plane forests . more specifically , an element @xmath115 is a directed planar graph with no loops on a possibly infinite and non empty set of vertices @xmath116 , with a _
finite _ number of connected components , such that each vertex has a finite inner degree and an outer degree equals to 0 or 1 .
the elements of @xmath114 will simply be called forests.the connected components of a forest are called the _ trees_. a forest consisting of a single connected component is also called a tree . in a tree @xmath117 , the only vertex with outer degree equal to 0 is called the _ root _ of @xmath117 .
it will be denoted by @xmath118 .
the roots of the connected components of a forest @xmath119 are called the roots of @xmath119 . for two vertices @xmath120 and @xmath121 of a forest @xmath119 , if @xmath122 is a directed edge of @xmath119 , then we say that @xmath120 is a _ child _ of @xmath121 , or that @xmath121 is the _ parent _ of @xmath120 .
the set @xmath123 of vertices of each forest @xmath119 will be enumerated according to the usual breadth first search order , see figure [ fig2 ] .
we emphasize that we begin by enumerating the roots of the forest from the left to the right .
in particular , our enumeration is not performed tree by tree .
if a forest @xmath119 contains at least @xmath124 vertices , then the @xmath124-th vertex of @xmath119 is denoted by @xmath125 . when no confusion is possible
, we will simply denote the @xmath124-th vertex by @xmath126 . +
a @xmath0-type forest is a couple @xmath127 , where @xmath115 and @xmath128 is an application @xmath129 $ ] . for @xmath130 , the integer @xmath131
is called the _ type _ ( or the _ color _ ) of @xmath121 .
the set of finite @xmath0-type forests will be denoted by @xmath132 .
an element @xmath133 will often simply be denoted by @xmath119 .
we assume that for any @xmath134 , if @xmath135 have the same parent , then @xmath136 .
moreover , if @xmath137 are the roots of @xmath119 , then @xmath138 . for each @xmath21 $ ] we will denote by @xmath139 the @xmath124-th vertex of type @xmath24 of the forest @xmath134 , see figure [ fig2 ] .
+ the aim of this subsection is to obtain a bijection between the set of multitype forests and some particular set of integer valued sequences which has been introduced in @xcite .
this bijection , which will be called a _ coding _ , depends on the breadth first search ordering defined at the previous subsection .
we emphasize that this coding is quite different from the one which is defined in @xcite .
[ 9076 ] let @xmath140 be the set of @xmath141^d$]-valued sequences , @xmath142 , such that for all @xmath18 $ ] , @xmath143 is a @xmath144-valued sequence defined on some interval of integers , @xmath145 , @xmath146 , which satisfies @xmath147 and if @xmath148 , then * for @xmath53 , the sequence @xmath149 is nondecreasing , * for all @xmath24 , @xmath150 , @xmath151 .
a sequence @xmath152 will sometimes be denoted by @xmath153)$ ] and for more convenience , we will sometimes denote @xmath154 by @xmath155 .
the vector @xmath156 , where @xmath157 will be called the length of @xmath101 . recall the definition of the order on @xmath15 , from the beginning of section [ main ] and let us set @xmath158:s_i<\infty\}$ ] , for any @xmath159 .
then the next lemma extends lemma 2.2 in @xcite to the case where the smallest solution of a system such as @xmath160 in ( [ 4379 ] ) may have infinite coordinates . [
1377 ] let @xmath152 whose length @xmath161 is such that @xmath162 for all @xmath24 @xmath58i.e .
@xmath163@xmath59 and let @xmath164
. then there exists @xmath165 such that @xmath166 @xmath58we will say that @xmath167 is a solution of the system @xmath58r , x@xmath168 and such that any other solution @xmath169 of @xmath160 satisfies @xmath170 .
moreover we have @xmath171 , for all @xmath172 .
the solution @xmath167 will be called _ the smallest solution _ of the system @xmath160 .
we emphasize that in @xmath173 , we may have @xmath174 , so that according to this definition , the smallest solution of the system @xmath160 may be infinite , that is @xmath175 for all @xmath21 $ ] .
note also that in @xmath173 it is implicit that @xmath176\setminus u_{\rm s}}x^{i , j}(s_i)<\infty$ ] , for all @xmath177 .
this proof is based on the simple observation that for fixed @xmath32 , when at least one of the indices @xmath178 s for @xmath53 increases , the term @xmath179 may only increase and when @xmath180 increases , the term @xmath181 may decrease only by jumps of amplitude @xmath182 . + first recall the notation @xmath183 , for @xmath159 introduced before lemma [ 1377 ] and set @xmath184 and for @xmath185 , @xmath186 where @xmath187 .
set also @xmath188 and @xmath189 $ ] . then note that since for @xmath53 , the @xmath190 s are positive and increasing , we have @xmath191 moreover , for each @xmath192 , @xmath193 define @xmath194 where in this definition , we consider that @xmath195 is satisfied for all @xmath196 if @xmath197 .
note that @xmath198 and @xmath199 , for all @xmath200 .
the index @xmath201 can be infinite and in general , we have @xmath202
. then the smallest solution of the system @xmath160 in the sense which is defined in lemma [ 1377 ] is @xmath203 .
+ indeed , ( [ 4379 ] ) is clearly satisfied for @xmath204 .
then let @xmath205 satisfying ( [ 4379 ] ) , that is @xmath206 we can prove by induction that @xmath207 , for all @xmath185 .
firstly for ( [ 5389 ] ) to be satisfied , we should have @xmath208 , for all @xmath209 , hence @xmath210 .
now assume that @xmath207 .
then @xmath211 and from ( [ 7399 ] ) and ( [ 5389 ] ) for each @xmath209 , @xmath212 , hence @xmath213 . + finally the fact that @xmath171 , for all @xmath172 readily follows from the above construction of @xmath214
. let @xmath215 , @xmath216 and denote by @xmath217 the number of children of type @xmath24 of @xmath120 .
for each @xmath18 $ ] , let @xmath218 be the number of vertices of type @xmath24 in @xmath123 .
then we define the application @xmath219 from @xmath132 to @xmath140 by @xmath220 where @xmath221 and for all @xmath21 $ ] , @xmath222 is the @xmath0-dimensional chain @xmath143 , with length @xmath223 , whose values belong to the set @xmath144 , such that @xmath224 and if @xmath148 , then @xmath225 we recall that @xmath226 is the @xmath124-th vertex of type @xmath24 in the breadth first search order of @xmath119
. we will see in theorem [ 5678 ] that @xmath227 is actually the smallest solution of the system @xmath160 , where @xmath228 is the number of roots of type @xmath24 of the forest @xmath119 .
this leads us to the following definition .
[ 9076 ] fix @xmath164 , such that @xmath229 .
* we denote by @xmath230 the subset of @xmath140 of sequences @xmath101 with length @xmath156 such that @xmath231 is the smallest solution of the system @xmath232 , where for all @xmath21 $ ] , @xmath233 , if @xmath234 and @xmath235 , if @xmath236 .
we will also say that @xmath231 is the smallest solution of the system @xmath160 .
* we denote by @xmath237 , the subset of @xmath132 of @xmath0-type forests containing exactly @xmath228 roots of type @xmath24 , for all @xmath21 $ ] .
the following theorem gives a one to one correspondence between the sets @xmath237 and @xmath230 .
[ 5678 ] let @xmath164 , be such that @xmath229 .
then for all @xmath238 , the chain @xmath239 belongs to the set @xmath230 .
moreover , the mapping @xmath240 is a bijection . in this proof , in order to simply the notation
, we will identify the sequence @xmath101 with its extension @xmath241 introduced in definition [ 9076 ] .
let @xmath127 be a forest of @xmath237 and let @xmath165 , where @xmath214 is the number of vertices of type @xmath24 in @xmath119 .
we define a subtree of type @xmath21 $ ] of @xmath127 as a maximal connected subgraph of @xmath127 whose all vertices are of type @xmath24 . formally , @xmath117 is a subtree of type @xmath24 of @xmath127 , if it is a connected subgraph whose all vertices are of type @xmath24 and such that either @xmath118 has no parent or the type of its parent is different from @xmath24 . moreover , if the parent of a vertex @xmath242 belongs to @xmath243 , then @xmath244 .
let @xmath21 $ ] and assume first that @xmath245 ( i.e. @xmath172 ) and let @xmath246 be the number of subtrees of type @xmath24 in @xmath119
. then we can check that the length @xmath214 of the sequence @xmath247 corresponds to its first hitting time of level @xmath248 , i.e. @xmath249 indeed , let us rank the subtrees of type @xmath24 in @xmath119 according to the breadth first search of their roots , so that we obtain the subforest of type @xmath24 : @xmath250 and let @xmath251 be the lukasiewicz - harris coding path of this subforest ( see , @xcite or @xcite for a definition of the lukasiewicz - harris coding path ) .
then we readily check that both sequences end up at the same level , i.e. @xmath252 note that if @xmath175 , then relation ( [ 4191 ] ) , whether or not @xmath178 is finite .
now let us check that @xmath167 is a solution of the system @xmath160 , that is @xmath253 let @xmath177 , then @xmath254 clearly represents the total number of vertices of type @xmath32 in @xmath123 which are either a root of type @xmath32 or whose parent is of a type different from @xmath32 , i.e. it represents the total number of subtrees of type @xmath32 in @xmath119 . on the other hand , from ( [ 4191 ] ) , @xmath255 is the number of these subtrees . we conclude that equation ( [ 5153 ] ) is satisfied .
it remains to check that @xmath167 is the smallest solution of the system @xmath160 . as in lemma
[ 1377 ] , set @xmath188 and for all @xmath256 $ ] , @xmath257 then from the proof of lemma [ 1377 ] , we have to prove that @xmath258 .
recall the coding which is defined in ( [ 6245 ] ) .
for all @xmath256 $ ] , once we have visited the @xmath259 first vertices of type @xmath32 which are actually roots of the forest , we have to visit the whole corresponding subtrees , so that , if the total number of vertices of type @xmath32 in @xmath127 is finite , i.e. @xmath177 , then the chain @xmath260 first hits @xmath261 at time @xmath262 .
then at time @xmath263 , an amount of @xmath264 more subtrees of type @xmath32 have to be visited .
so the chain @xmath260 has to hit @xmath265 at time @xmath266 .
this procedure is iterated until the last vertex of type @xmath32 is visited , that is until time @xmath267 ( note that the sequence @xmath268 is constant after some finite index ) . besides , from ( [ 7337 ] )
, we have @xmath269 on the other hand , if the total number of vertices of type @xmath32 in @xmath127 is infinite , then @xmath270 tends to @xmath271 ( it can be infinite by some rank ) . so that we also have @xmath272 in this case .
therefore , @xmath167 is the smallest solution of @xmath160 .
conversely let @xmath273 with length @xmath167 , then we construct a forest @xmath238 , generation by generation , as follows . at generation @xmath274 , for each @xmath21 $ ] , we take @xmath228 vertices of type @xmath24 .
we rank these @xmath275 vertices as it is defined in subsection [ space ] .
then to the @xmath276-th vertex of type @xmath24 , we give @xmath277 children of type @xmath256 $ ] if @xmath278 and @xmath279 children of type @xmath24 .
we rank vertices of generation @xmath280 and to the @xmath281-th vertex of type @xmath24 , we give @xmath282 children of type @xmath256 $ ] , if @xmath278 and @xmath283 children of type @xmath24 , and so on . proceeding this way , until the steps @xmath284 , @xmath55 $ ]
, we have constructed a forest .
indeed the total number of children of type @xmath32 whose parent is of type @xmath53 is @xmath285 , hence , the total number of children of type @xmath32 which is a root or whose parent is different from @xmath32 is @xmath254 .
moreover , each branch necessarily ends up with a leaf , since @xmath286 , for all @xmath53 and @xmath287 .
therefore we have constructed forest of @xmath237 .
finally , the mapping which is described in @xmath288 is clearly invertible , so we have proved that @xmath219 is a bijection from @xmath237 to @xmath230 .
@xmath289 to each @xmath290 we associate the chain @xmath291 indexed by the successive generations in @xmath119 , where for each @xmath21 $ ] , and @xmath185 , @xmath292 is the size of the population of type @xmath24 at generation @xmath124 .
more formally , we say that the ( index of the ) generation of @xmath216 is @xmath124 if @xmath293 , where @xmath294 is the tree of @xmath119 which contains @xmath120 , @xmath295 is the root of this tree and @xmath0 is the usual distance in discrete trees .
let us denote by @xmath296 the index of the highest generation in @xmath119 .
then @xmath297 is defined by @xmath298 let @xmath299 be the subset of @xmath123 of vertices which are not roots of @xmath119 .
we denote by @xmath300 the parent of any @xmath301 .
we also define the chains @xmath302 , for @xmath55 $ ] , as follows : @xmath303 if @xmath53 , @xmath304 , for all @xmath55 $ ] , @xmath305 if @xmath306 , and for @xmath185 , @xmath307 in words , if @xmath53 then @xmath308 is the total number of vertices of type @xmath32 whose parent is of type @xmath24 in the @xmath124 first generations of the forest @xmath119 .
if @xmath309 then we only count the number of vertices of type @xmath24 with at least one brother of type @xmath24 and whose parent is of type @xmath24 in the @xmath124 first generations . to this number
, we subtract the number of vertices of type @xmath24 with no children of type @xmath24 , whose generation is less or equal than @xmath310 .
then it is not difficult to check the following relation : @xmath311\,.\ ] ] we end this subsection by a lemma which provides a relationship between the chains @xmath101 and @xmath312 and @xmath302 .
this result is the deterministic expression of the lamperti representation of theorem [ 5891 ] below and its continuous time counterpart in theorems [ 2379 ] and [ 7462 ] . [
6209 ] the chain @xmath302 may be obtained as the following time change of the chain @xmath190 : @xmath313\,,\;\;i\neq j\,,\label{8377}\\ z^{i , i}(n)&=&r_i+x^{i , i}(\mbox{$\sum_{k=0}^{n-1}z^{(i)}(k)$})\,,\;\;\;n\ge1\,,\;\;\;i\in[d]\,.\label{8517}\end{aligned}\ ] ] in particular , we have @xmath314\,.\ ] ] moreover , given @xmath190 , @xmath55 $ ] , the chains @xmath302 , @xmath55 $ ] are uniquely determined by equations @xmath315 , @xmath316 and @xmath317 .
it suffices to check relations ( [ 8377 ] ) and ( [ 8517 ] ) .
then ( [ 9537 ] ) will follow from ( [ 2671 ] ) .
but ( [ 8377 ] ) and ( [ 8517 ] ) are direct consequences of the definition of the chains @xmath190 and @xmath302 in ( [ 6245 ] ) and ( [ 2472 ] ) respectively .
@xmath289 recall that @xmath26 is a progeny distribution , such that the @xmath27 s are any probability measures on @xmath6 , such that @xmath28 .
let @xmath318 be some measurable space on which , for any @xmath319 such that @xmath229 , we can define a probability measure @xmath320 and a random variable @xmath321 whose law under @xmath320 is this of a branching forest with progeny law @xmath35 .
then we construct from @xmath322 the random chains , @xmath40 , @xmath2 and @xmath97 , @xmath55 $ ] , exactly as in ( [ 6245 ] ) , ( [ 8477 ] ) and ( [ 2472 ] ) respectively .
in particular , @xmath323 . we can check that under @xmath320 , @xmath7 is a branching process with progeny distribution @xmath35 . more specifically ,
recall from ( [ 4590 ] ) the definition of @xmath324 , then the random processes @xmath8 and @xmath7 satisfy the following result . [ 5891 ] let @xmath319 be such that @xmath229 and let @xmath322 be an @xmath237-valued branching forest with progeny law @xmath35 under @xmath320 .
let @xmath325 , where @xmath326 is the number of vertices of type @xmath24 in @xmath91.then , * the random variable @xmath327 is almost surely the smallest solution of the system @xmath328 in the sense of definition @xmath329 .
if @xmath35 is irreducible , non degenerate and @xmath58sub@xmath59critical , then almost surely , @xmath330 , for all @xmath21 $ ] . if @xmath35 is irreducible , non degenerate and super critical , then with some probability @xmath78 , @xmath331 , for all @xmath21 $ ] and with probability @xmath80 , @xmath330 , for all @xmath21 $ ] . * on the space @xmath318 , we can define independent random walks @xmath332 , @xmath21 $ ] , with respective step distributions @xmath324 , @xmath21 $ ] , such that @xmath333 and if @xmath334 is the smallest solution of the system @xmath335 , then the following identity in law @xmath336){\stackrel{\mbox{\tiny $ ( d)$}}{=}}(\tilde{x}^{(i)}_k,\,0\le k\le \tilde{n}_i,\,i\in [ d])\ ] ] holds . *
the joint law of @xmath8 and @xmath327 is given as follows : for any integers @xmath223 and @xmath337 , @xmath30 $ ] , such that @xmath338 , @xmath339 , for @xmath53 , @xmath340 and @xmath341 , @xmath342 $ and $ n={\rm n}$}\big)=\\ & & \qquad\qquad\frac{\mbox{\rm det}(-k_{ij})}{n_1n_2\dots n_d}\prod_{i=1}^d\nu_i^{*n_i}\big(k_{i1},\dots , k_{i(i-1)},n_i+k_{ii},k_{i(i+1)},\dots , k_{id}\big)\,.\label{5478}\end{aligned}\ ] ] * the random process @xmath7 is a branching process with progeny law @xmath35 , which is related to @xmath8 through the time change : @xmath343 the fact that @xmath327 is the smallest solution of the system @xmath328 is a direct consequence of theorem [ 5678 ] and the definition of @xmath8 , that is @xmath344 .
assume that @xmath35 is irreducible , non degenerate and ( sub)critical .
then since the forest @xmath91 contains almost surely a finite number of vertices , all coordinates of @xmath327 must be finite from theorem [ 5678 ] .
if @xmath35 is irreducible , non degenerate and supercritical , then with probability @xmath78 the forest @xmath91 contains an infinite number of vertices of type @xmath24 , for all @xmath21 $ ] and with probability @xmath80 its total population is finite
. then the result also follows from theorem [ 5678 ] . in order to prove part 2 ,
let @xmath345 with @xmath346 , be a sequence of independent and identically distributed forests .
let us define @xmath347 and then let us concatenate the processes @xmath348 in a process that we denote @xmath349 .
more specifically , let us denote by @xmath350 the length of @xmath351 , then the process obtained from this concatenation is @xmath352 , where @xmath353 , @xmath354 and @xmath355 note that @xmath349 is obtained by coding the forests @xmath345 , @xmath185 successively .
then it readily follows from the construction of @xmath349 and the branching property that the coordinates @xmath332 are independent random walks with step distribution @xmath324 . moreover @xmath327 is a solution of the system @xmath335 , so its smallest solution , say @xmath356 , is necessarily smaller than @xmath327 .
this means that @xmath356 is a solution of the system @xmath328 , hence @xmath357 .
the third part is a direct consequence of the first part and the multivariate ballot theorem which is proved in @xcite , see theorem 3.4 therein .
then part 4 , directly follows from the definition of @xmath7 and lemma [ 6209 ] .
@xmath289 + conversely , from any random walk whose step distribution is given by ( [ 4590 ] ) we can construct a unique branching forest with law @xmath320 . the following result is a direct consequence of theorems [ 5891 ] and [ 5678 ] . [ 1259 ] let @xmath332 , @xmath21 $ ] be @xmath0 independent random walks defined on @xmath318 , whose respective step distributions are @xmath324 , and set @xmath352 .
let @xmath319 such that @xmath229 and let @xmath358 be the smallest solution of the system @xmath335 .
we define the @xmath359-valued process @xmath40 by @xmath360 .
then @xmath361 is a @xmath237-valued branching forest @xmath322 with progeny distribution @xmath35 .
let @xmath362 be the underlying random walk of the compound poisson process @xmath8 , that is the random walk such that there are independent poisson processes @xmath363 , with parameters @xmath43 such that @xmath364 and @xmath365)$ ] are independent .
then from lemma [ 1377 ] , there is a random time @xmath159 , such that almost surely , for all @xmath177 , @xmath366 .
moreover , if @xmath367 is any time satisfying this property , then @xmath368 . for @xmath21\setminus
u_{\rm s}$ ] , the latter equality implies that @xmath369 . since @xmath370 is a renewal process , it is possible only if @xmath371 , a.s .
, so that we can write : almost surely , @xmath372 then the first part of the theorem follows from the construction of @xmath8 as a time change of @xmath373
. more formally , the coordinates of @xmath66 are given by @xmath374 , so that in particular , @xmath375 .
additivity property of @xmath66 is a consequence of lemma [ 1377 ] and time change . from this lemma
, we can deduct that for all @xmath68 , if @xmath167 is the smallest solution of @xmath376 , then conditionally to @xmath245 , for all @xmath21 $ ] , the smallest solution @xmath377 of @xmath378 , and satisfies @xmath379 and @xmath380 is the smallest solution of the system @xmath381 , where @xmath382 , @xmath383 .
moreover , @xmath384)$ ] has the same law as @xmath373 and is independent of @xmath385 . using the time change ,
we obtain , @xmath386 where @xmath387 has the law of @xmath66 conditionally on @xmath388 , for all @xmath21 $ ] and @xmath389 is an independent copy of @xmath390 . then identity ( [ 6290 ] ) follows .
a @xmath0 type forest with edge lengths is an element @xmath391 , where @xmath215 and @xmath392 is some application @xmath393 . for @xmath216 , the quantity @xmath394
will be called the life time of @xmath120 .
it corresponds to the length of an edge incident to @xmath120 in @xmath391 whose color is this of @xmath120 .
this edge is a segment which is closed at the extremity corresponding to @xmath120 and open at the other extremity .
if @xmath120 is not a leaf of @xmath127 then @xmath394 corresponds to the length of the edge between @xmath120 and its children in the continuous forest @xmath391 . to each tree of @xmath127
corresponds a tree of @xmath391 which is considered as a continuous metric space , the distance being given by the lebesgue measure along the branches . to each forest @xmath391
we associate a time scale such that a vertex @xmath120 is born at time @xmath93 if the distance between @xmath120 and the root of its tree in @xmath391 is @xmath93 .
time @xmath93 is called the birth time of @xmath120 in @xmath391 and it is denoted by @xmath395 .
the death time of @xmath120 in @xmath391 is then @xmath396 .
if @xmath397 then we say that @xmath120 is alive at time @xmath398 in the forest @xmath391 .
we denote by @xmath399 the smallest time when no individual is alive in @xmath391 .
the set of @xmath0 type forests with edge lengths will be denoted by @xmath400 .
the subset of @xmath400 of elements @xmath391 such that @xmath238 will be denoted by @xmath401 .
elements of @xmath400 will be represented as in figure [ fig1 ] .
+ to each forest @xmath402 , we associate the multidimensional the step functions , @xmath403 that are defined as follows : we now define the discretisation of forests of @xmath400 , with some span @xmath408 .
let @xmath409 , then on each tree of @xmath409 , we place new vertices at distance @xmath410 , @xmath411 of the root along all the branches .
a vertex which is placed along an edge with color @xmath24 has also color @xmath24 .
then we define a forest in @xmath132 as follows .
a new vertex @xmath121 is the child of a new vertex @xmath120 if and only if both vertices belong to the same branch of @xmath391 , and there is @xmath192 such that @xmath120 and @xmath121 are respectively at distance @xmath410 and @xmath412 from the root .
this transformation defines an application which we will denote by @xmath413 note that with this definition , the roots of the three forests @xmath127 , @xmath391 and @xmath414 are the same and more generally , a vertex of @xmath414 corresponds to a vertex @xmath120 of @xmath391 if and only if @xmath120 is at a distance equal to @xmath410 from the root .
the definition of the discretisation of a forest with edge lengths should be obvious from figure [ fig3 ] .
recall from section [ main ] that @xmath415 are positive , finite and constant rates , and that @xmath26 , where @xmath27 , @xmath21 $ ] are any distributions in @xmath6 . from the setting established in subsection [ 2312 ]
, we can define a branching forest with edge lengths as a random variable @xmath416 , where @xmath417 is the sigma field of the borel sets of @xmath400 endowed with the gromov - hausdorff topology ( see section 2.1 in @xcite ) and where the law of @xmath321 under @xmath320 is this of a discrete branching forest with progeny distribution @xmath35 , as defined in subsection [ 6477 ] .
besides , let @xmath326 be the number of vertices of type @xmath24 in @xmath322 , then for all @xmath156 , conditionally on @xmath418 , @xmath21 $ ] , @xmath419 are sequences of i.i.d .
exponentially distributed random variables with respective parameters @xmath43 , and @xmath420,\,(f , c_f)]$ ] are independent .
+ then we have the following result which is straightforward from the above definitions . [ 7242 ] let @xmath421 be a branching forest with edge lengths with progeny distribution @xmath26 on @xmath6 and life time rates @xmath422 .
then for all @xmath423 , under @xmath320 , the process @xmath424\,,t\ge0)$ ] which is defined as in @xmath425 with respect to @xmath421 is a continuous time , @xmath6-valued branching process starting at @xmath426 , with progeny distribution @xmath26 and life time rates @xmath415 .
[ 4166 ] let @xmath421 be a branching forest with edge lengths , with progeny distribution @xmath26 on @xmath6 and life time rates @xmath422 . for @xmath408 ,
the forest @xmath427 is a @xmath58discrete@xmath59 branching forest with progeny distribution : @xmath428 where @xmath23 is the @xmath24-th unit vector of @xmath6 .
the fact that @xmath427 is a discrete branching forest is a direct consequence of its construction .
indeed , at generation @xmath124 , that is at time @xmath410 , the vertices of this forest inherits from the branching property of @xmath421 the fact they will give birth to some progeny , independently of each other and with some distribution which only depends on their type and @xmath429 . then its remains to determine the progeny distribution @xmath430 .
but it is obvious from the construction of @xmath427 that @xmath431 is the law of the total offspring at time @xmath429 of a root of type @xmath24 in the forest @xmath421 .
this is precisely the expression which is given in the statement .
let @xmath421 be a branching forest issued from @xmath101 , with edge lengths , with progeny distribution @xmath26 on @xmath6 and life time rates @xmath422 .
then from proposition [ 7242 ] , the process @xmath432,\,t\ge0)$ ] which is defined as in @xmath425 with respect to @xmath421 is a continuous time , @xmath433-valued branching process with progeny distribution @xmath26 and life time rates @xmath415 .
+ let @xmath408 and consider the discrete forest @xmath427 whose progeny distribution is given by lemma [ 4166 ] .
let @xmath434 be the associated ( discrete time ) branching process and let @xmath435 , @xmath21 $ ] be the decomposition of @xmath436 , as it is defined in ( [ 2472 ] ) .
then it is straightforward that @xmath437)\,,t\ge0)\rightarrow ( z^{i , j}_t\,,t\ge0)\,,\ ] ] almost surely on the skohorod s space of cdlg paths toward the process @xmath97 , as @xmath429 tends to 0 , for all @xmath55 $ ] , where @xmath97 is the decomposition of @xmath7 as it is defined in ( [ 2847 ] ) .
+ now , let @xmath438)$ ] be the coding random walk associated to @xmath427 , as in theorem [ 5891 ] .
we will first assume that @xmath439 is actually the coding random walk of a sequence of i.i.d .
discrete forests with the same distribution as @xmath427 in the same manner as in the proof of part 2 of theorem [ 5891 ] , so that in particular , @xmath439 is not a stopped random walk .
+ for @xmath21 $ ] , let @xmath440 and @xmath441 be the sequences of jump times of the discrete time processes @xmath442 and @xmath443 .
that is the ordered sequences of times such that @xmath444 and @xmath445 and @xmath446 , for @xmath185 .
fix @xmath447 , then since two jumps of the process @xmath7 almost surely never occur at the same time , there is @xmath448 , sufficiently small such that for all @xmath449 , the sequences @xmath450 and @xmath451 are a.s .
identical , for all @xmath21 $ ] .
moreover , from lemma [ 4166 ] and theorem [ 5891 ] , the jumps @xmath452 have law @xmath453 where @xmath454 .
the measure @xmath455 converges weakly to @xmath105 defined in ( [ 7458 ] ) , as @xmath456 .
hence from ( [ 5792 ] ) , the sequence @xmath457 converges almost surely toward the sequence @xmath458 of jumps of the process @xmath459\,,t\ge0)$ ] , which is therefore a sequence of i.i.d .
random variables , with law @xmath105 .
+ on the other hand , it follows from ( [ 8377 ] ) and ( [ 8517 ] ) in lemma [ 6209 ] and the fact that two jumps of the process @xmath7 almost surely never occur at the same time , that for all @xmath460 , there is @xmath461 sufficiently small , such that for all @xmath462 and @xmath463 , @xmath464 from lemma [ 4166 ] , the latter is a sequence of i.i.d .
geometrically distributed random variables with parameter @xmath465 . hence from ( [ 5792 ] ) , the sequence @xmath466 converges almost surely toward the sequence @xmath467 as @xmath456 , where @xmath468 is the sequence of jump times of @xmath469 .
moreover the variables of this sequence are i.i.d . and exponentially distributed with parameter @xmath470 .
+ then since @xmath471 is a random walk , the sequences @xmath472 , @xmath473 , @xmath21 $ ] are independent . therefore , from the convergences proved above , the sequences @xmath458 and @xmath474 are independent . then we have proved that the process @xmath475 is a compound poisson process with parameter @xmath43 and jump distribution @xmath105 .
moreover , it follows from the independence between the random walks @xmath471 , @xmath21 $ ] that the processes @xmath476 , @xmath21 $ ] are independent .
+ now from part 1 of theorem [ 5891 ] , if @xmath477 is the total population of type @xmath21 $ ] in the forest @xmath427 , then @xmath478 is the smallest solution of the system @xmath479 .
moreover it follows from the construction of @xmath427 , that @xmath480 , almost surely .
note that @xmath481 represents the total length of the branches of type @xmath24 in the forest @xmath421 .
then from the definition of @xmath5 in ( [ 1547 ] ) it appears the these compound poisson processes are stopped at @xmath482 and that @xmath483 satisfies ( [ 3678 ] ) .
+ the fact that @xmath111 is invertible is a direct consequence the first part of the following lemma .
[ 1592]let @xmath60 and @xmath484\}$ ] be a family of cdlg @xmath49-valued , step functions such that for @xmath53 , @xmath485 , @xmath190 are increasing , @xmath486 and @xmath247 are downward skip free , i.e. @xmath487 , for all @xmath51 , with @xmath488 .
then there exists a @xmath58unique@xmath59 time @xmath489 such that @xmath490 and if @xmath491 is any time satisfying @xmath492 , then @xmath493 .
moreover , the system @xmath494 admits a unique solution @xmath495\}$ ] and the system @xmath496\ ] ] admits a unique solution @xmath497)$ ] .
these solutions are functionals of stopped functions @xmath498\}$ ] .
the proof of the first part of the lemma can be done by applying lemma [ 1377 ] to the discrete time skeleton of the functions @xmath484\}$ ] , exactly as for the proof of the first part of theorem [ 6803 ] , see subsection [ 1842 ] .
then the proof of the existence and uniqueness of the solutions of both systems is straightforward .
let @xmath499 , @xmath185 be the discrete , ordered sequence of jump times of the processes @xmath484\}$ ] ( note that two functions @xmath190 can jump simultaneously ) . then for each of these two systems the solution can be constructed in between the times @xmath499 and @xmath500 in a unique way .
@xmath501 let @xmath502 be a family of independent random variables , such that for each @xmath276 , @xmath503 is uniformly distributed on @xmath504 $ ] .
we assume moreover that the family @xmath502 is independent of the compound poisson process @xmath40 . then let us construct a multitype branching forest with edge lengths in the following way .
let @xmath505 , @xmath185 be the sequence of ordered jump times of the process @xmath5 .
we first start with @xmath90 vertices of type @xmath21 $ ] .
let @xmath24 such that @xmath506\}$ ] .
then we construct the branches issued from each vertex of type @xmath256 $ ] and , in the scale time of our forest in construction , at time @xmath507 , we choose among the @xmath90 vertices of type @xmath24 according to @xmath508 the vertex who gives birth
. then the construction is done recursively .
let @xmath509 be the number of vertices of type @xmath256 $ ] in the forest at time @xmath507 and let @xmath362 such that @xmath510 corresponds to @xmath511 shifted at time @xmath512 , i.e. @xmath513 . then let @xmath514 , @xmath185 be the sequence of ordered jump times of the process @xmath510 , and let @xmath276 such that @xmath515\}$ ] .
then we continue the construction of the branches issued from each vertex of type @xmath256 $ ] and at time @xmath516 , we choose among the @xmath517 vertices of type @xmath276 according to @xmath518 the vertex who gives birth .
this construction is performed until all processes @xmath5 are stopped at time @xmath481 .
it is clear from this construction that the forest which is obtained is a multitype branching forest with edge lengths , with the required distribution and such that the processes @xmath97 defined as in definition [ 2593 ] with respect to this forest satisfy equation ( [ 3521 ] ) .
athreya and p. e. ney : _ branching processes .
die grundlehren der mathematischen wissenschaften , band 196 .
springer - verlag , new york - heidelberg , 1972 .
caballero , a. lambert and g. uribe bravo : proof(s ) of the lamperti representation of continuous - state branching processes .
_ 6 , 6289 , ( 2009 ) .
caballero , j.l .
prez and g. uribe bravo : a lamerti - type representation of continuous - state branching processes with immigration .
3a , 15851627 , ( 2013 ) .
l. chaumont and r. liu : coding multitype forests : application to the law of the total population of branching forests . to appear in _ transactions of the american mathematical society _ , ( 2014 ) .
harris : _ the theory of branching processes .
dover phoenix editions .
dover publications , inc . ,
mineola , ny , 2002 .
j. lamperti continuous state branching processes .
_ 73 , 382386 , ( 1967 ) .
le gall : random trees and applications .
_ , * 2 * , 245311 , ( 2005 ) .
d. khoshnevisan and y. xiao : harmonic analysis of additive lvy processes .
_ probab . theory related fields _ , 145 , no .
3 - 4 , 459515 , ( 2009 ) .
mode : _ multitype branching processes . theory and applications .
modern analytic and computational methods in science and mathematics , no .
34 american elsevier publishing co. , inc .
, new york 1971 .
s. watanabe : on two dimensional markov processes with branching property . _ trans .
_ 136 , 447466 , ( 1969 ) . | we obtain a bijection between some set of multidimensional sequences and this of @xmath0-type plane forests which is based on the breadth first search algorithm .
this coding sequence is related to the sequence of population sizes indexed by the generations , through a lamperti type transformation .
the same transformation in then obtained in continuous time for multitype branching processes with discrete values .
we show that any such process can be obtained from a @xmath1 dimensional compound poisson process time changed by some integral functional .
our proof bears on the discretisation of branching forests with edge lengths .
dedicated to the memory of marc yor |
we consider , for arbitrary but fixed @xmath5 @xmath6 the reaction
diffusion problem @xmath7 for @xmath8 @xmath9 with initial conditions @xmath10
@xmath11 satisfying @xmath12 the choice of the initial time @xmath13 and a class of initial conditions @xmath14 @xmath15 will be described later on , but for the purpose of this introduction it is useful to have in mind the `` natural '' case : @xmath16 @xmath17 for @xmath18 @xmath19 for @xmath20 and @xmath21 for @xmath20 @xmath22 for @xmath23 this initial value problem models the time evolution of a chemical system of two ( initially separated ) substances @xmath0 and @xmath24 that diffuse in some substratum and react according to @xmath25 with a substance @xmath26 that is supposed not to participate in the reaction anymore .
the model is a mean field description of such a situation where the functions @xmath27 and @xmath28 represent the densities of the substances @xmath0 and @xmath29 for more details see @xcite . equations ( [ system1 ] ) and ( [ system2 ] ) are best studied in terms of the sum @xmath30 and the difference @xmath31 which satisfy the equations @xmath32 with initial conditions @xmath33 and @xmath34 ( at time @xmath35 that satisfy @xmath36 and @xmath37 for initial conditions @xmath38 @xmath15 with @xmath39 the functions @xmath33 and @xmath34 are even and odd , respectively , and the equations ( [ equ ] ) , ( [ eqv ] ) preserve this symmetry . furthermore , for the special initial condition
@xmath40 with @xmath41 defined by the equation @xmath42 equation ( [ equ ] ) has the explicit solution @xmath43 we note that the initial condition ( [ initsym ] ) for @xmath44 ( at time @xmath45 ) is simply the solution of equation ( [ equ ] ) with the `` natural '' initial condition , @xmath46 for @xmath18 @xmath47 for @xmath20 evaluated at @xmath48 to keep this paper as simple as possible we now restrict the discussion to this case , i.e. , we consider from now on equation ( [ eqv ] ) with initial conditions satisfying ( [ limitv ] ) , and @xmath44 given by ( [ uxt ] ) .
we note , however , that more general ( asymmetric ) initial conditions for @xmath44 could be treated as well .
this would lead to corrections to @xmath44 of the order @xmath49 and such corrections do not change in any way the discussion of the equation for @xmath50 that follows .
the reaction diffusion problems considered here develop , in addition to the built in diffusive length scale @xmath51 an additional shorter length scale , on which the reaction takes place .
the function @xmath52@xmath53 is called the reaction term or reaction front , and we are interested in describing the asymptotic behavior of the function @xmath54 for large times .
the knowledge of this behavior is useful , since it appears to be universal , in the sense that it is largely independent of the choice of the initial conditions and of the details of the model under consideration . as mentioned above , if @xmath33 is an even function , then @xmath50 and as a consequence @xmath54 are even functions of @xmath55 we will see that the critical point of @xmath54 at @xmath56 is a maximum , and that @xmath54 decays ( rapidly ) for large @xmath55 before proceeding any further we note that the factor of @xmath57 in ( [ system1 ] ) , ( [ system2 ] ) and ( [ reaction001 ] ) is just a normalization , and has been chosen for convenience to make the equation ( [ eqv ] ) for @xmath50 look simple .
in fact , any system of the form @xmath58 with positive @xmath59 @xmath60 @xmath61 and @xmath62 and with initial conditions such that @xmath63 can be reduced , by scaling space and time and the amplitudes , to the problem @xmath64 with @xmath65 and with initial conditions such that @xmath66 in this paper we have limited the discussion to the case @xmath67 and @xmath68 the case @xmath69 leads to a moving reaction front .
a change of coordinates to a moving frame complicates the analysis , but the problem could still be treated with the methods presented here .
choosing @xmath70 makes the equations mathematically simpler . as a consequence , as we have seen
, the two equations for @xmath27 and @xmath28 can be reduced to just one equation for the sum @xmath71 since the equation for the difference @xmath72 can be solved explicitly . even though we do not expect the asymptotic behavior of the solution to change in any relevant way if @xmath73 , the strategy of proof would have to be changed considerably , since the equations can not be decoupled anymore in that case .
before we state our results , we briefly discuss the expected dependence of the results on the parameter @xmath74 the case @xmath3 has been studied in detail in @xcite , where it is proved that in this case the reaction term ( [ reaction001 ] ) satisfies , for all @xmath75 @xmath76 where @xmath77 and @xmath78 and where @xmath79 is a smooth function that decays like @xmath80 for large values of @xmath81 it follows furthermore from the results in @xcite that the function @xmath54 is very small on the diffusive scale in the sense that for @xmath82 @xmath83 and all @xmath84@xmath85 the smallness of @xmath54 on the diffusive scale is easily understood by realizing that , for @xmath3 and for positive values of @xmath86 on the diffusive scale , i.e. , for @xmath87 equation ( [ system1 ] ) essentially reduces to @xmath88 with @xmath89 therefore , the function @xmath27 decays exponentially fast to zero on this scale , and similarly for @xmath28 for negative values of @xmath55 for @xmath90 however , equation ( [ system1 ] ) reduces , for @xmath87 essentially to @xmath91 with @xmath89 the solution of ( [ a002 ] ) has an asymptotic behavior that is radically different from the solution of ( [ a001 ] ) .
in particular , for @xmath92 , the solution may even blow up in finite time if @xmath27 is not a positive function . note that , for @xmath93 odd , the nonlinear term in ( [ a002 ] ) is always a `` friction term '' , independent of the sign of @xmath94 and the case of @xmath93 odd will therefore turn out to be easier to treat than the case of @xmath93 even .
it is well known @xcite that for @xmath95 and small bounded integrable initial conditions , the nonlinearity in ( [ a002 ] ) becomes irrelevant for large times in the sense that the solution converges to a multiple of @xmath96 which solves the linear equation @xmath97 we would therefore expect that , for @xmath4 the function @xmath54 is of the order @xmath98 on the diffusive scale .
this turns out to be wrong . as we will prove below
, @xmath54 is of the order @xmath99 for @xmath4 because @xmath54 converges on this scale pointwise to a function that is not integrable at the origin .
this corresponds to a solution of ( [ a002 ] ) for which the nonlinear term is a marginal perturbation , i.e. , a solution with an amplitude of the order @xmath100 .
we will see that one can take advantage of this fact , and a diffusive stability bound will be good enough to prove convergence of @xmath54 to its limit , but as a consequence , our results will be limited to the case @xmath101 the cases @xmath92 and @xmath102 are special and will not be discussed any further .
the following theorem is our main result .
[ main]for arbitrary but fixed @xmath5 @xmath103 there exist @xmath104 functions @xmath105
@xmath106 @xmath107 @xmath108 and a class of initial conditions ( specified at @xmath109 such that ( [ eqv ] ) has a unique solution @xmath50 that satisfies for all @xmath110 the bound @xmath111 where @xmath112 @xmath113 @xmath114 and @xmath115 this theorem is a local result , in the sense that the class of initial conditions will be a set of functions in a ( small ) neighborhood of the function @xmath116 @xmath117 in particular , our methods do not allow us to show that the solution with the `` natural '' initial condition @xmath118 at @xmath119 belongs to this set at @xmath48 we do expect , however , that this is the case , as has been proved for @xmath3 in @xcite .
we note that , if an initial condition @xmath33 is such that @xmath120 for a certain @xmath121 then @xmath122 , if @xmath20 or @xmath123 if @xmath23 a priori , we do not need to consider such initial conditions , since in our model @xmath27 and @xmath28 represent particle densities , and the solutions @xmath27 and @xmath28 are positive if the initial conditions @xmath124 and @xmath15 are positive .
as we will see , for @xmath103 it will not be necessary to impose that @xmath124 and @xmath15 be positive everywhere , and it will neither be necessary to impose that @xmath125 be an even function . as we will see
, the functions @xmath126 and @xmath127 are small on the diffusive scale , i.e. , for @xmath128 @xmath129 and @xmath130 large , @xmath131 where @xmath132 if @xmath133 and @xmath134 if @xmath135 @xmath136 using the definition ( [ defv ] ) , ( [ defu ] ) for @xmath50 and @xmath137 we therefore find that for @xmath138 and @xmath130 large , @xmath139 and similarly for @xmath140 for @xmath141 in contrast to the case @xmath82 where only exponentially few particles reach the diffusive scale , the amount of particles decays only slowly for @xmath101 our results imply that , for large times , the density of the remaining particles is given by the function @xmath142 i.e. , it is independent of the initial conditions . as a corollary to theorem [ main ]
we get a precise description of the reaction front @xmath54 on the reactive and the diffusive scale .
this description will be given in section 4 , once we have defined the functions @xmath143
@xmath144 @xmath126 and @xmath127 in section 3 .
in section 2 we explain our strategy for proving theorem [ main ] .
this strategy is implemented in section 5 and section 6 .
the appendix contains the proof of the existence of the functions @xmath145 @xmath126 and @xmath127 .
consider functions @xmath50 of the form @xmath146 with @xmath147 as in theorem [ main ] , and @xmath148 for some @xmath149 with @xmath150 substituting ( [ ansatz ] ) into ( [ eqv ] ) leads to an equation for the function @xmath151 of the form @xmath152 for certain functions @xmath153 and @xmath154 and for @xmath155 some nonlinear map .
we will show that if @xmath147 is defined correctly , @xmath156 large enough and @xmath157 small enough , then @xmath153 can be chosen positive and @xmath155 will be small , so that the solution of equation ( [ eqpsi ] ) will be bounded for large times by the corresponding solution of the inhomogeneous heat equation @xmath158 we will find that , with the right choice of @xmath159@xmath160 from which the bound ( [ bound ] ) will follow .
we note that @xmath161 for @xmath162 so that contributions of initial conditions will become irrelevant for large times , i.e. , the solution @xmath151 becomes what is called `` slaved to the inhomogeneous term '' .
in order to implement the strategy outlined in section 2 , we need a function @xmath163 that approximates the solution @xmath50 for large times sufficiently well , uniformly in @xmath55 since we would like to control the time evolution of equation ( [ eqpsi ] ) on @xmath164 this function @xmath163 needs to satisfy @xmath165 in order for @xmath50 to satisfy the boundary conditions ( [ limitv ] ) .
furthermore , the inhomogeneous term @xmath166 in equation ( [ eqpsi ] ) contains second derivatives of @xmath167 and the function @xmath166 can therefore only be in @xmath168 if @xmath163 is at least twice differentiable .
we now construct a function @xmath163 satisfying these requirements through a two length scale asymptotic expansion . to simplify the notation later on we use the convention that , unless stated otherwise , @xmath169 and @xmath170 and we will refer to @xmath171 as the diffusive length scale and to @xmath172 as the reactive length scale . the function @xmath147 is given by the first and second order terms of a so called `` matched asymptotic expansion '' .
the `` matched '' refers to the fact that such an expansion contains functions that can not be associated uniquely with one of the length scales and can therefore be used to `` match '' the behavior at large distances of the shorter scale with the behavior at short distances of the larger scale .
let @xmath173 and let @xmath174 @xmath175 and @xmath176 be as in theorem [ main ] .
then , the functions @xmath177 @xmath178 @xmath179 and @xmath180 are of this form and will naturally show up in the function @xmath181 as a consequence , the representation ( [ vinfty ] ) for @xmath147 is not unique .
if we choose ( as we will ) to compute the expansion for @xmath147 in the order of decreasing amplitudes , i.e. , if we first compute the term of order @xmath182 then the term of order @xmath183 and finally the term of order @xmath184 we get a representation of @xmath147 of the form @xmath185 where @xmath186 with @xmath187 a certain constant to be determined later .
we note that , by definition , @xmath147 is a symmetric function , and it is therefore sufficient to consider positive values of @xmath86 if we choose appropriate boundary conditions at @xmath56 to ensure regularity . finally , since we will need to describe the asymptotic behavior of various functions near zero and infinity , we introduce the following notation .
let @xmath188 be a continuous function from @xmath189 to @xmath190 @xmath191 a positive integer and @xmath192 real numbers .
then , we say that @xmath193 near @xmath194 if @xmath195 and we say that @xmath196 near @xmath197 if @xmath198 to lowest order the function @xmath50 is asymptotic to @xmath200 , with @xmath41 as defined in ( [ defmu ] ) .
we note that @xmath41 has near @xmath201 the expansion @xmath202 where @xmath203 and @xmath204 .
furthermore , @xmath205 we make the ansatz @xmath207 which we substitute into equation ( [ eqv ] ) .
we multiply the resulting equation with @xmath208 and take then the limit @xmath209 keeping @xmath172 fixed .
this leads to the differential equation @xmath210 where @xmath172 is now considered a variable in @xmath211 since @xmath212 the correct boundary condition for @xmath213 at infinity is @xmath214 and at @xmath215 we impose @xmath216 which makes the function @xmath217 twice differentiable at @xmath194 since @xmath41 and @xmath218 are twice differentiable at zero and @xmath219 a proof of the following proposition can be found in the appendix .
[ theoremeta]for @xmath220 @xmath221 there exists a unique function @xmath222 that satisfies equation ( [ eqeta ] ) with the boundary conditions ( [ etainfinity ] ) and ( [ etazero ] ) .
the function @xmath223 is positive , and has near @xmath215 the expansion @xmath224 with positive coefficients @xmath225 @xmath226 and @xmath227 for @xmath172 large , @xmath223 is of the form @xmath228 for a certain constant @xmath229 with @xmath230 as in ( [ defdelta007 ] ) , @xmath231 and @xmath232 we note that @xmath233 the constants @xmath234 @xmath235 @xmath236 and @xmath237 are given in the appendix .
the function @xmath238 is defined in ( [ etaphi1 ] ) in terms of @xmath239 we make the ansatz @xmath241 which we substitute into equation ( [ eqv ] ) .
we multiply the resulting equation with @xmath242 and take then the limit @xmath209 keeping @xmath171 fixed . since @xmath243 and @xmath244 this leads to the differential equation for the function @xmath245 @xmath246 where @xmath171 is now considered as a variable in @xmath211 at @xmath201
we impose the boundary condition @xmath247 which removes the leading singularity of the function @xmath248 at @xmath249 as we will see , the sub
leading singularity is proportional to @xmath250 which is not a twice differentiable function at @xmath201 ( except for @xmath251 where @xmath252 this singularity will be cancelled by imposing appropriate boundary conditions for the function @xmath253 the correct choice of boundary conditions for @xmath254 at infinity is somewhat less obvious . in the appendix we show that the condition @xmath255 is not sufficient to single out a unique function @xmath256 if @xmath254 does converge to zero at infinity , then it is asymptotic to a solution of the equation @xmath257 this linear equation is compatible with a ( very slow ) algebraic decay , @xmath258 or with a modified gaussian decay , @xmath259 with the algebraic decay being the generic case .
it will be essential in later sections that @xmath260 decays rapidly at infinity , and we therefore impose the boundary condition @xmath261 a proof of the following proposition can be found in the appendix .
[ cmu2]for all @xmath262 there exists a unique positive function @xmath263 that satisfies equation ( [ mu2 ] ) with the boundary conditions ( [ bmu2 - 1 ] ) and ( [ bmu2 - 2 ] ) . for @xmath171 small
, the function @xmath254 is of the form @xmath264 with @xmath265 with @xmath266 and with @xmath267 as in proposition [ theoremeta ] .
for @xmath171 large , the function @xmath254 decays rapidly in the sense that @xmath268 for some constant @xmath269 we make the ansatz @xmath271 which we substitute into equation ( [ eqv ] ) .
we multiply the resulting equation with @xmath272 and take then the limit @xmath273 keeping @xmath274 fixed .
this leads to the ( linear ) differential equation for @xmath275 @xmath276 ~. \label{varphi2}\ ] ] in order to compensate the sub leading singular behavior of @xmath277 near @xmath278 we make the ansatz @xmath279 which we substitute into equation ( [ varphi2 ] ) . for the function
@xmath280 we get the equation @xmath281 since the function @xmath218 is regular near @xmath282 the solution @xmath280 turns out to be regular near @xmath282 too .
therefore , the function @xmath283 is twice differentiable near @xmath56 if we impose at @xmath215 the boundary condition @xmath284 at infinity we need that @xmath285 we therefore require that @xmath286 a proof of the following proposition can be found in the appendix .
[ cfi22]for all @xmath262 there exists a unique function @xmath287 that satisfies equation ( [ h ] ) with the boundary conditions ( [ bh-1 ] ) and ( [ bh-2 ] ) . near @xmath282
the function @xmath280 is of the form @xmath288 with certain coefficients @xmath289 and @xmath290 and for @xmath172 large @xmath280 is of the form @xmath291 with @xmath292 as defined in ( [ lambda0 ] ) , for a certain constant @xmath293 , and with @xmath294 as defined in proposition [ theoremeta ] .
using the properties of the functions @xmath295 @xmath106 @xmath238 and @xmath108 we get from theorem [ main ] the following behavior of the reaction front @xmath296 let @xmath50 be as in theorem [ main ] , and @xmath54 as defined in ( [ reaction001 ] ) .
then , for all @xmath297@xmath298 and for all @xmath129@xmath299 here , @xmath300 @xmath301 are as defined in proposition [ theoremeta ] and @xmath26 is as defined in ( [ asymmu2 ] ) .
in order to simplify the notation we define the function @xmath303 @xmath304 the function @xmath305@xmath306 the function @xmath307@xmath308 and the function @xmath309@xmath310 the function @xmath147 in theorem [ main ] and in ( [ repres1 ] ) can then be written as @xmath311 let now @xmath312 then , @xmath313 therefore , substituting the ansatz @xmath314 into ( [ eqv ] ) leads to the following equation for the function @xmath315@xmath316 with the function @xmath317 @xmath318 the function @xmath154@xmath319 and the map @xmath320@xmath321 the function @xmath323 @xmath324 is strictly decreasing on @xmath325 and therefore @xmath326 for @xmath327 furthermore , the functions @xmath213 and @xmath41 are strictly positive and @xmath41 is strictly increasing .
these properties imply that , for @xmath328 @xmath329 @xmath330 @xmath331 @xmath332 where @xmath333 next , since the functions @xmath334 and @xmath335 are bounded and since @xmath336 for @xmath337 we have that @xmath338 and as a consequence @xmath339 and @xmath340 are positive functions of @xmath86 for all fixed @xmath341 if @xmath342 large enough .
[ propvodd]for @xmath93 odd , @xmath343 there exists @xmath344 such that for all @xmath345 the function @xmath346 is positive .
the function @xmath347 is positive , for @xmath93 odd . as a consequence , for @xmath93 odd , equation ( [ equpsi1 ] )
is of the form indicated in section 2 , with @xmath348 and @xmath349 the rest of this section treats the case of @xmath93 even , which , as indicated in the introduction , is slightly more delicate .
it can be skipped in a first reading or if the reader is only interested in the case of @xmath93 odd .
so let @xmath93 be even .
the idea is to split @xmath346 into its positive part @xmath350 @xmath351 and its negative part @xmath352 , and to show that @xmath353 is small enough so that it can be treated together with the nonlinear term .
consider the function @xmath354 defined in ( [ defphi ] ) .
the problem is that @xmath354 becomes negative for large values of @xmath86 , and that therefore @xmath346 becomes negative for large values of @xmath55 to understand why @xmath354 becomes negative , we note that the leading order term @xmath355 in the large @xmath172 asymptotics of @xmath356 is compensated by the leading order term @xmath357 in the large @xmath171 asymptotics of @xmath358 the leading order of @xmath354 at @xmath86 large is therefore given by the second order term in the large @xmath172 asymptotics of @xmath359 and the leading term in the large @xmath172 asymptotics of @xmath360 the first of these terms is proportional to @xmath361 and the second one is proportional to @xmath362 the corresponding proportionality constants @xmath363 and @xmath364 can be computed for @xmath365 and turn out to be negative .
for @xmath366 these constants can not be obtained from asymptotic expansions , but numerical results show that they are in fact also negative in these cases .
we do not need a proof of this numerical fact , because the following proposition is also correct for positive @xmath346 .
[ propv]for @xmath93 even , @xmath103 there exists @xmath367 , such that the function @xmath368 , satisfies for all @xmath369 the bound @xmath370 the idea is to write @xmath354 as the sum of a function @xmath371 that is positive and a function @xmath372 that absorbs the asymptotic behavior at infinity . since @xmath373 for @xmath171 small , with @xmath374
there exists @xmath375 such that @xmath376 for all @xmath377 let @xmath378 to be chosen below , and let @xmath379 be the heaviside step function , i.e. , @xmath380 for @xmath381 and @xmath382 for @xmath23 then , we define the function @xmath383 by the equation @xmath384 and we set @xmath385 in order to prove that @xmath386 is positive , for @xmath387 large enough , we write @xmath388 where @xmath389 and @xmath390 @xmath391 is positive for @xmath392 since in this case @xmath393 and @xmath394 is positive for @xmath395 by definition of @xmath396 next we consider @xmath397 for @xmath398 we have that @xmath399 . but @xmath400 for all @xmath401 and all @xmath402 if @xmath403 is sufficiently large , since @xmath404 since @xmath127 is bounded , and since @xmath405 for @xmath172 large enough . finally , using the asymptotic properties of @xmath223 and @xmath406 we see that @xmath407 for @xmath408 if @xmath387 is chosen large enough .
we now estimate the function @xmath409 from the definition of @xmath410 we get that @xmath411 and therefore , since @xmath371 is positive , we have the lower bound @xmath412 for some constant @xmath413 , from which ( [ v1bound ] ) follows .
[ thm : inhomo ]
let @xmath166 be as defined in ( [ equi ] ) , and let @xmath414 then , there exists a constant @xmath415 such that for all @xmath416@xmath417 the function @xmath166 is even , and it is therefore sufficient to bound it for @xmath418 the strategy of the proof will be to rewrite the function @xmath166 as a sum of functions of the form @xmath419 with @xmath420 and with @xmath421 and @xmath54 functions with appropriate asymptotic behavior at zero and infinity .
each of the terms in the sum can then be estimated with the help of lemma [ lem : estimations ] below .
in order to keep the notation as simple as possible , we suppress in what follows the arguments of functions whenever there is no risk of confusion .
for @xmath422 the function @xmath423 is of the form @xmath424 where @xmath425 where @xmath426 and where @xmath427 in terms of the functions ( [ t1])([t6 ] ) we get that , for @xmath20 @xmath428 and therefore @xmath429 , where @xmath430 and where @xmath431 @xmath432 and @xmath433 are as defined above .
since @xmath434 it remains to be shown that @xmath435 using the differential equations for @xmath145 @xmath223 and @xmath436 we find that @xmath437 where @xmath438 as defined above , where @xmath439 and where @xmath440 \\ & \hspace{3cm}+n(t^{-2\gamma } t_{1})^{n-1}(t^{-4\gamma } t_{3})~.\end{aligned}\ ] ] the functions @xmath441 and @xmath442 can be further decomposed as follows @xmath443 where @xmath444 and @xmath445 are as defined above , and where @xmath446 it remains to be shown that @xmath447 but this follows using the definitions . in order to characterize the behavior of a function near zero and infinity we introduce the following family of vector spaces .
let @xmath448 and @xmath449 be two real numbers with @xmath450 then , we define @xmath451 to be the vector space of continuous functions @xmath54 from @xmath452 to @xmath190
for which the norm @xmath453 is finite .
note that , if a function is in @xmath454 , then it is also in @xmath455 for any pair of numbers @xmath456 for which@xmath457 furthermore , if @xmath458 is in @xmath459 , and @xmath460 is in @xmath461 , then the product @xmath462 is in @xmath463 .
the following lemma provides the tool that we use to estimate the terms on the right hand side of ( [ repi ] ) .
[ lem : estimations ] [ lem : bornes ] let @xmath464 and @xmath465 , and assume that @xmath466 and that @xmath467 then , there is a constant @xmath468 such that for all @xmath416 @xmath469 where @xmath470 from ( [ cb3 ] ) it follows that either @xmath471 or @xmath472 in the first case we get using ( [ cb1 ] ) that @xmath473 @xmath474 @xmath475 and therefore @xmath476 and in the second case we get using ( [ cb2 ] ) that @xmath477 and therefore @xmath478 we now show that the right hand side in ( [ repi ] ) can be bounded by a sum of terms of the form @xmath479 for each of these terms we then show that the corresponding functions @xmath421 and @xmath54 satisfy the hypothesis of lemma [ lem : estimations ] .
this then implies that @xmath480 @xmath481 for a certain @xmath482 depending on @xmath54 and @xmath483 it is therefore sufficient to prove that @xmath484 for all these terms in order to prove the inequality ( [ eqn : thminhomo ] ) . for @xmath485 @xmath486
we have the bounds @xmath487 and @xmath488 , @xmath489 , @xmath490 , @xmath491 , @xmath492 , @xmath493 , @xmath494 .
the inequalities ( [ inequalities ] ) follow by using the triangle inequality and the asymptotic properties of the functions @xmath41 , @xmath495 , @xmath218 and @xmath238 .
[ [ bound - on - the - function - a_2 . ] ] bound on the function @xmath496 .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + we have the bound @xmath497 where @xmath498 the function @xmath499 is in @xmath500 , and the function @xmath501 is in @xmath502 @xmath503 . since @xmath504 the inequalities ( [ cb1 ] ) and ( [ cb2 ] ) are satisfied and , since @xmath505 for @xmath506@xmath507 therefore , @xmath508
@xmath509 @xmath510as required .
[ [ bound - on - the - function - a_3 . ] ] bound on the function @xmath432 .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + we have that @xmath511 where @xmath512 @xmath513 .
since @xmath514 , and @xmath515 we have the bound @xmath516 with @xmath517 the function @xmath518 is in @xmath519 and the function @xmath520 is in @xmath521 . since @xmath522
the inequalities ( [ cb1 ] ) and ( [ cb2 ] ) are satisfied and @xmath523 . therefore @xmath524
@xmath525 @xmath526 as required .
similarly , we have that @xmath527 with @xmath528 the function @xmath529 is in @xmath530 and the function @xmath531 is in @xmath532 .
the inequalities ( [ cb1 ] ) and ( [ cb2 ] ) are satisfied and @xmath533 @xmath534 therefore , @xmath535
@xmath536 @xmath526 as required .
finally , @xmath537 where @xmath538 the function @xmath539 is in @xmath540 , and the function @xmath541 is in @xmath542 . the inequalities ( [ cb1 ] ) and ( [ cb2 ] ) are satisfied and @xmath543 .
therefore , @xmath535
@xmath544 @xmath526 as required .
[ [ bound - on - the - function - a_4 . ] ] bound on the function @xmath433 .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + since the functions @xmath545 and @xmath546 are bounded , @xmath547 and @xmath548 we have that @xmath549 where @xmath550 the function @xmath539 is in @xmath551 , and the function @xmath552 is in @xmath553 .
the inequalities ( [ cb1 ] ) and ( [ cb2 ] ) are satisfied and @xmath554 .
therefore , @xmath524
@xmath555 @xmath526 as required .
[ [ bound - on - the - function - a_5 . ] ] bound on the function @xmath438 .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + we have the bound @xmath556 where @xmath550 the function @xmath539 is in @xmath551 , and the function @xmath557 is in @xmath558 .
the inequalities ( [ cb1 ] ) and ( [ cb2 ] ) are satisfied and @xmath543 .
therefore , @xmath524
@xmath559 @xmath560 as required .
[ [ bound - on - the - function - a_6 . ] ] bound on the function @xmath561 .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + we have the bound @xmath562 where @xmath563 the function @xmath564 is in @xmath565 and the function @xmath566 is in @xmath567 the inequalities ( [ cb1 ] ) and ( [ cb2 ] ) are satisfied and @xmath568 .
therefore , @xmath524
@xmath569 @xmath526 as required .
[ [ bound - on - the - function - a_7 . ] ] bound on the function @xmath570 .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + we have the bound @xmath571 where @xmath572 .
the function @xmath573 is in @xmath574 and the function @xmath575 is in @xmath576 the inequalities ( [ cb1 ] ) and ( [ cb2 ] ) are satisfied and @xmath577 .
therefore , @xmath524
@xmath578 @xmath526 as required .
[ [ bound - on - the - function - a_8 . ] ] bound on the function @xmath579 .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + we have the bound @xmath580 where @xmath581 the function @xmath582 is in @xmath583 and the function @xmath584is in @xmath585 the inequalities ( [ cb1 ] ) and ( [ cb2 ] ) are satisfied and @xmath586 . therefore ,
@xmath587 @xmath588 @xmath589 as required .
[ [ bound - on - the - functions - a_pq . ] ] bound on the functions @xmath590 .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + we have the bound @xmath591 where @xmath592 the function @xmath593 is in @xmath594 and the function @xmath595 is in @xmath596 .
the inequalities ( [ cb1 ] ) and ( [ cb2 ] ) are satisfied , and @xmath597 therefore , @xmath598 , for @xmath599 and @xmath600 , for @xmath601 as required .
this completes the proof of theorem [ thm : inhomo ] .
@xmath602 equation ( [ equpsi1 ] ) is of the form ( [ eqpsi ] ) if we define the map @xmath603 by the equation @xmath604 with @xmath605 as defined in ( [ that ] ) and @xmath368 as defined in section [ sectionv ] . using the definitions
, we see that @xmath155 can be written as , @xmath606 with @xmath607 let @xmath608 as in ( [ vpq ] ) .
then , for all @xmath416 @xmath609 where @xmath610 the case @xmath611 follows from ( [ v1bound ] ) .
let now @xmath612 since @xmath613 for all @xmath614 @xmath615 we find that @xmath616 furthermore , since @xmath617 near @xmath618@xmath619 since the function @xmath620 is bounded , it follows that @xmath621 @xmath622 with @xmath623 for @xmath624 we improve this bound using additional properties of the function @xmath625 namely , since @xmath626 we have that @xmath627
for functions @xmath628 in @xmath629 we use the norms @xmath630 , @xmath631 and @xmath632 , and we denote by @xmath633 the banach space of functions @xmath634 in @xmath635 for which the norm @xmath636 , @xmath637 is finite .
let @xmath638 as in proposition [ propvodd ] and @xmath639 as in proposition [ propv ] , and consider , for fixed @xmath640 functions @xmath641 of the form @xmath642 with @xmath643 let @xmath644 be the fundamental solution of the differential operator @xmath645 and let , for given @xmath646 , the map @xmath647 be defined by the equation @xmath648 where @xmath649 and where @xmath650 with @xmath651 the integral equation @xmath652 is equivalent to the differential equation ( [ eqpsi ] ) with initial condition @xmath653 we note that , since the function @xmath153 is positive , the kernel @xmath644 is bounded pointwise by the fundamental solution @xmath654 of the heat equation , @xmath655 the following proposition makes theorem [ main ] precise .
[ maint]let @xmath656 with @xmath657 as defined in ( [ eqn : thminhomo ] ) , and let @xmath403 be sufficiently large . then , for all @xmath646 with @xmath658 , the equation @xmath652 has a unique solution @xmath659 in the ball @xmath660 the solution of the integral equation will be dominated by @xmath661 and , as we will see , @xmath662 has been chosen such that @xmath663 the idea is therefore to show that , if @xmath403 is large enough to make the nonlinear part of the map @xmath647 small , and if @xmath664 , then the map @xmath647 contracts the ball @xmath665 into itself , which by the contraction mapping principle implies the theorem . we first show that @xmath647 maps the ball @xmath666 into itself . for the contribution coming from the initial condition we have @xmath667 and
therefore @xmath668 we next estimate the norm @xmath669 let @xmath670 then , @xmath671 and therefore @xmath672 it remains to be shown that the nonlinearity is also bounded by @xmath673 for @xmath403 large enough .
for @xmath674 we have , @xmath675 for @xmath611 we get , since @xmath676 @xmath677 and for @xmath624 we get @xmath678 and for the other cases we have @xmath679 and therefore @xmath680 if @xmath403 is large enough . using the triangle inequality
we get that @xmath681 which proves that @xmath682 as claimed .
we now show that @xmath647 is lipschitz .
let @xmath126 and @xmath127 be in @xmath683 we have @xmath684 and therefore we get , using the same estimates as for ( [ normcase ] ) , that @xmath685 provided @xmath403 is large enough .
this completes the proof of theorem [ maint ] .
we first prove the existence of a unique positive solution of equation ( [ eqeta ] ) satisfying the boundary conditions ( [ etainfinity ] ) and ( [ etazero ] ) .
then , we derive the results on the asymptotic behavior near zero and infinity .
let , for @xmath687 , @xmath688 be the solution of the initial value problem on @xmath189 , @xmath689 then , there exists a unique @xmath690 such that the function @xmath691 is positive and satisfies @xmath692 .
we first prove that @xmath693 is unique . given a function @xmath218 from @xmath189 to @xmath694 we define the function @xmath695 @xmath696 .
assume that there are two values @xmath697 such that the functions @xmath698 and @xmath699 are positive and satisfy @xmath700 @xmath701 we first show that the function @xmath702 is positive for all @xmath703 namely , if we assume the contrary , then because @xmath704 , there must be a first @xmath705 such that @xmath706 furthermore , if @xmath707 then @xmath708 and therefore @xmath709 a contradiction .
therefore @xmath710 and as a consequence @xmath711 are positive for all @xmath121 from which it follows that @xmath712 in contradiction with @xmath713 to prove the existence of a @xmath690 for which @xmath714 is positive and for which @xmath715 , we use the so called shooting method . note that , for any @xmath687 , the initial value problem ( [ 1 ] ) has a unique solution @xmath716 and since @xmath717 the function @xmath688 is strictly decreasing on @xmath718 for @xmath719 small enough .
we will show that for small enough @xmath687 , the graph of @xmath688 intersects the real axis and @xmath688 becomes negative , whereas for @xmath720 large enough , @xmath688 has a minimum and then diverges to plus infinity .
the ( unique ) point between those two sets is @xmath690 .
define the two subsets @xmath721 and @xmath722 of @xmath325 @xmath723 we note that if @xmath724 and @xmath725 for some @xmath726 , then @xmath727 on any interval @xmath728 on which @xmath688 is defined , and a function @xmath729 with @xmath730 can therefore not converge to zero at infinity .
furthermore , since the function @xmath731 is a solution of the differential equation ( [ 1 ] ) , it follows , since solutions are unique , that @xmath732 if @xmath733 and therefore the intersection of @xmath721 with @xmath722 is empty .
the sets @xmath721 and @xmath722 are open , by continuity of the solution @xmath688 as a function of the initial data @xmath734 .
we now show that @xmath721 is non empty and bounded , which shows that @xmath735 this @xmath736 is neither in @xmath721 nor in @xmath722 , and therefore the function @xmath737 is at the same time strictly positive and strictly decreasing , and therefore @xmath715 . to prove that @xmath721 is non empty
, we fix any @xmath738 positive and choose @xmath739 small enough such that on @xmath740 $ ] the solution @xmath741 exists and is strictly decreasing . then , @xmath742 choose now @xmath743 and let @xmath744 be the corresponding solution .
as before , we have that the function @xmath745 and its second derivative @xmath746 are positive on the interval @xmath747 and therefore , since @xmath748 @xmath749 @xmath750 we find that @xmath751 using the definition of @xmath752 we therefore find that @xmath753 therefore @xmath754 .
we now prove that @xmath721 is bounded . for @xmath755 ,
let @xmath719 be the largest value ( possibly infinite ) such that on @xmath718 the solution @xmath688 exists and is strictly positive .
then , @xmath756 is positive on @xmath757 and , therefore @xmath758 for @xmath759 . as a consequence ,
if the function @xmath688 exists on @xmath760,$ ] then @xmath761 using again that @xmath762 we then find that @xmath763 for @xmath764 $ ] , and therefore @xmath765 on @xmath766 $ ] , which implies that @xmath767 which is positive if @xmath768 therefore @xmath769 must be equal to zero for some @xmath770 any such @xmath734 therefore belongs to @xmath771 if the function @xmath688 ceases to exist before @xmath772 it must have been diverging to plus infinity for some @xmath773 which again implies that @xmath774 must have been equal to zero for some @xmath775 and the corresponding @xmath734 is in @xmath771 the function @xmath223 is regular at zero , and the coefficients of its taylor series at zero can be computed recursively .
we have @xmath777 and @xmath778 and therefore we get using the differential equation that @xmath779 @xmath780 and @xmath781 the asymptotic behavior of @xmath223 at infinity is obtained as follows . assuming that @xmath223 behaves like @xmath782 at infinity we get from the differential equation that @xmath230 and @xmath267 are as defined in ( [ defdelta007 ] ) and ( [ deflambda007 ] ) , respectively . that this is indeed the correct leading behavior of @xmath223 at infinity
can now be proved by using standard techniques based on repeated applications of lhpital s rule .
see for example @xcite .
since the proof is simple , but lengthy and quite uninteresting , we do not give the details here .
once the leading behavior of @xmath223 at infinity has been established we make the ansatz @xmath783 . to
leading order we get for the function @xmath784 the linear equation @xmath785 there is a certain constant @xmath786 such that the function @xmath787 @xmath788 is a particular solution of equation ( [ eqs ] ) .
the solutions of the homogeneous equation associated with ( [ eqs ] ) are of the form @xmath789 , and using the definition ( [ deflambda007 ] ) for @xmath267 we find that @xmath790 for @xmath791 we have that @xmath792 and the asymptotic behavior of @xmath784 is therefore for @xmath791 of the form @xmath793 with @xmath794 , and of the form @xmath795 with some unknown coefficient @xmath363 for @xmath796 .
it is tedious , but not difficult , to prove that this is indeed the correct second order behavior of @xmath223 at infinity .
we omit the details . in order to study equation ( [ mu2 ] ) with boundary conditions ( [ bmu2 - 1 ] ) and ( [ bmu2 - 2 ] )
, we make the ansatz @xmath797 for the function @xmath798 we get the differential equation @xmath799 and the boundary conditions for @xmath798 are @xmath800 as indicated in section [ smu2 ] , a solution of equation ( [ mu2 ] ) that is defined on @xmath189 behaves at infinity either like @xmath802 or like @xmath803 the proof is similar to the one in @xcite .
we omit the details .
given the asymptotic behavior of @xmath254 at infinity , we find for the function @xmath798 at infinity either a behavior proportional to @xmath804 , or a behavior proportional to @xmath805 since @xmath806 we find that @xmath807 if and only if the boundary condition ( [ bcm1 ] ) is satisfied , and we will impose ( [ bcm3 ] ) from now on .
we now discuss the asymptotic behavior of the function @xmath798 near zero . from equation ( [ eqh ] ) we see that @xmath808 exists if and only if @xmath809 , i.e. , if @xmath810 and if @xmath811 we then find , that @xmath812 , with @xmath292 as defined in ( [ lambda0 ] ) . by taking derivatives of equation ( [ eqh ] )
we find that @xmath813 , and that @xmath814 for some constant @xmath815 by taking further derivatives , one can recursively compute the taylor coefficients of a solution @xmath816 of equation ( [ eqh ] ) that is regular ( in fact , analytic ) in a neighborhood of zero . the solution @xmath816 does however not satisfy the boundary condition ( [ bcm3 ] ) .
the solution of ( [ eqh ] ) that does satisfy ( [ bcm3 ] ) is of the form @xmath817 where @xmath818 with @xmath819 as defined in ( [ eqn : critexp ] ) . here , @xmath820 , with @xmath821 and with @xmath822 to be determined .
the asymptotic form ( [ msing ] ) can be obtained by substituting the ansatz ( [ msing ] ) for @xmath798 into equation ( [ eqh ] ) . since @xmath823
we find from ( [ msing ] ) that near zero @xmath824 .
we omit the details of the proof that the asymptotic behavior is as indicated .
we now prove the existence of a function @xmath798 that satisfies equation ( [ eqh ] ) with the boundary conditions ( [ bcm0 ] ) and ( [ bcm3 ] ) . since the second derivative of the solution @xmath798 at zero is positive , and since @xmath798 converges to zero at infinity , there must be a first @xmath825 such that @xmath826 the basic idea is now to use this position @xmath827 and the value @xmath828 of @xmath798 at @xmath482 , as parameters in shooting arguments towards zero and infinity . the first shooting argument will allow us to define a curve @xmath829 of initial conditions @xmath830 in @xmath831 for which the boundary condition at zero is satisfied , and the second shooting argument will allow us to find on this curve an initial condition for which the boundary condition at infinity is satisfied as well .
so , let @xmath832 be an initial condition .
locally , i.e. , near @xmath833 there exists a solution @xmath834 of equation ( [ eqh ] ) . by definition , @xmath835 @xmath836 and
therefore we get for the second derivative of @xmath834 at @xmath833 @xmath837 where @xmath838 and @xmath839 for initial conditions such that @xmath840 where @xmath841 and @xmath842 we therefore have that @xmath843 see fig . 1 for the graph of the function @xmath844 the function @xmath845 has a maximum at the point @xmath846 that satisfies the equation @xmath847 and the line @xmath845 divides the set of initial conditions into two subsets , a subset @xmath0 where @xmath848 and a subset @xmath849 where @xmath850 for initial conditions on @xmath845 we can compute @xmath851@xmath852 comparing with ( [ defxim ] ) we find that @xmath853 and we have that @xmath854 for @xmath855 we now construct the line @xmath829 for @xmath855 the proof is similar to the one in section [ seta777 ] . define the two subsets @xmath721 and @xmath722 of the interval @xmath863 , @xmath864 the intersection of @xmath721 with @xmath722 is by definition empty , and the sets @xmath721 and @xmath722 are open , by continuity of the solution @xmath865 as a function of the initial data @xmath734 .
we now show that all @xmath720 sufficiently close to @xmath267 are in @xmath866 and that all @xmath734 sufficiently close to @xmath867 are in @xmath868 this implies that @xmath869 and @xmath870 is neither in @xmath721 nor in @xmath722 , and therefore the function @xmath871 satisfies @xmath872 for all @xmath873 and therefore @xmath874 since @xmath875 so let @xmath832 be an initial condition . then , @xmath834 satisfies the integral equation @xmath876 where @xmath877 and where @xmath878 @xmath879 is strictly negative for @xmath880 and @xmath881 and therefore we find , like in the proof in section [ theoremeta ] that any solution with an initial condition @xmath734 sufficiently close to @xmath882 will cross the line @xmath883 similarly , for an initial condition @xmath832 close to @xmath884 we can use that @xmath885 @xmath886 and that @xmath887 is strictly negative to show that the corresponding solution will cross the line @xmath844 this completes the proof of the existence of @xmath888 to prove uniqueness it is sufficient to use that @xmath889 for @xmath890 in the set @xmath26 ( see fig .
1 ) , and to integrate the difference of two solutions from their respective initial condition to zero , which leads to a contradiction , since both solutions have to be equal to @xmath267 at zero . finally ,
that @xmath829 is a continuous function follows from the continuity of @xmath834 as a function of @xmath734 and @xmath891 using the uniqueness of @xmath888 we now prove with a second shooting argument that solutions with initial conditions @xmath892 with @xmath893 become negative somewhere in the interval @xmath894 and that solutions with initial conditions @xmath895 , with @xmath896 stay positive and diverge to plus infinity .
define the two subsets @xmath721 and @xmath722 of the interval @xmath900 , @xmath901 by definition , the intersection of @xmath721 with @xmath722 is empty , and the sets @xmath721 and @xmath722 are open , by continuity of the solution @xmath902 as a function of the initial data @xmath891 .
we now show that all @xmath891 sufficiently close to @xmath903 are in @xmath866 and that all @xmath891 sufficiently close to @xmath846 are in @xmath868 this implies that @xmath904 is neither in @xmath721 nor in @xmath722 , and therefore the function @xmath905 is positive and decreasing for @xmath906 which implies that @xmath907 so let @xmath908 be an initial condition with @xmath909 with @xmath910 the proof that such an initial condition is in @xmath721 is rather lengthy and we therefore do not give the details here , but on a heuristic level it is easy to understand why such a solution is in @xmath721 .
namely , near zero the asymptotics of the solution @xmath871 is @xmath911 where @xmath818 and where @xmath912 is such that @xmath913 . neglecting higher order terms
we find that @xmath914 , and we conclude that @xmath915 therefore , @xmath916 and @xmath917 if @xmath891 is small enough . therefore , since @xmath918 for all @xmath919 we find that @xmath920 and therefore @xmath921 for some @xmath922 as claimed .
next , let @xmath923 be an initial condition with @xmath924 with @xmath925 to be chosen below .
we now show that such an initial condition is in @xmath771 for all @xmath926 we have the lower bounds @xmath927 and @xmath928 @xmath929 where @xmath930 @xmath931 and @xmath932 and for @xmath933 we therefore have the lower bound @xmath934 and it follows , using again the integral equation ( [ inteq ] ) , that @xmath935 diverges at ( or before ) infinity , that therefore @xmath936 for some @xmath937 which implies that @xmath938 provided @xmath939 for @xmath940 small enough and for @xmath93 large enough ( [ cond ] ) can be verified without too much difficulty . with the help of a computer one can show that ( [ cond ] ) is satisfied for the remaining @xmath941 for @xmath251 ( [ cond ] ) is not satisfied , since the above bounds on @xmath942 and @xmath943 are too weak . sufficiently good bounds can be obtained by dividing the interval @xmath944 in two pieces and by integrating lower bounds on each of the subintervals .
we omit the details .
finally , uniqueness of @xmath945 can be proved by integrating the difference of two solutions from @xmath946 to infinity , which , using the positivity of @xmath947 leads to a contradiction with the fact that both of the solutions converge to zero at infinity .
the equation ( [ h ] ) for @xmath280 is linear .
we therefore first construct two linearly independent solutions @xmath948 and @xmath949 for the corresponding homogeneous equation , which we then use to construct , using standard methods , a solution of ( [ h ] ) that satisfies the boundary conditions ( [ bh-1 ] ) and ( [ bh-2 ] ) .
the homogeneous equation is @xmath950 where @xmath951 since the equation ( [ homh ] ) is linear , the integral equation for @xmath948 , @xmath952 has a positive solution that exists for all @xmath86 in @xmath211 by definition , we have near @xmath56 the behavior @xmath953 at infinity , the solution @xmath948 is asymptotic to a solution of the equation @xmath954 this equation is the same as the homogeneous part of equation ( [ eqs ] ) , and the leading order behavior of @xmath948 at infinity is therefore either proportional to @xmath955 or to @xmath956 with @xmath957 as defined in ( [ eqn : critexp ] ) . since @xmath948 is positive , we find using ( [ eqh1 ] ) , that @xmath958 for all @xmath86 in @xmath325 and therefore @xmath948 is near infinity of the form @xmath959 for some constant @xmath960 .
a second solution of the homogeneous equation ( [ homh ] ) is @xmath961 near @xmath56 we have that @xmath962 and near infinity we find that @xmath963 where @xmath964 , and @xmath965 @xmath966 we note that @xmath967 therefore , the function @xmath968 , @xmath969 where @xmath970 and where @xmath971 satisfies equation ( [ h ] ) . near zero , the function @xmath628 is of the form @xmath972 , and therefore , using the behavior of @xmath948 and @xmath949 near zero , we find that @xmath973 is near zero of order @xmath974 , and @xmath845 is near zero of order @xmath975 the function @xmath968 is therefore of order @xmath976 near zero . at infinity ,
the function @xmath628 is of the form @xmath977 where @xmath978 and therefore the function @xmath845 is near infinity of the form @xmath979 , and @xmath973 is near infinity of the form @xmath980 , for some constant @xmath981 using these asymptotic behavior for @xmath973 , @xmath948 , @xmath845 , and @xmath290 we find for the function @xmath982 near infinity the behavior , @xmath983 in the last equality we have used the definition ( [ lambda0 ] ) for @xmath984 the function @xmath985@xmath986 solves the equation ( [ h ] ) , satisfies the boundary condition ( [ bh-1 ] ) , and since , as we show in the next section , the higher order terms in ( [ hpinfty ] ) converge to zero at infinity , it also satisfies the boundary condition ( [ bh-2 ] ) . by construction ,
the leading behavior of @xmath280 at infinity is @xmath987 .
we therefore make the ansatz @xmath988 , and to leading order , we get for the function @xmath191 the linear equation @xmath989 for a certain constant @xmath990 .
the general solution of equation ( [ bvarphi2 - 2 ] ) is @xmath991 with a certain constant @xmath992 and with @xmath819 , @xmath993 as defined in ( [ eqn : critexp ] ) .
since @xmath994 but @xmath995 the coefficient of the term proportional to @xmath996 must be zero .
therefore , since@xmath997 for all @xmath998 the asymptotic behavior of @xmath999 is always given by @xmath1000 we omit the details of the proof that this is indeed the correct second order behavior of @xmath191 at infinity .
j. bricmont , a. kupiainen & g. lin : renormalisation group and asymptotics of solutions of nonlinear parabolic equations .
_ communications in pure and applied mathematics _ , new york , ny issn 0010 - 3640 , * 47 * , 893 , ( 1994 ) . | we analyze the long time behavior of initial value problems that model a process where particles of type @xmath0 and @xmath1 diffuse in some substratum and react according to @xmath2 the case @xmath3 has been studied before ; it presents nontrivial behavior on the reactive scale only . in this paper
we discuss in detail the cases @xmath4 and prove that they show nontrivial behavior on the reactive and the diffusive length scale . |
the @xmath7 redshift is a critical epoch at which both the global star formation rate ( sfr ) and the active galactic nucleus ( agn ) activity appear to have peaked ( e.g. * ? ? ?
* ; * ? ? ?
* ) , and the morphological differentiation of galaxies was fully underway ( e.g. * ? ? ?
* ; * ? ? ?
* ) . for a period of a few billion years , a large fraction ( at least 50% ) of massive galaxies with @xmath8 formed stars at sufficiently high sfrs ( @xmath9 ) to produce the bulk of their stellar populations in this span of time : i.e.
, this is the formation epoch of these galaxies .
therefore , learning as much as possible about the ( physical ) properties of these @xmath4 galaxies will help us to determine how galaxies accumulate their mass and stellar populations .
the specific sfr ( ssfr ) of most star - forming galaxies has been found to be a tight but weak function of stellar mass at all epochs up to at least @xmath10 , forming the so - called _ main sequence ( ms ) _ of galaxies ( e.g. , * ? ? ?
* ; * ? ? ?
the characteristic ssfr of the main sequence increases strongly with redshift , by a factor of 20 back to @xmath11 @xcite .
this implies that galaxies with sfrs that are an order of magnitude or more higher than those of local galaxies of the same stellar mass ( see eq . 1 in * ? ? ?
* ) are nevertheless the `` normal '' ms population of star - forming galaxies at @xmath4 .
the tightness of the ssfr on the ms suggests that the sfr is evolving in a quasi - equilibrium state ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . the _ chemical evolution _ of such galaxies , as traced by their gaseous o / h ratio , is a key diagnostic . in the local universe , there is a tight mass - metallicity relation ( mzr , * ? ? ?
* ) with , in sdss , a 1@xmath12 dispersion of o / h ranging from 0.20 dex at low mass , log(m/)@xmath13 , to 0.07 dex at high mass , log(m/)@xmath14 @xcite . considering their median errors in determining masses and metallicities of 0.09 and 0.03 dex , respectively , @xcite estimated that roughly half of the spread in the mass - metallicity relation could be attributed to observational uncertainties .
the sdss mzr has a relatively steep slope below @xmath15 , for a @xcite imf , but flattens at higher masses .
using chemical evolution models , @xcite interpreted the flattening in terms of efficient galactic winds that remove metals from galaxies with shallow potential wells ( lower - mass galaxies ) , an idea first introduced by @xcite .
other explanations are also possible .
flattening of the slope of the mzr at lower redshift has been interpreted as the chemical version of galaxy downsizing ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
on the other hand , @xcite claimed that the flattening of the mzr at later times could be due to a redshift evolution in the stellar mass at which the metallicities of galaxies begin to saturate .
it is also well - established that the mzr at @xmath16 evolves relative to that seen locally ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
. however , the observed shape and evolution of the mzr at high redshifts is likely to be affected by the selection of the samples and metallicity estimators used . in the local universe ,
the o / h at a given mass also depends on the sfr ( and thus also ssfr ) of the galaxy ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , i.e. , the sfr appears to be a `` second - parameter '' in the mzr .
the evidence for an inverse correlation between sfr and o / h is good at low masses , and the scatter in o / h can be reduced to about 0.07 dex across the ( m , sfr ) plane ( * ? ? ?
* ma10 in the following ) .
however , the form of the z(m , sfr ) relation at higher masses is less well - established , and in fact @xcite have claimed that the sfr - dependence reverses .
ma10 represented the local sdss z(m , sfr ) by a second - order polynomial in m and sfr ( their eq.2 describing a surface in the 3d space ) , and , by introducing the new quantity @xmath17 , they also defined a projection of the z(m , sfr ) that minimizes the metallicity scatter of local galaxies .
furthermore , ma10 claimed that the @xmath3 relation seen in measurements of low redshift sdss galaxies ( with @xmath18 ) is also applicable to much higher redshift galaxies , in which the typical sfrs are higher because of the evolution in the ms ssfr .
they coined the phrase fundamental metallicity relation ( fmr ) to denote this epoch - invariant @xmath3 relation , and claimed that , `` in practice , metallicity of star - forming galaxies of any mass , any sfr and at any redshift up to z=2.5 follow the following relation : @xmath19 with @xmath20 . ''
this eq.5 of ma10 describes a plane in 3d space , and is a simplification of their eq.4 . in this paper
we clearly distinguish between ( a ) the existence of a @xmath3 relation at a given epoch ( i.e. , whether the sfr is a second parameter in the mzr ) and ( b ) whether the @xmath3 relation is also epoch - independent .
we reserve the term fmr for an epoch-_independent _ @xmath3 relation . as we discuss further below
, it is clear that assessing the epoch-(in)dependence of the @xmath3 relation depends on the choice of extrapolation of any locally - defined @xmath3 relation into regions of the ( m , sfr ) parameter space that are poorly , if at all , populated in the local universe .
ma10 suspected that their fmr held up to @xmath21 , but evolved at @xmath22 , using the ( * ? ? ?
* eb06 in the following ) sample at @xmath4 , and the galaxy samples at @xmath22 of @xcite and @xcite .
initial explanations of why sfr is a second parameter in the mzr have generally involved ad hoc inflow of gas diluting the metallicity and increasing star - formation @xcite . in a recent paper , however , ( * ? ? ?
* li13 in the following ) showed that a @xmath3 relation is a natural outcome of a simple model of galaxies in which the sfr is regulated by the mass of gas present in a galaxy , and calculated the relation expected when a galaxy evolves in a quasi - equilibrium state .
they argued that this gas regulation would work at least since @xmath10 and possibly also at higher redshifts .
their derived z(m , sfr ) relation has a particular analytic form and depends on the internal parameters describing the regulator system , specifically the star - formation efficiency @xmath23 ( which is therefore the inverse of the gas depletion timescale @xmath24 ) , and the mass - loading of any wind that drives gas out of the system , their @xmath25 .
furthermore , the @xmath3 relation would only evolve to the extent that these two parameters change with epoch ( at fixed galactic mass ) , providing an attractive explanation for the claimed existence of an fmr .
li13 could represent the local sdss data from ma10 with their predicted @xmath3 relation with astrophysically plausible values of @xmath26 and @xmath27 and reasonable mass - dependences of these two parameters .
it should be noted that the analytic approximations employed by li13 have since been verified by more detailed calculations in @xcite .
the form and possible epoch-(in)dependence of the @xmath3 relation thus emerges , along with other measures of the state of high redshift galaxies , as a very powerful diagnostic of the regulation of star - formation in typical ms galaxies .
the definition of this relation at high redshifts is therefore a priority .
there are a number of ways to estimate the chemical abundances in galaxies , especially through the analysis of emission line ratios of gas in star - forming hii regions . over the years
, a number of diagnostics have been developed based on strong lines , including the well - known @xmath28 parameter first introduced by @xcite .
this is based on the ratio of oxygen lines @xmath29\,}{\lambda\,3727}$ ] and @xmath30\,}{\lambda\,5007}$ ] to h@xmath31 .
this ratio must be corrected for reddening using the h@xmath32 to h@xmath31 ratio .
a degeneracy between high and low o / h values that exists for most values of @xmath33 can be broken using , e.g. , the [ nii]6584/h@xmath32 ratio , or other means as summarized in , e.g. , @xcite .
finally , the o / h calibration also varies with ionization parameter , @xmath34 , which may well vary with redshift but which can be estimated from @xmath35}/{\rm [ o\,ii]\,}$ ] . in our earlier work @xcite , we extended the basic @xmath33 method by simultaneously fitting all of these five lines ( , , , , and ) for metallicity o / h , reddening @xmath36 and ionization parameter @xmath34 .
unfortunately , at high redshifts around @xmath10 , these five lines are seen in the atmospheric windows for a limited range of redshifts only , namely at @xmath37 . at higher redshifts , the h@xmath32 and [ nii ] lines which are critical for the practical implementation of the @xmath33 method
are shifted redward of the k - band window . even within this narrow redshift window ,
the chance of having all five lines clear of strong telluric oh emission is not very high .
in addition , near - infrared ( nir ) spectroscopy at these faint levels is difficult , even on 8-m class telescopes . as a result
, none of pioneering studies of metallicities of star - forming galaxies at @xmath4 @xcite has been based on measurements of all five lines , and only a handful , i.e. , four galaxies of eb06 , five of @xcite , two galaxies from lsd @xcite , and a few amaze objects @xcite have measurements of even three or four of the lines .
the derived o / h metallicities have therefore been based on a series of ad hoc calibrations of more limited line ratios , and often just [ nii]/h@xmath32 , as in the large sample of eb06 .
these suffer from significant scatter and systematic uncertainties arising from the poor knowledge of ionization parameter ( in case of the n2 calibration ) and/or reddening . even now
, there are only six @xmath4 galaxies with the five emission line fluxes measured in the literature , four of which are gravitationally lensed and all having rather low mass @xcite . there are therefore several caveats regarding the extensive eb06 study of metallicities of 87 galaxies at @xmath5 .
the ionization parameter could not be measured for any of these galaxies , and only coarse oxygen abundances using the [ nii]/h@xmath32 ratio were determined by applying the so - called n2 calibration @xcite , which does not take the ionization parameter into account . in fig.[n2_ionpar ] we show the [ nii]/h@xmath32 ratio vs. metallicity as a function of the ionization parameter @xmath34 ( dotted lines ) , based on fig .
7 in ( * ? ? ?
* kd02 in the following ) .
the reason why the n2 calibration works at low redshift is that ionization is correlated with metallicity .
however , if this is also true at higher redshift is a matter of debate , and it is not clear which range of ionization parameters one should adopt at each metallicity at higher redshift . to derive the mzr at @xmath4 , eb06 used the averaged [ nii]/h@xmath32 ratios in six stellar mass bins each containing @xmath38 galaxies , and estimated the o / h metallicities of these six averaged spectra using the n2 calibration .
the horizontal lines in fig.[n2_ionpar ] show the measurements of [ nii]/h@xmath32 and the inferred range in o / h ( the ionization parameter , i.e. , the location on a corresponding dotted line , is not known ) , while the filled blue circles show the o / h measurements derived by eb06 .
the magenta oblique line has been transformed from the n2 calibration to the kd02 calibration ( used in this paper ) by applying the conversion given in tab.3 of @xcite .
it is obvious from fig.[n2_ionpar ] that , without knowledge of the ionization parameter ( dotted lines ) , the metallicity estimates have very large uncertainties , especially at high metallicities : the six horizontal blue lines of possible o / h at a given [ nii]/h@xmath32 extend over @xmath39dex .
it should be noted that it was stated by @xcite that `` with the n2 calibrator it is possible to estimate the abundance of oxygen to within a factor of 2.5 at the 95 per cent confidence level . ''
factor 2.5 corresponds to about 0.4dex , which is comparable to the scatter in o / h at a given log([nii]/h@xmath32 ) in fig.[n2_ionpar ] ( indicated by the blue lines ) .
furthermore , only 4 of the 87 galaxies in the eb06 sample have [ oiii]@xmath40 and h@xmath31 line fluxes measured in addition to h@xmath32 and [ nii ] , i.e. , no information on type-2 agn contamination is available for all but 5% of the galaxies in the eb06 sample .
it is therefore unclear whether the ionization in these objects is dominated by an agn , which would lead to highly uncertain metallicity measurements .
unfortunately , these uncertainties in both the ionization parameter and agn contribution are now strong impediments to a clear understanding of these galaxies and of the chemical evolution at @xmath4 . to overcome these uncertainties , we have carried out nir follow - up spectroscopy with subaru - moircs and vlt - sinfoni of 20 star - forming galaxies at @xmath5 , selected from the zcosmos - deep sample . as well as providing a large enough parent sample , using zcosmos - deep galaxies enables us to also use the wealth of information on these galaxies that is available from the cosmos survey .
the aim of this paper is to use oxygen gas abundances , stellar masses and sfrs for a sample of zcosmos galaxies at @xmath5 to study the mzr and fmr at @xmath4 .
the paper is structured as follows : in sect .
2 we describe the selection of the targets and the subaru - moircs and vlt - sinfoni observations . in sect .
3 we present the measurement of emission line fluxes , and the derivation of sfrs and metallicities using a representation of the kd02 models .
we also discuss the possible contamination by agns using the bpt @xcite diagram . in sect .
4 we present the ssfr - mass , mzr and z(m , sfr ) relations at @xmath41 , and discuss our findings . finally in sect .
5 we present our conclusions .
concordance_-cosmology with @xmath42 , @xmath43 , @xmath44 is used throughout this paper .
note that _ metallicity _ and _ abundance _ will be taken to denote _ gas oxygen abundance _ throughout this paper , unless otherwise specified .
in addition , we use dex throughout to denote the anti - logarithm , i.e. , 0.3 dex is a factor of 2 .
the target galaxies for the nir spectroscopy with moircs and sinfoni were selected from the zcosmos - deep galaxy catalog .
zcosmos - deep is measuring several thousand redshifts in a broad redshift range @xmath45 .
we first selected a parent sample which we call _ elpar _ of @xmath46 zcosmos - deep galaxies with redshifts @xmath5 of high confidence classes
2.5 , 3 or 4 , i.e. , reliable redshifts ( see * ? ? ?
class 3 and 4 are very secure redshifts , while class 2.5 are less secure spectroscopic redshifts , but the decimal place .5 indicates that the spectroscopic and photometric redshift of objects with confidence class 2.5 are consistent to within @xmath47 .
x - ray detected objects were excluded from the sample elpar using the xmm @xcite , and chandra cosmos observations @xcite . for the nir spectroscopic follow - up
we only selected those elpar galaxies with redshift confidence classes 3 or 4 , i.e. , the most reliable redshifts based on unambiguous multiple spectral features that leave no room for any doubt about the redshift .
for each galaxy , we estimated an sfr using the model spectral energy distribution ( sed ) fits obtained , at the spectroscopic redshift , by the zebra ( zurich extragalactic bayesian redshift analyzer , * ? ? ? * ; * ? ? ? * ) code applied to the multi - band cosmos photometry in the bands u , b , v , g , r , i , z and ks . we used this sfr to predict an h@xmath32 flux by applying the @xcite conversion of h@xmath32 luminosity into sfr ( cf .
sect.[sfrs ] ) .
we then selected only those objects with sed - based sfrs larger than @xmath48/yr for the sinfoni observations , corresponding to an expected observed h@xmath32 flux larger than @xmath49ergs / s/@xmath50 at these redshifts , but also included targets with lower sed - based sfrs as fill - in objects during the subaru moircs mask preparations .
multi - object spectroscopic observations were performed with moircs @xcite in the second half of the nights of 2010 january 25 - 28 .
we used moircs in mos mode with the k1300 grism for the k - band , vph - h grism for the h - band , and vph - j grism for the j - band observations . the spectral resolution with a 0.8 slit
is @xmath51 in k , and @xmath52 in h and j. eight main target galaxies at @xmath4 were observed with one slit mask , with a total exposure time of @xmath53 hours in each of the bands j , h , and k. the seeing varied between 0.4 and 0.7 .
the target acquisition was done using 6 alignment stars , and the telescope was dithered along the slit by 2.0 .
the single exposure times were 1200s . at the end of each observing night ,
an a0v type star was observed for flux calibration with the same instrument configuration as the targets , and at different positions on the mask for the vph grisms , to take into account that the sensitivity curve shifts for the vph grisms , when the slits at different x positions ( in raw image ) are used .
it should be noted that this sensitivity curve dependence of the vph grisms on the position of the slits was seriously hampering the observations , in terms of restricting a lot the target selection and diminishing the s / n obtained .
the data reduction was done using the moircs mos data pipelines ( mcsmdp ) reduction script ( for more details see * ? ? ?
* their sect.3.1 ) .
this included flatfielding by a dome flat , rejection of bad pixels and cosmic - rays , sky - subtraction between each exposure pair , and distortion correction of the mos frames .
each 2-d spectrogram was then cut out based on the mask design file , wavelength calibration was done based on the oh sky emission lines , residuals of the sky subtraction ( due to time variation of oh emission during dithering observations ) were subtracted by fitting with a quadratic equation along the column , and finally all frames were co - added with appropriate offsets and weights .
the response spectra , including atmospheric transmission and instrument efficiency , were derived using the a0v - type stars hip 68713 and hip 74050 .
the data of these standard stars were reduced to a two - dimensional spectrum in the same way as for the galaxy targets , and the continuum was then traced with a polynomial to extract a one - dimensional spectrum with iraf : apall .
finally the one - dimensional spectrum of the standard star was divided by the best - fit model spectrum to obtain the response spectrum .
the absolute flux calibration was then performed by normalizing to the j- , h- or k - band total magnitude , respectively .
observations with the sinfoni integral field spectrograph @xcite in up to three bands ( j , h , and k ) were performed between the beginning of january and the end of april 2010 .
we used `` on - source '' dithering for all targets , which are sufficiently small to fit in the region of overlap of all science exposures for a target .
briefly , two successive exposures are dithered by roughly 1/2 of the field - of - view ( by about 4 arcsec ) , so that the frames can be subtracted pairwise for background subtraction .
exposure times were 40min for k - band , and 60min for j and h - band observations .
the spectral resolution is @xmath54 in k , @xmath55 in h , and @xmath56 in j - band .
the seeing varied between 0.7 and 1.2 .
the accuracy of the zcosmos - deep redshifts is about @xmath57 at @xmath41 , and they are generally absorption line redshifts from rest - frame ultraviolet ( uv ) features ( from the vimos optical spectroscopy ) , which often show a significant velocity offset with respect to the emission line gas ( see columns @xmath58 and @xmath59 in tab.[sinf_moircs ] ) . since all five emission lines are clear of oh lines for only a limited range of redshifts , our first step was to observe the h@xmath32 ( and [ nii ] ) lines in k - band for the 12 galaxies given in tab.[sinf_moircs ] .
these measurements allowed us to focus the h@xmath31 , [ oiii ] and [ oii ] observations ( in h- and j - band ) on those galaxies lying at optimal redshifts in terms of the oh emission contamination and emission line flux detections of ( most of the ) five lines .
tab.[sinf_moircs ] indicates the lines that could be observed for each object .
we reduced the sinfoni data with the spread software , especially developed for spiffi @xcite .
the data reduction is analogous to standard procedures applied to nir long - slit spectroscopy but with additional processing to reconstruct the three - dimensional ( 3d ) data cube .
the main reduction steps applied to each science target for a given instrument band are as described in sect.4.2 of @xcite . a special step in the reduction
is that the wavelength calibration and sky subtraction steps were repeated with optimization following the method described by @xcite .
the data of the standard stars were reduced in a similar way as the science data .
flux calibration was performed using the broadband magnitudes of the standards .
correction for atmospheric transmission was done by dividing the science cube by the integrated spectrum of the standard .
emission line fluxes were measured from the calibrated spectra following the procedure described in @xcite .
the package splot in iraf was used , and the flux errors were usually dominated by systematic uncertainties in establishing the local continuum . for each individual galaxy , tab.[sinf_moircs ] reports either the emission line fluxes or their upper limits .
all five emission line fluxes could be measured for only a few galaxies because of , e.g. , poor seeing during observations in one of the nir bands , low galaxy surface brightness , the overestimation of some sfrs derived from the sed , and atmospheric oh lines coinciding with the positions of some emission lines , as indicated in the respective columns of tab.[sinf_moircs ] .
we note in particular that : _ ( i ) _ we were able to measure for all 20 galaxies at @xmath5 reported in tab.[sinf_moircs ] . _
( ii ) _ the line was measured for three galaxies , affected by an oh line for seven galaxies , and too faint to be detected for ten objects .
we determined upper limits to the flux for the ten @xmath5 galaxies for which the flux was unaffected by a night sky line , and tab.[sinf_moircs ] lists the @xmath60 upper limits . _
( iii ) _ could be measured for 13 galaxies .
for other two galaxies in which the line was affected by a strong oh line but was not , the flux was derived using the theroretical line flux ratio between and , by multiplying the flux by a factor of three .
three galaxies do not have an h - band spectrum , resulting in no measurement of [ oiii ] and , and two galaxies have both and affected by night sky lines . _
( iv ) _ we were able to measure fluxes for eight of the @xmath61 galaxies . for the other galaxies the line was either too faint to be detected , or coincided with a strong oh line , or the h - band spectrum
was not observed , hence we were unable to measure the line flux .
in the problematic cases , we assumed that the @xmath62 ratio is in the range @xmath63 , corresponding to typical extinction values of star - forming galaxies of @xmath64 .
we used these values for to get an o / h estimate by simultaneously fitting all the lines for metallicity o / h , reddening @xmath36 and ionization parameter @xmath34 , as described in @xcite and in sect.[oxabund ] . _
( v ) _ could be measured for ten galaxies , and we obtained @xmath60 upper limits for two additional galaxies .
four galaxies do not have an j - band spectrum , implying no measurement of [ oii ] , and for one galaxy the region of the spectrum where [ oii ] should be detected is affected by a bad pixel area of the moircs instrument . for three galaxies the [ oii ]
line was affected by a strong oh skyline . to calculate sfrs for the sample we need to know the absolute flux of .
the sinfoni spectra were integrated in circular apertures with radii typically 1-1.5 for these seeing limited data . as mentioned in sect.[sinfredu ] ,
the sinfoni data were reduced with the same procedure as used by @xcite , so the uncertainties of the absolute flux calibration should be @xmath65 10% , as estimated by @xcite . in the good seeing conditions encountered for the moircs observations ( seeing between 0.4 and 0.7 ) , the resulting emission line fluxes from the moircs spectroscopy should approximate `` total '' fluxes , and the absolute calibration
was also checked versus the total j , h or k magnitude , as described in sect.[moircsredu ] .
additionally , we find a good agreement between the sfrs derived from from the sinfoni and moircs spectra and the sfrs derived from uv ( as described in sect.[sfrs ] ) , which reinforces the reliability of the absolute calibration of the nir data .
to proceed with the interpretation of the emission line properties , one has to establish whether the ionization is of stellar origin , or associated with agn activity . to identify galaxies
possibly dominated by an agn , we use the bpt @xcite diagram .
fig.[bpt_z23 ] shows those zcosmos galaxies with nir spectroscopy for which the [ nii]/ and [ oiii]/ ratios ( or upper limits of [ nii ] or ) were measured , as given in tab.[sinf_moircs ] : two galaxies with all four lines measured ( filled magenta circles with error bars ) , five galaxies with [ oiii ] , , and measured and an upper limit of [ nii ] ( filled magenta circles with arrows indicating the [ nii ] upper limits ) , and four galaxies for which could not be measured , but was estimated using the known flux as described in sect.[emline ] ( open circles ) .
galaxies with non - degenerate oxygen abundances ( see sect.[oxabund ] ) are overplotted as yellow crosses .
@xcite combined stellar population synthesis and photoionization models to build the first purely theoretical classification scheme for separating pure agns from galaxies containing star formation ( dashed black line ) . the empirical curve of ( * ? ? ? * solid black line ) separates star - forming galaxies in the local universe ( below / left of the curve ) from agns ( above / right of the curve ) .
the [ nii]/ and [ oiii]/ ratios have been reported in the literature for small numbers of galaxies at @xmath4 : only four galaxies in the eb06 sample have [ oiii]@xmath40 and h@xmath31 line fluxes measured in addition to h@xmath32 and [ nii ] , and these are shown as blue squares in fig.[bpt_z23 ] .
they lie between the theoretical curve of @xcite and the empirical curve of @xcite .
the two open triangles are measurements derived from stacked spectra by @xcite , and they also lie in the same region of the bpt diagram as the eb06 galaxies , and this also applies to the two galaxies from @xcite shown as blue star symbols .
most of the 11 zcosmos galaxies shown in fig.[bpt_z23 ] have larger ratios of the collisionally excited to recombination lines than present - day galaxies ( black circles ) .
all six galaxies with large [ oiii]/ close to the division between sf and agns have high ionization parameters .
as mentioned by @xcite , unusually high ionization parameters might cause the offset of high redshift galaxies in the bpt diagram compared to local ones .
@xcite found that local sdss galaxies with high equivalent widths ( a proxy for ssfr ) at a given mass , are displaced towards larger values of the two emission - line ratios used in the bpt diagram .
the median ssfr of the ms is higher by a factor of 20 at @xmath4 compared to the local ms ( cf .
eq . 1 in * ? ? ?
this implies that ms galaxies at @xmath4 , although displaced towards larger values of the two emission - line ratios in the bpt diagram , may still be ionized by recent star formation instead of agn activity .
recent work by @xcite shows that , for the more extreme ism conditions at higher redshifts ( e.g. , higher ionization parameters ) , the star - forming sequence may evolve compared to local galaxies . in fig.[bpt_z23 ] the open circles show local galaxies with @xmath66 from the nearby field galaxy sample ( nfgs ) of @xcite , and the two dashed blue lines in the left part of fig.[bpt_z23 ] show the upper and lower boundary of the evolved ( compared to local galaxies like the nfgs objects ) star - forming sequence at @xmath67 using the equations in @xcite .
galaxies containing agns form a mixing sequence with purely star - forming galaxies , and this mixing sequence is shown by the two blue dashed lines in the right part of the diagram using the values from tab.1 in @xcite for @xmath68 .
the position of the zcosmos @xmath4 galaxies ( and of galaxies at @xmath4 from the literature depicted in fig.[bpt_z23 ] ) in the region of the bpt diagram occupied by star - forming galaxies at @xmath67 ( indicated by the two blue dashed lines on the left ) implies that they are unlikely to be dominated by agns .
for the 20 zcosmos galaxies at @xmath5 with nir spectroscopy , the luminosities were corrected for extinction using , when available , the @xmath69 values computed as described in sect.[oxabund ] , or assuming an @xmath70 ( roughly the mean value of the galaxies with measured @xmath69 ) when no additional information was available .
these extinction corrected luminosities were used to calculate the sfrs , by applying the @xcite conversion of h@xmath32 luminosity into sfr : @xmath71 .
the sfrs for @xmath4 galaxies of the elpar parent sample were computed from their uv luminosities using the relation based on their 1500 luminosity , following eq.5 of @xcite , namley : @xmath72 , for a @xcite imf .
the extinction correction was done using eq.4 of @xcite , i.e. , @xmath73 .
the procedure is described in detail in sect.5.1.3 of @xcite . using a sample of zcosmos galaxies at @xmath4 with sinfoni nir spectroscopy of h@xmath32 and with a comparable range of stellar masses and sfrs as the galaxies studied here ,
@xcite compared the sfrs inferred from uv and from h@xmath32 , and showed this uv calibration of the sfr to be reliable at @xmath4 ( see , e.g. , their fig .
eb06 used the same value of e(b - v ) for the stellar uv continuum and for the nebular emission lines when calculating their published sfrs .
this is in agreement with several studies who found similar extinction for the stellar continuum and nebular emission ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
however , the relation between stellar and nebular exinction is uncertain , and other studies found that nebular regions are more extincted than the stellar continuum .
@xcite found that , when the additional dust extinction caused by nebular regions is calculated using @xcite , there is a closer agreement between the -derived sfrs and other sfr indicators for @xmath4 galaxies .
therefore , in this paper we re - calculate the ( s)sfrs of the @xmath4 eb06 sample from the published fluxes of @xcite , assuming @xmath74 , following @xcite who found that the balmer emission lines of hii regions are more extincted than the uv stellar emission by this amount .
in addition , in agreement with this prescription , @xcite also found that nebular regions are about two times more extincted than the stellar continuum in their sins sample of star - forming galaxies at @xmath4 .
the optical to irac photometric data from cosmos were used to perform an sed fitting analysis to estimate the galaxy stellar mass .
the _ hyperzmass _ software was used , which is a modified version of the public photometric redshift code _
@xcite , and details of the method are described in sect.2.3 of @xcite . briefly , _ hyperzmass _ fits photometric data points with the @xcite synthetic stellar population models , and picks the best - fit parameters by minimizing the @xmath75 between observed and model fluxes .
the best - fit results depend strongly on the assumptions regarding initial mass function ( imf ) , metallicity , stellar population models , extinction law , and star formation history ( sfh ) .
the stellar mass is obtained by integrating the sfr over the galaxy age , and correcting for the mass loss in the course of stellar evolution .
the choice of the different parameters was discussed and compared by @xcite for the @xmath76 redshift regime , and by @xcite for the @xmath77 redshift range , so we refer the reader to those papers for details . specifically , @xcite showed ( see their sect .
3 ) that cosmos galaxy stellar masses at @xmath77 derived from sed fitting can be considered reliable within @xmath78dex , i.e. , within a factor @xmath79 .
the stellar masses of the eb06 sample are the integral of the sfr over the lifetime of the galaxy ( without correction for the mass loss ) and so represent the total mass of stars formed rather than the mass in living stars at the time of observations .
they have been converted into the living stellar mass by assuming the usual instantaneous return to the ism of a fraction of the mass converted into stars .
we chose this fraction to be @xmath80 like in li13 , based on @xcite stellar population models .
we note that the stellar mass calculated in this paper is always the actual mass of long - lived stars , and when literature data is used we always correct to this definition if necessary .
for the zcosmos @xmath5 galaxies we derived values of o / h and dust extinction @xmath69 using the approach described in @xcite .
this adopts the models of kd02 , who developed a set of ionization parameter and oxygen abundance diagnostics based on the use of strong rest - frame optical emission lines .
the method of @xcite performs a simultaneous fit to all available emission line fluxes in terms of extinction parameter @xmath69 , ionization parameter @xmath34 , and o / h , and is described in detail in sect.3 of @xcite . an upper limit to
[ nii]@xmath81 often suffices to break the @xmath33 degeneracy and allows o / h determination based on this @xmath75 analysis , as shown in @xcite .
it should be noted that the point of the line - fitting approach developed in @xcite is to use all of the data and the best available emission line models ( in this case kd02 ) without applying other priors that may or may not be valid . for the reasons discussed above in sect.[emline ] , the fluxes of all five emission line fluxes could be reliably measured for only a few of the 20 observed @xmath4 galaxies . when was undetected , we estimated the flux as described in sect .
[ emline ] ( see also footnotes in tab.[sinf_moircs ] ) .
we do not correct for the possible effect of balmer stellar absorption lines on the measured line fluxes , because of the high equivalent widths ( ews ) of the emission lines of the galaxies studied . for a wide range of star formation histories and for plausible imfs ,
the stellar absorption at is always @xmath825 ( see , e.g. * ? ? ?
* ) , small compared to the ews of the emission lines in our studied galaxies .
the stellar absorption lines corrections are typically larger for the fluxes , but the line fluxes of the galaxies in our sample have quite large error bars , so this correction should be small compared to these uncertainties .
the error bars of o / h and @xmath69 that we report in tab.[sinf_moircs ] are the formal 1@xmath12 confidence intervals for the projected best - fitting values .
as described in sect.3.3 of @xcite , the error bars of the oxygen abundance are given by the range of oxygen abundance for those models with @xmath83 in the range @xmath84 ( where @xmath85 is the minimum @xmath75 of all allowed models for a given galaxy ) , corresponding to a confidence level of 68.3% for one single parameter .
we consider a possible second solution corresponding to a second ( higher ) peak , only if the minimum @xmath75 of the higher peak lies in the range @xmath86 ( cf .
fig.4 in * ? ? ?
our @xmath75 analysis takes into account all emission line fluxes measured and their error bars , and it turns out that only seven out of the 20 @xmath4 galaxies studied ( 35% of our sample ) get a non - degenerate o / h solution with the achieved quality of the data .
before discussing the mzr relation at @xmath6 , it is useful to consider the location of the @xmath4 galaxies in the ssfr@xmath87mass diagram . @xcite linked the ssfr measurements at @xmath88 and @xmath89 of @xcite and @xcite with the zcosmos data and the sdss relation to derive an equation of the evolution of the ssfr as a function of mass and time ( their eq.1 ) .
the dashed black ( almost horizontal ) lines in fig.[ssfrmass_zgt2 ] show the ssfrs calculated using their eq.1 for @xmath90 and @xmath68 , assuming a weak dependence of ssfr on mass ( as observed for local sdss galaxies ) , ssfr@xmath91m@xmath92 with @xmath93 . @xcite and others showed that the local ms relation has a dispersion of @xmath94dex about the mean relation , which is indicated by the dotted ( almost horizontal ) lines in fig.[ssfrmass_zgt2 ] .
the sfrs for the individual zcosmos galaxies with nir follow - up ( larger magenta symbols in fig.[ssfrmass_zgt2 ] ) were derived from , while for the parent elpar sample ( cyan circles ) we used sfrs derived from uv data as described in sect.[sfrs ] .
we used the values of the individual galaxies from tab.1 in @xcite to derive ssfrs as described in sect.[sfrs ] for the eb06 galaxies .
the oblique dashed lines show sfrs of 10/yr and 30/yr , respectively , which are roughly the selection thresholds of the zcosmos elpar ( cyan filled circles ) and eb06 sample ( blue filled circles ) . because of these sfr selection thresholds , when calculating the median ( s)sfrs at a given mass , the incompleteness at lower masses ( c.f .
* ) produces higher median ssfr values at lower masses than the relation from @xcite , because low mass lower sfr galaxies are missed from the sample ( fig.[ssfrmass_zgt2 ] ) .
this is shown in fig.[ssfrmass_zgt2 ] by the filled blue squares which indicate the mean ssfrs values in the mass bins used by eb06 .
we therefore defined ms objects as galaxies in the region delineated by eq.1 in @xcite , i.e. , the area between the dotted lines in fig.[ssfrmass_zgt2 ] .
we also note that massive galaxies which are highly obscured by dust are probably missed from both the eb06 and the zcosmos samples , as indicated by the lower ssfrs at the high mass end in fig.[ssfrmass_zgt2 ] .
this was also noted and discussed in sect.3.3 of @xcite .
since it affects both the eb06 and the zcosmos samples , this effect should not affect our findings about the difference in metallicities at higher masses between the two samples . in fig.[massoh
] we compare the mzr of galaxies at @xmath5 with the relation of sdss .
for all galaxies studied , the metallicity o / h was derived from nir spectroscopy of at least three emission lines , apart from the eb06 sample , whose metallicities were based on only and [ nii ] . for the local mzr , we used the median and @xmath95 values of sdss galaxies from tab.3 of @xcite , and converted their stellar masses to a @xcite imf and their oxygen abundances to the kd02 calibration by applying the conversion given in tab.3 of @xcite .
the 16th and 84th percentiles , and the medians ( 50th percentiles ) of the distribution of o / h in the respective mass bin are shown as black thick lines for sdss .
the three thinner black lines show the median sdss mzr , shifted downward by 0.3 , 0.5 and 0.7 dex , respectively .
our sinfoni and moircs based o / h measurements for the seven galaxies with non - degenerate metallicity measurements are shown as magenta filled circles .
literature data of four gravitationally lensed galaxies at @xmath5 @xcite , with o / hs converted to the kd02 calibration by applying the conversion given in tab.3 of @xcite , are depicted as black filled circles .
analysing their 3d - hst nir spectroscopy data , @xcite used [ oii ] , and [ oiii ] emission lines in stacked galaxy spectra to measure o / h metallicities at @xmath4 ; these data are shown in fig.[massoh ] as black open circles .
the mzr of eb06 , with o / hs from the n2-method @xcite converted to the kd02 calibration , is shown in blue .
the o / h metallicities of zcosmos galaxies and most literature values obtained using at least three emission lines are systematically lower by a factor of three to five , at a given mass , than the sdss relation . on the other hand ,
the [ nii]/-based metallicities from stacked spectra of the work by eb06 are lower by a factor of two compared to sdss .
this discrepancy can be due to : ( i ) the different selections of the samples ; ( ii ) type-2 agn contamination in the eb06 sample ; and/or ( iii ) different metallicity calibrations used . the first point ( i )
can be addressed using fig.[ssfrmass_zgt2 ] : both the elpar ( and the zcosmos galaxies with nir follow - up ) and the eb06 samples are quite representative of ms galaxies of relatively high mass . at lower masses
@xmath96 , elpar galaxies are more representative of ms galaxies , while eb06 galaxies lie predominantly above the ms , which could bias them towards _ lower _ metallicities .
however , the eb06 metallicity measurements are too high compared to zcosmos and the @xcite sample , and correcting a bias towards lower metallicity , by making the eb06 metallicities higher in their lowest two mass bins , would even increase the discrepancy at lower masses . regarding the second point ( ii ) , we note that only four out of 87 galaxies in the eb06 sample have [ oiii]@xmath40 and h@xmath31 line fluxes measured in addition to those of h@xmath32 and [ nii ] .
this means that no information on type-2 agn contamination is available for 83 galaxies .
recent work on the agn fraction at high redshift ( e.g. , * ? ? ?
* ) claims the existence of a non - negligible fraction of nuclear activity in higher mass galaxies at @xmath97 .
if any type-2 agns were among the spectra stacked by eb06 to measure metallicity , this would have systematically increased their measured [ nii]/ ratios and hence their derived o / h metallicities , when using the n2 calibration of @xcite .
related to the third point ( iii ) , as mentioned above , the oxygen abundances of eb06 have been converted from the @xcite to the kd02 calibration by applying the conversion given in tab.3 of @xcite . to verify
if the @xcite conversion , which was calibrated for local galaxies , works at @xmath4 , a larger sample of galaxies with five emission lines measured is needed at @xmath4 .
the ionization parameter dependence of the n2-method is addressed in both sect.[intro ] and fig.[n2_ionpar ] , and may be an important drawback of the analysis of eb06 based on the n2-method .
@xcite indeed found that the mzr at @xmath88 determined using the n2-method might be @xmath98 times too high in terms of metallicity , when the effects of both photoionization and shocks are not taken into consideration . assuming that the @xcite calibration really holds at @xmath4 , and that the eb06 metallicities are not strongly affected by either type-2 agn contamination or the ionization parameter
, this would imply that the gas - phase metallicities of galaxies at @xmath4 are between one half and one fifth of those of nearby galaxies of similar mass .
however , we think that the n2-method issues ( including the ionization parameter dependence ) and the type-2 agn contamination are serious issues influencing the eb06 metallicity measurements .
ma10 proposed the existence of an fmr based on comparing sdss measurements with a limited range in sfr and stellar mass , and higher redshift data which have higher values of sfr that are not observed in sdss ( @xmath99 ) . the claim that the eb06 @xmath4 galaxies follow the same fmr as sdss galaxies
is therefore based on an extrapolation because of the much higher median ssfr of the ms at @xmath4 compared to the local ms ( cf . * ? ? ?
eqs.2 , 4 and 5 of ma10 are used to predict the o / hs of high redshift galaxies following a redshift independent z(m , sfr ) given their sfr and masses , and ma10 claimed that these metallicities are observed up to @xmath21 ( but not at @xmath22 ) , i.e. , that a redshift independent z(m , sfr ) holds . for the convenience of the reader , we reproduce here eq.2 of ma10 : @xmath100 where @xmath101=log(m)10 and @xmath102=log(sfr ) in solar units .
because eq.5 of ma10 is quite similar to their eq.4 , we will not show their eq.4 for the comparison in the following and focus only on eq.5 , which was reproduced as eq.1 in sect.[intro ] .
we also reproduce eq.40 of li13 : @xmath103 where @xmath104 is the equilibrium value of the metallicity , @xmath105 is the metallicity of the infalling gas , @xmath106 the yield , @xmath107 the returned fraction , @xmath27 the mass - loading factor , @xmath26 the star formation efficiency .
the values of the different parameters @xmath26 , @xmath27 , @xmath106 , @xmath108 are given in tab.1 of li13 .
it is clear from the above that whether or not the high redshift galaxies fall on a non - evolving fmr depends on how the local z(m , sfr ) relation is extrapolated .
we therefore first explore how the non - evolving prediction up to @xmath109 differs when using the original formulations of the local z(m , sfr ) of ma10 compared to the more physically - based formulations in li13 . by construction
, these should agree in the part of the ( m , sfr ) plane that is well - sampled by sdss local galaxies .
we created a grid of oxygen abundances as a function of sfr and stellar mass , with log(sfr / m@xmath110/yr ) ranging from -2.0 to 3.0 in 51 steps of 0.1 dex , and log(m / m@xmath110 ) ranging from 8.0 to 12.0 in 41 steps of 0.1 dex .
at every ( sfr , mass ) grid point we then calculate the expected oxygen abundance using eqs.2 and 5 from ma10 and using eq.40 of li13 with a redshift - independent @xmath111 , i.e. , using in each case a non - evolving z(m , sfr ) but differing in the form of the relation . because ma10 uses a @xcite imf
, we converted the masses and sfrs to a @xcite imf ( as used in this paper ) after computation of the expected o / hs , by adding 0.23dex in mass and sfr , respectively .
the results are shown in fig.[fmrma10li13 ] , the ssfr - mass relation , color - coded according to the expected metallicities from the different z(m , sfr ) formulations of ma10 and li13 . as discussed in li13
, the form of the high redshift @xmath3 relation will depend on whether the parameters of the regulator system change with time .
in particular , there are theoretical reasons to suppose that the star - formation efficiency ( their @xmath26 parameter ) might scale roughly as @xmath112 if the gas depletion timescale scales as the dynamical time of the system .
observationally , this is an open question . in the spirit of investigating an epoch independent @xmath3
, we here consider henceforth the case that @xmath26 is epoch - independent , i.e. , the case represented by the dashed lines in fig.7 of li13 .
the black open circles show the parameter space in sfr and mass for which the predicted o / hs from li13 and ma10 differ by less than 0.05dex .
not surprisingly , this roughly corresponds to the range in observed sfr ( @xmath113 ) and mass values of sdss galaxies , from tab.1 in ma10 , which were both used to fit the parameters in eq .
40 of li13 and to derive eqs . 2 and 5 of ma10 .
it is obvious from fig.[fmrma10li13 ] that _ the extrapolation of eq .
40 of li13 and eqs . 2 and 5 of ma10 outside the sfr - mass parameter space area that was covered by sdss is rather different .
_ as an example , this can be seen in the _ grid _ color ( expected o / h ) for a typical star - forming massive galaxy at @xmath4 with log(m / m@xmath114 and @xmath115 , as indicated by the crossing point of the magenta lines in the four panels : the expected o / h decreases from panel d ) to b ) to a ) to c ) , as indicated by the green , blue and cyan colors of the grid points at the respective crossing point of the magenta lines .
fig.[sfrmass_obs ] shows the ssfr - mass diagrams with the grid color - coded in metallicities like in fig.[fmrma10li13 ] , but for a smaller range in mass and ssfrs corresponding to the range typically covered by galaxies at @xmath4 .
the galaxies with nir spectroscopy at @xmath4 are overplotted : the seven zcosmos galaxies with non - degenerate metallicities ( filled circles ) , the eb06 sample ( filled triangles ) , and the @xcite sample ( filled squares ) .
all data points are color - coded according to their metallicities ( e.g. , green color for @xmath116 , c.f .
fig.[fmrma10li13 ] ) , to explore if the measured metallicities ( color of the _ data _ points ) and the expected metallicity assuming z(m , sfr ) is redshift independent ( color of the _ grid _ points ) are in agreement . at a given mass ,
e.g. , at log(m / m@xmath117 , lower metallicities ( blue data points in fig.[sfrmass_obs ] ) are mainly observed for higher ( s)sfrs , in contrast to higher metallicities ( green data points ) for lower ( s)sfr . this explicitly confirms , albeit with only a few objects , that the _ sfr is still a second parameter in the mass - metallicity relation at @xmath4_. whether the z(m , sfr ) relation evolves or not compared to @xmath118 , i.e. , whether it is `` fundamental '' , depends on the extrapolation used . as in fig.[ssfrmass_zgt2 ] , ms galaxies are assumed to lie in the region between the black dashed / dotted ( almost horizontal ) lines in the panels of fig.[sfrmass_obs ] .
the black circles in all panels roughly indicate , as explained above , the highest ( s)sfrs observed , at a given mass , for sdss galaxies . it is obvious that all @xmath4 galaxies have higher sfrs ( @xmath99 ) than the sdss galaxies .
this means that _ trying to address the question if the local @xmath3 relation holds at @xmath4 , i.e. , if it is truly an fmr , depends on the extrapolation used . _ the zcosmos @xmath5 galaxies with metallicities measured in this study have log(m / m@xmath119 , i.e.
, they lie to the right of the magenta line in each panel of fig.[sfrmass_obs ] . using the ma10 z(m , sfr ) extrapolation shown in panel d
, a typical ms star - forming massive galaxy at @xmath41 with log(m / m@xmath119 can not have o / h@xmath120 if it follows a non - evolving fmr : for such ms galaxies , in the region between the dashed / dotted lines , the grid points of the predicted fmr in the panel d are green and red , i.e. , only o / h@xmath121 are expected if the fmr holds at @xmath4 following eq.5 of ma10 .
lower metallicities ( blue / cyan grid points ) are however predicted for ms @xmath41 galaxies up to log(m / m@xmath122 by li13 ( panels a and c ) .
o / h@xmath120 values ( blue grid points ) are also predicted by eq.2 in ma10 for ms galaxies with lower masses , log(m / m@xmath123 ( see panel b ) .
therefore , the @xmath41 zcosmos galaxies and the @xcite data points , which almost all have @xmath124 , are actually quite in agreement with the predictions for a non - evolving fmr based on the li13 formulation .
the metalllicities of some of the zcosmos galaxies are in agreement with the extrapolation from eq . 2 of ma10 , especially at lower masses , but they are not generally in agreement with the fmr predicted by the extrapolation from eq . 5 of ma10 .
this is shown in panel d by the disagreement between the blue / green color of most zcosmos data points and the green / red grid points .
for the eb06 sample , we do not show the lowest mass bin , which only has an upper limit for the metallicity .
the other five mass bins all have measured @xmath125 ( green and red filled triangles ) .
the measured metallicities in these five mass bins agree with the non - evolving predictions of eq.5 of ma10 , as shown in panel d by the agreement between the color of the grid points and the color of the data points ( triangle symbols ) .
however , the eb06 data agree with the ( non - evolving ) predictions of li13 or eq . 2 of ma10 ( panel b ) only for the highest mass bins .
this means that an fmr based on sdss and eb06 data , as claimed by ma10 , only exists when considering the particular form for the extrapolation represented by eq .
5 of ma10 . in general
, it should be noted that any @xmath124 metallicity measurement for a massive ( log(m / m@xmath114 ) @xmath4 galaxy will be considered as an `` outlier '' from a non - evolving fmr that is based on eq . 5 ( or the similar eq .
4 ) of ma10 .
this also applies to the metallicities of the @xcite and @xcite @xmath22 samples , which were therefore claimed by ma10 to need an evolving z(m , sfr ) relation . on the other hand ,
using the li13 extrapolation , such @xmath22 galaxies with low metallicities can be consistent with a non - evolving fmr .
the basic point is that the metallicity data of future larger samples of galaxies at @xmath22 that will soon become available with nir spectroscopy should be compared to both the ma10 and li13 predictions in order to establish whether the fmr evolves or not at @xmath22 .
the implied evolution , or lack of it , in the z(m , sfr ) relation will usually depend on the extrapolation of the local relation to higher sfrs .
this study of the mzr and fmr at @xmath4 differs from other studies by carefully addressing following issues : ( i ) a metallicity calibration using 23 based on up to five lines rather than the [ nii]/ ratio which has a strong ionization parameter dependence ( see fig.[n2_ionpar ] ) ; ( ii ) the location of the @xmath4 galaxies in the bpt @xcite diagram , which contains information about the conditions of the ism at these redshifts ( see fig.[bpt_z23 ] ) ; ( iii ) the discussion of the selection of the samples using the ( s)sfr - mass diagram ( see fig.[ssfrmass_zgt2 ] ) ; and ( iv ) noticing the different formulations of the fmr based on local galaxies data which give different extrapolations at @xmath4 ( see fig.[fmrma10li13 ] ) .
the mean mzr at @xmath5 is shifted to lower metallicities by a factor of three to five compared to the local sdss relation ( fig.[massoh ] ) .
this shift is larger than the factor of two evolution found by eb06 .
a possible cause for this is the unknown type-2 agn contamination for 83 out of 87 galaxies in the eb06 sample , with too high emission lines possibly mimicking too high metallicities .
another cause is the strong ionization parameter dependence of the @xcite method , especially at higher metallicities like the ones in the eb06 sample ( see fig.[n2_ionpar ] ) , which can produce large errors in the metallicity measurements .
the metallicities expected at high redshift for a non - evolving fmr up to @xmath109 depend on the formulation and thus the extrapolation of the low redshift z(m , sfr ) relation , since the high redshift galaxies lie in a different part of the ( m , sfr ) plane .
the metallicities reported here and in @xcite are more consistent with a non - evolving fmr if the physically - motivated z(m , sfr ) from li13 is used for the extrapolation rather than the empirical z(m , sfr ) relation from ma10 .
this study emphasizes the difficulties of measuring metallicities using three to five emission lines at @xmath4 : we could obtain non - degenerate oxygen abundances for only 7 out of 20 galaxies ( 35% of the sample ) with the achieved quality of the data .
the error bars of some emission lines , especially [ nii ] and , are quite large , resulting in quite large error bars for the oxygen abundances .
this indicates that longer exposure times and/or more sensitive detectors are necessary to better constrain the metallicities of high - redshift galaxies .
therefore , much larger samples with nir spectroscopy of galaxies at @xmath4 , which now become feasible with the advance of multi - object spectrographs like fmos on subaru , mosfire on keck , or kmos on vlt , are needed to put these findings on solid footings .
we are grateful the anonymous referee for his / her suggestions which have improved the paper .
we greatly acknowledge the contributions of the entire cosmos collaboration , consisting of more than 70 scientists .
cccccccccccc 401925&2.1394 & 2.1418 & oh & 7 @xmath126 2 & 42 @xmath126 3 & 20 @xmath126 2 & @xmath821 & 8.44@xmath127&0.03@xmath128&sinfoni&9.97 + 404360&2.2451 & 2.2453 & no j spectrum & no h spectrum & no h spectrum & 12 @xmath126 1 & @xmath821&-&-&sinfoni&10.23 + 404724&2.4750 & 2.4771 & 20 @xmath126 3 & 8 @xmath126 3 & 33 @xmath126 2 & 30 @xmath126 3 & oh&8.64@xmath129&0.49@xmath130&sinfoni&11.00 + 407711&2.1062 & 2.1080 & @xmath825 & 5.3 @xmath1261.7 & 22 @xmath126 3 & 20 @xmath126 2 & @xmath821.7&8.70@xmath131&0.85@xmath132&sinfoni&10.63 + 408147&2.1718 & 2.1746 & 14 @xmath126 1 & 11.5 @xmath126 4 & 14 @xmath126 3 & 45 @xmath126 3 & [email protected] & 8.53@xmath133&2.50@xmath134&sinfoni&10.78 + 408530 & 2.2841 & 2.2834 & 4.2 @xmath126 0.6 & 3.7 @xmath126 2 & 21 @xmath126 4 & 22 @xmath126 4 & @xmath822.9 & 8.60@xmath135&1.75@xmath136&moircs&10.29 + 408609 & 2.4425 & 2.4466 & 1.5 @xmath126 0.5 & oh & oh & 7.8@xmath126 1 & oh & -&-&moircs&10.17 + 408888 & 2.3757 & 2.3783 & 3.3 @xmath126 0.5 & oh & oh & 13.8 @xmath126 1.7 & oh & -&-&moircs&10.47 + 409130 & 2.2784 & 2.2837 & 2.8 @xmath126 0.3 & 1.5 @xmath126 0.5 & 8 @xmath126 2 & 6 @xmath126 1 & @xmath820.5&8.49@xmath137&0.61@xmath138&moircs&10.40 + 409585&2.1909 & 2.1943 & @xmath824.2 & 3.0 @xmath126 1.0 & 20.5 @xmath126 1 & 9 @xmath126 1.0 & @xmath820.8&8.40@xmath139&0.02@xmath140&sinfoni&10.18 + 409985&2.4505 & 2.4579 & 12 @xmath126 2 & 3.5 @xmath126 1.0 & 22 @xmath126 2 & 11.2@xmath126 0.8 & 1 @xmath126 0.2&8.59@xmath141&0.05@xmath142&sinfoni&10.42 + 410123&2.1993 & 2.1993 & 10 @xmath126 2 & 3.6 @xmath126 1.3 & 7 @xmath126 2 & 14 @xmath126 2 & oh&8.45@xmath129&1.13@xmath143&sinfoni&10.01 + 410432&2.4498 & 2.4556 & no j spectrum & no h spectrum & no h spectrum & 7 @xmath126 1 & @xmath820.9&-&-&sinfoni&10.81 + 411785&2.2429 & 2.2450 & 17 @xmath126 3 & 18 @xmath126 2 & 110 @xmath126 10 & 55 @xmath126 5 & 4 @xmath126 1.0&8.57@xmath144&0.25@xmath145&sinfoni&10.42 + 412952&2.3183 & 2.3211 & no j spectrum & 2.7 @xmath126 1.0 & 7 @xmath126 0.5 & 12 @xmath126 4 & oh&8.50@xmath146&1.41@xmath147&sinfoni&10.44 + 428671 & 2.2969 & 2.2953 & oh
& no h spectrum & no h spectrum & 2.8@xmath126 0.9 & oh & -&-&moircs & 9.80 + 429111 & 2.2750 & 2.2816 & oh & 1.2 @xmath126 0.5 & 3 @xmath126 0.5 & 4.8@xmath126 1 & @xmath820.6&8.35@xmath148&1.13@xmath149&moircs & 9.74 + 429152 & 2.2806 & 2.2811 & 6 @xmath126 2 & 2.1 @xmath126 0.2 & 13.5 @xmath126 0.5 & 7 @xmath126 0.3 & @xmath820.3&8.42@xmath150&0.11@xmath151&moircs & 9.85 + 429182 & 2.3750 & 2.3801 & bad pixels area & 1.4 @xmath126 0.5 & 5.1 @xmath126 0.5 & 5.5@xmath126 1.3 & oh & -&-&moircs & 9.97 + 429868&2.4414 & 2.4469 & no j spectrum & 2.5 @xmath126 1 & 9 @xmath1262 & 10 @xmath126 2 & @xmath821.7&8.58@xmath152&1.00@xmath153&sinfoni&10.07 + the [ nii]/h@xmath32 ratio vs. metallicity as a function of the ionization parameter @xmath34 ( different dotted lines indicate different ionization parameters ) based on fig.7 in kd02 . to derive the mzr at @xmath4 , eb06 used the averaged [ nii]/h@xmath32 ratios in six stellar mass bins and estimated the o / h metallicities of these six averaged spectra using the n2 calibration @xcite .
the horizontal lines indicate the measurements of [ nii]/h@xmath32 and the inferred range in o / hs , while the filled blue circles show the o / h measurements derived by eb06 , transformed to the kd02 calibration framework .
it is obvious that , without knowledge of the ionization parameter ( dotted lines ) , the metallicity estimates get very large error bars , especially at high metallicities : the six horizontal blue lines of possible o / hs at a given [ nii]/h@xmath32 extend over @xmath39dex .
, width=453 ] the bpt @xcite diagram .
zcosmos galaxies with sinfoni and moircs nir spectroscopy are shown , for which the [ nii]/ and [ oiii]/ ratios or upper limits of [ nii ] were measured ( arrows ) or was estimated from as described in sect.[emline ] ( open circles ) .
galaxies with non - degenerate oxygen abundances ( see sect.[oxabund ] ) are additionally overplotted with yellow crosses .
the theoretical curve of ( * ? ? ?
* dashed black line ) and the empirical curve of ( * ? ? ? * solid black line ) , which separate star - forming galaxies in the local universe ( below / left of the curves ) from agns ( above / right of the curves ) are shown .
both the [ nii]/ and [ oiii]/ ratios have been reported in the literature for small numbers of galaxies at @xmath4 ( blue open squares , triangles and stars ) .
recent work by @xcite shows that , for more extreme ism conditions at higher redshifts ( e.g. , higher ionization parameter ) , the star - forming sequence may evolve compared to local galaxies like the nfgs sample @xcite , as shown by the dashed blue lines in the left part of the diagram ( normal star - forming galaxies at @xmath67 should lie between the two blue lines ) .
this implies that all zcosmos @xmath4 galaxies depicted in the figure are unlikely to be dominated by agns . ,
width=529 ] the observed ssfr@xmath87mass relation for galaxies at @xmath4 .
sfrs for the individual zcosmos galaxies with nir follow - up ( magenta symbols ) were derived from , while for the parent zcosmos sample elpar ( cyan circles ) we used sfrs derived from uv as described in sect.[sfrs ] .
the larger filled magenta circles denote the 7 galaxies with non - degenerate metallicities , the 7 big open circles are for galaxies with double - valued metallicities , and the 6 open triangles for the zcosmos galaxies without an o / h estimate .
the parent sample elpar , the zcosmos galaxies with nir folllow - up , and the eb06 galaxies ( blue circles for individual measurements , blue squares for the mean values ) are quite representative for ms galaxies ( between dashed / dotted , almost horizontal lines ) , especially at the higher masses . , width=453 ] the mzr of @xmath5 galaxies , compared to sdss . for all galaxies o / hs were derived based on nir spectroscopy of at least three emission lines , except for the eb06 data ( filled blue circles ) .
our sinfoni and moircs o / h measurements are shown as magenta circles .
literature data of four gravitationally lensed galaxies at @xmath5 @xcite , are depicted as black filled circles .
metallicities from stacked spectra of galaxies with [ oii ] , and [ oiii ] lines measured , from a recent work by @xcite at @xmath67 , are shown as black open circles .
the three thinner black lines show the median local mzr of @xcite , shifted downward by 0.3 , 0.5 and 0.7 dex , respectively .
the metallicities of zcosmos objects and most galaxies from literature with at least three lines measured are shifted to lower metalllicities , at a given mass , by a factor of three to five compared to sdss , while the [ nii]/-based metallicities from stacked spectra of the work by eb06 are lower by a factor of two compared to sdss .
this can be due to the type-2 agn contamination in the eb06 sample and/or problems with the n2-metallicity calibration method used by eb06 ( see discussion in sect.[mzr ] ) .
, width=453 ] the ssfr - mass relation color - coded according to the expected metallicities derived using eq.40 of li13 , with a redshift independent @xmath154 and the parameters from tab.1 in li13 for different infall metallicities @xmath105 relative to the yield @xmath106 , @xmath155 ( panel a ) and @xmath156 ( panel c ) , and using eq.2 ( panel b ) and eq.5 ( panel d ) from ma10 . the black open circles in all panels show the parameter space in sfr and mass for which the predicted o / hs from li13 and ma10 differ by less than 0.05dex .
this roughly corresponds to the range in observed sfr and mass values of sdss galaxies , from tab.1 in ma10 , which were both used to fit the parameters in eq.40 of li13 and to derive eqs.2 and 5 of ma10 .
, width=529 ] the colors of the grid points indicate , like in fig.[fmrma10li13 ] , the metallicities expected for the different z(m , sfr ) extrapolations of ma10 and li13 . additionally , galaxies with nir spectroscopy at @xmath4 are overplotted : zcosmos galaxies ( filled circles ) , eb06 data ( filled triangles ) , and the @xcite stacked measurements ( filled squares ) .
all data points are color - coded according to their _ measured _ metallicities ( e.g. , green color for @xmath116 , see legend in fig.[fmrma10li13 ] ) , to explore if the measured metallicities ( color of the _ data _ points ) and the expected metallicity assuming a redshift independent z(m , sfr ) relation ( color of the _ grid _ points ) are in agreement . , width=529 ] | in the local universe , there is good evidence that , at a given stellar mass @xmath0 , the gas - phase metallicity @xmath1 is anti - correlated with the star formation rate ( @xmath2 ) of the galaxies .
it has also been claimed that the resulting @xmath3 relation is invariant with redshift - the so - called fundamental metallicity relation " ( fmr ) . given a number of difficulties in determining metallicities , especially at higher redshifts , the form of the @xmath3 relation and whether it is really independent of redshift is still very controversial . to explore this issue at @xmath4 , we used vlt - sinfoni and subaru - moircs near - infrared spectroscopy of 20 zcosmos - deep galaxies at @xmath5 to measure the strengths of up to five emission lines : , , , , and .
this near - infrared spectroscopy enables us to derive o / h metallicities , and also sfrs from extinction corrected measurements .
we find that the mass - metallicity relation ( mzr ) of these star - forming galaxies at @xmath6 is lower than the local sdss mzr by a factor of three to five , a larger change than found by @xcite using [ nii]/-based metallicities from stacked spectra .
we discuss how the different selections of the samples and metallicity calibrations used may be responsible for this discrepancy .
the galaxies show direct evidence that the sfr is still a second parameter in the mass - metallicity relation at these redshifts .
however , determining whether the @xmath3 relation is invariant with epoch depends on the choice of extrapolation used from local samples , because @xmath4 galaxies of a given mass have much higher sfrs than the local sdss galaxies .
we find that the zcosmos galaxies are consistent with a non - evolving fmr if we use the physically - motivated formulation of the @xmath3 relation from @xcite , but not if we use the empirical formulation of @xcite . |
after the breakthrough in the preparation and detection of bose einstein condensates of rubidium @xcite , sodium @xcite , and lithium @xcite , the second generation of experiments on ultracold , magnetically trapped bose gases now starts to probe condensate properties such as the temperature dependence of the condensate fraction @xcite or collective excitations of the condensate @xcite .
a matter - of - principle experiment has shown the feasibility of an output coupler for condensed trapped atoms @xcite , two different , overlapping condensates have been produced in a single trap @xcite , and interference fringes between two freely expanding condensates have been demonstrated @xcite , thus proving first - order coherence of `` mesoscopic '' matter waves .
higher - order coherence has been inferred from three - body recombination rates of condensed atoms @xcite , and even the propagation of sound in a bose condensate could be directly observed @xcite .
these developments naturally force the theorist to investigate a conceptually important question : when dealing with bose einstein condensates consisting of about @xmath0 to @xmath1 particles in a magnetic trap , to which extent can the different thermodynamical ensembles be considered as equivalent ?
a bose gas cooled below the condensation point and kept in a trap is neither in thermal contact with a heat bath , nor exchanging particles with some particle reservoir .
hence , the only ensemble that is appropriate for describing this situation is the microcanonical one . and
yet , the theoretical discussion of trapped condensates so far mainly relies on the grand canonical ensemble . within the grand canonical ensemble ,
the description of a trapped , ideal bose gas consisting of @xmath2 particles is comparatively simple .
we consider an isotropic harmonic trap with oscillator frequency @xmath3 , so that the degree of degeneracy of the @xmath4-th single - particle state is @xmath5 hence , the difference @xmath6 of the chemical potential and the single - particle ground state energy is determined by the equation @xmath7 - 1 } \ ; , \ ] ] where @xmath8 denotes the boltzmann constant , and @xmath9 is the temperature .
this sum can be evaluated approximately by introducing the density of states @xmath10 with @xmath11 , which follows immediately from eq .
( [ deg ] ) . converting the sum into an integral and applying standard arguments
, one then finds the condensation temperature @xcite @xmath12 \ ; , \label{tco}\ ] ] where @xmath13 denotes the condensation temperature pertaining to the large-@xmath2-limit @xcite ; @xmath14 is the riemann zeta function .
the lowering of @xmath15 with respect to @xmath16 is of order @xmath17 and results from the enhancement of the density of states ( [ rho ] ) above the leading `` volume''-term : since there are more states available , the need to condense arises only at lower temperatures .
it should be noted that , strictly speaking , there is no well - defined condensation temperature for a gas consisting of a finite number of particles .
however , fig .
1 clearly demonstrates that the onset of condensation in a three - dimensional harmonic trap becomes quite sharp already for particle numbers of the order of @xmath18 , and eq .
( [ tco ] ) describes this onset very well .
in contrast to the textbook case of the free bose gas , the heat capacity of the harmonically trapped bose gas exhibits a steep drop at the condensation point . as shown in fig . 2 , this drop becomes a discontinuous jump by about @xmath19 per particle in the large-@xmath2-limit .
remaining within the scope of the grand canonical ensemble , this analysis can be made formally more precise , and generalized to moderately anisotropic harmonic traps @xcite , which changes the value of @xmath20 .
but can one rely on the grand canonical ensemble ?
ziff , uhlenbeck , and kac have made a case that _ the grand canonical ensemble does not represent any physical situation for the condensed ideal bose gas _ , and advocate that _
its anomalous predictions should be ignored _
these `` anomalous predictions '' are related to what we call the `` grand canonical fluctuation catastrophe '' : within the grand canonical ensemble , the r.m.s .- fluctuations @xmath21 of the ground state occupation number @xmath22 are given by @xcite @xmath23 implying that @xmath21 is of the order of the total particle number @xmath2 below the condensation point .
however , if we consider an isolated bose gas in a trap , all particles occupy the ground state at zero temperature , so that the true fluctuations of @xmath22 vanish .
the grand canonical prediction for @xmath21 thus differs drastically from the microcanonical one .
since the magnitude of the fluctuations @xmath21 is related to the coherence properties of the condensate , it is of substantial interest to compute the true , i.e. , microcanonical fluctuations for a trapped ideal bose gas .
we will study an isotropic harmonic trapping potential with oscillator frequency @xmath3 in @xmath24 dimensions , and denote by @xmath25 the number of excitation quanta for some preassigned value of the excitation energy @xmath26 : @xmath27 the prime task now is to determine the number @xmath28 of microstates .
since there are generally many microstates where only a part of the @xmath2 particles carries all @xmath25 excitation quanta , leaving the other particles in the ground state , @xmath28 equals the number of possibilities for distributing the @xmath25 quanta over _ at most _
@xmath2 bose particles .
then the difference @xmath29 is the number of possibilities for distributing @xmath25 quanta over _ exactly _ @xmath30 particles , with @xmath30 ranging from @xmath31 to @xmath2 , so that @xmath32 is the probability for finding @xmath30 out of @xmath2 particles excited when the total excitation energy is @xmath33 . since the remaining @xmath34 particles occupy the ground state , the first moment @xmath35 of the distribution ( [ dis ] ) yields die microcanonical expectation value of the ground state occupation number according to @xmath36 the corresponding fluctuations follow from @xmath37
the case @xmath38 had been considered already in 1949 by temperley @xcite , and shortly thereafter by nanda @xcite .
it is of particular interest , since it can be treated in detail analytically with the tools furnished by the theory of the partion of integers : the number @xmath39 of microstates corresponds to the number of partitions of the integer @xmath25 into at most @xmath2 summands ; the commutativity of the summands reflects the indistinguishability of the bosons .
we now have to distinguish two cases : if @xmath40 , then the fact that the number of particles is finite has no consequences for the enumeration of microstates , and we are dealing with _ unrestricted _ partitions of @xmath25 . if @xmath41 , the partitions of @xmath25 are _ restricted _ by the requirement that the number of summands , corresponding to the number of excited particles , does not exceed @xmath2 .
let us first consider the case @xmath40 .
we follow the usual convention and denote the number of unrestricted partitions of @xmath25 as @xmath42 . introducing the dimensionless inverse temperature @xmath43 , it is an elementary exercise to show that @xmath44 which means that the generating function for @xmath42 corresponds , physically speaking , to the canonical partition function of a fictituous system of infinitely many distinguishable harmonic oscillators with frequencies that are integer multiples of @xmath3 . given this , it is natural to introduce the new variable @xmath45 and to extract the numbers @xmath42 by inverting eq .
( [ gen ] ) with the help of the saddle point approximation , after expressing @xmath46 in terms of @xmath25 .
but we have to be careful : for temperatures that may physically be considered as `` low '' , but are still high compared to @xmath47 , the variable @xmath48 is close to unity , so that we need to know the behaviour of @xmath49 in the vicinity of a singularity
. however , there exists the remarkable identity ( see eq .
( 1.42 ) in ref .
@xcite ) @xmath50 that links the low - temperature behaviour of our fictituous oscillator system to its high - temperature dynamics .
this allows us to derive the low - temperature approximation @xmath51 which , in turn , yields the desired energy - temperature relation : @xmath52 now we can compute the canonical entropy @xmath53 and apply `` bethe s theorem '' @xcite to get @xmath54 } { \left(-2\pi\frac{\partial n}{\partial \xi}\right)^{1/2 } } \ ; = \ ; \frac{1}{4\sqrt{3}n}\exp\!\left(\pi\sqrt{\frac{2}{3}n}\right ) \ ; .\ ] ] the right hand side is nothing but the celebrated hardy ramanujan formula for the number of unrestricted partitions @xcite , so we have just sketched a physicist s solution to a number - theoretical problem .
there is one detail that deserves particular attention : bethe s theorem amounts to the inversion of the canonical partition sum ( [ gen ] ) within the saddle point approximation , so that the dominant contributions to this sum are properly taken into account within the gaussian approximation . as a result , the microcanonical entropy @xmath55 differs from the canonical entropy @xmath56 by the saddle point correction @xmath57 .
such differences between thermodynamical quantities pertaining to different ensembles are characteristic for small systems . expressed the other way round , the usually assumed equality of thermodynamical quantites in different ensembles holds to the extent that such saddle point corrections can be neglected .
now we can turn to the case @xmath41 , where the number of quanta exceeds the number of particles , and have to determine the number @xmath58 of restricted partitions of @xmath25 . in principle , one can proceed as in the previous case , since there is the identity @xcite @xmath59 the generating function for @xmath60 , that is , the canonical partition function for @xmath2 ideal bosons trapped by a one - dimensional harmonic potential , equals the canonical partition function of a system of @xmath2 harmonic oscillators with frequencies @xmath3 , @xmath61 , @xmath62 , @xmath63 .
the corresponding asymptotic formula for @xmath60 , which turns out to be rather intricate , has been given by auluck and kothari @xcite .
however , a beautiful theorem due to erds and lehner @xcite helps to simplify the analysis : if , for some given @xmath25 and @xmath64 , the number @xmath2 obeys @xmath65 then @xmath66 hence , for given large values of @xmath25 and @xmath2 we _ define _ @xmath64 according to @xmath67 and obtain the approximation @xmath68 in order to check this approximation , we first determine the numbers @xmath69 by means of the saddle point inversion of eq . ( [ res ] ) , evaluate the distributions ( [ dis ] ) , and compute the relative fluctuations @xmath70 according to eq .
( [ flu ] ) .
the results for @xmath71 , @xmath72 , and @xmath18 are shown as full lines in fig .
3 . the corresponding data obtained with the help of the erds
lehner approximation ( [ elf ] ) are indicated by the dashed lines . evidently , the approximation is quite good already for @xmath73 .
the merit of the approximation ( [ elf ] ) lies in the fact that it allows us to derive an analytical expression for the microcanonical low - temperature ground state fluctuations of the ideal bose gas trapped by a one - dimensional harmonic potential @xcite : @xmath74 as already seen in fig . 3
, the low - temperature fluctuations vanish linearly with temperature .
most notably , they do _ not _ depend on the particle number @xmath2 .
this is immediately obvious for temperatures below the `` restriction temperature '' @xmath75 that is defined by the condition @xmath76 , since below @xmath75 the number of microstates becomes independent of @xmath2 , so that `` the condensate has no chance to know how many particles it consists of '' .
however , for @xmath77 this restriction temperature is merely of the order of @xmath78 , since @xmath79 , wheras the @xmath2-independence of @xmath21 actually persists almost up to @xmath80 , where the occupation of the ground state becomes significant @xcite .
this @xmath2-independence of @xmath21 appears to be characteristic for isolated condensates in general , as we shall indicate below .
the analysis for the case @xmath81 proceeds in close analogy to the case @xmath77 .
the restriction temperature , determined from @xmath82 , now reads @xmath83 which means that @xmath84 is about one fourth of the condensation temperature for a gas consisting of @xmath1 particles .
below @xmath84 the number of microstates , and hence all thermodynamical properties of the trapped gas , are independent of the particle number @xmath2 .
accordingly , we write @xmath85 for @xmath25 corresponding to temperatures @xmath9 less than @xmath84 , so that @xmath86 denotes the three - dimensional analogue of @xmath42 .
the generating function for @xmath86 again equals the canonical partition function of a system of infinitely many , distinguishable harmonic oscillators @xcite : @xmath87^{g_j } } \equiv z^{(3)}_{\infty}(\xi ) \ ; . \label{osc}\ ] ] the saddle point inversion of this generating function is straightforward , but a bit tedious .
the result has been given by nanda in 1951 @xcite : @xmath88 \widetilde{n}^{1/4 } + b \right ) \label{nan}\ ] ] with @xmath89 and @xmath90 this is one of the rare examples of a truly asymptotic formula for the number of microstates of a non - trivial bose system . for temperatures above @xmath84 the finiteness of the particle number
restricts the number of microstates , and we have to compute @xmath91 .
it turns out that the logarithm of @xmath86 provides a quite good approximation to @xmath92 even up to the condensation temperature @xcite , so that the entropy of the fictituous boltzmannian oscillator system described by the partition function ( [ osc ] ) approximates the entropy of the trapped bose gas for all temperatures below the onset of condensation .
this means that the distributions @xmath93 introduced in eq .
( [ dis ] ) must be well peaked ; those microstates where the @xmath41 quanta are actually `` spread out '' over all @xmath2 particles carry only a minor statistical weight below @xmath15 .
but still , this finding is of no help for computing the fluctuations @xmath21 , since this computation requires , according to eq .
( [ dis ] ) , to consider auxiliary systems consisting of @xmath94 particles and to determine the numbers @xmath95 .
the problem now is that an immediate analogue to the generating function ( [ res ] ) does not exist , that is , the canonical @xmath2-particle partition function @xmath96 is not known in closed form
. however , there exists the recursion formula @xcite @xmath97 that can numerically be evaluated and inverted within the saddle point approximation , so that we can determine the distributions @xmath93 at least numerically with high precision @xcite .
the results for @xmath98 are shown in fig . 4
: as anticipated , the distributions are well peaked , and moreover remarkably close to gaussians . from the expectation values of these distributions we obtain the microcanonical expectation values of the ground state occupation number , depicted as the full line in fig .
the agreement with the corresponding grand canonical data ( dashed ) is stunning , but in view of the preceding fig .
4 not unexpected : the predictions of the different ensembles will be the same as long as the most probable values of @xmath30 coincide with the expectation values , which is the case as long as the distributions @xmath93 stay symmetrical .
this , in turn , is guaranteed as long as the support of these distributions stays away from the border @xmath99 .
hence , there is a slight difference between the microcanonical and the grand canonical ground state fraction only in the immediate vicinity of the condensation point . for @xmath77
this difference is much more pronounced , since @xmath100 is considerably less well peaked @xcite than its counterpart for @xmath81 .
6 shows the microcanonical fluctuations of the ground state occupation number for @xmath81 and @xmath101 , @xmath102 , and @xmath103 .
remarkably , the low - temperature fluctuations for these three systems agree perfectly . more generally ,
if we compare two trapped bose gases with different @xmath2 under otherwise identical conditions , then their fluctuations @xmath21 are nearly identical for temperatures below the lower of their condensation temperatures .
this finding generalizes the formula ( [ gsf ] ) .
again , the @xmath2-independence of @xmath21 is immediately obvious for temperatures lower than the restriction temperature @xmath84 , but it actually persists almost up to the condensation temperature , as a result of the well - peakedness of the distributions @xmath93 .
we now come back to our initial question : can we rely on the grand canonical ensemble for bose einstein condensation in isolated traps ? as long as we are only interested in the usual thermodynamical quantities ,
the answer is yes .
5 has shown that the grand canonical prediction for the ground state occupation number in a three - dimensional trap is almost indistinguishable from the microcanonical one even if the particle number @xmath2 is as low as @xmath103 .
but in view of the fact that one - dimensional traps are not out of reach @xcite , it should be kept in mind that these differences are larger for @xmath77 @xcite . another comparison between different ensembles is presented in fig . 7 , which depicts the grand canonical ( long dashes ) , canonical ( short dashes ) , and microcanonical ( full line ) specific heat capacity for @xmath81 and @xmath98 .
again , noticeable differences can be found only close to the onset of condensation .
but when it comes to quantum statistical properties , the grand canonical ensemble is unsound @xcite . adopting the microcanonical spirit and starting from the enumeration of microstates ,
we have computed the microcanonical low - temperature fluctuations @xmath21 for harmonically trapped bose gases . while the case @xmath77 could be treated fully analytically , leading to eq .
( [ gsf ] ) , the case @xmath81 had to rely on the numerical evaluation of the recursion relation ( [ rec ] ) .
if one could derive an analytical approximation to the numbers @xmath91 for @xmath41 , with a quality similar to that of the erds
lehner approximation ( [ elf ] ) to @xmath58 , then the case @xmath81 could be brought into the same textbook status that the case @xmath77 already has by now .
it is also worthwhile to point out that very recently navez _
have proposed a new statistical ensemble , within which the ground state particles act as a reservoir , and exchange of particles with the excited - states subsystem without exchange of energy is possible @xcite .
the predictions for the fluctuations @xmath21 obtained with the help of this so - called `` maxwell s demon ensemble '' agree very favourably with our strictly microcanonical results , so that this ensemble might constitute a valuable tool for further studies . in closing , we reiterate the most important result of the present analysis : while the grand canonical ensemble predicts , for ideal bose gases at low temperatures , ground state number fluctuations @xmath21 of the order @xmath2 , and while usual thermodynamical fluctuations scale as @xmath104 , the low - temperature fluctuations @xmath21 for trapped , isolated ideal bose condensates are independent of the total particle number @xmath2 .
anderson , j.r .
ensher , m.r .
matthews , c.e .
wieman , and e.a .
cornell , science * 269 * , 198 ( 1995 ) .
davis , m .- o .
mewes , m.r .
andrews , n.j .
van druten , d.s .
durfee , d.m .
kurn , and w. ketterle , phys
lett . * 75 * , 3969 ( 1995 ) .
bradley , c.a .
sackett , and r.g .
hulet , phys .
* 78 * , 985 ( 1997 ) .
mewes , m.r .
andrews , n.j .
van druten , d.m .
kurn , d.s .
durfee , and w. ketterle , phys .
77 * , 416 ( 1996 ) .
ensher , d.s .
jin , m.r .
matthews , c.e .
wieman , and e.a .
cornell , phys .
lett . * 77 * , 4984 ( 1996 ) .
jin , j.r .
ensher , m.r .
matthews , c.e .
wieman , and e.a .
cornell , phys .
* 77 * , 420 ( 1996 ) .
m .- o . mewes , m.r .
andrews , n.j . van druten , d.m .
kurn , d.s .
durfee , c.g .
townsend , and w. ketterle , phys .
77 * , 988 ( 1996 ) .
jin , m.r .
matthews , j.r .
ensher , c.e .
wieman , and e.a .
cornell , phys .
lett . * 78 * , 764 ( 1997 ) .
mewes , m.r .
andrews , d.m .
kurn , d.s .
durfee , c.g .
townsend , and w. ketterle , phys .
78 * , 582 ( 1997 ) .
myatt , e.a .
burt , r.w .
ghrist , e.a .
cornell , and c.e .
wieman , phys .
* 78 * , 586 ( 1997 ) .
andrews , c.g .
townsend , h .- j .
miesner , d.s .
durfee , d.m .
kurn , and w. ketterle , science * 275 * , 637 ( 1997 ) .
burt , r.w .
ghrist , c.j .
myatt , m.j .
holland , e.a .
cornell , and c.e .
wieman , phys .
lett . * 79 * , 337 ( 1997 ) .
andrews , d.m .
kurn , h .- j .
miesner , d.s .
durfee , c.g .
townsend , s. inouye , and w. ketterle , phys .
rev . lett .
* 79 * , 553 ( 1997 ) .
s. grossmann and m. holthaus phys .
lett . a * 208 * , 188 ( 1995 ) .
s. grossmann and m. holthaus z. naturforsch .
* 50 a * , 921 ( 1995 ) .
w. ketterle and n.j .
van druten , phys .
a * 54 * , 656 ( 1996 ) .
de groot , g.j .
hooyman , and c.a .
ten seldam , proc .
london a * 203 * , 266 ( 1950 ) .
v. bagnato , d.e .
pritchard , and d. kleppner , phys .
a * 35 * , 4354 ( 1987 ) .
k. kirsten and d.j .
toms , phys .
a * 222 * , 148 ( 1996 ) .
k. kirsten and d.j .
toms , phys .
a * 54 * , 4188 ( 1996 ) .
h. haugerud , t. haugset , and f. ravndal , phys .
a * 225 * , 18 ( 1997 ) .
t. haugset , h. haugerud , and j.o .
andersen , phys .
a * 55 * , 2922 ( 1997 ) .
ziff , g.e .
uhlenbeck , and m. kac , phys . rep . *
32 * , 169 ( 1977 ) .
see p. 213 .
landau and e.m .
lifshitz , _ statistical physics _
( pergamon , london , 1959 ) .
pathria , _ statistical mechanics _ ( pergamon , oxford , 1985 ) . h.n.v .
temperley , proc .
london a * 199 * , 361 ( 1949 ) .
nanda , proc .
( india ) * 19 * , 681 ( 1953 ) .
hardy and s. ramanujan , proc .
* 17 * , 75 ( 1918 ) .
bethe , rev .
* 9 * , 69 ( 1937 ) .
see p. 81 .
auluck and d.s .
kothari , proc .
. soc . * 42 * , 272 ( 1946 ) .
p. erds and j. lehner , duke math .
* 8 * , 335 ( 1941 ) .
s. grossmann and m. holthaus , phys .
e * 54 * , 3495 ( 1996 ) .
s. grossmann and m. holthaus , _ fluctuations of the particle number in a trapped bose condensate _ , preprint ( 1997 ) .
nanda , proc .
. soc . * 47 * , 591 ( 1951 ) .
p. borrmann and g. franke , j. chem . phys . *
98 * , 2484 ( 1993 ) .
b. eckhardt , _ eigenvalue statistics in quantum ideal gases _ , preprint ( 1997 ) .
m. wilkens and c. weiss , _ universality classes and particle number fluctuations of trapped ideal bose gases _ , preprint ( 1997 ) .
n.j . van druten and w. ketterle , phys .
lett . * 79 * , 549 ( 1997 ) .
p. navez , d. bitouk , m. gajda , z. idziaszek , and k. rzaewski , _ the fourth statistical ensemble for the bose
einstein condensate _ , preprint ( 1997 ) . | we question the validity of the grand canonical ensemble for the description of bose einstein condensation of small ideal bose gas samples in isolated harmonic traps .
while the ground state fraction and the specific heat capacity can be well approximated with the help of the conventional grand canonical arguments , the calculation of the fluctuation of the number of particles contained in the condensate requires a microcanonical approach . resorting to the theory of restricted partitions of integer numbers , we present analytical and numerical results for such fluctuations in one and three - dimensional traps , and show that their magnitude is essentially independent of the total particle number .
-20 mm |
this paper considers the changes of the entropy @xmath0 and the spin susceptibility @xmath1 through @xmath2 for classic @xmath3-wave type - ii superconductors .
these quantities were calculated by maki@xcite in the dirty limit for superconducting alloys nearly 40 years ago .
however , detailed studies on clean systems are still missing even for @xmath3-wave superconductors . writing these quantities as [ alpha ]
@xmath16 @xmath17 the slopes @xmath7 and @xmath8 will be obtained quantitatively for arbitrary impurity concentrations .
the results near @xmath2 will also be useful for getting an insight into the behaviors over @xmath18 .
indeed , @xmath19 ( @xmath20 ) indicates overall field dependence which is convex downward ( upward ) , as realized from eq .
( [ alpha ] ) .
it seems to have been widely accepted that various physical quantities of classic @xmath3-wave type - ii superconductors follow the linear field dependence with @xmath21 at low temperatures . a theoretical basis for it is the density of states for a single vortex calculated by caroli , de gennes , and matricon.@xcite
however , few quantitative calculations have been carried out so far on the explicit field dependence .
recently , ichioka _
et al_.@xcite performed a numerical study on the density of states of clean two - dimensional @xmath3-wave superconductors with @xmath22 at @xmath23 .
they found the exponent @xmath24 for the overall field dependence of the zero - energy density of states .
also , experiments on the @xmath25-linear specific - heat coefficient @xmath26 for clean v@xmath27si,@xcite nbse@xmath28,@xcite and ceru@xmath28@xcite show marked upward deviations from the linear behavior @xmath29 .
even early experiments on @xmath26 for clean v and nb indicate similar deviations,@xcite although not recognized explicitly in those days due to the absence of a theory on clean systems .
these results indicate that the field dependence with @xmath20 is a general feature of clean @xmath3-wave superconductors , as suggested by ramirez.@xcite following the preceding works on the maki parameters@xcite and the flux - flow resistivity,@xcite which will be referred to as i and ii , respectively , i here present a detailed study on @xmath0 and @xmath1 near @xmath2 at all temperatures .
i thereby hope to clarify the @xmath30 and mean - free - path ( @xmath31 ) dependence of @xmath7 and @xmath8 .
calculations are performed for both two and three dimensional isotropic systems to see the dependence of @xmath7 and @xmath8 on detailed fermi - surface structures .
i also calculate the specific heat jump at @xmath2 for various values of @xmath30 and @xmath31 . to my knowledge , this kind of a systematic study has not been performed even for classic @xmath3-wave superconductors . unlike the convention
, i adopt the average flux density @xmath5 in the bulk as an independent variable instead of the external field @xmath32 . an advantage for it
is that the irrelevant region @xmath33 is automatically removed from the discussion on the field dependence .
this distinction between @xmath5 and @xmath32 becomes important for low-@xmath30 materials where @xmath33 occupies a substantial part of @xmath34 .
any experiment on the @xmath5 dependence should be accompanied by a careful measurement on the magnetization , especially for low-@xmath30 materials like nb and v. section [ sec : formulation ] provides the formulation , sec [ sec : results ] presents numerical results , and sec . [ sec : summary ] summarizes the paper .
i put @xmath35 throughout .
as before,@xcite i consider the @xmath3-wave pairing with an isotropic fermi surface and @xmath3-wave impurity scattering in an external magnetic field @xmath36 .
the formulation proceeds in exactly the same way for both the three dimensional system and the two - dimensional system placed in the @xmath37 plane perpendicular to @xmath38 .
the vector potential in the bulk can be written as@xcite @xmath39 where @xmath5 is the average flux density produced jointly by the external current and the supercurrent inside the sample , and @xmath40 expresses the spatially varying part of the magnetic field satisfying @xmath41 .
i first write down the expressions of the entropy and the magnetization in the presence of pauli paramagnetism . as can be checked directly,@xcite the effect
can be included in the eilenberger equations@xcite for the quasiclassical green s functions @xmath42 , @xmath43 , and @xmath44 by the replacement : @xmath45 where @xmath46 is the matsubara energy and @xmath47 is the bohr magneton .
the corresponding eilenberger functional@xcite for the free - energy difference between the normal and superconducting states is given by @xmath48 \biggr\ } .
\label{f}\end{aligned}\ ] ] here @xmath49 is the pair potential , @xmath50 is the density of states per spin and per unit volume at the fermi level , @xmath51 is the fermi wavevector , and @xmath52 denotes fermi - surface average satisfying @xmath53 .
the quantity @xmath54 is defined by@xcite @xmath55 + \hbar\,\frac{f\langle f^{\dagger}\rangle\!+\!\langle
f\rangle f^{\dagger}}{4\tau } \nonumber \\ & & \hspace{-1 mm } + \hbar\,\frac{g\langle g\rangle\!-\ ! 1}{2\tau } -\hbar\,\frac{f^{\dagger}\,{\bf v}_{\rm f}\!\cdot\ ! { \bm \partial}f -f\,{\bf v}_{\rm f}\!\cdot\ ! { \bm \partial}^{*}f^{\dagger } } { 2[g\!+\!{\rm sgn } ( \varepsilon_{n } ) ] } \ , , \label{i}\end{aligned}\ ] ] where @xmath56 is the relaxation time in the second - born approximation , @xmath57 is the fermi velocity , and @xmath58 denotes @xmath59 the quasiclassical green s functions @xmath42 and @xmath44 are connected by @xmath60 with @xmath61 .
the change of sign in @xmath47 is necessary here , because @xmath62 in the presence of pauli paramagnetism .
the functional derivatives of eq .
( [ f ] ) with respect to @xmath43 , @xmath63 , and @xmath40 lead to the eilenberger equation for @xmath42 , the self - consistency equation for @xmath64 , and the maxwell equation for @xmath40 , respectively .
the expression of the entropy @xmath0 is obtained from eq .
( [ f ] ) by the thermodynamic relation : @xmath65 .
considering the stationarity with respect to @xmath42 , @xmath49 , and @xmath40 , we only have to differentiate with respect to the explicit temperature dependence in @xmath66 .
we thereby obtain @xmath67 \ , , \label{s}\end{aligned}\ ] ] where @xmath68 with @xmath69 the volume of the system .
in contrast , the expression of the external field @xmath32 may be derived by applying the doria - gubernatis - rainer scaling to eq.([f]).@xcite the details are given in appendix a of i , and we obtain @xmath70 where @xmath71 denotes the normal - state magnetization due to pauli paramagnetism .
we thus arrives at the expression of the magnetization from pauli paramagnetism as @xmath72 when pauli paramagnetism is negligible compared with the orbital diamagnetism by supercurrent , we can take the limit @xmath73 in eqs .
( [ s ] ) and ( [ ms ] ) and retain only the leading - order terms .
this results in @xmath74 for eq .
( [ s ] ) . on the other hand , eq .
( [ ms ] ) is transformed by noting eq .
( [ varepsilon_n ] ) into @xmath75 \ , .
\label{ms1}\ ] ] if the zero - field expression @xmath76 is substituted into eq .
( [ ms1 ] ) with @xmath77 , the terms in the square bracket reduces to the yosida function.@xcite i now consider the cases where pauli paramagnetism is small and provide explicit expressions to eqs .
( [ s ] ) and ( [ ms1 ] ) near @xmath2 . from now on i adopt the units used previously@xcite where the energy , the length , and the magnetic field are measured by the zero - temperature energy gap @xmath78 at @xmath79 , the coherence length @xmath80 with @xmath81 the fermi velocity , and @xmath82 with @xmath83 the flux quantum , respectively , with @xmath84 .
first , @xmath42 , @xmath44 , and @xmath40 are expanded up to the second order in @xmath64 as @xmath85 substituting them into eqs .
( [ s ] ) and ( [ ms1 ] ) and using the eilenberger equations for @xmath86 and @xmath87 to remove terms with @xmath88 , we obtain [ s - m2 ] @xmath89 \ , , \label{s2}\end{aligned}\ ] ] @xmath90 further , @xmath64 and @xmath86 near @xmath2 can be expanded in the basis functions @xmath91 of the vortex lattice as@xcite [ expand ] @xmath92 @xmath93 where @xmath94 are the polar angles of @xmath57 with @xmath95 in two dimensions , @xmath96 denotes the landau level , and @xmath97 is an arbitrary chosen magnetic bloch vector characterizing the broken translational symmetry of the flux - line lattice and specifying the core locations.@xcite the coefficients @xmath98 and @xmath99 are both real for the relevant hexagonal lattice . substituting these expressions into eqs .
( [ s2 ] ) and ( [ ms2 ] ) and using the orthonormality of @xmath91 and @xmath100 , we obtain [ s - m3 ] @xmath101 \ , , \label{s3}\end{aligned}\ ] ] @xmath102 except @xmath103 , all the quantities in eqs .
( [ s3 ] ) and ( [ ms3 ] ) are to be evaluated at @xmath2 .
it is possible to give an alternative expression to eq .
( [ s3 ] ) using the equation for @xmath2 given by eq .
( 33 ) of i : @xmath104 = 0 \ , .
\label{hc2}\ ] ] differentiating eq .
( [ hc2 ] ) with respect to @xmath25 yields @xmath105 \ ! = 0 \ , .
\label{dhc2}\ ] ] the quantity @xmath106 has been calculated as eqs .
( 31)-(32 ) of i to be @xmath107 a similar procedure enables us to obtain the expressions of @xmath108 and @xmath109 in eq .
( [ ms3 ] ) as [ f_0n ] @xmath110 @xmath111 where @xmath112 is defined by eq . ( 25 ) of i. using eqs .
( [ dhc2 ] ) and ( [ f_0 ] ) in eq .
( [ s3 ] ) , we obtain @xmath113 with @xmath114 } { \displaystyle \pi t^{2 } \sum_{n=-\infty}^{\infty}\frac{\partial\langle \tilde{f}_{0}^{(1)}\rangle}{\partial h_{c2 } } } \ , . \label{dhc2dt}\ ] ] using eq .
( [ dhc2dt ] ) we can also calculate the specific - heat jump at @xmath2 .
it is given in conventional units as@xcite @xmath115 where @xmath116 is the maki parameter@xcite and @xmath117 .
equations ( [ s - m3 ] ) and ( [ deltac ] ) with eqs .
( [ df_0/db ] ) , ( [ f_0n ] ) , and ( [ dhc2dt ] ) are the main analytic results of the paper .
the quantities @xmath98 , @xmath118 , and @xmath116 have been obtained in i. the explicit expression of @xmath118 is given by @xmath119 where @xmath120 may be calculated efficiently by the procedure in sec .
iif of i , with a change of definition of @xmath121 as @xmath122 now it will be shown that eqs .
( [ s3 ] ) and ( [ ms3 ] ) reduce to an identical expression at @xmath9 for arbitrary impurity concentrations , which has the physical meaning of the zero - energy density of states .
let us start from eq .
( [ s3 ] ) where @xmath124 and @xmath125 yield the same contribution .
using this fact and eq .
( [ f_0 ] ) , it is transformed into @xmath126 . \label{s3a}\ ] ] the summation over @xmath127 for @xmath128 may be performed by using the euler - maclaurin formula and the asymptotic property @xmath129 ( @xmath130).@xcite for example , @xmath131 we thereby obtain [ s - m4 ] @xmath132 equation ( [ ms3 ] ) may be transformed similarly as @xmath133 thus , @xmath134 , or equivalently , @xmath135 at @xmath9 for arbitrary impurity concentrations .
equations ( [ s4 ] ) and ( [ ms4 ] ) have a simple physical meaning .
indeed , noting eq .
( [ fgaexp ] ) , ( [ fexpand ] ) , and eq .
( [ f_0 ] ) , we find an alternative expression at @xmath136 : @xmath137 which is nothing but the normalized density of states at @xmath138 .
thus , we have arrived at a simple result that the entropy and spin susceptibility at @xmath9 are both determined by the zero - energy density of states .
the coefficient of @xmath103 have been obtained in i. also , @xmath139 in eqs .
( [ s4 ] ) and ( [ ms4 ] ) may be calculated efficiently from eq .
( [ f_n ] ) by using the analytic expression:@xcite @xmath140 with @xmath141 .
hence , eqs .
( [ s4 ] ) and ( [ ms4 ] ) at @xmath9 can be evaluated fairly easily .
i here summarize analytic results in the dirty limit @xmath142 .
first , the key quantities @xmath143 are calculated by choosing @xmath144 in the procedure in sec .
iif of i. the results are given by @xmath145 since @xmath146 is of the order of @xmath147 , as shown below , @xmath148 may be approximated as @xmath149 . using this @xmath148 in eq .
( [ f_n ] ) and retaining only the leading - order contributions , we obtain @xmath150 notice that @xmath151 is smaller than @xmath152 by @xmath153 .
substitution of eq .
( [ f_ndirty ] ) into eq .
( [ hc2 ] ) leads to the equation for @xmath2 obtained by maki@xcite and de gennes:@xcite @xmath154 where @xmath155 is the digamma function , and @xmath156 is defined by @xmath157 with @xmath158 dimension of the system . as shown by maki,@xcite eq .
( [ hc2dirty ] ) can be solved near @xmath9 by using the asymptotic expression of @xmath159 as @xmath160\ , .
\label{hc2-dirty-0}\ ] ] thus @xmath161 , as assumed at the beginning .
differentiating eq .
( [ hc2dirty ] ) with respect to @xmath25 , we obtain @xmath162\ , .\ ] ] finally , @xmath116 and @xmath163^{2}$ ] are calculated from eqs .
( 34b ) and ( 36 ) of i as @xmath164 @xmath165 where @xmath166 is defined by @xmath167 with @xmath168 the thermodynamic critical field at @xmath9 .
equation ( [ kappa_2-dirty ] ) agrees with the result by caroli , cyrot , and de gennes.@xcite now , let us substitute eq .
( [ f_ndirty ] ) into eqs .
( [ ms3 ] ) and ( [ s3b ] ) and use eqs .
( [ hc2-dirty-0 ] ) .
we thereby obtain [ s - m - dirty ] @xmath169
@xmath170 thus , @xmath171 and @xmath172 is the same at @xmath9 , in agreement with eq .
( [ ms - dos ] ) ; they are both determined by the zero - energy density of states .
equation ( [ ms - dirty ] ) is the result obtained by maki.@xcite also , the expression @xmath173 for the normalized zero - energy density of states at @xmath9 agrees with the result for the local density of states obtained by de gennes.@xcite equation ( [ delta_0-dirty ] ) tells us that @xmath174 as @xmath128 for @xmath175 .
we hence find from eqs .
( [ alpha ] ) , ( [ s - dirty ] ) , and ( [ ms - dirty ] ) that the initial slopes at @xmath9 for @xmath175 are given by @xmath176 the results suggest the overall field dependence of @xmath0 and @xmath1 at @xmath9 which is convex downward .
notice that the flux - flow resistivity @xmath177 at @xmath9 also has the same initial slope @xmath178 in the dirty limit.@xcite these results strongly suggest that the density of states at @xmath179 is mainly relevant to the physical properties of the vortex state at @xmath9 .
if the @xmath180-wave impurity scattering is relevant , the following additional terms appears on the right - hand side of eq .
( [ i ] ) : @xmath181 where @xmath182 , for example , @xmath183 is the @xmath180-wave relaxation time , and @xmath184 is the unit vector along @xmath51 .
however , eqs .
( [ s ] ) , ( [ ms ] ) , and ( [ ms1 ] ) remain unchanged once @xmath54 is modified as above .
the corresponding calculations near @xmath2 may be performed as described in appendix a of i. it thereby follows that eq .
( [ s - m3 ] ) and ( [ deltac ] ) are also valid together with eqs .
( [ df_0/db ] ) , ( [ f_0 ] ) , and ( [ dhc2dt ] ) , where @xmath99 is now given by @xmath185\tilde{k}^{0}_{n}\,{\rm sgn}(\varepsilon_{n } ) \nonumber \\ & & \hspace{13 mm } + \frac{d}{4\tau_{1}}{\langle \tilde{k}^{0}_{1}\!\sin\theta'\rangle\tilde{k}^{1}_{n}\sin\theta } \biggr\ } \ , , \label{f(1)p1}\end{aligned}\ ] ] with @xmath186\!\ ! \left[1
\!-\!\frac{d}{4\tau_{1}}\langle \tilde{k}_{1}^{1}\!\sin^{2}\!\theta'\rangle\ , { \rm sgn}(\varepsilon_{n } ) \right ] \nonumber \\ & & \hspace{4 mm } + \frac{d}{8\tau\tau_{1}}\langle \tilde{k}_{1}^{0}\!\sin\theta'\rangle^{2 } \ , .
\label{f(1)pd}\end{aligned}\ ] ] in addition , eq .
( [ f_n ] ) is to be replaced by @xmath187 where @xmath188 with @xmath189 \tilde{k}^{1}_{n}\sin\theta\,{\rm sgn}(\varepsilon_{n } ) \nonumber \\ & & \hspace{13 mm } + \frac{1}{2\tau}\langle \tilde{k}^{0}_{1}\!\sin\theta'\rangle\tilde{k}^{0}_{n } \biggr\ } \ , .
\label{f(1)p2}\end{aligned}\ ] ] finally , the analytic results in the dirty limit are the same as those given in sec .
iid with a replacement of @xmath56 by the transport life time @xmath190 defined through @xmath191 i have adopted the same parameters as i and ii to express different impurity concentrations : @xmath192 numerical calculations of eqs .
( [ s - m3 ] ) and ( [ deltac ] ) with eqs .
( [ df_0/db ] ) , ( [ f_0n ] ) , and ( [ dhc2dt ] ) have been performed for each set of parameters by restricting every summation over the matsubara frequencies for those satisfying @xmath193 .
choosing @xmath194 is sufficient to obtain an accuracy of @xmath195 for eqs .
( [ ms3 ] ) and ( [ deltac ] ) , whereas @xmath196 ( @xmath197 ) is required for eq.([s3 ] ) in the dirty ( clean ) limit .
summations over landau levels have been truncated at @xmath198 where i put @xmath199 in the calculation of @xmath120 ; see sec .
iif of i for the details .
enough convergence has been obtained by choosing @xmath200 , @xmath201 , @xmath202 , @xmath203 , @xmath204 , and @xmath197 for @xmath205 , @xmath206 , @xmath207 , @xmath208 , and @xmath209 , respectively .
finally , integrations over @xmath210 have been performed by simpson s formula with @xmath211 integration points for @xmath212 .
figures [ fig:1 ] and [ fig:2 ] show temperature dependence of @xmath7 and @xmath8 defined by eqs .
( [ alphas ] ) and ( [ alphac ] ) , respectively , for different impurity concentrations parametrized by eq .
( [ parameters ] ) .
they have been calculated in three dimensions for @xmath213 and @xmath214 .
all the curves start from the same value @xmath215 at @xmath216 and develop differences among different impurity concentrations at lower temperatures .
the equality @xmath135 holds at @xmath9 , as shown by eq .
( [ ms - dos ] ) , and the value decreases from @xmath10 in the dirty limit to around @xmath217 for @xmath218 . according to eq .
( [ ms - dos ] ) , this variation in the slope at @xmath9 can be attributed to the dependence of the zero - energy density of states @xmath219 upon the impurity concentration . in particular , @xmath219 in the dirty ( clean ) limit decreases more rapidly ( mildly ) than the linear behavior @xmath220 near @xmath2 . from this result
, we expect the overall field dependence of the entropy and spin susceptibility at @xmath9 which is convex downward ( upward ) in the dirty ( clean ) limit , as realized from eq .
( [ alpha ] ) . as a function of @xmath221 for different impurity concentrations with @xmath222 , @xmath213 , and @xmath214 . ] as a function of @xmath221 for different impurity concentrations with @xmath222 , @xmath213 , and @xmath214 . ]
the difference between @xmath7 and @xmath8 at finite temperatures is rather small , as expected from @xmath135 holding at @xmath9 and @xmath223 .
in particular , the curves of @xmath7 and @xmath8 in the dirty limit depend neither on the dimensions nor @xmath224 .
however , the dependence develop gradually as the mean free path becomes longer . as a function of @xmath225 for @xmath158 , @xmath226 , and @xmath214 . ]
figure [ fig:3 ] shows the slope @xmath227 at @xmath9 as a function of @xmath225 for different combinations of dimensions and impurity scatterings .
the four curves start from the same value @xmath10 in the dirty limit , and decreases gradually through unity towards @xmath11 in the clean limit .
however , we observe only small dependence of @xmath228 on @xmath229 and @xmath224 .
we thus realize that the zero - energy density of states is mainly determined by the mean free path , and does not depend much on the dimensions nor the details of the impurity scattering .
it is interesting to note that the slope @xmath230 for the flux - flow resistivity @xmath177 , which was calculated previously@xcite for @xmath158 , @xmath213 , and @xmath214 , show a complete numerical agreement at @xmath9 with the corresponding @xmath7 and @xmath8 , i.e. , @xmath231 for arbitrary impurity concentrations at @xmath9 .
this fact indicates that @xmath177 at @xmath9 is also determined by the zero - energy density of states .
next , we examine the dependence of the slopes on the ginzburg - landau parameter @xmath232 .
figure [ fig:4 ] shows the same curves as fig .
[ fig:1 ] near the type - i - type - ii boundary of @xmath233 .
each curve is shifted upwards from the corresponding one in fig .
[ fig:1 ] for @xmath214 , but the quantitative difference is rather small .
this is also the case for @xmath8 .
thus , the slopes @xmath7 and @xmath8 as a function of @xmath5 do not have large @xmath232 dependence . as a function of @xmath221 for different impurity concentrations with @xmath222 , @xmath213 , and @xmath233 .
] finally , fig .
[ fig:5 ] plots the specific - heat jump @xmath234 over @xmath25 at @xmath2 as a function of @xmath221 for different impurity concentrations with @xmath222 , @xmath235 , and @xmath214 .
it is normalized by the corresponding quantity at @xmath216 and @xmath79 , i.e. , @xmath236 in the weak - coupling model .
the curves change gradually from almost @xmath25-linear overall temperature dependence in the dirty limit to @xmath237 dependence in the clean limit , and approach zero as @xmath238 at lowest temperatures.@xcite although the ratio near @xmath223 is strongly dependent on @xmath232 as@xcite @xmath239 the basic features pointed above are common among different values of @xmath232 , @xmath158 , and @xmath240 . at @xmath2 ,
normalized by the corresponding quantity at @xmath216 and @xmath79 , as a function of @xmath221 with @xmath222 , @xmath235 , and @xmath214 .
the curves are for @xmath241 , @xmath242 , @xmath243 , @xmath244 , and @xmath208 from the top to the bottom . ]
the entropy and the spin susceptibility near @xmath2 have been calculated for @xmath3-wave type - ii superconductors over all impurity concentrations .
the results have been expressed conveniently using the initial slopes @xmath7 and @xmath8 defined by eq .
( [ alpha ] ) .
the main conclusions are summarized as follows : ( i ) @xmath135 holds both at @xmath9 and @xmath216 .
( ii ) @xmath215 at @xmath216 for all impurity concentrations .
( iii ) at @xmath9 , the slope @xmath245 decreases from @xmath10 in the dirty limit to @xmath11 in the clean limit .
this change is due completely to the mean - free - path dependence of the zero - energy density of states .
the fact also suggests variation of the overall field dependence at @xmath9 from convex downward in the dirty limit to upward in the clean limit .
( iv ) the slopes have only small dependence on the dimensions and the details of the impurity scattering .
( v ) the slope @xmath230 of the flux - flow resistivity @xmath177 , which was calculated previously,@xcite also shows a complete numerical agreement at @xmath9 with @xmath7 and @xmath8 .
this fact indicates that the zero - energy density of states is also responsible for @xmath177 at @xmath9 .
the @xmath25-linear specific - heat coefficient @xmath246 observed in clean materials@xcite show curves of @xmath20 , which are in a qualitative agreement with the present calculation . on the other hand ,
@xmath246 for dirty samples@xcite follows well - accepted linear field dependence @xmath15 , which is apparently in contradiction with the present result in the dirty limit .
it should be noted however that a careful experiment@xcite on @xmath177 shows field dependence near @xmath9 which is convex downward , and experimentally obtained @xmath230 agrees quantitatively with the dirty - limit theory.@xcite detailed experiments on the mean - free - path dependence of @xmath246 and @xmath247 are desired to remove these discrepancies . | a theoretical study is performed on the entropy @xmath0 and the spin susceptibility @xmath1 near the upper critical field @xmath2 of @xmath3-wave type - ii superconductors with arbitrary impurity concentrations .
the changes of these quantities through @xmath2 may be expressed as @xmath4/[s_{\rm n}(t)\!-\!s_{\rm s}(t,0)]\!=\!1\!-\!\alpha_{s}(1\!-\!b / h_{c2})\!\approx\!(b / h_{c2})^{\alpha_{s}}$ ] , for example , where @xmath5 is the average flux density and @xmath6 denotes entropy in the normal state . it is found that the slopes @xmath7 and @xmath8 at @xmath9 are identical , connected directly with the zero - energy density of states , and vary from @xmath10 in the dirty limit to @xmath11 in the clean limit .
this mean - free - path dependence of @xmath7 and @xmath8 at @xmath9 is quantitatively the same as that of the slope @xmath12 for the flux - flow resistivity studied previously .
the result suggests that @xmath13 and @xmath14 near @xmath9 are convex downward ( upward ) in the dirty ( clean ) limit , deviating substantially from the linear behavior @xmath15 .
the specific - heat jump at @xmath2 also shows fairly large mean - free - path dependence . |
the physical properties of molecular clouds the birthplace of stars are of vital importance , since they reveal clues as to why stars are formed in certain regions , how many stars can form , how star formation proceeds and how it affects the next cycle of stellar production ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
therefore , the method we use to determine these cloud properties is a critical choice .
traditionally , the best quality data of molecular clouds has come from our own galaxy , where the small distances involved allow galactic molecular clouds with sizes down to @xmath0 10 pc to be resolved even with single dish telescopes .
these observations have offered the chance to study molecular clouds in some detail , with previous studies revealing our galactic molecular cloud population has a mass , size and velocity dispersion of 10@xmath1 m@xmath2 , 5 70 pc , and 2 10 km s@xmath3 , respectively ( e.g. , * ? ? ? * ; * ? ? ?
* ; * ? ? ?
despite the quality of this data , it provides information only on one type of galaxy , and the milky way is relatively quiescent .
studies of external galaxies are therefore required before we can reveal the whole picture of the relation between molecular clouds and star formation . of course
, the drawback of extragalactic studies is the large distances that limit the resolution .
however , observations of molecular clouds in these galaxies are now approaching the quality of that in the milky way , thanks to instruments and techniques such as the long - baseline interferometry employed with alma .
alma is designed to detect spectral lines such as carbon monoxide ( co ) which traces molecular clouds in a normal galaxy like the milky way at a redshift @xmath4 in less than 24 hours .
this rate means that the time required to map molecular clouds in a nearby galaxy will shrink to only a few hours . with the highest expected angular resolution of low transition co lines being about 0.04@xmath5 ,
a galaxy @xmath6 16 mpc away in the virgo cluster can be mapped with a physical resolution of @xmath64pc .
the highest velocity resolution of low transition co lines at 0.01 km s@xmath3 , so assuming the molecular clouds in nearby galaxies have comparable properties to those in the milky way , these resolutions should be sufficient to resolve these populations .
note that this assumption seems to be valid based on the current surveys of extragalactic clouds such as paws @xcite and canon @xcite .
these observations suggest that extragalactic and galactic molecular clouds share similar larson s scaling relation of cloud properties . with the spatial and velocity resolution of observations
about to resolve a wide population of molecular clouds , comparisons between observations and simulations become a key tool in understanding the results .
however , this can be challenging since the two techniques are based on different data structures
. simulations normally have data with three dimensional position ( @xmath7 , @xmath8 , @xmath9 ) and velocity ( @xmath10 , @xmath11 , @xmath12 ) coordinates , from which they directly calculate the physical properties of the molecular clouds such as mass , radius and volume density
. meanwhile observations have to take their data from the projected image on the sky .
this provides only two spatial dimensions ( ra and dec ) and typically a single velocity along the line - of - sight .
the resultant properties of the clouds are therefore computed in three independent dimensions , rather than six .
an additional issue is the choice for the molecular cloud boundary .
clouds are typically selected based on their density ( in the case of simulations ) or intensity ( from observations ) . yet
molecular clouds are not isolated entities and therefore exactly where their edges lie is not obvious .
simulators navigate this issue by selecting a fixed threshold for the cloud edge , while observers consider the ratio of the intensity to the rms noise . even within a single technique ,
there is no consensus for what such thresholds should be .
the problem is magnified when the molecular region is extended .
should a continuous contour mark a single cloud or should this be divided into separate peaks ? a major reason for this degeneracy is the absence of a ` correct ' answer .
molecular clouds are part of the interstellar medium ( ism ) ; a continuous blend of pressures and densities ( e.g. * ? ? ? * ) .
there is therefore no obvious edge to the clouds , which are typically thought to be borderline gravitationally bound at best @xcite .
likewise , it is not clear which properties should be used to identify the cloud .
simulators typically only use position , taking advantage of their three spatial dimensions . yet
, this excludes the velocity information which could identify two sections of an object moving together .
conversely , observers are plagued by projection effects which can result in well separated bodies being treated as a single object or regions of the same cloud being artifically split due to the internal motions of the cloud .
the combined result of different data structures and varying choices for cloud identification means that conclusions are being drawn about the star formation conditions in galaxies from results that are not considering the same objects .
previous work has tried to assess the extent on this problem on smaller scales , comparing star - forming regions found in synthetic observations of a single simulated molecular cloud with those identified using the full position data . @xcite
first noticed that the synthetic observed gmcs in their 3d simulations include spatially - unconnected structures due to the projection effect .
when examining dense clumps found in two- and three- dimensional position data , recently @xcite found that projection effects could overestimate the mass of the clumps within the cloud and falsely identify more diffuse regions as clumps .
the result was a shift by a factor of three in the clump mass function ( cmf ) , potentially explaining the difference between the observed cmf and the lower stellar initial mass function ( imf ) ( e.g. , * ? ? ?
@xcite did note that the addition of the velocity in their two dimensional data improved the match with the three dimensional clump properties compared to when the two dimensional data was used alone .
@xcite further considered the differences in the observed chemistry and gravity of the clump properties by matching clumps in the synthetic observations and three dimensional spatial data if they originate from the same density structure .
they concluded that the scatter in the clump properties causes significant uncertainty in the virial parameter ( estimating gravitational boundness ) of a cloud such that it is difficult to connect this with star formation .
in this paper , we present a comparison of the physical properties of the molecular clouds identified in three spatial co - ordinates ( position - position - position space or ppp ) and those in two projected spatial co - ordinates and a single line - of - sight velocity co - ordinate ( position - position - velocity space or ppv ) in the same simulated galaxy .
the goal is to assess whether the same objects can be identified in these two methods and if the properties are consistent .
this work is on a larger - scale than the previously mentioned studies , comparing the properties of a global cloud populations which extragalactic observations are starting to capture .
the simulated galaxy was modelled on the barred spiral ( sabc ) galaxy , m83 , using observational data from the 2mass k - band image @xcite to estimate the stellar potential .
the simulation was run using the three dimensional adaptive mesh refinement ( amr ) hydrodynamics code , ` enzo ` @xcite and is presented in @xcite , along with a full description of the run parameters .
the gas radiatively cooled down to 300k ( a limit designed to allow for the lack of small - scale pressure from unresolved turbulence or magnetic fields ) but no star formation or feedback was included . a projected face - on image of the gas density in the simulated galaxy
is shown in figure [ fig_sim_galaxies_full](a ) .
the galaxy s bar and two spiral arms are visible , with the arms extending to a radius of @xmath13kpc and the bar - end at 2.3kpc .
m83 is at a distance of 4.5mpc , corresponding to a resolution of 1@xmath5 @xmath14 20pc @xcite . in the simulation , the smallest cell size is @xmath15pc . in @xcite ,
the molecular clouds are identified in ppp space using two different methods .
the main cloud identification method uses a continuous contour at a density of @xmath16@xmath17 . using this technique , they found gmcs came in three different types : _ type a _ clouds were the most commn gmcs , dominating the cloud populations in all regions and having properties that agreed with typical observations .
_ type b _ clouds are massive giant molecular associations ( gmas ) created through multiple cloud interactions , and _ type c _ clouds are unbound , transient clouds , often found in tidal tails .
the second method divided clouds within this contour if there are two peaks separated by at least 20pc ; a technique used in @xcite for cloud identification .
we compare these two methods with clouds identified in a ppv synthetic observational data set of the same galaxy ( also known as a spectral line data cube ) .
this data set type is commonly used in millimetre to submillimetre observations , which trace the gas in the ism . the line - of - sight direction which defines the velocity co - ordinate is along the @xmath18axis of the galaxy , observing the simulation face - on .
the ppv data cube was created with a spatial and velocity spacing of 2pc ( slightly larger than the simulation cell size ) and 1km / s and included a gaussian thermal broadening with a width equal to the velocity dispersion of the gas .
when identifying the clouds in the ppv data set , we assume the galaxy is observed in @xmath19co ( 10 ) .
@xmath19co ( 10 ) is excited in low density molecular gas , making it is an ideal tracer of the entire molecular cloud .
it also has a critical excitation density of @xmath20@xmath17 , allowing us to use a cloud identification threshold of @xmath21 @xmath22 100 @xmath17 , in keeping with the ppp schemes . for the ppv cube ,
only cells with a density greater than the threshold value were included in the data .
such a cut - off produces a projected image shown in figure [ fig_sim_galaxies_full](b ) .
( a ) ( b ) to assess how close our simulated clouds are to those observed in the galaxy , we plotted the ppv column density distribution and overlaid the typical values found for the galactic clouds in figure [ fig_column_density ] .
the grey histogram shows the ppv distribution , with the typical column density marked by the black line .
the galactic star - forming molecular clouds ( color lines include two galactic surveys ( yellow and cyan lines , * ? ? ?
* ; * ? ? ?
* ) and three individual clouds : orion @xcite , perseus @xcite and taurus @xcite ( red lines ) , all estimated by co observations . the median column density for our ppv data ( @xmath23 = 21.7 @xmath24 ) is consistent to the typical values of the galactic clouds , where @xmath23 commonly ranges between 21.0 22.5 @xmath24 .
this suggests that our results should apply well to genine observations .
this paper is organized as follows : in section [ sec_cloud_identification ] , we introduce four different cloud identification methods using ppv and ppp , and present two of these as the main methods in this work . section [ sec_results ] presents the comparison of the global properties between ppv- and ppp - clouds , as well as the one - to - one match cloud properties between two sets .
classification of molecular clouds based on their physical properties is also shown in section [ sec_results ] . in section [ sec_diff_methods ] , we show a brief results of the two minor cloud identification methods .
summary of this work is given in section [ sec_summary ] .
overlaid are the typical column densities of the galactic clouds .
median column density of our ppv gas is @xmath25(h@xmath26 ) @xmath22 21.7 @xmath24 ( black line ) .
typical column densities that @xcite and @xcite suggest from their galactic surveys are plotted with yellow and cyan solid lines at @xmath6 21.9 and @xmath6 21.7 @xmath24 , respectively
. typical column density of the orion molecular cloud is @xmath6 22.3 @xmath24 ( red dashed line , * ? ? ?
clouds perseus and taurus , have similar column density of @xmath6 21.3 @xmath24 ( red dotted line , * ? ? ? * ; * ? ? ?
all the galactic values are measured with co observations.,scaledwidth=40.0% ]
for both our data sets in ppp and ppv , we used two cloud identification methods .
the first method identifies continuous structures of gas above our chosen density or flux threshold , which we refer to as _ islands_. the second method can further segregate these bodies into smaller objects such that each contains a single local maxima .
we refer to this as the _ decomposition _ method .
ppp clouds are identified in the original data from the simulation , while clouds are found in ppv space by first creating a ppv data cube from the simulation data and passing this in fits format to the cprops package @xcite .
cprops expects the data to be emission intensity , rather than the gas density followed by simulations . to convert between the two , we use a co - to - h@xmath27 conversation factor @xcite to change the physical unit ( g @xmath24 ) to observed @xmath19co ( 10 ) flux ( kelvin ) . since we do not consider chemistry ( to compute the exact co abundance ) nor radiative transfer and the excitation of molecular lines , this value is for an ideal circumstance
this also means that the conversion factor cancels when we derive the cloud mass , so its precise value does not affect our results . since the flux threshold in cprops
is determined by the ratio to the noise level ( @xmath28 ) , the so - called signal - to - noise ratio ( s / n ) , a low random noise with level of @xmath6 1 k is added to the ppv cube .
this sensitivity is sufficient to identify the smallest identified galactic molecular clouds at about 5 @xmath28 level with our spatial ( @xmath6 2 pc ) and velocity resolution ( @xmath6 1 km s@xmath3 ) .
these resolutions and sensitivity are higher than current alma observations , but technically achievable using the full - operation of alma and its long - baseline ability . while there is no doubt that resolution and sensitivity play a strong role in determining gmc properties , they are downstream of this work . instead , we focus on differences due to the techniques themselves in the absence of external factors . the cloud boundaries in the island method are defined slightly differently in the two techniques .
ppp works from the lowest density , drawing a contour at @xmath29@xmath17 and defining all cells within a closed section as the cloud .
ppv via cprops begins by masking the emission with a high s / n , picking out the cloud locations at densities much higher than the background .
it then extends this mask to the user defined lowest s / n , which outlines the observed cloud boundary .
cprops then assumes that the real cloud boundary is larger than the observed cloud boundary , since the cloud outer regions are being obscured by the background noise .
it therefore extrapolates from the observed boundary to a sensitivity of 0 k to form the real cloud boundary . in the decomposition method ,
the cloud boundaries are identified in a similar manner between ppp and ppv .
ppp - clouds are located by searching for local maxima and the adjoining surrounding cells that are above the threshold @xmath30@xmath17 , assigning these to a single cloud . in the ppp data ,
maxima closer than 20 pc arere merged into a single object .
ppv via the cprops decomposition method initially uses the island method to find the continuous structures and then searches these for separate peaks . for computational speed
, the search for island peaks is performed within a cube with a ( user - defined ) side of 22pc @xmath31 22pc @xmath31 7 km s@xmath3 .
this size was selected to be comparable to the average gmc diameter and velocity dispersion in our simulation , ensuring that a single cloud would not be accidentally divided by our numerical choice .
if multiple peaks are found within an island , cprops finds the lowest contour that surrounds both peaks and then separates them with a contour twice the rms noise above the shared boundary ( the default value suggested by @xcite and @xcite ) .
all pixels within the separated contours are assigned to the peaks .
cprops can potentially merge peaks within an island if their properties are sufficiently similar , but the high sensitivity of our analysis meant that this was not necessary @xcite .
the physical properties of the clouds are then derived once the cloud boundaries are set .
these derivations are not identical between the two methods , since the assumed raw data ( simulation or observation ) is measuring different quantities
. we do not attempt to correct for this , but adopt the original calculations that are used to describe clouds in simulation and observational studies as part of the comparison between the two methods . in both methods and data structures ,
cloud mass ( @xmath32 and @xmath33 ) are calculated by the sum of the cells or pixels in the clouds .
we note that a co - to - h@xmath26 conversion factor is required for the ppv data when moving between the observed flux and physical gas quantity .
a conversion factor of 2 @xmath31 10@xmath34 @xmath24 ( k km s@xmath3)@xmath3 is adopted in this work , which is the default value of cprops . however ,
this is used twice in the derivation of the cloud mass and cancels itself out due to the lack to chemistry , radiative transfer and excitation of molecular line in this work .
we confirmed this by adopting a conversation factor differing by a factor of ten and recalculating the mass to achieve the same result.the mass in both methods is therefore the total mass of gas in cells within our clouds . for the cloud radii and velocity dispersion properties ,
the ppp and ppv calculations differ more significantly .
the ppp data is calculated from three spatial and three velocity dimensions .
it therefore measures the average radius ( @xmath35 ) of cloud from its three projected axes , @xmath36 , @xmath37 , and @xmath38 , and computes the mass - weighted one - dimensional velocity dispersion from the average deviations between the gas velocity and the cloud s bulk velocity velocity in @xmath7 , @xmath8 , and @xmath9 directions .
in contrast to this , ppv must measure the properties projected along a single direction .
to do this , cprops first calculates the geometric mean of the square root of the spatial mass(flux)-weighted second centralized moment .
the root - mean - square ( rms ) ( standard deviation ) @xmath39 is the square root of the variance .
] of the intensity distribution along the major and minor axes of the projected cloud boundary .
quantitively , these are defined for each cloud as : @xmath40 where the @xmath41 and @xmath42 are the rms size of the intensity distribution along the @xmath7 and @xmath8 axis , @xmath43 is the observed flux of the @xmath44th pixel within the cloud , @xmath45 and @xmath46 are the position in the two spatial dimensions of the @xmath44th pixel and @xmath47 and @xmath48 are the flux - weighted mean positions of the cloud .
the rms size of the cloud ( @xmath49 ) is the geometric mean of @xmath41 and @xmath42 , @xmath50 .
the rms size is then converted into the effective radius ( @xmath51 ) by assuming a mass - centered density profile of a spherical cloud .
this is to allow for the fact that the true cloud edge extends beyond the identified boundary of the cloud .
the relation is simplified to @xmath52 , where the constant is an empirical correction derived from the resolved galactic gmcs in @xmath19co ( 10 ) observations @xcite .
note that unlike the ppp radii , the @xmath51 value is weighted by the flux . in a similar fashion , the velocity dispersion ( @xmath53 ) for the ppv clouds is the flux - weighted rms velocity dispersion within the defined cloud boundary : @xmath54 where @xmath55 is the velocity of the @xmath44th pixel within the cloud , @xmath56 is the flux - weighted mean velocity of the cloud . for both the ppp and ppv methods , the mass surface density ( @xmath57 and @xmath58 ) , virial parameter ( @xmath59 and @xmath60 ) and virial mass ( @xmath61 and @xmath62 )
are derived from the basic properties ( cloud mass , radius and velocity dispersion ) in the standard way : @xmath63 is defined as the mass per unit area and is therefore simply calculated from the cloud mass and radius .
@xmath64 measures the gravitational binding of a cloud , assume a spherical profile and no magnetic support or pressure confinement .
specifically , @xmath65 indicates that the cloud is gravitationally unbound while @xmath66 suggests a bound system .
it is defined from the mass , radius and velocity dispersion of the cloud as @xmath67 .
the virial mass comes from this parameter , @xmath68 .
a visual comparison of the clouds found in the ppp and ppv data sets for the island method is shown in figure [ fig_position_examples ] .
the panels each show a different region of the galactic disc , with the projected gas density overlaid with red circles for the ppv clouds and green plus signs for the ppp clouds .
the location of each symbol marks the cloud centre - of - mass , but the symbol size is not proportional to the cloud properties .
panel ( a ) shows the inner disc region , around 1.2kpc from the galactic centre and within the grand design bar shown in figure [ fig_sim_galaxies_full ] . due to the high number density of clouds in this region , this image size is @xmath69kpc , whereas panels ( b ) to ( d ) show a region of @xmath70kpc .
these three panels step outwards from the disc , with panel ( b ) showing a region at 5.7kpc from the galactic centre , panel ( c ) at 6.8kpc from the centre and panel ( d ) at 7.7kpc from the centre and outside the spiral arms .
notably , each ppp cloud typically has only one counterpart in ppv , even in the most crowded panel in figure [ fig_position_examples](a ) .
however , this is not universally true .
each panel of figure [ fig_position_examples ] displays examples of where either the ppp or ppv method has identified multiple clouds where the counter scheme finds a single object .
a unique ppv cloud may have multiple ppp counterparts that are physically close with similar velocities , causing them to merge in the ppv projected data space .
conversely , a ppp cloud may have multiple ppv counterparts if the level of the noise defining the cloud edge in ppv is higher ( or within a factor of two ) than the gas density between peaks within a single ppp cloud , causing the structure to split .
for the gas to be detected in both ppp and ppv , it must be higher than 100@xmath17 and @xmath71g@xmath24 in projection .
occassionally , this causes a ppp cloud to have no ppv counterpart at all , as shown in the top of panel ( a ) , where the small cloud diameter gave a projected density below the noise limit .
these splits may also lead to the centre - of - mass of the clouds in a region not overlapping , producing an off - set as seen in the lower left corner of panel ( c ) .
this agreement in cloud locations also extends to the cloud properties .
table [ tab_phy_prop_gal_str ] shows a comparison between the median properties of the clouds found by the ppv and ppp island methods .
the top row shows the properties for the entire cloud population , while the next three rows takes an separate look at the bar , spiral and disc populations ( see next section ) . in all areas ,
ppp finds slightly more clouds than in the ppv case , although the numbers differ by less than 10% .
this indicates that the effect of merging ppp clouds due to projection effects is slightly more pronounced than splitting ppp clouds due to noise .
both ppp and ppv clouds have masses around @xmath72@xmath73 and radii around 15pc .
the cloud surface density peaks at 525m@xmath74pc@xmath75 , with velocity dispersion around 4 - 5 km s@xmath3 and the clouds are found to be borderline gravitationally bound .
( a ) ( b ) ( c ) ( d ) .median physical properties of gmcs . [ cols="<,<,<,<,<,<",options="header " , ] the scaling relations for the decomposition method are shown in figure [ fig_pp2_prop_property_decomp ] . top panels ( a ) , ( b ) and ( c ) show the ppp cloud data set for the two larson relations and that of the virial parameter with radius .
the bottom three panels ( d ) , ( e ) and ( f ) show the same relationships for the ppv data . the symbols and colours for each cloud type match those in figure [ fig_pp2_prop_property ] . in general
, the trends in the ppp clouds are similar to that with the island cloud identification method , but the range of values is reduced due to the splitting of larger structures .
the boundary between _ type a _ and _ type c _ clouds in the mass radius relationship in panel ( a ) is no longer seen as a gap , although two different trends appear to be visible .
it vanishes entirely in the velocity radius relation in panel ( b ) .
the _ type c _ clouds remain the most unbound object in the simulation , with virial parameters @xmath76 .
the spread in the virial parameter value no longer depends greatly on radius , due to the more uniform cloud properties producing a much smaller spread in radius values . for clouds with virial parameter less than 10 , the _ type c _ and _ type a _
clouds overlap more strongly than in the island method , again due to the smaller range in cloud properties . in the ppv clouds using the decomposition method ,
the three relations are lost , with the radius , mass and velocity dispersion no longer well correlated . in the mass - radius relation in panel ( d ) , the _ type c _ clouds do show the same general trend at the ppp clouds and island method , but with a much wider scatter . by contrast , the _ type a _ clouds suggest a reverse trend , where the mass anti - correlates with radius .
this huge scatter is due to the watershed effect . in the cases were multiple peaks are within a massive island , all the peaks can share a high density envelope .
this causes a large amount of material to be discarded , often up to @xmath77% of the emission 10@xmath78 m@xmath2 in ppv , and 1.8 @xmath31 10@xmath79 m@xmath2 in ppp . for comparison ,
the total mass of clouds in the island method is 1.6 @xmath31 10@xmath79 m@xmath2 in ppv , and 1.8 @xmath31 10@xmath79 m@xmath2 in ppp . ] .
the result is small contours centered around a high density region , producing the trend in the _ type a _ clouds .
this effect is most notable in the splitting of the massive _ type b _ clouds , which often contain concetrated cores of material that become small objects .
extra scatter is also created from the mass - weighting of cprops on the radius , as discussed in previous sections . for the smaller _ type c
_ clouds , the watershed effect simply increases radius scatter as clouds are divided .
note that a significant number of _ type c _ clouds are split from the remains of the _ type b _ island method clouds , producing objects both larger and smaller than the original _ type c _ clouds .
the splitting of islands into multiple peaks also reduces the velocity dispersion of the clouds for ppv , as shown in panel ( e ) .
as only the partial structure ends up attacked to a peak , the velocity dispersion drops compared to the ppp case , where the extended envelope is not neglected . with the scatter in the radius , the _ type a _ and _ type c _ clouds strongly overlap and there is no visible trend with velocity dispersion . where the virial parameter is plotted against radius in panel ( f ) ,
the _ type a _ and _ type c _ clouds do occupy different region . _
type a _ clouds have a virial parameter of @xmath80 ( mostly @xmath81 ) , while the _ c _ remain the most unbound objects with @xmath82 .
this difference is actually more marked in the ppv data than the ppp , due to the reduced radius and velocity dispersion of the _ type a _ clouds which lowers the virial parameter .
_ type c _ clouds occupy approximately the same plot region in both ppv and ppp , but with a wider scatter in the radius as seen before .
exactly which method island or decomposition is more physical is up for debate .
there is no reason why a cloud should contain a single peak and indeed , star formation is unlikely to be centred in only one location of a turbulent cloud .
the use of the methods may depend on the quality of the observational data and the purpose of the study . for the extragalactic observations with typical resolution of @xmath83 20 pc ,
most studies adopt the decomposition method to select objects .
however , the resolution at these distances means it is hard to discern internal cloud structure , producing clouds that are similar in size to the island method in this work .
on the other hand , inspite of the higher resolutions ( @xmath84 20 pc ) , most of galactic studies continue to adopt the decomposition method .
this is due to the large dynamic range in the distance to the clouds , varying the resolution within a cloud s boundary .
both these choices originate from the lack of a clear definition for a gmc , leaving is ambiguous as to whether it is resolved .
our study emphasises the importance of this cloud definition and for high resolution data , the island method is the stronger choice for selecting similar objects in both simulation and observation data types ; a valuable asset in understanding the evolution of star - forming clouds . ( a ) ( b ) ( c ) ( a ) ( b ) ( c ) ( d ) ( e ) ( f )
we are rapidly reaching the point where simulations and observations will achieve comparable resolution of molecular cloud populations in a wide variety of galaxies .
this opens the door to truly constrain the mechanisms for cloud formation and thereby star formation in different galaxy environments .
the importance of this advance makes it imperative to examine carefully if the two techniques of observation and simulation are discussing the same star - forming molecular clouds .
after all , observational data is done in two spatial co - ordinates and one velocity co - ordinate , while simulation data typically uses all three spatial dimensions , but ignores the velocity components .
a direct match is not possible . in this paper , we compared the physical properties of giant molecular clouds formed in a simulation of a barred spiral galaxy using simulation and observational identification methods .
the two methods selected clouds from the data in position - position - position ( ppp ) space typical for simulation and position - position - velocity ( ppv ) space used in observations . the ppv data cube was assumed to be the product of @xmath19co ( 10 ) observations with optimal resolution and sensitivity .
the properties of the clouds found in both methods were compared and the clouds themselves matched to assess whether the two methods identified the same objects and if the properties were method dependent .
this process was repeated twice ; once where the clouds were identified as continuous structures within a contour ( `` island method '' ) and once where each peak in the density or emission was assigned to a seperate cloud ( `` decomposition method '' ) . 1 .
the total number of clouds identified was similar in the ppv and ppp methods ( 971 versus 1029 ) .
this was also true when comparing cloud numbers in each galactic environment of bar , spiral and disc .
the typical ( median ) cloud properties , such as their mass , radius and velocity dispersion , differ by @xmath85% between ppv and ppp .
the bimodal mass surface density distribution of ppp clouds ( peaks at @xmath6 100 and @xmath6 1000 m@xmath2 pc@xmath75 ) which had been found by @xcite is reproduced in the ppv clouds , with the boundary between the two populations being roughly consistent . 1 .
about 70% of clouds have single counterpart in both data sets .
this match rate was consisted in all three galactic environments .
the variation in properties between matched clouds typically lies within a scatter of a factor of two .
the largest scatter is seen for the derived properties , that depend on multiple cloud variables , such as the surface density and the virial parameter .
the differences in these properties suggest care should be taken when interpreting their physical meaning .
3 . smaller clouds ( @xmath86 10@xmath87 m@xmath2 )
have a larger scatter in their velocity dispersion because the velocity spacing of our ppv ( 1 km s@xmath3 ) data cube is too large to resolve the velocity dispersion of these small objects .
this provides a guide of the velocity ( instrumental spectral ) resolution needed for real observations . in the analysis of this simulation
presented in @xcite using the ppp cloud identification method , clouds were found to fall into three different populations : the _ type a _ clouds which have cloud properties agreeing with typical observations and account for the largest fraction of clouds in all galactic environments .
the _ type b _ massive associations that form through mergers of small clouds and are therefore seen in high - interaction environments like the bar more commonly than the disc .
finally , the _ type c _ clouds that are transient , unbound objects , forming in filaments and tidal tails , making them most common where the gravitationally dominant _ type b _ clouds are prevalent .
when classifying ppv clouds with the same definitions , the main results are : 1 . the fraction of each cloud type differs by @xmath84 10% between ppv and ppp in all environments .
2 . among the @xmath88% of clouds that have a direct match in both data sets , as high as @xmath89%
are categorised as same type in all environments .
the remaining 10% that differ change type due to the different mass and/or radius between two data sets . when we switched from identifying clouds as continuous islands to the peak - based decomposition method , the number of clouds found in both ppp and ppv increases significantly by @xmath6 3 times .
the resulting cloud properties for the overall population remained similar , but the one - to - one cloud match became significantly more difficult in the crowded environment and the match rate dropped to 40% .
the division of clouds such that they contain only a single peak also made the cloud properties more uniform , largely obscuring the three cloud types seen in the island method .
this was particularly true in the ppv data set , which had a large amount of scatter in cloud properties .
we conclude that the cloud identification method therefore plays a critical role in determining the cloud properties and therefore understanding the influence of galactic environment . finally , we emphasize that the resolution of our data means we are only considering the `` best possible '' situation .
our ppv data cube has minimal noise and much higher resolution than typical observations .
based on a simple estimation for alma observations using the simulation function of casa , it is unlikely we can achieve such a fine resolution and deep observation for the entirety of m83 within a reasonable time for cycle 3 ( 2015 2016 ) observations , although a partial galaxy may be possible . as an example , figure [ fig_m83_casa ] shows the simulated spectrum for the three types of cloud derived from a @xmath6 2-hour alma cycle 3 observations of m83 ( including calibrators ) using casa ( the common astronomy software applications package , * ? ? ?
* ) , assuming a 2.6 @xmath31 3.0 kpc observed area in the bar and spiral regions .
we have also ignored the inclination of the galaxy , chemical processes and radiation transfer in the ism .
such effects would alter the observed column density of molecular clouds and also raise the difficulty in identifying the clouds themselves . for our main results using the island method , we conclude that both ppp and ppv can potentially identify the same objects with close properties .
therefore , the techniques themselves are able to be compared well .
the question then becomes one of fighting down noise and achieving the necessary resolution to understand cloud and star formation in galaxies .
the next step in this work is to evaluate the projection effect as a function of different inclinations of galactic discs and physical resolutions .
this will provide a set of reference gmcs to help observers to declare the intrinsic bias in the their observed gmc properties .
2 hours ( @xmath6 5 hours is required if the aca total power observations are included .
this is certainly necessary in real observations ) .
the mapping area is 130@xmath5 @xmath31 150@xmath5 ( 2.6 @xmath31 3.0 kpc ) .
spatial , velocity resolutions and sensitivity are 0.7@xmath5 ( 14 pc ) , 1 km s@xmath3 and 0.012 jy , respectively.,scaledwidth=45.0% ] | as both simulations and observations reach the resolution of the star - forming molecular clouds , it becomes important to clarify if these two techniques are discussing the same objects in galaxies .
we compare clouds formed in a high resolution galaxy simulation identified as continuous structures within a contour , in the simulator s position - position - position ( ppp ) co - ordinate space and the observer s position - position - velocity space ( ppv ) .
results indicate that the properties of the cloud populations are similar in both methods and up to 70% of clouds have a single counterpart in the opposite data structure . comparing individual clouds in a one - to - one match
reveals a scatter in properties mostly within a factor of two .
however , the small variations in mass , radius and velocity dispersion produce significant differences in derived quantities such as the virial parameter .
this makes it difficult to determine if a structure is truely gravitationally bound .
the three cloud types originally found in the simulation in @xcite are identified in both data sets , with around 80% of the clouds retaining their type between identification methods .
we also compared our results when using a peak decomposition method to identify clouds in both ppp and ppv space .
the number of clouds increased with this technique , but the overall cloud properties remained similar .
however , the more crowded environment lowered the ability to match clouds between techniques to 40% .
the three cloud types also became harder to separate , especially in the ppv data set .
the method used for cloud identification therefore plays a critical role in determining cloud properties , but both ppp and ppv can potentially identify the same structures .
[ firstpage ] methods : numerical . techniques : image processing .
galaxies : ism .
ism : clouds .
ism : structure . |
there are various reasons why linear differential equations are easier than non - linear ones .
one is of course that the solutions of linear differential equations form a vector space over the underlying field of constants .
another important feature concerns the singularities .
while for a nonlinear differential equation the location of the singularity may depend continuously on the initial value , this is not possible for linear equations . instead , a solution @xmath7 of a differential equation @xmath8 where @xmath9 are some analytic functions , can only have singularities at points @xmath10 with @xmath11 .
in this article , we consider the case where @xmath9 are polynomials . in this case
, @xmath12 can have only finitely many roots .
we shall also consider the case of recurrence equations @xmath13 where again there is a strong connection between the roots of @xmath12 and the singularities of a solution . while every singularity of a solution leaves a trace in the leading coefficient of an equation ,
the converse is not true .
in general , the leading coefficient @xmath12 may have roots at a point where no solution is singular .
such points are called apparent singularities , and it is sometimes of interest to identify them .
one technique for doing so is called desingularization . as an example , consider the recurrence operator @xmath14 which is taken from ( * ? ? ?
* section 4.1 ) .
here , @xmath15 denotes the shift operator @xmath16 . for any choice of two initial values @xmath17
, there is a unique sequence @xmath18 with @xmath19 , @xmath20 and @xmath0 applied to @xmath21 gives the zero sequence . a priori , it is not obvious whether or not @xmath21 is actually an integer sequence , if we choose @xmath22 from @xmath23 , because the calculation of the @xmath24nd term from the earlier terms via the recurrence encoded by @xmath0 requires a division by @xmath25 , which could introduce fractions . in order to show that this division never introduces a denominator , the authors of @xcite note that every solution of @xmath0 is also a solution of its left multiple @xmath26 the operator @xmath27 has the interesting property that the factor @xmath25 has been `` removed '' from the leading coefficient
this is , however , not quite enough to complete the proof , because now a denominator could still arise from the division by @xmath28 at each calculation of a new term via @xmath27 . to complete the proof ,
the authors show that the potential denominators introduced by @xmath25 and by @xmath28 , respectively , are in conflict with each other , and therefore no such denominators can occur at all .
the process of obtaining the operator @xmath27 from @xmath0 is called desingularization , because there is a polynomial factor in the leading coefficient of @xmath0 which does not appear in the leading coefficient of @xmath27 . in the example above , the price to be paid for the desingularization was a new constant factor @xmath28 which appears in the leading coefficient of @xmath27 but not in the original leading coefficient of @xmath0 .
desingularization algorithms in the literature @xcite care only about the removal of polynomial factors without introducing new polynomial factors , but they do not consider the possible introduction of new constant factors . a contribution of the present paper is a desingularization algorithm which minimizes , in a sense , also any constant factors introduced during the desingularization .
for example , for the operator @xmath0 above , our algorithm finds the alternative desingularization @xmath29 which immediately certifies the integrality of its solutions . in more algebraic terms , we consider the following problem . given an operator @xmath30[{\partial}]$ ] , where @xmath31[{\partial}]$ ] is an ore algebra ( see section [ sect : pre ] for definitions ) , we consider the left ideal @xmath32l$ ] generated by @xmath0 in the extended algebra @xmath33 $ ] .
the contraction of @xmath34 to @xmath31[{\partial}]$ ] is defined as @xmath35[{\partial}]$ ] .
this is a left ideal of @xmath31[{\partial}]$ ] which contains @xmath31[{\partial}]l$ ] , but in general more operators .
our goal is to compute a @xmath31[{\partial}]$]-basis of @xmath36 . in the example
above , such a basis is given by @xmath37 ( see example [ ex : ah ] ) .
the traditional desingularization problem corresponds to computing a basis of the @xmath38[{\partial}]$]-left ideal @xmath39[{\partial}]$ ] .
the contraction problem for ore algebras @xmath38[{\partial}]$ ] was proposed by chyzak and salvy ( * ? ? ?
* section 4.3 ) . for the analogous problem in commutative polynomial rings ,
there is a standard solution via grbner bases ( * ? ? ?
* section 8.7 ) .
it reduces the contraction problem to a saturation problem .
this reduction also works for the differential case , but in that case it is not so helpful because it is less obvious how to solve the saturation problem .
a solution was proposed by tsai @xcite , which involves homological algebra and d - modules theory .
our work is based on desingularization for ore operators in @xcite . in particular , the @xmath40-removing operator in ( * ? ? ?
* lemma 4 ) provides us with a key to determine contraction ideals .
the algorithm developed in this paper is considerably simpler than tsai s and at the same time it applies to arbitrary ore algebras rather than only the differential case .
moreover , we compute a completely desingularized operator in a contraction ideal , which has minimal leading coefficient in terms of both degree and content . the rest of this paper is organized as follows . in section [
sect : pre ] , we describe ore polynomial rings over a principal ideal domain , and extend the notion of @xmath40-removed operators to them .
the notion of desingularized operators is defined and connected with contraction ideals in section [ sect : dc ] .
we determine a contraction ideal in section [ sect : algo ] , and compute completely desingularized operators in section [ sect : cd ] .
this section is divided into three parts .
first , we describe ore algebras that are used in the paper .
second , we extend the notion of @xmath40-removed operators in @xcite . at last
, we make some remarks on grbner basis computation over a principal ideal domain . throughout the paper ,
we let @xmath5 be a principal ideal domain . for instance
, @xmath5 can be the ring of integers or that of univariate polynomials over a field .
we consider the _ ore algebra _
@xmath4[{\partial } ; \sigma , \delta]$ ] , where @xmath41 \rightarrow r[x]$ ] is a ring automorphism that leaves the elements of @xmath5 fixed , and @xmath42 \rightarrow r[x]$ ] is a @xmath43-derivation , _ i.e. _ an @xmath5-linear map satisfying the skew leibniz rule @xmath44.\ ] ] the addition in @xmath4[{\partial}]$ ] is coefficient - wise and the multiplication is defined by associativity via the commutation rule @xmath45 for @xmath46.$ ] given @xmath47[{\partial}]$ ] , we can uniquely write it as @xmath48 with @xmath49 $ ] and @xmath50 .
we call @xmath51 the _ order _ and @xmath52 the _ leading coefficient _ of @xmath0
. they are denoted by @xmath53 and @xmath54 , respectively .
the ring @xmath4[{\partial } ; \sigma , \delta]$ ] is abbreviated as @xmath4[{\partial}]$ ] when @xmath43 and @xmath55 are clear from the context . for a subset @xmath56 of @xmath4[{\partial}]$ ] , the left ideal generated by @xmath56 is denoted by @xmath4[{\partial}]\cdot s$ ] .
let @xmath57 be the quotient field of @xmath5 .
then @xmath58 $ ] is an ore algebra containing @xmath4[{\partial}]$ ] . for @xmath47[{\partial}]$ ] , we define the _ contraction ideal _ of @xmath0 to be @xmath58 l \cap r[x][{\partial}]$ ] and denote it by @xmath36 .
we generalize some terminologies given in @xcite by replacing the coefficient ring @xmath59 $ ] with @xmath4 $ ] , where @xmath60 is a field .
[ def : premovable ] let @xmath47[{\partial}]$ ] with positive order , and @xmath40 be a divisor of @xmath54 in @xmath4 $ ] .
* we say that @xmath40 is _ removable _ from @xmath0 at order @xmath61 if there exist @xmath62 $ ] with order @xmath63 , and @xmath64 $ ] with @xmath65 in @xmath4 $ ] such that @xmath66[{\partial } ] \quad \text{and } \quad \sigma^{-k}({\operatorname{lc}}_{{\partial}}(pl ) ) = \frac{w}{vp}{\operatorname{lc}}_{{\partial}}(l).\ ] ] we call @xmath67 a _ @xmath40-removing operator for @xmath0 over @xmath4 $ ] _ , and @xmath68 the corresponding _
@xmath40-removed operator_. * @xmath40 is simply called _ removable _ from @xmath0 if it is removable at order @xmath63 for some @xmath69 .
otherwise , @xmath40 is called _ non - removable _ from @xmath0 .
note that every @xmath40-removed operator lies in @xmath36 . in the example of section [ sect : intro ] , @xmath70 is removable from @xmath0 at order 1 .
and @xmath27 is the corresponding @xmath70-removed operator for @xmath0 . in the differential ore algebra @xmath31[{\partial}]$ ] , where @xmath71 , let @xmath72 then @xmath73 is an @xmath1-removed operator for @xmath0 ( see ( * ? ? ?
* example 3 ) ) .
the authors of @xcite provide a convenient form of @xmath40-removing operators over @xmath59 $ ] in order to get the order bound .
we derive a similar form over @xmath4 $ ] and use it in section [ sect : cd ] .
[ premovable ] let @xmath47[{\partial}]$ ] with positive order .
assume that @xmath46 $ ] is removable from @xmath0 at order @xmath63 .
then there exists a @xmath40-removing operator for @xmath0 over @xmath4 $ ] in the form @xmath74 where @xmath75 belongs to @xmath4 $ ] , @xmath76 in @xmath4 $ ] , @xmath77 @xmath78 , @xmath63 , and @xmath79 .
by definition [ def : premovable ] , @xmath80 for some @xmath81 in @xmath4 $ ] with @xmath82 .
then we can write a @xmath40-removing operator for @xmath0 over @xmath4 $ ] in the form @xmath83 where @xmath84 $ ] , @xmath85 in @xmath4 $ ] , @xmath86 , @xmath79 .
let @xmath87 , @xmath88 , @xmath86 .
then @xmath89 where @xmath90 in @xmath4 $ ] , @xmath86
. moreover , @xmath91 by definition [ def : premovable ] , @xmath92 is a @xmath40-removing operator for @xmath0 over @xmath4 $ ] with the required form . in sections [ sect :
algo ] and [ sect : cd ] , we will make essential use of grbner bases in @xmath4[{\partial}]$ ] . when @xmath93 $ ] with @xmath6 being a field , the notion of grbner bases and buchberger s algorithm are available @xcite .
in our case , @xmath43 is an @xmath5-automorphism of @xmath4 $ ] , which implies that @xmath94 where @xmath95 are in @xmath5 and @xmath96 is a unit .
assume that @xmath97 is a term order on @xmath98 .
let @xmath67 be a nonzero operator in @xmath4[{\partial}]$ ] , and @xmath99 be the head coefficient of @xmath67 with respect to @xmath97 . by the commutation rule , @xmath100 has head coefficient @xmath101 , which is associated to @xmath99 , because @xmath102 is a unit .
this observation enables us to extend the notion of grbner bases and buchberger s algorithm in @xcite to ore case in a straightforward way .
in this section , we define the notion of desingularized operators , and connect it with contraction ideals . as a matter of notation , for an operator @xmath47[{\partial}]$ ] , we set @xmath103 note that @xmath104 is a left submodule of @xmath36 over @xmath4 $ ] .
we call it the @xmath63th submodule of @xmath36 .
when the operator @xmath0 is clear from context , @xmath104 is simply denoted by @xmath105 .
[ def : desingularization ] let @xmath47[{\partial}]$ ] with order @xmath106 , and @xmath107 where @xmath108 and @xmath109 , @xmath110 \setminus r$ ] are irreducible and pairwise coprime .
an operator @xmath111[{\partial}]$ ] of order @xmath63 is called a
_ desingularized operator for @xmath0 _ if @xmath112 and @xmath113 where @xmath114 , and @xmath115 is non - removable from @xmath0 for each @xmath116 , @xmath117 .
desingularized operators always exist by ( * ? ? ?
* lemma 4 ) .
[ lm : key ] let @xmath47[{\partial}]$ ] be of order @xmath106 , and @xmath69 with @xmath118 .
assume that @xmath27 is a desingularized operator for @xmath0 and @xmath119 .
* @xmath120 * @xmath121 is a desingularized operator for @xmath0 for each @xmath122 . *
set @xmath123 , where @xmath124 and @xmath125 $ ] is primitive .
then , for all @xmath126 of order @xmath127 with @xmath128 , @xmath129 divides @xmath130 in @xmath4 $ ] .
\(i ) let @xmath131 and @xmath132 suppose that @xmath133
. then there exists @xmath134 with @xmath135 . without loss of generality
, we can assume that @xmath136 , because the leading coefficients of @xmath137 and @xmath138 are of the same degree for all @xmath122 . by pseudo - division in
@xmath4 $ ] , we have that @xmath139 for some @xmath140 , @xmath141 $ ] , and @xmath142 or @xmath143 .
if @xmath144 were nonzero , then @xmath145 would be a nonzero operator of order @xmath63 in @xmath36 whose leading coefficient is of degree less than @xmath146 , a contradiction .
thus , @xmath147 . in particular , @xmath148 is positive , as @xmath133 .
it follows from that @xmath149 which belongs to @xmath4 $ ] .
hence , @xmath150 divides @xmath54 in @xmath4 $ ] .
consequently , there exists @xmath151 such that @xmath75 divides @xmath150 in @xmath4 $ ] .
this implies that @xmath152 is removable from @xmath0 , a contradiction .
( ii ) it is immediate from definition [ def : desingularization ] .
( iii ) let @xmath153 , where @xmath154 and @xmath7 is primitive in @xmath4 $ ] .
by ( ii ) , @xmath155 is a desingularized operator whose leading coefficient equals @xmath156 .
a similar argument used in the proof of the first assertion implies that @xmath157$.}\ ] ] by gauss s lemma in @xmath4 $ ] , @xmath158 .
we describe a relation between desingularized operators and contraction ideals .
let @xmath2 be a left ideal in @xmath4[{\partial}]$ ] , and @xmath124 .
the saturation of @xmath2 with respect to @xmath96 is defined to be @xmath159[{\partial } ] \mid a^i p \in i \,\
, \text{for some~$i \in { { \mathbb{n}}}$ } \right\}.\ ] ] since @xmath96 is a constant with respect to @xmath43 and @xmath55 , the saturation @xmath160 is a left ideal .
[ th : dc ] let @xmath47[{\partial}]$ ] with order @xmath161 .
assume that @xmath27 is a desingularized operator for @xmath0 .
let @xmath162 , where @xmath124 and @xmath163 is primitive in @xmath4 $ ] .
if @xmath27 belongs to @xmath105 for some @xmath69 , then @xmath164[{\partial } ] \cdot m_k ) : a^{\infty}.\ ] ] by lemma [ lm : key ] ( ii ) , we may assume that the order of @xmath27 is equal to @xmath63 .
set @xmath165[{\partial } ] \cdot m_k ) : a^{\infty}$ ] .
first , assume that @xmath166 .
then there exists @xmath167 such that @xmath168[{\partial } ] \cdot m_k$ ] .
it follows that @xmath169 l$ ] .
thus , @xmath126 by definition . next ,
note that @xmath170 and that @xmath171 .
it suffices to show @xmath172 .
we proceed by induction on @xmath173 . for @xmath174 .
@xmath175 by definition .
suppose that the claim holds for @xmath173 .
for any @xmath176 , @xmath177 . by lemma
[ lm : key ] ( iii ) , @xmath178 for some @xmath46 $ ]
. then @xmath179 .
it follows that @xmath180 . since @xmath181[{\partial } ] \cdot m_k \subseteq r[x][{\partial } ] \cdot m_i,\ ]
] we have that @xmath182[{\partial } ] \cdot m_i$ ] . on the other hand , @xmath183 by the induction hypothesis .
thus , @xmath182[{\partial } ] \cdot j$ ] , which is @xmath184 .
accordingly , @xmath166 by the definition of saturation .
first , we translate an upper bound for the order of a desingularized operator over @xmath185 $ ] to @xmath4 $ ] .
[ lm : bound ] let @xmath186[{\partial}]$ ] with order @xmath161 , and @xmath187 $ ] be a primitive polynomial and a divisor of @xmath54 .
assume that there exists a @xmath40-removing operator for @xmath0 over @xmath185 $ ] .
then there exists @xmath40-removing operator for @xmath0 over @xmath4 $ ] with order @xmath51 .
assume that @xmath188 $ ] is a @xmath40-removing operator for @xmath0 over @xmath185 $ ] .
let @xmath67 be of order @xmath63 .
then @xmath68 is of the form @xmath189 for some @xmath190 , b_i \in r$ ] , @xmath191 .
moreover , @xmath192 where @xmath193 $ ] with @xmath194 .
let @xmath195 in @xmath5 and @xmath196 .
then @xmath197[{\partial } ] \quad \text{and } \quad { { \sigma}}^{-k}({\operatorname{lc}}_{{\partial}}(pl ) ) = \frac{bw}{vp } { \operatorname{lc}}_{{\partial}}(l).\ ] ] since @xmath40 is primitive , we have that @xmath198 in @xmath4 $ ] .
thus , @xmath199 is a @xmath40-removing operator of order @xmath63 . by
the above lemma , an order bound for a @xmath40-removing operator over @xmath185 $ ] is also an order bound for a @xmath40-removing operator over @xmath4 $ ] .
the former has been well - studied in the literature .
order bounds for differential operators are given in ( * ? ? ?
* algorithm 3.4 ) and ( * ? ? ?
* lemma 4.3.12 ) .
those for recurrence operators are given in ( * ? ? ?
* lemma 4 ) and ( * ? ? ?
* lemma 4.3.3 ) .
desingularized operators are @xmath40-removing operators .
so we can find order bounds for them . by theorem [ th : dc ] , determining a contraction ideal amounts to finding a desingularized operator @xmath27 and an @xmath4$]-basis of @xmath105 , where @xmath63 is an upper bound for the order of @xmath27 .
next , we present an algorithm for constructing a basis for @xmath104 , where @xmath0 is a nonzero operator in @xmath4[{\partial}]$ ] and @xmath63 is a positive integer . to this end , we embed @xmath105 into the free module @xmath4^{k+1}$ ] over @xmath4 $ ] .
let us recall the right division in @xmath58 $ ] ( see ( * ? ? ?
* section 3 ) ) . for @xmath200
$ ] with @xmath201 , there exist unique elements @xmath202 $ ] with @xmath203 such that @xmath204 .
we call @xmath5 the _ right - remainder _ of @xmath205 by @xmath206 and denote it by @xmath207 .
let @xmath208[{\partial}]$ ] with order @xmath63
. then @xmath209 if and only if @xmath210[{\partial } ] l$ ] , which is equivalent to @xmath211 .
the latter gives rise to a linear system @xmath212 where @xmath213 is a @xmath214 matrix over @xmath215 .
clearing denominators of the elements in @xmath213 , we may further assume that @xmath213 is a matrix over @xmath4 $ ] .
we are concerned with the solutions of . set @xmath216^{k+1 } \mid ( f_k , \ldots , f_0 ) a = \mathbf{0 } \right\}.\ ] ]
we call @xmath217 the module of syzygies defined by . [
th : iso ] with the notation just specified , we have @xmath218 is a module isomorphism over @xmath4 $ ] .
let @xmath219 $ ] .
if @xmath205 belongs to @xmath105 , then @xmath211 , that is , @xmath220 belongs to @xmath217 .
hence , @xmath221 is a well - defined map .
clearly , @xmath221 is injective . for @xmath222
, we have @xmath223 as is induced by right division @xmath224 , @xmath205 belongs to @xmath105 .
so @xmath221 is surjective .
it is straightforward to see that @xmath221 is an @xmath4$]-module homomorphism .
by theorem [ th : iso ] , @xmath105 is finitely generated over @xmath4 $ ] .
to find an @xmath4$]-basis of @xmath105 , it suffices to compute a basis of the module of syzygies defined by . when @xmath5 is a field , we just need to solve over a principal ideal domain ( * ? ? ?
* chapter 5 ) .
when @xmath5 is the ring of integers or the ring of univariate polynomials over a field , we can use grbner bases of polynomials over a principal domain @xcite .
their implementations are available in computer algebra systems such as macaulay2 and singular .
we now consider how to construct a desingularized operator for @xmath0 . for @xmath225 , we define @xmath226 p \mid p \in m_k \right\ } \cup \{0\},\ ] ] where @xmath227 p$ ] stands for the coefficient of @xmath228 in @xmath67 .
it is clear that @xmath229 is an ideal of @xmath4 $ ] .
we call @xmath229 the @xmath63th coefficient ideal of @xmath36 . by the commutation rule , @xmath230 .
[ lm : dop ] let @xmath47[{\partial}]$ ] be of positive order . if the @xmath63th submodule @xmath105 of @xmath36 has a basis @xmath231 over @xmath4 $ ] , then the @xmath63th coefficient ideal @xmath232 b_1 , \ldots , [ { \partial}^k ] b_\ell \right\rangle.\ ] ] obviously , @xmath233 b_1 , \ldots , [ { \partial}^k ] b_\ell \rangle \subseteq i_k$ ]
let @xmath234 .
then @xmath235 for some @xmath209 with @xmath236 .
since @xmath105 is generated by @xmath231 over @xmath4 $ ] , @xmath237$.}\ ] ] thus , @xmath238 b_1 \right ) + \cdots + h_\ell \left ( [ { \partial}^k ] b_\ell \right).$ ] consequently , @xmath239 b_1 , \ldots , [ { \partial}^k ] b_\ell \rangle$ ] . [
th : dop ] let @xmath47[{\partial}]$ ] be of positive order .
assume that the @xmath63th submodule @xmath105 of @xmath36 contains a desingularized operator for @xmath0 .
let @xmath240 be a nonzero element in the @xmath63th coefficient ideal with minimal degree .
then an operator @xmath56 in @xmath105 with leading coefficient @xmath240 is a desingularized operator .
let @xmath27 be a desingularized operator in @xmath105 .
by lemma [ lm : key ] ( ii ) , we may assume that the order of @xmath27 is equal to @xmath63 .
let @xmath241
. then @xmath242 by lemma [ lm : key ] ( i ) .
let @xmath21 be the leading coefficient of @xmath240 with respect to @xmath1 and @xmath243 be that of @xmath244 .
then @xmath245 is zero .
otherwise , @xmath246 would be an operator of order @xmath63 whose leading coefficient with respect to @xmath15 has degree lower than @xmath247 , a contradiction to lemma [ lm : key ] ( i ) .
it follows from @xmath248 and definition [ def : desingularization ] that @xmath56 is also a desingularized operator .
let @xmath0 be an operator in @xmath249[{\partial}]$ ] of positive order .
we can compute a basis @xmath250 for the @xmath63th submodule of @xmath36 by theorem [ th : iso ] , where @xmath63 is an upper bound on the order of a desingularized operator for @xmath0 . by lemma [ lm : dop ] , we can obtain a basis @xmath251 for the @xmath63th coefficient ideal @xmath229 of @xmath36 . let @xmath252 be the extension ideal of @xmath229 in @xmath185 $ ] . using the extended euclidean algorithm in @xmath185 $ ] and clearing denominators , we find cofactors @xmath253 , @xmath254 $ ] and @xmath255 $ ] such that @xmath256 it follows from theorem [ th : dop ] that @xmath257 is a desingularized operator for @xmath0 with @xmath258 .
let @xmath96 be the content of @xmath240 with respect to @xmath1 . by theorem [
th : dc ] , @xmath36 is the saturation of @xmath4[{\partial } ] \cdot m_k$ ] with respect to @xmath96 . note that @xmath96 belongs to @xmath5 , which is contained in the center of @xmath4[{\partial}]$ ] .
so a basis of the saturation ideal can be computed in the same way as in the commutative case .
[ prop : saconst ] let @xmath2 be a left ideal of @xmath4[{\partial}]$ ] and @xmath99 be non - zero element in @xmath5 .
assume that @xmath184 is a left ideal @xmath259[{\partial } ] \cdot \left ( i \cup \ { 1 -c y \ } \right),\ ] ] where @xmath260 is a new indeterminate and commutes with every element in @xmath4[{\partial}]$ ]
. then @xmath261 = @xmath262[{\partial}]$ ] . since both @xmath260 and @xmath99 commute with @xmath15 , the argument in ( * ? ? ?
* , proposition 6.37 ) carries over .
we outline our method for determining contraction ideals .
[ algo : cont ] given @xmath47[{\partial}]$ ] , where @xmath263 or @xmath71 , compute a basis of @xmath36 . 1 .
derive an upper bound @xmath63 on the order of a desingularized operator for @xmath0 .
2 . compute an @xmath4$]-basis of @xmath105 .
3 . compute a desingularized operator @xmath27 , and set @xmath96 to be the content of @xmath264 with respect to @xmath1 .
4 . compute a basis of @xmath265[{\partial } ] \cdot m_k ) : a^{\infty}$ ] .
the termination of this algorithm is evident .
its correctness follows from theorem [ th : dc ] .
we assume that the commutation rule in @xmath4[{\partial}]$ ] is either @xmath263 or @xmath71 in @xmath4[{\partial}]$ ] , because we only know order bounds for those cases . in step 1
, the order bound is derived from ( * ? ? ?
* lemma 4 ) and ( * ? ? ?
* algorithm 3.4 ) . in step 2 , we need to solve linear systems over @xmath4 $ ] as stated in theorem [ th : iso ] .
this can be done by grbner basis computation in a free @xmath4$]-module of finite rank . in step 3 ,
@xmath27 is computed according to theorem [ th : dop ] and the extended euclidean algorithm in @xmath185 $ ] .
the last step is carried out according to proposition [ prop : saconst ] . in the shift ore algebra
@xmath266[n][{\partial}]$ ] , in which the commutation rule is @xmath267 . consider @xmath268 by ( * ? ? ?
* lemma 4 ) , we obtain an order bound @xmath269 for a desingularized operator .
thus , @xmath270 contains a desingularized operator for @xmath0 . in step 2 of algorithm [ algo : cont ]
, we find that @xmath270 is generated by @xmath271 where @xmath272 is a desingularized operator , @xmath273 .
using grbner bases , @xmath274[n][{\partial } ] \cdot m_2 ) : ( 2 + t)^{\infty}$ ] is generated by @xmath275 .
let us consider the example in section [ sect : intro ] .
[ ex : ah ] in the shift ore algebra @xmath276[{\partial}]$ ] , let @xmath277 by ( * ? ? ?
* lemma 4 ) , we obtain an order bound @xmath278 for a desingularized operator .
thus , @xmath279 contains a desingularized operator for @xmath0 . in step 2 of algorithm [ algo : cont ] , we find that @xmath279 is generated by @xmath280 , where @xmath281 is given in .
note that @xmath282 .
thus , @xmath281 is a desingularized operator .
consequently , @xmath283[{\partial } ] \cdot \{l , \tilde{t } \ } ) : 1^{\infty } = { { \mathbb{z}}}[n][{\partial } ] \cdot \{l , \tilde{t } \}.\ ] ] [ ex : bm ] in the differential ore algebra @xmath31[{\partial}]$ ] , in which the commutation rule is @xmath71 .
consider the operator @xmath284[{\partial}]$ ] in @xcite .
by ( * ? ? ?
* algorithm 3.4 ) , we obtain an order bound @xmath285 for a desingularized operator .
thus , @xmath286 contains a desingularized operator for @xmath0 . in step 2 of algorithm [ algo : cont ] , we find that @xmath286 is generated by @xmath287 , where @xmath288 .
note that @xmath289 .
thus , @xmath27 is a desingularized operator .
consequently , @xmath290[{\partial } ] \cdot \{l , { \partial}l , t \ } ) : 1^{\infty } = { { \mathbb{z}}}[x][{\partial } ] \cdot \{l , t \}.\ ] ]
as seen in section [ sect : intro ] , the shift operator @xmath291 has a desingularized operator @xmath27 with leading coefficient @xmath28 .
but the content of @xmath292 is 1 .
the redundant content @xmath28 has been removed by computing another desingularized operator @xmath293 in .
this enables us to see immediately that the sequence annihilated by @xmath0 is an integer sequence when its initial values are integers .
krattenthaler proposes a conjecture in @xcite : let @xmath294 and @xmath295 be two p - recursive sequences over @xmath23 with leading coefficients @xmath61 . then @xmath296 is also a p - recursive sequence over @xmath23 with leading coefficient @xmath61 . to test the conjecture for the two particular sequences ,
one may first compute an annihilator @xmath0 of @xmath296 , and then look for a nonzero operator in @xmath36 whose leading coefficient has minimal content with respect to @xmath61 . when the content is equal to @xmath297 , the conjecture is true for these sequences .
these two observations motivate us to define the notion of completely desingularized operators .
[ def : cd ] let @xmath47[{\partial}]$ ] with positive order , and @xmath137 a desingularized operator for @xmath0 . set @xmath298 , where @xmath99 is the content of @xmath299 with respect to @xmath1 and @xmath163 the corresponding primitive part .
we call @xmath137 a _ completely designularized operator _ for @xmath0 if @xmath99 is a divisor of the content of the leading coefficient of every desingularized operator for @xmath0 . to see the existence of completely designularized operators
, we assume that @xmath0 is of order @xmath51 . for
a desingularized operator @xmath27 of order @xmath63 , equations and in definition [ def : desingularization ] enable us to write @xmath300 where @xmath301 and @xmath302 note that @xmath163 is primitive and independent of the choice of desingularized operators .
[ lm : ideal ] let @xmath47[{\partial}]$ ] with order @xmath303 . set @xmath2 to be the set consisting of zero and @xmath304 given in for all desingularized operators for @xmath0 .
then @xmath2 is an ideal of @xmath5 . by definition [ def : desingularization ] ,
the product of a nonzero element of @xmath5 and a desingularized operator for @xmath0 is also a desingularized one .
so it suffices to show that @xmath2 is closed under addition .
let @xmath272 and @xmath305 be two desingularized operators of orders @xmath306 and @xmath307 , respectively .
assume that @xmath308 .
by , @xmath309 if @xmath310 , then there is nothing to prove .
otherwise , a direct calculation implies that @xmath311 thus , @xmath312 has leading coefficient @xmath313 accordingly , @xmath312 is a desingularized one , which implies that @xmath314 belongs to @xmath2 .
since @xmath5 is a principal ideal domain , @xmath2 in the above lemma is generated by an element @xmath99 , which corresponds to a completely desingularized operator .
let @xmath97 be a term order on @xmath315 .
for any non - zero operator @xmath316[{\partial}]$ ] , we define the _ head term _ of @xmath67 to be the highest term appearing in @xmath67 with respect to @xmath97 , and denote it by @xmath317 .
the next technical lemma serves as a step - stone to construct completely desingularized operators .
[ lm : sc ] let @xmath47[{\partial}]$ ] with order @xmath161 , and @xmath318 .
then @xmath4[{\partial } ] \cdot m_k { = } r[x][{\partial } ] \cdot m_{k + 1}$ ] if and only if @xmath319 assume that @xmath320 . since @xmath321 , it suffices to prove that @xmath322[{\partial } ] \cdot m_k$ ] . for each @xmath323 ,
we have that @xmath324 .
thus , there exists @xmath209 such that @xmath325 .
in other words , @xmath326 . consequently , @xmath111[{\partial } ] \cdot m_k$ ] .
conversely , assume that @xmath4[{\partial } ] \cdot m_{k + 1}= r[x][{\partial } ] \cdot m_k$ ] .
it suffices to prove that @xmath327 because @xmath328 by definition .
let @xmath329 be an @xmath4$]-basis of @xmath105 .
then @xmath329 is also a basis of the left ideal @xmath4[{\partial } ] \cdot m_k$ ] .
let @xmath97 be the term order such that @xmath330 if either @xmath331 or @xmath332 and @xmath333 .
since @xmath334 for each @xmath335 , s - polynomials and g - polynomials formed by elements in @xmath105 have orders no more than @xmath63 ( * ? ? ?
* definition 10.9 ) . by buchberger
s algorithm , there exists a grbner basis @xmath336 of @xmath4[{\partial } ] \cdot \mathcal{b}$ ] with respect to @xmath97 such that @xmath337 for each @xmath338 . for @xmath339 , there exists @xmath340 such that @xmath341 . since @xmath342[{\partial } ] \cdot m_{k + 1}$ ] , we have @xmath342[{\partial } ] \cdot m_k$ ] .
it follows that @xmath27 is reduced to zero by @xmath336 .
thus , @xmath343 by the choice of term order , @xmath344 .
if @xmath345 is of order @xmath346 , then @xmath347 , where @xmath348 is in @xmath4 $ ] and @xmath349 is the order of @xmath206 .
comparing the leading coefficients of operators in both sides of and noticing @xmath350 , we have @xmath351 it follows that @xmath352 on the other hand , @xmath353 implies that @xmath354 .
we have that @xmath355 by .
thus , @xmath356 . by the above lemma ,
@xmath357 whenever @xmath358 and @xmath359[{\partial}]\cdot m_\ell$ ] . in this case , a basis of @xmath360 can be obtain by shifting a basis of @xmath361 , which allows us to find a completely desingularized operator .
[ th : gbcd ] let @xmath47[{\partial}]$ ] with order @xmath161 .
assume that the @xmath362th submodule @xmath363 of @xmath36 contains a basis of @xmath36 .
let @xmath361 be the @xmath362th coefficient ideal of @xmath36 , and @xmath364 a reduced grbner basis of @xmath361 .
let @xmath365 be of the lowest degree in @xmath1 and @xmath205 be the operator in @xmath36 with @xmath366 .
then @xmath205 is a completely desingularized operator for @xmath0 .
by lemma [ lm : ideal ] , @xmath36 contains a completely desingularized operator @xmath56 .
let @xmath367 .
then @xmath368 is in @xmath360 for some @xmath358 . by lemma [ lm : sc ] , @xmath369 .
it follows that @xmath370 belongs to @xmath361 . by
, we have @xmath371 where @xmath372 and @xmath163 is a primitive polynomial in @xmath4 $ ] .
a direct calculation implies that @xmath373 . since @xmath374 , so does @xmath375 . note
that @xmath205 is a desingularized operator by theorem [ th : dop ] . by , @xmath376 where @xmath377 .
thus , @xmath378 .
since @xmath364 is a reduced grbner basis of @xmath361 , @xmath7 is the unique polynomial in @xmath364 with minimal degree .
moreover , @xmath379 is of minimal degree in @xmath361 .
so it can be reduced to zero by @xmath7 .
thus , @xmath380 . on the other hand , @xmath381 by definition [ def : cd ] .
thus , @xmath382 and @xmath383 are associated to each other .
consequently , @xmath205 is a completely desingularized operator for @xmath0 .
the construction in the above theorem leads to the following algorithm .
[ algo : cd ] given @xmath47[{\partial}]$ ] , where @xmath263 or @xmath71 , compute a completely desingularized operator for @xmath0 . 1 .
compute a basis @xmath384 of @xmath36 by algorithm [ algo : cont ] .
2 . set @xmath362 to be the highest order of the elements in @xmath384 .
compute an @xmath4$]-basis @xmath329 of @xmath363 .
3 . set @xmath385 .
compute a reduced grbner basis @xmath364 of @xmath386 4 .
set @xmath7 to be the polynomial in @xmath364 whose degree is the lowest one in @xmath1 . tracing back to the computation of step 3 , one can find @xmath387 $ ] such that @xmath388 5 .
output @xmath389 .
the termination of this algorithm is evident .
its correctness follows from theorem [ th : gbcd ] .
[ example2 ] consider two sequences @xmath294 and @xmath295 satisfying the following two recurrence equations @xcite @xmath390 respectively .
the sequence @xmath391 has an annihilator @xmath392[{\partial}]$ ] with @xmath393 in step 1 of the above algorithm , @xmath394[{\partial}]\cdot m_{14}$ ] . in steps 2 and 3
, we observe that @xmath395 is generated by @xmath396 .
in other words , we obtain a completely desingularized operator @xmath27 of order @xmath397 with @xmath398 . translating into the recurrence equations of @xmath399
, we arrive at @xmath400 where @xmath401 $ ] , @xmath402 .
this verifies krattenthaler s conjecture for the sequences @xmath403 and @xmath404 .
note that it is impossible to have a completely desingularized operator of order less than @xmath397 .
in fact , for some lower orders , one can obtain @xmath405 they can not produce a leading coefficient whose degree and content are both minimal . consider the following recurrence equations : @xmath406 let @xmath391 , which has an annihilator @xmath392[{\partial}]$ ] of order @xmath407 with @xmath408 , where @xmath409 $ ] and @xmath410 . by the known algorithms for desingularization in @xcite , we find that @xmath399 satisfies the recurrence equation @xmath411 where @xmath412 is an 853-digit integer , @xmath413 $ ] , @xmath414 .
on the other hand , algorithm [ algo : cd ] finds a completely desingularized operator @xmath27 for @xmath0 of order @xmath397 whose leading coefficient is @xmath415 .
translating into the recurrence equation of @xmath399 yields @xmath416 where @xmath417 $ ] .
let @xmath47[{\partial}]$ ] with positive order and @xmath27 a desingularized operator for @xmath0 .
then the degree of @xmath292 in @xmath1 is equal to @xmath418 where @xmath419 are given in definition [ def : desingularization ] .
hence , @xmath36 can not contain any operator whose leading coefficient has degree lower than @xmath146 .
we provide a lower bound for the content of the leading coefficients of operators in @xmath36 with respect to the divisibility relation on @xmath5 . to this end
, we write @xmath420 where @xmath421 and @xmath422 $ ] is primitive , @xmath423 .
we say that @xmath0 is _ @xmath5-primitive _ if @xmath424 .
gauss s lemma in the commutative case also holds for @xmath5-primitive polynomials .
[ lm : gauss ] let @xmath67 and @xmath137 be two operators in @xmath4[{\partial}]$ ] .
if @xmath67 and @xmath137 are @xmath5-primitive , so is @xmath425 .
first , we recall a result in ( * ? ? ?
* theorem 3.7 , corollary 3.8 ) .
assume that @xmath213 is a ring with endomorphism @xmath426 and @xmath427-derivation @xmath428 .
let @xmath429 be a @xmath427-@xmath55-ideal , that is , an ideal such that @xmath430 and @xmath431 .
then there exists a unique ring homomorphism @xmath432 \rightarrow ( a / i)[\tilde{{\partial } } ; \tilde { { { \sigma } } } , \tilde{\delta}]\ ] ] such that @xmath433 is the canonical homomorphism , and @xmath434 , where @xmath435 and @xmath436 are the homomorphism and @xmath435-derivation induced by @xmath427 and @xmath55 , respectively .
let @xmath40 be a prime element of @xmath5 and @xmath437 be the corresponding ideal in @xmath4 $ ] .
then @xmath2 is a @xmath427-@xmath55-ideal . from the above paragraph
, there exists a unique homomorphism @xmath438[{\partial } ; { { \sigma } } , \delta ] \rightarrow ( r[x]/i)[\tilde{{\partial } } ; \tilde { { { \sigma } } } , \tilde{\delta}]\ ] ] such that @xmath439 } : r[x ] \rightarrow r[x]/i$ ] is the canonical homomorphism , and @xmath434 .
note that @xmath440 , because , for @xmath441 with @xmath442 $ ] , @xmath443 .
it follows that @xmath435 is an injective endomorphism of @xmath444 . on the other hand , @xmath4/i$ ]
is a domain because @xmath2 is a prime ideal .
thus , @xmath265/i)[\tilde{{\partial } } ; \tilde { { { \sigma } } } , \tilde{\delta}]$ ] is a domain because @xmath4/i$ ] is a domain and @xmath435 is injective . since @xmath67 and @xmath0 are @xmath5-primitive , we have that @xmath445 .
so @xmath446 , because @xmath447 is a homomorphism .
consequently , @xmath448 is @xmath5-primitive .
there are more sophisticated variants of gauss s lemma for ore operators in ( * ? ? ?
* proposition 2 ) and ( * ? ? ?
* lemma 9.5 ) .
[ th : cremovable ] let @xmath47[{\partial}]$ ] with positive order and @xmath40 be a non - unit element of @xmath5 . if @xmath0 is @xmath5-primitive and @xmath449 , then @xmath40 is non - removable .
assume that @xmath40 is removable , then there exists a @xmath40-removing operator @xmath67 such that @xmath450[{\partial}]$ ] .
by lemma [ premovable ] , we can write @xmath451 where @xmath452 $ ] , @xmath453 in @xmath4 $ ] , @xmath86 and @xmath79
. let @xmath454 and @xmath455 .
then the content @xmath99 of @xmath456 with respect to @xmath15 is @xmath457 because @xmath453 , @xmath458 let @xmath459 then @xmath460 is the primitive part of @xmath456 . in particular
, @xmath460 is @xmath5-primitive . then @xmath461 .
since @xmath462 and @xmath450[{\partial}]$ ] , @xmath40 divides the content of @xmath463 with respect to @xmath15 .
since @xmath40 is a non - unit element of @xmath5 , @xmath463 is not @xmath5-primitive , a contradition to lemma [ lm : gauss ] .
[ example1 ] in the shift ore algebra @xmath276[{\partial}]$ ] , consider a @xmath23-primitive operator @xmath464 which annihilates @xmath465 .
we observe that @xmath278 is a constant factor of @xmath54 . by theorem
[ th : cremovable ] , @xmath278 is non - removable .
in this paper , we determine a basis of a contraction ideal defined by an ore operator in @xmath4[{\partial}]$ ] , and compute a completely desingularized operator whose leading coefficient is minimal in terms of both degree and content .
a more challenging topic is to consider the corresponding problems in the multivariate case .
our algorithms rely heavily on the computation of grbner bases over a principal ideal domain @xmath5 . at present , the computation of grbner bases over @xmath5 is not fully available in a computer algebra system . so the algorithms in this paper are not yet implemented . to improve their efficiency ,
we need to use linear algebra over @xmath5 as much as possible .
i am grateful to my supervisors manuel kauers and ziming li for initiating this research project , stimulating inspirational discussions and helping me revise the paper .
i thank shaoshi chen for suggestions on the proof of lemma [ lm : gauss ] and valuable information on literatures .
m. a. barkatou and s. s. maddah . removing apparent singularities of systems of linear differential equations with rational function coefficients . in _ proc . of issac15
_ , 5360 , new york , ny , usa , 2015 . | ore operators form a common algebraic abstraction of linear ordinary differential and recurrence equations . given an ore operator @xmath0 with polynomial coefficients in @xmath1 , it generates a left ideal @xmath2 in the ore algebra over the field @xmath3 of rational functions .
we present an algorithm for computing a basis of the contraction ideal of @xmath2 in the ore algebra over the ring @xmath4 $ ] of polynomials , where @xmath5 may be either @xmath6 or a domain with @xmath6 as its fraction field .
this algorithm is based on recent work on desingularization for ore operators by chen , jaroschek , kauers and singer . using a basis of the contraction ideal , we compute a completely desingularized operator for @xmath0 whose leading coefficient not only has minimal degree in @xmath1 but also has minimal content .
completely desingularized operators have interesting applications such as certifying integer sequences and checking special cases of a conjecture of krattenthaler . [ algorithms ] |
let @xmath0 be the riemann sphere , and @xmath1 .
a _ circle _ in @xmath2 is either a usual circle in @xmath3 , or a line in @xmath2 .
so the circles in @xmath3 are just the curves @xmath4 , where @xmath5 is a linear fractional function ( with @xmath6 ) .
long ago fatou suggested to study ( jordan ) curves @xmath7 which are invariant under a rational function of degree @xmath8 .
see @xcite for recent progress on this .
the case that @xmath9 lies in a circle @xmath10 is not interesting , because any rational function @xmath11 with @xmath12 leaves @xmath9 invariant , and there are no other rational functions with this property .
motivated by his results on invariant curves in @xcite , alexandre eremenko suggested to investigate the following source of invariant curves , and raised two questions about this family : let @xmath13 be non - constant rational functions , such that @xmath14 , so the curve @xmath15 is invariant under @xmath16 .
assume that @xmath9 is not contained in a circle .
* is it possible that @xmath9 is a jordan curve ?
( @xcite , @xcite ) * is it possible that @xmath17 is injective ?
( @xcite , and special case of @xcite ) note that @xmath15 is contained in a circle if and only if there is a linear fractional function @xmath18 such that @xmath19 . in this case
@xmath20 with @xmath21 .
so the decomposition of @xmath22 over @xmath3 essentially arises from a decomposition over @xmath23 .
there are rational functions @xmath24 whose decompositions do not come from a decomposition over the reals . on the other hand , it is known that decompositions of real polynomials over the complex numbers always arise from real decompositions .
see section [ s : rc ] for more about this .
the purpose of this paper is to give a positive answer to question ( a ) , and a negative answer to a slight weakening of ( b ) .
more precisely , regarding ( a ) , we show : [ t : ell ] for every odd prime @xmath25 there are rational functions @xmath13 both of degree @xmath25 , such that * @xmath14 . *
@xmath26 is injective , so @xmath27 is a jordan curve .
* @xmath27 is not a circle . in order to formulate the next two results ,
we define a weakening of injectivity of rational functions on @xmath23 .
a rational function @xmath28 is said to be _ weakly injective on @xmath23 _ , if there exists @xmath29 which is not a critical point of @xmath30 , and besides @xmath31 there is no @xmath32 with @xmath33 .
a partial answer to question ( b ) is [ t : b ] let @xmath13 be non - constant rational functions , such that @xmath24 .
assume that @xmath30 is weakly injective , and that the curve @xmath34 is not contained in a circle .
then the map @xmath35 is not injective .
a slight variant of this theorem shows that for a fairly large class of rational functions from @xmath36 , each decomposition over @xmath3 arises from a decomposition over @xmath23 .
[ t : rc ] let @xmath13 be non - constant rational functions such that @xmath37 is weakly injective .
then there is a linear fractional function @xmath18 such that @xmath38 . the main ingredient ( besides galois theory ) in the proof of the previous two theorems is the following group - theoretic result .
( see section [ s : gs ] for the notation . )
[ p : gs ] let @xmath39 be a transitive group of permutations of the finite set @xmath40 .
let @xmath41 be a permutation of @xmath40 of order @xmath42 which fixes exactly one element @xmath43 , and which normalizes @xmath39 , that is @xmath44 .
let @xmath45 be the stabilizer of @xmath43 in @xmath39 . then @xmath46 for each group @xmath47 with @xmath48 .
the proof of theorem [ t : ell ] uses elliptic curves .
the construction was motivated by a group - theoretic analysis similar to the one which led to the proofs of theorems [ t : b ] and [ t : rc ] .
[ [ acknowledgment . ] ] acknowledgment .
+ + + + + + + + + + + + + + + i thank alexandre eremenko for inspiring discussions about invariant ( jordan ) curves , and theo grundhfer for a careful reading of the somewhat intricate proof of proposition [ p : gs ] , and his suggestion which simplified its final step .
furthermore , i thank mike zieve for discussions about an example in the final section and for pointing out its connection with work by avanzi and zannier .
in this section we work out our sketch from @xcite and prove theorem [ t : ell ] .
let @xmath49 be an elliptic curve given by a weierstrass equation @xmath50 with @xmath51 .
by @xmath52 and @xmath53 we denote the complex and real points of @xmath49 . for @xmath54 we let @xmath55 be the complex conjugate of @xmath56 .
we use the structure of @xmath52 as an abelian group , with neutral element @xmath57 the unique point at infinity . for general facts about elliptic curves see e.g. @xcite .
let @xmath58 be a prime .
then there is a point @xmath59 of order @xmath25 , with @xmath60 .
let @xmath61\subset e({{\mathbb{c}}})$ ] be the group of @xmath25-torsion points .
then @xmath61 $ ] is isomorphic to the vector space @xmath62 , and the complex conjugation acts linearly on this space .
suppose that the claim does not hold , so the complex conjugation fixes each @xmath63-dimensional subspace of @xmath61 $ ] setwise .
then the complex conjugation acts as a scalar map .
therefore either @xmath61\subset e({{\mathbb r}})$ ] , or @xmath64 for each @xmath65 $ ] . in the latter case
write @xmath66 .
so @xmath67 is real and @xmath68 is purely imaginary .
thus , upon replacing @xmath49 with the twisted curve @xmath69 ( which is isomorphic over @xmath23 to @xmath70 ) , we obtain in either case an elliptic curve @xmath49 with @xmath61\subseteq e({{\mathbb r}})$ ] . on the other hand
, @xmath53 is isomorphic to @xmath71 or to @xmath72 ( see e.g. ( * ? ? ? * v. cor .
2.3.1 ) ) .
however , @xmath73 does not have a subgroup isomorphic to @xmath74 .
this proves the claim .
[ l : w ] suppose that @xmath75 has three distinct real roots .
then there are elements @xmath76 such that there is no @xmath77 with @xmath78 .
if @xmath75 has three distinct real roots , then @xmath79 , so any @xmath80 corresponding to @xmath81 , @xmath82 arbitrary , has the property that there is no @xmath77 with @xmath78 .
( in this case , @xmath53 has two connected components , and for each @xmath77 the element @xmath83 is on the connected component of @xmath57 . ) by an _
automorphism _ of an elliptic curve we mean a birational map of the curve to itself which need not fix the neutral element .
pick @xmath59 of order @xmath25 such that @xmath60 , and set @xmath84 .
let @xmath85 be the isogeny with kernel @xmath86 .
let @xmath87 be the dual isogeny .
then @xmath88 is the multiplication by @xmath25 map on @xmath49 . for @xmath76 as in the previous lemma define involutory automorphisms * @xmath89 of @xmath49 by @xmath90 , * @xmath91 of @xmath92 by @xmath93 , and * @xmath94 of @xmath49 by @xmath95 .
note that @xmath96 , so @xmath97 in the following we show that * @xmath98 , @xmath99 , and @xmath100 are projective lines , and * that there are degree @xmath42 branched covering maps @xmath101 , @xmath102 , and @xmath103 from the elliptic curves to these lines , * such that @xmath101 and @xmath103 are defined over @xmath23 , and * that after selecting uniformizing elements of the projective lines , there are unique rational functions @xmath104 and @xmath30 such that the following diagram commutes : @xmath105{rr}{\text{multiplication by $ \ell$ } } & e ' \arrow{r}{\phi ' } \arrow{d}{\psi ' } & e \arrow{d}{\psi''}\\ e/{\mathopen{<}\beta\mathclose { > } } \arrow{r}{g } & e'/{\mathopen{<}\beta'\mathclose { > } } \arrow{r}{f } & e/{\mathopen{<}\beta''\mathclose { > } } \end{tikzcd}\ ] ] we give an algebraic rather than a geometric description of the functions @xmath104 and @xmath30 .
this has the advantage that the method can be used to compute explicit examples , as we do at the end of this section .
let @xmath106 and @xmath107 be the function fields of @xmath49 and @xmath92 , respectively .
let @xmath108 and @xmath109 be the coordinate functions with @xmath110 and @xmath111 for @xmath112 .
so @xmath113 with @xmath114 . the comorphism @xmath115 is an automorphism of order @xmath42 of the real function field @xmath116 ( recall that @xmath117 ) .
we compute the fixed field of @xmath115 in @xmath116 : write @xmath118 , and set @xmath119 ( this choice of @xmath120 is taken from @xcite ) .
the addition formula for elliptic curves shows that @xmath121 from that we get @xmath122 so @xmath120 is in the fixed field of @xmath115 .
clearly @xmath123 .
let @xmath124 be the fixed field of @xmath115 . from @xmath125 and
@xmath126=2 $ ] we get @xmath127 once we know that @xmath128\le 3 $ ] . but this holds , as @xmath129 so @xmath108 has at most degree @xmath130 over @xmath36 .
now @xmath101 is just the rational function @xmath131 for which @xmath132 .
suppose ( without loss of generality ) that @xmath92 has also a weierstrass from @xmath133 , and let @xmath134 and @xmath135 be the associated coordinate functions .
note that @xmath136 for rational functions @xmath137 and all @xmath112 .
therefore @xmath138 , where @xmath139 is the comorphism of @xmath140 .
pick @xmath141 in @xmath142 such that @xmath143 is the fixed field of @xmath144 .
( here @xmath141 can not be taken from @xmath145 , because @xmath92 is not defined over @xmath23 . ) as before , let @xmath102 be the rational function with @xmath146 .
from we obtain @xmath147 , so @xmath148 and hence @xmath149 .
so @xmath150 for a rational function @xmath151 .
this is just the algebraic description of @xmath30 from above .
similarly one computes @xmath104 . for the rest of this section , we work with the morphisms rather than the comorphisms . recall that @xmath152 and @xmath153 are defined over @xmath23 .
so @xmath154 for all @xmath54 .
we now prove the required properties of @xmath104 and @xmath30 .
the assertion about the degrees follows from well - known facts about isogenies . * as the multiplication by @xmath25 map is defined over @xmath23 , and so are @xmath152 and @xmath153 , we have @xmath155 . *
we next show that @xmath30 is injective on @xmath156 .
suppose that there are distinct @xmath157 such that @xmath158 .
pick @xmath159 such that @xmath160 , @xmath161 .
then @xmath162 so @xmath163 or @xmath164 .
upon possibly replacing @xmath165 with @xmath166 we may and do assume @xmath163 , hence @xmath167 .
+ recall that @xmath152 is defined over @xmath23 and @xmath160 is real .
so @xmath168 , and therefore @xmath169 or @xmath170 .
likewise @xmath171 or @xmath172 .
recall that @xmath167 , and that @xmath173 by the choice of @xmath86 .
so we ca nt have @xmath174 , nor @xmath175 .
+ thus , without loss of generality , @xmath169 and @xmath172 .
so @xmath176 .
note that @xmath177 and @xmath178 both have order @xmath25 .
set @xmath179 .
then @xmath180 , and @xmath181 .
we obtain @xmath182 with @xmath183 , contrary to the choice of @xmath80 . * finally , we need to show that @xmath27 is not a circle .
suppose otherwise .
let @xmath184 be a linear fractional function which maps this circle to @xmath156 .
then @xmath185 maps @xmath23 to @xmath156 , so @xmath186 .
+ then @xmath187 is defined over @xmath23 , so @xmath188 for all @xmath54 . as
@xmath184 is bijective , @xmath189 respects @xmath91 , and @xmath86 is the kernel of @xmath140 , we get that for each @xmath190 either @xmath191 , or @xmath192 . + in the first case note that @xmath193 , so also @xmath194 , and therefore @xmath195 .
so @xmath176 if the first case happens .
+ we see that @xmath192 whenever @xmath196 .
recall that @xmath76 .
so @xmath197 , and therefore @xmath198 for all @xmath196 .
as @xmath199 for all @xmath200 , we get the absurd consequence that @xmath201 whenever @xmath202 , so @xmath53 is a subgroup of index @xmath42 in @xmath52 .
this final contradiction proves all the properties about the functions @xmath104 and @xmath30 . for fixed curves
@xmath49 , @xmath92 and isogeny @xmath140 as in the proof of theorem [ t : ell ] , and @xmath76 ( which need not fulfill the property of lemma [ l : w ] ) , let @xmath203 be the rational function constructed there . the case @xmath204 gives a latts function @xmath205 .
it is easy to see that @xmath206 for linear fractional functions @xmath207 .
if @xmath80 has the property from lemma [ l : w ] , then @xmath208 can not be chosen in @xmath36 .
so @xmath209 is a twist of @xmath210 over a quadratic field .
therefore , the relation of @xmath209 to the latts map @xmath210 is analogous to the relation of rdei functions to cyclic polynomials @xmath211 .
a construction like @xmath209 appeared in an arithmetic context in @xcite .
latts functions , which were known before latts work in 1918 , are classical objects in complex analysis .
see @xcite and @xcite for the relevance of these functions in complex dynamics , and especially @xcite for a lot of information about the history of these functions .
latts functions also appeared in 1877 in the context of approximation theory in work by zolotarev . today
they are called zolotarev functions in approximation theory .
here we explicitly compute an example for the case @xmath212 .
we aim to find an example where the elliptic curve @xmath49 is defined over @xmath213 , @xmath214 , and @xmath215 , where @xmath216 is an as small as possible number field .
let @xmath43 be a primitive third root of unity , so @xmath217 and @xmath218 .
as @xmath59 is required to be a non - real point , and the coordinates of the @xmath25-torsion group of an elliptic curve over @xmath213 generate the field of @xmath25-th roots of unity , we necessarily have @xmath219 . indeed , there are examples with @xmath220 .
the in terms of the conductor smallest elliptic curve @xmath49 over @xmath213 which has a @xmath130-torsion point in @xmath221 has the cremona label 14a2 and weierstrass form @xmath222 .
one computes that @xmath223 has order @xmath130 .
set @xmath84 .
then @xmath173 .
there is an isogeny @xmath224 with kernel @xmath86 , where @xmath92 is given by @xmath225 .
set @xmath226 .
the @xmath227-coordinates of @xmath228 with @xmath229 are roots of @xmath230 , so there is no @xmath77 with @xmath229 .
thus @xmath49 , @xmath86 and @xmath80 fulfill all the assumptions which we needed in the existence proof of @xmath231 and @xmath232 .
we now compute these functions .
let @xmath89 and @xmath91 be the automorphisms of @xmath49 and @xmath92 given by @xmath90 and @xmath93 . write @xmath118 and @xmath233 .
set @xmath119 and @xmath234 .
recall that @xmath138 , where @xmath136 for all @xmath235 . from
that we see also @xmath236 .
now recall that the function @xmath232 we are looking for fulfills @xmath237 .
we compute @xmath238 use this equation , and the equations @xmath119 and @xmath239 to eliminate the variables @xmath108 and @xmath109 .
so we are left with a polynomial equation in @xmath120 and the unknown function @xmath232 which we treat as a variable .
this polynomial has a factor of degree @xmath63 with respect to @xmath232 , from which we obtain @xmath232 .
analogously we get @xmath231 . after minor linear changes over @xmath213 ( which slightly simplify @xmath104 and @xmath30 )
we obtain @xmath240 the following plot shows the image of @xmath156 under @xmath241 . as expected , this curve is a jordan curve , but not a circle .
if @xmath39 acts on a set @xmath40 , then @xmath242 denotes the image of @xmath243 under @xmath244 .
furthermore , @xmath245 is the stabilizer of @xmath43 in @xmath39 . for @xmath246
we write @xmath247 for the conjugate @xmath248 of @xmath30 under @xmath249 . similarly , if @xmath250 is a subset or subgroup of @xmath39 , then @xmath251 .
if @xmath39 is transitive on @xmath40 , then @xmath252 is called a block if @xmath253 or @xmath254 for each @xmath255 . if this is the case , then @xmath40 is a disjoint union of sets @xmath256 for @xmath257 in a subset of @xmath39 .
these sets @xmath258 are called a block system .
note that @xmath39 acts by permuting these sets @xmath258 .
we assume that proposition [ p : gs ] is false , so there is a group @xmath47 with @xmath259 and @xmath260 . among the counterexamples with @xmath261 minimal we pick one with @xmath262 minimal .
in a series of lemmas we derive properties of such a potential counterexample , and eventually we will see that it does not exist .
note that @xmath263 fixes @xmath264 , hence @xmath265 and therefore @xmath266 , a fact we will use frequently . another trivial fact which we use throughout the proof is the following : if @xmath267 is a subgroup of @xmath39 , then @xmath41 normalizes @xmath268 and @xmath269 .
@xmath39 is transitive on @xmath40 .
set @xmath270 .
if @xmath271 then we are done .
so assume that @xmath272 .
if @xmath273 , then @xmath274 , and therefore of course @xmath275 . thus @xmath276 .
note that @xmath277 . by the assumption of a minimal counterexample
, we obtain that the proposition holds for the action of @xmath278 on @xmath279 , hence @xmath280 , a contradiction .
[ l : mms ] @xmath281 . set @xmath282 .
then @xmath283 , so if @xmath284 is a proper subgroup of @xmath39 , then @xmath280 by the minimality assumption of a counterexample . @xmath285 . set @xmath286 .
note that @xmath287 and @xmath288 .
therefore @xmath289 is a block for the action of @xmath278 on @xmath40 , and @xmath290 .
let @xmath291 be the block system which contains @xmath279 , so @xmath278 acts on @xmath291 . by the transitivity of @xmath39 all blocks in @xmath291 have the same size , and this size divides the odd number @xmath262 .
so the blocks have odd size , therefore @xmath41 has a fixed point in each block which is fixed setwise .
thus @xmath279 is the only block fixed by @xmath41 . for @xmath244 let @xmath292 be the induced permutation on @xmath291 .
the stabilizer of @xmath279 in @xmath293 is @xmath294 .
now suppose that @xmath295 , hence @xmath296 and therefore @xmath297 .
note that @xmath298 .
so the proposition applies and yields @xmath299 .
but the kernel of the map @xmath300 is contained in @xmath301 , so @xmath46 , a contradiction .
[ l : crucial ] if @xmath302 , then either @xmath303 , or @xmath304 .
set @xmath305 , and suppose that @xmath306 .
hence the proposition holds for @xmath284 , in particular @xmath307 .
thus @xmath308 . together with the previous lemma we get @xmath309 .
@xmath39 has even order .
suppose that the order of @xmath39 is odd .
pick @xmath310 . note that @xmath311 so @xmath41 acts on @xmath312 by inverting the elements .
as @xmath312 has odd order , there is @xmath313 with @xmath314 .
set @xmath315 .
first note that @xmath316 is fixed under @xmath41 : @xmath317 from this we obtain that @xmath316 permutes the fixed points of @xmath41 , so @xmath318 because @xmath43 is the only fixed point of @xmath41 .
another calculation shows @xmath319 hence @xmath320 contrary to @xmath285 and the choice of @xmath30 .
@xmath47 contains at least one involution which is not contained in @xmath45 . as @xmath321={\lvert\omega\rvert}$ ] is odd , there is a sylow @xmath42subgroup @xmath250 of @xmath39 contained in @xmath45 . by the previous lemma , @xmath322 .
as the number of sylow @xmath42subgroups of @xmath45 is odd and @xmath266 , we may and do pick @xmath250 with @xmath323 .
let @xmath324 be the set of involutions in @xmath250 . then @xmath325 .
suppose that the lemma is false .
then @xmath326 for all @xmath327 . as @xmath328 is a @xmath42group in @xmath45 , there is @xmath329 such that @xmath330 , and therefore @xmath331 .
so for each @xmath327 there is @xmath329 such that @xmath332 , where @xmath333 denotes the normalizer of @xmath324 in @xmath47 .
therefore @xmath334 from @xmath325 we obtain @xmath335 set @xmath336 .
if @xmath306 , then @xmath337 by lemma [ l : crucial ] , and therefore @xmath338 by , a contradiction . therefore @xmath339 . together with we obtain @xmath340
so all of @xmath39 normalizes @xmath324 .
then @xmath341 is a nontrivial normal subgroup of @xmath39 with @xmath342 , so @xmath343 fixes every point in @xmath40 , a contradiction .
we now obtain the final contradiction : let @xmath344 be an involution . set @xmath345 and let @xmath346 be the dihedral group generated by @xmath347 and @xmath348 . from lemma [ l : crucial ] , with @xmath349 ,
we get @xmath350 .
set @xmath351 . then @xmath352=2 $ ] ( because @xmath353 ) .
we claim that @xmath86 is transitive on @xmath40 .
if this were not the case , then , by the transitivity of @xmath39 and @xmath354 , @xmath86 would have exactly two orbits of equal size , so @xmath262 were even , a contradiction .
so @xmath355 , and @xmath356 , because transitive abelian groups act regularly .
the modular law yields @xmath357 and @xmath358 . from @xmath359
we get @xmath360 . but
the subgroups of the cyclic group @xmath86 are determined uniquely by their order , hence @xmath361 and finally @xmath280 .
for the rational function @xmath362 let @xmath363 be the function with complex conjugate coefficients . recall that @xmath27 is a circle in @xmath2 if and only if there is a linear fractional function @xmath18 such that @xmath364 .
the following lemma gives a useful necessary and sufficient criterion for this to hold .
by @xmath365 we mean the field of rational functions in @xmath232 .
[ l : cocycle ] let @xmath362 .
then @xmath366 for some linear fractional function @xmath18 if and only if @xmath367 . if @xmath366 , then @xmath368 , hence @xmath369 , and therefore @xmath370 . to prove the other direction ,
suppose that @xmath371 .
this assumption is preserved upon replacing @xmath30 with @xmath372 for a linear fractional function @xmath373 .
thus , without loss of generality , we may assume that there are @xmath374 with @xmath375 , @xmath376 , @xmath377 .
from @xmath367 we get @xmath378 for a linear fractional function @xmath379 . evaluating in @xmath380 , @xmath381 , and @xmath382 yields that @xmath383
fizes @xmath384 , @xmath385 and @xmath63 , hence @xmath386 .
so @xmath387 , and therefore @xmath28 .
the lemma holds more generally if we replace @xmath23 with a field @xmath216 and @xmath3 with a galois extension @xmath49 of @xmath216 , and @xmath367 by the condition @xmath388 for all @xmath389 . indeed ,
if @xmath216 is an infinite field , then we find @xmath390 such that the values @xmath391 , @xmath392 , and @xmath393 are distinct and therefore without loss of generality equal to @xmath384 , @xmath385 and @xmath63 .
so , as above , @xmath394 for all @xmath389 .
thus the coefficients of @xmath30 are fixed under @xmath395 and therefore contained in @xmath216 .
if @xmath216 is finite , we can argue as follows : we may assume that @xmath396 , so @xmath397 for relatively prime polynomials @xmath398 $ ] with @xmath399 .
in addition , we may assume that @xmath56 and @xmath165 are monic .
let @xmath41 be a generator of the cyclic group @xmath395 . from @xmath400 and @xmath401
we obtain @xmath402 for some @xmath403 .
repeated application of @xmath41 shows that @xmath404 .
so by the additive hilbert s theorem 90 there is @xmath405 with @xmath406 , hence @xmath407 and therefore @xmath408 .
( the same argument , except that hilbert s theorem 90 is a trivial fact for the extension @xmath409 , works as an alternative proof of the lemma too . )
theorem [ t : b ] is a direct consequence of theorem [ t : rc ] . for
if @xmath30 is weakly injective , and @xmath410 is injective on @xmath27 , then @xmath411 is weakly injective , so @xmath412 is weakly injective even more .
thus we only need to prove theorem [ t : rc ] .
let @xmath413 be a variable over @xmath3 , and @xmath414 be another variable over the field @xmath415 of rational functions in @xmath413 . if @xmath416 , then upon replacing @xmath417 with @xmath418 ( and @xmath419 with @xmath420 ) we may assume that @xmath421 . by further replacing @xmath417 with @xmath422 , we may and do assume that @xmath423 , where @xmath424 $ ] are relatively prime polynomials , @xmath425 , and @xmath426 , where the @xmath427 are pairwise distinct , @xmath428 , and @xmath429 for @xmath430 . by hensel
s lemma , @xmath431 , where @xmath432 $ ] has constant term @xmath427 .
as @xmath428 and @xmath433 $ ] , we actually have @xmath434 $ ] .
write @xmath435 .
the complex conjugation acts on the coefficients of the formal laurent series @xmath436 and fixes @xmath413 . under this action
, @xmath120 is fixed , and the @xmath437 s for @xmath430 are flipped in pairs .
note that @xmath438 for all @xmath439 .
@xmath440 is a galois extension of @xmath441 , and @xmath442 .
so the restriction of the complex conjugation action of @xmath436 to @xmath440 is an involution @xmath443 which fixes @xmath435 , and moves all @xmath444 with @xmath445 .
now write @xmath446 as in theorem [ t : rc ] . then also @xmath447 .
this yields the following inclusion of fields and the corresponding subgroups of @xmath448 by galois correspondence : @xmath449 { } & { { \mathbb{c}}}(z_1,\dots , z_n ) \arrow{d } & & & 1 \arrow{d } & \\ & { { \mathbb{c}}}(z ) \arrow{ld } \arrow{rd } & & & g_z \arrow{ld } \arrow{rd } & \\ { { \mathbb{c}}}(g(z ) ) \arrow{rd } & & { { \mathbb{c}}}(\bar g(z ) ) \arrow{ld } & m \arrow{rd } & & m^\sigma \arrow{ld}\\ & { { \mathbb{c}}}(t ) \arrow{d } & & & g \arrow{d } & \\ & { { \mathbb r}}(t ) & & & a & \end{tikzcd}\ ] ] here @xmath450 is the stabilizer of @xmath435 in @xmath39 . as @xmath47 is the stabilizer of @xmath232 in @xmath39 , and @xmath41 maps @xmath232 to @xmath363 ,
the stabilizer of @xmath363 is @xmath451 . by construction
, @xmath440 is the splitting field of @xmath452 over @xmath415 , hence @xmath39 acts faithfully on @xmath453 .
now @xmath39 is normal in @xmath448 , so @xmath454 normalizes @xmath39 .
furthermore , @xmath41 fixes exactly one of the @xmath444 .
so @xmath280 by proposition [ p : gs ] .
thus @xmath367 by the galois correspondence , and finally @xmath455 for some linear fractional @xmath18 by lemma [ l : cocycle ] .
this proves theorem [ t : rc ] .
if @xmath456 for polynomials @xmath457 $ ] , and @xmath458 $ ] , then it is well known that there is a linear polynomial @xmath459 $ ] such that @xmath460 $ ] .
see ( * ? ? ?
* theorem 3.5 ) , or ( * ? ? ?
2.2 ) for a down to earth proof .
a less elementary but more conceptual proof can be based on the fact that the galois group of @xmath461 over @xmath415 contains an element which cyclically permutes the roots of @xmath461 , and the other fact that subgroups of cyclic groups are uniquely determined by their orders .
note that if @xmath456 for a polynomial @xmath249 and rational functions @xmath462 , then there is a linear fractional function @xmath379 such that @xmath463 , and @xmath464 and @xmath465 are polynomials .
( this follows from looking at the fiber @xmath466 . )
so in order to get examples of rational functions @xmath467 which decompose as @xmath456 with @xmath13 such that there is no linear fractional function @xmath18 with @xmath364 , one has to assume that @xmath249 is not a polynomial . assume without loss of generality that @xmath30 is monic .
write @xmath470 with @xmath471 $ ] relatively prime and @xmath56 monic .
from @xmath24 we obtain @xmath472 clearly , both fractions are reduced , and the numerators of both sides are monic .
therefore @xmath473 , hence @xmath474 $ ] , and the claim follows from the polynomial case .
now we give some examples of rational functions @xmath467 with a decomposition @xmath456 with @xmath13 which is not equivalent to a decomposition over @xmath23 .
recall that this is equivalent to @xmath27 not being a circle .
of course , theorem [ t : ell ] gives many example for this .
but these examples are quite complicated and not explicit .
however , if one drops the requirement that @xmath27 is a jordan curve , then there are quite simple examples .
we give two series .
let @xmath475 $ ] be the polynomial with @xmath476 ( so @xmath477 is essentially a chebychev polynomial . ) set @xmath478 for an @xmath479-th root of unity @xmath480 .
then @xmath481 , while @xmath27 is not a circle if @xmath482 .
pick @xmath483 with @xmath484 , and set @xmath485 for @xmath486 .
a straightforward calculation shows that @xmath487 pick @xmath488 with @xmath489 , and set @xmath490 from @xmath491 and we get @xmath492 , hence @xmath24 . on the other hand ,
it is easy to see that , except for a some degenerate cases , @xmath27 is not a circle .
furthermore , we see that @xmath27 is nt even a jordan curve ( unless it is a circle ) , for if @xmath120 runs through @xmath156 , @xmath493 runs through the unit circle , so the numerator and denominator of @xmath494 vanish @xmath495 and @xmath479 times , respectively , so @xmath27 has several self intersections .
originally i had only found the cases @xmath496 , @xmath497 .
mike zieve observed the strong similarity of these examples with functions which turned up in work of avanzi and zannier .
in @xcite they classify triples @xmath498 $ ] , @xmath499 such that @xmath500 .
one of their cases ( ( * ? ? ?
4.7(3 ) ) ) is the above series with @xmath501 , and the series ( * ? ? ?
5.6(4 ) ) is essentially our series from above .
the connection with the work by avanzi and zannier is not a surprise : if we look for polynomials @xmath502 $ ] such that there is @xmath503 with @xmath504 , then @xmath505 with @xmath506 .
furthermore , note that if @xmath480 is an @xmath507-th root of unity , then @xmath508 . so upon setting @xmath509 , we have @xmath510 . robert e. macrae and pierre samuel
. subfields of index @xmath42 of elliptic function fields . in _ conference on commutative algebra ( univ .
kansas , lawrence , kan . , 1972 )
_ , pages 171193 .
lecture notes in math .
, vol . 311 .
springer , berlin , 1973 . | we consider the connection of functional decompositions of rational functions over the real and complex numbers , and a question about curves on a riemann sphere which are invariant under a rational function . |
a particular class of dwarf galaxies named blue compact dwarf galaxies ( bcdgs , sargent & searle , ) has seen increasing interest among astrophysicists because of their extreme current star forming activity which is in contrast to their apparent `` youth '' in terms of chemical evolution .
bcdgs represent about 5% of all dwarfs @xcite , @xcite and are among the smallest star forming galactic systems known .
one of their outstanding properties is that their optical spectra are dominated by lines characteristic of regions , which is the reason why they are frequently termed `` galaxies '' . from optical spectroscopy we know that many galaxies have low heavy element abundances , typically down by a factor of three up to more than twenty compared to the solar neighbourhood @xcite .
it quickly became clear that these objects must form stars in what is called a burst , otherwise the observed star formation rate would be in conflict with their total gas masses as derived from observations @xcite .
such a burst may last some @xmath65yrs , with a time span between bursts of the order of 10@xmath66 yrs .
it has been suggested that interaction with companions might trigger their star formation @xcite , but taylor et al .
found that only about 60% of galaxies have companions , often with masses about @xmath67 of the main galaxy .
one of the most interesting and important issues which has not been settled so far is the molecular gas content of these galaxies .
molecular hydrogen is believed to be the preponderant seed for star formation , so it is a natural assumption that large amounts of h@xmath0 should be present in bcdgs . yet the results have been anything but conclusive so far . following early attempts to detect the co line in bcdgs @xcite , there have been a number of observations to confirm or reject those inconclusive measurements ( e.g. sage et al .
, hereafter sslh ; gondhalekar et al . ; taylor , kobulnicky & skillman , hereafter tks ) .
surprisingly , the results remained partially contradictory , as for instance in the case of ii zw 40 ( arnault et al . ;
sslh ) , although improved instrumentation had been involved . prompted by the difficulty to detect the co line relied upon as a good tracer of molecular hydrogen content in bcdgs
, some of the pertinent publications prematurely concluded that molecular gas is deficient in these systems .
however , part of the difficulties to detect co might have been due to beam filling and sensivitity problems .
taking e.g. 30dor in the lmc as a template giant star - forming complex , it is clear that , if placed at some larger distance and covered by the beam , it could have escaped detection , as co brightness is rather low there , due to strong photo - dissociation in the high interstellar radiation field @xcite .
the same might be true for bcdgs . in this case
, high - sensitivity mapping might reveal previously undetected co emission .
we have therefore conducted a search for co in gas - rich ( based on ) galaxies , using the iram 30 m telescope .
in contrast to previous projects ( e.g. sslh , gondhalekar et al . ) , our observations not only consisted of single pointings towards the brightest position in the galaxies , but involved mapping a number of positions in them , to detect possible gas concentrations away from photodissociation regions . obtaining sensitive upper limits to the co luminosity in these systems
is as much a goal of this study as detecting emission . in section 2
we present details of our observations . in section 3
we present our data , and compare it with previous results . in the subsequent section ( sec .
4 ) we discuss possible causes for the detections and non - detections .
this section is divided into three subsections : section [ gas ] deals with the physical conditions of the gas derived from a lvg model ; in section [ metlum ] we analyze the relationship between metallicity and co luminosity ; finally , in section [ factor ] we discuss the @xmath5 factor problem , which has been heatedly debated in the past years and has not yet found a clear resolution .
our conclusions are presented in the last section ( sec .
the co observations were carried out in october 1996 with the iram 30 m telescope on pico veleta in spain . our target galaxies except haro2 have been selected from the sample examined by @xcite in the 21 cm line of atomic hydrogen .
pointing positions were chosen from peak column densities , as seen in their high angular resolution vla observations .
the basic properties of the galaxies are listed in table [ osservazioni ] . in the @xmath1 transition
( 115 ghz ) the hpbw is 22@xmath68 , whilst in the @xmath2 transition ( 230 ghz ) it is @xmath695 . at the distance of the galaxies , 14.7mpc to 23.3mpc ,
the beam size at 115ghz corresponds to 1.6kpc to 2.5kpc .
two independent sis receivers have been used simultaneously at each frequency .
more than 90% of our observations had @xmath70 600 k both at 230 and at 115 ghz .
an autocorrelator with a spectral resolution of 0.625 mhz at 115 mhz and 1.25 mhz at 230 mhz and a filter spectrometer with 1 mhz resolution were used .
a baseline of zeroth or first order was always subtracted , and the spectra were summed up to improve the signal - to - noise ratio ; the spectra were finally smoothed to roughly the same velocity resolution ( 5.2 for the @xmath1 and 4.8 for the @xmath2 transition ) .
spectra were obtained with a wobbling secondary mirror with a wobbler throw of @xmath71 in azimuth .
all temperatures throughout this article refer to main beam brightness with @xmath72 derived using @xmath73 and @xmath74 .
table [ osservazioni ] gives an overview of the observed galaxies and some of their known properties , and lists results of the galaxy - averaged co spectra .
we mapped most of the galaxies in our sample in order to cover most of the area where emission ( perhaps in `` hot spots '' ) might be present , and obtained low rms noise levels .
nevertheless , we detected co only in haro2 and in um465 .
for all the other galaxies we obtained low upper limits .
it is interesting to note that the smc , with an h@xmath0 mass of @xmath75 would be just detectable with our sensitivity if placed at a distance of 15mpc . in table
[ posizioni ] , all positions observed in all galaxies are listed , with the final rms obtained for single positions . in fig .
[ haromap ] and [ um465map ] , the @xmath1 maps of haro2 and um465 are displayed . in table
[ osservazioni ] we also give a summary of our results . for the detections , we give the parameters of a gaussian fit which we obtained for the lines averaged over all positions . in all other cases , upper limits to the co intensity were calculated , again based on the average of _ all _ positions .
these upper limits were derived as @xmath76 , where @xmath10 is the rms noise level obtained in the baseline range , @xmath77 is the velocity width of each channel and @xmath78 is the number of channels involved .
we always assumed a total line width of 70 , as this is the average velocity width one would expect from the co detections in the literature ( see for example sslh and gondhalekar et al . ) . in the case of detections
, @xmath11 has been calculated based on gaussian fits to the spectra .
based on the @xmath11 values , co luminosities ( @xmath12 ) were calculated .
those for um465 and haro2 are lower limits , because we did not completely map the co gas .
the other values are upper limits for the galaxies .
l c c c c c galaxy & @xmath79 & @xmath80 & @xmath11 & @xmath8 & @xmath9 + & [ @xmath68 ] & [ @xmath68 ] & [ k ] & [ ] & [ ] + haro2 & 00 & 00 & @xmath82 & @xmath83 & @xmath84 + & & & @xmath85 & @xmath86 & @xmath87 + & 10 & 00 & @xmath88 & @xmath89 & @xmath90 + & & & @xmath91 & @xmath92 & @xmath93 + & 00 & 10 & @xmath94 & @xmath95 & @xmath96 + & & & @xmath97 & @xmath98 & @xmath99 + um465 & 08 & 10 & @xmath100 & @xmath101 & @xmath102 + & & & @xmath103 & @xmath104 & @xmath105 + & 02 & 00 & @xmath106 & @xmath107 & @xmath108 + & & & @xmath109 & @xmath110 & @xmath111 + & 18 & 10 & @xmath112 & @xmath113 & @xmath114 + & & & no line + in the following sections we sumarize briefly our main results for the individual galaxies in comparison to previously published data where applicable .
we also give a short description of the galaxies as they appear on the digitized sky survey ( dss ) .
haro 2 is a relatively well studied bcdg .
its metallicity is about @xmath115 solar .
it has the shape and the brightness profiles of an elliptical galaxy , but possesses a brilliant blue nucleus which shows intense star formation .
the absolute blue magnitude is m@xmath116 @xcite .
a comparison between the uv , optical and fir spectra of haro2 with evolutionary population synthesis models has allowed to estimate the age of the youngest star formation episode to be 4 million years , followed by two older bursts , the younger of which was over 20 million years ago .
our co observations confirm the previous detection of co in both the @xmath1 and @xmath2 transitions by sslh and by knapp & rupen in the latter transition .
the emission is clearly extended , as seen in fig .
[ haromap ] , with significant lines in all positions ( see tab . [ fit ] ) .
moreover , fig .
[ haroratio ] suggests that the line in haro2 has two components which are seen at the same velocities in both transitions .
recently , observations with the iram interferometer fully confirmed this finding ( fritz et al . , in prep . ) .
the line ratio of the ( @xmath117 ) to the ( @xmath118 ) line seems to be independent of the velocity ( see fig .
[ haroratio ] ) .
following the path of sslh , we calculate two extreme line ratios , one assuming the source to be point - like , and one assuming a uniformly filled beam . in the latter case ,
since a beam filling factor of 1 is assumed , the ratio of the lines is effectively the ratio of integrated lines .
this is also true if the filling factor is @xmath119 , but equal for both transitions ( i.e. in the presence of large scale clumping ) . for a point source
the other extreme is considered : one assumes that the ( 2 - 1)/(1 - 0 ) intensity ratio is overestimated by exactly the ratio of the two beam areas , so that the maximum line ratio must be divided by the ratio of the squares of the two beam widths ( @xmath120 ) .
the results are shown in table [ ratios ] and range from 0.5 to 1.5 .
the differences between our ratios and those of sslh might be partly due to different main beam efficiencies used for the two transitions ( not specified by sslh ) . the dominant region of this dwarf galaxy is embedded in an extended faint irregular stellar body .
we have obtained 5 independent spectra ; none of them shows significant co emission .
um422 has also been observed with a single pointing by tks using the nrao 12 m telescope .
we confirm their non - detection with a significantly lower upper limit . in the optical
, um439 has a slightly elongated compact appearance with one prominent region south of the center .
high resolution observations by van zee , skillman & salzer reveal that the star formation is taking place in the peak of the extended gas distribution .
it was observed in co by sslh and by gondhalekar et al . .
we confirm their non - detections .
our upper limit to @xmath11 is a factor of 2.4 lower than that of sslh obtained with the nrao 12 m telescope and a factor of 2.5 lower than that obtained with the onsala 20 m telescope . because we additionally observed 4 independent positions with higher angular resolution than those just mentioned , our upper limit to @xmath12 is significantly lower than those previously published .
the stellar component of this galaxy , as detected in optical imaging , is very faint .
we have observed it only in the central position .
our upper limit for co is the first one ever published .
we have observed this galaxy towards one position only , where the optical emission is strongest .
the optical extent of the galaxy looks much smaller as compared to the .
the mass of is quite low , @xmath121 @xcite .
the star forming regions in this galaxy are confined to the center of an extended and distorted stellar component .
taylor et al .
have detected two companions of um456 .
um456 a seems to be a pure `` cloud '' with no optical counterpart , whereas um456 b is seen both in and on optical images .
both companions seem to be gravitationally bound to um456 .
none of them shows co ; with our better rms we do not confirm the `` marginal detection '' of um456 by sslh ; our upper limit is a factor of 6 lower than their claimed detection .
the two bcdgs um461 and um462 seem to form a bound system with a linear separation of about 70 kpc at a distance of 13.9 mpc .
two centrally located knots of star formation dominate the optical image of um462 .
they are associated with the peak of the column density @xcite .
the claimed detection of co in um462 by sslh could not be confirmed ; our upper limit is a factor of 2 lower .
this galaxy had also been observed by gondhalekar et al . with a higher upper limit obtained at lower angular resolution .
the optical appearance of this dwarf galaxy is circular in shape with an exponential law brightness distribution @xcite .
the nuclear starburst and extended dust lanes and patches are well resolved by hst imaging @xcite .
a faint nearby object was not detected in @xcite but confirmed as a physical neighbour @xcite using 6 m telescope spectroscopy .
the hst images of this companion reveal an irregular structure .
while sslh reported a marginal detection in um465 of co , the present work delivers a clear one .
the co emission is extended in this galaxy ( see fig . [ um465map ] ) , but with a lower intensity than haro2 . only in one of the three positions no @xmath2 line was detected .
the calculated line ratios are listed in table [ ratios ] , and for this galaxy they range between 0.4 and 1.3 .
c c c c source & point source & uniform filling & reference + & @xmath122 & @xmath122 & + & & & + haro2 & [email protected] & [email protected] & sslh + & [email protected] & [email protected] & this work@xmath53 + um465 & [email protected] & [email protected] & sslh + & [email protected] & [email protected] & this work +
our deep observations corroborate the difficulty to detect co in bcdgs . only two of the galaxies observed
show significant co emission ; in both cases it is extended in both transitions .
we note here that these two sources are those with the highest metallicity in our sample . in the other sources ,
even with observations towards several positions , we were unable to find co emission .
one could be tempted to say that these galaxies are void of molecular gas , but this conclusion is premature because the relationship of co emission and h@xmath0 content in a galaxy depends on many factors ( maloney & black , israel ) , some of which are not fully understood .
therefore , one can only conclude from the detection of co that h@xmath0 is present , whereas the absence of co does not necessarily imply a lack of h@xmath0 .
as a first step in our analysis , we try to infer some information on the physical properties of the co - emitting gas in bcdgs .
we make use of a large velocity gradient ( lvg ) model to predict ratios for the lowest co lines .
the basic lvg assumption is that of a systematic monotonic velocity gradient , which allows to treat the molecular excitation as a local problem ( see de jong , chu & dalgarno , and white for details ) .
this is certainly an idealization of extra - galactic cloud complexes ; however , an lvg code does not require detailed knowledge of the velocity field .
as a first rough estimate this model further assumes constant density and kinetic temperature within the molecular cloud . in fig .
[ lvg ] , we show an example of the dependence of line ratios on @xmath124(h@xmath0 ) and @xmath125 . for this figure , @xmath126}\over { |\nabla \vec{v}|}}=2\cdot 10^{-5}$ ] ( pc@xmath127 has been assumed .
this corresponds to a velocity gradient of 1 pc@xmath128 and an abundance of [ co / h@xmath0]@xmath130 or , due to how the lvg code is constructed , to a velocity gradient of 5 pc@xmath128 and [ co / h@xmath0]=10@xmath131 .
the figure shows the ratios of the intensities of the @xmath132 to @xmath133 transitions , @xmath134@xmath135 , and that of @xmath136 and the @xmath132 lines , @xmath134@xmath137 .
as discussed above , for haro2 and um465 the line ratios are @xmath138 and @xmath139 , respectively , depending on the filling of the sources in our beam . from our limited mapping of the two galaxies ,
we know that the sources are extended and the `` point source '' limit can be firmly excluded . on the other hand , observations of haro2 obtained with the plateu de bure interferometer ( fritz et al .
, in prep . )
show that the galaxy does not fill the beam of the iram 30 m telescope uniformly .
it is then reasonable to expect a value for the line intensity ratios in between those listed in table [ ratios ] , and may thus be close to unity . adopting a ratio @xmath140 between 0.8 and 1.0 , we can derive from fig .
[ lvg ] that the gas is either at high temperatures with medium densities ( @xmath141 a few hundred @xmath142 ) or at high densities ( @xmath143@xmath142 ) and low temperatures ( @xmath144k ) .
more stringent limits on volume density and kinetic temperatures require observations of higher @xmath64co transitions and/or @xmath145co transitions .
these transitions are expected to be very weak though , and thus difficult and time - consuming .
we subsequently examine the relation of the co emission and metallicity . because only the two galaxies in our sample with the highest metallicities are detected in co
, one could expect that co luminosity depends on metallicity .
therefore , we plot our co luminosities ( @xmath12 ) listed in table [ osservazioni ] vs. the metallicities of the galaxies of our sample .
since not all of the galaxies have known metallicities , we are left with only 6 galaxies .
these are shown in fig .
[ lumino ] . in this figure
we also include the data points given by tks because they represent the most comprehensive sample of co observations of dwarf galaxies with metallicity determinations .
we only selected those dwarf galaxies from the sample with metallicities better determined than 0.1 dex ; these galaxies are , however , not necessarily classical bcdgs .
we have chosen the co luminosity because it is largely independent of distance , although the error in the distance determination enters as the square in the co luminosity .
furthermore , the co luminosity is directly proportional to the h@xmath0 mass corrected for helium ( @xmath146 ) , with @xmath147 in and @xmath12 in kpc@xmath148 .
@xmath5 ( in units of @xmath149molecules @xmath150(k)@xmath128 ) is the well - known but poorly determined @xmath5 factor which relates the molecular hydrogen column density to the observed integrated co line intensity ( @xmath151 ) .
[ lumino ] shows a general trend that galaxies with higher metallicities have higher co luminosities , although no functional correlation is visible .
tks have recently proposed that galaxies with metallicities below 7.9 are basically undetectable in co. our data do not contradict this finding . for galaxies close to that limit we were only able to derive upper limits .
we note , however , that even above that limit there are galaxies not detected in co. although it is qualitatively expected that a higher metallicity leads to a higher @xmath12 because of the availability of the building blocks of the co molecule , fig .
[ lumino ] shows that the oxygen abundance can not be the only factor influencing the co luminosity . because bcdgs are actively star - forming galaxies
, the uv radiation field may be locally high .
this plays two conflicting roles for co : on the one hand , it heats the gas , so that the excitation temperature is higher and co is brighter ; on the other hand , if hard enough , it destroys co via photodissociation , thus the co emission becomes weaker .
pak et al . and bolatto , jackson & ingalls
have studied these effects and found that the co emitting regions are effectively smaller in low - metallicity environments and most of the carbon is present in atomic form .
the net effect in galaxies with low metallicities is that , due to low beam filling of the clouds , the molecular gas becomes invisible in the co lines and might be better traced in the fine structure lines of and .
this different beam filling might be the cause for the dependence of the integrated co line intensity @xmath11 on the metallicity found by tks . to study the effect of the radiation field on the co luminosity , we labelled the values in the co luminosity - metallicity plane with the absolute blue magnitudes of the galaxies ( right side of fig .
[ lumino ] ) .
again there is no simple relation between any two of these quantities .
however , there is evidence for an influence of both metallicity and absolute blue magnitude on the co luminosity , meaning that higher blue magnitudes lead to higher co luminosity for a given metallicity .
it also appears that at lower metallicities a higher absolute blue magnitude is necessary to reach a certain co luminosity .
clearly , more well - observed galaxies are necessary to study the relation between the three quantities . directly linked to the question of the dependence of the co luminosity on metallicity and radiation field
is the question of which @xmath5 factor is applicable to low - metallicity galaxies .
a number of studies have examined possible correlations between @xmath5 and the metallicity ( e.g. dettmar & heithausen , wilson , verter & hodge , arimoto , sofue & tsujimoto ) .
klein has proposed an additional dependence of @xmath5 on the cosmic ray flux as judged from the radio continuum brightness .
an independent determination of the @xmath5 factor for our bcdg sample would be useful to determine their molecular masses ; this is , however , beyond the scope of this paper .
a reliable value for @xmath5 has been established for the disk of the milky way , @xmath152 .
the currently best accepted value is @xmath153 molecules @xmath150 k@xmath128 @xcite .
@xmath5 factors determined for galaxies with lower metallicities are usually significantly higher ( e.g. cohen et al . , dettmar & heithausen ) .
l l l l source & @xmath5 & @xmath154h@xmath0 ) & @xmath147 ( ) + & & @xmath65 & @xmath65 + um422 & 18.6 & @xmath155 & 26 + um439 & 18.6 & @xmath156 & 3.5 + um456 & 25.7 & @xmath157 & 3.7 + um462 & 21.4 & @xmath158 & 2.9 + um465 & 5.4 & 0.3 & 0.39 + haro2 & 7.9 & 4.7 & 4.8 + in the following , we assume that the correlation between @xmath5 , derived from a virialization analysis of several galaxies , and the oxygen abundance obtained by arimoto et al . gives @xmath5 factors applicable to our galaxy sample .
h@xmath0 masses derived under this assumption
are listed in table [ xfactors ] . also given
are the masses as derived by @xcite for the um galaxies and by sslh for haro2 .
one remarkable result from this calculation is that those galaxies undetected in co have upper limits on the molecular gas mass significantly below the mass , whereas in the two galaxies where co is detected , and h@xmath0 masses are about the same .
the @xmath5 factors used here to derive molecular gas masses are _ global _ factors . studying the co emission of the magellanic clouds with different angular resolutions rubio ,
lequeux & boulanger noted that the derived @xmath5 factor depends on the linear resolution , implying that the _ local _
@xmath5 value is lower than the _ global _ one .
in order to calculate the _ local _ @xmath5 we use an lvg approximation assuming that h@xmath0 and co share the same volume .
the lvg approximation can then be used to derive the h@xmath0 column density , @xmath159 , from @xmath160 , where @xmath161 is the h@xmath0 volume density , @xmath9 is the line width and @xmath162 is the velocity gradient . since @xmath163 , one can derive @xmath5 from @xmath164 .
the dependence of the local @xmath5 on varying volume density and kinetic temperature is shown in fig .
[ xfile ] .
the velocity gradient is fixed to 2 pc@xmath128 . in the top panel ,
the abundance is [ co / h@xmath0]=10@xmath131 ; in the bottom panel [ co / h@xmath0]= 10@xmath165 .
these are extreme values for the metallicities of bcdgs .
@xmath152 is indicated by the thick solid line .
we note that , once the density becomes high enough , @xmath5 will not change significantly any more , due to the growth of the optical depth , and @xmath5 is independent of the abundance . on the other hand , at lower densities and kinetic temperatures , @xmath5 changes significantly depending on @xmath124(h@xmath0 ) and @xmath125 for a given metallicity . _ these simple calculations indicate that @xmath5 does not only depend on the metallicity of a galaxy . _
physical parameters , such as average volume density and kinetic temperature , play important roles , if the density is below @xmath166 @xmath142 and/or the kinetic temperature is below 50k , values found in many molecular clouds in the milky way .
the @xmath5-factor , that we find for a standard density of @xmath167@xmath142 and a low - metallicity environment , is low , i.e. close to galactic . in contrast
, @xmath5 derived from the formula of arimoto et al .
is higher by an order of magnitude .
this supports the concept of large amounts of hidden gas - either atomic or molecular .
interferometric observations of bcdgs might help to resolve this issue .
such a study of the nearby post - starburst dwarf ngc1569 , a galaxy that may be considered as a nearby bcdg in a post - starburst phase , yielded a rather high value of @xmath168 , based on virial masses of resolved gmcs @xcite .
this indicates that this method , which is also the basis of arimoto s formula , is sensitive to the ` hidden ' h@xmath0 and tends to yield global values for @xmath5 .
in contrast , we expect to find local values if line ratio studies encompassing several co isotopomers become available , since these studies directly probe the physics of the gas from which the co emission arises .
we have searched for emission from the @xmath64co ( @xmath1 and @xmath2 ) transitions in 10 dwarf galaxies , 8 of which are bcdgs and 2 are the companions of one of these .
we detected co in 2 of them ( haro2 and um465 ) and found it to be extended in both galaxies .
although we mapped part of the other galaxies , we were unable to detected co. we obtained very stringent upper limits .
we could not confirm the `` marginal detection '' of co in um456 and um462 previously reported by sslh
. the observed line ratios of the @xmath132 to @xmath133 transitions are not very sensitive to changes in the kinetic temperature .
modelling the ratio with a simple lvg code helps only to exclude low densities .
higher co transitions and/or observations of co isotopomers will help to get more stringent limits on these physical parameters .
we could not find any simple relation between metallicity and co luminosity .
molecular gas masses for the galaxies are derived assuming the relation between @xmath5 and metallicity given by @xcite .
we find that for those galaxies detected in the co lines the molecular gas mass is comparable to the mass , whereas for those galaxies undetected in co the mass is significantly larger than the limits on the molecular gas mass . even in the sources where co has not been detected , we do not argue against the presence of h@xmath0 . while it is certainly possible that in the extreme environment of a bcdg not just co but also h@xmath0 is destroyed , at least in regions close to young massive stellar clusters , a picture in which a large amount of h@xmath0 exists without co is attractive .
sensitive observations of and in these galaxies would thus be desirable in the future to shed light on this issue .
l.t.b . would like to thank prof .
loretta gregorini and the socrates / erasmus project which made this exchange possible and financed it , the faculty of science of the university of bologna and the c.n.a.a .
( _ consorzio nazionale per lastronomia e lastrofisica _ ) for grants supporting this work .
this project was supported by the deutsche forschungsgemeinschaft via the graduiertenkolleg _ `` the magellanic clouds and other dwarf galaxies '' . _ | blue compact dwarf galaxies ( bcdgs ) are currently undergoing strong bursts of star formation .
nevertheless , only a few of them have been clearly detected in co , which is thought to trace the `` fuel '' of star formation : h@xmath0 . in this paper
we present a deep search for co @xmath1 and @xmath2 emission lines in a sample of 8 bcdgs and two companions . only 2 of them ( haro2 and um465 ) are detected . for the other galaxies we have obtained more stringent upper limits on the co luminosity than published values .
we could not confirm the previously reported `` detection '' of co for the galaxies um456 and um462 . we analyze a possible relation between metallicity , co luminosity , and absolute blue magnitude of the galaxies .
we use previously determined relations between @xmath3(h@xmath4 and the metallicity to derive molecular cloud masses or upper limits for them . with these `` global '' @xmath5-values
we find that for those galaxies which we detect in co , the molecular gas mass is similar to the mass , whereas for the non - detections , the upper limits on the molecular gas masses are significantly lower than the mass .
using an lvg ( large velocity gradient ) model we show that @xmath5 depends not only on metallicity but also on other physical parameters such as , volume density and kinetic temperature , which rises the question on the validity of `` global '' @xmath5-factors .
= 3000 = = = = = = = = # 1 # 1 # 1 # 1 @mathgroup@group @mathgroup@normal@groupeurmn @mathgroup@bold@groupeurbn @mathgroup@group @mathgroup@normal@groupmsamn @mathgroup@bold@groupmsamn = `` 019 = ' ' 016 = `` 040 = ' ' 336 = " 33e = = = = = = = = # 1 # 1 # 1 # 1 = = = = = = = = # 1 latexl-.36em.3ex-.15em t-.1667em.7ex-.125emx [ section ] [ firstpage ] galaxies : dwarf galaxies : ism galaxies : starburst radio lines : ism l c c c c c c c c c c galaxy & @xmath6 & @xmath7 & distance & abundance & @xmath8 & @xmath9 & rms ( 1 @xmath10 ) & @xmath11 & @xmath12 + & ( 1950 ) & ( 1950 ) & mpc & 12+logo / h & & & & & + & 11 17 37 & + 02 48 16 & 21.3 & @xmath13 & 1600 & & 3.6 & @xmath14 & @xmath15 + & & & & & 1600 & & 7.5 & @xmath16 & @xmath17 + & 11 34 02 & + 01 05 38 & 14.7 & @xmath13 & 1097 & & 4.6 & @xmath18 & @xmath19 + & & & & & 1097 & & 11.6 & @xmath20 & @xmath21 + & 11 39 02 & 01 37 26 & 24.0 & & 1792 & & 21.9 & @xmath22 & @xmath23 + & & & & & 1792 & & 44.7 & @xmath24 & @xmath25 + & 11 44 26 & 00 00 57 & 19.2 & & 1439 & & 7.7 & @xmath26 & @xmath27 + & & & & & 1439 & & 11.6 & @xmath20 & @xmath28 + & 11 48 01 & 00 17 23 & 23.3 & @xmath29 & 1749 & & 4.3 & @xmath30 & @xmath31 + & & & & & 1749 & & 8.7 & @xmath32 & @xmath33 + & 11 48 00 & 00 15 30 & 23.3 & & 1749 & & 14.5 & @xmath34 & @xmath35 + & & & & & 1749 & & 27.1 & @xmath36 & @xmath37 + & 11 47 53 & 00 17 00 & 23.3 & & 1749 & & 9.1 & @xmath38 & @xmath39 + & & & & & 1749 & & 18.2 & @xmath40 & @xmath41 + & 11 50 13 & 02 11 26 & 13.9 & @xmath42 & 1051 & & 7.8 & @xmath43 & @xmath44 + & & & & & 1051 & & 14.4 & @xmath45 & @xmath26 + & 11 51 38 & + 00 24 56 & 15.4 & @xmath46 & 1144 & 17@xmath47 & 7.8 & 0.64@xmath48 & 3.12@xmath49 + & & & & & 1144 & 35@xmath50 & 18.8 & 1.47@xmath51 & 2.91@xmath52 + @xmath53 & 10 29 22 & + 54 39 24 & 20.5 & @xmath54 & 1401 & 19@xmath55 & 9.3 & 4.2@xmath56 & 36.7@xmath57 + & & & & & 1452 & 70.0 & & + & & & & & 1401 & 18@xmath58 & 14.0 & 6.44@xmath59 & 22.7@xmath60 + & & & & & 1452 & 70.0 & & + [ osservazioni ] l cccc source & @xmath61 & @xmath62 & transition & rms + & [ @xmath63 & [ @xmath63 & & [ mk ] + um422 & 0 & 0 & @xmath64co(1 - 0 ) & 7 + & & & @xmath64co(2 - 1 ) & 18 + & 20 & 20 & @xmath64co(1 - 0 ) & 7 + & & & @xmath64co(2 - 1 ) & 13 + & 20 , & 0 & @xmath64co(1 - 0 ) & 8 + & & & @xmath64co(2 - 1 ) & 22 + & 20 & 0 & @xmath64co(1 - 0 ) & 4 + & & & @xmath64co(2 - 1 ) & 4 + & 0 & 20 & @xmath64co(1 - 0 ) & 3 + & & & @xmath64co(2 - 1 ) & 9 + um439 & 0 & 0 & @xmath64co(1 - 0 ) & 5 + & & & @xmath64co(2 - 1 ) & 18 + & 10 & 20 & @xmath64co(1 - 0 ) & 5 + & & & @xmath64co(2 - 1 ) & 13 + & 10 & 20 & @xmath64co(1 - 0 ) & 7 + & & & @xmath64co(2 - 1 ) & 20 + & 10 & 0 & @xmath64co(1 - 0 ) & 7 + & & & @xmath64co(2 - 1 ) & 9 + um446 & 0 & 0 & @xmath64co(1 - 0 ) & 16 + & & & @xmath64co(2 - 1 ) & 22 + um452 & 0 & 5 & @xmath64co(1 - 0 ) & 5 + & & & @xmath64co(2 - 1 ) & 7 + um456 & 0 & 0 & @xmath64co(1 - 0 ) & 8 + & & & @xmath64co(2 - 1 ) & 13 + & 10 & 10 & @xmath64co(1 - 0 ) & 5 + & & & @xmath64co(2 - 1 ) & 9 + & 10 & 0 & @xmath64co(1 - 0 ) & 4 + & & & @xmath64co(2 - 1 ) & 9 + & 20 & 20 & @xmath64co(1 - 0 ) & 7 + & & & @xmath64co(2 - 1 ) & 11 + um456 a & 5&5 & @xmath64co(1 - 0 ) & 11 + & & & @xmath64co(2 - 1 ) & 13 + um456 b & 0 & 0 & @xmath64co(1 - 0 ) & 7 + & & & @xmath64co(2 - 1 ) & 11 + um462 & 10 & 5 & @xmath64co(1 - 0 ) & 5 + & & & @xmath64co(2 - 1 ) & 11 + um465 & 2 & 0 & @xmath64co(1 - 0 ) & 4 + & & & @xmath64co(2 - 1 ) & 8 + & 18 & 10 & @xmath64co(1 - 0 ) & 4 + & & & @xmath64co(2 - 1 ) & 9 + & 8 & 10 & @xmath64co(1 - 0 ) & 7 + & & & @xmath64co(2 - 1 ) & 11 + haro2 & 0 & 0 & @xmath64co(1 - 0 ) & 18 + & & & @xmath64co(2 - 1 ) & 24 + & 10 & 0 & @xmath64co(1 - 0 ) & 20 + & & & @xmath64co(2 - 1 ) & 31 + & 0 & 10 & @xmath64co(1 - 0 ) & 19 + & & & @xmath64co(2 - 1 ) & 33 + |
for a long time electromagnetically induced two - nucleon knockout has been considered the most direct tool for exploring the properties of nucleon pairs within nuclei and their interaction at short distance @xcite .
the cross section of an exclusive reaction contains the two - hole spectral function , which gives information on nuclear structure and correlations .
two nucleons can be naturally ejected by two - body currents , which effectively take into account the influence of subnuclear degrees of freedom like mesons and isobars . direct insight into dynamical short - range correlations ( src )
can be obtained from the process where the real or virtual photon hits , through a one - body current , either nucleon of a correlated pair and both nucleons are then ejected from the nucleus . a reliable and consistent treatment of these two competing processes , both produced by the exchange of mesons between nucleons , is needed .
their role and relevance , however , can be different in different reactions and kinematics .
it is thus possible to envisage situations where either process is dominant and various specific effects can be disentangled and separately investigated . first pioneering ( @xmath7 ) experiments were performed in bonn @xcite and tokio @xcite with poor statistical accuracy and energy resolution .
further experiments carried out at mami achieved much better results @xcite .
these experiments , with an energy resolution of about 6 - 9 mev , were able to resolve the major shells from which the two nucleons are emitted .
the first high - resolution @xmath1o(@xmath2)@xmath4n experiment able to separate final states was performed at maxlab in lund @xcite .
a high - resolution experiment for the @xmath1o(@xmath2)@xmath4 and @xmath1o(@xmath3)@xmath4c reactions in the photon - energy range between 90 and 270 mev , aiming not only at separating final states but also at determining the momentum distributions of each state , has been recently approved in mainz @xcite .
results for a first test and feasibility study are presented in ref .
@xcite .
reaction calculations performed with different microscopic models based on a two - nucleon knockout framework @xcite confirm that ( @xmath2 ) and ( @xmath3 ) reactions are generally dominated by two - body currents .
thus , they are useful to study the nuclear current , whose various components can be emphasized or suppressed in different conditions and kinematics .
however , they are also sensitive to the other theoretical ingredients and represent a suitable tool to study the conditions of nucleon pairs within a nucleus .
in fact , the angular distributions of the @xmath1o(@xmath2)@xmath4n and @xmath1o(@xmath3)@xmath4c cross sections for transitions to different final states have a different shape , essentially determined by the components of the center - of - mass ( c.m . )
orbital angular momentum of the initial pair @xcite , which can be explored by comparing the theoretical predictions with the experimental data .
interesting and complementary information is available from ( @xmath8 ) and ( @xmath7 ) reactions .
first measurements of the exclusive @xmath1o(@xmath9)@xmath4c reaction have been performed at nikhef in amsterdam @xcite and mami in mainz @xcite .
investigations on these data @xcite indicate that resolution of discrete final states provides an interesting tool to disentangle the two reaction mechanisms due to one - body currents , and thus to src , and to two - body currents .
recent calculations @xcite have shown that also the cross sections of the exclusive @xmath1o(@xmath10)@xmath4n reaction are sensitive to details of the nuclear correlations and in particular to the presence of the tensor component .
it is clear , however , that ( @xmath10 ) experiments are more difficult , as they require a triple coincidence measurement with the detection of a neutron .
thus , no data are available at present , but a proposal for the experimental study of the exclusive @xmath1o(@xmath10)@xmath4n reaction has been approved in mainz @xcite .
good opportunities for a more complete investigation of the properties of different pairs of nucleons in nuclei and of their interaction at short distance are offered by polarization measurements .
reactions with polarized particles give access to a larger number of observables , hidden in the unpolarized case and whose determination can impose more severe constraints to theoretical models . some of these observables are expected to be sensitive to the small components of the transition amplitudes , which are generally masked by the dominant ones in the unpolarized cross section and that often contain interesting information on subtle effects . in photoreactions , and in particular in ( @xmath2 ) reactions , the two - body contributions are always dominant in the cross section with respect to one - body ones . in order to study correlations , the polarization measurements are therefore necessary .
the asymmetry of the cross section for linearly polarized photons in ( @xmath3 ) and ( @xmath2 ) reactions was studied in refs .
@xcite . numerical results for exclusive @xmath6 and @xmath5 knockout from @xmath1o can be found in ref .
first measurements with polarized photon beams have been performed at legs on @xmath11he @xcite and @xmath1o @xcite and at mami on @xmath12c @xcite . in these first experiments , however , the energy resolution was not enough to separate specific discrete final states of the residual nucleus .
the nucleon recoil polarization in photon - induced and electron - induced two - nucleon knockout has been considered in refs . @xcite and @xcite , respectively .
the general formalism of the ( @xmath13 ) reaction and some theoretical predictions for the exclusive @xmath1o(@xmath14)@xmath4c knockout reaction are given in ref .
the results of these investigations indicate that a combined measurement of cross sections and polarization components would provide an interesting tool to disentangle the different reaction mechanisms and clarify the behaviour of nucleon pairs in the nuclear medium .
no data are available at present , but a measurement of the nucleon recoil polarization seems reasonably within reach of available experimental facilities . the general case of ( @xmath0 ) reactions , where the incident photon is linearly or circularly polarized and/or the outgoing nucleon polarization is detected is considered in this paper . besides the cross section , 15 new observables are obtained in the most general kinematics . the formalism and the polarization observables
are given in sect .
[ sec:2 ] .
some numerical results for the exclusive @xmath1o(@xmath15)@xmath4n@xmath16 and @xmath1o(@xmath17)@xmath4c@xmath16 reactions are presented and discussed in sects .
[ sec:3 ] and [ sec:4 ] , respectively .
conclusions are drawn in sect .
[ sec:5 ] .
the cross section of a reaction , where an incident photon , with momentum @xmath19 and energy @xmath20 , is absorbed by the nucleus and two nucleons are emitted , with momenta @xmath21 and @xmath22 and energies @xmath23 and @xmath24 , can be written in terms of four structure functions @xmath25 as @xmath26 , \label{eq : cs}\end{aligned}\ ] ] where @xmath27 and @xmath28 are the degrees of circular and linear polarization of the photon , @xmath29 is the angle of the photon polarization vector relative to the scattering plane , _
i.e. _ the ( @xmath30 ) plane , and @xmath31 with @xmath32 where @xmath33 and @xmath34 are the momentum and energy of the residual nucleus . in our reference frame
the @xmath35 axis is taken parallel to @xmath19 and the momentum @xmath21 lies in the ( @xmath36 ) plane .
the structure functions @xmath25 are obtained from the components of the hadron tensor @xmath37 , with @xmath38 , and are given by all the symmetrical and antisymmetrical independent combinations of these components , _
i.e. _ @xmath39 if the polarization of the nucleon 1 outgoing in the scattering plane is detected , the structure functions can in general be written as @xmath40 where @xmath41 is the unit vector in the spin direction of the recoil nucleon .
thus , when the polarization of the outgoing nucleon is considered , 16 structure functions appear : 4 spin independent , @xmath42 and @xmath43 , and 12 spin dependent , @xmath44 and @xmath45 , 4 for each one of the 3 components of @xmath46 and @xmath47 .
the explicit expressions of these structure functions in terms of the components of the hadron tensor @xmath48 , where @xmath49 and @xmath50 are the eigenvalues of the spin of the nucleon whose polarization is considered , can be easily obtained from eq .
( [ eq : fll ] ) , by simply substituting the quantities @xmath51 , wherever they appear , with the following expressions @xcite : @xmath52 note that in ref .
@xcite there is a misprint in the case @xmath53 .
usually , the quantities @xmath46 and @xmath47 are projected onto the basis of unit vectors given by @xmath54 ( parallel to @xmath55 ) , @xmath56 ( in the direction of @xmath57 ) and @xmath58 , and the structure functions are thus given for the components @xmath59 .
when the outgoing nucleon polarization is not detected , the cross section is summed over the spin quantum numbers of the outgoing nucleon and the spin independent structure functions @xmath42 go over to the structure functions @xmath25 of the unpolarized case @xcite .
thus , in this case only four structure functions are obtained , @xmath60 , and four independent quantities can be measured : the unpolarized cross section @xmath61 the circular asymmetry @xmath62 where @xmath63 and @xmath64 are the cross sections for a circularly polarized photon and @xmath28 = 0 , and the two independent linear asymmetries , @xmath65 and @xmath66 where @xmath67 , @xmath68 and @xmath69 are the cross sections for a linearly polarized photon in the directions @xmath70 , and @xmath71 , respectively , and @xmath27 = 0 . when the polarization of the outgoing nucleon is detected , the 12 new structure functions @xmath44 and 12 corresponding polarization observables , 4 for each direction @xmath72 , @xmath73 , and @xmath74 , are produced .
for an unpolarized photon , we have the components of the polarization @xmath75 where @xmath76 ( @xmath77 ) are the cross sections with a nucleon polarized parallel ( antiparallel ) to the component @xmath78 of the spin direction given in eq .
( [ eq : hll ] ) .
if the photon is circularly polarized and @xmath28 = 0 , we have the components of the polarization transfer coefficient @xmath79 where , here and in the following , the first argument in @xmath80 refers to the photon polarization and the second to the nucleon polarization in the direction @xmath78 . for a linearly polarized photon and @xmath27 = 0
, we have the components of the two polarization asymmetries @xmath81 and @xmath82 where the first argument in @xmath80 refers to the value of the angle @xmath29 . in the case of a coplanar kinematics , where both the emitted nucleons and the photon lie in the same plane , owing to parity conservation combined with the general properties of the hadron tensor , several structure functions vanish and only 8 of them survive : @xmath83 , @xmath84 , @xmath85 , @xmath86 , @xmath87 and @xmath88 and thus the quantities @xmath89 , @xmath90 , @xmath91 , @xmath92 , @xmath93 , and @xmath94
can be measured .
this result can be derived along the same lines as the one obtained in ref .
@xcite in the more general case of the ( @xmath13 ) reaction , and is similar to the result obtained for the ( @xmath95 ) reaction @xcite .
when the nucleon spin is not detected , in coplanar kinematics the only dynamical variable with a nonvanishing component in the @xmath96 direction is the pseudovector @xmath97 .
as no pseudovectors are available with a component along @xmath98 , the matrix elements of the hadron tensor @xmath99 and @xmath100 must vanish .
therefore , @xmath101 = @xmath102 = 0 and , consequently , @xmath103 = @xmath104 = 0 , and only @xmath91 survives . when the nucleon spin is detected , owing to the properties of the spin-@xmath105 matrix , only a linear dependence of the structure functions on the spin is allowed [ see eq .
( [ eq : hll ] ) ] .
as the spin is a pseudovector and in order to fulfill parity conservation , it must be coupled in the hadron tensor with the only available pseudovector independent of the spin , _
i.e. _ with @xmath97 , which in coplanar kinematics is directed along @xmath96 .
therefore , we have : @xmath106 = 0 , @xmath107 dependent on @xmath108 , and @xmath99 independent of @xmath108 . as a consequence
, one has : @xmath109 = @xmath110 = @xmath111 = @xmath112 = 0 .
thus , @xmath113 , @xmath114 , @xmath115 and @xmath116 vanish and only @xmath90 , @xmath92 , @xmath117 and @xmath118 survive .
the behaviour of the hadron tensor under time reversal and parity transformation has the property @xcite @xmath119 where @xmath120 is the spin vector in the ejectile rest frame , and the dependence on the final state boundary condition for incoming @xmath121 and outgoing @xmath122 scattered waves is shown . for nucleon knockout ,
the @xmath121 condition is appropriate . when the boundary conditions can be ignored , as in the plane - wave ( pw ) approximation , eq .
( [ eq : pick ] ) states that the symmetric part of @xmath51 is independent of @xmath120 and the antisymmetric part is proportional to @xmath120 .
therefore , when the hadron tensor is spin independent its antisymmetric part vanishes , while its symmetric part vanishes when it is spin dependent .
thus , in pw only @xmath83 , @xmath85 , @xmath102 , and @xmath123 survive and therefore @xmath89 , @xmath91 , @xmath104 , and @xmath124 , while @xmath103 = @xmath125 = @xmath126 = @xmath127 = 0 . if both coplanar kinematics and the pw approximation are considered , only @xmath89 , @xmath91 and @xmath117 survive .
the results are summarized in table 1 .
in this section numerical results are presented for the @xmath1o(@xmath2)@xmath4n reaction .
calculations have been performed within the theoretical framework of ref .
@xcite . in this model
it is possible to consider transitions to different low - lying discrete states of the residual nucleus , that are expected to be strongly populated by direct knockout . here , we present numerical examples for the transition to the 1@xmath129 ground state of @xmath4n , with @xmath130 .
this is the first state that can be experimentally separated with an energy resolution of a few mev and is therefore of particular interest .
the results are very sensitive to the theoretical treatment of the two - nucleon overlap between the ground state of the target and the final state of the residual nucleus .
an example is shown in fig .
[ fig : fig1 ] , where the calculated differential cross sections of the @xmath1o(@xmath2 ) reaction are displayed in a coplanar and symmetrical kinematics at the incident photon energies of @xmath131 mev and 400 mev . in this
kinematics the two nucleons are emitted in the scattering plane at equal energies and equal but opposite angles with respect to the beam direction .
then , for a fixed value of @xmath20 and changing the value of the scattering angle it is possible to explore all values of the recoil ( @xmath132 ) or missing momentum ( @xmath133 ) distribution
. one has @xmath134 where @xmath19 , @xmath21 , and @xmath22 are the momenta of the incident photon and of the two outgoing nucleons , respectively . if relative and c.m .
motions are factorized and final - state interactions are neglected @xcite , @xmath135 is opposite to the total momentum @xmath136 of the nucleon pair in the target .
thus , the shape of the momentum distribution is driven by the c.m .
orbital angular momentum of the knocked - out pair . as this feature
is not spoiled in an unfactorized approach , the symmetrical kinematics here considered is well suited to give information on the motion of the pair in different c.m .
angular momentum states .
calculations have been performed within the direct knockout framework of refs .
@xcite , where the final - state wave function includes the interaction of each one of the two outgoing nucleons with the residual nucleus by means of a phenomenological optical potential .
the nuclear current is the sum of a one - body part , including convective and spin currents , and of a two - body part , including terms corresponding to the lowest order diagrams with one - pion exchange , namely seagull , pion - in - flight and those with intermediate @xmath137 isobar configurations .
the results given by two different treatments of the two - nucleon overlap are compared in the left and right panels of fig .
[ fig : fig1 ] . the cross sections displayed in the left panels
are obtained with the overlap integral used in @xcite for the @xmath1o(@xmath10)@xmath4n reaction .
it includes the effects of short - range as well as tensor correlations , which are calculated within the framework of the coupled cluster method and with a correlation operator restricted to @xmath138 and @xmath139 excitations .
the argonne v14 potential @xcite is employed as a model for a realistic nucleon - nucleon interaction .
the explicit expression of the two - nucleon overlap is given by an expansion over relative and c.m .
wave functions .
the expansion coefficients , accounting for the global or long - range structure of the specific nuclear states , are determined from a configuration mixing calculation of the two - hole states in @xmath1o , which can be coupled to the angular momentum and parity of the requested final state and are renormalized to account for the spectroscopic factors of the s.p . states .
this is the best model which can be applied at present , as it contains all the complications of a many - body calculation and the effects of a realistic interaction .
the results in the right panels are calculated with the simpler prescription used in ref .
@xcite , _ i.e. _ by the product of the pair function of the shell model , described for the @xmath140 ground state of @xmath4n as a pure ( @xmath141)@xmath142 hole , and of a jastrow type central and state independent correlation function @xcite . with this simpler prescription the contribution of the one - body currents and thus of correlations is much lower than with the more refined model , where also tensor correlations are included . in both cases , however , and for both values of the photon energy ,
the calculated cross sections are dominated by two - body currents , _
i.e. _ by the seagull current at 100 mev and by the @xmath137 current at 400 mev .
the differences between the results shown in the left and right panels of the figure are anyhow large .
they are in particular due to the presence in the model of ref .
@xcite , together with the ( @xmath141)@xmath142 component , with a strength 0.60 , of an interference component ( @xmath143)@xmath144 equal to @xmath145 @xcite .
these differences , which are due only to the treatment of the two - nucleon overlap , clearly show that the shape and the size of the cross section are very sensitive to this treatment and that an accurate description of both aspects related to nuclear structure and correlations is needed to produce reliable numerical predictions .
this result , that is well established for ( @xmath8 ) reactions @xcite , is therefore confirmed also for two - nucleon knockout induced by real photons . the linear asymmetry @xmath91 , calculated with the two different prescriptions for the overlap integral and in the same conditions and kinematics as in fig .
[ fig : fig1 ] , is shown in fig .
[ fig : fig2 ] .
large differences are given also in this case by the two models , both at @xmath146 100 and 400 mev .
the contribution of the one - body current is generally small .
this confirms that even when two - body currents are dominant the results can be very sensitive to nuclear structure and correlation effects .
moreover , as it was already observed in previous investigations @xcite , the asymmetry appears particularly affected by the different terms of the two - body current and by their interference , in particular by the interference between the pion - in - flight and seagull currents .
this contribution is dominant on the asymmetry at @xmath147 100 mev , while the role of pion - in - flight on the cross section is generally small . at @xmath146 400
mev the @xmath137 current gives the main contribution , which , however , is not as dominant as in the cross section .
nonnegligible effects are given also by the other components of the two - body current .
for the following calculations we have chosen a coplanar kinematics at @xmath146 120 mev . at this value of the photon energy ,
far from the peak of the @xmath137 resonance , the contribution of the two - body @xmath137 current , that is an important but difficult ingredient @xcite of the theoretical model , is drastically reduced .
however , the photon energy is large enough to obtain outgoing nucleons with a reduced final state interaction .
the contribution of meson exchange currents can not be eliminated and the chosen energy is a compromise between the different contributions .
the energy and the scattering angle of the outgoing neutron are fixed at 45 mev and @xmath148 , respectively .
this value of the scattering angle is taken in order to minimise the mutual interaction between the outgoing nucleons .
different values of the recoil momentum can be obtained by varying the scattering angle of the outgoing proton .
this kinematics appears within reach of available experimental facilities .
the calculated differential cross section is displayed in fig .
[ fig : fig3 ] . in the left panel
the results given by the one - body current and by the addition of the different terms of the two - body current are shown . at the considered value of the photon energy the pion seagull current gives the dominant contribution to the cross section .
only a minor role is played by the one - body current and by the pion - in flight and @xmath137 currents .
the shape of the angular distribution , however , is practically the same as the one given by the one - body current . in the considered kinematics and in the calculated angular range , the recoil momentum varies between @xmath149 and 290 mev/@xmath150 , on both sides of the distribution .
the shape of the distribution is driven by the c.m .
orbital angular momentum @xmath73 of the knocked - out pair . in our model
the two - nucleon overlap function contains different components of relative and c.m .
motion @xcite .
for the transition to the @xmath140 ground state the relative waves are : @xmath151 ( the notation @xmath152 , for @xmath153 is used here for the relative states ) , combined with a c.m . @xmath154 and @xmath155 , @xmath156 , combined with @xmath157 , and @xmath158 . for this relative wave
we have separated the component already present in the uncorrelated wave function , which is combined with @xmath154 , and the one produced by tensor correlations and not present in the uncorrelated wave function , which is combined with @xmath154 and @xmath155 . as in ref .
@xcite we call these two terms @xmath158 and @xmath159 , respectively .
the contribution of @xmath159 emphasizes the role played in the calculations by tensor correlations , which are anyhow present also in the other components .
the separate contributions of the different waves of relative motion are shown in the right panel of fig .
[ fig : fig3 ] .
the shape of the different curves is determined by the corresponding values of @xmath73 .
the shape of the final cross section is driven by the component which gives the major contribution , _
i.e. _ by @xmath151 in the considered case .
only small effects are given by the other partial waves .
the component @xmath159 , in particular its part with @xmath160 , is nonnegligible at low values of the scattering angle , which correspond to larger values of the momentum .
the cross section is thus basically a combination of states with @xmath154 and @xmath155 .
the contribution of the state with @xmath157 for @xmath156 is very small .
the main role is played by the state with @xmath161 for @xmath151 , which gives the @xmath162-wave shape in the figure .
the contribution of states with @xmath155 for @xmath151 and @xmath159 becomes important at large values of the recoil momentum , where the contribution of the state with @xmath161 becomes much lower .
it has been shown in sect .
[ sec:2 ] that when the incident photon is polarized and/or the polarization of the outgoing nucleon is detected , 7 polarization observables survive in coplanar kinematics : the asymmetry @xmath91 , that can be measured in a reaction with a linearly polarized photon and where the nucleon recoil polarization is not detected , the component @xmath90 of the outgoing nucleon polarization , that can be measured in a reaction with an unpolarized incident photon , the components @xmath93 of the polarization transfer coefficient , that can be measured in a reaction with a circularly polarized photon , and the components @xmath92 and @xmath163 of the polarization asymmetries , that can be measured in a reaction with a linearly polarized incident photon .
the polarization observables calculated with our model and in the coplanar kinematics of fig .
[ fig : fig3 ] are shown in figs .
[ fig : fig4 ] and [ fig : fig5 ] .
the results indicate that all the observables are sizable .
they are sensitive to all the different terms of the nuclear current and also to their interference .
the main contribution is generally given by the seagull current , which , however , is not as dominant as in the cross section .
a large effect is given by the interference between the seagull and pion - in - flight terms , in particular for the asymmetry @xmath91 , where this effect gives the main contribution .
the role of the one - body current is not very important in general , but it can be meaningful in particular situations , for instance on @xmath90 . at the considered value of the photon energy , @xmath164 mev , a small although nonnegligible effect is generally given by the @xmath137 current .
we note that @xmath92 and @xmath163 vanish in the pw approximation , when final - state interactions ( fsi ) are neglected . indeed in fig .
[ fig : fig5 ] these quantities turn out to be smaller than the other observables which are present also in pw .
since they are produced by fsi , one might expect that they are sensitive to their treatment .
the results depend on kinematics and on the final state that is considered .
the numerical example in figs .
[ fig : fig3]-[fig : fig5 ] indicates that cross sections and polarization observables show a different sensitivity to the different terms of the nuclear current .
this can be understood if one considers that different structure functions contribute to the different quantities .
moreover , since the polarization observables can be expressed in terms of ratios between the various structure functions and @xmath83 , that gives the unpolarized cross section , one can expect that they are able to emphasize specific effects that can be smoothed out in the cross section .
thus , a combined experimental determination of cross sections and polarization observables would give a complete information on the reaction process .
it would impose constraints to the different theoretical ingredients and result in a stringent test of theoretical models . in figs .
[ fig : fig3]-[fig : fig5 ] we have shown the sensitivity to the different terms of the nuclear current . in figs .
[ fig : fig1 ] and [ fig : fig2 ] , in different kinematics , it has been shown that the cross section and the photon asymmetry @xmath91 exhibit a strong and different sensitivity to the theoretical treatment of the two - nucleon overlap .
a complete study of all the 16 observables defined in sect . [ sec:2 ] requires an out - of - plane kinematics .
a numerical example is shown in figs .
[ fig : fig6]-[fig : fig9 ] in a situation where the azimuthal angle @xmath165 of the outgoing nucleon whose polarization is not considered ( the proton in our case ) is @xmath166 . the other kinematical variables are taken as in the coplanar kinematics of fig . [ fig : fig3 ] .
the differential cross section in fig .
[ fig : fig6 ] confirms the dominant role of the seagull current and of the @xmath151 component .
the size , however , is an order of magnitude lower than in the peak region of the coplanar kinematics , while the shape indicates a less pronounced peak and a larger contribution of the components with @xmath160 .
the main reason of the differences can be attributed to a kinematic effect .
different values of the recoil momentum are obtained for the same angle @xmath167 in the two kinematics .
when the outgoing nucleon is taken out of the plane , values lower than @xmath168 mev/@xmath150 are forbidden and the region between 90 and 160 mev/@xmath150 , where the cross section in the coplanar kinematics has the maximum , is cut .
this explains the different size and shape of the results in figs .
[ fig : fig3 ] and [ fig : fig6 ] .
this kinematic effect gives also the main difference between the calculated polarization observables which are present both in coplanar and out - of - plane kinematics .
the other observables , which are nonvanishing only in out - of - plane kinematics , are sizable . in the considered situation
the main contribution is generally given by the seagull current , but also the other terms , in particular the interference between seagull and pion - in - flight , can be meaningful in some observables .
in this section numerical results are presented for the @xmath1o(@xmath3)@xmath4c reaction .
calculations have been performed within the theoretical framework of ref .
the model is basically the same used for the @xmath1o(@xmath2)@xmath4n reaction where the nuclear current is adapted to a @xmath6 pair .
thus , in the two - body current the charge - exchange terms vanish and only a part of the @xmath137 current contributes
. however , a different treatment of the two - nucleon overlap function is used , based on the results of the calculation of the two - proton spectral function of @xmath1o @xcite .
the two - nucleon overlaps for transitions to the lowest - lying discrete final states of the residual nucleus are obtained from a two - step procedure , where long - range and short - range correlations are treated in a separate but consistent way .
the calculation of long - range correlations is performed in a shell - model space large enough to incorporate the corresponding collective features which influence the pair removal amplitudes .
the single - particle propagators used for this dressed random phase approximation description of the two - particle propagator also include the effect of both long - range and short - range correlations . in the second step that part of the pair removal amplitudes which describes the relative motion of the pair
is supplemented by defect functions , accounting for src and obtained from the same g - matrix which is also used as the effective interaction in the rpa calculation .
the explicit expression of the overlap function is given in terms of a sum of products of relative and c.m .
wave functions .
different components contribute to different transitions . here
, we present numerical results for the transition to the 0@xmath129 ground state of @xmath4c , where the components are : @xmath170 , which is combined with @xmath154 , and @xmath171 , which is combined with @xmath157 .
the ground state has been already separated in recent high - resolution experiments for the @xmath1o(@xmath9)@xmath4c reaction @xcite .
the experimental cross sections are well reproduced by the results of our model and clear evidence for src has been obtained in the comparison .
thus , it seems interesting to investigate the same final state also for the ( @xmath3 ) reaction . the differential cross section of the @xmath1o(@xmath3)@xmath4c reaction , calculated in the same coplanar kinematics considered in sect
. [ sec:3 ] , is displayed in fig .
[ fig : fig10 ] .
the shape of the distribution is determined by the combination of the two components with @xmath154 and @xmath157 .
it is clearly shown in the right panel of the figure that @xmath170 with its @xmath162-wave c.m .
component prevails for angles corresponding to low values of the recoil momentum , where the @xmath172-wave has the minimum , while the contribution of @xmath171 knockout becomes important al large values of @xmath173 , where the @xmath162-wave decreases .
it was established in previous investigations that src and two - body current play a different role in different relative states @xcite .
while src are quite strong for the @xmath170 state and much weaker for the @xmath174 states , the contribution of the two - body @xmath137 current is strongly reduced in @xmath170 @xmath6 knockout @xcite and generally dominant in @xmath174 knockout .
therefore , @xmath170 @xmath6 knockout is generally dominated by the contribution of the one - body current , and thus by src , while @xmath174 @xmath6 knockout is generally dominated by the @xmath137 current . the selectivity of the reaction to different states was exploited in ( @xmath9 ) , where it was possible to envisage situations where either contribution is dominant and can thus be disentangled .
the result , however , depends on kinematics and on the reaction that is considered .
thus , it can not be directly applied to ( @xmath3 ) reactions , where only the transverse components of the nuclear current contribute and the longitudinal component , where the effects of src show up more strongly , is absent . on the other hand , in the considered kinematics , where @xmath164 mev , the contribution of the @xmath137 current should not be emphasized , as it is also demonstrated by the results of sect .
[ sec:3 ] for the ( @xmath2 ) reaction . in the present calculation the @xmath137 current
gives indeed the major contribution to @xmath171 knockout , but a sizable effect is also given by its interference with the one - body current .
the one - body current is dominant in @xmath170 knockout . in the region of the maximum , where @xmath171 has the minimum , the contribution of the @xmath137 current is larger in @xmath170 than in @xmath171 .
this explains the result in the left panel of fig .
[ fig : fig10 ] .
most of the contribution of the one - body current is given by @xmath170 knockout .
this determines the @xmath162-wave shape of the angular distribution .
the result for the @xmath137 current is due to the combined effect of @xmath171 , which gives the main contribution at large values of @xmath173 , and of @xmath170 , which enhances the cross section with the @xmath137 current at low values of @xmath173 . in the final cross section
both processes due to src and to the two - body current are important : src are of particular relevance at larger values of @xmath167 and the @xmath137 at lower values of @xmath167 .
the polarization observables are displayed in figs .
[ fig : fig11 ] and [ fig : fig12 ] .
all the calculated quantities are sizable .
both contributions of the one - body and of the two - body current are important , even though the relevance of the two reaction mechanisms can be different for different observables .
for instance , @xmath92 is basically given by the one - body current , while for other observables the final result , given by the sum of the one - body and the two - body current , is very different from the two separate contributions .
thus , a combined measurement of cross sections and polarization observables would be able to disentangle the two reaction processes and to test the ingredients and the approximations of a theoretical model .
the cross section and the polarization observables calculated for the @xmath1o(@xmath3)@xmath4c reaction in the out - of - plane kinematics with @xmath175 are given in figs .
[ fig : fig13]-[fig : fig16 ] . in comparison with the result for the coplanar kinematics , the cross section in fig .
[ fig : fig13 ] shows the same kinematic effect found in the @xmath1o(@xmath2)@xmath4n reaction .
values of @xmath173 between 90 and 160 mev/@xmath150 are forbidden , and the region that in the coplanar kinematics of fig .
[ fig : fig10 ] corresponds to the maximum of @xmath170 and to the minimum of @xmath171 is cut . as a consequence ,
the contribution of @xmath171 is always larger than the one of @xmath170 all over the angular distribution .
the size of the final cross section is reduced in the region of the maximum and a different shape is obtained .
the contributions of the one - body and of the two - body current displayed in the left panel are of about the same size and add up in the final cross section .
all the polarization observables in figs .
[ fig : fig14]-[fig : fig16 ] are sizable .
also in this case both the one - body and the two - body current are important . in some observables
either contribution prevails or appears even dominant in the calculation , but both reaction processes are in general important in the final result and their theoretical treatment can be tested in combined measurements of the different observables .
also the polarization observables which vanish in pw are sizable
. their determination might be of particular interest for the study of fsi , and in particular of the correlations between the two outgoing particles in the final state , that are neglected in the present model .
for the kinematics that are usually considered these effects should not be large in the cross section , but might be emphasized in particular polarization observables .
two - nucleon knockout reactions represent the best suited tool for exploring the behaviour of nucleon pairs in nuclei and their mutual interaction . in particular , electromagnetic reactions give a direct access to nuclear structure and correlations .
a combined analysis of cross sections and polarization observables is required for this study . in this paper
we have given the general properties of the ( @xmath0 ) reaction where the incident photon is linearly or circularly polarized and the polarization of one outgoing nucleon is detected . in the most general situation 16 measurable quantities , including the cross section , are obtained , each one corresponding to one different structure function .
these observables allow a complete investigation of all the components and the properties of the hadron tensor
. for a complete determination an out - of - plane kinematics is needed , _
i.e. _ a kinematics where one nucleon does not lie on the plane of the photon and of the other nucleon . in the more usual coplanar kinematics a situation similar to the one of the ( @xmath95 ) reaction is recovered , and only 8 observables are nonvanishing .
ten structure functions , and therefore the corresponding observables , vanish when the pw approximation is considered for the outgoing nucleons .
these quantities appear very well suited to investigate the effect of fsi between the ejected nucleons and the residual nucleus , and , possibly , the mutual interaction between the two outgoing nucleons .
numerical examples have been presented for the exclusive @xmath1o(@xmath2)@xmath4n and @xmath1o(@xmath3)@xmath4c reactions .
@xmath1o is a well suited target for this investigation . from the theoretical point of view ,
a great deal of work has been done and a large experience has been acquired on this nucleus . from the experimental point of view ,
precise measurements have been already carried out for these reactions or are planned in different laboratories , aiming at determining cross sections and polarization observables for individual final states .
it is of particular interest to consider exclusive reactions .
previous investigations have shown that different final states exhibit a different sensitivity to the theoretical ingredients of the calculations and to the various reaction processes .
thus , the experimental resolution of specific final states may act as a filter to disentangle and separately investigate different reaction processes and , in particular , to study nuclear correlations .
the results are very sensitive to the treatment of the nucleon pair wave function and to different aspects involving nuclear structure and correlations .
the comparison of theoretical predictions with experimental data can thus impose severe constraints to the theoretical models .
the polarization measurements , which are here particularly addressed , are expected to be sensitive to the small components of the transition amplitudes , which do not show up in the cross sections , where they are generally overwhelmed by the dominant components . in this
paper cross sections and polarization observables of the @xmath1o(@xmath2)@xmath4n and @xmath1o(@xmath3)@xmath4c reactions have been calculated with a similar approach and considering the transitions to the ground state of the residual nucleus .
the nuclear current contains a one - body part , which contributes to the transition only through the correlations , and a two - body part , given by the pion seagull , the pion - in - flight and the @xmath137 isobar contributions for ( @xmath2 ) reactions , and only by the contribution of the @xmath137 without charge exchange for ( @xmath3 ) reactions . in the @xmath5 pair wave function the effects of short - range and tensor correlations
are taken into account , while the @xmath6 pair includes a consistent treatment of short - range and long - range correlations .
as an example , a photon energy of 120 mev , which seems suitable for experiments , has been chosen in the calculations . at this energy value
the @xmath137 plays only a minor role .
thus , in the @xmath1o(@xmath2)@xmath4n reaction the seagull current is dominant .
the contribution of the one - body current is not very important in general , but can be meaningful in particular situations .
the contribution of tensor correlations is significant , particularly at high values of the recoil momentum .
both the one - body and the @xmath137 current are important in the @xmath1o(@xmath3)@xmath4c reaction .
the relevance of either contribution can be emphasized in particular kinematic conditions and in different observables .
the polarization observables are always sizable .
they are sensitive to all the different terms of the nuclear current and also to their interference . a combined experimental determination of cross sections and polarizations would give a complete information on the reaction process and result in a stringent test of the theoretical models .
this would make it possible to disentangle the different contributions to the reactions and shed light on the genuine nature of correlations in nuclei .
mcgeorge , i.j.d .
macgregor , s.n .
dancer , j.r.m .
annand , i. anthony , g.i .
crawford , s.j .
hall , p.d .
harty , j.d .
kellie , g.j .
miller , r.o .
owens , p.a .
wallace , d. branford , a.c .
shotter , b. schoch , r. beck , h. schmieden , and j.m .
vogt , phys .
c * 51 * ( 1995 ) 1967 ; + p. grabmayr , j. ahrens , j.r.m .
annand , i. anthony , d. branford , g.e .
cross , t. davinson , s.j .
hall , p.d .
harty , t. hehl , j.d .
kellie , th .
lamparter , i.j.d .
macgregor , j.a .
mackenzie , j.c .
mcgeorge , g.j .
miller , r.o .
owens , m. sauer , r. schneider , k. spaeth , and g.j .
wagner , phys .
b * 370 * ( 1996 ) 17 ; + p.d . harty , i.j.d .
macgregor , p. grabmayr , j. ahrens , j.r.m .
annand , i. anthony , r. beck , d. branford , g.e .
cross , t. davinson , s. j. hall , t. hehl , r.o .
owens , m. sauer , r. schneider , and k. spaeth , phys . lett .
b * 380 * ( 1996 ) 247 ; + th .
lamparter , j. ahrens , j.r.m .
annand , i. anthony , r. beck , d. branford , g.e .
cross , t. davinson , p. grabmayr , s.j .
hall , p.d .
harty , t. hehl , j.d .
kellie , i.j.d .
macgregor , j.a .
mackenzie , j.c .
mcgeorge , g.j .
miller , r.o .
owens , m. sauer , r. schneider , k. spaeth , and g.j . wagner , z. phys . a * 355 * ( 1996 ) 1 ; + i.j.d .
macgregor , t.t - h .
yau , j. ahrens , j.r .
m. annand , r. beck , d. branford , p. grabmayr , s.j .
hall , p.d .
harty , t. hehl , j.d .
kellie , th .
lamparter , m. liang , j.a .
mackenzie , s. mcallister , j.c .
mcgeorge , r.o .
owens , j. ryckebusch , m. sauer , r. schneider , and d.p .
watts , phys .
* 80 * ( 1998 ) 245 ; + d.p .
watts , i.j.d .
macgregor , j. ahrens , j.r .
m. annand , r. beck , d. branford , p. grabmayr , s.j .
hall , p.d .
harty , t. hehl , j.d .
kellie , th .
lamparter , m. liang , j.a .
mackenzie , s. mcallister , j.c .
mcgeorge , r.o .
owens , m. sauer , r. schneider , g.j .
wagner , and t.t - h yau , phys .
c * 62 * ( 2000 ) 014616 .
l. isaksson , _ ph.d .
thesis , university of lund _ , ( 1996 ) ; + l. isaksson , j - o .
adler , b - e . andersson , k.i .
blomqvist , a. sandell , b. schrder , p. grabmayr , s. klein , g. mauser , a. mondry , j.r.m .
annand , g.i .
crawford , v. holliday , j.c .
mcgeorge , and g.j .
miller , phys .
lett . * 83 * , ( 1999 ) 3146 .
j. ahrens , j.r.m .
annand , r. beck , d. branford , p. grabmayr ( spokesperson ) , s.j .
hall , t. hehl , d.j .
ireland , j.d .
kellie , i.j.d .
macgregor , m. mayer , j.c .
mcgeorge , f.a .
natter , r.o .
owens , m. sauer , g.j .
wagner , and s. wunderlich , mami proposal nr : a2/4 - 97
. p. grabmayr , _ proceedings of the fourth workshop on electromagnetically induced two - nucleon emission , granada , 1999 _ edited by c. garcia - recio , p. grabmayr , a.m. lallena and r. owens ( university of granada , granada , 1999 ) p. 331 .
onderwater , k. allaart , e.c .
aschenauer , th.s .
bauer , d.j .
boersma , e. cisbani , s. frullani , f. garibaldi , w.j.w .
geurts , d. groep , w.h.a .
hesselink , m. iodice , e. jans , n. kalantar - nayestanaki , w .- j .
kasdorp , c. kormanyos , l. lapiks , j.j.van leeuwe , r. de leo , a. misiejuk , a.r .
pellegrino , r. perrino , r. starink , m. steenbakkers , g. van der steenhoven , j.j.m .
steijger , m.a .
van uden , g.m .
urciuoli , l.b .
weinstein , and h.w .
willering , phys .
* 78 * , ( 1997 ) 4893 .
onderwater , k. allaart , e.c .
aschenauer , th.s .
bauer , d.j .
boersma , e. cisbani , w.h .
dickhoff , s. frullani , f. garibaldi , w.j.w .
geurts , c. giusti , d. groep , w.h.a .
hesselink , m. iodice , e. jans , n. kalantar - nayestanaki , w .- j .
kasdorp , c. kormanyos , l. lapiks , j.j .
van leeuwe , r. de leo , a. misiejuk , h. mther , f.d .
pacati , a.r .
pellegrino , r. perrino , r. starink , m. steenbakkers , g. van der steenhoven , j.j.m .
steijger , m.a .
van uden , g.m .
urciuoli , l.b .
weinstein , and h.w .
willering , phys .
* 81 * , ( 1998 ) 2213 .
r. starink , m.f .
van batenburg , e. cisbani , w .. h . dickhoff , s. frullani , f. garibaldi , c. giusti , d.l .
groep , p. heimberg , w.h.a .
hesselink , m. iodice , e. jans , l. lapiks , r. de leo , c.j.g .
onderwater , f.d .
pacati , r. perrino , j. ryckebusch , m.f.m .
steenbakkers , j.a .
templon , g .-
urciuoli , and l.b .
weinstein , phys . lett .
b * 474 * , ( 2000 ) 33 .
g. rosner,_conference on perspectives in hadronic physics _ , edited by s. boffi , c. ciofi degli atti , and m m. giannini ( world scientific , singapore , 1998 ) 185 ; + _ proceedings of the 10th mini - conference on studies of few - body systems with high duty - factor electron beams , nikhef , amsterdam 1999 _ , edited by h.p .
blok , e. jans , g. van der steenhoven ( nikhef , asterdam , 1999 ) p. 95 ; + m. kahrau , _ ph.d thesis , johannes gutenberg universitt , mainz _ ( 1999 ) .
annand , p. bartsch , d. baumann , j. becker , r. bhm , d. branford , s. derber , m. ding , i. ewald , k. fhl , j. friedrich , j.m .
friedrich , p. grabmayr ( spokesperson ) , t. hehl , d.g .
ireland , p. jennewein , m. kahrau , d. kndler , k.w .
krygier , a. liesenfeld , i.j.d .
macgregor , h. merkel , k. merle , p. merle , u. mller , h. mther , a. natter , r. neuhausen , th .
pospischil , g. rosner ( spokesperson ) , h. schmieden , a. wagner , g.j .
wagner , th .
walcher , m. weis , and s. wolf , mami proposal nr
: a1/5 - 98 . d.j .
tedeschi , g.s .
adams , g. audit , h. baghaei , a. caracappa , w.b .
clayton , a. dangelo , m.a .
duval , g. giordano , s. hobit , o.c .
kistner , j.m .
laget , r. lindgren , g. matone , l. miceli , w.k .
mize , m. moinester , c. ruth , a. sandorfi , c. schaerf , r.m .
sealock , l.c .
smith , p. stoler , p.k .
teng , c.e .
thorn , s.t .
thornton , k. vaziri , c.s .
whisnan , and e.j .
winhold , phys .
lett . * 73 * , ( 1994 ) 408 .
s. franczuk , i.j.d .macgregor , j. ahrens , j.r.m .
annand , j.f .
arneil , r. beck , d. branford , p. grabmayr , s.j .
hall , t. hehl , p.d .
harty , d.g .
ireland , j.d .
kellie , k. livingston , j.c .
mcgeorge , f.a .
natter , s. oberkirsch , r.o .
owens , c.j.y .
powrie , j. ryckebusch , m. sauer , a. settele , p.d .
watts , phys .
b * 450 * , ( 1999 ) 332 ; + i.j.d .macgregor , _ proceedings of the fourth workshop on electromagnetically induced two - nucleon emission , granada 1999 _ edited by c. garcia - recio , p. grabmayr , a.m. lallena and r. owens ( university of granada , granada , 1999 ) p. 339 . | the formalism of ( @xmath0 ) reactions is given where the incident photon is polarized and the outgoing nucleon polarization is detected .
sixteen structure functions and fifteen polarization observables are found in the general case , while only eight structure functions and seven polarization observables survive in coplanar kinematics .
numerical examples are presented for the @xmath1o(@xmath2 ) and @xmath1o(@xmath3 ) reactions .
the transitions to the ground state of @xmath4c and @xmath4n are calculated in a model where realistic short - range and tensor correlations are taken into account for the @xmath5 pair , while short - range and long - range correlations are included in a consistent way for @xmath6 pairs .
the effects of the one - body and two - body components of the nuclear current and the role of correlations in cross sections and polarizations are studied and discussed .
|
the linear boltzmann equation ( also referred to as the boltzmann - lorentz equation or kinetic lorentz equation ) is one of the fundamental transport equations that describe the macroscopic dynamics of a dilute gas in matter .
the equation was postulated by lorentz in 1905 @xcite by considering a gas of non - interacting point particles moving in an infinite , fixed array of hard sphere scatterers .
crucially , lorentz assumed that in the limit of small scatterers ( boltzmann - grad limit ) consecutive collisions become independent of each other and are solely determined by the single - scatterer cross section .
lorentz heuristic derivation of the linear boltzmann equation was put on a rigorous footing in the case of a _
random _ scatterer configuration in the seminal papers by gallavotti @xcite , spohn @xcite , and boldrighini , bunimovich and sinai @xcite . on the other hand , our recent studies of _ periodic _ scatterer configurations @xcite , @xcite
show that in this case the boltzmann - grad limit is governed by a transport equation which is distinctly different from the linear boltzmann equation .
one of the main features is here that ( contrary to lorentz assumption ) consecutive collisions are no longer independent : the collision kernel of our transport equation does not only depend on the particle velocity before and after each collision , but also on the flight time until the next collision and the velocity thereafter .
the collision kernel is thus significantly more complicated than in the linear boltzmann equation , and explicit formulas are so far only known in dimension @xmath0 @xcite ; cf . also @xcite , @xcite for different approaches .
the objective of the present paper is to focus on dimension @xmath1 and derive asymptotic estimates for the collision kernel for small and large inter - collision times .
these estimates yield in particular precise asymptotics for the distribution of the free path length in the periodic lorentz gas , and thus improve the upper and lower bounds obtained by bourgain , golse and wennberg @xcite , @xcite . to explain the setting of our results in more detail ,
let us fix a euclidean lattice @xmath2 , and assume ( without loss of generality ) that its fundamental cell has volume one .
we denote by @xmath3 the complement of the set @xmath4 ( the `` billiard domain '' ) , and @xmath5 its unit tangent bundle ( the `` phase space '' ) , with @xmath6 the position and @xmath7 the velocity of the particle at time @xmath8 . here
@xmath9 [ scrbdef ] denotes the open ball of radius @xmath10 , centered at the origin , and @xmath11 the unit sphere .
the dynamics of a particle in the lorentz gas is defined as the motion with unit speed along straight lines , and specular reflection at the balls @xmath12 .
we may in fact also permit other scattering processes , such as the scattering map of a muffin - tin coulomb potential ; cf .
@xcite for details .
a dimensional argument shows that in the boltzmann - grad limit @xmath13 the free path length scales like @xmath14 , i.e. , the inverse of the total scattering cross section of an individual scatterer .
it is therefore natural to rescale space and time by introducing the macroscopic coordinates ( see figure [ figlorentz ] ) @xmath15 the time evolution of a particle with initial data @xmath16 is then described by the billiard flow @xmath17 since the speed of our particle is a constant of motion we may assume without loss of generality that @xmath18 . for notational reasons it is convenient to extend the dynamics to the inside of each scatterer trivially , i.e. , set @xmath19 whenever @xmath20 is inside the scatterer .
that is , the relevant phase space is now the unit tangent bundle of @xmath21 , which will be denoted by @xmath22 .
let us fix a probability measure @xmath23 on @xmath22 . for random initial data @xmath24 with respect to @xmath23 , we can then view the billiard flow @xmath25 as a stochastic process .
the central result of @xcite , @xcite is that , if @xmath23 is absolutely continuous with respect to lebesque measure , the billiard flow converges in the boltzmann - grad limit to a random flight process @xmath26 , which is defined as the flow with unit speed along a random piecewise linear curve , whose path segments @xmath27 are generated by a markov process with memory two .
specifically , if we set @xmath28 and @xmath29 for @xmath30 , then the distribution of the first @xmath31 path segments is given by the probability density @xmath32 see theorem 1.3 and section 4 in @xcite . before describing the transition probability densities @xmath33 and @xmath34 in more detail , let us explain the relation of our limiting stochastic process @xmath35 with the macroscopic dynamics of a particle cloud discussed earlier .
the time evolution of an initial particle density @xmath36 in the lorentz gas with fixed scatterer radius @xmath10 is given by @xmath37 , where @xmath38 is the liouville operator defined by @xmath39(\vecq,\vecv ) : = f_0\big(f_{-t,\rho}(\vecq,\vecv)\big ) .\ ] ] the existence of the limiting stochastic process @xmath35 implies that for every @xmath40 there exists a linear operator @xmath41 , such that for every @xmath36 and any set @xmath42 with boundary of lebesgue measure zero , @xmath43(\vecq,\vecv)\ , d\vecq\ , d\!\vol_{\s_1^{d-1}}(\vecv ) = \int_{\scra } [ l^t f_0](\vecq,\vecv ) \ , d\vecq\ , d\!\vol_{\s_1^{d-1}}(\vecv ) .\ ] ] if @xmath44 is @xmath45 then its image under the limit operator @xmath46 is given by @xmath47(\vecq,\vecv)=\int_{{\mathbb{r}}_{>0}\times \s_1^{d-1 } } f(t,\vecq,\vecv,\xi,\vecv_+)\ , d\xi\ , d\!\vol_{\s_1^{d-1}}(\vecv_+),\ ] ] where @xmath48 is the unique solution of the differential equation @xmath49 f(t,\vecq,\vecv,\xi,\vecv_+ ) \\
= \int_{\s_1^{d-1 } } f(t,\vecq,\vecv_0,0,\vecv ) p_{\vecnull}(\vecv_0,\vecv,\xi,\vecv_+ ) \ , d\!\vol_{\s_1^{d-1}}(\vecv_0 ) \end{gathered}\ ] ] subject to the initial condition @xmath50 .
equation corresponds to the fokker - planck - kolmogorov equation of our limiting stochastic process @xmath35 , cf .
section 6.3 of @xcite , and may be viewed as a generalization of the linear boltzmann equation , cf .
* section 3 ) . 0.1 ( 10,8)(0,0 ) ( 0.5,1),title="fig:",scaledwidth=90.0% ] ( 0.1,3.2)@xmath51 ( 2.3,3.3)@xmath52 ( 6.6,5.8)@xmath52 ( 1.6,2.4)@xmath53 ( 6,2.5)@xmath54 ( 8.6,6.6)@xmath55 ( 6.8,6.8)@xmath56 we can express the probability densities @xmath57 and @xmath58 as @xmath59 @xmath60 where @xmath61 is the differential cross section , @xmath62 is the limiting probability density ( as @xmath13 ) of hitting , from a generic point in @xmath22 , the first scatterer at time @xmath54 ( in microscopic units ) with impact parameter @xmath63 , and @xmath64 is the limiting probability density of hitting , from a given scatterer with exit parameter @xmath65 , the next scatterer at time @xmath54 with impact parameter @xmath63 ( cf . figure [ figtp ] ) .
if the scattering map is given by specular reflection ( as in the original lorentz gas ) , we have explicitly @xmath66 for the scattering cross section , and @xmath67 for the exit and impact parameters . here
@xmath68 denotes the orthogonal projection of @xmath69 onto @xmath70 , and for each @xmath71 we have fixed a rotation @xmath72 with @xmath73 .
the study of the asymptotic properties of the probability densities @xmath74 and @xmath75 for @xmath76 and @xmath77 are the core objectives of the present paper .
precise formulas for @xmath78 and @xmath79 in terms of natural probability measures on the homogeneous space @xmath80 are given in section [ phidefsec ] below .
at this point we list the following useful facts : * if @xmath1 , the functions @xmath81 , \qquad \phi_\vecnull : \rr_{>0}\times \scrb_1^{d-1}\times \scrb_1^{d-1 } \to [ 0,1]\ ] ] are continuous .
* @xmath74 depends only on @xmath82 and @xmath83 ; we set @xmath84 with @xmath85 .
* @xmath75 depends only on @xmath86 the angle between @xmath87 ; we set @xmath88 with @xmath89 and @xmath90 $ ] .
* @xmath91 and thus @xmath92 .
* we have the formulas @xmath93 ( a)(d ) follow from ( * ? ? ?
* remark 4.5 ) and are proved in sections 8.1 and 8.2 of that paper .
( e ) follows from ( * ? ? ?
* remark 6.2 ) and ( * ? ? ?
* ( 4.16 ) ) . in dimension
@xmath0 we have the following explicit formula for the transition probability @xcite : @xmath94 with @xmath95 the same formula has recently been found independently by caglioti and golse @xcite and by bykovskii and ustinov @xcite , using different methods based on continued fractions .
all of the relations stated in ( a)(e ) above are also valid for the lorentz gas with _
random _ scatterer configuration . here the fundamental function @xmath96 is given by the explicit formula @xmath97 where @xmath98 denotes ( throughout this paper ) the volume of the unit ball in @xmath99 .
note that relation ( e ) implies with that in the random setting @xmath100 .
returning to the setting of the periodic lorentz gas , we first state the asymptotic formulas for @xmath78 and @xmath79 as @xmath77 . our main result in this direction is the following .
[ phi0zerosmallthm ] for @xmath101 and @xmath102 , @xmath103 that is , @xmath104 , where the remainder term is everywhere non - positive , and the implied constant is independent of @xmath105 and @xmath106 .
note that this estimate is consistent with formula in dimension @xmath0 , where we have the exact relation @xmath107 for @xmath108 . in the case of a random scatterer configuration , yields @xmath109 .
comparing this with theorem [ phi0zerosmallthm ] , we may conclude that the leading - order asymptotics of the transition probability density @xmath75 for @xmath77 is , in both the random and periodic set - up , independent of @xmath87 , and given by the relative density of scatterers which are completely visible from a given scatterer : in the random setting , this is the case for all scatterers close to the given scatterer ; in the periodic setting the same holds only for scatterers located on visible ( or primitive ) lattice points , whose relative density is given by @xmath110 .
theorem [ phi0zerosmallthm ] combined with immediately implies : [ phi0zerosmallcor ] for all @xmath101 and @xmath111 , @xmath112 where the remainder term is everywhere non - negative , and the implied constant is independent of @xmath105 . from our asymptotics for @xmath74 and @xmath75 we can also derive asymptotics for the limiting distribution of the free path length .
we have ( cf .
* remark 4.6 ) ) @xmath113 for the free path length between consecutive collisions , @xmath114 for the free path length from a generic initial point inside the billiard domain , and @xmath115 for the free path length of a particle starting at a lattice point ( with the scatterer removed ) .
only in dimension @xmath0 we have explicit formulas for the above limiting distributions , thanks to the work of dahlqvist @xcite , boca , gologan and zaharescu @xcite , and boca and zaharescu @xcite . using theorem [ phi0zerosmallthm ] and corollary [ phi0zerosmallcor ] in conjunction with , , , we obtain the following . for @xmath101 , @xmath116 here the first two remainder terms are @xmath117 and the last remainder term is @xmath118 , for all @xmath101 .
formulas , , are also valid in the random setting , and yield @xmath119 these asymtotics are the same as in the periodic setting , with the relative density of visible lattice points @xmath110 replaced by @xmath120 . in the case @xmath121 we are in fact able to compute @xmath75 _ explicitly _ for @xmath82 small . for @xmath122
we set @xmath123 also let @xmath124 be the continuous function given by @xmath125 , where @xmath126 is the largest possible volume of a tetrahedron which is contained in the closed cylinder @xmath127\times\overline{\scrb_1 ^ 2}$ ] and which has one vertex at @xmath128 and another at @xmath129 . [ d3explthm ] if @xmath121 , then for all @xmath130 and all @xmath131 , @xmath132 we will prove in lemma [ xi1boundslem ] below that @xmath133 for all @xmath130 , where both the lower and the upper bound are sharp . in particular the formula is always true when @xmath134 .
combining theorem [ d3explthm ] with we will prove the following : [ d3explthmcor1 ] if @xmath121 , then for all @xmath135 and all @xmath136 : @xmath137 where @xmath138\to{\mathbb{r}}_{>0}$ ] is the function @xmath139 the function @xmath140 is continuous and strictly increasing , and satisfies @xmath141 and @xmath142 .
furthermore , using theorem [ d3explthm ] in conjunction with , , , we will prove : [ d3explthmcor2 ] if @xmath121 , then @xmath143 figures [ phiplot ] and [ phi0plot ] show graphs of @xmath144 and @xmath145 for @xmath121 , obtained by numerical computations described in section [ numcompsec ] .
the case of @xmath76 is much more complicated than @xmath82 small . among other things ,
difficulties are caused by the fact that for large @xmath82 , @xmath146 _ vanishes _ unless both @xmath147 and @xmath148 are near @xmath120 .
the following theorem gives an asymptotic formula for @xmath146 as @xmath76 , for @xmath149 small .
the case of small @xmath149 is in a natural sense the most important one .
indeed , it was seen in @xcite that , for given large @xmath82 , the function @xmath146 takes its largest values when @xmath149 is small , and also that it has its largest support with respect to @xmath150 in this case . in particular , when integrating @xmath75 over @xmath151 , the main contribution comes from @xmath106 with @xmath152 .
this follows directly from theorem [ cylinder2ptsmainthm ] and proposition [ phi0supportthm ] below ( cf .
also ( * ? ? ?
* cor . 1.9 ) ) .
[ phi0xilargethm ] there exists a continuous and uniformly bounded function @xmath153 such that @xmath154 for all @xmath155 , where the error term is @xmath156 in dimension @xmath0 we have from ( * ? ? ?
* eq . ( 36 ) ) ( note that @xmath157 or @xmath158 in this case ) @xmath159 for @xmath1 , we will express @xmath160 as an integral of a function given by the probability that a random lattice in @xmath161 is disjoint from a union of two cut paraboloids ( cf .
section [ xidm1basicssec ] and below ) .
this integral representation will imply the following properties : * * symmetry : * @xmath162 ( cf . and ) . *
* support : * there is a continuous function @xmath163 , which we will define in terms of a lattice problem in dimension @xmath164 ( cf . below ) , such that @xmath165 this function @xmath166 satisfies the symmetry relation @xmath167 , and we also have @xmath168 uniformly over @xmath169 $ ] , @xmath170 ( cf . below ) .
it follows that there exist constants @xmath171 which only depend on @xmath172 such that @xmath173 in particular the support of @xmath160 is contained in @xmath174\times(0,c']\times{\mathbb{r}}_{\geq0}$ ] , and for any given @xmath175 , @xmath176 has compact support in the third variable . *
* bounds from above : * @xmath160 is uniformly bounded ( as mentioned ) , and we have the following more precise bound ( cf . and lemma [ sabvboundlemcor ] below ) : @xmath177 * * bounds from below : * for general @xmath172 it follows from theorem [ phi0xilargethm ] combined with ( * ? ? ?
7.8 ) that @xmath160 is uniformly bounded from below for @xmath178 near zero , viz .
there is a small constant @xmath179 which only depends on @xmath172 such that @xmath180 for @xmath121 we have a stronger result ( cf .
7.7 ) ) which says that is sharp in a natural sense ; but we expect that is _ not _
sharp for @xmath181 .
the lower bound implies that the main term dominates as @xmath76 in whenever @xmath182 , @xmath183 and @xmath184 . note that this @xmath185-regime contributes a positive portion to the integral . beyond this regime ,
our motivation for theorem [ phi0xilargethm ] is that even though we can not say _ exactly _ where the main term dominates , it dominates on a sufficiently large set so that we get a good asymptotics for @xmath186 via , see theorem [ phixiwasymptthm ] below . next we give a closely related result on the asymptotic shape of the _ support _ of @xmath79 for @xmath82 large and @xmath149 small .
note that by , the main term in is non - zero if and only if @xmath187 however , clearly theorem [ phi0xilargethm ] never helps us deduce @xmath188 , and furthermore since ( as discussed above ) it is difficult to state when the main term dominates the error term in theorem [ phi0xilargethm ] , we can not deduce @xmath189 in any reasonable region either . by using some intermediate formulas from the proof of theorem [ phi0xilargethm ]
( see section [ supportsection ] for details ) we are , however , able to prove that gives in fact a good approximation of the support of @xmath79 for @xmath82 large : [ phi0supportasymptthm ] there is a continuous function @xmath190\to{\mathbb{r}}_{>0}$ ] such that @xmath191 holds if and only if @xmath192 .
this function @xmath193 satisfies @xmath194 uniformly over all @xmath195 , @xmath196 $ ] .
( the implied constant depends only on @xmath172 . ) to complement the picture of @xmath79 for @xmath82 large , we recall the uniform bounds on the size and support of @xmath79 which the second author proved in @xcite . [ cylinder2ptsmainthm ] ( ( * ? ? ?
* thm . 1.8 ) ) let @xmath197 .
we then have , for all @xmath101 , @xmath102 , and writing @xmath198 , @xmath199 [ phi0supportthm ] ( ( * ? ? ?
1.9 ) ) let @xmath1 . we then have for all @xmath195 , @xmath200 $ ] , writing @xmath201 , @xmath202 where the implied constants only depend on @xmath172 .
( if @xmath157 then the right hand side of should be interpreted as @xmath203 . ) note that the case @xmath204 of theorem [ cylinder2ptsmainthm ] can be recovered from theorem [ phi0xilargethm ] and ; and the case @xmath204 of proposition [ phi0supportthm ] can be recovered from theorem [ phi0supportasymptthm ] combined with the fact that @xmath168 uniformly over @xmath169 $ ] , @xmath170 .
in @xcite it was also proved that the upper bound in theorem [ phi0supportasymptthm ] is _ sharp _ in a natural sense for @xmath121 , and also for general @xmath197 if either @xmath205 or @xmath206 .
next we give an asymptotic formula for @xmath186 for @xmath82 large .
[ phixiwasymptthm ] there exists a bounded continuous function @xmath207 such that @xmath208 as @xmath76 , uniformly over all @xmath209 . for @xmath0 ,
we have explicitly @xcite @xmath210 for general @xmath1 , we will express @xmath211 as an integral of a function given by the probability that a random lattice in @xmath161 is disjoint from a cut paraboloid ( cf .
this function has the following properties : the support of @xmath212 is exactly the interval @xmath213 , where @xmath214 is the same function as in . furthermore @xmath212 stays bounded from below as @xmath215 ; in fact the limit @xmath216 exists ; cf .
proposition [ limfdt2prop ] below .
this number has a natural interpretation in relation with theorem [ phixiwasymptthm ] : the function @xmath186 may be extended to a continuous function on @xmath217 $ ] ( cf . ,
e.g. , ( * ? ? ? * lemma 2.3 ) ) .
since holds uniformly with respect to @xmath209 we conclude by letting @xmath218 : @xmath219 we prove theorem [ phixiwasymptthm ] in section [ preliminariessec ] ; this proof also serves as a preparation for the proof of theorem [ phi0xilargethm ] , which we give in sections [ parabapprsec][phi0xiwzasymptsec ] . on the other hand
it is alternatively possible to obtain theorem [ phixiwasymptthm ] ( except we get a slightly worse error term when @xmath121 ) as a _ consequence _ of theorem [ phi0xilargethm ] and theorem [ cylinder2ptsmainthm ] , via the formula . in fact
this approach even gives an asymptotic formula for @xmath220 as @xmath76 ; cf .
theorem [ pphixiwasymptthm ] in section [ partialphixizsec ] . as a simple consequence of , theorem [ phi0supportasymptthm ] and proposition [ phi0supportthm ] , we obtain a precise understanding of the support of @xmath186 as @xmath76 .
recall that @xmath221 if and only if @xmath222 . [ phixiwsupcor ] there is a continuous function @xmath223 such that @xmath224 holds if and only if @xmath225 , and we have @xmath226 as @xmath227 .
the above results directly yield asymptotics for the different limiting distributions of the free path length , , , .
the first statement concerns the distribution function for the free path length of a trajectory with generic initial condition .
[ phixilargethm ] @xmath228 this result sharpens the upper bound given by bourgain , golse and wennberg @xcite and the lower bound of golse and wennberg @xcite . as to the distribution of the free path length between consecutive collisions
, we have : [ phibarxilargethm ] @xmath229 finally , regarding the distribution of the free path length starting at a lattice point ( with the scatterer removed ) , let us define @xmath230 [ deltastardef ] to be the maximal lattice packing density of any cylinder of the form @xmath231 ( cf .
@xcite or @xcite ) ; of course this number is independent of @xmath232 and @xmath233 .
[ phizeroasympt ] let @xmath234 be given .
then @xmath145 has compact support ; @xmath235 holds if and only if @xmath236 , where @xmath237 is the function from corollary [ phixiwsupcor ] .
@xcite for a brief listing of what is known about @xmath230 for general @xmath172 , in particular note that @xmath238 , where @xmath239 is the maximal packing density of balls in dimension @xmath164 , and @xmath240 is the corresponding lattice packing density . for dimension
@xmath241 we have explicitly : * for @xmath0 , we have @xmath242 , which implies @xmath243 ; cf .
* ( 30 ) ) .
* for @xmath121 , we have @xmath244 , which implies @xmath245 . * for @xmath246
, we have @xmath247 ( cf .
@xcite ) , which implies @xmath248 .
we begin by recalling the formulas of the probability densities @xmath74 and @xmath75 derived in @xcite . throughout this paper
we will write @xmath249 , @xmath250 and @xmath251 .
[ gmux1def ] we will view @xmath252 as the space of lattices in @xmath253 of covolume one , by letting @xmath254 correspond to the lattice @xmath255 .
we define @xmath256 to be the measure on @xmath257 coming from haar measure on @xmath258 , normalized to be a probability measure .
we will sometimes write @xmath259 , @xmath260 , @xmath261 and @xmath262 for @xmath258 , @xmath263 , @xmath252 and @xmath256 , if we need to emphasize the dimension .
we denote by @xmath264 the cylinder @xmath265 the function @xmath266 $ ] can be expressed as the probability that a random lattice in @xmath252 does not intersect the cylinder @xmath267 , i.e. , @xmath268 cf .
* ( 8.32 ) ) .
next , for each @xmath269 we set @xmath270 this space carries a natural probability measure @xmath271 , [ nuydef ] the properties of which are discussed in ( * ? ? ?
* sec . 7 ) and ( * ? ? ?
* sec . 5 ) .
the function @xmath272 $ ] is then given by @xmath273 where @xmath274 .
we now give the proof of theorem [ phi0zerosmallthm ] , viz .
the fact that @xmath275 for all @xmath101 , @xmath102 . using the @xmath258-invariance of @xmath271 ( (
* lemma 7.2 ) ) we may rewrite as @xmath276 with @xmath277 we will keep these choices of @xmath278 and @xmath279 from now on . recall the splitting @xmath280 where @xmath281 and where we may take @xmath282 , cf .
* ( 7.11)(7.12 ) ) .
note that if @xmath283 is such that @xmath284 with @xmath285 then @xmath286 ; hence may be rewritten as @xmath287 writing @xmath288 ( thus @xmath289 ) we now set @xmath290 so that @xmath291 .
then by ( * ? ? ?
* ( 7.15 ) ) we have @xmath292 where @xmath293 and the restriction of the measure @xmath271 to @xmath294 corresponds to the measure @xmath295 on @xmath296 under .
we also know that a fundamental domain for @xmath296 is given by @xmath297 , where @xmath298 is any fundamental domain for [ scrfdef ] @xmath299 .
hence @xmath300 where @xmath301 is the indicator function .
this relation immediately implies @xmath302 thus proving the upper bound in . on the other hand
, we claim that if @xmath303 is any matrix in @xmath304 with the property that @xmath305 for all @xmath306 , then @xmath307 holds for all @xmath308 .
indeed , note that for any @xmath309 , @xmath310 we have @xmath311 and if this vector lies in @xmath278 then its @xmath312-component must lie in @xmath313 , viz .
hence if we write @xmath315 and let @xmath316 be the projection @xmath317 , it follows that @xmath318 now if @xmath306 and if @xmath303 has the stated property then the above distance is @xmath319 and hence @xmath320 .
furthermore if @xmath321 then @xmath314 is impossible and we again conclude @xmath320 .
this proves the claim .
using the claim just proved together with we obtain @xmath322 and by a well - known bound ( cf .
e.g. ( * ? ? ?
* lemma 2.2 ) or @xcite ) the right hand side is @xmath323 this bound together with complete the proof of theorem [ phi0zerosmallthm ] .
@xmath324 we now turn to the case @xmath121 where we will prove the explicit formula for @xmath75 for @xmath82 small stated in theorem [ d3explthm ] . as a preparation
we first give a parametrization of @xmath294 .
let @xmath325 be given , and fix a matrix @xmath326 with @xmath327 . recall that @xmath328 ( cf .
* ( 7.15 ) ) ) .
now an arbitrary matrix @xmath329 in @xmath330 has the form @xmath331 where @xmath332 and @xmath333 are two real vectors satisfying @xmath334 , with `` @xmath335 '' denoting vector product .
such a pair of vectors may be parametrized by @xmath336 , via the map @xmath337 let us write @xmath338_{\vecy}$ ] for the matrix @xmath329 obtained in this way .
we have thus exhibited a ( surjective ) diffeomorphism @xmath339_{\vecy}\in hm_\vecy.\end{aligned}\ ] ] note that the lattice corresponding to @xmath338_{\vecy}$ ] is @xmath340_{\vecy}={\mathbb{z}}\vecy+{\mathbb{z}}\vecp+{\mathbb{z}}\vecq_{\vecy,\vecp}(\vecx).\end{aligned}\ ] ] recall that @xmath271 is a probability measure on @xmath341 ; by restriction this gives a measure on @xmath294 ( with @xmath342 ) ; we will write @xmath271 also for the lift of this measure to @xmath330 . [ mhyjaclem ] given @xmath343 , the measure @xmath271 on @xmath330 takes the following form in the @xmath338_{\vecy}$]-parametrization : @xmath344 let @xmath345 be the haar measure on @xmath346 normalized so that @xmath347 ; thus @xmath348 in the coordinates @xmath349 ( @xmath350 , @xmath351 ) . by definition @xmath271 on @xmath352
is the measure which corresponds to @xmath353 on @xmath346 under @xmath354 ( cf .
recall that both the set @xmath330 and the measure @xmath271 on this set are independent of the choice of @xmath355 ( subject to @xmath327 ) .
in particular if @xmath279 is fixed and @xmath356 is a fixed rotation then we may choose @xmath357 ; using this together with @xmath358 and the fact that @xmath359 preserves the lebesgue measure @xmath360 , one checks that if holds for @xmath279 then it also holds with @xmath279 replaced by @xmath361 .
hence it suffices to prove in the case @xmath362 , @xmath363 , and we may then assume @xmath364 to be the diagonal matrix @xmath365 $ ] . in this case we compute that @xmath366_\vecy m_\vecy^{-1}={\left ( \begin{matrix } 1 & { \mathbf{0}}\\ \trans\vecv & a \end{matrix } \right ) } \in h$ ] , with @xmath367 and , introducing variables @xmath368 and @xmath369 through @xmath370 , @xmath371 note that the last matrix product is the iwasawa decomposition of @xmath372 , in terms of which the ( normalized ) haar measure takes the form @xmath373 ( cf . below for the case of general @xmath172 ) . hence @xmath374 we keep @xmath121 .
let @xmath101 and @xmath130 be given , and assume @xmath375 .
( the case @xmath376 then follows by continuity . )
recall that @xmath377 where @xmath378 and @xmath274 .
if @xmath379 then we let @xmath380 be the line along the ( unique ) chord in @xmath381 with has midpoint @xmath382 , and let @xmath383 be the affine plane @xmath384 finally let @xmath385 be that open halfspace which has boundary @xmath126 and which contains the axis of @xmath278 , viz . @xmath386 .
if @xmath387 then we modify this definition by letting @xmath388 be an arbitrary diameter of @xmath381 and @xmath389 be any of the two open halfspaces determined by @xmath126 .
now consider the map @xmath390_\vecy\in x_1(\vece_1,\vecy).\end{aligned}\ ] ] let us first prove that the image of @xmath391 equals , up to a set of @xmath271-measure zero , @xmath392 indeed , every @xmath393 in the image of @xmath391 clearly satisfies @xmath394 , since @xmath395_\vecy$ ] . on the other hand , if @xmath329 is any given element in @xmath294 which satisfies @xmath394 and also @xmath396 ( this latter condition holds for @xmath271-almost all @xmath393 ) , then we will prove that @xmath397 for some @xmath398 . to this end , among the finitely many points in @xmath399 we pick one which has minimal distance to the line @xmath400 , and call it @xmath401 .
note that @xmath402 , due to our assumption @xmath396 . if @xmath403 then set @xmath404 ; otherwise set @xmath405 . in both cases
@xmath406 must hold ( this follows from the fact that among the two regions into which the line @xmath388 splits the unit disc @xmath381 , the smaller one is mapped into the larger one by reflection in the point @xmath382 ) ; also @xmath407 and @xmath408 has the same distance as @xmath401 to the line @xmath400 .
note also that @xmath409 , since @xmath410 forces @xmath411 while all points in @xmath412 have @xmath312-coordinates in @xmath413 .
now @xmath414 , for otherwise there would exist a point @xmath415 lying in the triangle @xmath416 ( the convex hull of @xmath417 ) ; this point @xmath418 would belong to @xmath278 since @xmath278 is convex , and @xmath418 would also lie closer to @xmath400 than @xmath408 does , thus causing a contradiction .
it follows from @xmath414 that there exists a point @xmath419 satisfying @xmath420 and @xmath334 .
then @xmath421 for some @xmath422 ( cf . ) , and replacing @xmath423 by @xmath424 with appropriate @xmath425 we may assume @xmath426
. now @xmath427_\vecy$ ] ( cf . ) , and this implies @xmath428_\vecy = j(\langle\vecp,\vecx\rangle)$ ] , thus completing the proof of our claim .
next we prove that @xmath391 is injective .
thus assume @xmath429 for some @xmath430 , @xmath431 .
then @xmath432 ( cf . ) .
we now claim that @xmath433 indeed , assume that there exists a point @xmath434 , for some non - zero integer @xmath31 .
then the tetrahedron with vertices @xmath435 has volume @xmath436 ( cf . )
; hence after a scaling and a translation we obtain a tetrahedron which is contained in the closed cylinder @xmath437\times\overline{\scrb_1 ^ 2}$ ] , which has one vertex at @xmath128 and another at @xmath129 , and which has volume @xmath438 .
this is impossible , since @xmath375 ( recall the definition of @xmath439 given just before the statement of theorem [ d3explthm ] ) ; hence is proved .
now @xmath440 , @xmath441 and imply @xmath442 .
similarly @xmath443 .
it follows that @xmath444 for some @xmath445 , @xmath446 . if @xmath447 then since both @xmath448 and @xmath449 lie in the interval @xmath450 we must have @xmath451 , viz .
it follows that the midpoint of the line segment between @xmath408 and @xmath401 is @xmath453 , and this contradicts the assumption that both @xmath408 and @xmath401 lie in @xmath389 .
hence we must have @xmath454 ; and by again using @xmath448 , @xmath455 we get @xmath456 , viz .
@xmath457 . finally using , and @xmath458 we see that also @xmath459 must hold .
this completes the proof that @xmath391 is injective .
it follows that @xmath391 is a diffeomorphism of @xmath460 onto an open subset of full ( @xmath271-)measure in @xmath461 .
hence by and lemma [ mhyjaclem ] , @xmath462 this completes the proof of theorem [ d3explthm ] .
@xmath324 @xmath324 @xmath324 using methods that are beyond the scope of the present paper we are able to prove that the function @xmath439 gives the true range of validity of the formula , i.e. for any @xmath130 , there exist triples @xmath463 arbitrarily near @xmath464 at which fails .
we next give sharp lower and upper bounds on @xmath439 .
recall that we defined ( on p. ) @xmath439 for @xmath130 as @xmath125 , where @xmath126 is the largest possible volume of a tetrahedron which is contained in the closed cylinder @xmath465\times\overline{\scrb_1 ^ 2}$ ] and which has one vertex at @xmath128 and another at @xmath129 . by applying the same formula to arbitrary @xmath466
we obtain an extension of @xmath467 to a continuous function @xmath468 ; we will write @xmath467 also for this extension . [ xi1boundslem ] we have @xmath469 here @xmath470 holds if and only if @xmath471 and @xmath472 ; and @xmath473 holds if and only if @xmath474 . note that this implies that @xmath133 for all @xmath130 , both bounds being sharp .
let @xmath475 be the maximal volume of a tetrahedron contained in @xmath476 .
this volume is clearly attained ; let us fix @xmath477 to be a tetrahedron of volume @xmath475 . by a simple variational argument , varying the vertices of @xmath478 one at a time
, we see that @xmath478 may be continuously deformed , keeping its volume @xmath475 fixed , into a tetrahedron @xmath479 which has all its vertices lying on the two circles @xmath480 and @xmath481 .
clearly each of these circles must contain at least one vertex .
if each circle contains two vertices , say @xmath482 and @xmath483 , then the same type of variational argument also shows that both vectors @xmath484 and @xmath485 must be orthogonal to @xmath486 , and both @xmath487 must be orthogonal to @xmath488 . in other words ,
the line segment @xmath489 must be a diameter of @xmath480 , and the line segment @xmath490 must be that diameter of @xmath481 whose direction is orthogonal to @xmath489 .
we compute that any such tetrahedron @xmath479 has volume @xmath491 . on the other hand
if one circle contains three of the vertices , say @xmath492 , then @xmath493 must be an equilateral triangle , and @xmath494 , a contradiction .
it follows that @xmath495 .
note also that for a tetrahedron @xmath479 with @xmath482 and @xmath483 and volume @xmath495 ( i.e. with @xmath489 and @xmath490 being diameters whose directions are orthogonal ) , any perturbation of @xmath496 inside @xmath476 with @xmath497 fixed makes @xmath498 _ strictly smaller _
( since the plane through @xmath496 orthogonal to @xmath499 contains _ only _ @xmath496 in its intersection with @xmath476 ) .
this implies that also the original tetrahedron @xmath478 necessarily had two vertices on @xmath480 and two vertices on @xmath481 , i.e. we have proved that a tetrahedron with vertices @xmath500 attains the maximal volume @xmath495 _ if and only if , _ up to a renaming of the vertices , @xmath489 is a diameter of @xmath480 and @xmath490 is a diameter of @xmath480 whose direction is orthogonal to @xmath489 .
this result immediately implies that @xmath501 for all @xmath466 , with equality if and only if @xmath471 and @xmath472 .
next , given any @xmath466 , let us consider the tetrahedron @xmath478 which has vertices @xmath128 , @xmath129 , @xmath502 , @xmath503 , with @xmath504 chosen so as to maximize the area of the triangle @xmath505 .
a simple variational argument , varying @xmath496 and @xmath506 one at a time , shows that these @xmath487 must satisfy @xmath507 and @xmath508 , and @xmath496 and @xmath509 must lie on the same side of the line @xmath510 , and @xmath506 and @xmath509 must lie on the same side of the line @xmath511 . if @xmath512 then this determines @xmath489 to be the unique chord of @xmath513 with midpoint @xmath514 where @xmath515 , and we get @xmath516}{{\textstyle \frac { 1}{3}}}(\alpha+\|\vecw\|)\sqrt{1-\alpha^2 } = { { \textstyle \frac { 1}{3}}}\bigl(f(\|\vecw\|)+\|\vecw\|\bigr)\sqrt{1-f(\|\vecw\|)^2}.\end{aligned}\ ] ] on the other hand if @xmath517 then the maximal area of @xmath505 is attained if and only if @xmath507 are orthogonal , and then @xmath518 , i.e. still holds .
it is clear from the first expression in that @xmath519 is a strictly increasing function of @xmath520 $ ] ; in particular we have @xmath521 , with equality if and only if @xmath517 .
this implies that @xmath522 , where equality is possible only if @xmath517 .
by instead taking @xmath478 to have vertices @xmath128 , @xmath129 , @xmath523 , @xmath524 , we see that @xmath473 can only hold if @xmath525 . to complete the proof of the lemma it now only remains to prove that @xmath526 .
thus let @xmath478 be a tetrahedron with vertices @xmath527 and of maximal volume subject to @xmath528 , @xmath529 and @xmath530 .
by the same type of variational argument as in the first half of this proof we may assume @xmath531 . if @xmath487 lie on the _ same _ circle then as in the discussion leading to we get @xmath518 .
on the other hand if @xmath487 lie on distinct circles , say @xmath532 and @xmath533 , then similarly as in the first half of this proof we must have @xmath534 , and this implies @xmath518 , again .
hence @xmath526 .
using and theorem [ d3explthm ] we have , whenever @xmath535 , @xmath536 substituting @xmath537 we get , writing @xmath538 : @xmath539 ( cf . ) . hence holds . to see that @xmath140 is a strictly increasing function of @xmath147 we may e.g. note that , for @xmath540 , @xmath541 and here @xmath542_{r=1-w}^{r=1+w } = 0,\end{aligned}\ ] ] with @xmath543 being negative for @xmath544 and positive for @xmath545 . furthermore @xmath546 is strictly increasing for @xmath547 .
hence it follows from that @xmath548 for all @xmath540 , as desired .
the formulas @xmath141 and @xmath142 follow by straighforward computations directly from the definitions ; for example for the computation of @xmath549 one uses the fact that the following is a primitive function of @xmath550 : @xmath551 to complete the proof of corollary [ d3explthmcor1 ] it now only remains to prove that @xmath552 it is clear from the definition of @xmath439 that @xmath553 , where @xmath126 is the largest possible volume of a tetrahedron which is contained in the closed cylinder @xmath127\times\overline{\scrb_1 ^ 2}$ ] and which has one vertex at @xmath129 . by simple variational arguments of the same type as in the proof of lemma [ xi1boundslem ] we see that this volume @xmath126 is attained either for a tetrahedron whose three other vertices lie on @xmath480 and form an equilateral triangle , or else for a tetrahedron with vertices @xmath554 where @xmath555 , the line segment between @xmath129 and @xmath496 contains the point @xmath556 , and
@xmath557 is that diameter of @xmath480 whose direction is orthogonal to @xmath558 .
a tetrahedron of the first type has volume @xmath559 and a tetrahedron of the second type has volume @xmath560 ; hence follows . by , theorem [ d3explthm ] and lemma [ xi1boundslem ] we have , whenever @xmath134 , @xmath561 writing @xmath537 and @xmath562 as in the proof of corollary [ d3explthmcor1 ] we have @xmath563 here the inner integral equals @xmath564 , and thus the above expression evaluates to @xmath565 ( the last step is by a straightforward computation , which is much simplified by re - using facts from the computation of @xmath566 and @xmath549 in the proof of corollary [ d3explthmcor1 ] . ) hence we obtain the formula for @xmath567 stated in corollary [ d3explthmcor2 ] .
furthermore , using @xmath568 ( cf . and )
, we also obtain the stated formula for @xmath144 , for all @xmath134 .
finally , by and theorem [ d3explthm ] we have , whenever @xmath569 , @xmath570 this gives the stated formula for @xmath145 , since @xmath571 by .
we now describe how the graphs of @xmath144 and @xmath145 in figures [ phiplot ] and [ phi0plot ] were obtained . for @xmath121 and
any @xmath572 we have , by ( * ? ? ?
* ( 4.3 ) and ( 3.8 ) ) , @xmath573 the first curve in figure [ phiplot ] was obtained by using this formula to evaluate @xmath574 for @xmath575 , @xmath576 ; taking this to be an approximation of @xmath577 , and drawing the piecewise linear curve connecting these points .
we stress that @xmath144 as well as @xmath145 are known to be continuous and decreasing functions , cf . , and
* lemma 7.11 ) .
the second curve in figure [ phiplot ] was obtained similarly , using @xmath578 , @xmath579 . in order to evaluate numerically , the integral over @xmath252
was replaced by an average over the hecke points corresponding to a large prime @xmath33 , shifted by a fixed rotation ; viz .
@xmath580 with fixed @xmath581 and @xmath478 running through the set @xmath582 also the integral over @xmath583 was replaced by an average over the @xmath584 points in @xmath585 , for some fixed @xmath586 and @xmath587 . for any fixed @xmath588
, this approximation is known to approach the correct value as @xmath589 ; cf .
, e.g. , @xcite , @xcite , @xcite . in our numerical experiments we noted that the rate of convergence seems to be improved by taking @xmath590 irrational and also taking @xmath591 to be `` sufficiently generic '' . for the curves in figure [ phiplot ] we used @xmath592 , @xmath593 , @xmath594 and @xmath595 .
we did not prove any error bounds for our approximation ; however as an indication of the error we mention that for @xmath31 with @xmath596 , the value which we obtained for @xmath574 always differed by less than @xmath597 from the known exact values of both @xmath574 and @xmath577 ( cf .
corollary [ d3explthmcor2 ] ) . also repeated runs with other choices of @xmath598 indicate that our values for @xmath574 are correct to within an absolute error @xmath599 , for all @xmath31 . similarly , for figure [ phi0plot ] we used the formula ( ( * ? ? ? * ( 4.3 ) and ( 3.8 ) ) ) @xmath600 to evaluate @xmath601 for @xmath575 , @xmath602 . for this case
our experiments suggest that , for a given number of sample points @xmath580 , we get a significantly better approximation of the @xmath252-integral by running @xmath478 through a _ random _ subset of @xmath603 with @xmath33 quite large , than by running @xmath478 through _ all _ of @xmath603 for a @xmath33 of more modest size .
the curve in figure [ phi0plot ] was obtained by using @xmath604 , @xmath591 as above , and letting @xmath478 run through @xmath605 randomly choosen points from @xmath603 .
comparison against the known @xmath79-values for @xmath606 ( cf .
corollary [ d3explthmcor2 ] ) , as well as comparisons versus the results of using other random seeds and/or other choices of @xmath591 and @xmath33 , indicate that our values for @xmath607 are correct to within an absolute error @xmath608 . regarding the support of @xmath609 , recall from proposition [ phizeroasympt ] and
the ensuing comments that @xmath235 holds if and only if @xmath610 . our numerics show that the function @xmath145 approaches zero quite quickly as @xmath82 approaches @xmath611 , and the largest value of @xmath612 which we saw among our sample points @xmath329 was @xmath613 .
in this section we prove theorem [ phixiwasymptthm ] on the asymptotic size of @xmath186 as @xmath76 . along the way we prove several lemmas which will also be useful later in our proof of theorem [ phi0xilargethm ] in sections [ parabapprsec][phi0xiwzasymptsec ] .
note that a second proof of theorem [ phixiwasymptthm ] will be given in section [ partialphixizsec ] , where we deduce theorem [ phixiwasymptthm ] as a consequence of theorem [ phi0xilargethm ] , using the integration formula .
recall that we write @xmath249 .
let @xmath303 be the subgroup of diagonal matrices with positive entries @xmath614 and let @xmath615 be the subgroup of upper triangular matrices @xmath616 every element @xmath254 has a unique iwasawa decomposition @xmath617 with @xmath618
. in these coordinates the haar measure takes the form ( @xcite ) @xmath619 where @xmath620 , @xmath621 , @xmath622 , are ( left and right ) haar measures of @xmath615 , @xmath303 , @xmath623 , normalized by @xmath624 , @xmath625 and @xmath626 . for @xmath627 one
has @xmath628 we set @xmath629 , \ :
1\leq j < k\leq d\bigr\}$ ] ; then @xmath630 is a fundamental region for @xmath631 .
we define the following siegel set : @xmath632 it is known that @xmath633 contains a fundamental region for @xmath634 , and on the other hand @xmath633 is contained in a finite union of fundamental regions for @xmath257 ( @xcite ) . given @xmath635 ,
its row vectors are @xmath636 thus @xmath637 is a basis of the lattice @xmath638 . if @xmath639 then we see that , for all @xmath591 , @xmath640 throughout the paper we will let @xmath641 denote certain constants which we fix once and for all and which only depend on @xmath172 ( or in some cases are absolute ) ; the @xmath172-dependence will mostly be suppressed but if necessary it will be made explicit by writing `` @xmath642 '' .
the bound implies that if @xmath639 and if the lattice @xmath638 has empty intersection with a large ball , then @xmath643 must be large : [ a1largelem ] for any @xmath644 such that the lattice @xmath638 is disjoint from some ball of radius @xmath645 in @xmath253 , we have @xmath646 . choose @xmath647 so that @xmath648 is the center of the given ball .
let @xmath649 be the integer nearest to @xmath650 .
then @xmath651 is a lattice point of @xmath638 , and has distance @xmath652 to @xmath408 .
this distance must be @xmath653 ; hence @xmath646 .
the set of lattices in @xmath252 which have a representative @xmath639 with @xmath643 larger than some large fixed number @xmath303 , may in an approximate sense be parametrized by the set @xmath654 . in this section
we prove a version of this fact , lemma [ fdcontainmentslem ] below , which we will make use of several times .
let us fix a function @xmath48 ( smooth except possibly at one point , say ) @xmath655 such that @xmath656 for all @xmath657 .
given @xmath635 , the matrices @xmath658 , @xmath659 and @xmath660 can be split uniquely as @xmath661 where @xmath662 , @xmath663 , @xmath664 , @xmath665 and @xmath666 , @xmath667 .
we set @xmath668 in this way we get a bijection between @xmath258 and @xmath669 ; we write @xmath670 $ ] for the element in @xmath258 corresponding to the 4-tuple @xmath671 .
in particular note that @xmath672\in g{\ : : \:}{\myutilde{m}}\in{\mathcal{s}}_{d-1},\ : { \utilde{a}}_1\leq { { \textstyle \frac { 2}{\sqrt 3}}}a_1^{\frac d{d-1 } } , \:\vecu\in(-{{\textstyle \frac { 1}{2}}},{{\textstyle \frac { 1}{2}}}]^{d-1}\bigr\}\qquad \\\label{sidsubssidm1 } \subset\bigl\{[a_1,\vecv,\vecu,{\myutilde{m}}]\in g{\ : : \:}{\myutilde{m}}\in{\mathcal{s}}_{d-1},\:\vecu\in(-{{\textstyle \frac { 1}{2}}},{{\textstyle \frac { 1}{2}}}]^{d-1}\bigr\}.\end{aligned}\ ] ] one checks by a straightforward computation using that the haar measure @xmath256 takes the following form in the parametrization @xmath670 $ ] : @xmath673 where @xmath674 is the @xmath675-dimensional volume measure on @xmath11 .
note that all of the above claims are valid also for @xmath0 , with the natural interpretation that @xmath676 with @xmath677 .
we will also need to know the explicit expression of the lattice @xmath638 in terms of @xmath678 : one computes that , for any @xmath309 and @xmath679 , @xmath680 in particular we always have @xmath681 let us fix a subset @xmath682 which contains exactly one of the vectors @xmath683 and @xmath684 for every @xmath685 .
let us also fix a ( set theoretical , measurable ) fundamental region @xmath686 for @xmath299 .
now for @xmath687 we set @xmath688\in g { \ : : \:}a_1>a,\ : \vecv\in\s_\pm^{d-1},\ : \vecu\in(-{{\textstyle \frac { 1}{2}}},{{\textstyle \frac { 1}{2}}}]^{d-1},\ : { \myutilde{m}}\in{\mathcal{f}}_{d-1}\bigr\}.\end{aligned}\ ] ] [ fdregionlongvector ] if @xmath689 satisfy @xmath690 for some @xmath691 , and if @xmath692 in the iwasawa decompositions @xmath693 , @xmath694 , then @xmath695 .
assume that @xmath696 $ ] and @xmath697 $ ] satisfy the assumptions of the lemma .
then @xmath698 , and this lattice has a basis @xmath637 ( cf . ) , and also a basis @xmath699 , the row vectors of @xmath700 .
now for each @xmath701 with @xmath702 the lattice vector @xmath703 has length @xmath704 , since @xmath705 . on the other hand , by a similar argument as in
, using @xmath706 , we have @xmath707 for each @xmath708 ; thus @xmath709 .
similarly @xmath710 for each @xmath711 .
hence @xmath712 .
let @xmath713 be the hyperplane spanned by this set of vectors .
now @xmath714 ( @xmath675-dimensional volume ) , and similarly for @xmath715 ; hence @xmath716 . also @xmath717 and similarly @xmath718
; hence since @xmath719 we conclude @xmath720 .
next by , if @xmath721 denotes the embedding @xmath722 , then @xmath723 and similarly for @xmath724 ; hence @xmath725 , and since @xmath726 we conclude @xmath727
. hence also @xmath728 , @xmath729 and @xmath730 ( in an obvious notation , cf . ) , and we now also obtain @xmath731 and @xmath732 , so that @xmath690 implies @xmath733 . but @xmath733 together with @xmath734 and @xmath735^{d-1}$ ] imply @xmath736 , thus @xmath737 .
[ fundlongdomainlem ] if @xmath644 has @xmath738 and @xmath739 , then there is some @xmath740 such that @xmath741 .
take any @xmath696\in{\mathcal{s}}_d$ ] with @xmath738 and @xmath739 .
we write @xmath742 as usual .
if @xmath743 then we replace @xmath329 with @xmath744 , where @xmath745\in\gamma$ ] and @xmath746 is chosen so that @xmath747 ; this new @xmath329 lies in @xmath633 and has the same @xmath659 component as before but @xmath683 negated .
hence from now on we may assume @xmath748 .
take @xmath749 so that @xmath750 ; let the iwasawa decomposition of this matrix be @xmath751 , and let @xmath105 be the unique vector in @xmath752 with @xmath753^{d-1}$ ] .
set @xmath754 .
we now claim @xmath755 . to prove this
, first note that @xmath756 has iwasawa decomposition @xmath757 from this we see by inspection that @xmath758 .
next , for each @xmath759 with @xmath702 we have @xmath760 , since @xmath761 .
hence any basis for the lattice @xmath762 must have at least one basis vector of length @xmath763 .
but as in we see that @xmath764 has a basis where each basis vector has length @xmath765 . hence @xmath766 .
using this fact together with @xmath767 we see that @xmath768 ( cf . and ) .
let us define @xmath769\in { \mathcal{s}}_d{\ : : \:}\vecv\in\s_\pm^{d-1}\bigr\}. \ ] ] then @xmath770 contains a fundamental region for @xmath771 ( viz .
@xmath772 ) , by the argument in the beginning of the proof of lemma [ fundlongdomainlem ] .
furthermore we have @xmath773 .
[ fdcontainmentslem ] there exists a ( set - theoretical , measurable ) fundamental region @xmath774 for @xmath771 which satisfies @xmath775 and @xmath776 where @xmath777\in g{\ : : \:}a_1>a,\ : a_2\geq(c_{{1}}^{(d-1)})^{-1}a,\ : \vecv\in\s_\pm^{d-1},\ : \vecu\in(-{{\textstyle \frac { 1}{2}}},{{\textstyle \frac { 1}{2}}}]^{d-1},\ : { \myutilde{m}}\in{\mathcal{s}}_{d-1}\bigr\}.\end{aligned}\ ] ] let @xmath778 ; let @xmath779 be an arbitrary measurable subset of @xmath780 which contains exactly one representative from each @xmath263-coset which intersects @xmath780 but does not intersect @xmath781 , and then let @xmath782 be an arbitrary measurable subset of @xmath770 which contains exactly one representative from each @xmath263-coset of @xmath258 which does not intersect @xmath783 .
( @xmath782 exists since @xmath772 . ) finally set @xmath784 .
we have @xmath785 since @xmath786 implies @xmath787 .
hence @xmath775 . by lemma [ fdregionlongvector ]
any two distinct elements in @xmath781 are @xmath263-inequivalent ; hence any two distinct elements in @xmath774 are @xmath263-inequivalent . on the other hand we have @xmath788 by construction ; hence @xmath774 is a fundamental region for @xmath771 .
finally the first inclusion in holds since @xmath789 by construction , and the second inclusion holds since @xmath790 by construction and @xmath791 by lemma [ fundlongdomainlem ] . in this section
we give a short proof of the following asymptotic formula .
[ phixilargethmint ] for any @xmath792 , @xmath793 note that @xmath794 gives the limit probability that a particle in the lorentz gas starting from a generic initial point inside the billiard domain travels length @xmath795 ( in macroscopic coordinates ) before its first collision .
in particular already theorem [ phixilargethmint ] sharpens the upper bound given by bourgain , golse and wennberg @xcite and the lower bound of golse and wennberg @xcite .
note that we will later prove theorem [ phixilargethm ] which gives an asymptotic formula for @xmath144 itself and which immediately implies theorem [ phixilargethmint ] .
however the following short direct proof of theorem [ phixilargethmint ] gives a first illustration of the usage of lemma [ fdcontainmentslem ] in a simple case . along the way we will prove some auxiliary results which we will need later anyway . by definition ( cf .
* ( 3.8 ) , ( 4.3 ) ) ) , @xmath796 where @xmath797 ; however since the right hand side of is invariant under @xmath798 where @xmath478 is an arbitrary volume preserving affine linear map , we may just as well take @xmath799 in
. then @xmath278 contains a ball of radius @xmath800 and hence by lemma [ a1largelem ] there is a constant @xmath801 which only depends on @xmath172 such that @xmath802 holds for every @xmath644 for which the inner integral in is nonzero . from now on we keep @xmath82 so large that @xmath687 .
fix @xmath775 to be a fundamental region for @xmath771 as in lemma [ fdcontainmentslem ] , applied with our @xmath803 .
now lemma [ fdcontainmentslem ] together with and @xmath804 imply @xmath805 if @xmath197 then using and for @xmath164 we obtain @xmath806 uniformly over all @xmath807 . hence the error term in
is @xmath808 . on the other hand
if @xmath0 then the error term vanishes for all sufficiently large @xmath82 .
the inner integral in the main term in remains the same if @xmath809 is replaced by any other fundamental region for @xmath810 .
since @xmath638 is spanned by the vectors @xmath637 ( cf . ) , one choice of a fundamental region for @xmath810 is @xmath811 .
hence , using and , we obtain @xmath812^{d-1 } } \int_{{\mathcal{f}}_{d-1 } } \int_{\prod_{j=1}^d [ -\frac{a_j}2,\frac{a_j}2 ) } i\bigl(({\mathbb{z}}^d[a_1,\vecv,\vecu,{\myutilde{m}}]+\vecw { \mathsf{k}})\cap \fz=\emptyset \bigr ) \\\label{phiirrasymptlarge2pf3 } \times d\vecw \ , d\mu({\myutilde{m}})\,d\vecu\,d\vecv\,\frac{da_1 } { a_1^{d+1 } } + o(\xi^{-2}),\end{aligned}\ ] ] where @xmath813 [ hsdef ] and @xmath814 ( cf . ) .
now for any @xmath643 , @xmath683 , @xmath815 , @xmath816 and @xmath817 appearing in the above integral we have , by and since @xmath818 , @xmath819+\vecw{\mathsf{k}}\subset \bigcup_{n\in{\mathbb{z } } } \bigl((na_1+w_1)\vecv+\vecv^\perp\bigr).\end{aligned}\ ] ] [ cylplanelemma ] let @xmath820 , so that the angle @xmath821 between @xmath683 and @xmath312 satisfies @xmath822 .
set @xmath823 as in .
then for any @xmath824 , the intersection @xmath825 is nonempty if and only if @xmath826 , and in this case we have @xmath827 after a rotation in the variables @xmath828 we may assume @xmath829 with @xmath830 and @xmath831 . by symmetry
we may also assume @xmath832 .
now @xmath833 , and we see that if @xmath834 lies in the intersection @xmath835 then so does @xmath836 . in particular @xmath835 is nonempty if and only if @xmath837 , thus proving the first assertion .
now assume @xmath838 .
let @xmath316 be the projection @xmath317 , and note that @xmath839 since @xmath840 , it follows that @xmath841 but note that @xmath842 scales volume with a factor @xmath843 ; hence @xmath844 here @xmath845 and @xmath846 here the third entry is redundant in the minimum , since @xmath847 .
also note that the minimum equals its first entry if and only if @xmath848 and in this case we have @xmath849 and thus @xmath850 ( since @xmath830 , @xmath831 , @xmath851 ) . furthermore the minimum equals its fourth entry if and only if @xmath852 and in this case we necessarily have @xmath853 . hence @xmath854 and
this concludes the proof of the first relation in .
the second relation in is obvious . by lemma [ cylplanelemma ] , the cylinder @xmath855 has nonempty intersection with the hyperplane @xmath856 if and only if @xmath857 .
it follows that if @xmath858 then none of the hyperplanes in intersect @xmath278 , and thus @xmath859+\vecw{\mathsf{k}})\cap\fz=\emptyset$ ] , independently of @xmath860 @xmath816 or @xmath861 .
hence , if we restrict the range of integration in further by @xmath862 then the resulting integral equals @xmath863 here in the last step we used the fact that @xmath864 for all @xmath851 . substituting @xmath865 the integral is seen to equal @xmath866 . hence we obtain the main term in theorem [ phixilargethmint ] , and to complete the proof of theorem [ phixilargethmint ] we now only have to prove that , if we denote by @xmath867 the remaining integral , viz . the integral in with range of integration further restricted by @xmath868 , then @xmath869 to prove this , for any @xmath870 appearing in the integral , we write @xmath871 with @xmath872 ; then @xmath873 , and hence using we see that @xmath874+\vecw{\mathsf{k}}\supset a_1^{-\frac1{d-1}}\iota\bigl({\mathbb{z}}^{d-1}{\myutilde{m}}+a_1^{\frac1{d-1}}\vecw'{\utilde{\mathsf{k}}}\bigr ) f(\vecv)+w_1\vecv,\end{aligned}\ ] ] where ( here and throughout the rest of the paper ) @xmath721 denotes the embedding @xmath875 it follows that @xmath876 here note that @xmath877 is isometric with @xmath878 , and for any @xmath879 with @xmath822 and any @xmath880 with @xmath868 , the set @xmath878 is convex and has @xmath675-dimensional volume @xmath881 , by lemma [ cylplanelemma ] .
hence for any @xmath738 we have @xmath882 note also that @xmath883 is a fundamental domain for @xmath884 for any given @xmath643 and @xmath816 ( cf . ) .
[ aslamboundlem ] for any @xmath792 and any convex subset @xmath885 , we have @xmath886 this is a straightforward modification of ( * ? ? ?
* lemma 2.5 ) .
in fact the bound holds for any _ measurable _ @xmath885 , as can be proved by using ( * ? ? ?
2.2 ) , after first rewriting the left hand side of as @xmath887 . if @xmath197 then by lemma [ aslamboundlem ] ( applied with @xmath164 ) and the discussion preceding it , we obtain @xmath888 i.e. we have proved . in the remaining case
@xmath0 we have @xmath889 and hence the integrand in vanishes whenever @xmath890 ; hence we obtain @xmath891 .
this completes the proof of theorem [ phixilargethmint ] .
@xmath324 @xmath324 we now start with the proof of the asymptotic formula for @xmath186 , theorem [ phixiwasymptthm ] .
note that if @xmath0 then theorem [ phixiwasymptthm ] ( with @xmath892 ) follows directly from the explicit formula in @xcite .
hence we will from now on assume @xmath197 .
from we have , using the right @xmath258-invariance of @xmath256 , @xmath893 where @xmath894 note in particular that @xmath278 does not denote the same cylinder as in the above proof of theorem [ phixilargethmint ] .
the choice of @xmath278 in will be in force for the rest of section [ preliminariessec ] .
it follows from proposition [ phi0supportthm ] and ( or more directly from ( * ? ? ?
1.4 ) ) that there is a constant @xmath895 which only depends on @xmath172 such that @xmath896 whenever @xmath897 .
the function @xmath212 appearing in the right hand side of will be defined in below , and it will be clear from this definition that @xmath212 vanishes for large @xmath8 ( cf .
lemma [ xisupboundlem ] or ) ; hence we may assume that @xmath898 is so large that @xmath899 for all @xmath900 .
this means that is automatic when @xmath897 .
hence from now on we will assume @xmath901 .
by lemma [ a1largelem ] there is a constant @xmath902 which only depends on @xmath172 such that @xmath903 holds for all @xmath639 with @xmath904 .
we will assume from start that @xmath905 ; in particular we have @xmath906 and @xmath687 .
applying lemma [ fdcontainmentslem ] in the same way as in we get @xmath907^{d-1 } } \int_{{\mathcal{f}}_{d-1 } } i\bigl({\mathbb{z}}^d[a_1,\vecv,\vecu,{\myutilde{m}}]\cap\fz = \emptyset\bigr)\ , d\mu^{(d-1)}({\myutilde{m } } ) \,d\vecu\,\frac{da_1}{a_1^{d+1}}\,d\vecv \\\label{phi2xizfirstasympt } + o\bigl(\xi^{-2}\bigr),\end{aligned}\ ] ] where ( again ) @xmath813 .
we parametrize a dense open subset of @xmath908 as follows ( recall that we are assuming @xmath197 ) : @xmath909 where @xmath851 , @xmath910 and @xmath911 .
thus @xmath821 is the angle between @xmath683 and @xmath312 , and @xmath912 is the angle between @xmath913 and @xmath312 in @xmath161 .
the @xmath675-dimensional volume measure on @xmath11 takes the following form in our parametrization : @xmath914 where @xmath915 is the @xmath916-dimensional volume measure on @xmath917 ( if @xmath121 : @xmath915 is the counting measure on @xmath918 ) . for any @xmath683 , @xmath643 , @xmath815 as above we have , using with @xmath919 only : @xmath920\cap\fz = \emptyset\bigr)\ , d\mu({\myutilde{m } } ) \leq\int_{{\mathcal{f}}_{d-1 } } i\bigl({\mathbb{z}}^{d-1}{\myutilde{m}}\cap a_1^{\frac1{d-1}}\fz_\vecv= \emptyset\bigr)\ , d\mu({\myutilde{m}}),\end{aligned}\ ] ] where @xmath921 ( recall the definition of @xmath721 in ) .
note that @xmath922 is isometric with @xmath923 .
now @xmath922 contains a @xmath675-dimensional relatively open cone of volume @xmath924 with @xmath509 in its base ( cf .
* lemma 7.1 ) ) .
hence by ( * ? ? ?
1.4 ) , @xmath925 let @xmath926 be an arbitrary constant .
( we will later impose some conditions on @xmath927 being sufficiently small , but it will be clear that it is possible to fix @xmath927 as an absolute constant satisfying these conditions . ) now in the first line of we may restrict the range of @xmath683 to @xmath928 at the cost of an error which , by and , is @xmath929 where we denote @xmath930 our goal in sec .
[ simplintsec][furthersimplsec ] is to prove that holds with the error term replaced by @xmath931 ; then in sec
. [ slightimprsec ] we will show how to improve the error term slightly for @xmath121 so as to complete the proof of theorem [ phixiwasymptthm ] . collecting our bounds so far we have , writing @xmath670 $ ] , @xmath932^{d-1 } } \int_{{\mathcal{f}}_{d-1 } } i\bigl({\mathbb{z}}^dm\cap\fz = \emptyset\bigr)\ , d\mu({\myutilde{m } } ) \,d\vecu\,\frac{da_1}{a_1^{d+1}}\,d\vecv + o(e).\end{aligned}\ ] ] [ phixizthmpflem1 ] for any @xmath664 and any @xmath683 as in with @xmath851 and @xmath933 $ ] , we have @xmath934 for all @xmath935 if and only if @xmath936 .
( recall that @xmath913 , thus @xmath937 . )
we have @xmath938 where @xmath823 .
hence @xmath939 if and only if @xmath823 has nonempty intersection with @xmath940 by lemma [ cylplanelemma ] this holds if and only if @xmath941 .
if this inequality holds for some nonzero integer @xmath31 then it must hold for some positive integer @xmath31 , since @xmath942 by our assumptions .
but for @xmath943 we _ always _ have @xmath944 .
hence we conclude that @xmath945 holds for all @xmath935 if and only if @xmath946 for all @xmath947 , viz . if and only if @xmath948 .
[ cylplanelemmacor ] for any @xmath683 as in with @xmath851 and @xmath933 $ ] , and any @xmath643 with @xmath949 , we have ( writing @xmath670 $ ] ) @xmath950^{d-1 } } \int_{{\mathcal{f}}_{d-1 } } i\bigl({\mathbb{z}}^dm\cap\fz=\emptyset\bigr ) \ , d\mu({\myutilde{m}})\,d\vecu \ll\xi^{-1}\bigl(v_1+wv_2+\|\vecv'\|-\xi^{-\frac1d}a_1\bigr)^{-\frac d2}.\end{aligned}\ ] ]
we use with @xmath951 and the fact that for any @xmath952 , @xmath953^{d-1}{\mathsf{a}}({\utilde{a}}){\utilde{\mathsf{k}}}$ ] is a fundamental domain for @xmath884 . writing @xmath954 ( a convex subset of @xmath161 ) we then obtain that the left hand side of is bounded from above by @xmath955 here the last inequality follows from lemma [ aslamboundlem ] .
now note that @xmath956 is isometric with @xmath957 , and hence applying lemma [ cylplanelemma ] ( cf .
the proof of lemma [ phixizthmpflem1 ] ) we get @xmath958 here the last relation is obvious if @xmath959 , while if @xmath960 then it follows from @xmath961 we now obtain the stated bound .
using , and lemma [ cylplanelemmacor ] it follows that the contribution from all @xmath643 with @xmath962 in is , writing @xmath963 and using @xmath964 , @xmath965 where @xmath966 is as in .
but for @xmath967 we have @xmath968\cap\fz=\emptyset$ ] if and only if @xmath969 , by , and lemma [ phixizthmpflem1 ] .
hence @xmath970 next we wish replace the lower integration limit for @xmath643 by @xmath971 .
note that for every @xmath972 we have @xmath973 hence using we see that the error when replacing `` @xmath974 '' with `` @xmath971 '' in is @xmath975 ( here we used our assumption @xmath901 . )
we thus conclude : @xmath976 next , for @xmath879 we define @xmath977 by @xmath978 where @xmath316 is the projection @xmath317 .
then @xmath979 is a linear map with determinant @xmath980 ( the same as the jacobian of @xmath33 restricted to @xmath981 ) .
thus @xmath982 is a map in @xmath304 , and hence by the invariance of the haar measure @xmath983 we have @xmath984 for any measurable subset @xmath985 .
note also that from the definition of the cylinder @xmath278 it is clear that @xmath986 is a subset of the cut ball @xmath987 , where @xmath988 hence @xmath989 and using this together with we get @xmath990 note that _ if _ @xmath991 ( say ) then @xmath986 is _ equal _ to @xmath987 ; hence equality holds in and hence _ equality also holds in . _
we wish to bound the contribution in from those @xmath972 which do _ not _ satisfy this condition .
[ goodintlem ] if @xmath683 satisfies @xmath851 , @xmath992 and @xmath993 , then @xmath994 the assumptions imply @xmath995 and thus there is some @xmath996 with @xmath997 ; thus @xmath998 , @xmath999 and @xmath1000 .
now @xmath1001 this gives the desired bound , since @xmath1002 .
it follows from lemma [ goodintlem ] that the contribution in from those @xmath972 which satisfy @xmath1003 is @xmath1004 this is bounded above by the expression in , and is thus @xmath1005 . in view of this bound , the inequality and the fact that equality holds in whenever @xmath991
, we obtain : @xmath1006 next we will replace @xmath1007 with a _
cut paraboloid_. we first prove three easy lemmas on the approximation of a part of a unit ball with a paraboloid .
( in the present section we will only need to take @xmath1008 in the following , but the case of general @xmath1009 is needed later in the treatment of @xmath75 . ) for any @xmath1010 with @xmath1011 , we let @xmath1012 be the paraboloid given by @xmath1013 where @xmath1014 this definition is motivated by the following lemma , which is proved by a direct calculation . [ puvdeflem ] if @xmath1015 , then @xmath1016 are the unique real numbers for which the parabola @xmath1017 is tangent to the unit circle @xmath1018 at the point @xmath1019 and also intersects the circle at @xmath1020 .
the following lemma gives the two fundamental facts which we will use about the relation between @xmath1021 and the unit ball @xmath1022 .
[ paraboloidballapprlem ] ' '' '' 1 .
for any @xmath1023 we have @xmath1024 .
2 . for any @xmath1025
we have @xmath1026 straightforward .
let us write @xmath1027 for the upper halfspace @xmath1028 also , for any @xmath1029 , we write @xmath1030 [ cutparaboloidincutballlem ] there is an absolute constant @xmath1031 such that for any @xmath1032 and @xmath1033 , if @xmath1034 then @xmath1035 we leave out the routine proof , since this lemma is a special case ( obtained when @xmath1036 ) of lemma [ cutcontainedlem2 ] which we will prove later . with these three lemmas in place
we may now approximate the integrand in from above and below . with @xmath1031 as in lemma [ cutparaboloidincutballlem ] , set ( for @xmath683 as in ) @xmath1037 from now on
we assume that @xmath82 is so large that @xmath1038 ; thus @xmath1039 .
we also require that our constant @xmath927 in should satisfy @xmath1040 .
then for all @xmath972 we have @xmath1041 , and hence by lemma [ paraboloidballapprlem](i ) and lemma [ cutparaboloidincutballlem ] , @xmath1042 we transform @xmath1043 to a `` standard '' cut paraboloid as follows .
set @xmath1044 and for @xmath1033 , @xmath1045 finally , for @xmath1046 and @xmath1047 , let @xmath1048 be the probability that a random lattice of covolume @xmath1049 is disjoint from @xmath1050 , viz.@xmath1051 by an obvious rotational symmetry the last expression is invariant under replacement of @xmath1050 by @xmath1052 for any @xmath1033 satisfying @xmath1053 .
now one checks by a quick computation that if we let @xmath1054 \in\gl(d-1,{\mathbb{r}}),\end{aligned}\ ] ] where @xmath1055 then @xmath1056 this transformation formula applies also with @xmath1057 in place of @xmath233 . note that @xmath478 scales volume with a factor @xmath1058 .
hence we obtain , for all @xmath972 , writing @xmath1059 , @xmath1060 where @xmath1061 using these bounds in we conclude @xmath1062 and @xmath1063 we will now simplify further the integral in . recall that @xmath1064 in this integral .
( note that the integral in can be viewed as a special case of the integral in , by replacing @xmath1065 by @xmath1057 . ) substituting @xmath1066 in the inner integral , and then using the parametrization and substituting further @xmath1067 and @xmath1068 , we get @xmath1069 where we use the notation @xmath1070 note that @xmath233 and @xmath1071 only depend on the integration variable @xmath10 , not on @xmath1072 or @xmath1073 . changing order between the two inner integrals gives @xmath1074 here note that for any @xmath1075 we have @xmath1076 and using this with @xmath1077 we conclude that the innermost integral in the above expression equals @xmath1078 .
hence the whole expression equals @xmath1079 next we replace @xmath10 with the new variable @xmath1080 in the outer integral , via the substitution @xmath1081 note that ( using @xmath1039 ) @xmath1082 uniformly over all @xmath1083 . here
the implied constant in the big-@xmath1084 term is absolute ( since @xmath1065 is an absolute constant ) .
hence so long as @xmath927 is sufficiently small , our substitution is a strictly increasing ( and smooth , thus with smooth inverse ) map from @xmath1085 to @xmath1086 where @xmath1087 , and furthermore , for all such @xmath1088 we have @xmath1089 hence our expression equals @xmath1090 now to bound the contribution from the error term we will use the following lemma .
[ xicvboundlem1 ] for all @xmath1091 and @xmath1047 we have @xmath1092 this follows from ( * ? ? ?
1.4 ) coupled with the fact that @xmath1050 contains a @xmath675-dimensional open cone of volume @xmath1093 with @xmath509 in its base . to give a slightly more detailed argument it is convenient to use a simple invariance relation which we will prove in section [ xidm1basicssec ] : by applied with @xmath1094 ( cf . ) , @xmath1095 and @xmath1096 , we have @xmath1097 where @xmath1098 now @xmath1099 contains the open cone with base @xmath1100 of radius @xmath1101 , and apex @xmath1102 , and this cone has volume @xmath1103 . hence the desired bound follows from ( * ? ? ?
1.4 ) . using lemma [ xicvboundlem1 ]
we see that the contribution from the error term in is ( since @xmath1104 over the whole range of integration ) @xmath1105 if @xmath181 then this is @xmath1106 , even after replacing the upper integration limit @xmath1107 by @xmath1108 .
( here we again used our assumption @xmath901 . )
however if @xmath121 then we obtain , using the fact that @xmath1109 , @xmath1110 hence in all cases the contribution from the error term in is @xmath1005 .
let us also note that for any fixed @xmath1111 , @xmath1112 is an increasing function of @xmath1113 , with derivative @xmath1114 if @xmath1115 then this derivative is @xmath1116 , by lemma [ xicvboundlem1 ] . since also @xmath1117
we conclude that the difference caused by replacing @xmath1071 by @xmath120 in is bounded by exactly the same expression as in .
hence from , we have proved that @xmath1118 ( we used the fact that the integrand is nonnegative to increase the @xmath1080-integration range from @xmath1119 to @xmath1120 . ) finally note that the same computations allow us to compute the integral in in _ exact _ terms ; we thus obtain @xmath1121 by lemma [ xicvboundlem1 ] we may here increase the upper range of @xmath1080 to @xmath1108 at the cost of an error @xmath1122 hence , using @xmath1123 and @xmath1124 , we finally obtain with @xmath1125 except that we get the error term @xmath931 ( cf . ) , which is slightly worse than the error term in when @xmath121 . note that lemma [ xicvboundlem1 ] implies that the function @xmath212 is uniformly bounded over @xmath1126 .
furthermore @xmath212 is easily seen to be continuous ; for instance this follows immediately from lemma [ xisupboundlem ] , and lemma [ xidm1contlem ] below . hence to complete the proof of theorem [ phixiwasymptthm ] it now only remains to improve the error term slightly in the case @xmath121 .
the following lemma is valid for arbitrary @xmath197 .
[ phixizthmpf2part2vanishlem ] there is a constant @xmath1127 which only depends on @xmath172 such that if @xmath820 satisfies either [ @xmath1128 and @xmath1129 or [ @xmath1130 and @xmath1131 , then the left hand side of vanishes for all @xmath1132 . for any @xmath738
, @xmath1133 contains a cone of volume @xmath1134 and edge ratio @xmath1135 ( cf .
7.1 ) ) ; hence by ( * ? ? ?
1.4 ) , there is a constant @xmath1136 such that the left hand side of vanishes whenever @xmath1137 .
the lemma follows from this .
( in the case @xmath1128 one also uses @xmath901 . ) now if @xmath1138 is sufficiently small so that @xmath1139 and @xmath1140 , then lemma [ phixizthmpf2part2vanishlem ] says that remains a valid bound on the error in question even if we restrict the integration range for @xmath1141 to @xmath1142
. this does not give any improvement if @xmath181 , but if @xmath121 it means that the error bound in is improved from @xmath1143 down to @xmath1144 .
keeping the assumption @xmath1139 , lemma [ phixizthmpf2part2vanishlem ] also says that we may restrict the range of @xmath912 in to @xmath1145 .
when @xmath121 this means that we improve the bound in down to @xmath1146 . in the same way also the bounds in and can be improved down to @xmath1146 . finally to improve the error bounds in section [ furthersimplsec ] we note the following
[ xisupboundlem ] there is a constant @xmath1147 which only depends on @xmath172 such that @xmath1148 holds whenever @xmath1149 , @xmath1150 .
this follows by noticing that the cone considered in the proof of lemma [ xicvboundlem1 ] has edge ratio @xmath1151 , and volume @xmath1152 , and applying ( * ? ? ?
1.4 ) . using lemma [ xisupboundlem ] ( and the fact that @xmath1104 for all @xmath1086 ) we see that the inner integral in vanishes unless @xmath1153 , and
thus the bound is valid also if we integrate over all @xmath1153 instead of @xmath1086 .
hence for @xmath121 we obtain the improved bound @xmath1154 . using all these improved bounds in sec .
[ simplintsec][furthersimplsec ] we conclude that , for @xmath121 , _ if _ @xmath1155 is sufficiently large so that @xmath1156 and @xmath1157 , then @xmath1158 we have already proved in sec .
[ furthersimplsec ] the same formula with the error bound @xmath1143 , valid for _ all _ @xmath1159 and large @xmath82 .
( recall that both @xmath186 and @xmath1160 _ vanish _ when @xmath1161 . ) taking these facts together it follows that holds .
this concludes the proof of theorem [ phixiwasymptthm ] .
@xmath324 @xmath324 @xmath324
we now introduce a general lattice probability function @xmath1162 involving two cut paraboloids in @xmath161 .
this is the function in terms of which we will later express our asymptotic formula for @xmath146 as @xmath76 , cf .
theorem [ phi0xilargethm ] and .
we keep @xmath197 throughout this section .
recall @xmath1163 for @xmath1033 and @xmath1164 we let @xmath1165 be the following cut translate of @xmath1166 : @xmath1167 thus @xmath1168 ( cf . ) .
now for @xmath1169 , @xmath1033 and @xmath1047 , we let @xmath1162 be the probability that a random lattice of covolume @xmath1049 has empty intersection with both @xmath1170 and @xmath1171 , i.e. @xmath1172 in the special case @xmath1173 we also write , for short : @xmath1174 we will in fact only consider @xmath1162 for @xmath1175 ; actually we will even have @xmath1176 throughout the paper except in proposition [ limfdt2prop ] below .
the function @xmath1162 satisfies an invariance relation under the simultaneous transformation of the couple @xmath1177 by any affine linear map preserving @xmath1166 .
let us write @xmath1178 for the group of non - singular affine linear transformations of @xmath161 .
we represent the elements of @xmath1178 by pairs @xmath1179 with @xmath1180 and @xmath1181 , where the action of @xmath1179 on @xmath161 is given by @xmath1182 . now note that for any @xmath1183 which maps @xmath1166 onto itself we have , directly from , @xmath1184 ( note here that @xmath1185 . indeed ,
for all sufficiently large @xmath40 we have @xmath1186 and thus @xmath1187 .
this implies @xmath1188 for some @xmath1189 , and hence @xmath1190 . ) from and the fact that @xmath256 is preserved under @xmath1191 , we get @xmath1192 as a special case of this relation we have @xmath1193 next note that for any @xmath1194 , @xmath1195 , the following affine linear map preserves @xmath1166 : @xmath1196 the set of these maps @xmath1197 forms a group , with multiplication laws @xmath1198 applying repeatedly with @xmath1199 as in and @xmath1094 as in we see that @xmath1200 may always be transformed to a pair of vectors with @xmath1201 and @xmath1202 .
it is now natural to define , for @xmath1203 , @xmath1204 , @xmath1205 we note that this function satisfies the symmetry relation @xmath1206 indeed , applying with @xmath1207 we get @xmath1208 and using now with @xmath1209 $ ] together with the fact that @xmath1162 only depends on the direction of @xmath1210 and not its length , we obtain .
note that we are now using the same sign @xmath1211 for four different but related functions : @xmath1212 , @xmath1162 , @xmath1213 and @xmath1214 ( cf . , , and ) .
there should be no risk of confusing these , since the number or types ( vector / scalar ) of the arguments are different in the four cases .
the relation between @xmath1212 and the other functions is of course : @xmath1215 the following lemma tells how to bring two points @xmath1216 to normal position : [ talfbetgennormalizelem ] assume that both @xmath1217 and @xmath1218 lie in @xmath1166 ( viz . ,
@xmath1219 and @xmath1220 ) .
set @xmath1221 then @xmath1222 and @xmath1223 .
this is verified by a direct computation . in this section
we prove the properties of the function @xmath212 stated in the paragraph below theorem [ phixiwasymptthm ] , and we also derive theorem [ phixilargethm ] as a consequence of theorem [ phixiwasymptthm ] and .
recall that we have defined , in , @xmath1224 we introduce the function ( for @xmath1203 , @xmath1204 ) @xmath1225 then in particular we have ( cf . ) @xmath1226 it follows easily from and ( * ? ? ?
* lemma 2.3 ) that @xmath1212 is a continuous function of @xmath1080 and @xmath1049 .
hence we conclude that , for @xmath40 , @xmath1227 ( cf .
below for the last relation ) , just as stated in the introduction .
next , it is easy to see that @xmath212 stays bounded from below as @xmath215 .
indeed , note that @xmath1228 ; hence @xmath1229 holds whenever @xmath1230 is smaller than some positive constant which only depends on @xmath172 ( cf .
* lemma 2.2 ) ) . in particular
if @xmath8 is sufficiently small then the integrand in is @xmath1231 for all @xmath1232 , @xmath1233 , and thus @xmath1234 .
however we even have that @xmath1235 exists and is positive : [ limfdt2prop ] for every @xmath197 we have @xmath1236 using with @xmath1094 , @xmath1237 and @xmath1238 we may rewrite as follows : @xmath1239 where @xmath1240 .
we now have , uniformly over all @xmath1241 $ ] and @xmath1047 @xmath1242 ( cf .
* lemma 2.3 ) ) , and also @xmath1243 ( cf .
* cor . 1.4 ) ) .
let us keep @xmath1244 for the rest of the proof . using the above two bounds we see that the error caused by replacing @xmath1245 by @xmath1246 in is : @xmath1247 to compute this we note that @xmath1248 furthermore note that for @xmath1249 we have @xmath1250 and @xmath1251 , while for @xmath1252 we have @xmath1253 , and @xmath1254 holds if and only if @xmath1255 , where @xmath1256 ( note @xmath1257 ) .
writing also @xmath1258 ( note @xmath1259 ) , we now obtain that is @xmath1260 hence we have proved @xmath1261 substituting @xmath1262 the double integral becomes : @xmath1263 here the inner integral equals @xmath1264 uniformly over @xmath1265 .
using this and the bound we obtain ( and we furthermore see that the rate of convergence is @xmath1266 ) .
we next turn to the proof of theorem [ phixilargethm ] using theorem [ phixiwasymptthm ] and .
the key step in this derivation is the following formula .
[ phixiwaveragepf3prop ] @xmath1267 we may evaluate the integral @xmath1268 by introducing polar coordinates @xmath1269 ( @xmath1270 , @xmath1271 ) and using theorem [ phixiwasymptthm ] together with the fact that @xmath899 for all @xmath1272 and @xmath896 unless @xmath1273 .
this gives , for large @xmath82 , @xmath1274 ( here for @xmath121 we used the fact that , for every fixed @xmath179 , @xmath1275 . ) replacing the factor @xmath1276 by @xmath120 in the integral in the right hand side of causes an error @xmath1277 , since @xmath212 is uniformly bounded and of compact support .
hence we get , via proposition [ phixiwaveragepf3prop ] , @xmath1278 this concludes the proof , in view of and @xmath1279 .
it now remains to prove proposition [ phixiwaveragepf3prop ] .
we need the following lemma .
[ phixiwasymptthmpflem ] @xmath1280 let @xmath1281 be the space of affine lattices ( i.e. translates of lattices ) of covolume one in @xmath161 , endowed with its invariant probability measure @xmath1282 and standard projection @xmath1283 .
let @xmath1284 be the set of @xmath1285 for which there is a unique point @xmath1286 which has minimal @xmath312-coordinate among the points in @xmath1287 ( note that @xmath1288 for every @xmath1285 ) .
then @xmath1289 is open and of full measure in @xmath1290 , and @xmath1291 is a smooth injective map from @xmath1289 onto a certain subset @xmath1292 of @xmath1293 , so that @xmath1294 now for given @xmath1295 and @xmath1296 we have @xmath1297 if and only if @xmath1298 and @xmath1299 .
the last condition holds for @xmath256-almost all @xmath1300 .
hence the inner integral in equals @xmath1301 , and this proves the lemma .
using with @xmath1302 , where @xmath1303 we get , for any @xmath1304 , @xmath1305 , @xmath1306 where ( for any fixed @xmath40 ) @xmath1307 one checks that the map @xmath1308 is a diffeomorphism from @xmath1309 onto @xmath1310 , with the inverse map @xmath1311 and the jacobian is @xmath1312 hence may be rewritten as follows : @xmath1313 substituting this formula in @xmath1314 and then changing the order of integration and using @xmath1315 we get @xmath1316 now use the fact that @xmath1317 for any @xmath1318 and integrate over this sphere ; this gives @xmath1319 and now is a consequence of lemma [ phixiwasymptthmpflem ] .
[ xidm1trivboundcor ] for all @xmath1203 , @xmath1204 , @xmath1320 and @xmath1047 we have @xmath1321 where @xmath1322 .
it is clear from the definitions and that @xmath1323 ; hence by lemma [ xicvboundlem1 ] we have @xmath1324 the lemma follows from this bound by also using .
[ paraboloidincllem ] for any @xmath1325 and @xmath1326 we have @xmath1327 the point @xmath1328 belongs to the set in the left hand side of if and only if @xmath1329 , and it belongs to the set in the right hand side of if and only if @xmath1330 hence our task is to prove @xmath1331 for all @xmath1332 .
this simplifies to @xmath1333 which is clear when @xmath1326 .
[ xidm1incrlem ] for any fixed @xmath1334 , @xmath1214 is an increasing function of @xmath1335 .
it suffices that @xmath1162 is an increasing function of @xmath1049 for fixed @xmath1336 .
this follows directly from the definition , since @xmath1337 whenever @xmath1338 .
[ xidm1contlem ] @xmath1214 is a continuous function on @xmath1339 .
it suffices to prove that @xmath1162 is continuous .
fix @xmath1340 and let @xmath1341 run through a sequence of tuples in @xmath1342 tending to @xmath1343
. then @xmath1344 and thus ( ( * ? ? ?
1.2.2 ) ) @xmath1345 hence by ( * ? ? ?
* lemma 2.3 ) , @xmath1346 [ xidm1contlem3 ] there is a constant @xmath179 which only depends on @xmath172 such that , for all @xmath1347 , @xmath1204 , @xmath1047 , @xmath1348 and @xmath1349 $ ] , we have @xmath1350 set @xmath1351 , @xmath1352 and @xmath1353 .
it suffices to prove the inequality for @xmath1354 since the case @xmath1355 then follows by continuity ( lemma [ xidm1contlem ] ) . by with @xmath1356
we have , for any @xmath1357 , @xmath1358 where @xmath1359 , @xmath1360 , @xmath1361 and @xmath1362 .
similarly , using the fact that @xmath1363 , @xmath1364 where @xmath1365 , @xmath1366 and @xmath1367 . now take @xmath1368
; then by lemma [ paraboloidincllem ] we have @xmath1369 and thus @xmath1370 holds so long as @xmath1371 , viz.@xmath1372 , which is certainly true for @xmath1373 with @xmath1374 sufficiently large .
just as in the proof of theorem [ phixiwasymptthm ] a crucial step was to approximate the probability of a random lattice avoiding a given cut ball by the corresponding probability for a cut paraboloid ( cf .
[ paraboloidapprsec ] ) , so it will be important in the proof of theorem [ phi0xilargethm ] to carry out the corresponding approximation for a union of _ two _ given cut balls . given @xmath1375 and @xmath102 we define @xmath1376 ( this agrees with if @xmath1377 ) and set @xmath1378 finally let @xmath1379 be the probability that a random lattice of covolume @xmath1049 has empty intersection with @xmath1380 , viz.@xmath1381 our goal in the present section is to approximate the function @xmath1379 from above and below with the function @xmath1214 .
when proving theorem [ phi0xilargethm ] we may without loss of generality assume @xmath1382 , since @xmath92 and also the right hand side of is symmetric in @xmath150 ( as will be seen from the definition , using ) . for given @xmath1383 and @xmath1384 , we now make an explicit choice of corresponding vectors @xmath1385 , @xmath1386 .
@xmath1387\label{case12formula } & \ : \\\notag & \text{if } \ : z\leq w\cos\varphi : & & \begin{cases}\vecz = z\vece_1 \\ \vecw = w(\cos\varphi)\vece_1-w(\sin\varphi)\vece_2 . \end{cases}\end{aligned}\ ] ] these @xmath1388 are easily verified to satisfy @xmath1389 , @xmath1390 and @xmath1391 .
note that @xmath1392 holds if and only if the triangle @xmath1393 is acute ; the point of our choice of @xmath1388 in this case is to make @xmath1394 hold . for @xmath1388 given as above
we now wish to approximate @xmath1380 by two cut paraboloids .
in fact we will use two translates of the same paraboloid @xmath1021 ( cf .
[ paraboloidapprsec ] ) , with @xmath1395 and appropriate @xmath233 .
this means that near the origin , @xmath1396 looks very much like @xmath1397 ( cf .
lemma [ puvdeflem ] and note @xmath1398 ) .
also @xmath1399 is in many cases a good approximation of @xmath1400 near the origin ; however if @xmath1401 is near @xmath1402 then @xmath1399 may even fail to contain the origin , and a much better approximation of @xmath1400 is given by @xmath1403 , where @xmath1404 denotes reflection in the hyperplane @xmath1405 ( viz .
@xmath1406 ) . with @xmath1407 , @xmath1408 as in section [ paraboloidapprsec ]
we have @xmath1409 in fact we will use @xmath1403 in place of @xmath1399 if and only if @xmath1410 . in view of the above relation , it is convenient to introduce the point @xmath1411 note that @xmath1410 can only happen in the first of the two cases in , since @xmath1412 implies @xmath1413 . note also that if @xmath1414 then @xmath1415 and @xmath1416 , and by lemma [ paraboloidballapprlem](i ) we have @xmath1417 the following lemma gives conditions for the opposite inclusion to hold : [ cutcontainedlem2 ] there is an absolute constant @xmath1418 such that whenever @xmath1419 , @xmath1420 , @xmath1320 , @xmath1421 , and @xmath1422 then we have , for @xmath87 as in and @xmath1423 as in , @xmath1424 recalling the definition of @xmath1380 we see that it suffices to prove @xmath1425 in view of and lemma [ paraboloidballapprlem](ii ) , in order to prove the first inclusion in it suffices to prove @xmath1426 without loss of generality we may rescale @xmath1210 so that @xmath1427 .
but we have by , @xmath1428 also the image of the halfspace @xmath1429 under the translation @xmath1430 is computed to equal @xmath1431 , where @xmath1432 hence is equivalent with @xmath1433 note that @xmath1434 ( cf . and recall @xmath1421 ) . using the we compute that the supremum of @xmath1435 taken over all points @xmath1436 lying in the set in the left hand side of is @xmath1437 ( this presupposes @xmath1438 ; in the opposite case the set in the left hand side of is empty , so that the desired inclusion holds trivially . )
now note that if @xmath1392 then @xmath1439 , and hence @xmath1440 the same computation also shows @xmath1441 using , and we obtain @xmath1442 this bound is obviously also true when @xmath1412 , cf . , i.e. it is true in general .
we also get from ( using @xmath1443 ) , @xmath1444 also note that @xmath1445 , @xmath1446 , @xmath1447 and @xmath1448 ( cf . and recall @xmath1421 ) , and @xmath1449 .
it follows from the above observations that is @xmath1450 this has been proved under the assumptions @xmath1419 , @xmath1420 , @xmath1033 , @xmath1427 and @xmath1421 .
now if we _ also _ assume @xmath1451 where @xmath1452 is a sufficiently large constant ( this in particular forces @xmath1453 to be small ; hence @xmath1454 ) , then it follows that the big-@xmath1084-term in is @xmath1455 , and since @xmath1456 this implies that holds .
this completes the proof of the first inclusion in ; the second inclusion is proved by a completely similar argument .
next , in order to relate and to the function @xmath1214 , we need to transform @xmath1457 by a linear map into a union of the form @xmath1458 , cf . and .
the following lemma gives a detailed description of the parameters occurring in this transformation .
[ normalizepuvlem ] there is an absolute constant @xmath1459 $ ] such that for any fixed @xmath185 with @xmath1460 , @xmath1461 , there exist a function @xmath1462\to\gl(d-1,{\mathbb{r}})$ ] and @xmath45 functions @xmath1463\to{\mathbb{r}}_{>0}$ ] , @xmath1464\to{\mathbb{r}}_{\geq0}$ ] , @xmath1465\to{\mathbb{r}}$ ] , such that for @xmath1388 as in , @xmath1423 as in , @xmath1395 and arbitrary @xmath1466 $ ] and @xmath1033 , we have , writing @xmath1467 , @xmath1468 , @xmath1469 , @xmath1470 , @xmath1471 : @xmath1472 where @xmath1473 and furthermore , with @xmath1474 and with absolute implied constants : @xmath1475 as an auxiliary lemma , let us first note the following regarding the functions @xmath1407 , @xmath1408 , @xmath1476 , @xmath1477 introduced in section [ paraboloidapprsec ] .
[ abcederlem ] the following bounds holds uniformly over all @xmath1010 with @xmath1478 : @xmath1479 the bound @xmath1480 is proved by a direct computation and the other three bounds follow trivially from this .
let @xmath185 ( and thus @xmath1481 ) be given as in the formulation of the lemma , and set @xmath1395 . also set @xmath1482 if @xmath1410 and otherwise @xmath1483 , so that @xmath1484 .
now for any @xmath1485 $ ] we let @xmath1407 , @xmath1408 , @xmath1476 , @xmath1477 be as in section [ paraboloidapprsec ] , define @xmath1486 by @xmath1487 and set @xmath1488 finally define @xmath1470 , @xmath1471 , @xmath1468 , @xmath1469 as in lemma [ talfbetgennormalizelem ] , applied with @xmath1489 as in ( we will see that @xmath1489 lie in @xmath1166 provided that @xmath1490 is sufficiently small ; cf .
and below ) , and set @xmath1491 and @xmath1492m_{\alpha,\beta}^{-1}.\end{aligned}\ ] ] then by construction we have @xmath1493 for some @xmath1181 ; @xmath1494 ; @xmath1495 , and @xmath1496 . also by a quick computation one
checks that @xmath1497 for all @xmath1498 , with @xmath1499 as in . hence holds .
it remains to verify that the functions @xmath1500 have the properties stated in .
writing @xmath1217 and @xmath1218 , we compute @xmath1501 note that @xmath1502 always holds , and recall in the proof of lemma [ cutcontainedlem2 ] .
it follows that @xmath1503 hence assuming that @xmath1490 has been taken sufficiently small we have @xmath1504 , and since also @xmath1505 , implies @xmath1506 we also get ( cf . )
@xmath1507 now if @xmath1392 then we obtain from : @xmath1508 and here if @xmath1509 then ( since @xmath1510 ) @xmath1511 while if @xmath1512 ( thus @xmath1513 ) then we still have @xmath1514 . hence always when @xmath1392 we get , using also and , @xmath1515 on the other hand if @xmath1412 then we have @xmath1516 as noted below , and also @xmath1517 and @xmath1518 hence using , and we see that again holds ; thus is true _ in general_. by a similar discussion , also using lemma [ abcederlem ] , we find that , both when @xmath1392 and when @xmath1412 : @xmath1519 by similar computations ( using @xmath1520 ) we also have @xmath1521 we next study the difference @xmath1522 . by , @xmath1523
if @xmath1392 , then assuming @xmath1490 sufficiently small and using and we have @xmath1524 note also that if @xmath1509 then @xmath1525 and thus by , and , @xmath1526 .
hence always when @xmath1392 we have @xmath1527 on the other hand if @xmath1412 then @xmath1512 and @xmath1528 ; hence again holds , i.e. is true in general . by similar considerations , also using lemma [ abcederlem ] , we find that , both when @xmath1392 and when @xmath1412 , @xmath1529 note that by assuming @xmath1530 to be sufficiently small we can force the big-@xmath1084 terms in and to be less than @xmath1531 in absolute value .
hence the first three lines of now follow from lemma [ talfbetgennormalizelem ] combined with , , , , .
the fourth line of follows from @xmath1532 , @xmath1533 , @xmath1534 , @xmath1535 , and lemma [ abcederlem ] .
finally the last line of follows by also using @xmath1536 , cf . and lemma [ talfbetgennormalizelem ] .
in the next two propositions we give the desired approximations of @xmath1379 in terms of the @xmath1211-function .
we start with the approximation from above , which is in some respects more complicated than the one from below .
[ upsilonmainupperboundprop ] there exist constants @xmath1537 $ ] and @xmath1538 which only depend on @xmath172 such that for any fixed @xmath185 with @xmath1539 , @xmath1540 , there exist @xmath45 functions @xmath1541\to{\mathbb{r}}_{>0}$ ] and @xmath1542\to{\mathbb{r}}$ ] which satisfy the bounds @xmath1543 for all @xmath1544 $ ] , and which have the property that , for @xmath1388 as in , and for all @xmath1047 and all @xmath1033 with @xmath1545 , @xmath1546 we will use the constants @xmath1490 from lemma [ normalizepuvlem ] and @xmath1547 from lemma [ cutcontainedlem2 ] . after possibly enlargening @xmath1452
we may assume that @xmath1548 whenever @xmath1460 , @xmath1461 ( cf . ) .
take @xmath1549 $ ] so small that @xmath1550 .
now let @xmath185 ( and thus @xmath87 ) be given subject to @xmath1539 and @xmath1540 .
let us write @xmath1551\to{\mathbb{r}}_{>0}$ ] , @xmath1464\to{\mathbb{r}}_{\geq0}$ ] , @xmath1552\to{\mathbb{r}}$ ] for the functions provided by lemma [ normalizepuvlem ] .
now for @xmath1544 $ ] we set @xmath1553 ( note that @xmath1554 for all @xmath1544 $ ] because of our choice of @xmath1555 ; in particular @xmath1556 , @xmath1557 are well - defined for all @xmath1544 $ ] . )
we set @xmath1395 as in lemma [ normalizepuvlem ] ; note that @xmath1558 for all @xmath1544 $ ] , because of our choice of @xmath1452 .
now let @xmath1033 be given , subject to @xmath1545 .
set @xmath1559 , @xmath1560 and @xmath1561 , where @xmath1462\to\gl(d-1,{\mathbb{r}})$ ] is as in lemma [ normalizepuvlem ] ; then @xmath1562 with @xmath1468 , @xmath1469 and @xmath1563 also by lemma [ cutcontainedlem2 ] we have @xmath1564 hence , recalling , , and the fact that @xmath256 is @xmath258-invariant , @xmath1565 using now also lemma [ xidm1contlem3 ] it follows , assuming that @xmath1566 and @xmath1567 ( if @xmath157 : @xmath1568 ) both lie in @xmath1569 $ ] , that @xmath1570 where @xmath1571 .
but from lemma [ normalizepuvlem ] we know that @xmath1572 ; hence @xmath1573 , and thus after possibly shrinking @xmath1555 we can ensure that @xmath1574 always hold for our @xmath1575 . hence we see , using also the fact that @xmath1576 ( cf .
lemma [ normalizepuvlem ] ) , that holds , if we define @xmath1577.\end{aligned}\ ] ] ( if @xmath157 then @xmath1578 , and we set @xmath1579 . ) finally the properties in follow directly from our definitions and lemma [ normalizepuvlem ] ( taking @xmath1555 sufficiently small ) .
[ upsilonmainlowboundprop ] there exist constants @xmath1580 $ ] and @xmath1581 which only depend on @xmath172 such that for any fixed @xmath185 with @xmath1582 , @xmath1583 , there exist real numbers @xmath1584 which satisfy the bounds @xmath1585 and which have the property that , for @xmath1388 as in , and for all @xmath1047 and all @xmath1033 , @xmath1586 where @xmath1587 .
this is very similar to the proof of proposition [ upsilonmainupperboundprop ] , except that we apply lemma [ normalizepuvlem ] with @xmath1588 ( independent of @xmath1210 ) ; in place of lemma [ cutcontainedlem2 ] we simply use to obtain @xmath1589 ; and finally we apply lemma [ xidm1contlem3 ] with @xmath1590 and @xmath1591 , and @xmath1592 , @xmath1593 in place of @xmath1594 .
we now start with the proof of theorem [ phi0xilargethm ] .
the proof involves approximating @xmath75 with an integral involving the @xmath1595-function ( cf .
below , as well as ) , which is then estimated from above and below in terms of the @xmath1211-function , using propositions [ upsilonmainupperboundprop ] and [ upsilonmainlowboundprop ] .
the resulting integral is then made cleaner in a series of steps , eventually resulting in the function @xmath160 which we define in below . note that if @xmath0 then theorem [ phi0xilargethm ] ( with @xmath1596 as in ) follows directly from the explicit formula in @xcite
hence we will from now on assume @xmath197 .
as pointed out in section [ apprparaboloidsec ] we may assume @xmath1382 without loss of generality .
let us fix the constant @xmath898 so that @xmath1597 and @xmath1598 whenever @xmath1599 ( cf .
proposition [ phi0supportthm ] ) ; these conditions are equivalent to the conditions imposed on @xmath898 at the start of section [ simplintsec ]
. it will be clear from the definition of @xmath160 in that @xmath1600 holds whenever @xmath1601 ( for recall , , and @xmath1602 ) .
thus is automatic when @xmath1599 .
hence from now on we will assume @xmath1603 .
let @xmath1604 be a positive constant which is smaller than both [ clowprequirements ] @xmath1555 from proposition [ upsilonmainupperboundprop ] and @xmath1605 from proposition [ upsilonmainlowboundprop ] .
( we will later impose some further conditions on @xmath1604 being sufficiently small , but we will see that it can be fixed in a way which only depends on @xmath172 . ) if @xmath1606 then by theorem [ cylinder2ptsmainthm ] and coupled with lemma [ sabvboundlemcor ] below , both @xmath146 and the main term in the right hand side of are @xmath1607 , and thus is automatically true .
hence from now on we will assume @xmath1608 . by lemma [ a1largelem ]
there is a constant @xmath1609 [ finalfixingofclowe ] which only depends on @xmath172 such that for any @xmath101 and any translate @xmath278 of the cylinder @xmath1610 , we have that @xmath903 holds for all @xmath639 with @xmath904 .
we will assume from start that @xmath1611 ; in particular we have @xmath1612 and @xmath687 . for later reference we recapitulate our main assumptions on @xmath185 : @xmath1613 we fix @xmath1388 as in , for our given @xmath185 .
let @xmath775 be a fundamental region for @xmath771 as in lemma [ fdcontainmentslem ] ( applied with our @xmath1614 ) . by ( * ?
* ( 7.32 ) ) we have @xmath1615 where @xmath1616 , @xmath1617 , @xmath1618 is the set of primitive vectors in @xmath1619 , and @xmath1620 . using the bound ( * ? ? ?
7.3 ) on the contribution from all @xmath1499 with @xmath1621 in , we get @xmath1622 where we write @xmath1623 , and where @xmath1624 using now lemma [ fdcontainmentslem ] and the fact that @xmath738 holds for all @xmath1625 with @xmath904 , we get : @xmath1626 following @xcite , we parametrize @xmath1627 ( for @xmath1623 ) as follows . for any @xmath1628 and @xmath1629 with @xmath1630
there is a unique choice of @xmath664 , @xmath662 such that @xmath1631\in g_{\veck,\vecy}$ ] , namely : @xmath1632 we write @xmath1633_{\veck,\vecy}$ ] for this element @xmath1631\in g_{\veck,\vecy}$ ] .
this gives a bijective map @xmath1634_{\veck,\vecy}\in g_{\veck,\vecy}.\end{aligned}\ ] ] let @xmath1635 be the lattice @xmath1636_{\veck,\vecy}. \ ] ] for any given @xmath1637 as above , @xmath1635 is in fact independent of @xmath1638 ( for note that @xmath1639 are independent of @xmath1640 , and so is the congruence class of @xmath1641 ; hence the claim follows from @xmath693 and ) . note also that for any @xmath1637 as above there is a unique choice of @xmath1638 which yields @xmath1642^{d-1}$ ] .
using now the definition of @xmath1643 ( cf . ) , @xmath804 , and the expression for the measure @xmath271 in the parameters @xmath1637 ( ( * ? ? ?
* lemma 5.2 ) ) , we conclude @xmath1644 where @xmath1645 here the first error term is @xmath1646 where @xmath1647 .
( for note that @xmath1648 implies @xmath1649 , i.e. @xmath969 ; cf . .
) but for each @xmath1650 , the set @xmath1651 contains an open @xmath675-dimensional right cone with @xmath509 in its base , of radius @xmath1652 and height @xmath1653 , where @xmath1654 is the angle between @xmath1655 and @xmath106 in @xmath161 , by ( * ? ? ?
* lemma 7.1 ) . hence using ( * ? ? ?
* lemma 7.4 ) and a parametrization similar to but rotated to have @xmath1654 in place of @xmath912 , we see that is @xmath1656 hence , since @xmath1657 , we conclude @xmath1658 next , by ( * ? ? ?
7.5 ) , at the cost of an error which is @xmath1659 we may restrict the range of integration in to the set @xmath1660 where @xmath1661 as in .
recall that @xmath1662 ( cf .
as before we write @xmath1663 and @xmath1664 ; then @xmath937 and @xmath1665 ( cf . ) .
note that @xmath972 forces @xmath1666 and thus @xmath1667 .
we also have @xmath1668 and thus also @xmath1669 we now impose the condition that @xmath1604 should be so small that forces @xmath1670 ( this is clearly possible , since both @xmath1671 and @xmath1672 are always @xmath1673 in ) . then also @xmath1674 , and since @xmath1617 it follows that , for all @xmath972 , @xmath1675 note that this was derived without ever using the condition @xmath1676 in ; hence that condition is in fact redundant ( since @xmath1677 ) , i.e. we have @xmath1678 recall that we always have @xmath1679 , cf . .
the following lemma implies that @xmath934 holds for all @xmath972 and all @xmath1680 .
[ goodintlem2 ] for any @xmath1629 with @xmath1681 and @xmath1682 , and for any @xmath1683 with @xmath1684 , @xmath1685 , @xmath1686 since @xmath1687 has nonempty intersection with @xmath1688 for @xmath1689 , and @xmath1688 is convex , it suffices to prove that holds for @xmath1690 and for @xmath1691 .
we have @xmath1692 where @xmath1693 ; hence @xmath1688 has nonempty intersection with @xmath1687 if and only if @xmath1694 has nonempty intersection with @xmath1695 hence by lemma [ cylplanelemma ] our task is to prove that @xmath1696 for @xmath1697 , i.e. it suffices to prove that @xmath1698 however this is clear since the assumptions of the lemma imply @xmath1699 and thus @xmath1700 , and similarly @xmath1701 .
let us also note that @xmath1702 indeed these relations follow from with @xmath1689 , if we also note that for any @xmath1623 ( @xmath1638 ) , if @xmath1703_{\veck,\vecy}$ ] then @xmath1704 , since @xmath1705 .
taking also lemma [ goodintlem2 ] into account we conclude that for any @xmath972 we have the equivalence : @xmath1706 note here that @xmath1707 hence by the same argument as we used to get , we have for every @xmath972 : @xmath1708 cf . and .
furthermore we have _ equality _ in whenever both @xmath991 and @xmath1709 hold .
now by lemma [ goodintlem ] ( applied after appropriate rotations ) , if @xmath972 does _ not _ satisfy both @xmath991 and @xmath1709 , then @xmath1710 where from now on we write @xmath1711 for the angle between @xmath1712 and @xmath105 , and ( as before ) @xmath1654 for the angle between @xmath1712 and @xmath106 . since @xmath1382
, we see that forces @xmath1713 to hold ( the implied constant depends only on @xmath172 ) . but recall the definition of @xmath1714 , , and note that @xmath1715 and @xmath1716 , by the triangle inequality in @xmath1717 .
hence by choosing @xmath1604 sufficiently small we can ensure that every @xmath972 which does not satisfy both @xmath991 and @xmath1709 must in fact satisfy @xmath1718 .
now the total contribution from these @xmath972 to the integral in can be bounded by following the proof of in ( * ? ? ?
7.2 ) ( the `` @xmath1719-part '' , for @xmath204 ) , but restricting the integration by @xmath1718 ; cf .
especially ( * ? ? ?
* ( 7.22)-(7.23 ) ) ; it follows that this contribution is : @xmath1720 by proposition [ phi0supportthm ] we know that @xmath191 implies @xmath1721 hence if @xmath1722 then is @xmath1723 , where @xmath1724 and it follows that : @xmath1725 this is in fact true _ in general _
, for in the remaining case when @xmath1726 , is an obvious consequence of and the inequality .
( in we may note that @xmath1727 if @xmath121 , but if @xmath181 then @xmath1727 holds if and only if @xmath1728 or @xmath1729 .
however in the end we will just use @xmath1730 , cf . . )
next , using , and their analogues for @xmath105 , we have for every @xmath972 : @xmath1731 since @xmath1732 and @xmath1733 for all @xmath972 ( cf . and ) , implies @xmath1734 now because of , and @xmath1735 , proposition [ upsilonmainupperboundprop ] can be applied to bound the integrand in from above , throughout the range of integration .
it follows ( using also and the monotonicity property of @xmath1211 in its last argument , lemma [ xidm1incrlem ] ) that there is a constant @xmath1736 which only depends on @xmath172 such that @xmath1737 where @xmath1738 and where @xmath1584 are functions in @xmath1739)$ ] satisfying .
similarly , by proposition [ upsilonmainlowboundprop ] , after possibly increasing @xmath1736 we _ also _ have @xmath1740 where @xmath1741 where , this time , @xmath1071 and @xmath1742 are real _ numbers _ satisfying .
[ lowerboundremark1 ] in fact the error term `` @xmath1743 '' in may be improved to `` @xmath1744 '' .
indeed , note that the error term in is non - negative ; hence by going through the proof of but only aiming for a lower bound we obtain @xmath1745 hence , using we get @xmath1746 and now our claim follows by applying proposition [ upsilonmainlowboundprop ] .
we wish to simplify the integrals in and .
we will first discuss ; the treatment of is similar but easier , as we explain at the end of this section . to start with ,
we introduce new variables of integration @xmath1747 via @xmath1748 a quick computation shows that this formula defines a diffeomorphism @xmath1749 with inverse given by @xmath1750 now @xmath1751 for all @xmath1752 , where we write @xmath1753 . also for all @xmath972 we have ( cf
. , and recall @xmath1754 ) : @xmath1755 and this forces @xmath1756 hence @xmath1757 and @xmath1758 and @xmath1759 for all @xmath972 . also note that for @xmath1760 related by the above diffeomorphism , we have for any @xmath711 , @xmath1761 : @xmath1762 hence for @xmath972 we have , using also and @xmath1763 : @xmath1764\text { or } [ j>2,\ell=1 ] \\ \frac12(v_1 + 2v_2)(\delta_{\ell , j-1}+o(\|\vecv''\|^2))&\text{if } \ : j>2,\ell>1 . \end{cases}\end{aligned}\ ] ] hence the jacobian is : @xmath1765 finally recall that if we parametrize @xmath908 with @xmath1766 then @xmath1767 . in view of the above observations ,
now becomes @xmath1768 where from now on we write @xmath1769 and where @xmath1770 , and @xmath1771 is the set of all [ sppdef ] @xmath1772 which correspond to @xmath972 under our diffeomorphism .
let us write @xmath1773 and @xmath1774 for short , and recall . also recall that the dependence of @xmath1211 on its third argument is only up to proportionality . hence
in we may take @xmath1775 we will next carry out one more substitution , taking @xmath1776 as new variables of integration .
note that defines a @xmath45 function @xmath1777 for all @xmath1778 in the open cone @xmath1779 ( since @xmath1780 ) .
it follows from that @xmath1781 .
we have @xmath1782 hence for all @xmath1783 we have , using also and @xmath1784 : @xmath1785 using these bounds and we obtain , for all @xmath1783 : @xmath1786 it follows that if @xmath1787 is sufficiently small then our map @xmath1777 is injective on all of @xmath1292 . because of and we can ensure this by requiring the constant @xmath1604 to be sufficiently small .
in particular it now follows that the map @xmath1777 restricts to a diffeomorphism from @xmath1771 onto some open set @xmath1788 .
[ spppdef ] using together with @xmath1789 we get @xmath1790 by requiring @xmath1604 to be sufficiently small we may assume that the big-@xmath1084 term in is @xmath1791 , say .
we now obtain from : @xmath1792 the treatment of is quite similar : by exactly the same argument as that leading to we obtain , after possibly increasing @xmath1736 , @xmath1793 where @xmath1771 is the same as in , and @xmath1210 is as in , where this time @xmath1584 are real _ numbers _ satisfying . just as before we now carry out one more substitution , taking @xmath1776 as new variables of integration .
since @xmath1584 are now constants independent of @xmath1778 , this transformation is much less complicated than before ; in fact it is just a ( non - singular ) _ linear _ map @xmath1794 , with jacobian @xmath1795 ( cf . ) .
hence we get @xmath1796 where ( again ) @xmath1788 is the image of @xmath1771 under our ( this time linear ) map @xmath1777 . in order to bound the contribution from the various error terms in
we first prove some auxiliary bounds on integrals involving the @xmath1211-function .
[ sabvboundlem ] we have , uniformly over all @xmath1203 , @xmath1797 , @xmath1047 , @xmath1798 ( when @xmath1355 the right hand side should be interpreted as `` @xmath1799 '' . )
note that @xmath1800 if and only if @xmath1801 .
by lemma [ xidm1trivboundcor ] we have @xmath1802 let us first assume @xmath181 . by noting the symmetry @xmath1803 , and using polar coordinates for the remaining variables @xmath1804 ( @xmath1805 ) , we get @xmath1806
this is always @xmath1807 where we used @xmath1808 in the last step . on the other hand
if @xmath1809 then we can do better as follows : @xmath1810 in the remaining case @xmath121 we get instead @xmath1811 which again agrees with .
[ sabvboundlemcor ] we have , uniformly over all @xmath1203 , @xmath1797 , @xmath1047 , @xmath1812 writing @xmath1813 we get @xmath1814 hence the lemma follows from lemma [ sabvboundlem ] . [ toughxiintlem ]
we have , uniformly over all @xmath1203 , @xmath1797 , @xmath1815 , @xmath1816 , @xmath1817 we first assume @xmath181 . substituting @xmath1818 with @xmath1805 , and using the bound on @xmath1211 from lemma [ xidm1trivboundcor ] , similarly as in the proof of lemma [ sabvboundlem ] , we see that the left hand side of is @xmath1819 we now prove an auxiliary result : [ technicallem ] for any fixed @xmath1820 we have , uniformly over all @xmath1821 , @xmath1822 : @xmath1823 first assume @xmath1824 . then the left hand side of is @xmath1825 i.e. holds .
next assume @xmath1826 .
then the left hand side of is @xmath1827 if @xmath1828 then this is @xmath1829 ; if @xmath1830 then it is @xmath1831 , and finally if @xmath1832 then it is @xmath1833 .
hence holds in all cases .
we now continue onwards with the proof of lemma [ toughxiintlem ] . by lemma [ technicallem ] (
used together with @xmath1834 ) , is @xmath1835 let us first assume @xmath1836 .
then @xmath1837 for all @xmath1838 , and we thus get , upon setting @xmath1839 , @xmath1840 by our assumptions we have @xmath1841 ; hence the above is @xmath1842 also by our assumptions we have @xmath1843 . using this and the fact that @xmath1844 for @xmath181 , we find by a quick computation that the above is @xmath1845 ( the last step follows since @xmath1846 and @xmath1847 , both of which are true by assumption . ) it now remains to treat the case @xmath1848 . in this case
@xmath1837 still holds whenever @xmath1849 or @xmath1850 , and hence the contribution from all such @xmath1851 in is still bounded by , and hence also bounded by .
hence it only remains to treat the contribution from @xmath1851 with @xmath1852 and @xmath1853 .
but for these @xmath1851 we have @xmath1854 and hence the contribution from these @xmath1851 in is @xmath1855 hence holds also when @xmath1848 .
finally we treat the case @xmath121 . in this case
the left hand side of is @xmath1856 if @xmath1836 then arguing as before we get @xmath1857 this is the same as `` with @xmath121 '' , and the analysis goes through as before except that there is an extra logarithm factor ( since @xmath1858 for @xmath121 ) , and we obtain the bound in .
the extension to the case @xmath1848 works as before .
we now bound the contribution from the error term in .
first , it follows from lemma [ sabvboundlemcor ] that the contribution from `` @xmath1859 '' in is @xmath1860 we next consider the contribution from @xmath1861 .
note that @xmath1862 for all @xmath1863 .
furthermore @xmath1864 where we recall that @xmath1773 , @xmath1774 .
hence @xmath1865 here @xmath1866 and @xmath1867 , by ( also recall ) .
hence if apply lemma [ toughxiintlem ] with @xmath1868 and an appropriate choice of @xmath1869 , using lemma [ xidm1incrlem ] , then we conclude ( since @xmath1870 ) that the contribution from @xmath1861 in is @xmath1871 ( this is the same @xmath966 as in . ) one checks by inspection that @xmath1872 , cf . , .
furthermore , using @xmath1603 we see that ( cf . ) @xmath1873 hence we conclude @xmath1874 here note that ( cf . )
@xmath1875 also recall and @xmath1876 ( cf . ) .
hence by requiring @xmath1604 to be sufficiently small we can force @xmath1877 to hold for all @xmath1863 .
now from we see that there is a constant @xmath1878 which only depends on @xmath172 such that @xmath1879 where @xmath1880 recall that @xmath1788 ; hence the above inequality remains true if we replace the range of integration by @xmath1881 . writing @xmath1813 we thus get @xmath1882 now in the inner integral in we substitute @xmath1883 , @xmath1884 $ ] . using @xmath1885 and lemma [ xidm1incrlem ]
we see that the contribution from @xmath1886 is @xmath1887 where the last bound follows from lemma [ toughxiintlem ] and lemma [ sabvboundlemcor ] .
we also have , for all @xmath1888 $ ] : @xmath1889 hence we obtain @xmath1890 here the contribution from the @xmath1891-term is again @xmath1005 , since it is bounded above by the first line in .
hence , setting @xmath1892 , we have finally proved : @xmath1893 in a very similar way we also obtain a similar _ lower _ bound : first , in almost exactly the same way as we got from , leads to @xmath1894 we will prove that the range of integration in may be replaced by @xmath1881 , at the cost of an error which is @xmath1005 . recall that we have a bijection @xmath1895 between @xmath1896 and @xmath1752 , and by definition @xmath1897 is the set of all @xmath1896 which correspond to @xmath972 .
hence for any @xmath1898 the corresponding vector @xmath683 has @xmath1899 , cf . .
using also @xmath1900 and @xmath1901 ( cf . )
, we see that there is a constant @xmath1902 such that the error caused by replacing @xmath1897 by @xmath1881 in is @xmath1903 using @xmath1904 we get @xmath1905 where to bound the second indicator function in we used the fact that if @xmath1906 then @xmath1907 . the contribution from the `` min''-term in is seen to be @xmath1005 using lemma [ toughxiintlem ] and with our present constants @xmath1584 , and the contribution from the `` @xmath301''-term in is , using @xmath1908 and lemma [ sabvboundlem ] : @xmath1909 ( cf . and ) .
hence we have proved that the range of integration in may indeed be replaced by @xmath1881 .
now by the same argument as when going from to we obtain @xmath1910 together , and imply that the relation in theorem [ phi0xilargethm ] holds , with @xmath1911 the fact that @xmath160 is uniformly bounded follows from lemma [ sabvboundlemcor ] . furthermore for any @xmath1820 there exists a bounded set @xmath1912 such that for all @xmath1913 the support of the integrand in is contained in @xmath1914 ( this follows from lemma [ xisupboundlem ] , since @xmath1915 ) .
hence lemma [ xidm1contlem ] implies that @xmath160 is continuous .
this completes the proof of theorem [ phi0xilargethm ] .
@xmath324 @xmath324 @xmath324
we continue to keep @xmath197 . recall that we have defined , for @xmath1203 , @xmath1204 ( cf . ):
@xmath1916 note that @xmath1917 , immediately from and . also by we
have the symmetry relation @xmath1918 it follows from that @xmath1919 in order to express this relation in a slightly cleaner way we introduce the function @xmath1920 then translates into the symmetry relation @xmath1921 and translates into , i.e. @xmath1922 we remark that we will prove below in corollary [ basicrholowboundcor ] that @xmath1923 holds uniformly over @xmath1924 , @xmath1797 .
this translates into the relation @xmath1925 we next prove the existence of the continuous function @xmath190\to{\mathbb{r}}_{>0}$ ] as stated in theorem [ phi0supportasymptthm ] .
[ xi0wzphidisc ] let us fix any @xmath102 .
now the function @xmath1926 is continuous , decreasing ( cf .
* lemma 7.11 ) ) , positive for @xmath82 small ( e.g. by theorem [ phi0zerosmallthm ] ) and vanishing for all sufficiently large @xmath82 ( e.g. by ( * ? ? ?
1.9 ) ) ; hence there exists a unique number @xmath1927 such that @xmath1928 .
this proves that there exists a unique function @xmath190\to{\mathbb{r}}_{>0}$ ] such that @xmath1929 . since @xmath75 is continuous ( jointly in all three variables )
it follows that @xmath193 is lower semicontinuous in @xmath1930 $ ] .
finally the fact that @xmath193 is upper semicontinuous ( and hence continuous ) follows from the following lemma , which is a slight generalization of ( * ? ? ?
* lemma 7.11 ) : [ phi0incrgenlem ] for any @xmath1931 and @xmath1932 , @xmath1933 holds whenever @xmath1934 , @xmath1935 and @xmath1936 follow the proof of ( * ? ? ?
* lemma 7.11 ) , but replace the matrix @xmath478 therein by @xmath1937 the following proposition gives one half of theorem [ phi0supportasymptthm ] .
[ phi0supportasymptthmhalfprop ] we have @xmath1938 uniformly over all @xmath195 , @xmath196 $ ] .
the proof depends on the following lemma ( with constants @xmath1605 , @xmath1939 as in proposition [ upsilonmainlowboundprop ] ) .
[ phi0supportasymptthmhalfproplem ] let @xmath185 be given with @xmath1582 and @xmath1583 .
then there exist numbers @xmath1584 satisfying and which have the property that for any @xmath101 with @xmath1940 and for @xmath1388 as in , we have @xmath1941 for all @xmath1942 with @xmath1943 and @xmath1944 .
let @xmath1945 be given with @xmath101 , @xmath1582 and @xmath1583 , and assume @xmath1940 .
take @xmath1388 as in , and set @xmath1616 and @xmath1617 as in the previous section .
then since @xmath258 is covered by a countable number of @xmath774-translates we must have @xmath1946 for every @xmath1947 .
in particular this holds for @xmath1948 , and recalling the definition of @xmath1635 in it follows that @xmath1648 for almost all @xmath1949 .
let us write @xmath1950 for the set of all @xmath1629 with @xmath1943 and @xmath1944 .
then for every @xmath1951 we have @xmath1952 so that @xmath1953 , and thus we have @xmath1648 for almost all @xmath1954 .
we next note that for every @xmath1951 the conclusion of lemma [ goodintlem2 ] holds , viz . for all @xmath1680 we have @xmath934 , with @xmath1955 .
indeed , as in the proof of that lemma it suffices to check that holds , and this is clear since @xmath1956 for @xmath1951 .
it now follows as in section [ pfphi0xilargethmsec ] that holds for all @xmath1951 .
but _ also _ , by proposition [ upsilonmainlowboundprop ] , there exist numbers @xmath1584 which only depend on @xmath185 ( and @xmath172 ) and which satisfy , such that @xmath1957 is larger than or equal to the left hand side of for all @xmath1951 . these observations together imply that the left hand side of must vanish for almost all @xmath1951 , and thus since @xmath1211 is continuous ( lemma [ xidm1contlem ] ) , it must vanish for _ all _ @xmath1951 . because of @xmath1958 and
, we may assume @xmath1959 without loss of generality .
note that is content - free unless both @xmath149 and @xmath1960 are small ; we may thus assume @xmath1582 and @xmath1583
. given @xmath185 we let @xmath1388 be the corresponding points as in . also let @xmath1584 be the corresponding numbers as in lemma [ phi0supportasymptthmhalfproplem ] .
set @xmath1961 $ ] and @xmath1962 .
fix any number @xmath1963 . then there is some @xmath1320 , which we fix from now on , such that @xmath1964 for @xmath1965 and thus for all @xmath1966 .
now let @xmath8 be a small positive parameter which we will later take to tend to @xmath1057 , and set @xmath1967 where @xmath1968 clearly for all sufficiently small @xmath8 the vector @xmath683 is well - defined , lies in @xmath11 , and has @xmath1943 and @xmath1944 .
also note that , for all @xmath1966 , @xmath1969 hence lemma [ phi0supportasymptthmhalfproplem ] implies that @xmath1970 for all @xmath101 with @xmath1971 letting now @xmath215 , and then using the fact that @xmath1972 was arbitrary with @xmath1963 , it follows that @xmath1970 for all @xmath101 with @xmath1973 this concludes the proof .
[ basicrhoboundcor ] we have @xmath1974 , uniformly over all @xmath1924 , @xmath1204 .
we may assume @xmath1797 , since @xmath1917 .
given any @xmath1975 $ ] and @xmath1797 we may find @xmath1976 and @xmath1977 satisfying @xmath1978 , @xmath1979 , and with both @xmath1960 and @xmath149 arbitrarily small . now proposition [ phi0supportthm ] says that @xmath1980 , and the corollary follows from this combined with proposition [ phi0supportasymptthmhalfprop ] .
we will now prove that _ if @xmath1981 is sufficiently large , then the formula , with range of integration @xmath1982 in place of @xmath1714 , holds without error terms .
_ we will use this result to complete the proof of theorem [ phi0supportasymptthm ] , but it is clearly also of independent interest ; for example we expect that when @xmath121 this result could be used to find completely explicit formulas for @xmath79 in certain parameter regimes with @xmath82 large ( we stress however that we anticipate any such explicit formula to be rather complicated ) . in view of this independent interest
we allow a more general choice of @xmath87 than in when stating the result ; this does not cause any extra difficulties in the proof .
[ exactformulaprop ] there is a constant @xmath1983 which only depends on @xmath172 such that for any @xmath185 with @xmath1984 , @xmath1985 , any @xmath1986 , and any @xmath102 with @xmath1389 , @xmath1390 , @xmath1987 and @xmath1988 , @xmath1989 , we have @xmath1990 where @xmath1617 and where @xmath1982 is as in .
( to be more precise : @xmath1991 , where the constant @xmath1992 is as on p. . )
the proof of proposition [ exactformulaprop ] basically consists in going through the reductions carried out in section [ pfphi0xilargethmsec ] , checking that at each step the error is in fact zero , provided that @xmath1981 is sufficiently large .
we start by proving a couple of auxiliary lemmas .
[ trivlinalglem2 ] assume @xmath1993 , let @xmath1994 be the ellipsoid @xmath1995 ( i.e. on half axes @xmath1996 ) , and let @xmath1997 be an arbitrary linear subspace of dimension @xmath591 . then @xmath1998 is larger than or equal to the volume of a @xmath591-dimensional ellipsoid with ( on ) half axes @xmath1999 .
indeed , a simple application of the min - max principle in linear algebra shows that @xmath2000 is an ellipsoid with on half axes @xmath2001 satisfying @xmath2002 , @xmath2003 .
[ a2smalllem ] for any @xmath151 and @xmath101 , if @xmath254 satisfies @xmath2004 and furthermore is generic in the sense that @xmath2005 , then we necessarily have @xmath2006 in the iwasawa decomposition of @xmath329 ( cf . , ) .
let @xmath2007 satisfy the stated assumptions .
after a rotation we may assume @xmath2008 , @xmath2009 .
set @xmath1616 .
note that @xmath904 implies @xmath2010 .
hence in view of the genericity assumption we have @xmath2011 now if @xmath1994 denotes the ellipsoid @xmath2012 then @xmath2013 indeed , if @xmath2014 then @xmath2015 , @xmath2016 and @xmath2017 for @xmath2018 , and thus also @xmath2019 which proves the claim .
it follows that @xmath2020 however @xmath2021 is a lattice of covolume @xmath2022 in the @xmath2023-dimensional subspace @xmath2024 ( cf . )
; furthermore @xmath2025 is an ellipsoid centered at @xmath509 which by lemma [ trivlinalglem2 ] has volume @xmath2026 . hence by minkowski s theorem ( cf . ,
e.g. , ( * ? ? ?
* thm . 10 ) ) we must have @xmath2027 .
this proves the lemma , since @xmath2028 .
recall equation in section [ pfphi0xilargethmsec ] .
it was proved in ( * ? ? ?
7.3 ) that if @xmath2029 is sufficiently large ( which we assume from now on ) then our assumption @xmath1986 implies that all terms with @xmath1621 in _ vanish .
_ hence @xmath2030 where we write @xmath1623 .
we next use lemma [ fdcontainmentslem ] with @xmath2031 .
( note @xmath2032 , so that @xmath687 certainly holds provided @xmath2029 is sufficiently large . ) for any @xmath1623 , by lemma [ a2smalllem ] we have @xmath2033 for @xmath271-almost all @xmath1705 with @xmath904 ; thus by taking @xmath2029 sufficiently large we can force @xmath2034 to hold for these @xmath329 , and it follows that the set @xmath2035 in satisfies @xmath2036 .
hence , recalling and the discussion between and , we have @xmath2037 thus , in order to prove it now suffices to prove that , with @xmath1955 , @xmath2038 holds for all @xmath1650 . note that if @xmath683 has the property that @xmath2039 for @xmath2040almost all @xmath2041 then @xmath2042 and also @xmath2043 , cf . and the argument we used to get .
hence from now on we may assume that @xmath683 does not have the above property , i.e. we may assume that @xmath683 satisfies @xmath2044 where @xmath1647 and @xmath2045 .
we keep @xmath1650 , and as before we write @xmath2046 and @xmath2047 . by (
* lemma 7.1 ) , @xmath922 contains an open right @xmath675-dimensional cone @xmath2048 with @xmath509 in its base , which has radius @xmath2049 , height @xmath2050 and edge ratio @xmath2051 , where @xmath2052 and , since @xmath2053 , @xmath2054 contains an open right @xmath675-dimensional cone @xmath2055 with @xmath509 in its base , which has radius @xmath2056 and height @xmath2057 , where @xmath2058 now by ( * ? ? ?
1.4 ) applied with @xmath2048 , and using @xmath2059 , we see that forces @xmath2060 .
if @xmath2061 then would give @xmath2062 which is impossible if @xmath2029 is sufficiently large .
hence we must have @xmath2063 ( in particular @xmath2064 ) .
we now obtain @xmath2065 hence we conclude that , for any @xmath1650 satisfying our assumption , @xmath2066 in particular by taking @xmath2029 large we can force @xmath1654 to be less than any fixed small constant of our choice . also since @xmath2067 we may from now on assume @xmath2068 .
note from the proof of ( * ? ? ?
* lemma 7.1 ) that the heights of the cones @xmath2048 and @xmath2055 are both parallel to the line @xmath2069 thus if we let @xmath2070 be the linear map which acts by scalar multiplication by @xmath2071 on every vector in @xmath2072 and multiplication by @xmath2073 on every vector in @xmath2074 , then @xmath2075 is a cone which has both height and radius @xmath2076 , and hence by lemma [ a1largelem ] any @xmath2077 with @xmath2078 must have @xmath2079 .
also @xmath2080 is a cone with radius @xmath2081 and height @xmath2082 , where @xmath2083 and edge ratio @xmath2051 as before .
let us temporarily assume @xmath2084 .
we then claim that @xmath2085 . indeed , using and , and recalling that @xmath2063 and @xmath2068
, this claim is seen to be equivalent with @xmath2086 which is true because of @xmath2087 and .
now since @xmath2085 we may just as well assume @xmath2088 , for if @xmath2089 then we may shrink the cone @xmath2090 by keeping the base fixed while decreasing @xmath2050 until @xmath2091 , and remains true .
now implies @xmath2092 and thus ( * ? ? ?
* lemma 7.4 ) applies to give @xmath2093 . hence using , ,
we conclude @xmath2094 this is of course true also when @xmath2095 , by ; hence holds for all @xmath1650 satisfying our assumption . by
and we can force both @xmath1654 and @xmath1711 to be less than any fixed small constant of our choice , by taking @xmath2029 large .
hence also @xmath149 and @xmath2096 are forced to be small , since @xmath2097 and @xmath2098 .
furthermore both @xmath148 and @xmath147 must be near @xmath120 , by and using @xmath1976 .
hence by lemma [ goodintlem2 ] , if @xmath2029 is sufficiently large then @xmath934 holds for all @xmath1650 satisfying and all @xmath1680 ; and hence @xmath2099 furthermore by we have @xmath2100 , and hence by taking @xmath2029 large we can force the ratio @xmath2101 to be smaller than any fixed constant of our choice ; similarly by we can also force @xmath2102 to be small . hence by lemma [ goodintlem ] ( applied after appropriate rotations ) , if @xmath2029 is sufficiently large then both @xmath991 and @xmath1709 must hold , for all @xmath1650 satisfying . hence by and the same argument as we used to get , it follows that holds for all such @xmath683 , and we are done
. we will now prove an upper bound on @xmath193 which together with proposition [ phi0supportasymptthmhalfprop ] will complete the proof of theorem [ phi0supportasymptthm ] .
we first prove a weak form of the desired statement .
[ phi0supupbdweakprop ] there is a constant @xmath2103 which only depends on @xmath172 such that for any @xmath1984 , @xmath2104 and any @xmath2105 we have @xmath1598 .
assume @xmath1984 , @xmath2104 , @xmath2106 ( where we will successively impose conditions on @xmath2107 being sufficiently large ) , and @xmath191 .
then by proposition [ phi0supportthm ] we have @xmath2108 . hence if @xmath2107 is sufficiently large then @xmath1539 and @xmath1540 , where @xmath1555 is the constant in proposition [ upsilonmainupperboundprop ] .
we also require @xmath2109 ; then by proposition [ exactformulaprop ] we have @xmath2110 where we now take @xmath102 as in .
we saw in the proof of proposition [ exactformulaprop ] that @xmath2111 can only hold for @xmath1650 with @xmath2112 , and then we must also have @xmath2113 and @xmath2114 ( where @xmath2115 is some constant which only depends on @xmath172 ) , since @xmath2097 and @xmath2098 .
thus by requiring @xmath2116 we force @xmath2117 ( and hence afortiori @xmath2118 ) to hold for all @xmath1650 with @xmath2111 .
it now follows from our assumption @xmath1722 together with and proposition [ upsilonmainupperboundprop ] that there exists some @xmath1650 satisfying @xmath2114 and @xmath2119 note that the computation in applies , and since both @xmath2120 and @xmath2121 are @xmath2122 , and @xmath2123 ( as follows from @xmath980 and @xmath2124 ) , we get @xmath2125 we require that @xmath2107 is so large that the big @xmath1084-term in the last expression has absolute value @xmath2126 ; it then follows that the left hand side of is @xmath2127 , and thus @xmath2128 hence the statement of the proposition follows , after increasing @xmath2107 if necessary so as to be larger than or equal to the implied constant in the last big @xmath1084-term .
[ basicrholowboundcor ] we have @xmath2129 , uniformly over all @xmath1924 , @xmath1204 . because of corollary [ basicrhoboundcor ] and @xmath1917
, it suffices to prove @xmath2130 for all @xmath1924 , @xmath1797 .
given any @xmath1975 $ ] and @xmath1797 we may find @xmath1976 and @xmath1977 satisfying @xmath1978 , @xmath1979 , and with both @xmath1960 and @xmath149 arbitrarily small . now proposition [ phi0supportthm ] says that @xmath1980 ; thus if we take @xmath1960 sufficiently small ( for our fixed @xmath1594 ) we have @xmath2131 . but
@xmath191 for all @xmath192 , and hence proposition [ phi0supupbdweakprop ] implies that @xmath2132 .
using here @xmath1980 we obtain @xmath2130 , as desired . using corollary [ basicrholowboundcor ] we are now able to make proposition [ phi0supupbdweakprop ] a bit more precise , as follows : [ phi0supupbdprop ] we have @xmath2133 uniformly over all @xmath195 , @xmath200 $ ] .
because of @xmath1958 and , we may assume @xmath1959 without loss of generality .
let us write @xmath1978 and @xmath1979 , as usual .
it follows from corollary [ basicrholowboundcor ] that there is a constant @xmath2134 which only depends on @xmath172 such that @xmath2135 holds whenever @xmath2136 and @xmath2137 .
hence for any such @xmath185 , if we let @xmath467 be the unique real positive solution to the equation @xmath2138 then @xmath2139 and hence by proposition [ phi0supupbdweakprop ] we have @xmath2140 . also @xmath2141 ; thus @xmath2142 and @xmath2143 which implies the desired bound . in the remaining case when @xmath2144 or @xmath2145 , the desired bound follows automatically from corollary [ basicrholowboundcor ] and proposition [ phi0supportthm ] .
note that theorem [ phi0supportasymptthm ] follows from proposition [ phi0supupbdprop ] and proposition [ phi0supportasymptthmhalfprop ] , together with the discussion at the end of section [ xi0wzphisec ] .
@xmath324 @xmath324 @xmath324 we may now also give the simple proof of corollary [ phixiwsupcor ] : the existence and continuity of @xmath237 is proved by a similar argument as for @xmath193 ( cf .
the end of section [ xi0wzphisec ] ) , working directly from the definition of @xmath74 , .
in particular , the upper semicontinuity of @xmath237 follows from the fact that @xmath2146 holds whenever @xmath2147 , where @xmath2148 with @xmath2149 .
this in turn follows from the inclusion @xmath2150 , where @xmath2151 .
next , in order to prove the asymptotic formula we note that , by , @xmath2152\bigr\ } , \qquad\forall w\in[0,1).\end{aligned}\ ] ] it follows from proposition [ phi0supportthm ] that there exists a constant @xmath1136 which only depends on @xmath172 such that , for any @xmath1159 , the supremum in remains unchanged if we restrict to the set @xmath2153{\ : : \:}1-z\leq c(1-w),\ : \varphi\leq c\sqrt{1-w}\bigr\}.\end{aligned}\ ] ] hence by theorem [ phi0supportasymptthm ] , as @xmath2154 we have @xmath2155 however it is immediate from the definition that @xmath2156 ; hence @xmath2157 for all @xmath1203 , @xmath1204 ( cf . and ) , so that the supremum in is attained at @xmath2158 , @xmath157 . hence holds .
changing to a slightly different topic , we now give the quick proof of proposition [ phizeroasympt ] .
it follows e.g. from and ( * ? ? ?
* lemma 7.11 ) that @xmath145 is a continuous and decreasing function of @xmath82 ; furthermore and imply @xmath2159 . hence we have @xmath235 if and only if @xmath2160 , and by this holds if and only if @xmath2161 using @xmath2162 and the fact that @xmath2163 holds for @xmath256-almost all @xmath1296 , we see that holds if and only if @xmath2164 but note that @xmath1296 satisfies @xmath2165 if and only if @xmath638 is a packing lattice of @xmath2166 ( cf . ,
e.g. , ( * ? ? ?
20 , thm 1 ) ) . hence holds if and only if @xmath2167 , i.e. if and only if @xmath2168 .
in this section we use theorem [ phi0xilargethm ] and to derive an asymptotic formula for @xmath2170 as @xmath76 .
in particular this results in a new proof of theorem [ phixiwasymptthm ] ( except for a slightly worse @xmath2171-factor ) , and it also gives an internal check of consistency of our asymptotic formulas .
[ pphixiwasymptthm ] for any @xmath197 we have @xmath2172 as @xmath76 , uniformly over all @xmath209 , where @xmath2173 the function @xmath2174 is a bounded continuous function from @xmath2175 to @xmath2176 .
note that the second equality in follows immediately from the definition of @xmath212 ( cf . ) and the fact that @xmath2177 for all @xmath40 .
this last identity is proved by substituting @xmath2178 in the outer integral in the left hand side , then carrying out the differentiation with respect to @xmath8 , and finally substituting back @xmath2179 .
the fact that @xmath2174 is bounded and continuous is proved by the same argument as below , and similar considerations also justify the differentiation in .
note in particular that it follows that @xmath212 is @xmath45 on all @xmath2175 .
note also that the main term in is what is obtained by differentiation of the main term in .
hence theorem [ pphixiwasymptthm ] indeed implies theorem [ phixiwasymptthm ] ( except for a slightly worse @xmath2171-factor ) upon integrating over @xmath2180 , and using the fact that for any fixed @xmath147 we have @xmath896 for all sufficiently large @xmath82 . the key fact needed for the deduction of theorem [ pphixiwasymptthm ] from theorem [ phi0xilargethm ] is the following : [ xiaveragelem ] for any @xmath1033 and @xmath1047 we have ( writing @xmath1322 as usual ) @xmath2181 by definition we have @xmath2182 hence by fubini s theorem the left hand side of equals @xmath2183 but for @xmath256-almost every @xmath1300 we have @xmath2184 , and for each such @xmath329 the set @xmath2185 is in fact a fundamental domain for @xmath2186 .
indeed , for every @xmath2187 the set @xmath2188 is infinite , and there is exactly one @xmath2189 for which @xmath2190 , namely that @xmath2189 for which @xmath2191 is minimal ( the uniqueness is guaranteed since @xmath2184 ) .
hence the volume of the set in equals @xmath1049 , and the lemma follows . as in section [ simplintsec ]
we fix the constant @xmath898 so that @xmath1597 ( cf . ) and @xmath896 whenever @xmath897 .
then @xmath2192 for all @xmath900 , and hence is automatic whenever @xmath897 .
hence from now on we will assume @xmath901 .
by we have : @xmath2193 let us write @xmath2194 and @xmath2195 . by proposition [ phi0supportthm ]
there is a constant @xmath179 which only depends on @xmath172 such that @xmath2196 holds for all @xmath2197 , where @xmath2198 now for @xmath2199 we apply theorem [ phi0xilargethm ] to @xmath2200 , while for @xmath2201 we apply the bound from theorem [ cylinder2ptsmainthm ] .
this gives @xmath2202 where @xmath966 is as in .
we parametrize @xmath151 as @xmath2203 then @xmath2204 let us first consider the contribution from the error term @xmath931 in .
for @xmath2199 we have @xmath2205 and @xmath2206 .
hence the contribution from the @xmath931-term in is : @xmath2207 we may assume that @xmath1374 is so large that @xmath2208 .
then @xmath2209 , since @xmath901 , and hence if @xmath181 then we get ( cf . )
@xmath2210 when @xmath121 we get the same bound except for an extra factor @xmath2211 .
the integral over @xmath2212 in is easily seen to be @xmath2213 .
( this bound was also pointed out in ( * ? ? ?
* cor . 1.10 ) .
) hence we have @xmath2214 where we write @xmath2215 for the error majorant in the second line of .
next note that in we have @xmath2216 uniformly over all @xmath2199 .
using this fact together with lemma [ sabvboundlemcor ] we see that the error caused by replacing @xmath2217 by @xmath2218 in is @xmath2219 and this is seen to be @xmath2220 by a computation which is very similar to the computation above bounding the contribution from @xmath931 in .
hence @xmath2221 now by corollary [ basicrhoboundcor ] and we see that by requiring that the constant @xmath1374 has been chosen sufficiently large ( in a way that only depends on @xmath172 ) we have that any tuple @xmath2222 for which @xmath2223 must satisfy both @xmath2224 and @xmath2225 . from now on we assume that @xmath82 is so large that @xmath2226 .
it then follows that in the outer integral in we may extend the range @xmath2227 to the larger set of all @xmath2228 without changing the value of the integral .
in fact if @xmath2229 then we may extend the range all the way to @xmath2230 without changing the value of the integral ; on the other hand if @xmath2231 then we may extend the range to @xmath2230 at the cost of an error which is @xmath2232 hence we obtain , after substituting @xmath2233 and using ( note that the following is correct also for @xmath121 , with the convention that @xmath2234 ) : @xmath2235 next , using we see that @xmath2236 for any @xmath2237 and any @xmath2238 with @xmath2239 . integrating over all @xmath2237
we get that the main term in equals @xmath2240 where @xmath2241 is the orthogonal complement of @xmath312 in @xmath161 .
substituting now @xmath2242 and then letting @xmath2243 , we get @xmath2244 applying now lemma [ xiaveragelem ] and fubini s theorem we get @xmath2245 where we substituted @xmath2246 ( @xmath1305 , @xmath2247 , @xmath2237 ) and used the fact that @xmath2248 . hence is proved .
[ cols= " < , < , > " , ] j. h. conway and n. j. a. sloane , _ sphere packings , lattices and groups _ ,
third edition , springer - verlag , new york , 1999 .
p. dahlqvist , the lyapunov exponent in the sinai billiard in the small scatterer limit .
nonlinearity * 10 * ( 1997 ) , 159173 . | the dynamics of a point particle in a periodic array of spherical scatterers converges , in the limit of small scatterer size , to a random flight process , whose paths are piecewise linear curves generated by a markov process with memory two .
the corresponding transport equation is distinctly different from the linear boltzmann equation observed in the case of a random configuration of scatterers . in the present paper we provide asymptotic estimates for the transition probabilities of this markov process .
our results in particular sharpen previous upper and lower bounds on the distribution of free path lengths obtained by bourgain , golse and wennberg . |
the macho project is monitoring millions of stars every night searching for the gravitational microlensing signature of massive compact halo objects ( machos ) .
this project has the dedicated use the great melbourne telescope at mount stromlo .
this endeavor was stimulated by paczyski s suggestion @xcite that gravitational microlensing was a possible way to detect baryonic dark matter in the halo of the milky way .
the principle of microlensing is simple ; if a macho lies near the line of sight to a background star ( the source ) , it will deflect light from the source and produce two images . for galactic scales
, these images can not be resolved even by hst ( @xmath12 arcsec and thus microlensing ) , but the two unresolved images combine to give an apparent increase in the source brightness . due to the relative motions of the observer ,
lens and source , this magnification is transient , so the effect appears as a symmetrical and unique brightening in an otherwise constant star .
the duration of the event is a function of the mass of the lens , the relative distances of the lens and source and the motion of the lens with respect to the line of sight .
the magnification is just a function of the distance of the lens from the line of sight .
the probability that a source will be microlensed is termed the optical depth to microlensing , @xmath13 .
the large magellanic cloud ( lmc ) was chosen as the primary target because its line of sight passes through much of the halo yet it is relatively nearby , and contains millions of stars resolvable in modest seeing with a small telescope .
our secondary target was chosen as the galactic bulge for two reasons .
although the lmc is circumpolar from mount stromlo , it is at too high an airmass to be usefully observed when the bulge is overhead .
calculations before we began taking data @xcite suggested that there should be an observable microlensing signal from known stellar populations in the disk and the bulge so observing the bulge would serve to prove whether microlensing could be detected with our system .
we also observe the small magellanic cloud ( smc ) at a lower priority .
the smc provides a different line of sight through the halo , but is sufficiently farther away and smaller than the lmc that we can only monitor a few million stars with our system .
= .45 the refurbishment of the 1.27-m great melbourne telescope has been described by hart _
figure [ camera ] shows in detail the corrector , beam splitter and ccd cameras which allow us to monitor 0.5 square deg simultaneously in two passbands .
these two passbands were chosen to maximally utilize the quantum efficiency of our thick ccds while cutting out much of the sky brightness due to oh emission in the red and scattering in the blue .
figure [ camera ] also shows the total throughput as a function of wavelength for our two cameras .
each camera contains a 2x2 mosaic of 2048x2048 ccds .
exposure times are 300 s toward the lmc , 150 s toward the bulge and 600 s toward the smc .
it takes about 70 s to read out the 32 mpixels .
details of the camera system can be found in @xcite .
images are flattened immediately upon acquisition , then simultaneously archived to 8-mm tape and photometry begun .
reductions and analysis are done in the dome using two sparc 1000s and a large disk farm ( currently @xmath14 200 gbyte ) .
macho field centers are defined so that the 0.5 square degree fields tile the denser regions of the lmc , smc and much of the southeast quadrant of the bulge within 10 degrees of the galactic center . these field centers ( as well as much additional information on the macho project )
can be found at url http://wwwmacho.anu.edu.au .
our observing strategy is to sequentially image each of these fields , spiraling outward from dense bar fields in the lmc and generally moving from west to east in the bulge . as of september 1996 , over 48,000 exposures
have been taken with the system , of which about @xmath15 are of the lmc .
the data covered in this presentation is derived from the central 22 fields in the lmc observed for the first two years of the project ( september 1992 to november 1994 ) and a central 24 fields in the bulge observed in the project s first bulge season ( 1993 ) .
deep images from the early months of the project have been reduced to create photometry templates .
a template is divided into 64 ` chunks ' in each color and relatively bright , unblended stars in each chunk are chosen as fiducials to define the point spread function ( psf ) and astrometric reference frame in future observations . a routine observation is reduced chunk by chunk by determining the geometric transformation from the template to the current chunk using the fiducial stars , defining a psf using the fiducials , and fitting at each transformed template position ( more details can be found in @xcite and @xcite ) . this scheme allows rapid reductions , and a whole night of observations can be reduced within a few hours of dawn .
although the microlensing search is effected using only relative photometry , it is useful to derive transformations to standard astronomical passbands .
this is complicated by two systematic effects : 1 ) the template observation for each field is subject to different extinction , seeing , and possible cloud obscuration , and 2 ) the photometric zero point varies from chunk to chunk ( see @xcite ) within a template observation due to systematic effects of the photometry code .
( this is due to the fact that there is no effort made to derive an aperture correction for individual chunk reductions . )
we have obtained roughly 30 photometric nights of ccd data using the ctio 0.9-m and sso 1-m to measure the necessary 1408 transformations ( 64 chunks per field and 22 fields ) .
although the analysis of this data is not complete , we have determined a mean transformation to johnson v and kron - cousins r which we will use in this paper .
this transformation yields magnitudes good to about @xmath16 mag and colors which are good to about @xmath17 mag ( the zero point offsets in the two passbands are correlated ) .
light curves of time - ordered photometry are then built for all template objects .
light curves are not created for objects which are not in the template .
these light curves are then analyzed for evidence of microlensing @xcite @xcite @xcite . because photometry is available within hours of an observation ,
a star s current magnitude can be compared to its mean magnitude determined from a previous analysis and if a previously constant star demonstrates a significant deviation , our software produces an alert .
these alerts have been used to notify observers around the world that microlensing events are in progress ( see http://darkstar.astro.washington.edu ) .
* table 1 .
* optical depth estimates [ cols="<,<,<",options="header " , ] = 0.7
we have detected significant microlensing toward both the lmc and the bulge .
our microlensing discoveries are important for three reasons : 1 ) we have detected significant microlensing by halo dark matter looking toward the lmc , 2 ) we have detected no short 1 microlensing events looking toward the lmc and so have ruled out a significant fraction of the halo being composed of objects with mass @xmath18 , and 3 ) we have detected a higher optical depth for microlensing toward the bulge than predictions based on simple milky way models , lending support to the presence of a significant bar .
it is important to note that the microlensing interpretation for the events we report has been strongly supported and can be considered proven .
the majority of events are well fit by the simple functional form needed to describe point source , point lens microlensing .
there have been spectroscopic observations through the course of microlensing toward the bulge @xcite and the lmc @xcite which have shown no spectral changes during apparent brightness changes of more than one magnitude in the span of a week .
the distribution of magnifications has been shown to be that expected from the uniform distribution of impact parameters which result from microlensing @xcite .
exceptions to simple lensing support the microlensing interpretation because they are predicted deviations due to the motion of the earth @xcite , due to the finite size of stars @xcite , and due to the existence of binary lenses @xcite @xcite .
while there are deviations from simple lensing due to the blending of lensed and unlensed images , this is well understood and can be accounted for by fitting an unmagnified fraction of the baseline flux , and thus adding an additional fit parameter in each color @xcite @xcite .
we have recently concluded an analysis of the first two years of data on 22 fields in the lmc .
these data have yielded nine microlensing candidates eight of which are shown in figure [ lmcmicro ] .
the position of the baseline color and magnitude for these 8 events is shown in a composite color - magnitude diagram of lmc data in figure [ microcmd ] .
the ninth event occurred in an unresolved source star which was lensed during the template observation . because our efficiency analysis does account for sources which are not represented by objects in our template ,
this event is not considered further .
the measured optical depths for various samples in both the lmc and the bulge are shown in table 1 .
a fraction more than one microlensing event would be expected from known populations in the milky way and the lmc after monitoring 8.6 x @xmath19 stars for two years .
thus , we considered two samples for analysis , a six event sample ( excluding one possible lmc - lmc event and one additional event ) and an eight event sample .
both samples produce similar conclusions .
we detect about half the predicted optical depth if the dark halo were composed of only machos .
a maximum likelihood analysis suggests that we are detecting @xmath1450% of the dark halo mass in objects of @xmath14 0.5 m@xmath20 .
a full account of this analysis is forthcoming @xcite .
this analysis and a companion analysis looking for evidence of short microlensing events in our data @xcite support and extend the eros negative result @xcite demonstrating the absence of short microlensing events .
our result limits objects in the mass range @xmath18 ( venus through brown dwarf masses ) to contributing less than 20% of the dark halo .
= .45 although the macho project has only analyzed one season of bulge data to date , this season yielded 42 microlensing candidates .
we have detected an additional 75 bulge events with our real time alert system since late 1994 .
we have used both the full 1993 sample and a subset derived from bright , clump giant source stars to calculate an optical depth to the bulge . because detailed efficiency simulations have not been completed for our bulge data , the corrections for faint source stars are somewhat uncertain , but the clump giant sample is not subject to significant efficiency corrections .
both samples yield a higher optical depth than the expected @xmath21 optical depth to the bulge @xcite @xcite .
ogle and duo report similar measurements of optical depth @xcite @xcite .
these data support the hypothesis of a massive bar in the milky way and may also provide evidence of significant disk dark matter @xcite @xcite .
microlensing of apparent disk stars as evidenced by the microlensing events along the upper blue sequence in figure [ microcmd ] underscores the possibility of significant disk dark matter .
= .45 = .3 the color - magnitude diagrams ( cmds)of figure [ microcmd ] demonstrate the photometry product of the microlensing search .
the mean photometric error at v = 20 for hundreds of epochs is about 25% for the lmc and about 40% for the bulge .
the lmc cmd is created from less than 1% of the lmc light curves while the bulge cmd is created from about 1% of the bulge light curves .
much of the scatter in these diagrams is due to the uncertainty in the transformations as discussed earlier . as part of the analysis of the light curves in the search for microlensing ,
a variability statistic is calculated for each light curve .
the @xmath22/dof for a fit to a constant is calculated using only the 80% of points
closest to the median for each light curve .
the light curve is defined as variable if this , robust , @xmath22/dof is greater than the regular @xmath22/dof which can occur by chance 1% of the time for a constant light curve . the nightly sampling for extended periods of time results in an effective sampling even for variables with periods much shorter than a day .
this is because the sampling interval constantly changes since we attempt to observe each field at its lowest airmass in a night . of the @xmath14 9 million lmc light curves from the first two years of lmc data
37,051 light curves were identified as variable . of the @xmath14 12 million bulge light curves currently analyzed from 1993 ,
38,718 were identified as variable .
an additional analysis of lmc light curves was conducted to identify those which were rejected because of relatively rare excursions below the robust envelope defined above .
this analysis identified about 1000 more eclipsing binaries , 3000 more low amplitude rr lyrae ( primarily due to confusion ) , and about 6000 low amplitude , low signal - to - noise variable candidates .
these 85,000 light curves were tested for periodic behavior using a code designed specifically for the macho project @xcite .
when the red and blue light curve yielded the same period to within 0.1 % the light curve was defined as representing a periodic variable .
figure [ lmcvars ] shows period - amplitude and period - magnitude plots for the 23,589 periodic variables identified in the lmc .
figure [ blgvars ] shows the period - amplitude and period - magnitude plots for the 14,776 periodic variables identified in the bulge .
a preliminary description of the lmc variables has been presented @xcite .
these figures hold a wealth of information about the variable stellar populations of the lmc and the bulge .
there are also certain systematic features in this data set which are apparent .
there are vertical lines in the distributions for two reasons due to the phasing code we have used .
the phasing code will return an alias of an integer fraction of a day when it tries to phase a low signal - to - noise ( low amplitude ) aperiodic variable . for high signal - to - noise aperiodic or very long period variables , it will fold them with a period of the light curve length .
there are multiple , long period vertical lines because different fields yielded different light curve durations .
we have used these figures to define regions in period , magnitude , amplitude space which contain highly purified samples of particular types of variables .
we have also used the blue - to - red amplitude ratio to discriminate pulsating from eclipsing variables @xcite .
one can clearly see the rrab and rrc in the lmc as distinct clumps near log(period ) = -0.3 and -0.5 , respectively . in the bulge
, the rrc locus is clearly contaminated by short period , low amplitude variables , mostly eclipsing binaries .
the fundamental and overtone cepheid period - luminosity ( pl ) loci stand out in the lmc data and are absent in the bulge data as would be expected because of its mostly old population .
perhaps the most striking feature of these diagrams is the appearance of clear , multiple pl relations for bright , long period variables
. these are all giants lying just beyond the tip of the giant branch .
as can be seen , they have low amplitudes with a suggestion of increasing amplitude with increasing period . eclipsing binaries are found scattered across these diagrams .
more detailed discussions of the @xmath14 10,000 rr lyrae @xcite , 1500 cepheids in the lmc @xcite , and type ii cepheids and rv tauri in the lmc @xcite contained in the macho database are presented elsewhere in this volume .
examples of some of the more interesting periodic variables from our database are shown in figure [ varfig ] .
eclipsing binaries from the eros and ogle survey are discussed in detail elsewhere in this volume @xcite @xcite @xcite .
examples of extremes in period for eclipsing systems in the macho data are shown in the left panels .
about a dozen eclipsing , red giant pairs with periods over 100 days have been identified in our lmc data so far .
two examples of the low amplitude , periodic red giant variables with periods less than 100 days are shown in the middle panels .
there seems to be quite a range in the constancy of the amplitude of stars in this region of the period - amplitude diagram .
the relatively short periods and the sinusoidal light curves are suggestive of overtone oscillations .
although there are few classical cepheids along our line of sight to the bulge , we have discovered about 20 new bl her and w vir stars .
examples of two of these are shown in right panels . we have also begun a detailed examination of the aperiodic blue variables found in the lmc data .
our interest in these stars was stimulated by the observation of rare , small amplitude outbursts in mostly constant stars which we have termed bumpers .
we have compiled a catalog of about 400 aperiodic blue variables which demonstrate some truly odd photometric behavior .
these stars were chosen from about 250,000 light curves with @xmath23 and @xmath24 @xcite .
characteristic light curves of two major classes of variability found in the catalog have been previously presented @xcite .
figure [ bump ] shows three other interesting blue , aperiodic variables : an odd one , a background seyfert and a nova . in this set of aperiodic blue variables
we have also found supersoft x - ray sources , and planetary nebulae .
the unevenly spaced sampling , the wide field of view , the simultaneous color information and the long , densely sampled light curves make the macho database a treasure trove for the study of time variability of astrophysical sources . while we have only begun to examine this database outside the context of the search for machos , the early returns promise much for the future .
the macho project is also moving rapidly toward achieving its goal of determining the baryonic content of the halo .
the macho project plans to continue operation at least through 1999 . at the end of this time
, we expect to have a sample of about 50 microlensing events across the face of the lmc , about 1000 events toward the bulge , and a database of light curves for more than 50 million stars in the lmc , the smc and the bulge , spanning 8 years .
990.4 cm alard , c. 1995 , eso messenger , * 80 * , 31 alcock , c. , et al .
( the macho collaboration ) 1995 , apj , * 454 * , l125 alcock , c. , et al . ( the macho collaboration ) 1996 , apj , * 461 * , 84 alcock , c. , et al . ( the macho collaboration ) 1996 , apj , _ in press _ and astro - ph/9512146 alcock , c. , et al . ( the macho collaboration ) 1996 , apj , _ in press _ and astro - ph/9604176 alcock , c. , et al . ( the macho collaboration ) 1996 , apj , _ in press _ and astro - ph/9605148 alcock , c. , et al . ( the macho collaboration ) 1996 , in preparation auborg , e. , et al . ( the eros collaboration ) , 1995 , aa * 301 * , 1 bennett , d. et al .
( the macho collaboration ) 1996 , astro - ph/960612 bennetti , s. , et al . ,
1995 aa , * 294 * , l37 cook , k. h. et al .
( the macho collaboration ) 1995 , asp conf .
proc . , * 83 * , 221 cowley , a. , et al , 1984 , aj , * 286 * , 196 griest , k. , et al ( the macho collaboration ) 1991 , apj , * 366 * , 412 hart , j. et al ( the macho collaboration ) , 1996 , pasp , * 108 * , 220 liller , w. , 1992 , iauc 5651 marshall , s. ph.d .
thesis , uc santa barabara minniti , d. , et al ( the macho collaboration ) these proceedings paczyski , b. 1986 , apj , * 304 * , 1 paczyski , b. 1991 , apj , * 371 * , l63 paczyski , b. , 1994 , apj , * 435 * , l113 paczyski , b. 1996 , these proceedings pollard , k. a. , et al ( the macho collaboration ) , these proceedings purdue , p. , et al ( the macho collaboration ) , 1995 , baas * 27 * , 1402 pratt , m. , et al .
( the macho collaboration ) 1995 , baas , * 27 * , 1351 reimann , j. ph.d . thesis , uc berkeley rucinski 1996 , these proceedings stubbs , c. , et al ( the macho collaboration ) , 1993 , spie proc . * 1900 * , 192 tobin , w. , 1996 , these proceedings udalski , a. , et al . 1994 , acta astron . , * 44 * , 165 udalski , a. , et al . 1994 ,
apj , * 435 * , l113 welch , d. , et al ( the macho collaboration ) , these proceedings | 0.4 cm the macho project monitors millions of stars in the large magellanic cloud , the small magellanic cloud and the bulge of the milky way searching for the gravitational microlensing signature of baryonic dark matter .
this project has yielded surprising results .
an analysis of two years of data monitoring the large magellanic cloud points to @xmath0 of the mass of the milky way s halo in compact objects of @xmath1 .
an analysis of one year of monitoring the bulge has yielded more microlensing than predicted without the invocation of a massive bar or significant disk dark matter .
the huge database of light curves created by this search is yielding information on extremely rare types of astrophysical variability as well as providing temporal detail for the study of well known variable astrophysical phenomena . the variable star catalog created from this database
is previewed and example light curves are presented .
@xmath2 lawrence livermore national laboratory , livermore , ca 94550 + @xmath3 center for particle astrophysics , university of california , berkeley , ca 94720 + @xmath4 supercomputing facility , australian national univ . ,
canberra , act 0200 , australia + @xmath5 department of physics , university of california , davis , ca 95616 + @xmath6 department of physics and astronomy , university of washington , seattle , wa 98195 + @xmath7 department of physics , university of california san diego , la jolla , ca 92093 - 0350 + @xmath8 european southern observatory , d-85748 garching bei mnchen , germany + @xmath9 department of physics , university of oxford , oxford ox1 3rh , u.k . + @xmath10 dept . of physics and astronomy , mcmaster univ . ,
hamilton , ontario , canada l8s 4m1 + @xmath11 mount stromlo and siding springs obs .
, australian natl .
univ . ,
weston , act 2611 , australia |
infrared ( ir ) emission bands reveal the presence of polycyclic aromatic hydrocarbon ( pah ) molecules in the interstellar medium ( ism ) @xcite .
these so - called aromatic infrared bands ( aibs ) represent a collective fingerprint of the interstellar pahs , which can not be resolved to lead to the identification of individual molecules .
as the relative intensities and profiles of the aibs vary depending on the region where the bands are observed , different populations of pahs are proposed to account for the variations .
the populations are defined in terms of charge state , size distribution , presence of heterocycles and side groups , and clusterization state ( for a review , see * ? ? ?
@xcite measured ir spectra of synthetic carbon nanoparticles containing @xmath6-bonded chains , both polyynes and cumulenes . having found similarities between the laboratory spectra and interstellar ir emission bands , they conjectured the presence of pah molecules carrying @xmath6-bonded chains in ir emission regions . to date , this conjecture awaits verification . the failure to find the molecules in ir sources , however ,
would not demonstrate their absence from regions in which different conditions prevail and , therefore , another chemistry takes place , such as the diffuse ism .
the search for molecules in the diffuse ism is conducted by looking for absorption bands corresponding to their electronic transitions , which typically occur at near ultraviolet ( uv ) and visible ( vis ) wavelengths . since little
is known about the electronic spectra of pahs carrying @xmath6-bonded chains , it appears worthwhile to undertake a study of such species .
we report on the uv / vis absorption spectra of polyynyl - substituted pahs isolated in ne matrices at 6 k and on the ir spectra of these molecules obtained with samples embedded in salt pellets .
quantum chemical calculations have been performed to assist in the interpretation of the laboratory spectra . in the light of the new data , we discuss the relevance of polyynyl - substituted pahs to astrophysics .
figure [ fig1 ] shows the structures of the molecules studied . alongside the non - substituted pahs anthracene , phenanthrene , and pyrene
, they include a selection of derivatives obtained by replacing an h atom with either an ethynyl ( @xmath0c@xmath1h ) or a butadiynyl ( @xmath0c@xmath2h ) group .
several of these substances were obtained from commercial providers .
anthracene ( aldrich , purity 97% ) , phenanthrene ( aldrich , purity 98% ) , 9-ethynylphenanthrene ( aldrich , purity 97% ) , pyrene ( aldrich , purity 99% ) , and 1-ethynylpyrene ( abcr , purity 96% ) were used as received . the other substances , namely 9-ethynylanthracene , 9-butadiynylanthracene , 9-butadiynylphenanthrene , and 1-butadiynylpyrene , were synthesized in - house .
they were prepared via negishi coupling @xcite of the corresponding bromoarenes with an appropriate silylated alkyne followed by protodesilylation .
the structure of these molecules was verified through nmr measurements . moreover ,
the purity of the in - house samples was controlled by hplc .
the chromatograms yielded purities of 96 , 93.5 , @xmath799 , and @xmath799% for 9-ethynylanthracene , 9-butadiynylanthracene , 9-butadiynylphenanthrene , and 1-butadiynylpyrene , respectively .
these four substances were kept under ar atmosphere in vials stored in a dewar flask filled with co@xmath1 ice and were used as synthesized . for convenience ,
the names of the substances are abbreviated .
anthracene , phenanthrene , and pyrene are referred to as ant , phn , and pyr , respectively .
each derivative is indicated with a prefix comprising the number that identifies the carbon atom carrying the side chain followed by ety for ethynyl or buy for butadiynyl .
for instance , 9etyant stands for 9-ethynylanthracene .
the ir absorption spectra of the six polyynyl - substituted pahs were measured from 2.5 to 67 @xmath3 m at room temperature on grains embedded in csi pellets .
each pellet was prepared by pressing 200 to 400 mg of a mixture of the sample with csi powder . in every mixture ,
the mass ratio of sample to csi was 1:500 .
the ir spectra were obtained by transmission using a fourier transform ir spectrometer ( bruker 113v ) with a resolution of 2 @xmath8 and an averaging over 32 measurements .
spectra were measured in two steps as different detectors had to be used , one for the 150660 @xmath8 range , another for the 4004000 @xmath8 domain .
a pellet of pure csi was used to provide a reference spectrum for background subtraction .
further correction was necessary and was carried out by subtracting a polynomial function fitted to the background .
the matrix isolation spectroscopy ( mis ) setup was described in previous studies @xcite .
briefly , it consists of a commercial spectrophotometer ( jasco v-670 ex ) coupled to a vacuum chamber by means of optical fibers .
the chamber is equipped with a closed - cycle he cryocooler ( advanced research systems inc .
de-204sl ) .
we used ne ( linde , purity 99.995% ) as the matrix material .
in the course of each experiment , a caf@xmath9 substrate was placed into the vacuum chamber and cooled to a temperature of 6 to 7.5 k by the action of the cryocooler .
the rare gas was fed into the vacuum chamber at a flow rate of 5 sccm ( standard cubic centimeter per minute ) , and it condensed on the cold substrate . in order to produce pah - doped ne matrices ,
the rare gas was simply seeded with pah molecules . under normal conditions , most pahs , including those presently studied , are in the solid phase .
still , molecules are released into the gas phase and the associated vapor pressure is temperature - dependent .
thus , the sample under study was kept in a small reservoir connected to the rare - gas line and its vapor pressure was optimized by varying its temperature .
so as to obtain the spectra of ant , 9etyant , and 9buyant , the samples were kept at 1824 , 5.08.5 , and 63 @xmath10c , respectively .
the deposition of the respective matrices took 24 , 105 , and 95 min .
for phn , 9etyphn , and 9buyphn , temperatures of 20.5 , 13.5 , and 61 @xmath10c were used , respectively , and the corresponding matrices were formed in 3 , 60 , and 115 min .
note that , as an exception , the flow of ne in the case of phn was 7 sccm .
regarding pyr , the sample was left at room temperature ( 21 @xmath10c ) and the deposition of the matrix took 45 min .
the temperature applied to 1etypyr was 40 @xmath10c and the matrix was grown in 50 min .
attempts to obtain spectra of 1buypyr by following the procedure described above were unsuccessful . whereas the heated 9buyant and 9buyphn samples released enough molecules to dope ne matrices , the 1buypyr powder did not , as inferred from the absence of 1buypyr bands from the measured spectra .
temperatures as high as 105 @xmath10c had been applied .
it did not seem reasonable to consider higher temperatures in view of the behavior of the other two butadiynyl - substituted species compared to those of the corresponding ethynyl - substituted and parent pahs .
for this reason , we applied laser vaporization to 1buypyr . a little amount ( @xmath111 mg ) of 1buypyr powder
was spread over the surface of a caf@xmath1 substrate . in order to make the 1buypyr grains stick to the substrate ,
drops of methanol , one at a time , were deposited on them and let to evaporate . in the mis chamber , the sample - carrying substrate was placed with its back to the ne gas inlet and its sample - covered side facing the subtrate on which the matrix would form .
vaporization of 1buypyr was then obtained by directing a pulsed laser beam toward the sample .
the laser source ( continuum minilite ii ) emitted photons with a wavelength of 532 nm , in pulses of @xmath1110 ns duration and 1 mj energy .
it was operated with a repetition rate of 10 hz . the laser spot at the surface of the sample had an approximate diameter of 2.7 mm , giving a fluence of 18 mj @xmath12 .
as the ne flow was 5 sccm , the sample was exposed to the laser beam for a total duration of 20 min .
every 2 min , the position of the laser spot on the sample was shifted . due to the limited size of the sample , some overlap of the irradiated spots could not be avoided . the sample darkened and blackened during the experiment , especially where the irradiated spots overlapped , and was not entirely vaporized . when the bands of the @xmath13 transition were too weak for being
clearly observed in the extended spectra , new measurements were carried out after further growth of the matrix , with an increased temperature of the oven when it was necessary .
all our measurements were carried out in steps of 2 , with a scanning speed of 110 min@xmath14 , and at a resolution of 2 .
the spectrophotometer was calibrated with an accuracy of 3 .
the spectra presented result from the averaging of up to nine measurements .
theoretical calculations were carried out to assist in the interpretation of the various spectra .
we have determined the theoretical structures of the molecules in their electronic ground state . for this purpose , calculations based on the density functional theory ( dft ) were performed with the gaussian 09 software @xcite .
the b3lyp functional was used in conjunction with the cc - pvtz basis set .
the tight convergence criteria implemented in the software were applied when optimizing the molecular geometries .
optimization procedures were conducted while forcing ant and pyr into @xmath15 geometry , phn , 9etyant , and 9buyant into @xmath16 , and the other molecules into @xmath17 .
vibrational frequencies were computed to verify that the optimized structures corresponded to minima in the potential surfaces .
in addition , the results included the values of the permanent electric dipole moment .
the predefined ultrafine grid of gaussian 09 was used in all dft - based calculations .
electronic states and transitions were calculated at the geometries determined for the electronic ground states .
both the zindo semi - empirical model and the time - dependent density functional theory ( td - dft ) approach were used .
the td - dft calculations were carried out with the b3lyp functional and the cc - pvtz basis set already employed for the geometry optimization of the electronic ground states .
the ultrafine grid was used in these calculations as well .
to our knowledge , the uv / vis spectra of the butadiynyl - substituted pahs are reported here for the first time ever , independently of the experimental conditions . while the spectra of the ethynyl - substituted pahs were measured before in solutions at room temperature @xcite , our low - temperature , matrix - isolated spectra are the first to be published . finally , the full uv / vis spectrum of ant isolated in a ne matrix is also presented for the first time .
the spectra we obtained for phn and pyr are in agreement with previous measurements @xcite .
the ir spectra of the butadiynyl - substituted pahs are also presented for the first time .
our spectra of 9etyant and 9etyphn are in very good agreement with results obtained from films condensed on substrates kept at 77 k @xcite .
we have not found detailed reports on the ir spectrum of 1etypyr , preventing comparison with earlier results .
the ir spectra of the non - substituted pahs , measured under various conditions , can be found in the literature , e.g. , in @xcite for pahs isolated in ar matrix and in @xcite for measurements in kbr pellets and in solutions .
figure [ fig2 ] shows the electronic spectra of ant and its derivatives . in the spectrum of ant , the features of the well - known @xmath13 transition are clearly seen .
they consist of the origin band at 3634 and the associated vibronic bands that form a progression on its blue side to approximatly 3000 . at shorter wavelength ,
a broad feature dominates the spectrum , exhibiting multiple peaks with a maximum at 2374 . in accordance with clar s notation @xcite
, it is denoted by @xmath18-band .
the spectrum of 9etyant is similar in its general aspect .
it shows a series of bands with its origin at 3868 , which we assign to the @xmath13 transition .
one can note that the band pattern differs slightly from the one given by the @xmath13 transition of ant , due to the presence of a new band next to the origin .
the frequency of the corresponding vibrational mode is @xmath11209 @xmath8 . with 9etyant too ,
the spectrum is dominated by a broad feature attributed to the @xmath18-band , which reaches its maximum at 2440 . in comparison with the spectrum of ant ,
the bands are redshifted , by 234 in the case of the origin of the @xmath13 transition .
the bands are also broadened , and the peaks surmounting the @xmath18-band are less separated .
finally , the ratio of the peak intensity of the origin band of the @xmath13 transition to that of the @xmath18-band is higher . in the case of 9buyant
, one observes again a spectrum similar to that of ant .
it is affected by the differences already noted in the spectrum of the ethynyl - substituted molecule , only more strongly .
the redshifts are larger as the origin band of the @xmath13 transition is found at 3998 , that is 364 more to the red than in the spectrum of ant , and the @xmath18-band arises at approximately 2500 . a single maximum surmounts the @xmath18-band suggesting a larger broadening than in the spectrum of 9etyant .
the ratio of the peak intensities of the origin band of the @xmath13 transition relative to the @xmath18-band has further increased .
a new band observed next to the origin of the @xmath13 transition of 9etyant is also present in the spectrum of 9buyant , indicating a vibrational mode with a frequency of @xmath11216 @xmath8 .
other minor peaks are detected , notably at 2200 , 2746 , and 3236 , which belong neither to the @xmath13 transition nor to the @xmath18-band .
the vibrational mode with a frequency of @xmath11210 @xmath8 that gives rise to a band in the @xmath13 transition of 9etyant and 9buyant , but not in that of ant , is an in - plane ccc bending motion .
selection rules forbid the excitation of this mode in the @xmath13 transition of ant , which is a molecule with @xmath15 symmetry .
the rules change for 9etyant and 9buyant as the molecular symmetry becomes @xmath16 upon the addition of the side chains .
according to the ir spectra introduced in section [ res : ir ] , the mode has a frequency of 232 and @xmath11220 @xmath8 in the ground states of 9etyant and 9buyant , respectively , as measured on grains dispersed in csi pellets .
we found unscaled ( scaled ) theoretical values of 234 ( 228 ) and 218 ( 212 ) @xmath8 , respectively , with the calculations presented in sections [ ex : theory ] and [ res : theo ] .
the electronic spectra of phn and its derivatives are displayed in figure [ fig3 ] . in the spectrum of phn
, the @xmath13 transition gives extremely weak features , with the origin band arising at 3412 . at shorter wavelength
, one finds the strong bands of the @xmath19 transition , which has its origin at 2844 .
then one finds the features of the @xmath18-band , the strongest of which dominates the whole spectrum and peaks at 2428 .
it is attributed to the origin of this electronic transition .
taking into account the accuracy of the measurements , the positions we report are in excellent agreement with those of @xcite , which were 3411 , 2834 , and 2430 , respectively .
regarding 9etyphn , one finds a spectrum somewhat similar to that of phn , with the obvious difference that the @xmath18-band no longer dominates the spectrum .
the main peak in the region of the @xmath18-band , at 2476 , is redshifted by 48 in comparison with the @xmath18-band of phn . attributing this peak to the origin of the @xmath18-band is tentative as several maxima arise around this position .
they most likely belong to several electronic transitions as they can not be easily interpreted as a unique vibrational progression .
the dominating feature in the spectrum of 9etyphn corresponds to the origin band of the @xmath19 transition , which is found at 3010 , 166 more to the red than the same transition in phn .
the vibronic pattern is very similar to what is observed for phn , only the bands are slightly broadened .
at longer wavelength , a series of weak bands is attributed to the @xmath13 transition .
we assign the band with the longest wavelength , at 3504 , to the origin of the transition .
the vibrational pattern is complex , with a weak origin band , which is different from what is observed in phn .
the uv / vis spectrum of 9buyphn resembles that of 9etyphn . in the region
where the @xmath18-band is expected , however , one finds a very broad feature , for which it is not possible to determine the origin wavelength . due to this broadening , in comparison with the origin band of the @xmath19 transition , the @xmath18-band has an even lower strength than in the spectrum of 9etyphn .
the origin band of the @xmath19 transition dominates the spectrum at 3156 . like the other bands of the same transition , it is broader and redshifted by 146 compared to the corresponding features in the 9etyphn spectrum . as for the @xmath13 transition ,
its spectrum is very weak and consists of a very small number of features .
we assign the band at 3544 to the origin of the transition .
one notes that the bands of the @xmath13 transition are roughly twice as intense with respect to the @xmath19 bands in the spectra of phn and 9etyphn .
figure [ fig4 ] depicts the electronic spectra of pyr and its derivatives .
the spectrum of pyr is dominated by the features of three systems , which are the @xmath19 transition , the @xmath18-band , and the @xmath20-band .
they give rise to maxima at 3236 , 2652 , and 2328 , respectively , which mark the origins of the systems .
these values are in agreement with those obtained by @xcite from a less resolved spectrum .
we found the origin band of the very weak @xmath13 transition at 3674 .
the spectrum of 1etypyr is similar to that of pyr with broadened and redshifted transitions . in this instance , however , the spectrum is clearly dominated by the peak that marks the origin of the @xmath19 transition at 3438 .
the origins of the @xmath18- and @xmath20-bands are found at 2736 and 2384 , respectively , while the origin of the @xmath13 transition is at 3776 .
one also notes , in comparison with the spectrum of the parent pah , that the relative peak intensity of the origin band of the @xmath13 transition is higher .
finally , in the spectrum of 1buypyr , the origins of the @xmath19 transition , @xmath18-band , and @xmath20-band are found at 3578 , 2794 , and ( tentatively ) 2394 , respectively . in comparison with the spectrum of 1etypyr ,
the features are further broadened and redshifted .
the origin band of the @xmath13 transition is found at 3820 .
its relative peak intensity has become much stronger .
one can note in the spectrum of 1buypyr that the origin of the @xmath19 transition arises as a complex vibronic structure .
a matrix site effect or phonon side wings can be ruled out as they would affect the origin band of the @xmath13 transition in the same way .
we propose that the structure is caused by an interaction between the @xmath21 state and the vibrational manifold of the @xmath22 state .
studies in molecular beams have shown that this interaction takes place in pyr ( see @xcite and references therein ) .
the resulting vibronic structure in the @xmath19 transition is not noticeable in matrix - isolated spectra of pyr because the matrix - induced band broadening makes the bands overlap .
nothing forbids the interaction to occur in 1etypyr and 1buypyr . in the case of 1buypyr
, the difference between the energies of the @xmath21 and @xmath22 states is small .
it is approximately 1770 @xmath8 , as compared with 3684 and 2604 @xmath8 for pyr and 1etypyr , respectively .
consequently , the interaction is strong and it involves the vibrational manifold of the @xmath22 state where fundamental vibrational levels are found .
thus , in the vicinity of the origin of the @xmath19 transition , it gives rise to a strong and discrete vibronic structure .
the increased strength of the origin band of the @xmath13 transition is also an effect of this interaction . considering all uv / vis spectra of figures [ fig2][fig4 ] ,
the lower - energy transitions rather resemble those of the parent pah molecules in terms of vibrational pattern and bandwidth .
in contrast , the higher - energy transitions are strongly broadened , causing the lowering of their peak intensities . as this broadening increases with the length of the side group , we relate it to the number of its vibrational modes .
thus , the broadening would correspond to fast deexcitation through internal conversion and not to a matrix effect .
figure [ fig5 ] shows the ir absorption spectra of the polyynyl - substituted pahs .
the spectra have been first normalized with respect to the mass of sample powder in each pellet and the molecular weight of each species , and then with respect to the strongest band .
some of the bands characteristic of the pah family are not significantly affected by the substitution of a h atom with a much heavier polyynyl chain .
thus , the nonresolved group of weak bands at @xmath113050 @xmath8 ( 3.28 @xmath3 m ) corresponds to the aromatic ch stretching modes . in spite of the substitution , their frequencies are similar to those observed for the ch stretching modes of the parent pahs @xcite .
another feature characteristic of pahs is a strong band , the strongest or second strongest in our spectra , which is found between 700 and 900 @xmath8 ( 14.3 and 11.1 @xmath3 m ) .
it is due to the ch bending mode in which all h atoms move out - of - plane in phase .
while the position of the band varies depending on the pah structure , it is not much affected by the length of the side group .
in contrast , the substitution has a dramatic effect on other modes as it modifies the symmetry of the molecules .
for instance , ant has an out - of - plane aromatic ch bending mode with a frequency of @xmath11824 @xmath8 ( value of the theoretical harmonic frequency scaled by a factor of 0.97 ) , which is inactive with respect to ir absorption because it conserves the inversion center of the molecule .
after substitution of one of the h atoms on the central ring for a much heavier side group , this mode becomes active , giving rise to a band of medium intensity at 849 @xmath8 in the spectra of 9etyant and 9buyant .
other bands , which are indicated with arrows in figure [ fig5 ] , are specific to polyynyl side groups and correlate with the vibrational frequencies in the free molecules , ethyne and butadiyne @xcite .
accordingly , all spectra show more or less clearly a pair of bands of medium strength in the 625 @xmath8 ( 16 @xmath3 m ) region .
they correspond , in the order of increasing frequency , to the c@xmath23c@xmath0h out - of - plane and in - plane bending modes , which are shared by all polyynyl side groups .
all spectra also display a band of high intensity in the 32603300 @xmath8 ( 3.0303.068 @xmath3 m ) interval which is caused by the acetylenic ch stretching vibration . except for 9etyant and 9buyant ,
the band is stronger for the butadiynyl - substituted pahs than for the ethynyl - substituted ones .
as the band has a similar intensity in the spectra of 9buyant , 9buyphn , and 1buypyr , the normalization of the spectrum of 9etyant may be incorrect due to an inaccurate estimation of the mass of this sample .
on the other hand , ethynyl- and butadiynyl - substituted pahs can be distinguished by examining the bands arising from the acetylenic cc stretching modes .
the ethynyl group possesses a single c@xmath23c bond and , as a consequence , a single stretching mode , which causes a very weak band at @xmath112095 @xmath8 ( 4.77 @xmath3 m ) .
with two c@xmath23c bonds , the butadiynyl group has two stretching modes , antisymmetric and symmetric , that cause bands at @xmath112050 and @xmath112200 @xmath8 ( 4.88 and 4.55 @xmath3 m ) , respectively .
the band due to the antisymmetric stretching is very weak while the other has medium intensity .
another means to distinguish between ethynyl- and butadiynyl - substituted pahs is to look for the band arising from the in - plane waving mode of the butadiynyl chain near 200 @xmath8 ( 50 @xmath3 m ) .
although this band is weak , it can be clearly seen in our spectra . according to calculations ,
the band caused by the corresponding out - of - plane motion would be weaker , in various proportions , and would lie between 120 and 160 @xmath8 ( 62 and 83 @xmath3 m ) , at the limit of the range covered by our measurements .
therefore , it can not be found in our spectra . in sections [ res : uv ] and [ res :
ir ] , the measured spectra have been interpreted with the assistance of the results of theoretical calculations and with the knowledge obtained from the study of non - substituted pah molecules .
our interpretation of the laboratory ir spectra takes into account the theoretical ir spectra displayed in figure [ fig6 ] .
these spectra reproduce the measurements well enough to allow us to assign the experimental ir bands .
still , as they were obtained with dft - based calculations , they differ from the laboratory spectra in ways that were already reported for the case of other pahs ( see @xcite and references therein ) .
first , the theoretical frequencies are larger than the observed frequencies .
this is not unexpected since the theoretical values do not take anharmonicity into account .
it is common to correct the theoretical frequencies by applying a scaling factor , if necessary by taking the aromatic ch stretching frequencies into account separately .
detailed discussions of this issue , first addressed by @xcite , and dft spectra for many pahs can be found in @xcite and @xcite . in this study ,
a value of 0.959 is found by fitting the scaling factor as the set of data includes simultaneously frequencies of the six polyynyl - substituted pahs . when the bands are divided in two groups , those at frequencies below 2000 @xmath8 and those at frequencies above 2000 @xmath8 , scaling factors of 0.974 and 0.953 are obtained for the low- and high - frequency domains , respectively .
the theoretical and measured spectra also differ in their intensity patterns as the relative intensities of the aromatic ch stretching bands are much higher in the theoretical spectra than in the measured spectra ( see @xcite and references therein ) .
we find that this effect applies to the acetylenic cc and ch stretching bands as well .
another property of the molecules that has been determined is their permanent electric dipole moment in the electronic ground state .
table [ table1 ] gives the theoretical values for the parent pahs and their derivatives . among the parent pahs
, only phn possesses a permanent electric dipole moment as the molecule does not have an inversion center .
its value , however , is extremely small and has not been determined experimentally to our knowledge .
a significant dipole moment appears when an ethynyl side group is present . for comparison , the theoretical values we have obtained for 9etyant ( 0.5893 d ) , 9etyphn ( 0.7466 d ) , and 1etypyr ( 0.9251 d ) are close to the value measured for ethynylbenzene , that is 0.656 @xmath24 0.005 d @xcite .
our results show that the electric dipole moment is larger with a butadiynyl side group than with an ethynyl side group by a factor of two approximately .
electronic states and vertical transition energies are given in table [ table2 ] .
while theoretical vibrational frequencies are calculated with an accuracy of a few @xmath8 , the uncertainty on theoretical electronic transition energies is far greater , i.e. , of the order of 0.5 ev for the values computed with the td - dft approach @xcite . in the case of pah molecules ,
the theoretical determination of electronic transitions is usually challenging , because of the close proximity of the electronic states .
interactions , which are not taken into account in the usual models , may take place between them .
as noticed in the uv / vis spectra of the phenanthrenes and pyrenes , the energy spacing between the first two excited states decreases upon substitution with the ethynyl group and decreases further with the butadiynyl group .
the calculation of the electronic states and vertical transition energies gives qualitatively reasonable results for the anthracenes , even though quantitative differences exist between the results obtained with the zindo and td - dft approaches .
this is also true for phn and 9etyphn . for 9buyphn and the pyrenes
, the calculation fails to give the correct order for the two electronic states with the lowest energies ( with the exception of the zindo model applied to pyr ) .
we have examined the molecular orbitals calculated with the zindo model and also those obtained with the td - dft approach .
the electronic transitions we have observed are attributed to @xmath4-@xmath5 transitions . in every substituted species we have studied ,
the pah moiety and the side group contribute together to a number of @xmath4 orbitals , notably the highest occupied and lowest unoccupied molecular orbitals ( homo and lumo ) .
for example , the homo and the lumo of 9buyant are depicted in figure [ fig7 ] , along with the next orbitals , i.e. , the homo@xmath01 and the lumo@xmath251 .
while the latter two orbitals do not comprise contributions from the side group in the substituted anthracenes and phenanthrenes , they do in the substituted pyrenes .
because electronic transitions in substituted pahs involve molecular orbitals with contributions from both the pah moiety and the side group , more electrons are involved in the transitions of a substituted pah compared to the parent pah .
thus , larger oscillator strengths can be expected for the substituted molecules .
such a trend is observed in table [ table2 ] , except for the @xmath18-band of the anthracenes , for which the oscillator strength decreases as the length of the side group increases .
the growth of the oscillator strength remains however modest as it barely reaches an order of magnitude .
ideally , the interpretation of observational ( astrophysical ) spectra makes use of reference measurements obtained in the laboratory under the conditions of the astrophysical object . in general , it is not possible to exactly reproduce these conditions .
therefore , either they are mimicked or information is extracted from laboratory measurements obtained under various conditions .
although interstellar pahs have been revealed by their collective ir emission features , the aibs , the ir spectroscopy of pahs is most often performed in the laboratory in absorption , either at cryogenic temperature on molecules isolated in a rare gas matrix ( such as in mis ) or , more simply , at room temperature on grains and crystallites embedded in salt pellets .
as we have opted for the latter technique , the experimental conditions do not allow us to compare directly our measurements with the aibs , which are attributed to populations of free - flying pah molecules , nor do they allow us to compare them with the absorption bands caused by other free - flying molecules such as benzene , the only pah - like species identified in space so far @xcite .
indeed , due to the interaction of the pah molecules with their neighbors , the ir absorption bands measured in salt pellets are affected by shifts and variations of the relative intensities in comparison with the gas - phase absorption spectra . moreover , the absorption bands presently measured in csi pellets with a full width at half maximum ( fwhm ) of @xmath1156 @xmath8 are narrower than the emission bands , which are expected to show a fwhm of @xmath1130 @xmath8 @xcite .
finally , the band intensities observed in absorption and emission may differ as , in the latter case , they depend on the energy distribution among the vibrationally excited modes in addition to the transition moment . after they were found in ir sources , pah molecules were proposed as candidates for the carriers of the diffuse interstellar bands ( dibs ) , which are interstellar absorption features observed between 4000 and 18000 against background stars ( see @xcite and references therein ) .
the uv / vis absorption spectra of pahs can be measured more easily under conditions that mimic those of the ism , e.g. , by applying cavity ring - down laser absorption spectroscopy on jet - cooled molecules @xcite .
such spectra can be directly compared with the dibs .
the drawback of laser - based spectroscopy techniques is the usually narrow wavelength range they can cover . as a consequence ,
they represent the ultimate step in testing candidates for the carriers of the dibs . for a first approach to the spectral properties of a new species , it is preferable to apply mis since measurements can cover very large wavelength ranges in a single scan while using little sample .
the uv / vis absorption spectra we present are measured on pah molecules isolated in ne matrices at 67.5 k. compared to spectra measured under conditions relevant to the ism , the bands are expected to be redshifted by an amount approximately proportional to the variation of the polarizability of the molecule upon the corresponding transition ( see , e.g. , * ? ? ?
* ) . due to the low polarizability of ne ,
the matrix - induced shifts are small , of the order of 1@xmath26 or less .
therefore , matrix - isolated spectra represent valuable data because they reveal the general aspect of the spectra , which can be sufficient to demonstrate that a molecule can not be a dib carrier .
the present study was motivated by the conjecture that pahs carrying @xmath6-bonded chains may be present in ir emission regions .
it was put forward by @xcite following their observation of both aromatic structures and @xmath6-bonded chains in laboratory analogs of cosmic carbonaceous grains .
these analogs were produced by laser ablation of graphite and subsequent condensation of the ablated carbon atoms in the gas phase @xcite .
another study on carbonaceous grains synthesized in a similar fashion not only revealed aromatic structures and @xmath6-bonded chains , it also disclosed acetylenic ch bonds , though in a very low amount compared to the @xmath6-bonded chains @xcite .
the lack of acetylenic ch bonds in the former case and its very small amount in the latter indicated that the chains were essentially bridges between aromatic structures @xcite or fullerene fragments @xcite rather than polyynyl side groups .
following laser ablation of graphite , carbonaceous grains form in the gas phase at high temperatures , in the range 40006000 k @xcite .
they are considered as analogs of grains expected in the hot environment of carbon - rich stars or in the material ejected from supernovae @xcite .
we can conclude that the conditions in these environments are not favorable to the formation of polyynyl - substituted pahs .
however , polyynyl - substituted pahs would rather form at lower temperatures .
@xcite studied mixtures of molecules extracted from soots produced by ir laser - induced pyrolysis of ethylene , either pure or mixed with acetylene or benzene .
temperatures lower than 1800 k were found essential to the formation of pah - rich soots @xcite .
the absorption spectra of the mixtures extracted with chloroform ( chcl@xmath27 ) showed bands at 2104 and 3288 @xmath8 ( 4.75 and 3.04 @xmath3 m , respectively ) .
figure [ fig8 ] reproduces the spectrum of one of the extracted mixtures ( sample cp85a of * ? ? ?
* ) and compares it with the ir spectrum of 9etyphn from figure [ fig5 ] .
the comparison leads us to attribute the bands at 4.75 and 3.04 @xmath3 m to the stretching modes of c@xmath23c bonds and acetylenic ch bonds , respectively .
their positions and relative intensities indicate the presence of ethynyl groups carried by pahs .
since the formation of the molecules contained in the soot took place under conditions relevant to the environment of evolved stars @xcite , e.g. , asymptotic giant branch stars , it can be concluded that pahs carrying ethynyl groups can form in such regions .
conclusions can not be drawn concerning longer polyynyl side groups which were not detected in the soot extract .
the formation of ethynyl - substituted pahs in the environment of stars can be explained by the h - abstraction - c@xmath1h@xmath1-addition ( haca ) mechanism that contributes to the formation of larger pahs @xcite .
indeed , ethynyl - substituted pahs are intermediate products during the growth process @xcite .
it can take place in the circumstellar envelopes of carbon - rich stars , in regions where the temperatures are in the range between 900 and 1100 k @xcite .
interestingly , polyynyl - substituted pahs may form in the ism itself , at temperatures much lower than those required by the haca mechanism . in a combined theoretical and experimental study
, @xcite showed that benzene can form under conditions found in cold molecular clouds through the reaction of the ethynyl radical ( c@xmath1h ) with 1,3-butadiene ( ch@xmath1chchch@xmath1 ) .
starting with benzene , the so - called ethynyl addition mechanism ( eam ) introduced by @xcite would lead to the formation of larger aromatic species @xcite , first of all ethynylbenzene ( also known as phenylacetylene ) @xcite .
it would also lead to the formation of polysubstituted benzenes carrying ethynyl and butadiynyl groups , and , possibly , similarly polysubstituted pahs .
this mechanism is of great interest , for ethynyl is abundant in the ism @xcite and in diffuse clouds @xcite .
as the polyynyl radicals butadiynyl c@xmath2h @xcite , hexatriynyl c@xmath28h @xcite , and octatetraynyl c@xmath29h @xcite have been identified in space by microwave emission observations , various other products can be formed if they can also react with pah molecules at low temperature .
the relevance of the eam and the other reactions just mentioned to the chemistry of the ism needs to be examined especially with regard to timescales if more information on the corresponding reaction rates and reactant densities is available .
molecules formed in circumstellar envelopes or in diffuse molecular clouds are exposed to uv photon bombardment .
therefore the question of the photostability of the polyynyl - substituted pahs arises .
data regarding this particular class of molecules are not available . on the other hand , laboratory studies on the photostability of other pahs
have been reported @xcite .
@xcite found that methyl- and vinyl - substituted pahs ( h@xmath27c@xmath0 and h@xmath1c@xmath30ch@xmath0pahs ) have a lower photostability than their parent pahs .
this suggests that the photostability of a pah molecule is weakened upon substitution with a hydrocarbon group .
the extent of the effect , however , depends on the electronic structure of the substituent .
it was observed that the effect was less pronounced upon substitution with a vinyl group than with a methyl group , probably on account of the cc double bond of the vinyl group .
similarly , the experiments revealed that the photostability of diphenylacetylene , which consists of two phenyl moieties connected by a @xmath0c@xmath23c@xmath0 chain ( h@xmath31c@xmath32c@xmath23c@xmath0c@xmath28h@xmath31 ) , was higher than that of biphenyl , which consists of two phenyl moieties linked by a cc single bond ( h@xmath31c@xmath32c@xmath28h@xmath31 ) @xcite .
the higher photostability was attributed to the presence of the cc triple bond , which increased the overall conjugation of the @xmath4 orbitals @xcite .
thus , we can assume that polyynyl - substituted pahs , like other substituted pahs , have a lower photostability than their parent molecules . taking into account how the conjugation of the @xmath4 orbitals affects the photostability , we can also assume a higher photostability of polyynyl - substituted pahs compared to methyl - substituted pahs and , by generalizing , pahs carrying @xmath33-bonded aliphatic chains .
precisely these latter species may exist in ir sources despite the local flow of uv photons .
indeed , an ir emission band at 3.4 @xmath3 m has been attributed to the stretching vibration of an aliphatic c(@xmath33)h bond , which may be a component of side groups attached to pah molecules ( see * ? ? ?
thus , polyynyl - substituted pahs , which have a higher photostability , may withstand the conditions that prevail in ir sources , and those in the ism as well .
neither circumstellar nor interstellar polyynyl - substituted pahs have been observed yet .
the spectra of various ir sources do not show clear signs of their presence , neither as components of grains nor as free - flying molecules . according to the results of our ir measurements ( figure [ fig5 ] ) and theoretical calculations ( figure [ fig6 ] ) ,
the frequencies of the acetylenic cch bending and ch stretching modes do not vary greatly with the length of the polyynyl chain and the mass of the pah moiety . as a consequence , collective contributions to the ir emission bands
are expected from polyynyl - substituted pahs .
this is illustrated in figure [ fig9 ] , which represents , in one panel , the normalized sum of all the experimental ir spectra of figure [ fig5 ] and , in a second panel , the normalized sum of the theoretical ir spectra of figure [ fig6 ] .
for a better comparison with the experimental sum spectrum , the theoretical vibrational frequencies have been scaled with two factors depending on their value , as described in section [ res : theo ] . moreover , in order to lend the theoretical absorption spectrum the aspect of an emission spectrum , the bands have been given a gaussian profile with a fwhm of 30 @xmath8 @xcite .
first , the out - of - plane and in - plane acetylenic cch bending modes would ideally give rise to two close features in the 16 @xmath3 m region .
it is however more likely that a single broad feature would be produced , as seen in figure [ fig9 ] , due to the slight variations of the wavelength positions from species to species and , considering emission spectra , due to the expected fwhm of @xmath1130 @xmath8 of each band @xcite .
a search for this emission feature would be hindered by bands related to the deformation of aromatic rings which are present in the same wavelength region .
as a fact , observations revealed a well - defined emission band at 16.4 @xmath3 m @xcite and an underlying plateau between 15 and 20 @xmath3 m @xcite , and both were assigned to ccc bending modes of pah molecules @xcite .
still , a contribution of acetylenic cch bending bands to the plateau can not be ruled out .
second , the acetylenic ch stretching modes of polyynyl - substituted pahs would cause a band at @xmath113280 @xmath8 ( 3.05 @xmath3 m ) . to date , however , the only emission feature reported at a close wavelength is pfund s hydrogen recombination band pf@xmath34 at 3.04 @xmath3 m .
finally , as the number and energies of the acetylenic cc stretching modes vary depending on the polyynyl group , different pahs carrying a specific group would produce collective features characteristic of that group . in ethynyl - subtituted pahs , the unique acetylenic cc stretching mode gives rise to a weak band at @xmath114.8 @xmath3 m .
the antisymmetric and symmetric acetylenic cc stretching modes of butadiynyl - substituted pahs give two bands , a weak one at @xmath114.9 @xmath3 m and a medium one at @xmath114.5 @xmath3 m , respectively .
there is no report in the literature of emission bands corresponding to acetylenic cc stretching modes .
@xcite suggested the possibility for cc stretching modes of @xmath6-bonded chains , either acetylenic or cumulenic , to contribute to weak emission bands found at 5.25 and 5.7 @xmath3 m @xcite , although a detailed study by @xcite attributed these emissions to various combinations of vibrations in pure non - substituted pahs .
it should be mentioned that a search for ir emission bands of carbon chains , though not specifically as side groups , was undertaken by @xcite . through the examination of the 5 @xmath3 m region , where the bands of cc stretching modes are expected
, they concluded that carbon chains in general contribute at best extremely little to the ir emission features . while the features of carbon chains are not found in ir emission spectra , absorption bands observed toward ir sources
have been attributed to acetylene ( c@xmath1h@xmath1 , at 13.7 @xmath3 m , * ? ? ?
* ) , diacetylene ( c@xmath2h@xmath1 , at 8.0 and 15.9 @xmath3 m , * ? ?
* ; * ? ? ?
* ) , and triacetylene ( c@xmath28h@xmath1 , at 8.1 and 16.1 @xmath3 m , * ? ? ?
* ; * ? ? ?
another band of acetylene was observed at 2.44 @xmath3 m @xcite .
an absorption feature measured at 3.0 @xmath3 m , very close to the position of the acetylenic ch stretching band of polyynyl - substituted pahs , was attributed to overlapping bands of hydrogen cyanide ( hcn ) and acetylene @xcite .
thus , ir observations do not disclose bands that can be assigned with certainty to polyynyl - substituted pahs , neither in the regions that exhibit aromatic ir emission features , nor toward ir sources that reveal absorption bands of carbon chains .
still , the presence of polyynyl - substituted pahs in circumstellar environments or in the ism , i.e. , in regions where conditions may allow their formation , even in low abundance , can not be ruled out
. in general , non - substituted pahs possess at best a very weak permanent electric dipole moment , supporting the theoretical values in the first line of table [ table1 ] .
for this reason , laboratory measurements in the radiofrequency ( rf ) domain are not feasible , as reported by @xcite who attempted to measure the spectrum of phn .
thus , a search for their rf lines is impracticable . on the contrary ,
our calculations indicate that polyynyl - substituted pahs show a significant permanent electric dipole moment , which depends strongly on the length of the side chain according to the values in table [ table1 ] .
because these molecules possess a large moment of inertia , one must expect large partition functions and , as a consequence , weak lines that make detection difficult @xcite .
as far as the polyynyl - substituted pahs are concerned , laboratory measurements in the rf domain are missing , preventing a successful search for specific molecules with radiotelescopes .
such measurements can be envisaged , as demonstrated by the study of ethynylbenzene ( phenylacetylene ) by @xcite .
assuming polyynyl - substituted pahs are present in the diffuse ism , there is a possibility to detect their absorption bands at uv / vis wavelengths .
bands of interstellar non - substituted pahs were not discovered by searches in the uv / vis region , leading to conclude that such bands were either absent or too weak for detection at the given signal - to - noise ratio of the observations @xcite .
this conclusion is valid if one assumes that the bands carried by all pahs present are separate .
it is not valid , however , if one presumes that the bands overlap . in that case
, they could give rise to a smooth contribution to the interstellar extinction curve , including its uv bump at 2175 and its far - uv rise @xcite .
the strongest absorption bands of some of the polyynyl - substituted pahs presented here ( 9buyphn , 1etypyr , and 1buypyr ) are in the 30503850 wavelength range within which a search for dibs was fruitless @xcite .
thus , either their amount in the ism is lower than the limit given by the signal - to - noise ratio of the observations or their features contribute to the interstellar extinction curve . even though the spectral features of the species we have studied do not coincide with observed emission or absorption bands , neither at ir nor at uv / vis wavelengths
, they may point to a characteristic property of the elusive carriers of the dibs , for which pahs are candidate @xcite .
non - substituted pahs that show absorption bands at visible wavelengths , however , usually present features of significant intensity in the uv region as well .
this is not consistent with the fact that dib carriers apparently exhibit a single strong band which is found at a visible wavelength .
moreover , the pah family , even limited to the non - substituted pahs , comprises , so to say , an infinite number of species .
in contrast , as noted by @xcite , the number of dibs of moderate to strong intensity is about 50 . because these dibs do not show any evidence of correlated variations when observed toward various lines of sight , the existence of a specific carrier for each dib has been inferred . assuming that the carriers are pah molecules , they must stand out among the members of their family . in that case , criteria such as size distribution and charge state may not be sufficient to limit the number of potential carriers .
another selection mechanism , based on the formation and destruction processes of the carriers , was discussed by @xcite . led by the present study , we may propose that dib carriers are characterized by specific combinations of aromatic cores with side groups .
these pah molecules carrying side groups would be intermediate products in the destruction or formation routes of interstellar pahs .
the number of dib carriers would be limited due to the selection mechanism imposed by their formation and destruction processes and by the specificity of the combinations of aromatic cores with side groups .
due to the side groups , electronic transitions at uv wavelengths would give rise to broadened features that would blend in the interstellar extinction curve , while transitions in the visible region would be associated with well - defined bands .
their oscillator strength would be greater than those of the parent pahs since more electrons are involved .
it could also be that another ingredient has to be considered in order to obtain a molecule that can carry a strong dib since side groups may not be sufficient to obtain transitions in the vis to near ir spectral regions .
the dibs correlate neither with the aibs , attributed to pah molecules , nor with the uv bump at 2175 @xcite , to which interstellar pahs may contribute ( see @xcite and references therein ) .
this could be construed as an argument against pahs being carriers of the dibs . as an answer to such an interpretation
, we may consider that two populations of pah molecules exist in space .
the first population would be produced in carbon - rich circumstellar envelopes through the haca mechanism .
this population would essentially consist of a great diversity of non - substituted pahs
. it would be denser close to the stars , giving the aibs , and very diffuse in the ism , causing a smooth contribution to the uv extinction curve , the population of each species being too low for individual detection .
the second population would be produced in molecular clouds by the eam phenomenon and would consist of pahs with side groups , some of which being dib carriers .
we have examined the relevance of polyynyl - substituted pah molecules to astrophysics .
it appears that ethynyl - substituted pahs may form in circumstellar envelopes through the haca mechanism .
polysubstituted pahs carrying ethynyl and butadiynyl groups may form in molecular clouds , even cold ones , through neutral - neutral reactions of pahs with ethynyl radicals ( eam ) . as polyynyl radicals have been identified in space , it would be interesting to know whether they can react with pahs as easily as ethynyl , which would result in the formation of polyynyl - substituted pahs .
one can expect that the photostability of polyynyl - substituted pahs is lower than that of the non - substituted pahs .
due to conjugation of the @xmath4 orbitals over both the pah and polyynyl moieties , the molecules most likely withstand the uv radiation field .
moreover , the dramatic broadening that we have observed for the absorption bands in the uv region suggests a very short lifetime of the electronically excited states , which is attributed to fast internal conversion .
therefore , it is reasonable to assume the existence of stationary populations of polyynyl - substituted pahs in regions where pah growth takes place . like the populations of non - substituted pahs , populations of polyynyl - substituted pahs would cause collective features in the ir wavelength region , as illustrated by the spectra we have measured and calculated . since the observation of these features has not been reported yet , it can be concluded that the populations are below the detection level at ir wavelengths .
it is the case in ir sources , where ir emission bands of non - substituted pahs are observed , and nearby these sources , toward which ir absorption bands of hydrocarbons are observed . in the uv / vis region
, the molecules would not give rise to collective features .
nevertheless , as the absorption cross section is in general larger for electronic transitions than for vibrational transitions , searches for individual species can be proposed .
although the present spectra were measured in rare gas matrices and hence can not be directly compared with interstellar spectra , all the bands we have observed would lie at uv wavelengths under the conditions of the ism . since surveys of the uv region have not revealed any band that could be attributed to a pah molecule , we conclude that the amount of each of the species we have studied is too small for detection with the current limitations .
our most interesting observation concerns the intensity profile of the absorption bands in the electronic spectra .
the bands in the upper uv range are dramatically broadened while those in the lower range , arising from the first two electronic transitions , are barely affected .
we interprete this broadening of the bands with shorter wavelengths as the effect of very fast deexcitation through internal conversion , in which the polyynyl group plays a crucial role .
such broad bands would be difficult to identify in interstellar spectra and would only contribute to the extinction curve of the ism . on the other hand , the strongest band in the spectra of the first two electronic transitions may stand out .
thus , the electronic spectra of polyynyl - substituted pahs present similarities with the spectra of the dib carriers .
although the molecules we have studied are not dib carriers , our study may represent a new step toward their identification by introducing the cause for the lack of dibs in the uv region and the limited number of strong dibs .
this work was carried out within a cooperation between the max - planck - institut fr astronomie , the friedrich - schiller - universitt jena , and the technische universitt dresden .
the support of the deutsche forschungsgemeinschaft ( contract no .
hu 474/24 ) is gratefully acknowledged .
we are thankful to dr .
h. mutschke for giving us access to the ftir spectrometer and to g. born for preparing the csi pellets and performing the hplc measurements .
our thanks also go to the anonymous referee for suggestions that helped to improve the quality of this article .
allamandola , l. j. , tielens , a. g. g. m. , & barker , j. r. 1985 , , 290 , l25 allamandola , l. j. , bregman , j. d. , sandford , s. a. , et al . 1989 , , 345 , l59 allamandola , l. j. , hudgins , d. m. , bauschlicher , c. w. , & langhoff , s. r. 1999 , , 352 , 659 allamandola , l. j. , hudgins , d. m. , & sandford , s. a. 1999 , , 511 , l115 bauschlicher , c. w. , boersma , c. , ricca , a. , et al . 2010 , , 189 , 341 boersma , c. , mattioda , a. l. , bauschlicher , c. w. , et al . 2009 , , 690 , 1208 cernicharo , j. & gulin , m. 1996 , , 309 , l27 cernicharo , j. , heras , a. m. , pardo , j. r. , et al . 2001 , , 546 , l127 cernicharo , j. , heras , a. m. , tielens , a. g. g. m. , et al . 2001 , , 546 , l123 chiar , j. e. , pendleton , y. j. , geballe , t. r. , & tielens , a. g. g. m. 1998 , , 507 , 281 clar , e. 1950 , spectrochim .
acta , 4 , 116 clayton , g. c. , gordon , k. d. , salama , f. , et al .
2003 , , 592 , 947 cox , a. p. , ewart , i. c. , & stigliani , w. m. 1975 , j. chem .
faraday trans . 2 , 71 , 504 crawford , m. k. , tielens , a. g. g. m. , & allamandola , l. j. 1985 , , 293 , l45 dierksen , m. & grimme , s. 2004 , j. phys .
a , 108 , 10225 duley , w. w. 2006 , faraday discuss . , 133 , 415 duley , w. w. & hu , a. 2009 , , 698 , 808 frenklach , m. , clary , d. w. , gardiner , w. c. , & stein , s. e. 1985 , symposium ( international ) on combustion , 20 , 887 frenklach , m. & feigelson , e. d. 1989 , , 341 , 372 frisch , m. j. , trucks , g. w. , schlegel , h. b. , et al . 2009 , gaussian 09 , revision a.02 ( wallingford ct : gaussian , inc . ) geballe , t. r. , najarro , f. , figer , d. f. , schlegelmilch , b. w. , & de la fuente , d. 2011 , nature , 479 , 200 goto , m. , gaessler , w. , hayano , y. , et al .
2003 , , 589 , 419 goulay , f. & leone , s. r. 2006 , j. phys .
chem . a , 110 , 1875 gredel , r. , carpentier , y. , rouill , g. , et al .
2011 , , 530 , a26 gulin , m. , cernicharo , j. , kahane , c. , gomez - gonzalez , j. , & walmsley , c. m. 1987 , , 175 , l5 gulin , m. , green , s. , & thaddeus , p. 1978 , , 224 , l27 hudgins , d. m. & sandford , s. a. 1998 , j. phys .
a , 102 , 329 jger , c. , mutschke , h. , henning , th . , & huisken , f. 2008 , , 689 , 249 jger , c. , huisken , f. , mutschke , h. , llamas - jansa , i. , & henning , th .
2009 , , 696 , 706 jger , c. , huisken , f. , mutschke , h. , et al .
2007 , carbon , 45 , 2981 jochims , h. w. , baumgrtel , h. , & leach , s. 1999 , , 512 , 500 jochims , h. w. , rhl , e. , baumgrtel , h. , tobita , s. , & leach , s. 1997 , int .
j. mass spectrom .
ion processes , 167/168 , 35 jones , b. , zhang , f. , maksyutenko , p. , mebel , a. m. , & kaiser , r. i. 2010 , j. phys .
a. , 114 , 5256 jones , b. m. , zhang , f. , kaiser , r. i. , et al .
2011 , proc .
, 108 , 452 karcher , w. , fordham , r. j. , dubois , j. j. , glaude , p. g. j. m. , & ligthart , j. a. m. 1985 , spectral atlas of polycyclic aromatic compounds ( dordrecht : d. reidel publishing company ) keady , j. j. & hinkle , k. h. 1988 , , 331 , 539 lacy , j. h. , evans , n. j. , achtermann , j. m. , et al .
1989 , , 342 , l43 langhoff , s. r. 1996 , j. phys .
chem . , 100 , 2819 lger , a. & dhendecourt , l. 1985 , , 146 , 81 leger , a. & puget , j. l. 1984 , , 137 , l5 lucas , r. & liszt , h. s. 2000 , , 358 , 1069 malloci , g. , joblin , c. , & mulas , g. 2007 , chem .
, 332 , 353 marsh , n. d. & wornat , m. j. 2000 , poly .
, 19 , 263 mebel , a. m. , kislov , v. v. , & kaiser , r. i. 2008 , j. am .
soc . , 130 , 13618 motylewski , t. , linnartz , h. , vaizert , o. , et al .
2000 , , 531 , 312 moutou , c. , verstraete , l. , lger , a. , sellgren , k. , & schmidt , w. 2000 , , 354 , l17 negishi , e. & anastasia , l. 2003 , chem .
, 103 , 1979 rouill , g. , krasnokutski , s. , huisken , f. , et al .
2004 , j. chem .
phys . , 120 , 6028 rouill , g. , steglich , m. , jger , c. , et al .
2011 , chemphyschem , 12 , 2131 salama , f. 2008 , in iau symp .
251 , organic matter in space , ed .
s. kwok & s. sandford ( cambridge : cambridge university press ) , 357 salama , f. & allamandola , l. j. 1995 , adv .
space res .
, 15 , 413 salama , f. , bakes , e. l. o. , allamandola , l. j. , & tielens , a. g. g. m. 1996 , , 458 , 621 salama , f. , galazutdinov , g. a. , kreowski , j. , et al .
2011 , , 728 , 154 salama , f. , joblin , c. , & allamandola , l. j. 1994 , j. chem .
phys . , 101 , 10252 shimanouchi , t. 1972 , nat .
stand . ref .
data ser . , nat
( u.s . ) , 39 , 1 steglich , m. , bouwman , j. , huisken , f. , & henning , th .
2011 , , 742 , 2 steglich , m. , carpentier , y. , jger , c. , et al .
2012 , , in press ( arxiv:1203.2026 ) steglich , m. , huisken , f. , dahl , j. e. , carlson , r. m. k. , & henning , th .
2011 , , 729 , 91 steglich , m. , jger , c. , rouill , g. , et al .
2010 , , 712 , l16 suzuki , h. , ohishi , m. , kaifu , n. , ishikawa , s. , & kasuga , t. 1986 , pasj , 38 , 911 thorwirth , s. , theul , p. , gottlieb , c. a. , mccarthy , m. c. , & thaddeus , p. 2007 , , 662 , 1309 tielens , a. g. g. m. 2008 , ara&a , 46 , 289 tucker , k. d. , kutner , m. l. , & thaddeus , p. 1974
, , 193 , l115 van der zwet , g. p. & allamandola , l. j. 1985 , , 146 , 76 van kerckhoven , c. , hony , s. , peeters , e. , et al .
2000 , , 357 , 1013 visser , t. , sarobe , m. , jenneskens , l. w. , & wesseling , j. w. 1998 , fuel , 77 , 913 wang , h. & frenklach , m. 1994 , j. phys .
, 98 , 11465 xiang , f. y. , li , a. , & zhong , j. x. 2011 , , 733 , 91 llllllll molecule & transition & & & + & & @xmath36/ & strength & @xmath36/ & strength & @xmath36/ & assignment + ant & 1 & 3691 & 0.2222 & 3855 & 0.0525 & 3634 & @xmath13 + & 2 & 2631 & 2.4307 & 2372 & 1.9722 & 2374 & @xmath18-band + 9etyant & 1 & 3984 & 0.3314 & 4133 & 0.1096 & 3868 & @xmath13 + & 2 & 2692 & 2.1803 & 2466 & 1.5623 & 2440 & @xmath18-band + 9buyant & 1 & 4212 & 0.4358 & 4329 & 0.2157 & 3998 & @xmath13
+ & 2 & 2718 & 2.0456 & 2547 & 1.1209 & 2490 & @xmath18-band + phn & 1 & 3096 & 0.1424 & 2937 & 0.0580 & 2844 & @xmath19 + & 2 & 3335 & 0.0011 & 3113 & 0.0015 & 3412 & @xmath13 + 9etyphn & 1 & 3282 & 0.3114 & 3139 & 0.1878 & 3010 & @xmath19 + & 2 & 3400 & 0.0029 & 3238 & 0.0004 & 3504 & @xmath13 + 9buyphn & 1 & 3621 & 0.2794 & 3371 & 0.3635 & 3156 & @xmath19 + & 2 & 3427 & 0.0028 & 3327 & 0.0470 & 3544 & @xmath13 + pyr & 1 & 3526 & 0.5175 & 3366 & 0.2518 & 3236 & @xmath19 + & 2 & 3629 & 0.0027 & 3300 & 0.0003 & 3674 & @xmath13 + 1etypyr & 1 & 3747 & 0.6753 & 3621 & 0.4012 & 3438 & @xmath19 + & 2 & 3707 & 0.0107 & 3417 & 0.0012 & 3776 & @xmath13 + 1buypyr & 1 & 3960 & 0.7316 & 3851 & 0.6080 & 3578 & @xmath19 + & 2 & 3736 & 0.0063 & 3495 & 0.0047 & 3820 & @xmath13 + | we report on the absorption spectra of the polycyclic aromatic hydrocarbon ( pah ) molecules anthracene , phenanthrene , and pyrene carrying either an ethynyl ( @xmath0c@xmath1h ) or a butadiynyl ( @xmath0c@xmath2h ) group .
measurements were carried out in the mid infrared at room temperature on grains embedded in csi pellets and in the near ultraviolet at cryogenic temperature on molecules isolated in ne matrices .
the infrared measurements show that interstellar populations of polyynyl - substituted pahs would give rise to collective features in the same way non - substituted pahs give rise to the aromatic infrared bands .
the main features characteristic of the substituted molecules correspond to the acetylenic ch stretching mode near 3.05 @xmath3 m and to the almost isoenergetic acetylenic cch in- and out - of - plane bending modes near 15.9 @xmath3 m .
sub - populations defined by the length of the polyynyl side group cause collective features which correspond to the various acetylenic cc stretching modes .
the ultraviolet spectra reveal that the addition of an ethynyl group to a non - substituted pah molecule results in all its electronic transitions being redshifted .
due to fast internal energy conversion , the bands at shorter wavelengths are significantly broadened . those at longer wavelengths are only barely affected in this respect .
as a consequence , their relative peak absorption increases .
the substitution with the longer butadiynyl chain causes the same effects with a larger magnitude , resulting in the spectra to show a prominent if not dominating @xmath4-@xmath5 transition at long wavelength . after discussing the relevance of polyynyl - substituted pahs to astrophysics ,
we conclude that this class of highly conjugated , unsaturated molecules are valid candidates for the carriers of the diffuse interstellar bands . |
since the discovery of hot jupiters @xcite and their detection in transit @xcite , the study of atmospheric dynamics on these objects has attracted substantial interest
. accordingly , over the last decade a vast hierarchy of global circulation models ( gcms " , largely adopted from numerical codes aimed at simulating the earth s climate ) has been established ( see @xcite and the references therein ) .
although distinct models exhibit subtle differences ( see @xcite ) , hydrodynamical simulations show broad agreement on the qualitative features of the flow .
specifically , the overwhelming majority of 3d gcms obtain super - rotating eastward zonal jets with characteristic maximal speeds of order @xmath0 few km / s at photospheric ( and somewhat higher ) pressures .
because atmospheric temperatures on hot jupiters can reach values high enough to ionize alkali metals such as k and na @xcite , it has been realized that magnetohydrodynamic ( mhd ) effects may carry substantial repercussions for global circulation . to simulate the coupling between the atmospheric flow and the background field ,
a number of authors have augmented their gcms with a rayleigh drag aimed at mimicking the coupling between the large - scale circulation and the background field @xcite .
although a natural first step , this approach is not adequate generally since magnetic torques inherently depend on the geometry of the field and exhibit phase signatures of damping that differ from those of a rayleigh drag @xcite .
accordingly , the first self - consistent boussinesq magnetohydrodynamic simulations were performed by @xcite and recently , improved anaelastic mhd calculations were presented by @xcite . by and large ,
this aggregate of magnetic simulations suggests that zonal jets are damped by lorentz forces , in agreement with analytical studies @xcite .
a simplifying assumption that is consistently employed when studying mhd effects in hot atmospheres is the alignment between the magnetic axis and the spin axis of the planet .
although this is synonymous to the case of saturn @xcite , other giant planets in the solar system posses oblique magnetic fields , and it is reasonable to expect that misalignments between the magnetic and rotational axes will be common to the population of close - in planets .
correspondingly , in this paper we address the effects of a tilted dipole magnetic field on global circulation in a partially ionized atmosphere .
the paper is organized as follows . in section 2 , we describe the considered physical setup . in section 3 , we calculate the currents as well as the associated lorentz forces , induced by the interactions between the background field and the imposed flow . in section 4 , we examine the perturbations to the zonal jets that arise from the lorentz forces .
we conclude and discuss our results in section 5 . throughout the paper ,
exclusively analytical perturbative techniques are employed .
there exists a clear trade - off between model simplicity and realism in the study of dynamical meteorology . accordingly , while computationally intensive 3d gcms have been utilized for constructing the most detailed representations of the atmospheric states of hot jupiters @xcite , simplified 2d simulations @xcite have been used to elucidate and analyze specific phenomena .
as this study finds itself in the latter category , we shall also restrict our treatment to a thin shell ( envisioned to reside at the photospheric pressure level ) of thickness @xmath1 where @xmath2 is boltzmann s constant , @xmath3 is temperature , @xmath4 is mean molecular weight , @xmath5 is surface gravity , and @xmath6 is the planetary radius . ) is taken to coincide with the @xmath7-axis , while the background magnetic field ( shown as a blue vector field ) is tilted towards the @xmath8-axis by a small angle , @xmath9 . as a consequence of the atmosphere s finite conductivity , zonal jets ( shown as green hoops ) are envisioned to be perturbed away from their default azimuthally symmetric state . ] for tractability , we assume that the inclination of the background magnetic field relative to the spin - axis , @xmath9 , is small .
thus , we introduce : @xmath10 as an inherent small parameter of the problem .
the background dipole field then takes the form @xcite : @xmath11,\end{aligned}\ ] ] where @xmath12 is the surface field strength , @xmath13 is the associated legendre polynomial and the @xmath8-axis of the coordinate system is chosen to correspond to the field s line of nodes .
meanwhile , following @xcite we adopt a kinematic prescription for the zonal jet : @xmath14 where @xmath15 is an unknown perturbation to the background flow .
we assume that radial flow is prohibited in the region of interest , and that @xmath15 is independent of @xmath16 ( i.e. @xmath15 represents an azimuthally asymmetric 2d flow ) .
additionally , we assume that the magnetic diffusivity , @xmath17 , is constant at the pressure - level of interest .
a cartoon depicting the geometrical setup of the problem is shown in figure ( [ toy ] ) .
the steady state induction equation , relevant to partially ionized hot jupiter atmospheres , is written as follows @xcite : @xmath18 note that in the above expression , the hall and ambipolar diffusion terms have been neglected , as justified by @xcite .
the total magnetic field is composed of the sum of the background and induced fields : @xmath19 .
however , since @xmath20 is defined as a gradient of a scalar function , it is curl - free by construction meaning @xmath21 on the other hand , for clarity we may take the @xmath22 term to be dominated by @xmath20 @xcite . conventionally , this assumption implies a magnetic reynolds number @xmath23 somewhat smaller than unity . as discussed by @xcite ,
magnetic reynolds numbers of order unity or less are expected for a sizable fraction of hot jupiters , especially under the assumptions of strong ( e.g. @xmath24 g ) magnetic fields ( see also @xcite ) .
however , it should also be noted that the aforementioned assumption may be justified even in the regime where @xmath25 is substantial , provided a circulation geometry that is not complex enough to directly influence @xmath22 ( a simple example is an unperturbed kinematic zonal flow , that induces a purely toroidal field , which in turn drops out of the induction term in equation ( [ induction ] ) - see e.g. @xcite ) . consequently .
to first order in @xmath26 , equation ( [ induction ] ) reads : @xmath27.\end{aligned}\ ] ] in the above expression , @xmath28 represents the field induced by the interaction of the background flow ( @xmath29 ) with the the aligned component of the field ( @xmath30 ) , @xmath31 depicts the field induced by the interaction of the background flow ( @xmath29 ) with the non - axisymmetric component of the field ( @xmath32 ) , and @xmath33 designates the field arising from the coupling between the ( unknown ) perturbed component of the circulation ( @xmath15 ) and the axisymmetric field ( @xmath30 ) . as a first step towards solving equation ( [ inductionfo ] ) , we shall assume that magnetic induction associated with the unknown velocity field @xmath15 gives rise to a negligible lorentz force and need not be addressed .
as shown in the appendix , this simplification is equivalent to assuming that the elsasser number , defined as : @xmath34 where @xmath35 is the atmospheric density and @xmath36 is the permeability of free space , is much less than unity .
the first and second lines of equation ( [ inductionfo ] ) can then be solved sequentially and independently .
the leading order solution to equation ( [ inductionfo ] ) can be trivially obtained by separation of variables and has been discussed elsewhere ( see e.g. @xcite ) .
we shall briefly rehash it here for completeness .
the expression @xmath37 possesses azimuthal symmetry ( i.e. only has a @xmath38 component and carries no @xmath39dependence ) .
moreover , its angular part is an eigenfunction of the @xmath40 operator .
correspondingly , we immediately find that the solution for @xmath28 is satisfied by @xmath41 thus , the first line of equation ( [ inductionfo ] ) becomes a second - order ode for @xmath42 : @xmath43 following @xcite and @xcite , we adopt zero radial current boundary conditions at the outer @xmath44 and inner @xmath45 edges of the shell : @xmath46 .
the solution for @xmath42 then takes the form : @xmath47 . in this figure ,
the radial component of the current is neglected as it is much smaller than its azimuthal and meridional counterparts .
note that the vector field s symmetry axis corresponds to the line of nodes of the tilted background magnetic field . ] written out explicitly , the lorentz force reads : @xmath48 + in keeping with the notation introduced above , the ubiquitous @xmath49 superscripts in equation ( [ forceref1 ] ) imply that both terms arise from currents associated with the coupling between the axisymmetric components of the flow and the field .
meanwhile , the @xmath50 and @xmath26 subscripts signify the interactions between these currents and the axisymmetric and non - axisymmetric components of the field respectively .
because our treatment is quasi-2d , we are primarily interested in the vertically averaged lorentz force : @xmath51 moreover , we shall take advantage of the smallness of the aspect ratio @xmath52 and expand @xmath53 as a power series . to leading order in @xmath54 ,
the relevant expressions take the form : @xmath55 note that the @xmath56 term is omitted , since it governs the conventional magnetic damping effect ( see e.g. @xcite ) and is of no interest to us , as it is envisioned to only modulate the magnitude of @xmath57 ( which we take as an adjustable parameter ) .
the term on the second line of equation ( [ fl ] ) provides a small ( of order @xmath58 ) correction to the aforementioned effect , and is also of little importance . on the contrary , the term on the first line of equation ( [ fl ] )
may give rise to a qualitative alteration of the zonal jet .
the solution to the second line of equation ( [ inductionfo ] ) is not as trivial as that considered above because the associated induction term lacks azimuthal symmetry .
as such , it is difficult to obtain an analytical solution simply by inspection and we shall address this equation in a somewhat alternative manner .
that is , rather than solving for the induced field @xmath31 , we shall directly seek a solution for the induced current @xmath59 . specifically , we uncurl the second line of equation ( [ inductionfo ] ) to recover ohm s law : @xmath60 where @xmath61 is the electric potential . as before ,
the induced radial current is constrained to be null at @xmath62 and @xmath63 .
consequently , we choose the angular part of @xmath61 to correspond to that of @xmath64 ( since it is not affected by differentiation with respect to @xmath16 ) and retain an unspecified radial dependence : @xmath65 taking advantage of the divergence - free nature of the induced current @xmath66 we obtain a second - order ode for @xmath67 : @xmath68 with the aforementioned boundary conditions ( i.e. @xmath69 & @xmath70 ) , we obtain : @xmath71 because the thickness of the atmospheric shell in this calculation is taken to be small compared to the planetary radius , the current density @xmath59 is primarily azimuthal and meridional . neglecting the small radial component ,
the 2d geometry of @xmath59 is shown in figure ( [ current ] ) .
the corresponding lorentz force then takes the form : @xmath72 concurrently , to leading order in @xmath54 , its vertically averaged counterpart reads : @xmath73 cumulatively , there are two constituents of the lorentz force , given by equations ( [ fl ] ) and ( [ flsecond ] ) .
their ratio is of order : @xmath74 consequently , in subsequent analysis , @xmath75 can be neglected in favor of @xmath76 .
with the perturbing forces specified , we now seek a solution for the non - axisymmetric perturbation to the zonal flow . within the context of the thin - shell geometry considered here ,
we can make use of the boussinesq approximation , and neglect density variations in the domain of interest . under this assumption
, the continuity equation reduces to the incompressibility condition : @xmath77 owing to our 2d treatment of the global circulation , equation ( [ incompress ] ) can be automatically satisfied by specifying a stream - function , @xmath78 , from which @xmath15 arises @xcite : @xmath79 in order to obtain the correction to the velocity field , @xmath15 , we must ( perturbatively ) solve the imvicid navier - stokes equation @xcite : @xmath80 here , @xmath81 is the material derivative , @xmath82 is the spin vector and @xmath83 is pressure .
firstly , as in the previous section , we assume steady state , meaning @xmath84 . secondly , choosing a characteristic velocity scale of order @xmath85 km / s , a characteristic length scale of order @xmath86 and a spin - period of order @xmath87 day ( which translates to a coriolis parameter of @xmath88 ) , we obtain a rossby number of order @xmath89 meaning that advective accelerations can be neglected in favor of the coriolis effect @xcite .
thus , the entire left hand side of equation ( [ navierstokes ] ) is set to zero .
finally , we assume that the atmosphere is in hydrostatic equilibrium and that the background zonal flow is geostrophic @xcite . accordingly , to zeroth order in @xmath26 ( i.e. neglecting the dipole tilt ) , equation ( [ navierstokes ] ) reads : @xmath90 it remains to solve the residual terms in the navier - stokes equation .
given the supposed smallness of @xmath26 , it is likely that thermal advection associated with @xmath15 does not alter the global pressure profile significantly ( although this assertion depends on the thermodynamic properties of the atmosphere ) .
consequently , an approximate force - balance between coriolis and lorentz forces can be envisaged at first order in @xmath26 . in the specified regime , to leading order in @xmath54 , the individual components of of the navier - stokes equation take the form : @xmath91 it follows from direct integration of the above expressions that they are satisfied by the stream - function : align = v_max r ( ) ( ) .
evidently , the magnitude of the perturbation is controlled by the product of the field obliquity and the elsasser number .
figure ( [ flow ] ) shows the geometry of the perturbed zonal jet with @xmath92 ( left panel ) and @xmath93 ( right panel ) .
in this paper , we have considered the global circulation of a weakly ionized hot jupiter atmosphere , subject to interaction with an inclined magnetic field . using a perturbative approach
, we identified the form of the induced current primarily responsible for the distortion of zonal winds .
subsequently , we obtained a closed - form expression for the stream - function of the non - axisymmetric flow .
the above discussion demonstrates that on hot jupiters that possess sizable , significantly tilted fields , atmospheric jets are perturbed away from purely zonal states , and the fractional deviation from axisymmetric circulation scales as @xmath94 .
naturally , in order to obtain an analytic solution , it was necessary to make a series of simplifying assumptions .
specifically , our treatment relies on the dominance of the dipole component of the field over higher - order harmonics , the smallness of the field s obliquity @xmath26 , a small elsasser number @xmath95 , a rotationally dominated force balance , confinement of the current to a thin - shell geometry , and the validity of a kinematic ( as opposed to dynamic ) treatment of the large - scale circulation .
the last of these assumptions is arguably most significant , since in the contrary case of @xmath96 , the electromagnetic skin - depth effect inherent to the atmospheric differential rotation may act to attenuate the oblique component of the field @xcite .
cumulatively , this means that the derived solution is not appropriate for all choices of system parameters , and more exotic behavior is indeed possible beyond the scope of the specified limitations .
still , the derived solution may ( at least qualitatively ) be informative beyond its formal range of applicability term in the induction equation will alter the behavior of the perturbed flow to not increase without bound with @xmath95 , but to force the obliquity of the jet to correspond to the obliquity of the field ( and not exceed it ) .
in other words , it may be speculated that instead of scaling as @xmath97 , the flow s obliquity will scale as @xmath98 ( or something similarly behaved ) within the context of a more complicated calculation . ] .
suitably , the onset of complex behavior should be investigated in detail using a numerical non - ideal mhd gcm @xcite .
an important quality of the performed calculation is that it is observationally significant . in the recent years
, the forefront of the observational study of hot jupiters has transitioned away from unembellished detection towards direct characterization . to this end ,
@xcite and @xcite performed the first observational surveys of thermal phase variations on close - in gas giants . while the focus of the former study was aimed at measuring the disparity in brightness between the dayside and the nightside on the planets hd209458b & hd179949b ( see also @xcite ) , the latter study successfully derived a longitudinal thermal map of hd189733b .
the interpretation of both sets of results has been instrumental in constraining the true state of hot jupiter atmospheric dynamics ( see e.g. @xcite ) . unlike purely zonal flows that transport heat longitudionally , magnetically - supported perturbations discussed in this work
give rise to latitudinal flows as well as the associated meridional heat transport .
hence , ( with the exception of specifically chosen field geometries ) the dayside hot spot may be advected not only away from the sub - solar point but away from the equator .
moreover , under the assumption of tidal spin - synchronization , such perturbations are time - independent in a frame co - rotating with planet .
although the aforementioned phase mapping approach @xcite does not posses latitudinal resolution , the newly proposed eclipse mapping method @xcite is likely to prove instrumental in constructing an aggregate of 2d exo - atmospheric profiles . to this end , the calculations of show that the localization of the planetary hot spot within the framework of conventional analysis is model - dependent .
however , they have also pointed out that multi - wavelength scanning of the eclipse may yield exoplanetary portraits of enhanced resolution and the characterization of the atmospheric time - variability may be feasible .
such maps , combined with transit observations in the near - uv , that place constraints on the planetary magnetic field @xcite , may directly probe the phenomena described in this work . accordingly
, observational efforts aimed at quantification of magnetically perturbed jets will be greatly aided by the commencement of next - generation space - based missions such as jwst .
* acknowledgments*. we thank dave stevenson , nick cowan , tami rogers , adam showman , chris spalding , vivien parmentier and tristan guillot for inspirational discussions .
additionally , we are thankful to the anonymous referee , whose thoughtful report led to a substantial improvement of the manuscript . in our computation of the the induced field
, we explicitly assumed that terms involving @xmath15 contribute negligibly to the solution of equation ( [ induction ] ) .
having derived a functional form of the stream - function associated with @xmath15 , we can now deduce the conditions under which this assumption holds . to do so
, we adopt the same approach taken in section 3.2 . the current density induced by the interactions between @xmath15 and @xmath30
is : @xmath99 where in direct analogy with @xmath61 , @xmath100 is the electric potential .
employing zero radial current boundary conditions at the edges of the atmospheric shell as above , the latitudinal and longitudinal dependencies of @xmath100 must conform to that of the radial component of @xmath101 .
suitably , we adopt the following functional form for @xmath100 : @xmath102 as before , the continuity equation for the current yields an ode for @xmath103 : @xmath104 with the aforementioned boundary conditions , the solution for @xmath103 takes the form : @xmath105 having obtained an expression for @xmath106 , we can now write down the associated lorentz force , which arises from the interaction of this current with the axisymmetric component of the field : @xmath107 upon vertical averaging , to leading order in @xmath54 we obtain : @xmath108 consequently , the fractional magnitude of the effect we neglected in section 3 is of order @xmath109 evidently , the derived solution applies when the elsasser number is substantially smaller than unity . | giant planets that reside in close proximity to their host stars are subject to extreme irradiation , which gives rise to thermal ionization of trace alkali metals in their atmospheres . on objects where the atmospheric electrical conductivity is substantial , the global circulation couples to the background magnetic field , inducing supplementary fields and altering the nature of the flow .
to date , a number of authors have considered the influence of a spin - pole aligned dipole magnetic field on the dynamical state of a weakly - ionized atmosphere and found that magnetic breaking may lead to significantly slower winds than predicted within a purely hydrodynamical framework . here , we consider the effect of a tilted dipole magnetic field on the circulation and demonstrate that in addition to regulating wind velocities , an oblique field generates stationary non - axisymmetric structures that adhere to the geometry of the magnetic pole . using a kinematic perturbative approach , we derive a closed - form solution for the perturbed circulation and show that the fractional distortion of zonal jets scales as the product of the field obliquity and the elsasser number .
the results obtained herein suggest that on planets with oblique magnetic fields , advective shifts of dayside hotspots may have substantial latitudinal components .
this prediction may be tested observationally using the eclipse mapping technique . |
in the recently discovered iron - based superconductors @xcite , the superconducting temperatures are found to be in close proximity to an antiferromagnetic ( afm ) and a tetragonal - orthorhombic ( t - o ) structural transitions .
it turns out that the suppression of both the afm and the t - o structural transitions , by doping or by pressure , eventually induces superconductivity .
these phenomena beg the question about the role of spin and lattice degrees - of - freedom in the emergence of superconductivity@xcite .
several theoretical descriptions have been proposed to interpret the relationship between the structural and afm transitions on the basis of orbital ordering or by introducing an intermediate spin - nematic phase resulting from an effective @xmath11 local - spin model or from an itinerant model both with equivalent consequences.@xcite although different in details , these descriptions emphasize the importance of the magnetoelastic coupling in driving the two transitions simultaneously or separately .
cano__et al . _ _
@xcite studied the interplay between the elastic and spin degrees - of - freedom in iron pnictide superconductors using a ginzburg - landau approach , indicating that the magnetoelastic coupling can bring about the particular features of the structural and magnetic transitions in these systems including the emergence of the collinear stripe - type afm ordering .
recently , a microscopic study @xcite of a simple symmetry - allowed model hamiltonian demonstrated that due to the effect of magnetoelastic coupling , the considerable orthorhombic elastic softening is caused by critical spin fluctuations present in the system before magnetic order occurs .
this may explain why the afm transition is often preceded by the t - o structural transition .
it should be pointed out that this picture is similar to the nematic phase model@xcite . to date
, there are very few experimental reports on the magnetoelastic effect in iron pnictides and such reports investigated the role that magnetoelastic effect plays in the structural and magnetic transitions under the application of external driving forces .
magnetoelastic effects have been demonstrated by applying pressure to cafe@xmath12as@xmath12 and inducing an o - t and afm - to - nonmagnetic transitions@xcite , or by applying shear stress to bafe@xmath2(as@xmath0p@xmath1)@xmath2 that shifts the magnetic transition and superconducting critical temperatures significantly.@xcite therefore , it is of interest to study the possibly intrinsic magnetoelastic effect in iron - based superconductors , without introducing any external driving force . the co - doped bafe@xmath2as@xmath2 system exhibits a rich phase diagram with a complex interplay between the structural , magnetic , and superconducting phases.@xcite in the parent bafe@xmath2as@xmath2 compound , the afm ordering transition at _ _
t__@xmath9 coincides with a t - o structural transition at _ _ t__@xmath4 . upon doping both transitions
gradually separate , such that _ _
t__@xmath4 @xmath13 _ _ t__@xmath9 , accompanied with the appearance of superconductivity above @xmath14 in ba(fe@xmath0co@xmath15)@xmath12as@xmath2 .
it is interesting to point out that the orthorhombic distortion @xmath16 in ba(fe@xmath0co@xmath15)@xmath12as@xmath2 with lower co content ( @xmath17 ) shows one clear anomaly at the magnetic transition temperature _ _
t__@xmath9 , but it is absent at _ _
t__@xmath9 for @xmath18 superconductor with intermediate co content . for @xmath19
higher than @xmath200.066 , both the magnetic and structural transitions are completely suppressed and superconducting transition is the only transition observed . only in an intermediate composition region of @xmath21 , does ba(fe@xmath0co@xmath1)@xmath2as@xmath2 exhibit a coexistence of superconductivity , o - structure and afm phases providing potential candidates to investigate the effects of the magnetoelastic coupling .
however , the tendency of these crystals to form twinned orthorhombic domains has hampered definitive determination of inherent features of the intermediate phase between _
_ t__@xmath4 and _ _ t__@xmath9 where the presumed nematic phase exists .
therefore , there have been extensive efforts to de - twin these crystals @xcite to establish the underlying electronic , structural , and magnetic anisotropies that characterize this intermediate phase .
motivated by these issues , we set out to investigate anisotropic features of the crystal structure and domain formation over a wide range of temperatures using high resolution x - ray diffraction methods that reveals the intrinsic magnetoelastic coupling in the ba(fe@xmath0co@xmath1)@xmath2as@xmath2 ( @xmath18 ) superconductor .
the high resolution allows us to separately monitor bragg reflections of different orthorhombic twin domains and study their temperature evolution , as has been done recently on cefeaso@xcite .
the ba(fe@xmath0co@xmath1)@xmath2as@xmath2 ( @xmath18 ) crystal was grown using a self - flux solution method as described previously@xcite .
the crystal has been characterized by x - ray diffraction , neutron scattering , magnetization , resistivity and heat capacity , identifying reported three transitions , structural at _ _
t__@xmath22 k , magnetic at _ _
t__@xmath23 k and sc at _ _
t__@xmath24 k.@xcite a plate - like piece of crystal with its @xmath25-axis perpendicular to its surface was chosen for investigations by high - resolution x - ray scattering technique using the six - circle diffractometer of the 6-id - b beamline at the advanced photon source ( aps ) , argonne national laboratory ( x - ray energy kept at 8 kev ) .
the scattering geometry of our experiment is similar to that used previously by li _ _
et al.__@xcite .
we use orthorhombic indices @xmath26 at all temperatures so that the tetragonal ( _ hkl_)@xmath27 indices are provided in terms of the twin domains in the orthorhombic structure with the following conversion @xmath28 and @xmath29 .
the crystal was mounted at the end of the cold - finger of a displex cryogenic refrigerator with access to ( @xmath30)@xmath10 and high index ( @xmath31)@xmath10 bragg reflections .
flux intensity on the sample was optimized to eliminate beam heating effects of the sample while maintaining a reasonable signal to noise ratio . to achieve that , slit setups and attenuations yielding bragg reflection intensities that scaled with beam attenuations were chosen . in this regard
, it should be noted that the low thermal conductivity in the sc state required a significant increase of beam attenuation to prevent sample - heating during measurements .
these considerations limited the choice of setups , i.e. , resolution , but as discussed below , by adequate analytical tools we captured the intrinsic behavior ( i.e. , charge correlation lengths , in particular ) of this system .
as demonstrated on the cefeaso@xcite , the twin domains are uniformly rotated and separated in reciprocal space by a microscopic shear - angle enabling the characterization of each domain .
1 shows schematically a limited in - plane reciprocal zone with the ( 110)@xmath32 of the untwinned crystal ( a ) that transforms to the ( 200)@xmath10 and ( 020)@xmath10 in the orthorhombic symmetry notation ( b ) .
the misfit angle between the two domains is determined by the orthorhombic distortion ( @xmath33).@xcite the arrows in fig .
[ fig1 ] show typical scans performed at bragg reflections . in the present study
, we monitored the ( 208)@xmath10/(028)@xmath10 systematically ( corresponding to ( 118)@xmath27 at high temperatures ) and also the ( 008)@xmath10 , as a function of temperature .
2(a ) shows the evolution of the orthorhombicity as the temperature decreases with a sharp splitting of the ( 118)@xmath32 into the ( 208)@xmath10 and ( 028)@xmath10 at _ _
t__@xmath34 k. with further decrease in temperature , the orthorhombicity increases without displaying an anomaly at _ _
t__@xmath35 k ( see the inset of fig .
2(a ) ) , which is consistent with the report by kim
_ @xcite . a slight decrease in the splitting
is observed below the sc temperature _ _
t__@xmath3 , indicating that superconductivity and the orthorhombic distortion are coupled.@xcite .
note that for higher co substitution , the suppression of orthorhombic order parameters becomes larger .
for example , the orthorhombic distortion in @xmath36 is completely suppressed and the reentrant transition to tetragonal structure occurs below _ _ t__@xmath3 .
when @xmath19 increases to 0.066 , the orthorhombic distortion vanishes and no t - o transition is observed.@xcite from fig .
2 ( a ) , we point out that the in - plane lattice constant shrinks linearly in the tetragonal phase of ba(fe@xmath0co@xmath1)@xmath2as@xmath2 with decreasing temperature at a rate of @xmath37 / k per unit cell ( linear thermal expansion parameter @xmath38/k ) . by contrast , the thermal expansion along the @xmath25-axis is weakly quadratic ( fig .
2(b ) ) with no abrupt anomaly at _ _ t__@xmath4 and displays a deviation from the quadratic form near _ _
t__@xmath9 ( a weak minimum is observed in _
c_-axis lattice parameter for ( 20__l__)@xmath10 reflection as shown in the inset of fig.2(b ) ) .
the effect of magnetic transition at _ _
t__@xmath9 on the lattice parameter @xmath25 implies a coupling between them , which will be discussed below .
3(a ) shows representative in - plane scans along @xmath39 for the ( 208)@xmath10 bragg reflection at 150 k ( above _ _
t__@xmath4 ) and 46 k ( below _ _
t__@xmath4 ) .
it is clear that the full width at half maximum ( fwhm ) of the bragg peak is broader below _
_ t__@xmath4 .
variation in the line broadening can be due to changes in coherence length , domain size , mosaic distributions , and more likely a combination of all three . to obtain quantitative evaluation of peak line - widths , these and other scans were initially modeled as a gaussian , a lorentzian , their linear combination , or pseudo - voigt line - shapes , but none of these lineshapes yielded satisfactory agreement with the data .
we therefore adopted a standard convolution method by systematically folding a gaussian resolution function and a lorentzian function that reflects an exponentially decaying charge ( chemical ) coherence length as follows , @xmath40 where @xmath41 and @xmath42 , such that @xmath43 and @xmath44 are the fwhm of the gaussian and lorentzian functions in reciprocal space , respectively .
our resolution was high enough to resolve the twin domains separately , i.e. , optimizing peak intensity of the ( 208)@xmath10 from one twin domain and that of the ( 028)@xmath10 of the other domain required sample rotation between the two peaks as has been done for cefeaso@xcite .
while the fwhm of the lorentzian function ( @xmath44 ) represents the intrinsic width @xmath45 of the sample , we should note that the gaussian function in eq .
( [ eq1 ] ) has two contributions : one from geometrical setup ( i.e. , incident and scattered beam divergence ) and the other from the mosaic spread of the studied crystal ( see a detailed discussion in resolution function in ref .
it is by now well established that the mosaic spread of typical pnictides undergoing shear induced o - t transition exhibit mosaic spread changes due to stresses during the transitions .
we therefore attribute the temperature dependence of the gaussian width as arising primarily from the variation in the mosaic distributions .
3(b ) shows the fwhm of the gaussian function for the ( 208)@xmath10 along @xmath39 as a function of temperature indicating anomalies that can be related to the stresses introduced by the structural and magnetic transitions in the system .
the temperature evolutions of the intrinsic width @xmath45 ( @xmath46 ) extracted from the lorentzian functions of ( 208)@xmath10 and ( 028)@xmath10
bragg peaks are shown in fig.4(a ) .
it is evident that anomalies are observed at both _ _ t__@xmath4 and _ _ t__@xmath9 .
similar observation by using high - resolution x - ray diffraction was reported in different systems tbv@xmath0as@xmath1o@xmath47 ( x=0 and 1 ) @xcite , where the intrinsic width of the bragg peaks shows a clear peak at their _
_ t__@xmath3 that is reminiscent of @xmath48 anomalies in the heat capacity .
note that the intrinsic width @xmath45 in unit of @xmath49 , i.e. , the fwhm of the lorentzian profiles from x - ray scattering in reciprocal space corresponds to the inverse charge correlation length @xmath50 @xcite : @xmath51 thus , the temperature dependence of the in - plane charge correlation lengths along and normal to the in - plane component of afm propagation wavevector ( 101)@xmath10 , i.e. , longitudinal charge correlation length @xmath6 ( along the afm bond direction ) and transverse charge correlation length @xmath8 ( along the ferromagnetic bond direction ) , can be derived from the intrinsic widths of ( 208)@xmath10 and ( 028)@xmath10 bragg peaks . as illustrated in fig.4 ( b ) ,
both @xmath6 and @xmath8 show two clear peaks at _ _ t__@xmath4 and _ _ t__@xmath9 , respectively . with the decrease of temperature to _ _
t__@xmath4 , the in - plane charge correlation lengths increase gradually , followed by a rapid decrease below _ _ t__@xmath4 . when the temperature approaches _
_ t__@xmath9 , the in - plane charge correlation lengths increase again .
it is worth emphasizing that the bragg reflections used to extract these data are not allowed by the symmetry of the magnetic structure of the ordered iron moments and they are strictly the result of charge ( nuclei ) ordering .
interestingly , the intrinsic width @xmath45 and charge correlation length obtained from the high - resolution x - ray data show a clear anomaly at the magnetic transition temperature , suggesting these charge bragg reflections are sensitive to the spin - structure and fluctuations .
this is presumably due to the strong magneto - elastic coupling that exerts secondary effects on charge correlations , domain formation and their shape .
below around _ _
t__@xmath9 , we notice that the charge correlation lengths along and normal to the in - plane component of afm propagation wavevector ( 101)@xmath10 are different ( @xmath52 ) , displaying an anisotropic behavior .
the anisotropy in charge correlation length is defined as@xcite @xmath53 @xmath54=0 indicates isotropic correlations ( @xmath6=@xmath8 ) , whereas @xmath54=1 ( @xmath54=-1 ) corresponds to the extreme case of @xmath55 ( @xmath56 ) for structural domains consisting of long linear stripes .
based on this equation , we have derived the anisotropy of the charge correlation length as a function of temperature , as shown in the inset of fig .
the anisotropy is most pronounced below _
_ t__@xmath9 , with values ranging from 0.1 to 0.8 as the temperature is lowered . a possible scenario to interpret the anisotropic charge correlation lengths along @xmath5 and @xmath7 below around _ _
t__@xmath9 in @xmath18 may be related to the presence of the antiphase magnetic domains @xcite .
mazin and johannes @xcite first proposed that antiphase domains and their dynamic fluctuations are central for understanding the high-__t__@xmath3 ferropnictides . very recently , li _ et.al _
@xcite observed surface - pinnned antiphase domains in bafe@xmath2as@xmath2 using high - resolution scanning tunneling microscopy .
since the energy differences between the afm stripe magnetic structure and other afm patterns are small @xcite , it is highly possible that many antiphase magnetic boundaries are formed .
the antiphase domains are pinned at _
@xmath57 _ _ t__@xmath9 , and show dynamic fluctuations in the region of _ _
t__@xmath9 @xmath57 _ t _ @xmath57 _ _ t__@xmath4 .
there are two kinds of simple antiphase domains ( labeled a and b ) with boundaries along @xmath5 and @xmath7 axes , as shown in fig .
due to the same magnetoelastic interactions that lead to a difference in the ferromagnetic and afm bond lengths in the orthorhombic structure , we propose that the formation of such antiphase domains are accompanied by elastic distortions at their boundaries .
the antiphase boundaries along @xmath5 axis influence the magnitude of transverse charge correlation length @xmath8 , whereas the antiphase boundaries along @xmath7 axis affect the longitudinal charge correlation length @xmath6 .
differences in the density of antiphase boundaries in the two directions eventually leads to the anisotropy in @xmath6 and @xmath8 below _ _ t__@xmath9 .
the dynamic fluctuations of the antiphase domains may be responsible for the anisotropy in charge correlation lengths in the small temperature region above _ _ t__@xmath9 ( but lower than _ _
t__@xmath4 ) , as shown in the inset of fig .
4 ( b ) . based on such antiphase magnetic domain scenario and the @xmath11 model ,
the anisotropic charge correlation lengths can be used to estimate the ratio of the magnetic exchange parameters @xmath58 and @xmath59 . as shown in fig .
5 , the energies per - spin ( s ) for forming these two kinds of antiphase boundaries are given by @xmath60 @xmath61 since @xmath58 is afm , the number of antiphase domain boundaries @xmath62 and @xmath63 should scale inversely with the magnetic energy of the domain wall so that @xmath64 the charge correlation length scales inversely proportional to the number of boundaries , i.e. , @xmath65 .
thus , @xmath66 in fig .
4 ( b ) , we observe @xmath67 at low temperatures and from eq .
( 6 ) we can get @xmath68 , which is consistent with the previous calculations or experiments on other iron pnictides , such as lafeaso @xcite and ba(fe@xmath0co@xmath1)@xmath2as@xmath2 ( @xmath69)@xcite .
this is also reasonable for producing a stripe - type afm structure that requires @xmath70 .
it is worthwhile noting that the in - plane charge correlation lengths show gradual changes at temperatures significantly above _
_ t__@xmath4 , probably due to magnetic fluctuations that are known to persist above _
_ t__@xmath4 in similar pnictides@xcite .
we point out that this feature may support the nematic model . in the nematic model@xcite ,
the nematic order coincides with the structural transition with the notion that the driving force for the @xmath71 transition is not elastic in origin but magnetically driven by ising - like interpenetrating afm domains@xcite .
nematic ( magnetic ) fluctuations remain at higher temperature above _ _ t__@xmath4 , which has been suggested by various techniques , such as susceptibility anisotropy @xcite , shear modulus@xcite , inelastic neutron scattering@xcite , and anisotropic in - plane resistivity@xcite .
we now turn to a discussion of the @xmath25-axis charge correlation length . fig .
6(a ) and ( b ) show the temperature dependence of the intrinsic width of the ( 008)@xmath10 bragg reflection and the corresponding out - of - plane charge correlation length along @xmath25 axis ( @xmath72 ) , respectively .
a single anomaly at the magnetic transition temperature _ _ t__@xmath9 and a gradual change above and through _ _
t__@xmath4 , are observed .
the absence of sharp anomaly in @xmath72 suggests that atomic distortions resulting from the t - o structural transition mainly occur in the @xmath73 plane .
this is consistent with the fact that the in - plane lattice parameters @xmath5 and @xmath7 change significantly , but @xmath25 changes weakly around _ _
t__@xmath4 , as can also be seen from fig .
both the out - of - plane charge correlation length @xmath72 and lattice parameter @xmath25 show an anomaly at _ _
t__@xmath9 , showing close correlation between the afm magnetic transition and the modification of structure along @xmath25 axis .
recent experiments and calculations reveal that the fe local moment is very sensitive to the fe - as distance in iron pnictides .
@xcite performed density functional theory ( dft ) calculations within the generalized gradient approximation and found the fe - fe transverse exchange coupling is strongly dependent on both the afm symmetry and the fe - as distance .
belashchenko _ _ et al.__@xcite demonstrated that in layered iron - pnictide compounds , as the fe - as distance is decreased , the degree of itinerancy of fe moments increase .
moreover , the coupling between the local moment and the fe - as distance is controlled by strong covalent fe - as bonding .
recently , neutron diffraction studies of cefeas@xmath0p@xmath1o @xcite and dft calculations @xcite demonstrated that a decrease in fe - as distance induces strong hybridization between fe 3__d _ _ and as 4__p _ _ orbitals , leading to quenched fe magnetic moments
. therefore , the afm transition at _ _
t__@xmath9 is coupled to the change of fe - as distance , which leads to the anomalies in the out - of - plane charge correlation length and lattice parameter @xmath25 . in summary ,
high resolution x - ray diffraction studies on structural bragg reflections of the sc and afm ba(fe@xmath0co@xmath1)@xmath2as@xmath2 ( @xmath19=0.047 ) single crystal reveal secondary effect stemming from the magnetic properties of the system , which is understood to result from intrinsic and strong magnetoelastic coupling .
in addition to showing anomalies around the structural and magnetic transitions , the in - plane charge charge correlation lengths along @xmath5 and @xmath7 axes show anisotropy below around _ _
t__@xmath9 , which probably results from the effect of antiphase boundaries formed along @xmath5 and @xmath7 axes .
employing our anisotropic charge correlation lengths , we are able to estimate the ratio of @xmath74 to be around 1 on the basis of such antiphase magnetic domain scenario and @xmath11 model .
the out - of - plane charge correlation length @xmath72 and lattice parameter @xmath25 exhibit a single anomaly at _ _
t__@xmath9 , which can be associated with the modification of fe - as distance when the afm transition occurs .
our results also show gradual evolution of the bragg peak widths and in - plane charge correlation length above _ _
t__@xmath4 , which is presumably induced by the nematic magetic fluctuations up to almost 200 k.
research at ames laboratory is supported by the us department of energy , office of basic energy sciences , division of materials sciences and engineering under contract no .
de - ac02 - 07ch11358 .
use of the advanced photon source at argonne national laboratory was supported by the us department of energy , office of science , office of basic energy sciences , under contract no .
de - ac02 - 06ch11357 y.kamihara,y .
kamihara , t. watanabe , m. hirano , and h. hosono , _ j. am .
soc . _ * 130 * , 3296 ( 2008 ) .
p m aswathy , j b anooja , p m sarun and u syamaprasad , _ supercond .
technol . _
* 23 * , 073001 ( 2010 ) .
r. m. fernandes , l. h. vanbebber , s. bhattacharya , p. chandra , v. keppens , d. mandrus , m. a. mcguire , b. c. sales , a.s .
sefat , j. schmalian , _ phys .
lett . _ * 105 * , 157003 ( 2010 ) .
v. barzykin and l. p. gorkov , _ phys .
b _ * 79 * , 134510 ( 2009 )
. a. cano , m. civelli , i. eremin , and i. paul , _ phys .
b _ * 82 * , 020408 ( 2010 ) .
r. fernandes and j. schmalian , _ supercond .
technol . _ * 25 * , 084005 ( 2012 ) .
i. paul , _ phys .
lett . _ * 107 * , 047004 ( 2011 ) .
a. kreyssig , m. a. green , y. lee , g. d. samolyuk , p. zajdel , j. w. lynn , s. l. budko , m. s. torikachvili , n. ni , s. nandi , j. b. leao , s. j. poulton , d. n. argyriou , b. n. harmon , r. j. mcqueeney , p. c. canfield , and a. i. goldman , _ phys .
b _ * 78 * , 184517 ( 2008 ) .
kuo , james g. analytis , j .- h .
chu , r. m. fernandes , j. schmalian , i. r. fisher , _ phys .
b _ * 86 * , 134507 ( 2012 ) .
s. nandi , m. g. kim , a. kreyssig , r. m. fernandes , d. k. pratt , a. thaler , n. ni , s. l. budko , p. c. canfield , j. schmalian , r. j. mcqueeney , and a. i. goldman , _ phys .
lett . _ * 104 * , 057006 ( 2010 ) .
d. k. pratt , w. tian , a. kreyssig , j. l. zarestky , s. nandi , n. ni , s. l. budko , p. c. canfield , a. i. goldman , and r. j. mcqueeney , _ phys .
_ * 103 * , 087001 ( 2009 ) .
n. ni , m. e. tillman , j. q. yan , a. kracher , s. t. hannahs , s. l. budko , and p. c. canfield , _ phys .
b _ * 78*,214515 ( 2008 ) . j .-
chu , j.g .
analytis , c. kucharczyk , and i. r. fisher , _ phys .
b _ 79 , 014506 ( 2009 ) .
g. analytis , d. press , k. degreve , t. d. ladd , y .
yamamoto , and i. r. fisher , _ science _ * 329 * , 824 ( 2010 ) .
li , j .- q .
yan , j. w. kim , r. w. mccallum , t. a. lograsso1 , and d. vaknin , _ phys .
b _ * 84 * , 220501(r ) ( 2011 ) .
m. g. kim , r. m. fernandes , a. kreyssig , j. w. kim , a. thaler , s. l. budko , p. c. canfield , r . j. mcqueeney , j. schmalian , and a. i. goldman , _ phys .
b _ * 83 * , 134522 ( 2011 ) .
r. a. cowley , _ acta cryst .
_ * a43 * , 825 ( 1987 ) .
k. c. rule , m. j. lewis , h. a. dabkowska , d. r. taylor , and b. d. gaulin , _ phys .
b _ * 77 * , 134116 ( 2008 ) .
yejun feng , jiyang wang , r. jaramillo , jasper van wezel , s. haravifard , g. srajer , y. liu , z .- a .
xu , p. b. littlewood , t. f. rosenbaum , _ proc .
sci . _ * 109 * , 7224 ( 2012 ) .
j. e. lorenzo , c .
mazzoli , n. jaouen , c. detlefs , d. mannix , s. grenier , y .
joly , and c. marin,_phys .
lett . _ * 101 * , 226401 ( 2008 )
. g. s. tucker , r. m. fernandes , h .- f .
li , v. thampy , n. ni , d. l. abernathy , s. l. budko1 , p. c. canfield , d. vaknin , j. schmalian , and r. j. mcqueeney , _ phys .
b _ * 86 * , 024505 ( 2012 ) . i.i.mazin and m. d. johannes,_nat . phys . _ * 5 * , 141 ( 2009 ) .
z. p. yin and w. e. pickett , _ phys .
b _ * 80 * , 144522 ( 2009 ) guorong li , xiaobo he , jiandi zhang , rongying jin , a. s. sefat , m. a. mcguire , d. g. mandrus , b. c. sales , and e. w. plummer , _ phys .
b _ * 86 * , 060512(r ) ( 2012 ) . f. ma , z .- y
. lu , and t. xiang , _ phys .
b _ * 78 * , 224517 ( 2008 ) h .- f .
li , c. broholm , d. vaknin , r. m. fernandes , d. l. abernathy , m. b. stone , d. k. pratt , w. tian , y. qiu , n. ni , s. o. diallo , j. l. zarestky , s. l. budko , p. c. canfield , and r. j. mcqueeney , _ phys .
b _ * 82 * , 140503 ( 2010 ) h .- f .
li , w. tian , j. l. zarestky , a. kreyssig , n. ni , s. l. budko , p. c. canfield , a. i. goldman , r. j. mcqueeney , and d. vaknin , _ phys .
b _ * 80 * , 054407 ( 2009 ) . s. o. diallo , d. k. pratt , r. m. fernandes , w. tian , j. l. zarestky , m. lumsden , t. g. perring , c. l. broholm , n. ni , s. l. budko , p. c. canfield , h .- f .
li , d. vaknin , a. kreyssig , a. i. goldman , and r. j. mcqueeney , _ phys .
b _ * 81 * , 214407 ( 2010 ) .
s. kasahara , h.j .
shi , k. hashimoto , s. tonegawa , y. mizukami , t. shibauchi , k. sugimoto , t. fukuda , t. terashima , andriy h. nevidomskyy , y. matsuda , _ nature _ * 486 * , 382 ( 2010 ) .
l. w. harriger , h.q .
luo , m.s .
liu , c. frost , j .-
hu , m. r. norman , p. dai , _ phys . rev . b _ * 84 * , 054544 ( 2011 ) .
z. p. yin , s. lebegue , m. j. han , b. p. neal , s.y .
savrasov , and w. e. pickett , _ phys .
lett . _ * 101 * , 047001 ( 2008 ) .
k. d. belashchenko and v. p. antropov , _ phys .
b _ 78 , 212505 ( 2008 ) . c. de la cruz , w. z. hu , shiliang li , q. huang , j.w .
lynn , m. a. green , g. f. chen , n. l. wang , h. a. mook , qimiao si , and p. dai , _ phys .
* 104 * , 017204 ( 2010 ) .
y. lee , d. vaknin , h .- f .
li , w. tian , j. l. zarestky , n. ni , s. l. budko , p. c. canfield , r. j. mcqueeney , and b. n. harmon , _ phys .
b _ * 81 * , 060406 ( 2010 ) . | the interplay between structure , magnetism and superconductivity in single crystal ba(fe@xmath0co@xmath1)@xmath2as@xmath2 ( x=0.047 ) has been studied using high - resolution x - ray diffraction by monitoring charge bragg reflections in each twin domain separately .
the emergence of the superconducting state is correlated with the suppression of the orthorhombic distortion around _ _
t__@xmath3 , exhibiting competition between orthorhombicity and superconductivity . above _
_ t__@xmath4 , the in - plane charge correlation length increases with the decrease of temperature , possibly induced by nematic fluctuations in the paramagnetic tetragonal phase . upon cooling ,
anomalies in the in - plane charge correlation lengths along @xmath5 ( @xmath6 ) and @xmath7 axes ( @xmath8 ) are observed at _
_ t__@xmath4 and also at _ _
t__@xmath9 indicative of strong magnetoelastic coupling .
the in - plane charge correlation lengths are found to exhibit anisotropic behavior along and perpendicular to the in - plane component of stripe - type afm wave vector ( 101)@xmath10 below around _ _ t__@xmath9 .
the temperature dependence of the out - of - plane charge correlation length shows a single anomaly at _ _
t__@xmath9 , reflecting the connection between fe - as distance and fe local moment .
the origin of the anisotropic in - plane charge correlation lengths @xmath6 and @xmath8 is discussed on the basis of the antiphase magnetic domains and their dynamic fluctuations . |
in this paper a ( lattice ) _ region _ is a connected union of unit squares in the square lattice .
a domino is the union of two unit squares that share an edge .
a ( domino ) _ tiling _ of a region @xmath0 is a covering of @xmath0 by dominos such that there are no gaps or overlaps .
denote by @xmath1 the number of tilings of the region @xmath0 .
the _ aztec diamond _ of order @xmath2 is defined to be the union of all the unit squares with integral corners @xmath3 satisfying @xmath4 .
the aztec diamond of order @xmath5 is shown in figure [ qa2](a ) .
in @xcite it was shown that the number of tilings of the aztec diamond of order @xmath2 is @xmath6 . , and its division into two congruent parts.,width=377 ] next
, we consider three related families of regions name _ quartered aztec diamonds _ , that were introduced by jockusch and propp @xcite . divide the aztec diamond of order @xmath2 into two congruent parts by a zigzag cut with 2-unit steps ( see figure [ qa2](b ) for an example with @xmath7 ) . by superimposing two such zigzag cuts that pass the center of the aztec diamond we partition the region into four parts , called quartered aztec diamonds .
up to symmetry , there are essentially two different ways we can superimpose the two cuts . for one of them
, we obtained a fourfold rotational symmetric pattern , and four resulting parts are congruent ( see figure [ qa](a ) ) .
denote by @xmath8 these quartered aztec diamonds .
for the other , the obtained pattern has klein 4-group reflection symmetry and there are two different kinds of quartered aztec diamonds ( see figure [ qa ] ( b ) ) ; they are called _ abutting _ and _ non - abutting _ quartered aztec diamonds . denote by @xmath9 and @xmath10 the abutting and non - abutting quartered aztec diamonds of order @xmath2 , respectively .
.,width=377 ] the number of tilings of a region three kinds of quartered aztec diamond can be obtained by the theorem stated below ( @xcite , theorem 1 ) [ main ] @xmath11 @xmath12 @xmath13 @xmath14 @xmath15 @xmath16 in @xcite , juckusch and propp presented a proof for theorem 1 by investigating properties of _ antisymmetric monotone triangles _ "
. we will prove theorem [ main ] by a visual way in the next section
we have 4 recurrences that were proved by m. ciucu in@xcite , theorem 4.1 .
the _ dual graph _ of a region @xmath0 is the graph whose vertices are unit square in @xmath0 and whose edges connect precisely two unit squares sharing an edge .
a _ perfect matching _ of a graph @xmath22 is a collection of edges such that each vertex of @xmath22 is adjacent to exactly one selected edge .
denote by @xmath23 the number of perfect matchings of @xmath22 . by a well - known bijection between tilings of a region and perfect matchings of its dual graph , we enumerate perfect matchings of the dual graph of a region rather than enumerating its tilings directly .
since we are considering only regions in the square lattice , one can view the dual graphs of those regions as subgraphs of the infinite square grid @xmath24 .
an edge in a graph @xmath22 is called a _ forced edge _ , if it is in every perfect matching of @xmath22 .
one can remove some forced edges from a graph to get a new graph with the same number of perfect matchings .
we have the following lemma by considering forced edges in the dual graphs of quartered aztec diamonds .
instead of comparing the numbers of tilings of the regions , we compare the numbers of perfect matchings of their dual graphs . in each of the four equalities , the dual graph of the region on the left
is obtained from the dual graph of the region on the right by removing forced edges .
the proofs of ( [ eq5])-([eq8 ] ) are illustrated by figures [ force1 ] ( a)-(d ) , respectively . in these figures , the forced edges are represented by the bold edges , and the dual graph of the region on the left of each equality is represented by the graph consisting of shaded unit squares .
next , we consider a well - known family of graphs as follows . consider a @xmath29 rectangular chessboard and suppose that the corners are black .
the @xmath30 _ aztec rectangle _ is the graph whose vertices are the white square and whose edges connect precisely those pairs of white squares that are diagonally adjacent ( see figure [ holeyar](a ) for an example with @xmath31 and @xmath32 ) .
we are interested in the the number of perfect matchings of two families of _ holey aztec rectangles _ as follows .
[ lem4]the number of perfect matchings of a @xmath30 aztec rectangle , where all the vertices in the bottom - most row , except for the @xmath33-st , the @xmath34-nd , @xmath35 , and the @xmath36-th vertex , have been removed ( see figure [ holeyar](b ) for an example with @xmath31 , @xmath32 , @xmath37 , @xmath38 , @xmath39 ) , equals @xmath40 [ lem5 ] the number of perfect matchings of a @xmath30 aztec rectangle , where all the vertices in the bottom - most row have been removed , and where the @xmath33-st , the @xmath34-nd , @xmath35 , and the @xmath36-th vertex , have been removed from the resulting graph ( see figure [ holeyar](c ) , for and example with @xmath31 , @xmath32 , @xmath41 , @xmath42,@xmath43 ) , equals @xmath44 let @xmath22 be a connected subgraph of @xmath24 symmetric about a diagonal lattice @xmath47
. assume all the vertices of @xmath22 on @xmath47 are consecutive lattice points on that line .
go along the line @xmath47 from left to right , and alternate between deleting the edges of @xmath22 that touch them from below , and deleting the edges of @xmath22 that touch them from above ( see figure [ factor ] for an example ) .
let @xmath48 and @xmath49 be the connected components of the resulting graph that are above and below @xmath47 .
it is easy to see that the number of vertices of @xmath22 on @xmath47 must be even if @xmath22 has perfect matchings , let @xmath50 be half of this number . then ciucu s factorization theorem @xcite implies that @xmath51 apply the factorization theorem to the graph @xmath58 with the symmetric axis @xmath47 .
there are @xmath59 vertices of @xmath22 on @xmath47 , so @xmath60 .
it is easy to see that @xmath48 is isomorphic to the dual graph of @xmath61 , and @xmath49 is isomorphic to the dual graph of @xmath62 ( see figure [ qa8](a ) for the case @xmath63 )
. then ( [ eq9 ] ) follows .
again , we apply the factorization theorem to the graph @xmath64 with the symmetric axis @xmath65 .
it is easy to see that @xmath66 is isomorphic to the dual graph of @xmath62 , and @xmath67 is isomorphic to the dual graph of @xmath68 ( the case @xmath63 is shown in figure [ qa8](b ) ) .
moreover , it is easy to see @xmath69 .
this implies ( [ eq10 ] ) .
similarly , two equalities ( [ eq11 ] ) and ( [ eq12 ] ) can be obtained from applying the factorization theorem to @xmath70 and @xmath71 ; and the proofs are illustrated in figures [ qa7](a ) and ( b ) , respectively .
assume @xmath72 and @xmath73 are two sets defined in lemma [ lem3 ] .
denote by @xmath74 the product is taken over all elements @xmath75 s in @xmath72 .
similarly we denote by @xmath76 the corresponding product with elements in @xmath73 .
since the dual graph @xmath22 of @xmath8 is a bipartite graph , the numbers of vertices in two vertex classes of @xmath22 must be the same if @xmath22 admits perfect matchings . by enumerating vertices in each vertex class
we can prove ( [ main1 ] ) ( see example for @xmath63 in figure [ main1eq ] ; the difference between the numbers of vertices in two classes are 1 ) .
similarly , we get the ratio @xmath89 by dividing ( [ eq9 ] ) by ( [ eq10 ] )
. then from ( [ eq7 ] ) and the formula ( [ main3 ] ) for @xmath86 , we can verify ( [ main5 ] ) for @xmath86 .
again , two equalities ( [ eq11 ] ) and ( [ eq12 ] ) imply the ratio @xmath90 ; then from ( [ eq8 ] ) and the equality ( [ main4 ] ) for @xmath86 , we get ( [ main6 ] ) for @xmath86 .
c. krattenthaler .
schur function identities and the number of perfect matchings of holey aztec rectangles .
_ q - series from a contemporary perspective _ ( south hadley , ma ) : 335349 , 1998 .
, 254 , amer .
soc . , providence , ri , 2000 . | we get four quartered aztec diamonds by dividing an aztec diamond region by two zigzag cuts passing its center .
w. jockusch and j. propp ( in an unpublished work ) found that the number of tilings of quartered aztec diamonds is given by simple product formulas . in this paper
we present a simple proof for this result .
* keywords : * aztec diamond , domino , tilings , perfect matchings |
the motion of a particle in temporally and spatially periodic potential has attracted renewed attention as a model of molecular motors @xcite . in order to make it a motor , that generates unidirectional motion out of periodic motion ,
the potential has to be a `` ratchet '' , _
i.e. _ its reflection symmetry in both spatial and temporal direction has to be broken , a fact generally known as a curie s theorem . traditionally , the friction has supplied the source of temporal symmetry breaking @xcite .
there are several recent works @xcite in which the frictionless `` inertial '' regime is studied where the time - asymmetric perturbation is the driving force of the unidirectional motion . through these studies ,
it has become evident that the characeristics of the phase space plays a dicisive role .
the exploration of the full phase space of a dynamical system , however , requires considerable computational resources .
this naturally leads to models of ratchet dynamics in which the differential equation governing the time evolution is reducible to the discrete difference equations , namely the _ map _
one would then like to have a model of a system with ratchet dynamics , which is free from _ ad hoc _ assumptions and parameters that tend to obscure the underlying simple physics . in this article , we propose a simple model of a particle motion under temporally and spatially periodic potential whose key feature is in that the obtained descrete map is a natural extention of the chirikov - taylor map , also known as the standard map @xcite whose dynamical properties have been thouroughly analyzed .
the unidirectional motion of the system is understood in the context of the deterministic diffusion found in the standard map .
also , as a practical benefit of the discrete evolution of the map dynamics , the computational burden is substantially reduced , and it becomes possible to study , for the first time , the transportation properties of a system in a broad region in the controlling parameter space .
a very intricate dependence of the unidirectional motion on the controlling parameter is revealed though a numerical investigation .
we consider the motion of a classical particle described by a time - periodic hamiltonian with instantaneous potential action , or @xmath0-function kick , which has space periodic part and alternating - current type driving motion ; @xmath1 \delta ( t - nt\ { 1+\varepsilon s_n\ } ) } , \end{aligned}\ ] ] where , for now , the trapping potential @xmath2 is set to be symmetric and given by @xmath3 and @xmath4 is the parity of the iteration number @xmath5 .
the evolution of the system is described by @xmath6 \delta ( t - nt\ { 1+\varepsilon s_n\ } ) } , \end{aligned}\ ] ] one can integrate the motion in between the @xmath0-function kicks and obtain the map which describe the evolution of position @xmath7 and momentum @xmath8 at discrete time @xmath9 ; @xmath10 this set of equations can be regarded as a three - parameter extension of the standard map @xcite . the parameter @xmath11 , which is present in original standard map is the strength of the periodic potential that traps the particles in each periodic location .
the parameter @xmath12 is the amplitude of alternating swing motion , and @xmath13 is the measure of its time asymmetry . as with the case of the standard map
, the dynamical properties of the system is studied through the analysis of phase space @xcite .
the transport property of standard map , @xmath12 = @xmath13 = 0 , is characterized in terms of phase space average of functions of position and momentum @xmath14 and @xmath15 .
it is well known that the direct averages @xmath16 and @xmath17 are zeroas a direct result of the fact that the standard map is both time - reversal and space - reflectional symmetric
. however the mean square @xmath18 can be non zero when one has the chaotic motion unbounded by kam tori in @xmath15-direction .
this happens with sufficiently large value of @xmath11 .
one then has deterministic diffusion through chaotic motion which is characterized by the random - walk behavior , @xmath19 @xmath20 @xmath21 . in the extended standard map ( [ eq5])-([eq6 ] )
, a unidirectional motion can be realized if one has by the non - zero value for the @xmath17 itself in the diffusive motion .
the reflection symmetries play an essential role in the existence ( or non - existence ) of the unidirectional motion @xcite .
one can rewrite ( [ eq5])-([eq6 ] ) as @xmath22 these become identical in form to ( [ eq5])-([eq6 ] ) themselves in case of @xmath12 = 0 .
therefore , in this case , one has @xmath23 which means that there can not be any directed transport .
we now define @xmath24 and @xmath25 , obtain a reverse map @xmath26 therefore , if one sets @xmath13 = 0 , the map for @xmath27 and @xmath28 are identical , and one has @xmath29 therefore , one needs both @xmath30 and @xmath31 to have directed motion .
it is important to note that the map ( [ eq5])-([eq6 ] ) in fact atcs differently for odd and ven @xmath5 because of the existence of @xmath32 term .
it is advantageous to consider the phase spaces for odd @xmath5 and even @xmath5 separately . the reflection symmetry between ( [ eq5])-([eq6 ] ) and ( [ eq7])-([eq8 ] )
should then be interpreted as one between two phase space profiles each representing odd @xmath5 and even @xmath5 . .
the strength of periodic potential is set to be @xmath33 .
the temporal asymmetry parameter @xmath13 and amplitude of alternating swing @xmath12 are varied .
left ( right ) column is for odd ( even ) @xmath5 .
, title="fig:",width=340 ] .
the strength of periodic potential is set to be @xmath33 .
the temporal asymmetry parameter @xmath13 and amplitude of alternating swing @xmath12 are varied .
left ( right ) column is for odd ( even ) @xmath5 .
, title="fig:",width=340 ] for the standard map , @xmath34 @xmath35 , as is well known , the map is @xmath36-periodic in both @xmath14 and @xmath15 directions .
the periodicity for @xmath14 direction is kept intact for non - zero @xmath12 and @xmath13 : when one has a map @xmath37 @xmath38 @xmath39 , one also has @xmath40 on the other hand , original @xmath36 periodicty for @xmath15 - direction is lost when non - zero @xmath13 is present . if , however , @xmath13 is a rational number , @xmath41 where @xmath42 and @xmath43 are mutually incomensurate positive integers , a new @xmath44-periodicty for @xmath15-direction emerges , because in this case , one has @xmath45 ( top ) and @xmath46 ( bottom ) started from a set of random initial conditions confined in the region @xmath47
@xmath48 @xmath49 and @xmath50 @xmath48 @xmath51 .
the strength parameters are set to be @xmath52 and @xmath33 . for the case of @xmath35 ,
one has @xmath53 @xmath54 . , title="fig:",width=302 ] ( top ) and @xmath46 ( bottom ) started from a set of random initial conditions confined in the region @xmath47 @xmath48 @xmath49 and @xmath50 @xmath48 @xmath51 .
the strength parameters are set to be @xmath52 and @xmath33 .
for the case of @xmath35 , one has @xmath53 @xmath54 .
, title="fig:",width=302 ] the change of the @xmath15-periodicity has a direct impact on the transportation properties of the extended standard map .
the @xmath44-periodicity in @xmath15-direction for a rational number @xmath55 ( where @xmath42 and @xmath43 are two mutually incommensurate natural numbers ) means that the kam - tori ( when they exit ) appear at least once in @xmath56 period .
thus the chaotic diffusion is bounded within the region in between two adjacent tori .
this fact gives an estimate for the average momentum @xmath57 within a region that include @xmath58 in the form @xmath59 note that only the upperbound is obtained since the exact location of the two adjacent kam tori near @xmath60 can not be determined _ a priori_. this is evident when one considers @xmath35 case , in which the map is @xmath36-periodic in @xmath15-direction , and a similar estimate to above gives @xmath57 @xmath61 @xmath62 , while , as we have lernt in ( [ eq12 ] ) , it is zero because kam tori is placed symmetrically around @xmath60 . as the function of asymmetry parameter @xmath13 and the strength of swing motion @xmath12 for the extended standard map with asymmetric kick ( 5)-(6 ) , title="fig:",width=226 ] as the function of asymmetry parameter @xmath13 and the strength of swing motion @xmath12 for the extended standard map with asymmetric kick ( 5)-(6 ) , title="fig:",width=226 ] as the function of asymmetry parameter @xmath13 and the strength of swing motion @xmath12 for the extended standard map with asymmetric kick ( 5)-(6 ) , title="fig:",width=226 ] the situation
is illustrated in fig.1 where we draw phase space profiles for the frictionless extende standard map with various @xmath13 and @xmath12 .
the strength parameter is chosen to be @xmath11 @xmath63 in these examples .
the effect of the breaking of phase space symmetries on the transportation properties of the system becomes evident by the inspection of fig .
2 , where we plot the portion of the phase space profile that is reached from a set of initial conditions clustered around in the vicinity of @xmath58 . with the onset of non - zero @xmath12 and
non - zero @xmath13 terms , the average momenta in the chaotic region around @xmath58 is shifted from zero , and the directed tranport is realized .
this is essentially in accordance with the conclusion of flach et al .
@xcite , in which they identify this phenomenon asymmetrical levy flight .
we are now ready to ready to look at the directed transport not just in one parameter set , but in a global region in the parameter space . in fig .
3 , we plot the average momentum @xmath57 of the chaotic band that include @xmath58 as the function of swing parameter @xmath12 and time asymmetry @xmath13 . the coupling strength @xmath11 is taken to be @xmath64 ( top ) , @xmath33 ( middle ) and @xmath65 ( bottom ) .
we start from a ensemble of 800 random initial configuration satisfying @xmath47 @xmath48 @xmath49 and @xmath50 @xmath48 @xmath51 .
we symmetrize the ensemble in terms of @xmath66 and @xmath67 to guarantee @xmath68 exactly at the start .
each configuration is evolved with ( 5)-(5 ) for 12,000 times , and the momentum @xmath15 is averaged over that period and over the ensemble to obtain @xmath57 for a given @xmath12 and @xmath13 .
calculation on the grid points of @xmath69 in the space of @xmath70 is used for each of the three figures . from these figures
, it appears that the direction of the asymmetric transport ( positivity or negativity ) is controllable , having step - like structure of increasing @xmath57 with decreasing @xmath13 , which is in broad accordance to our arguments leading to ( [ eq15 ] ) .
an intriguing feature is the direction - reversal of @xmath57 with parameter variation , whose occurence is not at all obvious in the outset .
note also that the speed of unidirectional transport can be rather large : @xmath57 is almost @xmath36 in the peak region , which means that the particle moves one spatial period in one temporal period .
our model can be extended to incorporate the ratchet dynamics with damping through the addition of friction . in place of eqs .
( [ eq5])-([eq6 ] ) , we set @xmath71 } \\
\nonumber & & \times { \delta ( t - nt\ { 1+\varepsilon s_n\ } ) } -\gamma p\end{aligned}\ ] ] where we now add spatially asymmetric term to the trapping potential as @xmath72 the map obtained from these is written as @xmath73 e^{-\gamma ( 1+\varepsilon s_n)}\end{aligned}\ ] ] ( top figures ) , and the periodic attractor @xmath74 ( bottom figures ) .
other parameters are set to be @xmath75 and @xmath76 . , title="fig:",width=302 ] ( top figures ) , and the periodic attractor @xmath74 ( bottom figures ) .
other parameters are set to be @xmath75 and @xmath76 .
, title="fig:",width=302 ] there are now six parameters in the model : @xmath11 , @xmath12 , @xmath77 , @xmath0 , @xmath78 , and @xmath13 .
we reduce this number by setting @xmath13 @xmath79 @xmath80 .
since the temporal asymmetry is already broken by @xmath78 alone , we do not expect to loose out much .
we now obtain @xmath81 e^{-\gamma}\end{aligned}\ ] ] as the function of @xmath78 and @xmath12 for the extended standard map with damping ( 21)-(22 ) the parameters controling the spatial asymmetry are set to be @xmath75 and @xmath76 .
, title="fig:",width=226 ] as the function of @xmath78 and @xmath12 for the extended standard map with damping ( 21)-(22 ) the parameters controling the spatial asymmetry are set to be @xmath75 and @xmath76 .
, title="fig:",width=226 ] as the function of @xmath78 and @xmath12 for the extended standard map with damping ( 21)-(22 ) the parameters controling the spatial asymmetry are set to be @xmath75 and @xmath76 .
, title="fig:",width=226 ] as in the frictionless case , we consider the reflection symmetry of the map ( [ eq21])-([eq22 ] ) . with with the diffinition @xmath24 and @xmath25 , one obtains a reverse map @xmath82e^{-\gamma}\end{aligned}\ ] ] in the elastic limit @xmath78 @xmath83 @xmath80 ,
the map for @xmath27 and @xmath84 are identical , and one has @xmath85 irrespective to the symmetry property of @xmath86 .
this result , puzzling in first inspection , is of course in accordance with curie s theorem .
if , on the other hand , we have @xmath87 , the motion is overdamped , and in sufficient steps , it will come to stand still , namely , @xmath88 with friction , after sufficiently large steps @xmath5 , all trajectories fall into the atractors .
they can be either stable attractors with finite period , or the strange attractors with fractal dimension depending on the parameter value @xcite .
two examples from each case are shown as the top and bottom figures in figs .
4 . as is well known , these two cases alternate with minute variation of parameters .
if one has @xmath89 @xmath90 , the shape of attractor respects the spatial reflection symmetry , when both even and odd @xmath5 are considered .
therefore , one can only have @xmath57 in this case . on the other hand ,
as long as @xmath91 and @xmath92 , one has no spatial mirror symmetry in the phase space , and one should have @xmath93 in general .
it is expected that the value of @xmath57 can vary wildly as fuctions of parameter because of the afore - mntioned alternation of periodic and strange attractors with minute change of he parameter .
the transport property of the extended standard map with friction can be overviewd in fig . 5 ,
in which we show the three - dimensional plot of @xmath57 as functions of two controlling parameters @xmath12 and @xmath78 .
the spatial asymmetry parameters are fixed to be @xmath75 and @xmath94 .
the coupling strength @xmath11 is varied from @xmath95 to @xmath96 from the top to bottom .
as before , we start from a @xmath66 and @xmath67symmetrized ensemble of 1600 initial configuration satisfying @xmath47 @xmath48 @xmath49 and @xmath50 @xmath48 @xmath51 .
each configuration is evolved with ( [ eq21])-([eq22 ] ) for 40,000 times , and the momentum @xmath15 is averaged over that period and over the ensemble to obtain @xmath57 for a given @xmath12 and @xmath13 .
the grid points of @xmath97 in the parameter space of @xmath98 in the range @xmath99 $ ] and @xmath100 $ ] are used for each of the three figures .
one can see from these figures , that the overdamped motion in ratchet potential shows directed transport , but is less controllable than the inertial counterpart , as expected from our preceeding arguments . at the moment
, we do not have any clue as to the physical origin of the intriguing structure found in fig .
however , we feel that the simplicity of our model ( [ eq21])-([eq22 ] ) should yield to a mathematical analysis eventually to give us such clue .
in this work , an extended standard map which has spatial and temporal asymmetry is formulated .
we have outlined its phase space properties , and have studied the unidirectional diffusion , which is thought to be relevant for molecular and biological motors . there
exist several studies of great mathematical rigor on the standard map @xcite .
our extended standard map may hopefully become a subject of similar research .
one immediate extension of the current model is to introduce the randomness .
namely , one may consider @xmath13 not as a fixed quantity , but a fluctuating number for each @xmath5 .
this is a simple way to include the thermal fluctuation into the model .
lastly , the study of quantum version of the current model may be of great interest .
this is particularly so , because the quantum version of standard map @xcite is again a standard tool in the study of quantum chaotic diffusion .
tc expresses his thanks to prof .
izumi tsutsui for helpful dicsussions .
he also thanks members of the theory group of high energy accelerator research organization ( kek ) at tsukuba for the hospitality extended to him during his stay .
this work has been supported in part by the grant - in - aid ( c ) ( no .
10640413 ) of the japanese ministry of science and education , and also by gaas and the czech ministry of education within the projects a1048101 and me170 .
99 f. julicher , a. adjari and j. prost , rev .
* 69 * ( 1997 ) 1269 .
r. bartussek , p. h " anggi , j.g .
kissner , euro .
lett * 28 * ( 1994 ) 459 .
p. jung , j.g .
kissner , p. h " anggi , phys .
lett * 76 * ( 1996 ) 3436 .
m. porto , m. urbakh , j. klafter , phys .
* 85 * ( 2000 ) 491 .
popescu , c.m .
arizmendi , a.l .
salas , f. family , phys .
lett . * 85 * ( 2000 ) 3321 .
s. flach , o. yevtushenko , y. zolotaryuk , phys .
* 84 * ( 2000 ) 2358 .
h. schanz , m .- f .
otto , r. ketzmerick and t. dittrich , phys .
* 87 * ( 2001 ) 070601 .
chirikov , phys . rep .
* 52 * ( 1979 ) 263 .
t. cheon , bussei kenkyu ( kyoto ) * 76 * ( 2001 ) 248 .
see , for example , l.e .
reichl , _ the transition to chaos : in conservative classical systems : quantum manifestations _ ( springer , heidelberg , 1991 ) .
e. ott , _ chaos in dynamical systems _ ( cambridge univ .
press , cambridge , 1993 ) .
greene j. math .
phys . * 20 * ( 1979 ) 1183 .
mackay , j.d .
meiss , _ hamiltonian dynamical systems _
( adam hilger , bristol , 1987 ) .
g. casati , b. chirikov , f. izrailev , j. ford , leture notes phys .
* 93 * , ed . by g. casati and j. ford ( springer , berlin , 1979 ) p.334 .
s. fishman , d.r .
grempel , r.e.prange , phys .
* 49 * ( 1982 ) 509 . | we analyze the transport properties of a set of symmetry - breaking extensions of the chirikov taylor map .
the spatial and temporal asymmetries result in the loss of periodicity in momentum direction in the phase space dynamics , enabling the asymmetric diffusion which is the origin of the unidirectional motion .
the simplicity of the model makes the calculation of the global dynamical properties of the system feasible both in phase space and in controlling - parameter space .
we present the results of numerical experiments which show the intricate dependence of the asymmetric diffusion to the controlling parameters .
keywords : deterministic diffusion , chaotic transport , molecular motor |
outflows of photo - ionised gas are ubiquitous in active galactic nuclei ( agn ) .
recent surveys with _ xmm - newton_and _ suzaku_by @xcite and @xcite ( hereafter ` paper i ' ) have shown that fek - shell outflows are observed in a significant fraction ( @xmath21 ) of active galaxies .
these potentially massive outflows ( @xmath22@xmath23 ) , which are identified through blue - shifted resonant absorption from fexxv he@xmath24(@xmath25kev ) and fexxvi ly@xmath24(@xmath26kev ) , can posses outflow velocities ( @xmath27 ) ranging from a few thousand kms@xmath6 up to mildly relativistic values ( @xmath28c ; @xcite ) which indicates a substantial mass transport into the host galaxy .
the large wind velocity which indicates an origin directly associated with the accretion disk , hence leading them to be dubbed ` disk - winds ' implies that the ensuing outflow may be energetically significant in terms of feedback ( i.e. , @xmath29 , @xcite ) .
the winds are observed in both radio - quiet ( @xcite , paper i ) and radio - loud ( paper i , @xcite ) agn which suggests that they may also be an important addition to the traditional agn unified model ( e.g. , ) . in paper
i we performed a systematic search for fek absorption in 51 _
suzaku_-observed agn which were heterogeneously selected from the heasarc data archive ( see paper i for selection criteria ) .
our main results from that paper were : ( i ) blue - shifted fek - shell absorption is present at the @xmath30 level in 20/51 ( @xmath1 ) of the selected agn ( in 28/73 individual observations ) ; ( ii ) the absorbing gas is typified by mean values of @xmath31 and @xmath32 , while the outflow velocity spans a continuous range between @xmath33kms@xmath6 and @xmath34kms@xmath6 , respectively , with a mean and median velocity of @xmath35c and @xmath36c . ultimately , these results are all consistent with those found by @xcite on the basis of their _ xmm - newton_sample . in this work ,
we build upon our previous results to assess the properties of the disk - wind itself .
the remainder of this paper is laid out as follows : in [ sec : wind parameters ] we outline the equations through which assess the wind properties , in [ sec : parameter_constraints ] we use the inferred values to perform a qualitative assessment of the wind properties , before going on to conduct a quantitative correlation analysis of how the various wind properties scale with the agn bolometric luminosity in [ sec : correlation_analyses ] . in [ sec : discussion ] we then compare our results to those already present in the literature , discuss the wind acceleration mechanism which may be responsible for the observed outflows , and finally assess whether the observed wind is likely to be energetically significant in terms of feedback .
we summarise our results and present our overall conclusions in [ sec : summary_conclusions ] .
l c c c c c & & & & & + & & & & & + 3c111@xmath37 & @xmath38 & 44.7 & @xmath39 & @xmath40 & @xmath41 + 3c390.3@xmath42 & @xmath43 & 44.7 & @xmath44 & @xmath45 & @xmath46 + 4c+74.26@xmath47 & @xmath48 & 47.0 & @xmath49 & @xmath50 & @xmath51 + apm08279@xmath52 & @xmath53 & 45.4 & @xmath54 & @xmath55 & @xmath56 + cbs126@xmath57 & @xmath58 & 44.2 & @xmath44 & @xmath59 & @xmath60 + eso103-g035@xmath61 & @xmath62 & 43.9 & @xmath63 & @xmath64 & @xmath65 + mcg-6 - 30 - 15@xmath66 & @xmath67 & 43.2 & @xmath68 & @xmath69 & @xmath70 + mr2251 - 178@xmath71 & @xmath72 & 45.2 & @xmath73 & @xmath74 & @xmath75 + mrk279@xmath76 & @xmath77 & 43.4 & @xmath78 & @xmath79 & @xmath80 + mrk766@xmath81 & @xmath82 & 43.2 & @xmath83 & @xmath84 & @xmath85 + ngc1365@xmath86 & @xmath87 & 42.8 & @xmath88 & @xmath89 & @xmath90 + ngc3227@xmath91 & @xmath92 & 42.5 & @xmath83 & @xmath93 & @xmath94 + ngc3516@xmath91 & @xmath95 & 43.6 & @xmath96 & @xmath97 & @xmath98 + ngc3783@xmath91 & @xmath77 & 43.6 & @xmath99 & @xmath100 & @xmath94 + ngc4051@xmath91 & @xmath101 & 42.5 & @xmath102 & @xmath103 & @xmath104 + ngc4151@xmath91 & @xmath105 & 42.9 & @xmath106 & @xmath107 & @xmath108 + ngc4395@xmath109 & @xmath110 & 40.7 & @xmath111 & @xmath112 & @xmath113 + ngc5506@xmath114 & @xmath115 & 43.7 & @xmath116 & @xmath117 & @xmath118 + pds456@xmath119 & @xmath120 & 45.3 & @xmath54 & @xmath121 & @xmath122 + swj2127@xmath123 & @xmath124 & 43.7 & @xmath111 & @xmath125 & @xmath126 + columns: ( 1 ) smbh mass , in units of @xmath127 ; ( 2 ) absorption - corrected 11000 rydberg ionising luminosity ( in units of ergs@xmath6 ) , as determined from the best - fit continuum model outlined in paper i ; ( 3 ) measured wind column density , in units of @xmath23 ; ( 4 ) wind ionisation parameter , in units of ergcms@xmath6 ; ( 5 ) inferred wind outflow velocity , as measured in paper i. + references: @xmath128@xcite ; @xmath129@xcite ; @xmath130@xcite ; @xmath131@xcite ; @xmath132@xcite ; @xmath133@xcite ; @xmath134@xcite ; @xmath135@xcite ; @xmath136@xcite ; @xmath137@xcite ; @xmath138@xcite ; @xmath139@xcite ; @xmath140@xcite ; @xmath141@xcite ; @xmath142@xcite ; @xmath143@xcite ; @xmath144@xcite .
+ notes: @xmath145 is taken as the mean of extreme values found in the literature ; @xmath146errors on @xmath147 not present in literature ; @xmath148full designation apm08279 + 5255 ; @xmath149full designation swiftj2127.4 + 5654 . where a source has an fek wind detected in more than one observation , the mean @xmath150 , @xmath151 and @xmath27 values
are reported here ; see paper i for further details .
the maximum column density @xmath150 of gas along the line - of - sight ( los ) is given by @xmath152 , where @xmath153 is the average gas number density and @xmath154 is the maximum observed distance of the absorber .
this , combined with the definition of the ionisation parameter @xmath155 ( @xcite ) , yields the maximum distance that the absorber can be located from the ionising source given its observed column density and ionisation state , @xmath156 where @xmath157 is the source ionising luminosity integrated between @xmath158 rydberg .
conversely , a lower limit on @xmath159 can be inferred by considering the escape radius of the gas given its observed velocity . for a simple keplerian disc orbiting a black hole
the escape velocity at distance @xmath159 is @xmath160 . in the limit that @xmath161 , i.e. , assuming that the measured outflow velocity along the los is equal to the escape velocity at observed radius @xmath159 , we can set a lower limit on the location of the wind , @xmath162 these relations hence allow upper / lower limits to be placed on the wind location given the measured parameters of the absorbing gas , albeit with large uncertainties .
l c c c c c c & & & & & & + & & & & & & + 3c111 & @xmath163 & @xmath164 & @xmath165 & @xmath166 & @xmath167 & @xmath168 + 3c390.3 & @xmath169 & @xmath170 & @xmath171 & @xmath172 & @xmath173 & @xmath174 + 4c+74.26 & @xmath175 & @xmath176 & @xmath177 & @xmath178 & @xmath179 & @xmath180 + apm08279 & @xmath181 & @xmath182 & @xmath183 & @xmath184 & @xmath185 & @xmath186 + cbs126 & & @xmath187 & & @xmath188 & & @xmath189 + eso103-g035 & @xmath190 & @xmath191 & @xmath192 & @xmath193 & @xmath194 & @xmath195 + mcg-6 - 30 - 15 & @xmath196 & @xmath197 & @xmath198 & @xmath199 & @xmath200 & @xmath201 + mr2251 - 178 & @xmath163 & @xmath202 & @xmath203 & @xmath204 & @xmath167 & @xmath205 + mrk279 & @xmath206 & @xmath207 & @xmath203 & @xmath208 & @xmath209 & @xmath210 + mrk766 & @xmath211 & @xmath212 & @xmath213 & @xmath214 & @xmath215 & @xmath216 + ngc1365 & & @xmath217 & & @xmath218 & & @xmath219 + ngc3227 & & @xmath220 & & @xmath221 & & @xmath222 + ngc3516 & & @xmath223 & & @xmath224 & & @xmath225 + ngc3783 & @xmath226 & @xmath227 & @xmath228 & @xmath229 & @xmath230 & @xmath231 + ngc4051 & @xmath232 & @xmath233 & @xmath234 & @xmath235 & @xmath236 & @xmath237 + ngc4151 & @xmath238 & @xmath239 & @xmath240 & @xmath241 & @xmath242 & @xmath243 + ngc4395 & & @xmath244 & & @xmath245 & & @xmath246 + ngc5506 & @xmath247 & @xmath248 & @xmath249 & @xmath250 & @xmath167 & @xmath251 + pds456 & @xmath252 & @xmath253 & @xmath254 & @xmath255 & @xmath256 & @xmath257 + swj2127 & @xmath258 & @xmath259 & @xmath260 & @xmath261 & @xmath262 & @xmath263 + the mass outflow rate of the wind , @xmath264 , is a crucial parameter and is the main means through which the overall flow energetics are assessed .
the mass outflow rate for an arbitrary wind is given by @xmath265 , where @xmath266 and @xmath267 are the density and velocity profile of the wind and @xmath268 is a factor which accounts for the geometry of the system . assuming that the flow has constant terminal velocity @xmath269 and that the absorbing gas has cosmic elemental abundances ( i.e. , @xmath270 of its mass by hydrogen and @xmath271 by helium ) , @xmath272 where @xmath273 is the proton mass and @xmath153 is the electron number density of the plasma . for a thin spherically symmetric isotropic wind @xmath274 , where the product @xmath275 is known as the _ global filling factor _ and accounts for both the solid angle occupied by the flow ( @xmath276 ) and how much of the flow volume is filled by gas ( @xmath277 ) .
thus , @xmath278 where we have neglected the constant factor of order unity .
the value @xmath277 is extremely difficult to determine because it depends on the ionisation and clumpiness of the gas . at low
intermediate ionisation states the flow is likely to be clumpy / filamentary , while at high ionisation states it can be considered largely smooth and of low density because the vast majority of elements are stripped of electrons . in the clumpy case
, the column density of the wind can be given by @xmath279 , with @xmath280 implicitly allowing for inhomogeneities in the flow .
alternatively , at high ionisation states @xmath281 and @xmath282 . substituting for @xmath283 and @xmath154 then leads to algebraic upper and lower limits on the mass outflow rate : @xmath284 the major remaining uncertainty is the opening angle of the system @xmath276 . here
, we adopt the average opening angle @xmath285 as inferred from the global detection fraction of fek winds reported in the literature ( @xmath286 ; @xcite , paper 1 ) .
the assumption of a uniform wind geometry is clearly an oversimplification of real systems ; in reality , numerous factors such as source luminosity , gas density , etc . , will contribute to the shaping of an x - ray disk - wind and therefore the wind geometry is going to differ on an object - by - object basis .
even so , we note that the average opening angle of @xmath285 is comparable to the wind opening measured from the p - cygni - like fek profile in pds456 ( @xmath276 is resolved to be @xmath287 ; see @xcite ) .
this suggests that it is a good approximation to the geometry of real disk - winds .
provided that the wind has already reached a steady terminal velocity by the point at which it is observed , the mechanical power imparted by expelling mass at a rate @xmath13 with velocity @xmath27 is simply equal to its kinetic energy : @xmath288 .
we determine the range of likely kinetic power by substituting for @xmath289 , leading to @xmath290 similarly , the rate at which the outflow transports momentum into the environment of the host galaxy is given by @xmath291 .
substituting again for @xmath13 leads to a plausible range between @xmath292 our method of estimating the inner radius of the outflow , @xmath283 , is proportional to the escape radius of the black hole , @xmath293 , which in turn is proportional to the mass of the central black hole ( bh ) , @xmath147 . by extension
, this also means that the other lower limiting quantities in equations ( [ eq : rmin ] ) , ( [ eq : moutmin ] ) , ( [ eq : kemin ] ) and ( [ eq : poutmin ] ) are also proportional to @xmath147 .
estimates for @xmath147 gathered from the literature are collated in table [ table : measured_wind_properties ] .
we obtain most of the @xmath147 values through the numerous reverberation mapping studies available in the literature ( i.e. , @xcite ) which tend to offer relatively tight constraints on the mass of the central object ; in the cases where an agn had been subject to reverberation mapping multiple times , we adopt the most recent estimate on @xmath147 only .
there are several empirically determined mass estimates for the bhs in 3c111 , eso103-g035 , ngc1365 , ngc5506 and pds456 . for these
agn we report the mean value , with the associated errors taken as half of the range between the minimum and maximum values to account for the uncertainty in the individual estimates on @xmath147 . for cbs126 , where we were unable to find any robust mass estimates in the literature , we infer @xmath147 from the blr line width and luminosity scaling relation of @xcite using the appropriate spectral values listed in @xcite . for 3c390.3
we use the range of values inferred by @xcite . in all other sources
we take @xmath147 directly from the reference listed in the table footnote .
the sample encompasses almost 6 orders of magnitude in black hole mass , with @xmath294 ranging from @xmath295 in the dwarf - seyfert galaxy ngc4395 ( @xcite ) , all the way up to an estimated @xmath296 in the high luminosity bal quasar apm08279 + 5255 ( @xcite ) .
we estimate the bolometric luminosity for each agn : @xmath297 , where @xmath298 is the unattenuated source luminosity integrated between @xmath299kev and @xmath300 is the bolometric correction factor .
various studies have shown that the spread of @xmath300 amongst individual agn is quite large ( e.g. , @xcite ) , whilst the correction appropriate for a particular agn can be a function of luminosity ( @xcite ) , eddington ratio ( @xcite ) , or both ( @xcite ) , such that the uniform application of a single bolometric correction factor may be inappropriate for a heterogeneously selected sample .
we therefore primarily use the @xmath300 values listed in the works of @xcite and @xcite which are empirically determined on the basis of the broad - band spectral energy distributions of each agn .
several of the agn in the sample are not listed in these works . for these sources
, we searched the literature for an appropriate bolometric luminosity and used that instead ( see table [ table : measured_wind_properties ] caption ) . only for 3c111 and
cbs126 were we unable to locate either a bolometric correction factor or an definite and empirically measured bolometric luminosity ; for these two sources we simply assume that @xmath301 , which is similar to that found previously for other sources which harbour ufos ( e.g. , @xcite .
values for @xmath298 ( hereafter denoted @xmath302 ) , @xmath300 and the resultant estimate of @xmath16 are all listed in table [ table : normalised_table ] . for completeness
, we also compute the likely eddington ratio of each agn : @xmath303 , where @xmath304 .
several of the agn in the have fek absorption detected in multiple epochs ( see paper i ) , such as is the case for 3c111 , mrk766 , ngc1365 , ngc3227 , ngc3783 , ngc4051 , or have more than one absorption trough which comprises a multi - velocity system ( pds456 and apm08279 + 5255 ) . for these agn considering the mean parameter value and accounting for intrinsic outflow variability between epochs by folding the uncertainties on each of the individual measurements into the error bar . in the majority of cases we find the individual measurements to be largely consistent within the errors , suggesting that the outflow is persistent across the different epochs . only in 3c111 ,
mrk766 and ngc4051 do there appear to be significant differences at the 90% level ( see table [ table : measured_wind_properties ] ) , and this is reflected in the broad error bars for these sources .
for pds456 and apm08279 + 5255 , which both appear to have two fek absorption systems at different velocities , we adopt the mean velocity and again fold the range of possible velocities into the error bar .
table [ table : measured_wind_properties ] summarises the measured parameters of the winds detected in the _ suzaku_sample .
using the values in table [ table : measured_wind_properties ] we first computed the wind parameters in standard units and then normalised them to the appropriate values for a given black hole mass ; the resultant values are noted in tables [ table : main_results ] and [ table : normalised_table ] , respectively.we normalised the specific parameters as follows : ( i ) the distance @xmath305 to units of schwarzchild radius @xmath306 , ( ii ) the kinetic luminosity @xmath307 to @xmath308 , ( iii ) the mass outflow rate @xmath309 to the eddington accretion rate @xmath310 ( assuming @xmath311 for the accretion efficiency ) , and ( iv ) the momentum rate to the eddington momentum rate @xmath312 .
algebraically , this yields a set of 8 normalised equations : @xmath313 errors on the normalised parameters were determined by both propagating the uncertainties on the measured xstarvalues through the various equations , and also taking into account the error on @xmath147 where applicable . for cbs126 , ngc1365 , ngc3227 , ngc3516 and ngc4395 , which all have relatively slow wind velocities of the order @xmath314c ( see table [ table : measured_wind_properties ] ) , we found that @xmath315 ( within the errors ) which indicates that either : ( i ) the outflow may not have achieved the requisite escape speed for its observed location , in which case the wind material may not escape into the host galaxy unless it is subject to an additional acceleration mechanism once it has crossed the los , or ( ii ) that the wind steam line flows almost perpendicularly along the los such that we are only seeing the tangential component of @xmath27 . while the associated error bars in these slow sources may overlap ( suggesting that the upper and lower limits may be formally consistent ) we choose to only report the upper limiting quantities for these five agn in tables [ table : main_results ] and [ table : normalised_table ] so that we do not introduce an artificially tight ( and potentially misleading ) constraint on the properties of the wind . in the remainder of our analysis we separate the outflow sample into two groups based on their velocity : those with @xmath316c ( @xmath317kms@xmath6 ) are hereafter classified as ` slow ' , whilst those with @xmath2c are ` fast ' . whilst the chosen velocity threshold is essentially arbitrary
, this classification scheme provides a useful means to distinguish between those pristine ` disc - winds ' from those which may just be the higher - ionisation component of the inner - blr or the more distant warm absorber , with which they probably share an overlap in velocity space and location ( e.g. , see @xcite ) .
l r c c c r r r r r r & & & & & & & & + & & & & & & & & + 3c111 & @xmath318 & @xmath319 & @xmath320 & @xmath321 & @xmath322 & @xmath323 & @xmath324 & @xmath325 + 3c390.3 & @xmath326 & @xmath327 & @xmath328 & @xmath329 & @xmath330 & @xmath331 & @xmath332 & @xmath333 + 4c+74.26 & @xmath334 & @xmath335 & @xmath336 & @xmath337 & @xmath338 & @xmath339 & @xmath340 & @xmath341 + apm08279 & @xmath342 & @xmath91 & @xmath343 & @xmath344 & @xmath345 & @xmath346 & @xmath347 & @xmath348 + cbs126 & @xmath349 & @xmath350 & & @xmath351 & & @xmath352 & & @xmath353 + eso103-g035 & @xmath354 & @xmath355 & @xmath356 & @xmath357 & @xmath358 & @xmath359 & @xmath360 & @xmath361 + mcg-6 - 30 - 15 & @xmath362 & @xmath363 & @xmath364 & @xmath365 & @xmath366 & @xmath367 & @xmath368 & @xmath369 + mr2251 - 178 & @xmath370 & @xmath371 & @xmath328 & @xmath372 & @xmath373 & @xmath374 & @xmath375 & @xmath376 + mrk279 & @xmath377 & @xmath378 & @xmath379 & @xmath380 & @xmath381 & @xmath382 & @xmath383 & @xmath384 + mrk766 & @xmath385 & @xmath386 & @xmath387 & @xmath388 & @xmath322 & @xmath389 & @xmath390 & @xmath391 + ngc1365 & @xmath392 & @xmath393 & & @xmath394 & & @xmath395 & & @xmath396 + ngc3227 & @xmath397 & @xmath398 & & @xmath399 & @xmath400 & @xmath401 & & @xmath402 + ngc3516 & @xmath403 & @xmath404 & & @xmath405 & & @xmath406 & & @xmath407 + ngc3783 & @xmath408 & @xmath409 & @xmath410 & @xmath411 & @xmath412 & @xmath413 & @xmath414 & @xmath415 + ngc4051 & @xmath416 & @xmath417 & @xmath418 & @xmath419 & @xmath420 & @xmath421 & @xmath422 & @xmath423 + ngc4151 & @xmath424 & @xmath425 & @xmath356 & @xmath426 & @xmath427 & @xmath428 & @xmath429 & @xmath430 + ngc4395 & @xmath431 & @xmath432 & & @xmath433 & & @xmath434 & & @xmath435 + ngc5506 & @xmath436 & @xmath437 & @xmath438 & @xmath439 & @xmath440 & @xmath441 & @xmath442 & @xmath443 + pds456 & @xmath444 & @xmath109 & @xmath343 & @xmath445 & @xmath446 & @xmath447 & @xmath448 & @xmath449 + swj2127 & @xmath450 & @xmath335 & @xmath379 & @xmath451 & @xmath452 & @xmath453 & @xmath454 & @xmath455 + notes: @xmath128the bolometric luminosity of apm08279 + 5255 is taken as @xmath456 ( @xcite ) , assuming a conservative magnification factor of @xmath457 ( @xcite ) ; @xmath458 for pds456 taken from @xcite ; @xmath459 correction factors taken from @xcite or @xcite , respectively ; @xmath460 estimated from the @xmath461 correlation of @xcite using the best - fit @xmath462 values that we found in paper i. @xmath146@xmath302 here corresponds to the unattenuated source luminosity between @xmath299kev . in figure
[ fig : constraint_diagrams ] we show constraint diagrams for the various wind parameters because it is simply proportional to @xmath463 and therefore follows a similar overall distribution to the one for @xmath14 ] . from figure
[ fig : constraint_diagrams]a b it is clear that the absorbing material is distributed across several orders of magnitude in distance from their black hole , with @xmath159 spanning @xmath465 cm or @xmath466@xmath467 . by considering the mean range for the fast outflows ( see caption for more details ) ] we see that the fast outflows appear to be located between @xmath468 cm ( @xmath469@xmath467 ) distance range .
this corresponds to @xmath470pc and implies that whilst there may be some limited overlap between the fast fek outflows and the traditional soft x - ray warm absorber ( which is usually inferred to be on parsec - scale distances , e.g. , and references therein ) , the former are generally located much closer to their central black hole .
this is in agreement with other results in the literature which find the highly - ionised fek outflows most likely originate in a wind from the inner regions of the accretion disc ( e.g. , @xcite ) , and is also consistent with what @xcite found on the basis of the _ xmm - newton_outflow sample .
following our estimates on @xmath159 , figure [ fig : constraint_diagrams]c shows an analogous plot for @xmath13 .
the constraints on @xmath13 also vary significantly , spanning almost 8 orders of magnitude , and ranging from @xmath471gs@xmath6 in the dwarf seyfert galaxy ngc4395 all the way up to @xmath472gs@xmath6 in the massive quasar apm08279 + 5255 .
the other winds have @xmath13 estimates distributed between these two extremes , with the mean range for the fast outflows falling between @xmath473gs@xmath6 ( @xmath474yr ) .
interestingly , and while there remains considerable uncertainty in each case , @xmath475 is of the order of the eddington rate , i.e. , @xmath476 , while even @xmath477 is still @xmath478 of @xmath479 .
this immediately suggests that the winds are transporting a substantial amount of material into their host galaxies .
the ejected mass has a large kinetic power : figure [ fig : constraint_diagrams]e shows that while the range of @xmath14 again spans several orders of magnitude , the overall mean for the fast flows is constrained to between @xmath480 and @xmath481ergs@xmath6 .
the normalised values show that , on average , @xmath482 is @xmath483 , with a mean range of @xmath484 .
the wind mass outflow rate and the associated kinetic power are therefore significant fractions of the eddington limited values . in [ subsec : energetic_significance ] we extend these results to assess whether the wind is energetically significant in terms of feedback .
as mentioned previously , our estimates on @xmath159 are subject to large uncertainties .
it can therefore be useful to compare a few of our inferred values with some more robust measurements available in the literature .
we consider here the exemplar cases of pds456 and mrk766 .
currently , pds456 is the only agn where the location of the fek - shell wind has been robustly constrained from discrete line variability . on the basis of a recent
_ suzaku_campaign @xcite constrained the fek absorber in pds456 to @xmath485@xmath467 .
this is entirely consistent with what we infer using equations ( [ eq : rmax][eq : rmin ] ) .
similar is also true for mrk766 : @xcite and @xcite estimates the absorbing gas to be located on distances of the order @xmath486@xmath467 , while @xcite later showed that both the broad - band spectral variability and the fek absorption could also be attributed to a stratified , inhomogenous and clumpy absorber which occults the agn at a distance of @xmath487@xmath467 .
these estimates are again consistent with what we find here ( see figure [ fig : constraint_diagrams]a - b ) , suggesting that the two studies are perhaps probing the same layer of gas . elsewhere in the literature
, @xcite used a detailed spectral timing analysis to de - convolve the x - ray spectrum of ngc4051 into its constituent absorption- and reflection - dominated components , with the hard x - ray reverberation signal suggesting that the reflecting gas is @xmath488@xmath467 from the black hole ( at the 90% level , see @xcite ) .
this is around one order of magnitude closer than where we estimate the fek absorbing gas to be in this source .
one possibility is that the reflecting layer detected by @xcite is physically distinct to the fek absorber , possibility associated with the compton - thick occluding clouds posited by ( * ? ? ?
* , in prep ) .
versus : ( a ) @xmath489 , ( b ) @xmath490 and ( c ) @xmath491 , as noted in table [ table : main_results ] .
the red and blue data - points correspond to agn with @xmath492kms@xmath6 and @xmath493kms@xmath6 , respectively .
the solid black line corresponds to the ` best - fit ' linear regression to the fast winds , as estimated from 100,000 mcmc realisations of the dataset using the linmix_err bayesian regression algorithm ( @xcite ) , while the grey shaded area denotes the region which contains 90% of the posterior probability distribution .
the parameters of the fits are reported in table [ tab : linear_regression ] .
, title="fig:",scaledwidth=45.0% ] versus : ( a ) @xmath489 , ( b ) @xmath490 and ( c ) @xmath491 , as noted in table [ table : main_results ] .
the red and blue data - points correspond to agn with @xmath492kms@xmath6 and @xmath493kms@xmath6 , respectively .
the solid black line corresponds to the ` best - fit ' linear regression to the fast winds , as estimated from 100,000 mcmc realisations of the dataset using the linmix_err bayesian regression algorithm ( @xcite ) , while the grey shaded area denotes the region which contains 90% of the posterior probability distribution .
the parameters of the fits are reported in table [ tab : linear_regression ] .
, title="fig:",scaledwidth=45.0% ] versus : ( a ) @xmath489 , ( b ) @xmath490 and ( c ) @xmath491 , as noted in table [ table : main_results ] .
the red and blue data - points correspond to agn with @xmath492kms@xmath6 and @xmath493kms@xmath6 , respectively .
the solid black line corresponds to the ` best - fit ' linear regression to the fast winds , as estimated from 100,000 mcmc realisations of the dataset using the linmix_err bayesian regression algorithm ( @xcite ) , while the grey shaded area denotes the region which contains 90% of the posterior probability distribution .
the parameters of the fits are reported in table [ tab : linear_regression ] .
, title="fig:",scaledwidth=45.0% ] supplementing the constraint diagrams we also carried out a correlation analysis .
this analysis bears notable similarities to the one recently conducted by @xcite for the _ xmm - newton_outflow sample .
there , the authors combined the parameters of the hard - band absorber , as measured in their fek absorption survey with _ xmm - newton _ , with those for the soft x - ray warm absorber collated from the literature to determine how the agn wind parameters vary globally in relation to their distance from the black hole ( @xcite ) . here , noting that _ suzaku_does not have the soft x - ray energy resolution necessary to constrain the velocity of the soft x - ray absorbing gas , we consider only the fek absorbers and do not consider the warm absorber .
whilst this leads to a relatively small dynamic range in terms of their distance from the black hole , the heterogeneous nature of the _ suzaku_sample encompasses a significantly broader range of agn in terms of their bolometric luminosity ( @xmath16 ) .
suzaku_sample therefore offers a useful opportunity to probe how the wind parameters may vary in comparison to the luminosity of their host agn .
llrrrr & & & & + ( r.75eml)1 - 2 & & & & & + @xmath16 & @xmath494 & @xmath495 & @xmath496 & @xmath497 & @xmath498 + & @xmath499 & @xmath500 & @xmath501 & @xmath502 & @xmath503 + & @xmath27 & @xmath504 & @xmath505 & @xmath497 & @xmath506 + & @xmath159 & @xmath507 & @xmath508 & @xmath509 & @xmath510 + & @xmath13 & @xmath511 & @xmath512 & @xmath513 & @xmath514 + & @xmath14 & @xmath515 & @xmath516 & @xmath517 & @xmath518 + & @xmath15 & @xmath519 & @xmath520 & @xmath521 & @xmath522 + & @xmath523 & @xmath524 & @xmath525 & @xmath526 & @xmath527 + & @xmath528 & @xmath529 & @xmath530 & @xmath531 & @xmath532 + & @xmath533 & @xmath534 & @xmath535 & @xmath526 & @xmath536 + & @xmath537 & @xmath538 & @xmath539 & @xmath540 & @xmath541 + @xmath542 & @xmath15 & @xmath543 & @xmath520 & @xmath521 & @xmath522 + notes : all regressions were performed using the linmix_err bayesian regression routine ( @xcite ) which takes into account both measurement and upper limits in the @xmath544-variable .
the noted best - fit values correspond to the median value of the posterior probability distribution , as simulated from 100,000 mcmc realisations of the data , with errors taken as the range of values which encompass 90% of the posterior probability .
see @xcite for more details .
nominal values correspond to a fit to the entire sample , while those in brackets are for an analogous fit with ngc4395 removed .
@xmath148regression computed with lower limits excluded .
+ columns : ( 1 ) slope of the linear regression ; ( 2 ) normalisation / intercept of the best - fit regression line ; ( 3 ) standard deviation of the intrinsic scatter in the data ; ( 4 ) best - fit linear correlation coefficient .
@xmath545 denote perfect positive / negative correlations , respectively .
we compute linear regressions of the form @xmath546 , where @xmath24 and @xmath547 are the intercept and slope of the straight line fit , respectively , using the linmix_err bayesian regression algorithm ( @xcite ) .
this routine employs markov chain monte carlo ( mcmc ) techniques to self - consistently account for measurement errors and intrinsic scatter in the data , whilst also allowing for limited censorship ( upper - limits only ) in the independent variable .
we computed regression parameters from 100,000 mcmc realisations of the data by the linmix_err routine .
this yields posterior distributions for the intercept @xmath24 , slope @xmath547 , standard deviation ( scatter ) of the data @xmath548 , and the pearson linear correlation coefficient @xmath549 .
we adopt the median value of the posterior distribution as our ` best - fit ' to the data , and estimate the parameter errors as the range of simulated values which encompass 90% of the posterior distribution about the median .
we note the slow winds have @xmath550kms@xmath6 which is similar to that measured in the traditional soft x - ray warm absorber .
it is therefore unclear whether these winds are associated with a _ bona fide _ disk - wind or alternatively with another layer of gas which is more distant from the black hole , e.g. , with a higher ionisation component of the warm absorber .
for this reason , we fit all subsequent regression analyses to the fast winds only , @xmath492kms@xmath6 . owing to their high velocities , these faster winds are kinematically distinct from the soft x - ray absorbing gas and thus may more obviously represent the signature of a pristine disk - winds .
this limits our analysis to the winds which are capable of escaping the gravitational potential of the central black hole ( based on our los ) , and are therefore more likely to have an effect on the host galaxy in terms of feedback . in figure
[ fig : lbol_absorber ] we show how @xmath489 , @xmath490 and @xmath491 , as measured in paper i and summarised in table [ table : main_results ] , vary with @xmath16 . from figures [ fig : lbol_absorber]a and [ fig : lbol_absorber]b we find no discernible relationships between either @xmath150 or @xmath151 and @xmath16 ( @xmath551 at 90% confidence ) : the @xmath489 values cover a wide range for a given value of @xmath16 , while @xmath490 appears to cluster at around @xmath552 , mirroring the median of the parameter distributions that we found in paper i. this also appears to be true for both the fast and the slow wind sub - samples .
conversely , there is a correlation present between @xmath27 and @xmath16 , with the faster winds tending to be being observed in more luminous agn .
the fast winds have a constrained slope of @xmath553 , with @xmath554 and @xmath555 .
in contrast , the slower systems , i.e. , those with @xmath314c , appear to be isolated to the lower left quarter of the plot which suggests that low velocity winds are preferentially located in lower luminosity systems . this is consistent with the overall picture that wind velocity is correlated with @xmath16 ; fitting a correlation to the entire sample yields @xmath556 .
this is formally consistent within the errors but does suggest that the inclusion of the slow winds may skew the relationship to a slightly steeper slope .
figure [ fig : lbol_raw ] then shows how the raw wind parameters vary with @xmath16 . here ,
and in subsequent figures , each data point corresponds to the mid - point in the range for each parameter , calculated as , e.g. , : @xmath557 where nerr and perr are the negative and positive error , respectively .
the associated error bars then denote the range between the maximum and minimum values , including the uncertainties arising from the input variables . in equation
[ eqn : mean_value ] we are essentially only considering the upper- and lower - limits for _ all _ derived values in the sample , such that in the cases where the minimum and maximum values are themselves unconstrained ( i.e. , only and upper- or lower - limit is available with no associated error ) we are able to use the censored value when calculating the mid - point .
this ensures that all of the available data is included in our analysis .
to first order , we find that all of the raw wind parameters are plausibly correlated with @xmath16 to some degree . whilst the slope for @xmath558 is tentative , and marginally consistent with @xmath551 at the 90% level ( figure [ fig : lbol_raw ] ,
the overall distribution is skewed towards a positive relationship among the data which suggests that the winds are observed at larger radii in brighter agn , as would be expected given the larger typical size - scales in these systems .
figure [ fig : lbol_raw ] ) then shows that there is a positive relationship between @xmath559 ( @xmath560 ) , such that the winds in more luminous agn are correspondingly more massive . in both of these cases
the slow winds appear to follow the same overall trend as the fast ones , despite not formally being included in the regression computation .
similar is also true for @xmath561 and @xmath562 , which share strong positive slopes of @xmath563 ( @xmath564 ) and @xmath565 ( @xmath566 ) , respectively .
the observed slopes here likely stem from their mutual dependence on @xmath27 : @xmath567 and @xmath568 .
indeed , the @xmath569 and @xmath570 relationships implied here are formally consistent with the @xmath571 found earlier in figure [ fig : lbol_absorber]c .
this ultimately suggests that the wind velocity may be the driving factor in the observed relationships , with the uncertainties on the other wind parameters only adding to the underlying scatter of the data . in conclusion , these results indicate that disk - winds in higher luminosity sources are ( tentatively ) located farther away from their central black hole in terms of absolute distance ( as expected for larger @xmath147 ) , they contain more material , and they are accordingly more energetic .
the driving factor behind these relationships appears to be the wind velocity , which is faster in more luminous objects .
similar scatter plots for the normalised wind parameters are shown in figure [ fig : lbol_norm ] .
in contrast to before , this time we are unable to discern any linear correlations in the normalised @xmath572 and @xmath573 plots .
in fact , and whilst there is a weak negative correlation suggested in the case of the former , both of the relationships are formally consistent with @xmath574 such that the @xmath523 and @xmath528 ratios are largely independent of @xmath16 . the fast winds
are therefore observed at similar relative distances from the smbh regardless of the source luminosity , and their mass outflow rate is a similar fraction of @xmath479 .
the slow winds also appear to follow consistent relationships .
both @xmath533 ( figure [ fig : lbol_norm]d ) and @xmath537 ( figure [ fig : lbol_norm]e ) are also formally consistent with a slope of zero , although weak positive correlations are suggested . in both figure [ fig : lbol_norm]c and [
fig : lbol_norm]d , the slow winds appear to lie below the computed regression line for the fast winds . their location in these plots appears to mirror that seen in figure [ fig : lbol_norm ] for the @xmath575 correlation , which is again consistent with the wind velocity being the parameter behind the relationships . for completeness
, we also tried to determine whether the parameters of the wind scale with the eddington ratio of the agn .
unfortunately , however , the relatively tight clustering in the data at @xmath576 , compounded by the lack of fek wind parameters in low eddington ratio sources ( i.e. , @xmath577 ) sources , meant that we were unable to discern any plausible relationships among the data .
as a consistency check , we also checked for evidence of relationships amongst the raw @xmath578 values finding that the same correlations are still present in the data , albeit at lower significance level .
the presence of the same correlations in both the contiguous @xmath578 and the non - simultaneous @xmath16 indicates that the correlations are not driven by uncertainties in @xmath300 .
in paper i we showed that @xmath1 of the agn in the _ suzaku_sample exhibit evidence for highly - ionised winds in their fek band .
this is in agreement with the detection fraction reported by @xcite for _ xmm - newton _ , and corroborates their conclusion that such winds are either ( i ) persistent over the active phase and thus have a large covering fraction , or ( ii ) transient in nature and have a covering fraction of unity for only a fraction of the time .
we also showed that the @xmath489 and @xmath490 distributions are also entirely consistent between the two samples , covering ranges of @xmath579 and @xmath580 in both , respectively , with the outflow velocity @xmath27 also being largely consistent on the basis of a kolmogorov - smirnov ( ks ) test . by comparing the distributions of wind parameters found in this work to those determined by @xcite we similarly find that the fek absorbers detected in both the _ suzaku_and _ xmm - newton_studies are typified by the same range of physical parameters and occupy the same parameter space in terms of their overall location and energetics : @xmath581 ( @xmath582pc ) , @xmath583 ( @xmath584yr ) and @xmath585 .
their medians are also similar with @xmath586 , @xmath587 and @xmath588 for _ suzaku_(_xmm - newton _ ) , respectively , in natural units .
the _ suzaku_outflow sample therefore robustly confirms that fek absorption are a real physical component of emergent x - ray spectrum , that the implied winds likely have a large covering fraction , and that they are typified by large column densities and high ionisation .
the range of outflow velocities , which are typically @xmath589 , but can be as low as @xmath590 in some cases , also suggests that the highly - ionised winds share an overlap in velocity space with the traditional soft x - ray warm absorber .
this is in line with the results of @xcite who argue that some of the soft x - ray warm absorbers could be the artefacts of accretion disc winds which have propagated farther away from the smbh .
another interesting study that we can consider is that of @xcite , who compare the kinetic power of both the warm absorber and relativistic jets to @xmath16 of their respective source .
they consider both black - hole binaries ( bhbs ) and agn which means that their study samples an extremely broad spectrum of both black hole mass @xmath591 $ ] and bolometric luminosity [ @xmath592 .
even though the @xcite study deals primarily with the low - velocity soft x - ray warm absorber they do include a few ` ultra - fast outflows ' ( which they define as those with @xmath593c , as we have done in this work ) as a matter of comparison , although they do not fit them directly as part of their regression analysis ; two of their ` ultra - fast ' sources , namely 3c111 and apm08279 + 5255 , are also included in our _ suzaku_sample . in principle
we could compare the properties of the _ suzaku_-detected outflows with those of king et al . to determine if our high velocity fek winds scale in a similar manner to the warm - absorber .
however , this is not possible in practice because the two studies compute the wind kinetic power according to different assumptions ( king et al .
give their quantities in units of per - covering - fraction ) , which means the intercept of any linear regressions will therefore by intrinsically offset .
even so , the _ slope _ of any linear fit which is the important parameter as it measures how the kinetic power varies with bolometric luminosity can still be compared because it is largely independent of the normalisation .
king et al .
find the kinetic power of warm absorber to be strongly correlated with @xmath16 , scaling with a tight global slope of @xmath594 , while they found that those in agn have a much flatter local slope of @xmath595 when considered in isolation ( although that this could be driven by there being a smaller number of agn ) . in figure
[ fig : lbol_raw]d we showed that kinetic power of the _ suzaku_-detected fek winds are also strongly correlated with the bolometric luminosity with a slope in the range @xmath516 ( @xmath596 ) .
this result , which is driven by the @xmath597 dependence , is in good agreement with the slope reported by king et al . for their global sample and suggests that the fast fek winds may scale with @xmath16 in a similar manner to the warm absorber .
to summarise , by comparing our results for those obtained by @xcite we see that the fast fek winds ( @xmath2c ) is similar to that found for slow warm absorbers . in this work
we shown that the observed wind velocity , and subsequently the overall wind energetics , are proportional to the agn bolometric luminosity .
it is therefore tempting to conclude that the winds are accelerated by radiation pressure . in this section
we investigate whether radiation pressure alone can account for the observed correlations .
in contrast to the undoubtedly line - driven winds which are prevalent in the uv spectra of agn , the high ionisation state of the fek winds suggests that line - driving is unlikely to be the dominant acceleration mechanism here .
an alternative means of radiatively accelerating high ionisation gas is through thomson / compton scattering of the continuum x - ray photons . while line - driving may play a role in large black holes accreting near the eddington limit ( hagino et al .
2014 , in prep ) we concentrate on the latter mechanism here . in comparison to fast line - driven winds in the uv , the comparatively small interaction cross - section in highly ionised material means that high - ionisation winds , such as those considered here , couple much less efficiently with the incident radiation and , as a result , require either much larger source luminosities or higher column densities to achieve equivalent outflow velocities . a general characteristic of these ` continuum - driven winds ' is that they are accelerated by photons scattering off free electrons in the absorbing gas which transfers a portion of the photon momentum to the material , hence causing a wind .
this scenario has been considered extensively in the literature ( e.g. , @xcite ) .
we outline the general theory again here in an effort to search for relevant relationships with @xmath16 which we can relate to the results of our regression analysis .
the efficiency through which the photon momentum of the incident radiation field is transferred into the wind depends upon the plasmas optical depth to electron scattering , @xmath598 , which determines how many times each electron is likely to scatter in the flow .
the total electron scattering optical depth of a homogeneous wind viewed from infinity down to radius @xmath159 is given by @xmath599 where @xmath600 is the thomson cross - section for electron - scattering , @xmath601 is the electron number density in a mass - conserving spherical flow ( c.f .
, [ sub : mass_outflow_rate ] ) and @xmath277 is a geometrical factor which takes into account the global covering of the gas .
normalising to the eddington mass - accretion rate , @xmath602 , then yields @xmath603 as has been shown by @xcite ( see also @xcite for a recent review ) .
if we further assume that the measured flow velocity roughly corresponds to the escape velocity at the radius where the wind was launched , i.e. , @xmath604 , we have @xmath605 which , after substituting for @xmath310 and re - arranging , becomes @xmath606 where @xmath607 , the geometric factor @xmath608 and we have substituted from @xmath609 in the final step to express the relation in terms of observable quantities .
equation ( [ eqn : pout_pbol ] ) therefore shows that a wind accelerated by electron scattering should have an outward momentum - rate which is proportional to that of the incident radiation field .
this result ( and the associated derivation ) is directly equivalent to the one presented by @xcite .
note that while @xmath610 has thus far been described as the optical depth to electron scattering it can in principal also account for additional sources of opacity in the flow , e.g. , though bound - free or bound - bound absorption , which can also be further boosted by multiple scattering events .
a reasonable way of testing for the continuum - driven wind scenario is to therefore see whether @xmath15 and @xmath611 follow a linear trend .
figure [ fig : pbol_pout]a shows that the two are clearly correlated , with a slope of @xmath565 .
this is consistent with the order of unity expected from equation ( [ eqn : pout_pbol ] ) , and is also agreement with the slope of @xmath612 found by @xcite .
figure [ fig : pbol_pout]b is a constraint diagram for @xmath613 showing the deviations from the expected ratio of unity ; in total , most of the sources ( 13/20 ; @xmath614 of the sample ) are formally consistent with the ratio of unity expected for a continuum - driving scenario
. another useful relationship can be obtained by multiplying equation ( [ eqn : pout_pbol ] ) again by @xmath27 : @xmath615
i.e. , the wind kinetic power is predicted to be proportional to the bolometric luminosity of the agn . in figure
[ fig : lbol_raw]d we showed that @xmath14 and @xmath16 are indeed correlated ( table [ tab : linear_regression ] ) , although the slope ( @xmath616 ) appears to be slightly steeper than ( but still marginally consistent with ) that predicted from equation ( [ eqn : lk_lbol ] ) .
the steeper slope here is probably due to the @xmath617 relationship observed before . indeed , taking this into account yields @xmath618 , which is exactly what we observe in these data .
even so , it is worth noting that other acceleration mechanisms are not conclusively ruled out for these winds . if the winds are launched with zero initial radial velocity , then the terminal velocity of the wind in terms of the eddington limit is : @xmath619 , where @xmath620 is the local terminal velocity and @xmath621 is the force multiplier ( @xmath622 for thomson scattering ) . as a result ,
any wind accelerated purely by thomson - scattering will not be able to greatly exceed the local terminal velocity without either : ( a ) additional sources of opacity in the flow ( such that @xmath623 ) , or ( b ) additional mechanisms transferring momenta to the flow . in this case ,
additional opacity in the flow may be attributed to the outflowing gas if it is stratified and contains clumps of lower - ionisation gas .
this extra opacity could perhaps be associated with the partially - covering gas required by many of our models in paper 1 . alternatively
, magnetic processes may play a role in the initial acceleration which provides additional momentum to the flow , thereby allowing it to escape the system .
indeed , have been proposed along these lines in the literature ( e.g. , @xcite .
it is not possible to distinguish between these possibilities on the basis of existing data .
considering again the continuum - driven scenario , an interesting thing of note from figure [ fig : pbol_pout]b is that while all of the fast fek winds are consistent with the unity ratio within the errors , all of the slow winds fall below the limit ( even when only the upper - limits are considered ) .
this possibly reflects subtle changes in the @xmath624 ratio which allows for the possibility of continuum - driven winds in sub - eddginton agn , whilst also allowing for subtle differences in the coupling efficiency of the gas between the two groups . as mentioned in
[ table : measured_wind_properties ] , the agn considered here are relatively tightly clustered around a mean @xmath625 which means that we are unable to reliably test how the eddington ratio affects the @xmath626 ratio .
nevertheless , we can still try to gauge how @xmath598 changes across the sample . from equation ( [ eqn : pout_pbol ] )
we can see that when normalised to eddington @xmath627 , with @xmath628 ( see [ sub : mass_outflow_rate ] ) and therefore that @xmath629 to within an order of magnitude .
this means that @xmath630 , and that the @xmath537 ratio can be in principal be used as a proxy for the gas opacity when the wind was launched .
figure [ fig : pout_tau ] shows how @xmath631 varies with @xmath27 .
the higher velocity systems are largely consistent with @xmath632 ( @xmath633 ) within the errors , with a mean of @xmath634 , which is consistent with what is argued by @xcite .
in contrast , the slow systems ( blue ) all appear to have @xmath537 ratios around an order of magnitude lower
. taken at face value , this possibly indicates that they have a lower opacity to scattering .
there also appears to be a transitionary phase between the two groups , with @xmath635 a few thousand @xmath636 , which may be consistent to the idea of stratified winds launched over a wide range of radii with varying velocities ( see @xcite ) .
it is worth noting , however , that the opacity estimated from @xmath537 could be lower than that of the bulk flow if ( a ) we view the wind from an acute angle where the observed velocity is not representative of the bulk flow velocity , and ( b ) the wind has a component of completely ionised gas which contributes to the scattering opacity but has no spectroscopic signatures .
one possibility is that all of the winds have @xmath637 when integrated over all lines of sight and across all ionisation phases .
alternatively , the fact that the slow winds have @xmath638 , coupled with the fact that they are preferentially observed in lower luminosity agn , may indicate that they are accelerated by another mechanism such as magnetic pressure , for example , which may have a larger contribution at lower luminosities ( e.g. , see @xcite ) regardless , the fast winds which are the ones most likely to affect their host galaxy due to their large velocity and therefore the most interesting in terms of agn host - galaxy feedback are consistent with a wind launched by continuum scattering at the eddington limit ( e.g. , @xcite ) . ) and the radiation field ( @xmath542 ) momentum rates .
_ top _ : a scatter plot showing the clear linear relationship between the two quantities ( see table [ tab : linear_regression ] for parameters ) .
plot components are the same as previous .
_ bottom _ : constraint diagram showing the @xmath626 ratio for each source .
almost all of the agn in the sample are consistent with a ratio of unity ( denoted by the dotted black line ) . , title="fig:",scaledwidth=40.0% ] ) and the radiation field ( @xmath542 ) momentum rates . _
top _ : a scatter plot showing the clear linear relationship between the two quantities ( see table [ tab : linear_regression ] for parameters ) .
plot components are the same as previous .
_ bottom _ : constraint diagram showing the @xmath626 ratio for each source .
almost all of the agn in the sample are consistent with a ratio of unity ( denoted by the dotted black line ) .
, title="fig:",scaledwidth=40.0% ] versus @xmath639 which can be used as a diagnostic for @xmath598 ( see text for details ) .
the majority of fast agn ( @xmath492kms@xmath6 ; red points ) are consistent with @xmath632 , thereby implying that @xmath633 , while all of the slow winds ( blue points ) are well below a ratio of unity.,scaledwidth=40.0% ] the independent _ suzaku_and _ xmm - newton_wind samples have conclusively shown highly ionised fek winds are both frequently observed in agn and that they contain a large amount of mechanical energy .
however , a fundamental question still remains : is the energy imparted by the mass flow likely to play a role in terms of agn host - galaxy feedback scenarios ?
agn feedback models have postulated that @xmath20 of an agns bolometric radiative output needs to be converted to mechanical power in order to have a notable effect on the growth of a central smbh and the host galaxy ( e.g. , @xcite ) .
however , recent numerical simulations by @xcite have shown that should the effects of ` secondary ' feedback such as cloud ablation be taken into account , the required energy can be significantly lower , of the order @xmath640 .
moreover , hydrodynamic simulations have also shown that even relatively modest fek outflows ( e.g. , those with only @xmath641c , @xmath642yr@xmath6 and @xmath643ergs@xmath6 ; see @xcite ) can impart significant feedback upon the host galaxy on @xmath644kpc scales should these effects be taken into account .
investigating what fraction of an agns bolometric output is conveyed through mechanical power therefore provides a means of qualitatively assessing the likely energetic significance of the mass flow . a plot of @xmath645 to this effect
is shown in figure [ fig : lbol_kout ] . to the @xmath16 of each agn in the _
the horizontal lines denote @xmath646 fractions of @xmath647 and @xmath648 which are the theoretical thresholds for feedback .
almost all of the fast absorption systems ( red points ) have @xmath14 in excess of @xmath647 of @xmath16 which suggests that they may be significant in terms of agn host - galaxy feedback scenarios.,scaledwidth=40.0% ] a large fraction of the outflows have a kinetic luminosity either in excess of or consistent with the @xmath647 of @xmath16 , with 10/20 also exceeding @xmath648 as well .
the mean kinetic power considering only the fast winds is @xmath649 .
this is well above the supposed @xmath640 threshold for significant feedback . in the most conservative case ,
i.e. , by considering only @xmath650 given the available errors , we estimate @xmath651 .
this is comparable to the @xmath640 threshold and consistent with what was found by @xcite on the basis of the _ xmm - newton_outflow sample .
therefore , even in the most conservative case , the energetic output of the fast _ suzaku_-detected fek winds is comparable to the threshold thought necessary for feedback .
importantly , these estimates are derived solely from the wind kinematics and are therefore independent of the initial driving mechanism .
the fek winds detected here may therefore impart enough energy into the host galaxy to play a roll in galaxy - scale feedback . with this in mind , it is useful to highlight the recent work of @xcite . to set the context of their work , it is first useful to explain the difference between momentum- and energy - driven outflows .
as a wind propagates into the host galaxy it will sweep up a shell of ambient gas , the dynamics of which are determined by how efficiently the shocked wind material behind the shell can cool .
initially , the shocked gas can cool efficiently ( by compton scattering ) so the shell stays thin and is driven outwards directly by the force of the wind ( cf .
our equation [ eqn : pout_pbol ] ; also @xcite ) .
this is momentum - driven feedback . as the outflow reaches larger distances , however , the cooling efficiency drops and the shocked gas becomes geometrically thick and hot .
the outwards force on the shell then comes from the wind thrust in equation ( [ eqn : pout_pbol ] ) mediated by the thermal ppressure and work done on and by the shocked gas .
this is energy - driven feedback ( e.g. , @xcite ) .
the transition between the two phases can occur at relatively small radii ( see @xcite ) , and energy - driven outflows are therefore expected to be dominant on large galactic scales .
@xcite consider purely energy - driven outflows and investigate their link to the observed @xmath652 relationship .
they showed that , for an energy - driven shell to reach the escape speed from an isothermal sphere of dark matter and gas with velocity dispersion @xmath548 requires a critical bh mass given by @xmath653 such that the velocity of the wind from the black hole ( @xmath27 ) enters explicitly .
mcquillin & mclaughlin used this equation to infer the wind speed that would have been needed during the active protogalactic phase of 51 quiescent ( @xmath654 ) galaxy spheroids ( taken from the sample of @xcite ) in order to explain the observed values of @xmath655 and @xmath548 in these galaxies as a result of ( energy - driven ) feedback .
their main result is a distribution of bh wind speeds , derived in essence from the scatter of the @xmath652 relation , that bears a remarkable resemblance to the empirical distributions of @xmath27 in the agn samples of both @xcite and paper i. relation ( c.f . , @xcite )
is induced by energy - driven feedback , with @xmath656 .
the blue and black arrows correspond to the median outflow velocites of @xmath657c and @xmath658c in the _
suzaku_-observed agn and quiescent sub - samples , respectively .
the two samples are formally indistinguishable by the kolmogorov - smirnov ( ks ) test . see text for further discussion.,scaledwidth=40.0% ] figure [ fig : mcquillin ] compares the velocity distribution inferred by @xcite to the one that we measured for _ suzaku_-detected fek outflows ( note that we consider all of the outflows in this panel ) . as noted by mcquillin & mclaughlin , there is a close resemblance between the two , with both samples tracing a similar overall distribution .
their overall range are similar , as is their median : @xmath659c in the quiescent galaxies , and @xmath660c for the _ suzaku_-detected outflows ( see paper i ) .
formally , the two distributions can not be distinguished : a kolmogorov - smirnov ( ks ) test only rules their being from the same parent population at @xmath661 confidence . the fact that there are similarities between the inferred and observed @xmath27 distributions at all is in itself extremely interesting , not only because the independent samples of quiescent galaxies and agn are drawn from two different sub - populations of galaxy , but because it implies a close relationship between the emergent wind and the evolution of the host galaxy .
overall , the work of @xcite is encouraging for the general notion that high - velocity outflows contribute to the feedback which is invoked to explain the @xmath652 relationship in local quiescent galaxies , and , more importantly , for the idea that agn play a key role in galaxy formation and evolution .
in paper i we formed a detailed spectroscopic study of fek absoption in a heterogeneous sample of 52 _ suzaku_-observed agn .
there , we found that @xmath1 of the sources in the sample harboured statistically significant fexxvi he@xmath24and / or fexxvi ly@xmath24absorption lines with velocities ranging from a few thousand kms@xmath6 to @xmath662c .
we also measured their column density ionisation parameter of the absorber using the xstarphotoionisation code . here , we build upon these results to compute the location @xmath159 , mass outflow rate @xmath13 , kinetic power @xmath14 and momentum flux @xmath15 of the implied wind , and use these results to inform an exploratory analysis of how the wind parameters scale with the bolometric luminosity of their host agn .
we focus our attention on the fastest winds , defined here as @xmath663 , because these ones likely represent genuine disk - winds which are likely distinct from the slower ( and more distant ) warm absorber .
the main results of our study are summarised below . 1 .
the mean radial distance to the fast winds is @xmath664 cm ( typically @xmath4 ) from the black hole .
this corresponds to @xmath665pc , such that the fek absorbers are located much closer to the black hole than the traditional parsec - scale warm absorber
. the mean wind mass outflow rate and mean kinetic power are constrained to @xmath666gs@xmath6 ( @xmath667 ) and @xmath668ergs@xmath6 , respectively .
the average upper limit is @xmath669 , but is still an appreciable @xmath478 of @xmath479 if only the lower limits are considered , whereas the kinetic power is constrained to @xmath670 .
these properties are consistent with wind at a substantial fraction of the eddington limit which carries a large amount of mass into the host galaxy .
the fek absorber column density and ionisation parameter are largely independent of @xmath16 but there is a significant correlation between the wind velocity @xmath27 and @xmath16 , with a slope @xmath553 .
this tentatively indicates that more luminous agn launch winds with a larger observed velocity along the line - of - sight .
the winds in more luminous agn contain more material , and they are subsequently more energetic
. the median slope of the @xmath561 and @xmath562 regressions are formally consistent with unity .
there are no significant correlations in the @xmath572 and @xmath573 planes .
therefore , the winds are located at similar distances in schwarzschild units and contain a similar fraction of @xmath479 . + 5 . the momentum flux of the radiation field ( @xmath542 ) and that of the wind ( @xmath15 ) are strongly correlated , with a slope @xmath565 .
this is quantitatively consistent with what is expected from a continuum - driven wind .
a significant fraction of the sample ( @xmath671 , @xmath672 ) exceed the minimum @xmath18 threshold thought necessary for feedback ( @xcite ) , while @xmath673 ( @xmath19 ) also exceed the less conservative @xmath20 threshold as well ( @xcite ) . in the most conservative case ,
the mean @xmath646 ratio is @xmath674 .
this suggests that the winds may be sufficiently energetic in terms of feedback .
these results enforce those recently obtained with _ xmm - newton_and provide additional evidence in favour of fek absorption being a prevalent feature of the agn x - ray spectrum , and that the ensuing wind may be energetically significant in terms of feedback .
we thank the referee for their thorough reading of the manuscript and useful comments which greatly enhanced the clarity of this analysis .
m. cappi acknowledges support from contracts asi / inaf n.i/037/12/0 and prin inaf 2012 .
f. tombesi acknowledges support for this work by the national aeronautics and space administration ( nasa ) under grant no .
nnx12ah40 g issued through the astrophysics data analysis program , part of the roses 2010 . | ongoing studies with _ xmm - newton_have shown that powerful accretion disc winds , as revealed through highly - ionised fek - shell absorption at @xmath0kev , are present in a significant fraction of active galactic nuclei ( agn ) in the local universe ( @xcite ) . in @xcite
we analysed a sample of 51 _
suzaku_-observed agn and independently detected fek absorption in @xmath1 of the sample , and we measured the properties of the absorbing gas . in this work
we build upon these results to consider the properties of the associated wind . on average , the fast winds ( @xmath2c )
are located @xmath3 cm ( typically @xmath4 ) from their black hole , their mass outflow rates are of the order @xmath5yr@xmath6 or @xmath7 and kinetic power is constrained to @xmath8ergs@xmath6 , equivalent to @xmath9 .
we find a fundamental correlation between the source bolometric luminosity and the wind velocity , with @xmath10 and @xmath11 ( @xmath12 confidence ) , which indicates that more luminous agn tend to harbour faster fek winds . the mass outflow rate @xmath13 , kinetic power @xmath14 and momentum flux @xmath15 of the winds are also consequently correlated with @xmath16 , such that more massive and more energetic winds are present in more luminous agn
. we investigate these properties in the framework of a continuum - driven wind , showing that the observed relationships are broadly consistent with a wind being accelerated by continuum - scattering .
we find that , globally , a significant fraction ( @xmath17 ) of the sample can plausibly exceed the @xmath18 threshold thought necessary for feedback , while @xmath19 may also exceed the less conservative @xmath20 of @xmath16threshold as well .
this suggests that the winds may be energetically significant for agn host - galaxy feedback processes .
[ firstpage ] galaxies : active galaxies : nuclei x - rays : galaxies line : identification |
in the last years , we have seen increased efforts of statistical physicists to tackle stochastic dynamical processes in networks in order to study various phenomena @xcite such as ordering processes , the spreading of epidemics and opinions , synchronization , collective behavior in social networks , stability under perturbations , or avalanche dynamics .
a drastic simplification can be achieved when short cycles in the network , defined by interaction terms , are very rare .
this is the case for locally tree - like graphs such as random regular graphs , erds - rny graphs , and gilbert graphs . for such random graphs with @xmath0 vertices , almost all cycles have length @xmath1 such that their effect is negligible in the thermodynamic limit @xcite .
for static problems , this has been exploited in the so - called cavity method @xcite , where conditional nearest - neighbor probabilities are computed iteratively within the bethe - peierls approximation .
the method was very successfully applied to study for example equilibrium properties of spin glasses @xcite , computationally hard satisfiability problems @xcite , and random matrix ensembles @xcite .
this big success has motivated the generalization of the cavity method to dynamical problems @xcite , which is known as the dynamic cavity method or dynamic belief propagation .
as the number of possible trajectories and , hence , the computational complexity increase exponentially in time , applications have however been quite restricted to either very short times @xcite , oriented graphs @xcite , or unidirectional dynamics with local absorbing states @xcite . in the latter case , one can exploit that vertex trajectories can be parametrized by a few switching times .
another idea has been to neglect temporal correlations completely as in the one - step method @xcite or to retain only some @xmath2 correlations as in the 1-step markov ansatz @xcite . while this works well sometimes for stationary states at high temperatures , such approximations are usually quite severe for short to intermediate times or low temperatures . in this paper , we present an efficient novel algorithm for the solution of the parallel dynamic cavity equations for generic ( locally tree - like ) graphs and generic bidirectional dynamics . the central objects in the dynamic cavity method are conditional probabilities for vertex trajectories of nearest neighbors the so - called edge messages .
as temporal correlations are decaying in time and/or time difference @xmath3 , we can approximate each edge message by a matrix product , i.e. , there is one matrix for every edge , edge state , and time step , encoding the temporal correlations in the corresponding part of the evolution .
it turns out that the dimensions of these matrices do not have to be increased exponentially in time .
one can obtain quasi - exact results with much smaller matrix dimensions .
computation costs and precision can be tuned by controlling the dimensions in truncations .
the idea of exploiting the decay of temporal correlations to approximate edge messages in matrix product form is in analogy with the use of matrix product states @xcite for the simulation of strongly correlated , mostly one - dimensional , quantum many - body systems .
these have been used very successfully in algorithms like the density - matrix renormalization group @xcite to compute for example quantum ground - state properties , often with machine precision @xcite . besides lifting the restrictions of the aforementioned approaches , the matrix product edge - message ( mpem ) algorithm can also outperform monte carlo simulations ( mc ) of the dynamics in important respects . in particular , besides allowing for the simulation of single instances , alternatively , one can work directly in the thermodynamic limit .
perhaps more importantly , it has a favorable error scaling .
while statistical errors in mc decay very slowly with the number of samples @xmath4 as @xmath5 , mpem yields also observables with absolutely small expectation values with very good precision which is essential for the study of decay processes and temporal correlations .
[ fig : msgsubgraph ] .,title="fig : " ] let @xmath6 denote the state of vertex @xmath7 at time step @xmath8 , and @xmath9 the state of the full system at time @xmath8 . given the state probabilities @xmath10 for time @xmath8 , we evolve to the next time step , @xmath11 , by applying the stochastic transition matrix @xmath12 . as vertex @xmath7 interacts only with its nearest neighbors @xmath13 , the probability for @xmath14 only depends on the states @xmath15 of these vertices at the previous time step such that the global transition matrix @xmath12 is a product of local transition matrices @xmath16 , @xmath17 here @xmath18 , and @xmath19 is the state of the nearest neighbors of vertex @xmath7 at time @xmath8 . in the cavity method @xcite , one neglects cycles of the ( locally tree - like ) graph according to the bethe - peierls approximation to reduce this computationally complex evolution to the dynamic cavity equation @xcite @xmath20\\ \times \big[\prod_{k\in\partial i\setminus\{j\ } } \mu_{k\to i}({\bar{\sigma}}_k^{t}|{\bar{\sigma}}_i^{t-1})\big]\end{gathered}\ ] ] which only involves the so - called edge messages @xmath21 for the edges of a single vertex @xmath7 . for simplicity
, we have assumed that vertices are uncorrelated in the initial state such that @xmath22 .
the edge messages @xmath23 in the dynamic cavity equation are conditional probabilities for the trajectories @xmath24 and @xmath25 on edge @xmath26 .
specifically , if we consider a tree graph and cut off everything `` right '' of vertex @xmath27 as indicated in figure [ fig : msgsubgraph ] by the dashed line , @xmath23 denotes the conditional probability of a trajectory @xmath28 on vertex @xmath7 , given the trajectory @xmath29 on vertex @xmath27 .
. constructs @xmath30 out of the edge messages @xmath31 of the previous time step .
this is exact for tree graphs and covers locally tree - like graphs in the bethe - peierls approximation .
although we have gained a lot in the sense that the computational complexity is now linear in the system size , it is still exponential in time @xmath8 , if we were to encode the edge messages without any approximation .
to circumvent this exponential increase of computation costs , we can exploit the decay of temporal correlations and approximate the exact edge message by a matrix product @xmath32\\ \times a^{(t)}_{i\to j}(\sigma_i^{t-1})a^{(t+1)}_{i\to j}(\sigma_i^{t}).\end{gathered}\ ] ] the particular choice of assigning vertex variables @xmath33 and @xmath34 to the @xmath35 matrices @xmath36 occurring in the matrix product , is advantageous for the implementation of the recursion relation for mpems as will become clear in the following . in order for the matrix product to yield a scalar ,
we set @xmath37 .
[ fig : mpem ] the time - evolution starts at @xmath38 with @xmath39 . using the dynamic cavity equation , we iteratively build matrix product approximations for edge messages for time @xmath40 from those for time @xmath8 .
it is simple to insert the matrix product ansatz for the edge messages in the dynamic cavity equation , but not trivial to bring the resulting edge message again into the canonical mpem form as required for the subsequent evolution step . the specific assignment of the vertex variables to matrices in eq .
has been chosen such that all contractions ( products and sums over vertex variables ) occurring in the cavity equation are time - local in the sense that , given mpems @xmath31 in canonical form for all neighbors @xmath41 , the resulting @xmath42 can be written in ( non - canonical ) matrix product form as @xmath43.\ ] ] as depicted in figure [ fig : mpem]b , the tensors @xmath44 for @xmath45 are obtained by contracting a local transition matrix @xmath46 with tensors @xmath47 from the time-@xmath8 mpems .
this contraction entails a sum over the @xmath48 common indices @xmath49 , where @xmath50 is the vertex degree .
assuming for the simplicity of notation that the matrix dimensions @xmath51 for all time-@xmath8 mpems are identical , the resulting matrices @xmath52 $ ] have left and right indices of dimensions @xmath53 and @xmath54 , respectively .
the contraction for tensor @xmath55 is very similar and @xmath56 $ ] .
in preparation for the next time step , we now need to bring the evolved edge message back into canonical form .
furthermore , we need to introduce a controlled approximation that reduces the matrix dimensions because they would otherwise grow exponentially in time . both , a reordering of the vertex variables @xmath57 and @xmath58 in the matrix product and a controlled truncation of matrix dimensions can be achieved by sweeping through the matrix product and doing certain singular value decompositions ( svd ) @xcite of the tensors .
let us shortly explain the notion of truncations at the example of a matrix product @xmath59 , where @xmath60 is an @xmath35 matrix and @xmath61 . in order to reduce in a controlled way , e.g. , the left matrix dimension @xmath62 of @xmath63
, we suggest to do an svd of the matrix product such that @xmath64 where we have grouped the variables to @xmath65 and @xmath66 . @xmath67 and @xmath68 are isometric matrices ; @xmath69 and @xmath70 .
now , truncating some of the singular values @xmath71 , such that only the @xmath72 largest are retained , we obtain the controlled approximation @xmath73 note that this truncation scheme guarantees the minimum possible two - norm loss @xmath74 for the given new matrix dimension @xmath72 .
while it seems very desirable to discard unimportant information and control the growth of computation cost through such truncations , the svd appears to be an insurmountable task . assuming that each variable @xmath75 can take @xmath76 different values and that @xmath77 , the cost for the svd would scale exponentially in time like @xmath78 @xcite .
however , the beauty of matrix products is that such an svd can in fact be done sequentially with linear costs of order @xmath79 ( @xmath80 ) as follows .
first , we do an exact transformation of the matrix product to bring it to the form @xmath81 , where tensors @xmath82 and @xmath83 obey the left and right orthonormality constraints @xmath84 respectively .
this is achieved through a sequence of svds .
it starts with the svd @xmath85 , where @xmath86 is a diagonal matrix of singular values , @xmath87 is isometric according to @xmath88 and @xmath89 obeys eq . .
the sweep continues with the svd @xmath90 and so on until the computation of @xmath91 .
analogously , we do a second sequence of svds starting from the right and ending with @xmath92 and , finally , define the central tensor as @xmath93 .
after this somewhat laborious preparation , we can do the actual truncation , based on the svd @xmath94 with the same singular values @xmath95 as in eq . .
with the @xmath96 matrix @xmath97_{kk'}:=\delta_{kk'}\lambda_k$ ] , the truncated matrix product takes the form @xmath98 [ fig : glauber ] with this tool in hand , we can now truncate the evolved edge message and bring it back into canonical form . in a first sweep from right ( @xmath99 ) to left ( @xmath100 ) , using svds , we can sequentially impose the right orthonormality constraints [ see eq . ] on the @xmath101-tensors . in a subsequent sweep from left to right , again based on svds , we can now truncate the tensors to decrease bond dimensions from @xmath102 to something smaller .
what is left , is to reorder the indices @xmath103 and @xmath104 of the vertex variables . in a sweep from right to left
, we go from the variable assignment @xmath105 in the truncated and orthonormalized version @xmath106 of the mpem to the assignment @xmath107 in the matrix product @xmath108 d^{(t+1)}_{i\to j}(\sigma_i^{t+1}).\ ] ] at the right boundary , we start with an svd and controlled truncation @xmath109 , and continue with @xmath110 and so on until ending at @xmath100 . in an analogous final sweep from left to right , we change to the canonical variable assignment @xmath111 as in eq . .
after executing these steps for all edge messages , the next evolution step from @xmath40 to @xmath112 can follow .
the joint probability of trajectories @xmath113 and @xmath29 for the vertices of an edge @xmath26 is given by the product of the two corresponding edge messages .
after marginalization , one obtains for example the probability for the edge state @xmath114 at time @xmath8 as @xmath115 in the mpem approach , this can be evaluated efficiently , as indicated in figure [ fig : mpem]c , by executing the contractions sequentially from left ( @xmath100 ) to right ( @xmath116 ) .
similarly , one can for example also compute temporal correlators @xmath117 from probabilities @xmath118 .
figure [ fig : glauber ] compares the simulation of glauber dynamics @xcite using our mpem algorithm to mc simulations and to the 1-step markov approximation @xcite .
specifically , we have ising spins interacting ferromagnetically on @xmath119 random regular graphs , with local transition matrices @xmath120 . in the initial state ,
all spins have magnetization @xmath121 , i.e. , @xmath122 .
besides being applicable for single instances of finite graphs , the mpem approach gives also direct access to the thermodynamic limit .
for disordered systems this can be done in a population dynamics scheme . the homogeneous case , considered here , is particularly simple as all edges of the graph are equivalent in the thermodynamic limit .
hence , one can work with a single mpem .
figure [ fig : glauber]a shows the evolution of the magnetization . in the ferromagnetic phase ( @xmath123 ) , it approaches a finite equilibrium value , whereas it decays to zero in the paramagnetic phase ( @xmath124 ) . as shown for @xmath123 , mc simulations contain finite size effects which become small for the system with @xmath125 sites .
mc errors decrease slowly when increasing the number of samples @xmath4 as @xmath5 .
this is problematic for observables with small absolute values where cancellation effects make it difficult to get a precise estimate .
this is , e.g. , apparent in the magnetization decay for @xmath124 which , in contrast , is very precisely captured with mpem . in these simulations
, we control the mpem precision by keeping only singular values @xmath126 above a threshold , specified by @xmath127 .
decreasing @xmath128 , increases precision and computation costs .
the 1-step markov approximation @xcite is not suited to handle temporal correlations . at long times it performs well for @xmath124 and fairly good for @xmath123 , but deviates rather strongly at earlier times .
figure [ fig : glauber]a shows the connected temporal correlation function @xmath129 for @xmath123 as a function of @xmath130 for several times .
its decay behavior can be difficult to impossible to capture with mc . in the example , mc deviations are often orders of magnitude above those of the numerically cheaper mpem simulations .
the novel mpem algorithm , based on matrix product approximations of edge messages allows for an efficient and precise solution of the dynamic cavity equations .
besides lifting restrictions of earlier approaches , mentioned in the introduction , it gives direct access to the thermodynamic limit , and its error scaling is favorable to that of mc simulations .
we think that it is a very valuable tool , particularly , as it yields temporal correlations and other decaying observables with unprecedented precision and gives access to low - probability events .
this opens a new door for the study of diverse dynamic processes and inference or dynamic optimization problems for physical , technological , biological , and social networks . | we describe and demonstrate an algorithm for the efficient simulation of generic stochastic dynamics of classical degrees of freedom defined on the vertices of a locally tree - like graph .
networks with cycles are treated in the framework of the cavity method .
such models correspond for example to spin - glass systems , boolean networks , neural networks , or other technological , biological , and social networks .
building upon ideas from quantum many - body theory , the algorithm is based on a matrix product approximation of the so - called edge messages conditional probabilities of vertex variable trajectories .
the matrix product edge messages ( mpem ) are constructed recursively .
computation costs and precision can be tuned by controlling the matrix dimensions of the mpem in truncations .
in contrast to monte carlo simulations , the approach has a better error scaling and works for both , single instances as well as the thermodynamic limit . as we demonstrate at the example of glauber dynamics , due to the absence of cancellation effects , observables with small expectation values can be evaluated reliably , allowing for the study of decay processes and temporal correlations . |
in several local analysis problems arising in mathematical physics , control theory , dynamical systems e.t.c .
one is led to consider the classification problem for pairs @xmath0 , where @xmath1 is a germ of a closed 2-form on a manifold @xmath5 and @xmath2 is a function germ , with or without singularities .
the most studied case is when the 2-form is nondegenerate , i.e. it defines a symplectic structure on @xmath5 .
then @xmath2 can be viewed as a hamiltonian function and the classification problem reduces to the well known problem of symplectic classification of singularities of functions ( c.f .
@xcite , @xcite , @xcite ) . in this direction ,
the 2-dimensional problem is drastically different than the higher dimensional one .
indeed , a symplectic structure on the plane is an area ( volume ) form and the classification of hamiltonian functions on the plane becomes a problem of volume - preserving geometry .
there , analytic normal forms exist in abundance ( c.f .
@xcite , @xcite , @xcite as well as @xcite , @xcite , @xcite , @xcite for related results ) in contrast to the symplectic category @xcite . in this paper
we consider a generalisation of the classification problem for pairs @xmath0 on a 2-manifold @xmath5 , where now the 2-form @xmath1 is allowed to have singularities and thus it does not define a symplectic structure everywhere on @xmath5 .
this situation is typical when we consider hamiltonian systems with constraints ( c.f .
@xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite ) . in analogy with the unconstrained case
, we may define a constrained hamiltonian system ( chs ) on a 2-manifold @xmath5 , simply as a pair @xmath0 consisting of a function @xmath2 and a 2-form @xmath1 on @xmath5 as above .
one may obtain such a system as the restriction of a hamiltonian system in a general ambient symplectic manifold , on a 2-dimensional surface representing the constraints .
let @xmath6 be the `` hamiltonian vector field '' associated to the pair @xmath0 through the equation : @xmath7 this vector field is in general not defined and smooth everywhere on @xmath5 .
the obstruction to the existence and/or uniqueness of @xmath6 is obviously the set of zeros @xmath3 of the 2-form @xmath1 . in the theory of singularities of constraints systems ( c.f .
@xcite , @xcite ) it is called the _ impasse hypersurface _ , while in the theory of differential systems is usually called the _ martinet hypersurface _ ( c.f .
@xcite ) , in honor of j. martinet who was the first who studied systematically singularities of differential forms @xcite .
the problem is thus to classify chs at impasse points ( away from the impasse points the problem reduces to the ordinary symplectic classification of functions ) .
it is easy to see that the germ of a generic singular 2-form @xmath1 on the plane can be reduced to martinet normal form @xcite : @xmath8 the geometric invariants of the 2-form @xmath1 on the plane are just its martinet curve of zeros @xmath9 , along with an orientation ( in the real analytic ( smooth ) case ) induced by the two symplectic structures in its complement .
the orientation of the martinet curve plays no role in the initial definition of singularity classes for the pair @xmath10 and thus , all the singularities are defined by the relative positions of the germ @xmath2 with the martinet curve . in particular , the pair @xmath11 can be viewed as defining a germ of a `` boundary singularity '' at the origin of the plane i.e. such that : * either @xmath2 has an isolated critical point at the origin , * or @xmath2 is non - singular but its restriction @xmath12 on the martinet curve has an isolated critical point at the origin .
thus , in order to study the singularities of pairs @xmath0 , one may fix an arbitrary boundary singularity @xmath13 and study possible normal forms of degenerate 2-forms @xmath1 , whose zero set @xmath3 is exactly the curve @xmath14 , under the action of the pseudogroup @xmath15 of diffeomorphisms preserving the boundary singularity .
the study of isolated boundary singularities has been initiated by v. i. arnold in @xcite ( see also @xcite , @xcite , @xcite for general references ) where he extended the @xmath16 , @xmath17 , @xmath18 classification of simple singularities to include also the @xmath19 , @xmath20 , @xmath21 series of weyl groups in the scheme of singularity theory . the list of simple normal forms obtained by arnold is given for convenience in table [ tab:1 ] .
llll @xmath22 & @xmath23 & @xmath24 & @xmath25 + @xmath26 & @xmath27 & @xmath28 & @xmath29 + @xmath30 & @xmath31 & @xmath31 & @xmath32 + the number @xmath33 in the list is an important invariant .
it is called the milnor number , or the multiplicity of the boundary singularity and it is intimately related with the classification problem of pairs @xmath0 we will study , as is the ordinary milnor number of an isolated singularity in the symplectic case ( c.f . @xcite and also @xcite ) .
the manifestation of the milnor number in symplectic ( isochore in higher dimensions ) classification problems in the ordinary ( without boundary ) case , is due to the following fact : the corresponding deformation space of a germ of a symplectic form @xmath1 on the plane ( relative to diffeomorphisms preserving the germ @xmath2 ) is exactly equal to the brieskorn module of the singularity @xmath2 and in particular , according to the brieskorn - deligne - sebastiani theorem @xcite , @xcite , any such deformation space will be a free module of rank equal to the milnor number of @xmath2 , over the ring @xmath34 of analytic functions on ( the values of ) @xmath2 .
thus , the classification problem reduces to a problem of relative de - rham cohomology .
the classification problem in the martinet case which we will study here , is not much different than the symplectic classification problem , despite the fact that the corresponding relative de - rham cohomology theory for isolated boundary singularities has not yet been established in the literature .
in particular one may conjecture , in analogy with the ordinary case , that the number of functional invariants in the classification of pairs @xmath0 is exactly equal to the milnor number of the boundary singularity @xmath4 and moreover , there exists a normal form for the pair involving exactly these invariants ( which by the way will be analytic functions of 1-variable ) . here
we will prove this statement only for the special case where the pair @xmath4 is a quasihomogeneous boundary singularity ( theorem [ t2 ] ) and in particular for any of the simple germs in arnold s list in table [ tab:1 ] ( as well as for other unimodal , e.t.c .
boundary singularities which are still quasihomogeneous c.f .
@xcite , @xcite and @xcite for the corresponding lists ) .
our proof for the quasihomogeneous case is very similar ( almost identical in some parts ) to the one proposed by j. p. francoise in @xcite , @xcite for the ordinary case .
while not that general , in order to include all isolated boundary singularities ( i.e. away from arnold s list ) , francoise s method has the advantage that it is algorithmic in nature : one may construct step by step the characteristic invariants of the pair @xmath0 as well as the bounds of the corresponding norms .
also it has minimum prerequisites , such as the relative poincar lemma and the relative de - rham s division lemma ( see section 2 ) , while a general proof should go through several sheaf theoretic techniques .
thus , a main part of the paper is to prove ( theorem [ t1 ] ) that the corresponding deformation module of a martinet 2-form @xmath1 ( relative to diffeomorphisms preserving a simple boundary singularity @xmath4 ) is a free @xmath34-module of rank equal to the milnor number of the boundary singularity , i.e. equal to the corresponding number @xmath33 in table [ tab:1 ] . one final remark : the classification of functions @xmath2 and martinet 2-forms @xmath1 in higher dimensions is a much more complicated problem and the results of this paper do not generalise in this case .
probably , formal normal forms do exist but we do nt discuss this here . instead , we give an application of the 2-dimensional results in a problem arising in the geometric theory of hamiltonian systems with constraints , that is , the problem of classification of generic singular lagrangians ( of first order in the velocities ) on the plane , under variational ( gauge ) equivalence ( theorem [ tl ] ) .
such lagrangians , already studied by dirac @xcite for the purposes of quantisation and which appear in high energy physics ( c.f .
@xcite , @xcite , @xcite ) , in hydrodynamics and general vortex theory ( c.f .
@xcite , @xcite and references therein ) and also in control theory and sub - riemannian geometry @xcite ,
@xcite , @xcite and several other instances , give rise to euler - lagrange equations which define a constrained hamiltonian system @xmath0 and thus , they are subjectable to the analysis in this paper .
we fix a coordinate system @xmath35 at the origin of @xmath36 such that the boundary is given by the equation @xmath37 . to a boundary singularity @xmath13 we associate its local algebra ( c.f .
@xcite ) @xmath38 where the ideal @xmath39 in the denominator is the tangent space to the @xmath40-orbit of @xmath2 , i.e. under diffeomorphisms ( right - equivalences ) preserving the boundary @xmath14 ( as usual @xmath41 is the algebra of germs of analytic functions at the origin ) .
we call it , in analogy with the ordinary case the _ jacobian ideal of the boundary singularity @xmath13_. its codimension , i.e. the @xmath42-dimension @xmath33 of the vector space @xmath43 is called the _ multiplicity _ or _ milnor number of the boundary singularity _ @xmath13 and it is an important invariant : it is related to the milnor number @xmath44 of @xmath2 : @xmath45 and the milnor number @xmath46 of its restriction @xmath47 on the boundary : @xmath48 by the formula : @xmath49 topologically it can be interpreted as the rank of the relative homology group @xmath50 of the pair @xmath51 , @xmath52 , of smooth milnor fibers of @xmath2 and @xmath53 , as one may easily deduce from the exact homology sequence induced by the inclusion @xmath54 ( the fiber @xmath55 consists of a finite number of points ) . in particular , according to a theorem of arnold @xcite which generalises milnor s theorem @xcite for the boundary case , the space @xmath56 has the homotopy type of a bouquet of @xmath33 circles .
it follows that the pair @xmath51 is a riemann surface with betti number @xmath44 and @xmath57 distinguished points . for @xmath58 , the @xmath44 cycles of @xmath59 and the @xmath46 segments joining the distinguished points shrink at the origin :
they are called vanishing cycles and half - cycles respectively and they form a basis of @xmath50 ( see @xcite ) .
denote now by @xmath60 , the complex of germs of analytic differential forms at the origin ( where @xmath61 the algebra of analytic functions ) and by @xmath62 the subcomplex of forms that `` vanish on @xmath37 '' , in the sense that their coefficients belong in the ideal @xmath63 generated by the equation of @xmath14 .
notice that any form vanishing on @xmath14 vanishes automatically when evaluated at tangent vectors of @xmath14 ( i.e. it has zero pull back by the embedding @xmath64 ) , but the converse does not hold ( take for example the 1-form @xmath65 ) .
we distinguish by writing @xmath66 for the space of @xmath67-forms with zero pull - back on @xmath14 .
notice that with this notation @xmath68 , @xmath69 ( identically ) and @xmath70 .
the space @xmath71 may be identified with the space of 2-forms whose zero set contains the curve @xmath37 .
we write @xmath72 for the space of martinet 2-forms ( with zero set exactly equal to @xmath14 ) .
we will need the following local version of a type of relative poincar lemma for the complex @xmath73 ( c.f .
@xcite , @xcite ) : [ l1 ] for any closed @xmath67-form @xmath74 there exists an @xmath75-form @xmath76 such that @xmath77 . by the classical poincar lemma the 1-parameter family of maps @xmath78 , @xmath79 $ ] , is a contraction at the origin , it preserves @xmath14 , @xmath80 and is such that : @xmath81 , @xmath82 and @xmath83 , where @xmath84 is defined by @xmath85 and the vector field @xmath86 is defined as the generator of @xmath87 .
notice that by definition @xmath88 is tangent to @xmath14 for all @xmath89 .
now , since @xmath90 vanishes on @xmath14 to second order , the @xmath75-form @xmath91 vanishes also on @xmath14 and since @xmath88 is tangent to @xmath14 it follows that @xmath92 . by the fact that @xmath93 , it follows that @xmath76 .
fix now a pair @xmath13 , where @xmath2 has an isolated singular point at the origin of finite multiplicity @xmath94 . the differential @xmath95 defines an ideal in the algebra @xmath60 of germs of differential forms at the origin , which induces an ideal in all the subalgebras @xmath62 , and @xmath96 we consider .
the lemma below gives necessary and sufficient conditions for the ideal membership problem .
it is an analog in the relative case of de - rham s division lemma @xcite ( see also @xcite ) .
[ l2 ] * let @xmath97 .
then @xmath98 holds for some @xmath99 if and only if @xmath100 .
* for any 1-form @xmath101 such that @xmath102 , there exists a function germ @xmath103 such that @xmath104 .
* the proof is an obvious calculation * it suffices to show that the relative de - rham s division lemma is true for the complex @xmath105 : indeed , if we write @xmath106 for some @xmath107 , then @xmath108 and since @xmath109 is a non - zero divisor it will follow from the relative de - rham division lemma in @xmath105 that there exists a function @xmath110 such that @xmath111 .
but then @xmath112 and the germ @xmath113 as we wanted . to show now that the relative division lemma is true in @xmath114 it suffices to notice that the condition @xmath115 for any 1-form @xmath116 forces it again to be of the form @xmath117 for some @xmath118-form @xmath119
. indeed , this follows from restriction @xmath120 and the fact that @xmath121 ( because @xmath2 has isolated singularities ) .
then , as before , the equation @xmath122 implies that @xmath123 and thus , by the ordinary de rham division lemma , there exists a function germ @xmath124 such that @xmath125 .
thus @xmath126 and the lemma is proved .
it is important to notice here that the condition @xmath115 implied on a 1-form @xmath90 vanishing on @xmath14 ( i.e. for @xmath116 or @xmath101 ) prevents it from being closed ( = exact ) .
indeed , @xmath127 means that the function @xmath128 is constant on the fibers of @xmath2 and thus it is a holomorphic function on the values of @xmath2 : @xmath129 .
but then , since @xmath128 should vanish on @xmath130 it means that it should vanish everywhere in the neighborhood under consideration ( because the fibers of @xmath2 are almost everywhere transversal to @xmath130 ) .
by the extension of tougeron s theorem on the finite determinacy of boundary singularities @xmath13 proved by v. i. matov @xcite we may suppose that @xmath37 and @xmath2 is polynomial of sufficiently high degree ( @xmath131 ) .
write @xmath15 for the pseudogroup of symmetries of the pair @xmath13 . in the following lemma
we identify the set of ( infinitesimal ) trivial deformations of martinet 2-forms relative to @xmath15-action : [ l3 ] let @xmath1 be a germ of a martinet 2-form .
the tangent space to the orbit of @xmath1 under the @xmath15-action consists of all 2-forms of the form : @xmath132 let @xmath133 be an element of the lie algebra @xmath134 .
the infinitesimal deformation of @xmath1 associated to @xmath133 is by definition an element of the form @xmath135 ( where @xmath136 is the lie derivative ) .
we have that @xmath137 and thus , by de - rham s division with @xmath95 , there exists a function germ @xmath138 such that @xmath139 . since the 1-form @xmath140 vanishes on @xmath14 to second order ( because both @xmath133 and @xmath1 vanish on @xmath14 ) we conclude that @xmath103 .
it follows that @xmath141 , @xmath103 .
conversely , let @xmath103 be such that there exists a vector field @xmath133 with @xmath142 . in fact define @xmath133 as the dual of the 1-form @xmath143 through @xmath1 : @xmath144 ( this is possible because @xmath138 vanishes on @xmath14 ) .
obviously @xmath145 and @xmath133 is vanishes on @xmath14 since @xmath138 vanishes on @xmath14 to second order .
thus @xmath146 and the lemma is proved .
it follows that the quotient space @xmath147 consists of the nontrivial infinitesimal deformations of the martinet 2-form @xmath1 relative to the symmetries of the boundary singularity .
along with the @xmath42-linear space structure , the deformation space @xmath148 ( which we will denote simply by @xmath149 ) has a natural @xmath34-module structure with multiplication by @xmath2 .
we call it the _ deformation module of the martinet germ @xmath1_. in the next section we will show that this module is a free module of rank @xmath94 over @xmath34 .
this statement is analogous to the classical brieskorn - deligne - sebastiani theorem @xcite , @xcite for ordinary singularities ( i.e. without boundary ) .
this finiteness result along with the following proposition are cornerstones in the classification problem . [ p1 ]
fix a boundary singularity @xmath13 .
let @xmath1 and @xmath150 be two germs of martinet 2-forms at the origin , such that @xmath151 .
then there exists a diffeomorphism @xmath152 such that @xmath153 .
the proof is by the homotopy method .
consider a 1-parameter family of martinet 2-forms connecting @xmath1 and @xmath150 : @xmath154 so that @xmath155 and @xmath156 .
we seek a 1-parameter family of vector fields @xmath157 such that @xmath158 for all @xmath79 $ ] .
choose @xmath159 by @xmath160 .
it preserves @xmath161 and it is also in @xmath134 by the same reasoning as in the previous lemma .
it follows that the time 1-map of the flow @xmath162 of @xmath159 sends @xmath163 to @xmath164 .
the classical brieskorn - deligne - sebastiani theorem for an isolated singularity @xmath2 @xcite , @xcite and also @xcite , states that the quotient space @xmath165 is a free @xmath34-module of rank equal to the milnor number of the singularity @xmath2 .
it is proved using sheaf theoretic techniques and the properties of the gauss - manin connection on the vanishing cohomology bundle associated to the singularity .
for some special cases of singularities though , other simpler proofs exist ( c.f .
@xcite , @xcite for the morse case ) . for the class of all quasihomogeneous singularities a proof was given by j. p. franoise in @xcite
( see also @xcite ) .
franoise s proof , while not covering the whole class of isolated singularities , it has the advantage that it is algorithmic , i.e. one may construct term by term the coefficients in the expansion of a form in the brieskorn module as well as their bounds .
franoise s algorithm is rather general and it holds also in the global , polynomial case ( c.f .
@xcite ) . here
we will provide the necessary modifications to include the boundary case as well .
in particular we will prove : [ t1 ] let @xmath13 be a quasihomogeneous boundary singularity at the origin of milnor number @xmath33 and let @xmath1 be a germ of a martinet 2-form whose zero set is exactly the curve @xmath14 .
then the deformation module of @xmath1 is a free module of rank @xmath33 over @xmath34 : @xmath166 to prove the theorem , suppose that @xmath2 is a quasihomogeneous polynomial of type @xmath167 , @xmath168 , i.e. such that @xmath169 .
denote by @xmath170 the euler vector field of @xmath2 , i.e. such that @xmath171 .
write also @xmath172
. then the following division lemma holds : [ divlem ] if @xmath13 is a quasihomogeneous boundary singularity at the origin of the plane , then the following identity holds : @xmath173 it suffices to find , for a given 1-form vanishing on the boundary @xmath99 , a 2-form @xmath174 and a function @xmath175 vanishing on the boundary to second order , such that @xmath176 but @xmath177 and so the equality above reduces to @xmath178 i.e. to @xmath179 taking exterior differential , it suffices to find @xmath180 such that : @xmath181 we view @xmath182 as an operator in formal series : @xmath183 this is obviously an invertible operator since for any monomial @xmath184 , @xmath185 , @xmath186 we have : @xmath187 where @xmath188 never vanishes .
thus given @xmath180 we can find a formal solution @xmath189 .
this solution can easily be extended to an analytic solution @xmath99 in a fundamental system of neighborhoods of the origin ( see lemma [ rdl ] below ) . the lemma is proved .
now we describe the analog of franoise s algorithm for the generation of the coefficients in the decomposition of @xmath1 in the deformation module .
first we construct a formal basis of the @xmath190$]-module @xmath191 ( i.e. of the formal deformation module ) : choose a monomial basis @xmath192 of the local algebra @xmath43 of the boundary singularity and lift it to a basis of monomial 2-forms @xmath193 in @xmath194 .
any 2-form @xmath97 can be written as @xmath195 for some 2-form @xmath196 .
decompose now @xmath197 in @xmath198 : @xmath199 where @xmath200 is a 1-form vanishing on @xmath14 ( and defined uniquely by @xmath1 modulo terms of the form @xmath143 ) and @xmath201 for @xmath202 .
this decomposition induces also a decomposition , after multiplication with the function @xmath109 , in the space @xmath203 , in the sense that : @xmath204 write now @xmath205 and decompose the 2-form @xmath206 according to the division lemma [ divlem ] and plug it to equation ( [ de11 ] ) above : @xmath207 where @xmath208 and @xmath209 . continuing that way , decompose @xmath210 in @xmath211 : @xmath212 where again @xmath213 ( is defined by @xmath210 modulo terms @xmath143 ) and the @xmath214 are constants .
multiplying by @xmath109 and plugging this back to equation ( [ de22 ] ) we obtain the new decomposition : @xmath215 where @xmath216 .
now use again the division lemma [ divlem ] for the 2-form @xmath217 to obtain new @xmath218 , @xmath219 such that : @xmath220 and plug it back to ( [ de33 ] ) above to get : @xmath221 continuing that way with the 2-form @xmath222 , e.t.c .
we obtain at the @xmath223-th iterate a decomposition of the form : @xmath224 the term @xmath225 belongs , for @xmath226 , to the intersection of all maximal ideals @xmath227 , i.e. it goes to zero in the @xmath190$]-module @xmath228 .
thus , the algorithm converges in the krull topology . for convenience in notation , we change coordinates from @xmath35 to @xmath229 ( so that @xmath230 ) .
we write in this notation @xmath231 for a monomial @xmath184 and a vector @xmath185 in @xmath232 .
let @xmath233 and let @xmath234 be a polycylinder in @xmath36 .
we consider the pseudo - norm @xmath235 in @xmath41 defined by : @xmath236 where @xmath237 is an analytic function .
we denote by @xmath238 the subset of @xmath41 for which the pseudo - norm @xmath235 is finite and thus defines a norm . for @xmath239
we have accordingly : @xmath240 identify now @xmath241 with @xmath238 and @xmath242 with @xmath243 .
obviously @xmath244 can be identified with the subspace @xmath245 and @xmath246 with @xmath247 .
the map @xmath248 is a @xmath41-linear map and induces a map : @xmath249 a section of @xmath250 is a @xmath42-linear map : @xmath251 such that @xmath252 .
we can see from the definition that @xmath253 is division with @xmath95 .
we say that a section @xmath254 is adapted to the polydisc @xmath255 ( or to @xmath256 ) if @xmath254 is a continuous mapping between banach spaces , i.e. there exists a constant @xmath257 such that : @xmath258 for all @xmath259 .
a consequence of malgrange s priviledged neighborhoods theorem is the following : given @xmath250 there is a section @xmath254 such that the set of polydiscs @xmath255 onto which is adapted , forms a fundamental system of neighborhoods of the origin .
now we continue the proof by making precise the choise of the section .
we will use the following two lemmata , whose proofs are exactly the same as in @xcite .
the first concerns the bounds obtained by division with @xmath95 and the second the corresponding bounds obtained by the relative poincar lemma .
[ rdl ] if @xmath259 is such that @xmath260 , then : @xmath261 where @xmath262 and @xmath263 .
it is exactly the same as in @xcite , lemma 3.1.2 .
for completeness , write @xmath264 with @xmath265 , @xmath266 , where , since @xmath267 vanishes on @xmath230 , the vector @xmath268 in @xmath269 is of the form @xmath270 , @xmath271 . then by direct computation : @xmath272 where @xmath273 , @xmath274 are unit vectors in @xmath232 . from this
the lemma follows .
[ rpl ] for any closed 1-form @xmath275 , there exists a function @xmath276 such that : @xmath277 where @xmath278 .
it follows by the the proof of the relative poincar lemma [ l1 ] .
indeed , the function @xmath279 is defined by @xmath280 where @xmath281 is a contraction mapping at the origin and @xmath282 is the euler vector field : @xmath283 .
the bound on the norm follows then by a direct computation .
now , the finiteness of the deformation module in the analytic category may be stated in the following form : let @xmath255 be a polycylinder onto which the section @xmath254 is adapted .
then there exists a smaller polycylinder @xmath284 such that for any @xmath285 there exist an analytic function @xmath286 and @xmath33 analytic functions @xmath287 such that : @xmath288 with the following explicit bounds : @xmath289 @xmath290 the proof is again the same as in the ordinary case @xcite ( theorem 3.2 ) .
we present it for completeness .
start with the first decomposition @xmath291 with the bound @xmath292 .
then we solve the equation @xmath293 and we find a @xmath180 such that , according to lemma [ rdl ] : @xmath294 then by the relation @xmath295 and lemma [ rpl ] we obtain the bounds : @xmath296 decomposing this way we obtain at the @xmath223-th iterate : @xmath297 @xmath298 choose now @xmath299 such that @xmath300
. then , for the term @xmath301 ( where @xmath302 ) we have the bounds : @xmath303 and for the term @xmath304 ( where @xmath305 ) we have the corresponding bounds : @xmath306 from this the theorem follows .
again as in @xcite , it suffices to show that the @xmath34-module @xmath149 is torsion free .
suppose then that there exists a function germ @xmath307 such that @xmath308 for some 2-form @xmath97 .
we will need to show that there exists an analytic function @xmath309 such that @xmath310 .
but by assumption , we have that @xmath311 and thus , by the relative de - rham division lemma , there exists a germ @xmath312 such that @xmath313 . taking the exterior differential ,
this relation reads : @xmath314 take now quasihomogeneous decomposition of the form @xmath1 , @xmath315 . for any @xmath316 we have @xmath317 and since the lie derivative commutes with the differential , we obtain the existence of an analytic function : @xmath318 such that @xmath310 .
obviously @xmath309 and freeness is proved . to construct a basis of @xmath149 we consider the @xmath34-module @xmath319 .
we have a natural inclusion of @xmath34-modules @xmath320 .
multiplication by @xmath2 in @xmath149 gives obviously elements inside @xmath321 and so @xmath322 .
on the other hand , by the quasihomogeneous division with @xmath95 ( lemma [ divlem ] ) we obtain that the class of any 2-form of the form @xmath206 , @xmath99 , can be represented by the class of a 2-form @xmath323 , @xmath174 . from this
it follows that : @xmath324 thus we obtain a sequence of isomorphisms of @xmath42-vector spaces : @xmath325 which is again a @xmath33-dimensional vector space .
thus , by nakayama lemma a basis of monomials @xmath326 , @xmath202 of the local algebra @xmath43 lifts to a basis @xmath327 in the deformation module @xmath149 .
the following theorem is the analog of a theorem of franoise @xcite in the ordinary ( symplectic , or volume - preserving ) case .
it concerns the local normal forms of martinet 2-forms under the action of the boundary - singularity preserving diffeomorphism group @xmath15 .
[ t2 ] let @xmath13 be a quasi - homogeneous boundary singularity of finite multiplicity @xmath33 . then
, for any germ of a martinet 2-form @xmath1 at the origin , there exist @xmath33 analytic functions @xmath328 and a diffeomorphism @xmath152 , such that @xmath1 is reduced to the normal form : @xmath329 where the classes of the monomials @xmath326 form a basis of the local algebra @xmath43 . moreover , the @xmath33 functions @xmath330 are uniquely determined by the pair @xmath10 . the existence of the normal form is obtained immediately by the homotopy method of proposition [ p1 ] and the finiteness theorem [ t1 ] .
thus , it suffices to prove only the uniqueness of the coefficients @xmath331 . for this
, we will need the following lemma : [ l0 ] for any @xmath332 and any @xmath152 , there exists an @xmath333 such that : @xmath334 briefly , interpolate @xmath335 by a 1-parameter formal subgroup @xmath336 , i.e. @xmath337 , @xmath338 and @xmath339 this is always possible due to general properties of the group @xmath15 ( the proof goes as in @xcite for the ordinary case ) .
then @xmath340 where @xmath341 is the formal vector field generated by the 1-parameter subgroup @xmath162 .
since @xmath162 preserves @xmath342 for all @xmath79 $ ] we have that @xmath343 for some formal function @xmath344 and in particular : @xmath345 now if we consider the decomposition of the @xmath346-form @xmath347 in the deformation module @xmath149 : @xmath348 then this it can be read as a decomposition in the formal module @xmath349 . comparing these two decompositions
we immediately obtain @xmath350 for all @xmath202 . _
( continuation of the proof of theorem [ t2 ] ) . _
decompose first @xmath1 in the deformation module @xmath149 : @xmath351 and take the difference of @xmath1 with @xmath352 : @xmath353 then from lemma ( [ l0 ] ) above it immediately follows that @xmath354 . the complete geometric description of the @xmath33 functional invariants @xmath355 associated to the pair @xmath0 is a difficult task and it is a part of what is usually called gauss - manin theory . below we will consider only the nondegenerate case @xmath356 , i.e. for the pair : @xmath357 where @xmath358 .
let us consider first a small ball at the origin of @xmath36 such that the fibers of @xmath359 are transversal to the boundary of this ball over the points @xmath89 of a sufficiently small disc in @xmath42 , centered at the origin ( the critical value of the restriction @xmath47 ) . modifying the neighborhoods under consideration sufficiently
, we may suppose that the fibers of the restriction @xmath47 on the martinet curve ( which consists of two points away from the origin , for @xmath360 ) are also transversal to the restriction of the boundary of the initial ball on the martinet curve @xmath14 .
the latter consists of two points and thus transversality with the fibers of @xmath47 means simply that they do not meet on @xmath14 , i.e. the fibers @xmath361 are bounded within a sufficiently small segment of the martinet curve .
the intersection of each of the fibers of @xmath2 with the interior of the chosen ball , is an open riemann surface @xmath362 with a set of distinguished points @xmath363 .
let @xmath364 be a 1-parameter family of relative cycles on the pair of fibers representing a relative homology class in @xmath365 , so that @xmath364 is obtained by continuous deformation of some relative cycle @xmath366 over a smooth pair @xmath367 .
as is easily seen , for @xmath89 real and positive , the pair of fibers is contractible to its real part and as @xmath368 the fiber @xmath363 shrinks to a point ( see figure [ fig:1 ] ) .
arnold called the relative cycle @xmath364 arising this way , vanishing half cycle @xcite : @xmath369 obviously , if @xmath1 is a germ of a martinet 2-form and @xmath90 is a primitive of @xmath1 , then the integral @xmath370 is an invariant of the pair @xmath0 . in a realisation of the 2-form as a magnetic ( curvature ) 2-form , the integral @xmath371 is nothing more that the magnetic flux on the 2-cell enclosed by the vanishing half - cycle .
it is easy to see that this integral is a holomorphic function of @xmath89 and thus we may consider its derivative @xmath372 : @xmath373 the integrand of this integral , the so called gelfand - leray form @xcite , @xcite , is defined as follows : let @xmath374 be the standard martinet 2-form . then if we denote by @xmath375 the euler vector field of @xmath2 the following relation holds : @xmath376 where @xmath377 and of course @xmath378 .
now , since @xmath379 we have that @xmath380 and thus : @xmath381 now the latter integral @xmath382 can be evaluated immediately , @xmath383 and it can be interpreted as the magnetic flux enclosed by the vanishing half - cycle @xmath364 .
thus we have : @xmath384 from this equation we obtain the expression for the invariant : @xmath385 its geometric explanation is direct : it measures the rate of change of the magnetic flux on the family of fibers , enclosed by the vanishing half - cycle @xmath364 for @xmath89 varying close to zero .
[ fig:1 ] , while the magnetic flux is @xmath386.,width=264,height=188 ] the problem of classification of functions @xmath2 under the action of the group @xmath387 of diffeomorphisms preserving the martinet germ , is not so easy as is the corresponding classification of martinet 2-forms @xmath1 relative to the @xmath15-action ( where as usual @xmath388 ) . to see
this it suffices to check that the infinitesimal equation @xmath389 where @xmath390 is a family of functions and @xmath159 a family of vector fields , does not admit any solutions inside the algebra @xmath391 of vector fields preserving the martinet germ @xmath1 .
despite this fact , for the nondegenerate case @xmath356 , a normal form involving exactly 1 functional invariant exists .
fix the martinet germ @xmath392 .
the following theorem describes the @xmath387-orbit of the @xmath393-boundary singularity @xmath394 .
it is the analog of vey s isochore morse lemma @xcite in the martinet case : [ t3 ]
let @xmath395 be a function germ such that the origin is a regular point for @xmath2 but nondegenerate ( morse ) critical point for the restriction @xmath396 on the martinet curve .
then there exists a diffeomorphism @xmath397 and a uniquely defined function @xmath398 , @xmath399 , @xmath400 such that @xmath401 by theorem [ t2 ] above we may choose a coordinate system @xmath35 such that @xmath402 and @xmath403 , where @xmath404 is a function , nonvanishing at the origin .
we may suppose that @xmath405 .
we will show that there exists a change of coordinates @xmath406 such that the pair @xmath407 goes to @xmath408 for some function @xmath409 and @xmath1 is reduced to martinet normal form . to do this , we set @xmath410 , @xmath411 , where @xmath412 is some function with @xmath413 ( so @xmath414 is indeed a boundary - preserving diffeomorphism ) . with any such function @xmath133 we have @xmath415 for the function @xmath416 , with @xmath399 and @xmath400 .
now it suffices to choose @xmath133 so that the map @xmath335 satisfies @xmath417 , i.e. such that the following initial value problem is satisfied for the function @xmath418 : @xmath419 as is easily verified this admits an analytic solution given by the formula @xmath420
we consider an example from lagrangian mechanics which concerns the motion of a charged particle on a riemann surface in the strong coupling ( quantisation ) limit with an electromagnetic field , or more generaly with an abelian gauge field ( @xcite , @xcite , @xcite ) .
there are several approaches and reformulations of the problem , the most appropriate for our case being that of a generalised particle ( c.f .
@xcite , @xcite ) , a variant of which we present below .
we fix a 2-dimensional riemannian manifold @xmath5 .
we consider the family of mechanical systems described by a regular lagrangian function @xmath421 `` quadratic in the velocities '' , i.e. such that in any local trivilisation of the tangent bundle with coordinates @xmath422 it can be expressed as : @xmath423 where the functions @xmath424 are homogeneous in the velocities @xmath425 of degree @xmath67 and :
* @xmath426 is a nondegenerate quadratic form @xmath138 on @xmath5 ( the riemannian metric ) representing the kinetic energy of the system * @xmath427 corresponds to a 1-form ( vector potential ) @xmath90 of gyroscopic forces
( such as magnetic forces e.t.c . ) represented by the 2-form @xmath428 , * @xmath429 is independent in the velocities and represents the scalar potential of other external forces acting upon the system ( such as electric e.t.c . ) , * @xmath430 , @xmath431 and @xmath432 are the coupling constants ( @xmath433 is the mass , @xmath434 is the charge e.t.c . ) , which may be viewed as formal parameters .
it is known that motions of the generalised particle emanating from a pont @xmath435 are smooth curves @xmath436 on @xmath5 , @xmath437 , which satisfy the generalised euler - lagrange equations : @xmath438 where @xmath439 is the covariant derivative associated to the riemannian metric @xmath138 and @xmath440 is the interior multiplication of a vector field with the 2-form @xmath428 . the right hand - side of equation ( [ el ] )
is known in the theory of electromagnetism as the lorentz force . in particular , the motions @xmath441\rightarrow m$ ] of the particle are exactly the critical points of the action functional : @xmath442 for the case @xmath443 , i.e. in the absence of external forces , the geometry of this variational problem with lagrangian @xmath444 , has been studied extensively in terms of subriemannian geometry and control theory ( c.f .
@xcite , @xcite , @xcite ) . there
, the eventual singularities @xmath445 of the 2-form @xmath446 play a special role : they are the _ abnormal geodesics _ of the corresponding pontryagin maximum principle and they can be obtained formally by the euler - lagrange equations ( [ el ] ) for @xmath447 . for the case @xmath448 , the geometry of this variational problem for the values @xmath449 can be studied in terms of jacobi metrics and for most of the cases where the 1-form @xmath450 of gyroscopic forces is nonsingular , in the sense that the 2-form @xmath428 is nondegenerate ( symplectic ) on @xmath5 .
it is important to notice also that for any @xmath449 the legendre transform @xmath451 of @xmath136 is a diffeomorphism and thus the phase space of the lagrangian system can be identified with cotangent bundle @xmath452 with hamiltonian @xmath453 and symplectic form @xmath454 the natural symplectic form of the cotangent bundle .
the generalised euler - lagrange equations ( [ el ] ) transform in that way to the canonical hamilton s equations : @xmath455 we will be interested here in the _ quantisation limit equations _ ,
i.e. for @xmath456 ( or @xmath457 , @xmath458 ) . notice that for @xmath447 the lagrangian @xmath459 is linear in the velocities and thus its legendre tranform @xmath451 is not a diffeomorphism .
for this reason , lagrangians linear in the velocities are called _ singular ( or constrained ) _ and the dynamics that they define through the euler - lagrange equations : @xmath460 is also called _ singular ( or constrained ) lagrangian dynamics_. the equation ( [ els ] ) above can be viewed as a constrained hamiltonian system @xmath10 with 2-form @xmath428 the form of gyroscopic forces and `` hamiltonian '' @xmath2 defined by the potential energy @xmath461 of the initial system . to describe this geometrically notice that in the limit @xmath447 , the image of the legendre tranform @xmath462 defines naturally a constraint submanifold in the phase space which is exactly equal to the image of the 1-form @xmath450 , viewed as a local section @xmath463 of the cotangent bundle : @xmath464 and it is thus diffeomorphic to the configuration space @xmath5
one has thus defined a diagram of maps : @xmath465 whose left arrow is the embedding of @xmath5 in @xmath452 through @xmath90 and the right arrow is the hamiltonian @xmath321 .
this shows that the motions of the generalised particle can be viewed inside the general scheme of the theory of hamiltonian systems with constraints .
as is easily verified , the euler - lagrange equations ( [ el ] ) for @xmath447 , correspond to the restriction of the hamiltonian system @xmath466 on the constrained submanifold , i.e. on the image of the 1-form @xmath90 .
indeed , one has @xmath467 and thus the restriction @xmath468 defines the constrained hamiltonian system : @xmath469 which is exactly the system of euler - lagrange equations ( [ els ] ) ( were we have put @xmath470 ) .
now let as consider the problem to determine the motions of the particle in the quantisation limit .
fix a point @xmath471 and identify the germs of singular lagrangians @xmath472 at @xmath473 with the germs of the corresponding pairs of potentials @xmath474 .
a singular lagrangian @xmath136 will be called `` generic '' if the corresponding pair of potentials @xmath475 is in general position ( relative to diagram ( [ d1 ] ) ) or equivalently , the codimension of the singularities of the pair @xmath476 is less or equal to @xmath477 .
the fact that this definition is correct is veryfied by the darboux - givental theorem @xcite .
replace now the 1-form @xmath90 by a 1-form @xmath478 , where @xmath479 is some arbitrary function , and the potential @xmath2 by @xmath480 , @xmath481 some constant .
then the form of the euler - lagrange equations ( [ els ] ) does not change and so there is naturally defined an equivalence relation between singular lagrangians : two germs @xmath482 and @xmath483 of singular lagrangians at @xmath471 will be called variationally ( or gauge ) equivalent if their euler - lagrange equations ( [ els ] ) are equivalent , i.e. there exists a diffeomorphism germ @xmath335 , @xmath484 , an arbitrary function germ @xmath479 and a constant @xmath481 such that : @xmath485 from the results of the previous sections on the classification of the pair @xmath476 we immediately obtain : [ tl ] the germ of a generic analytic lagrangian @xmath136 at a point @xmath473 on a manifold @xmath5 , is variationally equivalent to one of the following four invariant normal forms : @xmath486 @xmath487 @xmath488 @xmath489 where the functions germs @xmath409 and @xmath490 are analytic functions of one variable with a simple zero at the origin and they are the unique functional invariants of the ( variational orbits of the ) corresponding lagrangians . normal forms ( [ lnf0 ] ) , ( [ lnf1 ] ) correspond to the well known regular and morse cases respectively at a point @xmath491 in the symplectic plane . local normal forms ( [ lnf2 ] ) and @xmath492 correspond to regular , transversal points @xmath493 of @xmath2 on the martinet curve ( open and dense ) and to ( isolated ) points of first order tangency of @xmath2 with the martinet curve respectively .
normal form ( [ lnf3 ] ) is obtained immediately from corollary [ t3 ] of theorem [ t2 ] .
the normal form ( [ lnf2 ] ) is a simple exercise ; the @xmath494 sign comes from the fact that in the real case the martinet curve @xmath495 has an invariant orientation induced by the two symplectic structures in its complement in @xmath5 .
in fact , there is no real analytic diffeomorphism preserving @xmath496 and sending @xmath497 to @xmath498 .
the author whould like to thank jeroen lamb , dmitry turaev as well as michail zhitomirskii and jean - pierre francoise for useful discussions and their attention on the work .
55 r. abraham , j. e. marsden , _ foundations of mechanics _ , benjamin / cummings , ( 1978 ) v. i. arnold , v. v. kozlov .
a. i. neishtadt , _ mathematical aspects of classical and celestial mechanics _ ,
encyclopaedia of mathematical sciences , springer - verlag v. i. arnold , s. m. gusein - zade , a. n. varchenko , _ singularities of differentiable maps _ , monographs in mathematics , volume i , birkhuser , ( 1985 ) v. i. arnold , s. m. gusein - zade , a. n. varchenko , _ singularities of differentiable maps _ , monographs in mathematics , volume ii , birkhuser , ( 1988 ) v. i. arnold , v.v .
goryunov , o.v .
lyashko , v. a. vasilev , _ singularity theory i & ii _ , dynamical systems vi , viii , encyclopaedia of mathematical sciences , springer - verlag v. i. arnold , _ critical points of functions on a manifold with boundary , the simple lie grous @xmath499 , @xmath500 and @xmath501 and singularities of evolutes _ , russian mathematical surveys , 33:5 , ( 1978 ) , 99 - 116 g. d. birkhoff , _ dynamical systems _ , american mathematical society , vol . 9 , ( 1927 ) e. brieskorn , _ die monodromie der isolierten singularitten von hyperflnchen _ , manuscripta math . , 2 ( 1970 ) , 103 - 161 y. colin de verdiere ,
_ singular lagrangian manifolds and semi - classical analysis _ , duke math .
journal 116 , ( 2003 ) , 263 - 298 y. colin de verdiere , j. vey , _ le lemme de morse isochore _ ,
topology , vol .
18 , ( 1979 ) , 283 - 293 g. de rham , _ sur la division de formes et de courants par une forme linaire _ , commentarii mathematici helvetici , vol
. 28 , ( 1954 ) , 346 - 352 p. a. m. dirac , _ generalized hamiltonian dynamics _ , canadian journal of mathematics , vol . 0 ,
2 , ( 1950 ) , 129 - 148 w. domitrz , s. janescko , m. zhitomirskii , _ relative poincar lemma , contractibility , quasihomogeneity and vector fields tangent to a singular variety _ , illinois journal of mathematics , vol . 48 , no .
3 , ( 2004 ) , 803 - 835 , d. eisenbud , _ commutative algebra with a view towards algebraic geometry _ , graduate texts in mathematics , springer - verlag , 150 l.faddeev , r. jackiw , _ hamiltonian reduction of unconstrained and constrained systems _ , physical review letters , vol . 60 , no . 17 , ( 1988 ) , 1692 - 1694 j. p.
franoise , _ modle local simultan dune fonction et dune forme de volume _ , astrisque , s.m.f .
59 - 60 , ( 1978 ) , 119 - 130 j. p. franoise , _ relative cohomology and volume forms _ , singularities , banach center of publications , vol . 20 , ( 1988 ) , 207 - 222 j. p. franoise , _ le thorme de m. sebastiani pour une singularit quasi - homogne isole _ ,
annales de l institute fourier , vol .
2 , ( 1979 ) , 247 - 254 m. d. garay , _ an isochore versal deformation theorem _ , topology , vol .
43 , ( 2004 ) , 1081 - 1088 m. d. garay , _ analytic geometry and semi - classical analysis _ , proceedings of the steklov institute of mathematics 259 ( 2007 ) , 3559 .
a. b. givental , _ singular lagrangian manifolds and their lagrangian mappings _ , itogi nauki tekh .
probl . mat .
, novejshie dostizh .
33 ( 1988 ) , 55 - 112 , english translation : j. soviet math . 52.4 ( 1990 ) ,
3246 - 3278 r. jackiw , _ constrained quantization without tears _ , diverse topics in theoretical and mathematical physics , world scientific , ( 1995 ) k. kourliouros , _ local classification of constrained hamiltonian systems on 2-manifolds _ , journal of singularities , vol . 6 , ( 2012 ) , 98 - 111 v. v. kozlov , _ general vortex theory _ , dynamical systems x , encyclopaedia of mathematical sciences , springer , 67 , ( 2003 ) , 76 - 107 a. lichnerowicz , _ les varits de poisson at leurs algbres de lie associes _ , j. differential geometry , 12 , ( 1977 ) , 253 - 300 b. malgrange , _ frobenious avec singularits : codimension 1 _ , revue de lihes , no . 46 , ( 1976 ) , 163 - 173 b. malgrange , _ intgrales asymptotiques et monodromie _ , ann .
scient . ec .
7 , ( 1974 ) , 405 - 430 j. martinet , _ sur les singularits des formes diffrentielles _ , ann .
fourier(grenoble ) , 20 , no .
1 , ( 1970 ) , 95 - 178 v. i. matov , _ singularities of the maximum function on a manifold with boundary _ , translated form trudy seminara imeni i. g. petrovskogo , no . 6 , ( 1981 ) , 195 - 222 j. milnor , _ singularities of complex hypersurfaces _ , princeton university press and the tokyo university press , princeton , new jersey , ( 1968 ) r. montgomery , _ abnormal minimizers _ , siam j. control and opt .
32 , no . 6 ,
( 1994 ) , 1605 - 1620 r. montgomery , _ a tour of subriemannian geometries , their geodesics and applications _ ,
mathematical surveys and monographs , vol .
91 , american mathematical society r. montgomery , _ hearing the zero locus of a magnetic field _ , comm .
168 , ( 1995 ) , 651 - 675 f. pham , _ singularities of integrals , homology , hyperfunctions and microlocal analysis _ ,
universitext , springer s. pnevmatikos , _ evolution dynamique dun system mecanique en presence de singularites generiques _ , singularities & dynamical systems , north holland mathematical studies , ( 1983 ) , 209 - 218 s. pnevmatikos , d. pliakis , _ gauge fields with generic singularities _ , mathematical physics , analysis and geometry , 3 , no .
4 , ( 2000 ) , 305 - 321 m. sebastiani , _ preuve dune conjecture de brieskorn _ , manuscripta math . vol .
2 , ( 1970 ) , 301 - 308 c. l. siegel , j. moser , _ lectures on celestial mechanics _ , springer - verlag , ( 1971 ) j. sotomayor , m. zhitomirskii , _ impasse singularities of differential systems of the form @xmath502 _ , journal of differential equations , 169 , ( 2001 ) , 567 - 587 a. n. varchenko , _ local classification of volume forms in the presence of a hypersurface _ , translated from funktsionalnyi analiz i ego prilozheniya , vol . 19 , no .
4 , ( 1985 ) , 23 - 31 j. vey , _ sur le lemme de morse _ , inventiones mathematicae , 40 , ( 1977 ) , 1 - 9 j. vey , _ un problme de cohomologie relative _ , arkiv fr matematik , 15 , ( 1977 ) , 21 - 31 s. yakovenko , _ bounded decomposition in the brieskorn lattice and pfaffian picard - fuchs systems for abelian integrals _ , bulletin des sciences mathmatiques , vol .
127 , issue 7 , ( 2002 ) , 535 - 554 m. zhitomirskii , _ local normal forms for constrained systems on 2-manifolds _ , bul .
2 , ( 1993 ) , 211 - 232 . | we consider here the analytic classification of pairs @xmath0 where @xmath1 is a germ of a 2-form on the plane and @xmath2 is a quasihomogeneous function germ with isolated singularities .
we consider only the case where @xmath1 is singular , i.e. it vanishes non - degenerately along a smooth line @xmath3 ( martinet case ) and the function @xmath2 is such that the pair @xmath4 defines an isolated boundary singularity . in analogy with the ordinary case ( for symplectic forms on the plane )
we show that the moduli in the classification problem are analytic functions of 1-variable and that their number is exactly equal to the milnor number of the corresponding boundary singularity .
moreover we derive a normal form for the pair @xmath0 involving exactly these functional invariants .
finally we give an application of the results in the theory of constrained hamiltonian systems , related to the motion of charged particles in the quantisation limit in an electromagnetic field , which in turn leads to a list of normal forms of generic singular lagrangians ( of first order in the velocities ) on the plane .
singular symplectic structures , boundary singularities , normal forms , constrained hamiltonian systems , singular lagrangians |
the recently and long - awaited direct detection of a gravitational waves source ( gw150914 ) by the laser interferometer gravitational wave observatory ( ligo ) collaboration with the virgo collaboration @xcite was shown to be consistent with numerical simulations of the merger of a black hole ( bh ) binary system with masses @xmath2 @xcite .
it has been already stressed that this result is groundbreaking in several respects : it shows that binaries of gravitationally collapsed objects exist and merge within the age of the universe .
the merger of binary supermassive bhs has been shown by a number of numerical simulations @xcite to produce a collimated electromagnetic flash accompanying the gw event
. however , standard scalings in the pioneeristic work by @xcite are inconsistent with the large luminosity ( @xmath3 erg@xmath4s ) of the transient signal detected by gamma - ray burst monitor ( gbm ) on board the _ fermi _ satellite at photon energy @xmath5 kev @xcite , as discussed by @xcite .
owing to the exceptional impact of such an association , it seems worth exploring at least qualitatively a possible scenario for the joint events as already done , e.g. , in @xcite .
the widely accepted blanford - znajek mechanism envisages supermassive rotating bhs ( typically @xmath6 ) powering electromagnetic jets with luminosity @xmath7 erg@xmath4s and with circum - bh disks currents sustaining a magnetic field near the bh of order of @xmath8 gauss .
the luminosity of the gbm transient combined with @xmath9 bhs makes the association of gw150914 with the gbm transient unlikely as it would imply implausibly high magnetic fields .
we propose that the compenetration of the two magnetospheres amplifies by orders of magnitude the magnetic energy and drive , through reconnection , the release of a relativistically hot and optically thick pair plasma .
we argue in this short note that the latter scenario might be a commonplace electromagnetic source associated with merger of collapsed objects .
the source of gw 150914 very likely comprises of two objects at the final stage of their gravitational collapse , close to their respective horizons , inspiralling and moving at a good fraction of the speed of light . even if only mass ( @xmath10 ) , charge ( @xmath11 ) and spin ( @xmath12 ) are bhs observables , it is customary to assume that stationary bhs in the universe have no substantial charge ( @xmath13 in geometric units ) for a far - away observer .
we allow for each of the collapsing objects to carry a non - vanishing global electric charge @xmath13 and spin angular momentum @xmath12 .
we have no argument to support that the charge can not be quenched as it is customarily assumed ; we merely examine here the consequences of a magnetosphere supported by an intrinsic electric charge during the inspiral . if the two spins @xmath12 are anti - aligned and approximately orthogonal to the orbital plane , the compression of the anti - parallel magnetic field lines within the current sheet separating the two magnetospheres will drive a violent turbulent magnetic reconnection , which can eject a relativistically hot pair plasma that later emits the hard @xmath0-rays signal detected by fermi .
the problem of gravitational collapse of an electrically charged and spherically symmetric shell has been long investigated @xcite .
astrophysical implications of a charged collapsing object have been explored in the past two decades in connection with grbs prompt and afterglow high - energy light curves ( see , e.g. , @xcite ) and @xmath14 and @xmath15-ray band spectral evolution @xcite . for a stationary isolated kerr - newman bh , it is well - known that , upon developing the general relativistic solution in a far - away observer rest frame , the leading components to order @xmath16 of the electromagnetic stress - energy tensor @xmath17 behave , as for the magnetic field @xmath18 , as magnetic dipole : @xmath19 , @xmath20 and @xmath21 , where @xmath22 is the angular momentum of spin per mass unit and @xmath23 are @xcite coordinates .
the space - time geometry within a binary bh circular orbit is accessible only through numerical simulations and an analytical formulation is not known to date and possibly inexistent .
our first ansatz is that the magnetic dipole structure is approximately preserved within the inspiralling orbit .
turbulent magnetic reconnection has been repeatedly invoked in the past decade to explain the production of non - thermal particles populations , beside thermal heating , for those observations that were not complying with time - space constraints of well - understood mechanisms or with ideal mhd , traditionally in solar flares @xcite and more recently in the intensity peak of energetic particles at solar termination shock @xcite , and in relativistic astrophysical outflows ( see @xcite and references therein ) .
we argue that , during the last phase of the magnetospheres coalescence , turbulent compenetration at the current sheet layer dissipates magnetic energy into relativistically hot thermal pair plasma and non - thermal particles ejected in a highly collimated relativistic outflow .
a cartoon illustration in fig .
[ fig ] shows the current sheet layer that ejects the thermal pair plasma into the downstream cone along @xmath24-axis .
the electric field @xmath25 accelerates particles ( electrons along @xmath26 and positrons along @xmath27 ) as long as the particles remain confined within the boundary layer and gives rise to a non - thermal distribution ( see , e.g. , recent simulations of relativistic reconnection in flat space - time in @xcite ) .
assuming @xmath28 does not alter the main process in this simplified scenario . the relativistic highly collimated thermal plasma of pairs electron - positrons and photons outflowing the reconnecting region will propagate ( bulk lorentz @xmath29 ) as far as the plasma rest frame opacity of the high - energy @xmath15-ray to the electron - positron pair production drops to @xmath30 ( as discussed for grbs in @xcite )
; at this point the electromagnetic radiation is released in a single pulse detected at earth as the gbm transient of arrival time @xmath31 duration given by @xmath32 sec .
the gw signal emitted at the binary merger travels undisturbed presumably at the speed of light with infinite mean free path .
it has been already estimated @xcite that the observed arrival time delay @xmath33 is compatible with a distance travelled by an ultra - relativistic plasma ( @xmath29 ) of order of @xmath34 cm . at such a distance from the coalescence location ,
much larger than the clean circum - merger medium , the ejected plasma emits the gbm flash .
axis and charged particles that move back and forth across the reconnection layer while being accelerated by @xmath25 along the speiser orbits . [ fig],width=340 ]
the scenario qualitatively described above requires obviously a detailed demonstration through analytic estimates or extended resistive general relativistic mhd simulations .
we only argue here that the process of magnetic reconnection at the compenetration of magnetospheres of compact objects might be considerable source of electromagnetic radiation .
the reconnection growth rate in force - free relativistic reconnection @xcite can be cast as @xmath35 where @xmath36 is the alfvn time - scale and @xmath37 the resistive time - scale , both measured in the local lorentz frame . by using the estimate of @xmath15 for magnetars magnetosphere @xcite
, one can approximate the reconnection rate for a current sheet thickness of @xmath38 cm as @xmath39 the time - scale @xmath40 is measured in the local reference frame of the reconnecting layer ; thus , it does not account for the gravitational redshift in the unknown space - time geometry ( we remind that the frequency at the observer frame @xmath41 is shifed from the frequency in the source frame @xmath42 by a factor @xmath43 , for the schwarzschild metric , that can be very small close to the horizon ) , the dilation of @xmath40 due to gravitational redshift is weaker . ] .
however , on the very short merging time - scale due to highly relativistic flow speed involved , the actual reconnection can be limited to a current sheet thickness @xmath44 cm , greatly reducing @xmath40 @xcite .
the conversion of @xmath40 to the observer frame is also affected by doppler shift ( @xmath45 $ ] , where @xmath46 is the source speed and @xmath47 is the angle of the direction of photon emission with respect to the line of sight to the observer located far away from the source ) , arrival time dilation due to the fact that two photons emitted with a certain time - lag @xmath48 in the plasma comoving frame are detected with a different time - lag by the observer at earth @xmath33 because the source is moving ( @xmath49 ) and cosmological redshift ( @xmath50 for gw150914 ) . as an order of magnitude estimate , a radiation pulse emitted when the pair plasma at @xmath29 reaches transparency can easily produce a pulse @xmath1 sec delayed from gw150914 @xcite . by assuming that the magnetic field @xmath51 upstream of the reconnecting current sheet inflows at speed @xmath52 in the late merging phase near the horizon ( @xmath53 ) , one can equate @xmath54 to the gbm transient luminosity ( @xmath55 erg@xmath4s ) ; the resulting magnetic field of order @xmath56 gauss is not unusual for magnetars although it rules out magnetic field sustained by an accretion disk in a configuration _ _
la @xcite . in this treatment
the value of the @xmath57 is not relevant as long as the merging magnetospheres can be sustained during the coalescence process .
also , we did not discuss here the process leading to the formation of the collapsing rotating stars . before the detection of gw150914 ,
binary bhs were not deemed to exist in reality : models of gravitational collapse of massive stars in an already formed binary system and model of coalescence of two bhs moving one with respect to the other with a sufficiently small impact parameter have both unresolved issues .
we have used here magnetosphere of a kerr - newman bh for an object close to the final stage of its gravitational collapse .
it is expected that in the local lorentz frame where the reconnecting magnetic field lines come close to each other , the current sheet layer is dragged along the orbital motion of the compact objects and stretched by differential rotation , developing strong shear instabilities ( see @xcite for the shear instability for an isolated neutron star ) . those instabilities will presumably grow faster and faster as the system merge .
a noteworthy caveat on the scenario here proposed is that the current modelling of turbulent magnetic reconnection is still under development .
moreover , _ in situ _ solar wind measurements @xcite did not confirm that field dissipation efficiently transfers magnetic energy to a non - thermal particle population in non - relativistic flows .
however , several recent numerical experiments seemingly support that reconnection produces energetic particles power - laws in various astrophysical contexts @xcite .
this scenario will be further investigated within either the framework of tearing mode instabilities , or petschek model of two slow mode shocks @xcite , both cases in the incompressible regime as usually assumed in the literature .
we hope that the scenario here proposed for the claimed association of the directly detected gw signal with a hard @xmath0-ray transient signal , and the expected future joint detections , will stimulate fresh thoughts and prompt detailed resistive general relativity numerical investigations .
we acknowledge useful comments and feedbacks by d. arnett , m. begelman , m. rees , l. rezzolla , b. zhang .
this work was supported , in part , by nasa under grants nnx13ag10 g and nnx15aj71 g . | we propose a qualitative scenario to interpret the argued association between the recent direct measurement of the gravitational wave event gw150914 by ligo - virgo collaborations and the hard @xmath0-ray transient detected by fermi - gbm @xmath1 sec after . in a binary system of two
gravitationally collapsing objects with a non - vanishing electric charge , the compenetration of the two magnetospheres occurring during the coalescence , through turbulent magnetic reconnection , produces a highly collimated relativistic outflow that becomes optically thin and shines in the gbm field of view .
we propose that this process should be expected as a commonplace in the future joint gravitational / electromagnetic detections .
[ firstpage ] physical data and processes : turbulence ; ism : cosmic rays , magnetic fields . |
the microwave response is a source of important information in superconductors . in high-@xmath5 superconductors ( htcs )
various microwave techniques have been used to get information , among the others , on the symmetry of the order parameter , on the vortex parameters such as the vortex viscosity and pinning frequency @xcite and on the temperature dependence of the superfluid fraction ( via the measurement of the temperature dependence of the london penetration depth @xcite ) . in short ,
microwave measurements are a very powerful tool to study the superconducting state .
+ while yba@xmath0cu@xmath1o@xmath2 ( ybco ) and , to a minor extent , bi@xmath0sr@xmath0cacu@xmath0o@xmath6 ( bscco ) have been the subject of intensive experimental investigation , other htcs did not receive the same attention .
in particular , a very few reports dealt with the vortex - state microwave response in rare - earth substituted 123 compounds , mainly on dy @xcite and gd - substituted materials.@xcite in fact , re - substituted 123 compounds present potentially interesting features in the vortex state : in particular , they often give enhanced irreversibility lines @xcite and in some case a @xmath5 slightly higher than in ybco . to what extent those features can change the microwave properties in the vortex state is not known , nor it is clear what effect substitutions can have on the pinning at high frequencies .
aim of this paper is to present an extensive study of the microwave response of smba@xmath0cu@xmath1o@xmath2 ( smbco ) in the vortex state in moderate fields , up to the critical temperature .
it will be shown that the complex resistivity shows up an unexpectedly strong field dependence .
in particular , the imaginary part increases noticeably with the applied magnetic field at low temperatures , becomes nearly field - independent at a sample - dependent temperature around @xmath7 0.9 and then decreases with the application of the field up to @xmath5 .
in addition , both the real and imaginary parts of the resistivity present clear sublinear field dependences . we show that all those findings can not be easily understood within the conventional framework for vortex motion at high frequencies .
by contrast we find that , by assuming a significant effect of the magnetic field on the superfluid depletion , all the temperature and field dependences are well described by the simple viscous vortex motion supplemented with the two - fluid conductivity .
we obtain almost the same temperature - dependent vortex viscosity in two different samples , in agreement with its intrinsic origin .
consistently , the temperature dependence of the vortex viscosity in smbco is identical to the one measured in ybco .
the smba@xmath0cu@xmath1o@xmath2 films were deposited on ( 100)k laalo@xmath1 substrates by a high oxygen pressure sputtering technique .
even if these substrates show higher mismatch in comparison with the srtio@xmath1 , the use of laalo@xmath1 is mandatory for microwave studies and applications due to its low loss tangent .
sintered targets@xcite of 35 mm diameter and of 2 mm thickness were used .
the nominal composition of the metal elements ( sm : ba : cu ) in the targets was 1:2:3 .
all the films were deposited at high temperatures , with the substrates placed onto an inconel heater whose temperature was monitored by a thermocouple embedded into the furnace .
the films were grown in a vacuum chamber at constant oxygen pressure @xmath8 and substrate temperature @xmath9 during the deposition . the tetragonal to orthorhombic phase transition was accomplished after the deposition process by cooling the samples in an oxygen pressure of one atmosphere from @xmath9 to the annealing temperature , @xmath10 , kept constant for the time of the annealing process , @xmath11 .
the samples were finally cooled down to the room temperature .
the main deposition parameters are : @xmath12 910@xmath13c , @xmath14 2.6 mbars , @xmath15 560@xmath13c , @xmath16 15 min .
the crystal structure of the films was examined by x - ray @xmath17 diffraction .
the x - ray diffraction pattern showed that the growth of the film onto a laalo@xmath1 substrate was single phased with preferential orientation along the @xmath4 axis .
the rocking curve of the ( 005 ) peak exhibits a typical full width at half maximum ( fwhm ) of less than 0.2@xmath13 .
the surface morphology of the films was observed by atomic force microscopy ( afm ) at room temperature .
2d - nucleation growth was observed , and roughness was found to be @xmath18 2 nm on 1 @xmath19 m @xmath20 1 @xmath19 m area .
film thickness was typically @xmath21 200 nm . two films were prepared , with @xmath22 86.5 k , as determined from the inflection point of the temperature - dependent real resistivity .
the dc resisistivity at 100 k was @xmath23(100)@xmath24300@xmath25 cm in both samples .
further details on sample preparation and structural characterization have been reported elsewhere.@xcite + the microwave response was measured by the end - wall cavity technique at 48.2 ghz with a moderate magnetic field applied along the @xmath4-axis .
the maximum attainable field was @xmath260.8 t. @xmath27 factor and resonance frequency @xmath28 were measured as a function of the magnetic field at various fixed temperatures in the range 60 - 90 k. from the field induced change of @xmath27 and @xmath28 the field induced change of the effective surface impedance @xmath29 could be obtained.@xcite due to the small thickness of the film , the thin film approximation applies to the measurements,@xcite so that @xmath30 , where @xmath31 is the complex resistivity .
it must be stressed that the absolute value of @xmath31 is affected by errors in the evaluation of the geometrical factor and of the film thickness , and in the calibration of the cavity.@xcite the first two are eliminated and the third strongly reduced by working with the reduced complex resistivity change , @xmath32 , with @xmath33(100 k ) . in the following
we will work with the reduced quantity , since ( as it will be clear later ) the absolute values will be useful only in the discussion of the details of the parameters .
+ we now present the main features of our data .
two samples were investigated , and both presented the same experimental behaviour . in the following ,
we focus on measurements taken more systematically in sample a. in figures [ r1vsb ] and [ r2vsb ] we report the field - sweeps at several temperatures for the real and imaginary parts .
the real part , @xmath34 , increases with the field , with increasing amplitude up to a typical temperature @xmath35 82 k ( figure [ r1vsb]a ) , above which the application of the field has a lower and lower effect by increasing the temperature ( figure [ r1vsb]b ) .
the imaginary part , @xmath36 , increases with the field at lower temperatures , becomes approximately field independent at @xmath37 80 k and then starts to decrease with the field ( figure [ r2vsb ] ) above @xmath38 .
the same features are observed in sample b , with @xmath37 77 k. both @xmath34 and @xmath36 exhibit a sublinear behaviour with respect to the applied field .
the explicit field dependences of @xmath34 and @xmath36 can be directly identified by replotting the same data of figures [ r1vsb ] and [ r2vsb ] vs @xmath39 ( figures [ r1vssqrtb ] and [ r2vssqrtb ] ) .
it is immediately seen that @xmath40 , while @xmath34 has an upward curvature . within the experimental uncertainty ,
@xmath41 is the sum of a linear and a square - root term in the applied field . summarizing , the full body of our data in smbco can be described by the empirical relations : @xmath42 @xmath43 interestingly , these field dependences do not vary , neither when @xmath36 changes from positive to negative , nor when the temperature raises so that the measurements cross the irreversibility line ( by increasing the temperature from 65 k to @xmath5 we certainly cross @xmath44 ) , nor when the temperature is raised above @xmath45 . only very close to @xmath5 ( approximately 1 k below , where strong fluctuational effects become predominant ) the field dependences empirically found appear to slightly modify and the empirical equations above no longer describe the data very accurately .
all the obtained coefficients in the two samples studied have the same behavior as a function of the temperature , as reported in figure [ a1a2 ] for @xmath46 and @xmath47 and in figure [ b1 ] for @xmath48 .
this is a clear indication ( a ) against the relevance of pinning at our measuring frequency , and ( b ) in favour of some `` intrinsic '' origin for the behaviour of the resistivity in the mixed state .
+ given the very clear field dependences of the complex resistivity , we search for a theoretical frame able to describe our data , taking into account that no features appear when crossing the irreversibility line . in the following section we briefly summarize the reference frame that we adopt for the theoretical description of the microwave response in the vortex state .
in principle , the response of a superconductor in the mixed state is dictated by the motion of vortices and of charge carriers .
the latter originates from superfluid as well as normal carriers , and field - induced changes of the quasiparticle ( qp ) density have a direct effect on both the real and the imaginary part . a general frame that includes all those contributions is the coffey - clem ( cc ) theory @xcite ( a very similar model has been independently derived by brandt @xcite ) , based on the assumption of the two - fluid conductivity and of a periodic vortex pinning potential .
+ the general cc expression , including vortex , superfluid and qp contributions , is rather complex . in particular , the combined effect of pinning and creep make even the vortex motion alone of difficult description . in order to clarify the role of pinning , we first focus the attention below the irreversibility line where its effect is maximum . in the latter case
, the characteristic frequency for the vortex diffusion reduces to the pinning frequency @xmath49 , where @xmath50 is the vortex viscosity and @xmath51 the pinning constant ( labusch parameter ) .
the resulting expression for the microwave complex resistivity @xmath52 in the vortex state can be then cast in the form : @xmath53\ ] ] and @xmath54\ ] ] where @xmath55 is the measuring frequency , @xmath56 is the so - called flux- resistivity , and @xmath57 is the complex conductivity in absence of motion of vortices .
the latter is temperature , field and frequency dependent , and contains the contributions of superfluid and qp , through their fractional densities and the qp scattering time .
+ these expressions have been sometimes used for the analysis of the microwave response in the vortex state.@xcite however , in most cases it was judged that the field - dependence of the qp and superfluid densities was negligible , and that at low enough temperatures @xmath58 . within these approximations ,
the coffey - clem ( cc ) prediction for the field - induced changes to the resistivity reduces to the well - known gittleman - rosenblum ( gr ) model @xcite ( in the gr model the qp and superfluid contributions are not considered , so the response is described by the vortex motion alone ) : @xmath59 and @xmath60 the latter model was in fact extensively used for the analysis of the microwave resistivity or surface impedance ( see ref.@xcite and references therein ) .
when it is possible to apply this conventional analysis , the vortex parameters @xmath61 and @xmath62 can be easily obtained by inverting the data through equations [ gr1],[gr2 ] . the vortex viscosity is calculated as @xmath63 .
finally , the pinning constant comes from the definition of @xmath61 .
+ in the following section , we discuss first our data below the irreversibility line in terms of the gr model , and then within the cc model allowing for the inclusion of a possibly relevant effect of the magnetic field on the qp and superfluid fractional densities .
the emerging frame strongly points toward the latter interpretation .
when discussing our data on smbco we are faced with two main experimental features : the field dependences of @xmath64 and @xmath65 , which are clearly sublinear , and the very relevant increase of the imaginary part in even moderate fields .
since the role of pinning reveals itself mostly on the imaginary part , one might be tempted to assign the strong field dependence of @xmath36 simply to a strong pinning . within this quite conventional view
, one would apply the gr model and , following the procedure explained in section [ th ] , would obtain @xmath61 , @xmath50 , @xmath51 directly from the data .
the result of the procedure is reported in figure [ conventional ] for several temperatures . as can be seen
, the so - obtained pinning frequency would be a very weak function of the applied field , in agreement with commonly reported dependences.@xcite however , the viscosity @xmath50 would present a clear increase with the field ( approximately as @xmath66 ) , as reported in figure [ conventional ] ( panel b ) . this dependence is not easily explained .
in fact , a similar behaviour at high fields has been tentatively explained in bi@xmath0sr@xmath0cuo@xmath67 ( ref.@xcite ) in terms of a peculiar field dependence of the quasiparticle relaxation time in a d - wave superconductor , appearing when the intervortex distance becomes smaller than the mean free path .
however , the model would explicitly predict the usual _ field independence _ at low fields .
due to our field and temperature ranges ( by far lower than the temperature - dependent upper critical field ) , this picture does not seem very convincing .
the anomalies become even more evident when , within the same gr model , we try to obtain the pinning constant @xmath51 from the vortex pinning frequency and the vortex viscosity : as reported in figure [ conventional ] ( panel c ) , the so - obtained pinning constant would _ increase _ with the field ( again , approximately as a square root ) .
this behaviour does not seem reasonable .
in particular , we note that in ref.@xcite @xmath51 was found to be constant at low fields , and to decrease at higher fields , as expected in high - frequency measurements .
we conclude that an explanation of our data in terms of the gr model is at least questionable .
+ we now show that the empirical field dependences for the complex resitivity can be immediately derived within a cc - like model , by including a field dependence of the qp fractional density in agreement with the existence of lines of nodes in the superconducting gap .
+ we first critically discuss the assumptions underlying the cc model .
the first assumption , the applicability of the two - fluid model , is reasonably justified by comparison with microwave studies@xcite of ybco , which is the superconducting cuprate most similar to smbco . in particular , the imaginary part of the conductivity can be written as : @xmath68 , where @xmath69 is the zero - temperature london penetration depth , and @xmath70 is the temperature - dependent superfluid fraction , which is found@xcite to be @xmath71 , with @xmath72 , to a good approximation in the full temperature range from low temperatures up to very close to @xmath5 . moreover , it was found that the normal fluid fractional density @xmath73 decreased with decreasing temperature below @xmath5 .
in addition , no significant frequency dependence was observed @xcite in the real conductivity up to 75 ghz for temperatures @xmath74 60 k indicating that at 50 ghz and above @xmath75 the real part of the conductivity can be written as @xmath76 , where @xmath77 is the quasiparticle relaxation time .
+ regarding the chosen description of the vortex motion , the assumption of periodic pinning potential is believed not to strongly affect the main features of the response , until the probing frequency is high enough to displace the vortices a small fraction of the intervortex distance , as is the case at high microwave frequencies .
in fact , swept - frequency corbino disk measurements in the frequency range 6 - 20 ghz@xcite have shown that the frequency dependence of @xmath78 was well described by the coffey - clem expression for vortex dissipation , and gave estimates of the pinning frequency @xmath79 5@xmath2010@xmath80 at 50 ghz and above 60 k.@xcite + assuming that the main features of ybco apply also to the closely related compound smbco , we compare our data to a cc - like model where the field - induced superfluid depletion is explicitly included , and we assume @xmath81 .
the latter assumption is further supported by the experimental fact that our data do not show any detectable feature with increasing temperature from below to above the irreversibility line .
in addition , at small enough fields , not too close to @xmath5 we have @xmath82 and @xmath83 1 , so that eq.s [ cc1],[cc2 ] simplify considerably and we obtain the approximate expressions : @xmath84 and @xmath85 it can also be shown that , within the approximations described above , eq.s [ cc1appr],[cc2appr ] remain valid above and below the irreversibility line : these approximate equations represent ( a ) the microwave response as given by free oscillations of flux lines ( analytically described by the flux flow resisitivity ) , given by the first term in eq.[cc1appr ] , and ( b ) the contribution from superfluid and qp conductivity , as given by eq.[cc2appr ] and by the second term in eq.[cc1appr ] . +
an explicit field dependence of the two - fluid conductivity has to be inserted in these equations in order to compare them to the data . in a clean , conventional s - wave superconductor
the superfluid density in the mixed state would decrease , and the qp fraction would increase , due to the density of states in the vortex cores .
the effect would be proportional to the number of vortices , thus linear in @xmath86 .
however , it has been shown that in a superconductor with lines of nodes in the superconducting gap delocalized states ( outside of the vortex cores ) give a very significant contribution , exhibiting finite density of states at the fermi level.@xcite the dos is proportional to the number of vortices and to the spacing between them , thus leading to @xmath87 increase of the qp fractional density .
this effect has been experimentally observed in the specific heat of ybco,@xcite and in measurements of the qp fractional increase with sub - terahertz spctroscopy in bscco thin films.@xcite + we then incorporate the field dependence in the two - fluid conductivity through the superfluid and qp fractional densities@xcite as @xmath88 and @xmath89 , with @xmath90 , @xmath91 ( ref.@xcite ) and @xmath92 as appropriate for a superconductor with lines of nodes in the superconducting gap.@xcite + interestingly , taking into account that @xmath93 is a small parameter , one can expand the two - fluid terms in eq.s [ cc1appr ] and [ cc2appr ] .
one finally gets for the field variations of the normalized resistivity : @xmath94 \left(\frac{b}{cb_{c2}}\right)^{\frac{1}{2}}\\ \label{cc1final } \equiv b_{1}(t ) b + a_{1}(t ) b^{\frac{1}{2}}\end{aligned}\ ] ] and @xmath95 \left(\frac{b}{cb_{c2}}\right)^{\frac{1}{2}}\\ \label{cc2final } \equiv a_{2}(t ) b^{\frac{1}{2}}\end{aligned}\ ] ] where we have defined @xmath96 .
these two equations give exactly the functional field dependences that have been empirically found in the data .
this is one of the main results of this paper : the experimental field dependence of the complex resistivity , not explainable in the framework of fluxon motion alone , is recovered by a model that includes simple free flux line oscillation ( corresponding to flux flow resistivity ) and the essential role of superfluid depletion in a superconductor with lines of nodes in the gap .
we note that the inclusion of a vortex motion term in the imaginary resistivity would have given an additional _ linear _ term in eq.[cc2final ] . to a great accuracy
, no @xmath86-linear term is observed in the data , giving further consistency to the approximations used to derive eq.s [ cc1final],[cc2final ] . as a further relevant result ,
we note that the @xmath46 and @xmath47 coefficients here defined present the same qualitative features as in the experimental values : on one side , the study of eq.[cc1final ] reveals that @xmath46 can present a peak ; more important , it is immediately seen that @xmath47 undergoes a sign change at a temperature such that : @xmath97^{-\frac{1}{2 } } \label{vincolo}\ ] ] summarizing , the model here developed contains all the experimental features present in the data : the field dependences of the complex resistivity and the behavior with the temperature of @xmath46 and @xmath47 . + for a quantitative fit of the data with the theoretical expressions , one has to determine the temperature dependences of the parameters contained in eq.s [ cc1final],[cc2final ] . exploiting the similarities with ybco , and consistently with the indication of the existence of lines of nodes in the gap , we have taken @xmath98 and @xmath99 , and @xmath100 . to our knowledge
there are no detailed studies of the finite frequency conductivity in smbco that could give indications on the temperature dependence of the qp scattering time : we then make the very crude assumption that the qp scattering time does not change much from above to below @xmath5 . with this choice , one can write ( apart small corrections due to the temperature dependence of the normal state resistivity , that we neglect here ) @xmath1011 , and @xmath102 . with this choice , on the basis of the constraint given by eq.[vincolo ]
we have @xmath1030.146 and 0.212 , in sample a and b , respectively.@xcite the simultaneous fits of @xmath104 and @xmath105 contain only @xmath106 as a common scale factor . in figure [ a1a2 ]
we plot the coefficients @xmath46 and @xmath47 , compared with the theoretical curves computed on the basis of eq.s [ cc1final],[cc2final ] with @xmath107330 t and 230 t as scale factors for sample a and b , respectively . the numerical values of @xmath108 are larger by a factor @xmath18 30 than expected with @xmath109 280 @xmath25 cm and @xmath110 2000 .
neglected effects , such that a temperature dependent quasiparticle scattering time,@xcite a zero - temperature superfluid fraction less than unity , a two - component order parameter@xcite might be responsible for this enhancement .
nevertheless , given the crudeness of the model and the absence of fit parameters , the theoretical expectations describe with surprising accuracy our data .
we believe that the essential physics of the field dependent microwave response is related to the effects extensively described in this paper .
+ finally , we discuss the fluxon dynamics . within the present interpretation ,
it is entirely described by the coefficient @xmath48 .
using for @xmath111 the determinations obtained from @xmath46 and @xmath47 above , we immediately get the fluxon viscosity @xmath50 , as reported in figure [ b1 ] . as can be seen
, the data for @xmath50 attain the same value in both samples , indicating that the physics in the vortex core is related to sample - independent processes . +
further insights can be gained by examining the data within the bardeen - stephen model for the viscosity:@xcite in this case @xmath112 . since ( within 10% )
@xmath113 is the same in both samples , we deduce that also the upper critical field is the same . from that
, we get for the constant @xmath4 the relation @xmath1141.44 ( subscripts refer to the sample ) .
comparison with vortex viscosity data taken in ybco@xcite shows that : ( a ) the temperature dependence is the same in smbco and ybco ( minor differences appear only close to the critical temperature ) , and ( b ) the data of ybco scale with a factor of 2 on the corresponding data in smbco , as we show in figure [ etainsieme ] .
since the ratio of the normal state resistivities at 100 k in the ybco sample here reported and in the smbco samples is @xmath115(smbco)/@xmath115(ybco)@xmath24 3 , and @xmath116(ybco)@xmath24 165 t,@xcite using again the bardeen - stephen model we obtain @xmath116(smbco)@xmath24 250 t , @xmath1171.32 , @xmath1180.92 , in agreement with the requirement @xmath91 .
it should be also noted that the vortex viscosities measured in ybco are in excellent agreement@xcite with those measured in ybco crystals at three different frequencies,@xcite suggesting a common behavior for the electronic states in the vortex cores of smbco films , ybco films and ybco crystals . + as a final remark , we remind that the quantitative description of the vortex state microwave response here presented could be accomplished only with the essential inclusion of the enhanced qp increase in a magnetic field .
a description in terms of fluxon motion alone would have given very unlikely dependences of the vortex parameters .
we have presented the first extensive vortex - state microwave characterization in smba@xmath0cu@xmath1o@xmath2 thin films as a function of the temperature and magnetic field .
the field - dependence of the 48 ghz complex resistivity exhibits a markedly sublinear behaviour .
the analysis in terms of strong pinning with weak field dependent superfluid depletion leads to internal contradictions , so that alternative interpretations have to be found .
the experimental results give clear indications of the irrelevance of pinning at our measuring frequency . a framework including free oscillation of flux lines ( whose resistivity is given by the flux flow expression ) and an enhanced role of the field - induced pair breaking ,
due to the existence of lines of nodes in the superconducting gap , is able to describe the data qualitatively and quantitatively .
the temperature dependences of the qp and superfluid contributions to the complex resistivity are fitted by the model with no free parameters . within this frame
, we estimate the fluxon viscosity and we find the same values in both smbco samples .
the temperature dependence is the same as that measured in ybco .
moreover , the viscosity data in smbco and ybco scale together with a numerical factor . using the bardeen - stephen model
we estimate the upper critical field in smbco and the @xmath4 constants , which we consistently find @xmath119 .
we conclude that the vortex state resitivity in smbco is well described by vortex motion ( that at our measuring frequencies takes the form of flux flow ) , with the essential contribution of the strong field dependence of the qp and superfluid fractional densities .
the latter is connected to the structure of the superconducting gap .
useful discussions with a. maeda and r. wrdenweber are warmly acknowledged . support from the esf `` vortex '' program is acknowledged .
this work has been partially supported by italian miur under firb `` strutture semiconduttore / superconduttore per lelettronica integrata '' .
l.m . acknowledges financial support from miur under the same project . c. beneduce , f. bobba , m. boffa , a.m. cucolo , m.c .
cucolo , a. andreone , c. aruta , m. iavarone , f. palomba , g. pica , m. salluzzo , r. vaglio , _ int . j. mod
. phys . b _ * 13 * , 1333 ( 1999 ) ; m.a .
boffa , f. bobba , a.m. cucolo , r. monaco , _ int . j. mod
b _ * 17 * , 768 ( 2003 ) ; m. boffa , m.c .
cucolo , r. monaco , a.m. cucolo , _ physica c _ * 384 * , 419 ( 2003 ) .
e. silva , m. lanucara , r. marcon , _ supercond .
technol . _ * 9 * , 934 ( 1996 ) ; n. pompeo , r. marcon and e. silva , _ proceedings of 6th european conference on applied superconductivity - eucas 2003 _ , 14 - 18/9/2003 , sorrento ( italy ) , to be published , paper 709 . d.a .
bonn , r. liang , t.m .
riseman , d.j .
baar , d.c .
morgan , k. zhang , p. dosanjh , t.l .
duty , a. macfarlane , g.d .
morris , j.h .
brewer , w.n .
hardy , c. kallin , a.j .
berlinsky,_phys . rev .
b _ * 47 * , 11314 ( 1993 ) .
we have taken for simplicity a field - independent quasiparticle scattering time . however , it is easy to show that this choice changes the subsequent equations only by some numerical factor .
for example , allowing for an electron - vortex scattering like the one proposed in @xcite , one would only redefine the quantity @xmath120 in eq.s[cc1final],[cc2final ] .
e. silva , r. fastampa , m. giura , r. marcon , d. neri and s. sarti , _ supercond .
_ * 13 * 1186 ( 2000 ) ; r.rogai , r.marcon , e. silva , r. fastampa , m. giura , s. sarti , m. boffa , a.m. cucolo , _ int . j. mod
b _ * 14 * , 2828 ( 2000 ) s. sarti , c. amabile , n. tosoratti , e. silva , cond - mat/0307143 ; c.amabile , r. fastampa , m. giura , s. sarti , e. silva , v. ferrando , c. ferdeghini , _ proceedings of 6th european conference on applied superconductivity - eucas 2003 _ , 14 - 18/9/2003 , sorrento ( italy ) , to be published . with a different choice of the temperature dependence of @xmath73 and @xmath70
one would have different @xmath108 .
as an example , in sample a we have tried @xmath121 and @xmath122 , with @xmath1231,2,4 ( and correspondingly @xmath1030.075 , 0.146 , 0.278 for our values of @xmath38 ) , obtaining nearly identical fits ( with the same scale factor ) .
h. srikanth , b. a. willemsen , t. jacobs , s. sridhar , a. erb , e. walker , and r. flkiger _
b _ * 55 * , r14733 ( 1997 ) ; h. srikanth , z. zhai , s. sridhar , a. erb , and e. walker _ phys . rev .
b _ * 57 * , 7986 ( 1998 ) | we report measurements of the microwave complex resistivity at 48 ghz in smba@xmath0cu@xmath1o@xmath2 thin films .
measurements are performed with a moderate magnetic field , @xmath3 0.8 t , applied along the @xmath4-axis .
we find that the complex resistivity presents clear sublinear field dependences , and that the imaginary part is remarkably sensitive to the moderate magnetic field .
interpretation considering an unusually strong pinning leads to very anomalous field dependences of the single - vortex viscosity and of the pinning constant . by contrast , allowing for a significant effect of the magnetic field on the depletion of the condensate , the data are quantitatively described by the simple free - flow - like expression , supplemented with two - fluid conductivity . in this frame
, we obtain the vortex viscosity from the data .
we compare vortex viscosity in smba@xmath0cu@xmath1o@xmath2 and in yba@xmath0cu@xmath1o@xmath2 . 2 |
multiple hypergeometric functions @xcite play an important role in many branches of science . in particular , a large class of feynman diagrams are expressed in terms of horn - type hypergeometric functions @xcite .
let us consider a multiple series : @xmath7 with @xmath8 the sequences @xmath9 and @xmath10 are called _ upper _ and _ lower _ parameters of the hypergeometric function , respectively .
let @xmath11 denote the unit vector with unity in its @xmath12 entry , and define @xmath13 for any integer multi - index @xmath14 .
two functions of type ( [ h ] ) with sets of parameters shifted by unity , @xmath15 and @xmath16 , are related by a linear differential operator : @xmath17}^+ h(\vec{\gamma},\vec{\sigma } , \vec{x } ) \;.
\label{do1}\end{aligned}\ ] ] similar relations also exist for the lower parameters : @xmath18}^- h(\vec{\gamma};\vec{\sigma};\vec{x } ) \;.
\label{do2}\end{aligned}\ ] ] the linear differential operators @xmath19 , @xmath20 are called the _ step - up _ and _ step - down _
operators for the upper and lower index , respectively . if additional step - down and step - up operators @xmath21 , @xmath22 satisfying @xmath23}^- u_{[\gamma_c \to \gamma_c+1]}^+ h(\vec{\gamma},\vec{\sigma},\vec{x } ) = l_{[\sigma_{c}-1 \to \sigma_{c}]}^+ l_{[\sigma_c \to \sigma_c-1]}^- h(\vec{\gamma},\vec{\sigma},\vec{x } ) = h ( \vec{\gamma},\vec{\sigma},\vec{x})\ ] ] ( _ i.e. _ , the inverses of @xmath24 , @xmath25 ) are constructed , we can combine these operators to shift the parameters of the hypergeometric function by any integer .
this process of applying @xmath26 to shift the parameters by an integer is called * differential reduction * of a hypergeometric function . in this way
, the horn - type structure provides an opportunity to reduce hypergeometric functions to a set of basis functions with parameters differing from the original values by integer shifts : @xmath27 where @xmath28 and @xmath29 are polynomials with respect to @xmath30 and @xmath31 , and @xmath32 are lists of integers .
algebraic relations between the functions @xmath33 with parameters shifted by integers are called * contiguous relations*. the development of systematic techniques for solution of contiguous relations has a long story .
it was started by gauss who described the reduction for the @xmath34 hypergeometric function in 1823 @xcite .
numerous papers have since been published @xcite on this problem .
an algorithmic solution was found by takayama in ref .
@xcite , and those methods have been extended in a later series of publications @xcite ( see also @xcite ) .
let us recall that any hypergeometric function can be considered to be the solution of a proper system of partial differential equations ( pdes ) . in particular , for a horn - type hypergeometric function , the system of pdes can be derived from the coefficients of the series . in this case , the ratio of two coefficients can be represented as a ratio of two polynomials , @xmath35 so that the horn - type hypergeometric function satisfies the following system of differential equations : @xmath36 h(\vec{\gamma},\vec{\sigma } , \vec{x } ) \ ; , \label{diff}\ ] ] where @xmath37 .
it was pointed out in several publications @xcite that ( i ) the differential reduction algorithm , eq .
( [ reduction ] ) , can be applied to reduction of feynman diagrams to some subset of basis hypergeometric functions with well - known analytical properties @xcite ; ( ii ) the system of differential equations , eq .
( [ diff ] ) , can be used also for construction of , so - called , @xmath38-expansion of hypergeometric functions around rational values of parameters via direct solution of the system of differential equations @xcite .
this is another motivation for creating the package for manipulation with parameters of horn - type hypergeometric functions .
the aim of this paper is to present the _
@xcite based package * hyperdire * for the differential reduction of the horn - type hypergeometric function with arbitrary values of parameters to a set of basis functions .
the current version consists of two parts : one , * pfq * , for manipulation with hypergeometric function , @xmath0 , and the second one , * appellf1f4 * , for manipulations with appell functions , @xmath1 .
the algorithm of differential reduction for other functions can be implemented as an additive module .
in contrast to the recent programs written by members of computational particles physics @xcite , the aim of our package is the manipulation with hypergeometric functions without construction of @xmath38-expansion @xcite . the preliminary version of @xmath39 was presented in @xcite and is available in @xcite .
the latest version is available in @xcite .
let us consider the generalized hypergeometric function , @xmath41 , defined around @xmath42 by a series @xmath43 where @xmath44 is a pochhammer symbol , @xmath45 .
the sequences @xmath46 and @xmath47 are called the upper and lower parameters of hypergeometric functions , respectively . in terms of the operator @xmath48 : @xmath49 the differential equation for the hypergeometric function @xmath50 can be written as @xmath51 { } _ pf_q(\vec{a } ; \vec{b } ; z ) = 0 .
\label{diffpfq}\ ] ] the differential reduction for these functions was analyzed in details in @xcite .
here we remind some of the main relations relevant to our program .
the universal differential operators , eqs .
( [ do1]),([do2 ] ) , have the following form : @xmath52 where the operators @xmath53 are called the step - up ( step - down ) operators for the upper ( lower ) parameters of hypergeometric functions .
this type of operators were explicitly constructed for the hypergeometric function @xmath40 by takayama in ref .
@xcite . for completeness
, we reproduce his result here : @xmath54 where @xmath55 \right|_{a_i \to a_i - 1 } \ ; , \nonumber \\ c_i & = & -a_i \pi_{j=1}^{p } ( b_j-1-a_i ) \ ; , \nonumber \\ t_i ( x ) & = & \frac{x \pi_{j=1}^p ( x + b_j - 1 ) - c_i } { x+a_i } = \sum_{j=0}^p p^{(p)}_{p - j}(\{b_r\!-\!1\ } ) \frac{\left [ x^{j+1 } \!-\ ! ( -a_i)^{j+1 } \right]}{x\!+\!a_i } \nonumber \\ & = & \sum_{j=0}^p p^{(p)}_{p - j}(\{b_r\!-\!1\ } ) \sum_{k=0}^j x^{j - k}(-a_i)^k \ ; , \label{diff : oper:1}\end{aligned}\ ] ] and @xmath56 \right|_{b_i \to b_i + 1 } \ ; , \nonumber \\ d_i & = & \pi_{j=1}^{p+1 } ( 1+a_j - b_i ) \ ; , \nonumber \\ s_i ( x ) & = & \frac{\pi_{j=1}^{p+1 } ( x + a_j ) - d_i } { x+b_i-1 } \sum_{j=0}^{p+1 } p^{(p+1)}_{p+1-j}(\{a_r\ } ) \frac{\left [ x^{j } \!-\ ! ( 1\!-\!b_i)^{j } \right]}{x\!-\!(1\!-\!b_i ) } \nonumber \\ & = & \sum_{j=0}^p p^{(p+1)}_{p - j}(\{a_r\ } ) \sum_{k=0}^j x^{j - k}(1\!-\!b_i)^k , \label{diff : oper:2}\end{aligned}\ ] ] where @xmath57 means substitution of @xmath58 by @xmath59 , and polynomials @xmath60 are defined as @xmath61 @xmath62 is a polynomial of an order of @xmath63 with respect to variables @xmath64 , and @xmath65 for example , @xmath66 and @xmath67 . keeping in mind that @xmath68 we get that these polynomials satisfy the following relations : @xmath69 where @xmath70 . in particular
, @xmath71 the differential reduction has the form of a product of several differential step - up / step - down operators , @xmath72 and @xmath73 , respectively : @xmath74 so that maximal power of @xmath48 in this expression is equal to @xmath75 . since the hypergeometric function @xmath76 satisfies the differential equation of an order of @xmath77 ( see eq .
( [ diff ] ) ) : @xmath78 \theta^r + z \pi_{k=1}^{p+1 } a_k \biggr\ } { } _ { p+1}f_{p}(\vec{a};\vec{b } ; z ) \;,\end{aligned}\ ] ] it is possible to express all terms containing higher power of the operator @xmath79 , where @xmath80 , in terms of rational functions of parameters and argument @xmath81 times by @xmath82 , where @xmath83 . in this way ,
any function @xmath84 is expressible in terms of the basic function and its first @xmath85-derivative : @xmath86 where @xmath87 is the set of integer numbers , and @xmath88 and @xmath89 are polynomials in the parameters @xmath90 and @xmath81 . from eq .
( [ diff : oper:1 ] ) follows that if one of the upper parameters @xmath91 is equal to unity , then the application of the step - down operator @xmath92 to the hypergeometric function @xmath40 will produce unity , @xmath93 taking into account the explicit form of the step - down operator @xmath94 , @xmath95 \;,\end{aligned}\ ] ] we get the differential identity @xmath96 \theta^j \biggr\ } { } _ { p+1}f_{p}(1 , \vec{a};\vec{b } ; z ) = \pi_{k=1}^p(b_k\!-\!1 )
\;. \label{relation}\end{aligned}\ ] ] the case when two or more upper parameters are equal to unity , @xmath97 , does not generate any new identities .
let us consider the system of linear differential equations of the second order for the functions @xmath98 : @xmath99 where @xmath100 and @xmath101 are variables , @xmath102 are rational functions , @xmath103 for @xmath104 and @xmath105 the exploring of @xmath106 instead of the standard @xmath107 is explained by our applications . taking the derivative of eq.([eq1 ] ) with respect to @xmath108 , using the well - known property @xmath109 and applying eq .
( [ eq2 ] ) , we rewrite the first equation as follows : @xmath110 \theta_{12 } \omega(\vec{z } ) \nonumber \\ & & = \biggl\ { \left [ \theta_2 p_0 \!+\ ! p_1 \!+\ !
p_2 r_0 \right ] \theta_{12 } \!+\ ! \left [ p_2 r_1 \!+\ ! \theta_2 p_1 \right ] \theta_1 \!+\ ! \left [ p_2 r_2 \!+\ ! \theta_2 p_2 \!+\ !
p_3 \right ] \theta_2 \!+\ !
p_2 r_3 \!+\ ! \theta_2 p_3 \biggr\ } \omega(\vec{z } ) \ ; , \nonumber \\ & & \label{eq1a}\end{aligned}\ ] ] repeating a similar operation for eq .
( [ eq2 ] ) we get @xmath111 \theta_{12 } \omega(\vec{z } ) \nonumber \\ & & = \biggl\ { \left [ \theta_1 r_0 \!+\ ! r_2 \!+\ ! r_1 p_0 \right ] \theta_{12 } \!+\ ! \left [ p_1 r_1 \!+\ ! \theta_1 r_1 \!+\ !
r_3 \right ] \theta_1 \!+\ ! \left [ p_2 r_1 \!+\ ! \theta_1 r_2 \right ] \theta_2 \!+\ !
r_1 p_3 \!+\
! \theta_1 r_3 \biggr\ } \omega(\vec{z } ) \ ; , \nonumber \\ & & \label{eq2a}\end{aligned}\ ] ] it is well known @xcite that under condition @xmath112 there are 4 independent solutions of system ( [ eq1]),([eq2 ] ) . in this case , eqs .
( [ eq1a]),([eq2a ] ) can be solved so that @xmath113 \theta^2_{12 } \nonumber \\ & & \!+\ ! \left [ p_2 r_1 \!+\ ! \theta_2 p_1 \!+\ !
p_0 \left ( p_1 r_1 \!+\ !
\theta_1 r_1 \!+\ !
r_3 \right ) \right ] \theta_1 \nonumber \\ & & \!+\ ! \left [ p_2 r_2 \!+\ ! \theta_2 p_2 \!+\ ! p_3 \!+\ !
p_0 \left ( p_2 r_1 \!+\ !
\theta_1 r_2 \right ) \right ] \theta_2 \nonumber \\ & & + p_2 r_3 \!+\ ! \theta_2 p_3 \!+\ !
p_0 \left ( r_1 p_3 \!+\ !
\theta_1 r_3 \right ) \biggr\ } \omega(\vec{z } ) \ ; , \label{eq1b } \\ ( 1 \!-\ ! p_0 r_0 ) \theta_{122 } \omega(\vec{z } ) & = & \biggl\ { \left [ \theta_1 r_0 \!+\ ! r_2 \!+\ !
r_1 p_0 \!+\ !
r_0 \left ( \theta_2 p_0 \!+\ !
p_1 \!+\ ! p_2 r_0 \right ) \right ] \theta^2_{12 } \nonumber \\ & & \!+\ ! \left [ r_0 \left ( p_2 r_1 \!+\ !
\theta_2 p_1 \right ) \!+\ ! p_1 r_1 \!+\ !
\theta_1 r_1 \!+\ ! r_3 \right ] \theta_1 \nonumber \\ & & \!+\ ! \left [ p_2 r_1 \!+\ ! \theta_1 r_2 \!+\ ! r_0 \left ( p_2 r_2 \!+\ ! \theta_2 p_2 \!+\ ! p_3 \right ) \right ] \theta_2 \nonumber \\ & & + r_0 \left ( p_2 r_3 \!+\ !
\theta_2 p_3 \right ) + r_1 p_3 \!+\ ! \theta_1 r_3 \biggr\ } \omega(\vec{z } ) \;. \label{eq2b}\end{aligned}\ ] ]
the condition of complete integrability is defined from the relation @xmath114 , or in terms of the quantities ( [ eq1b ] ) and ( [ eq2b ] ) it has the following form : @xmath115\right ) - \theta_1 \left ( \frac{1}{1-p_0 r_0 } \left[\mbox{r.h.s.eq.}~(\ref{eq2b } ) \right]\right ) \nonumber \\ & & = \frac{1}{1-p_0 r_0 } \left[\mbox{r.h.s.eq.}~(\ref{eq1b } ) \right ] - \frac{1}{1-p_0 r_0 } \left[\mbox{r.h.s.eq.}~(\ref{eq2b } ) \right ] \nonumber \\ & & \hspace{5 mm } + \theta_2 \left ( \theta_{12 } \omega(\vec{z } ) \right ) - \theta_1 \left ( \theta_{12 } \omega(\vec{z } ) \right ) \ ; , \label{full}\end{aligned}\ ] ] where we have used the following relation @xmath116 . if the condition ( [ full ] ) is valid , eqs .
( [ eq1 ] ) and ( [ eq2 ] ) can be reduced to the pfaff system of four differential equations @xmath117 where @xmath118 in the case @xmath119 @xmath120 are expressible in terms of three other elements : @xmath121 . in particular , in terms of the notation of eqs.([eq1a ] ) , ( [ eq2a ] ) , @xmath122 \theta_{12 } \omega(\vec{z } ) \nonumber \\ & = & \biggl\ { - \left [ p_0 \left ( \theta_1 r_1 + r_3 + r_1 p_1 \right ) + p_2 r_1 + \theta_2 p_1 \right ] \theta_1 \nonumber \\ & & - \left [ p_0 \left ( \theta_1 r_2 + r_1 p_2 \right ) + p_2 r_2 + \theta_2 p_2 + p_3 \right ] \theta_2 \nonumber \\ & & - \left [ p_0 \left ( \theta_1 r_3 + r_1 p_3 \right ) + p_2 r_3 + \theta_2 p_3 \right ] \biggr\ } \omega(\vec{z } ) \;.
\label{aux}\end{aligned}\ ] ] in this case , the integrability conditions are valid and eqs .
( [ eq1 ] ) and ( [ eq2 ] ) can be reduced to the pfaff system ( [ pfaff ] ) of three differential equations @xmath123 and the system has 3 solutions . for appell s hypergeometric functions @xmath124 and @xmath125 the values of coefficients of eqs .
( [ eq1 ] ) , ( [ eq2 ] ) are collected in table [ tab:1 ] .
@xmath126 .[tab:1 ] [ cols="^,^,^,^,^",options="header " , ] @xmath126 it is not surprising that the set of exceptional values of the parameters coincides with the set of parameters defining the condition of irreducibility of monodromy group of the corresponding hypergeometric functions ( see @xcite and references therein ) .
we are expecting that the condition of irreducibility of the mellin - barnes integral @xcite can be used as a criterion of irreducibility of feynman diagram .
in this section , we will present the mathematica . ] based package * pfq * for differential reduction of the hypergeometric function @xmath127 .
in contrast to the version presented in @xcite the current version deals with non - exceptional and exceptional values of parameters . the takayama algorithm is implemented in file `` * pfq.m * '' , and an example of application is given in the file `` example - pfq.m '' .
the program could be loaded in a standard way : @xmath128 and includes two routines : * togroebnerbasis*[@xmath129 and * explicitform*[@xmath129 .
the main routine @xmath130 , \label{togroebnerbasis}\ ] ] calculates explicitly the ratio of the functions @xmath131 and @xmath88 of eq .
( [ decomposition ] ) .
the `` argumentsvector '' in eq .
( [ togroebnerbasis ] ) is the set of parameters of the hypergeometric function in the l.h.s of eq .
( [ decomposition ] ) with an explicit set of integer numbers , by which each parameter should be shifted : @xmath132 where @xmath133 are integers , @xmath134 are any symbols , and @xmath135 denotes the argument of the hypergeometric function .
the output of * togroebnerbasis * [ ] has the following structure : @xmath136 where * @xmath137 are the set of parameters and the argument of the resulting hypergeometric function entering in r.h.s . of eq .
( [ decomposition ] ) ; * @xmath138 are the rational functions where the numbering corresponds to power of @xmath139 , @xmath140 ; * @xmath141 is an overall factor . the explicit form of eqs .
( [ togroebnerbasis])-([output:1 ] ) is the following : @xmath142 1 .
* example 1 * ] .
[ simpl ] ] : reduction of @xmath143 + [ \{\{1+a@xmath144,2+a@xmath145 , a@xmath146},\{1+b@xmath144 , b@xmath145 + 2},x } ] + , + @xmath147 + @xmath148 + corresponds to @xmath149 { } _ { 3}f_{2}\left(\begin{array}{c| } a_{1 } , a_2 , a_3 \\
b_{1}+1 , b_{2}+1 \end{array } ~z
\right ) \;.\end{aligned}\ ] ] 2 .
* example 2 * : reduction of @xmath150 + [ \{\{1+a@xmath144,1+a@xmath145 , a@xmath146,a@xmath151},\{1+b@xmath144 , b@xmath145 + 1,b@xmath146},x } ] , + + @xmath152 + corresponds to @xmath153 { } _ { 4}f_{3}\left(\begin{array}{c| } a_1,a_2,a_3,a_4 \\ b_1 + 1,b_2 + 1,b_3 + 1 \end{array }
~x \right ) \;.\end{aligned}\ ] ] in both these examples , @xmath154 means the set of integer values of parameters in the original hypergeometric function , e.g. @xmath155 in example 1 , and @xmath156 corresponds to values of the parameters which should be changed , e.g. @xmath157 in example 2 .
routine * explicitform * [ ] convert the results of reduction , eq .
( [ basis ] ) , to mathematica - standard expression for generalized hypergeometric function .
+ + * example 3 * : reduction of @xmath158 .
+ = * togroebnerbasis * [ \{1+@xmath159,1+@xmath160},\{1+@xmath161},x } ] ; + + [ ] + @xmath162 for non - exceptional values of the parameters the differential reduction is performed with the help of eqs .
( [ universal : a]),([universal : b]),([diff : oper:1]),([diff : oper:2 ] ) .
the higher powers of the operators @xmath79 are expressed with the help of the differential equation for the hypergeometric function , eq .
( [ diffpfq ] ) . also , the following relation is used in some cases : @xmath163 when some of the upper parameters of the initial hypergeometric function are integer , then the higher power of the differential operator can be excluded with the help of eq .
( [ relation ] ) . in this case , the input of * togroebnerbasis * [ ] does not change , and the output of * togroebnerbasis * [ ] has the following structure ( case when only one upper parameter is integer ) : @xmath164 where * @xmath165 are the set of parameters and argument of the resulting hypergeometric function ; * * if upper parameter is integer , the proper @xmath91 is equal to zero * ; * @xmath166 are the rational functions where the numbering corresponds to power of @xmath139 , @xmath167 ; * @xmath141 is overall factor ; * @xmath168 is the resulting polynomial entering in the r.h.s .
( [ relation ] ) . in the explicit form @xmath169 where @xmath170 is a set of integers .
+ + * example 4 * : reduction of @xmath143 with an integer parameter .
+ [ \{\{3,1+a@xmath145,1+a@xmath146},\{2+b@xmath144,2+b@xmath145},x } ] ; + + @xmath171 + @xmath172 in an explicit form it is @xmath173 { } _ { 3}f_{2}\left(\begin{array}{c| } 1,a_2 + 1,a_3 + 1 \\
b_1 + 2,b_2 + 2 \end{array } ~x
\right ) \nonumber \\ & & - \frac{\left(b_1 + 1\right ) \left(b_2 + 1\right)}{2 ( x-1 ) } \;. \nonumber\end{aligned}\ ] ] other useful relations which are implemented in the function * togroebnerbasis * [ ] are the relations derived in @xcite : @xmath174 where @xmath175 are positive integers , @xmath176 , and @xmath177 + + * example 5 * : reduction of @xmath143 with integer difference values of parameters .
+ [ \{\{3+b@xmath144,1+a@xmath145,1+a@xmath146},\{2+b@xmath144,2+b@xmath145},x } ] ; + + @xmath178 + in an explicit form : @xmath179 { } _ { 2}f_{1}\left(\begin{array}{c| } a_2,a_3 \\ b_2 \!+\ ! 1 \end{array } ~x \right ) \;. \nonumber\end{aligned}\ ] ] the following relations : @xmath180 \;,\end{aligned}\ ] ] where @xmath181 and @xmath182 are implemented in mathematica 7.0 .
in this section , we will present the mathematica . ] based program * appellf1f4 * for differential reduction of the appell hypergeometric functions @xmath183 , @xmath184 , @xmath185 and @xmath125 .
the program is available in @xcite .
the current version , @xmath186 , deals with non - exceptional values of a parameters only .
the program could be loaded in a standard way : @xmath187 the package includes the following basic routines : @xmath188 , \label{input : f1 } \\ & & { \bf f2indexchange}[\mbox{changingvector } , \mbox{parametervector } ] , \label{input : f2 } \\ & & { \bf f3indexchange}[\mbox{changingvector } , \mbox{parametervector } ] , \label{input : f3 } \\ & & { \bf f4indexchange}[\mbox{changingvector } , \mbox{parametervector } ] .
\label{input : f4}\end{aligned}\ ] ] and @xmath189 \;.\ ] ] the @xmath190 in eq .
( [ input : f1])-([input : f4 ] ) is the set of integers at which we want to change the values of parameters of the appell functions ( vector @xmath191 in eqs .
( [ reductionf1 ] ) , ( [ reductionf234 ] ) ) .
the set of parameters of appell s function are defined in the list @xmath192 ( corresponds to vector @xmath193 and arguments @xmath194 in the l.h.s .
( [ reductionf1 ] ) , ( [ reductionf234 ] ) ) .
we want to point out that enumeration of parameters and arguments in the @xmath192 correspond one - to - one to eqs .
( [ definition : f1 ] ) , ( [ definition : f2 ] ) , ( [ definition : f3 ] ) , ( [ definition : f4 ] ) .
the output of @xmath195 $ ] and @xmath196 $ ] , has a little different structure .
the structure of output of @xmath195 $ ] is the following : @xmath197 where 1 .
@xmath198 is the set of new parameters of appell function @xmath183 ; 2 .
@xmath199 are the rational functions corresponding to the ratio of @xmath200 and @xmath201 of functions entering in eq .
( [ reductionf1 ] ) .
* example 6 * : .
] reduction of @xmath183 .
+ + * f1indexchange[*\{@xmath202,@xmath203,@xmath204,@xmath204},\{@xmath58,@xmath161,@xmath205,@xmath206,@xmath207,@xmath208 } * ] * + + @xmath209 + + in an explicit form : @xmath210 \nonumber \\ & & \times f_1(a+1,b_1 - 1,b_2,c ; z_1,z_2).\end{aligned}\ ] ] routine * explicitformf1*[@xmath211 converts the result of reduction to mathematica standard expression for the appell function @xmath183 .
+ + * example 7 * : reduction of @xmath212 .
+ = * f1indexchange[*\{@xmath202,@xmath203,@xmath204,@xmath204},\{@xmath58,@xmath161,@xmath205,@xmath206,z@xmath144,z@xmath145 } * ] * + + * explicitformf1 * [ ] + + @xmath213 + @xmath214 + @xmath215 the output of @xmath216 $ ] , @xmath217 $ ] , and @xmath218 $ ] , is similar and has the following structure : @xmath219 where 1 .
@xmath220 is the name of appell functions under reduction : @xmath221 or @xmath222 ; 2 .
@xmath198 is the set of new parameters of the appell function ; 3 .
@xmath223 are the rational functions corresponding to the ratio of @xmath224 and @xmath225 of functions entering in eq .
( [ reductionf234 ] ) .
* example 8 * : reduction of @xmath226 .
+ + * f2indexchange[*\{@xmath204,@xmath204,@xmath202,@xmath202,@xmath204},\{a , b@xmath144,b@xmath145,c@xmath144,c@xmath145,z@xmath144,z@xmath145 } * ] * + + @xmath227 + @xmath228 + + in an explicit form : @xmath229 f_2(a , b_1,b_2 + 1,c_1 + 1,c_2;z_1,z_2 ) \;.\end{aligned}\ ] ] the appell functions @xmath184 , @xmath185 and @xmath125 , are not built in the current version of mathematica ( version 7.0 ) , so that the * appellf1f4 * package does not include a function similar to * explicitformf1[]*.
the standard way of analytical evaluation of multiloop feynman diagrams is based on exploring of integration by parts relations @xcite for the reduction of original diagram to set so - called master - integrals .
this reduction can be performed in an arbitrary space - time dimension ( before construction of @xmath230-expansion ) .
now there are a few ways of computer realization of this algorithm @xcite . in a series of papers
@xcite it was pointed out that the hypergeometric representation of feynman diagram in conjunction with the differential reduction algorithm allow one to make reduction directly in terms of hypergeometric functions . in particular , it was formulated the conjecture about the structure of the hypergeometric representation of feynman diagram ( see as review @xcite ) .
let us consider the hypergeometric representation of the feynman diagram @xmath231 , @xmath232 where @xmath233 are the horn - type hypergeometric functions defined by eq .
( [ h ] ) , @xmath234 is a list of the powers of the propagators in the feynman diagram , @xmath235 is the space - time dimension , @xmath236 are the arguments of the hypergeometric functions which are related to the kinematic invariants of the feynman diagram , @xmath237 are the linear combinations of @xmath234 and @xmath235 with a polynomial coefficients , and @xmath238 are the rational functions of the variables @xmath239 with the coefficients depending on @xmath235 and @xmath234 . in general , the r.h.s . of eq .
( [ f - h ] ) can include the different types of the hypergeometric functions ( e.g. @xmath240 and @xmath127 , where @xmath85 is integer , are allowed ) .
our proposition is that : ( i ) regardless the type of functions in the r.h.s . of eq .
( [ f - h ] ) the number of basis elements is the same ( up to a module of rational functions ) for each term of the r.h.s . of eq .
( [ f - h ] ) ; ( ii ) the number @xmath241 of nontrivial master integrals following from integration - by - parts relations which are not expressible in terms of gamma functions is then equal to the number of basis elements @xmath242 for each term of the r.h.s .
( [ f - h ] ) .
as illustration of the differential reduction algorithm and our conjecture , let us consider the diagrams shown in figs.[diagrams ] , [ diagrams2 ] and [ diagrams3 ] .
let us start from consideration of the two - loop propagator diagram @xmath243 depicted in fig .
( [ diagrams ] ) with one massive line and three - massless lines : @xmath244^{j_1 } [ ( k_1-k_2)^2]^{\sigma_1 } [ k_1 ^ 2]^{\sigma_2 } [ k_2 ^ 2\!-\!m^2]^{j_4 } } \;. \nonumber \\ \label{v000m}\end{aligned}\ ] ] such diagrams appear as one of the master - integrals in pole mass of heavy quarks .
the mellin - barnes representation of eq .
( [ v000 m ] ) reads : @xmath245 ^ 2 } { ( p^2)^{j_1\!+\!\sigma_1\!+\!\sigma_2\!+\!j_4\!-\!n } } \frac{\gamma(\frac{n}{2}\!-\!j_1 ) \gamma(\frac{n}{2 } \!-\ !
\sigma_1 ) \gamma(\frac{n}{2 } \!-\ !
\sigma_2 ) \gamma(\sigma_1\!+\!\sigma_2\!-\!\frac{n}{2 } ) } { \gamma(j_1 ) \gamma(\sigma_1 ) \gamma(\sigma_2 ) \gamma(j_4 ) \gamma(n\!-\!\sigma_1\!-\!\sigma_2 ) } \nonumber \\ & & \hspace{-5 mm } \times \int dt \left ( - \frac{m^2}{p^2 } \right)^t \frac { \gamma(-t ) \gamma(t\!+\!j_4 ) \gamma(n \!-\ ! \sigma_1 \!-\ ! \sigma_2 \!-\
! j_4 \!-\ !
t ) \gamma(j_1 \!+\
! \sigma_1 \!+\ ! \sigma_2 \!+\ ! j_4 \!-\ !
n \!+\ ! t ) } { \gamma \left(\sigma_1 \!+\ ! \sigma_2 \!+\ ! j_4 \!-\ !
\tfrac{n}{2 } \!+\ ! t \right ) \gamma \left(\tfrac{3n}{2 } \!-\
! \sigma_1 \!-\
! \sigma_2 \!-\
! j_4 \!-\ !
t \right ) } \;. \label{mb : v000m}\end{aligned}\ ] ] closing the contour of integration in eq .
( [ mb : v000 m ] ) on the right , we obtain @xmath246 ^ 2 } { ( p^2)^{\!j_1\!+\!\sigma_1\!+\!\sigma_2\!+\!j_4\!-\!n } } \frac{\gamma(\frac{n}{2}\!-\!j_1 ) \gamma(\frac{n}{2 } \!-\ !
\sigma_1 ) \gamma(\frac{n}{2 } \!-\ !
\sigma_2 ) \gamma(\sigma_1\!+\!\sigma_2\!-\!\frac{n}{2 } ) } { \gamma(j_1 ) \gamma(\sigma_1 ) \gamma(\sigma_2 ) \gamma(j_4 ) \gamma(n\!-\!\sigma_1\!-\!\sigma_2 ) } \nonumber \\ & & \biggl\ { \left(-\frac{p^2}{m^2 } \right)^{j_4 }
\frac{\gamma(j_4 ) \gamma(n\!-\!\sigma_1\!-\!\sigma_2 ) \gamma(j_1\!+\!\sigma_1\!+\!\sigma_2\!-\!n ) } { \gamma(\sigma_1 \!+ \ !
\sigma_2-\frac{n}{2 } ) \gamma(\frac{3n}{2}\!-\!j_1\!-\!\sigma_1\!-\!\sigma_2 ) } \nonumber \\ & & \hspace{10 mm } \times \left .
_ 3f_2\left ( \begin{array}{c } j_4 , \ ; n\!-\!\sigma_1\!-\!\sigma_2 , \ ; \frac{n}{2}\!+\!1\!-\!\sigma_1\!-\!\sigma_2 \\ n\!+\!1\!-\!j_1\!-\!\sigma_1\!-\!\sigma_2 , \ ; \frac{3n}{2}\!-\!j_1\!-\!\sigma_1\!-\!\sigma_2 \end{array } \right| \frac{p^2}{m^2 } \right ) \nonumber \\ & & + \left(-\frac{p^2}{m^2 } \right)^{j_1\!+\!\sigma_1\!+\!\sigma_2\!+\!j_4\!-\!n } \frac{\gamma(j_1 ) \gamma(n\!-\!j_1\!-\!\sigma_1\!-\!\sigma_2 ) \gamma(j_1\!+\!\sigma_1\!+\!\sigma_2\!+\!j_4\!-\!n ) } { \gamma(\frac{n}{2 } ) \gamma(\frac{n}{2}\!-\!j_1 ) } \nonumber \\ & & \hspace{10 mm } \times \left .
_ 3f_2\left ( \begin{array}{c } j_1 , \ ; j_1\!+\!\sigma_1\!+\!\sigma_2\!+\!j_4\!-\!n , \ ; 1\!+\!j_1\!-\!\frac{n}{2 } \\ 1\!+\!j_1\!+\!\sigma_1\!+\!\sigma_2\!-\!n , \frac{n}{2 } \end{array } \right|\frac{p^2}{m^2 } \right ) \biggr\ } \;.
\label{v0002}\end{aligned}\ ] ] both hypergeometric functions in eq . (
[ v0002 ] ) are reducible to the @xmath247 functions with one unity upper parameter , namely , @xmath248 respectively , plus rational functions of @xmath81 . in accordance with section [ count
] , there is only one nontrivial master - integral in the framework of integration - by - parts relations and a few integrals expressible in terms of the product of @xmath249-functions . in accordance with tarasov s algorithm @xcite
there is one master integral of the same topology and two integrals expressible in terms of the product of @xmath249-functions @xcite . for completeness
, we will present a few first terms of @xmath230-expansion of the master - integral . for @xmath250 , where @xmath230 is the parameter of dimensional regularization @xcite and @xmath251
, we have @xmath252 where we use @xmath253 for the second @xmath254 hypergeometric function the all - order @xmath230-expansion are available : @xmath255 where @xmath256 and @xmath257 are classical and nielsen polylogartihms , respectively . for the first hypergeometric function
we apply the following relation : @xmath258 the first coefficients of the @xmath230 expansion read : @xmath259 \biggr \ } \nonumber \\ & & + { \varepsilon}^3 \biggl\ { 9 + \frac{4}{3 } \ln^3 ( -z ) + 8 \zeta_3 + 8 \zeta_2 \ln ( -z ) \nonumber \\ & & \hspace{10 mm } + 2 \frac{1-z}{z } \biggl [ 4 { { \mbox{s}}_{1,2\!}\left(z\right ) } \!+\ !
\frac{2}{3 } \ln^3 ( 1\!-\!z ) \!+\ !
4 \ln ( 1\!-\!z ) { { \mbox{li}}_{2}\left(z\right ) } \!-\ ! 5 { { \mbox{li}}_{3}\left(z\right ) } \nonumber \\ & & \hspace{15 mm } \!+\ ! 2 \zeta_2 \ln ( 1\!-\!z ) \!-\ !
\ln ( -z ) { { \mbox{li}}_{2}\left(z\right ) } \!-\ ! \ln ( -z ) \ln^2 ( 1\!-\!z ) \!+\ !
\ln^2 ( -z ) \ln ( 1\!-\!z ) \biggr ] \biggr \ } \nonumber \\ & & + { \varepsilon}^4 \biggl\ { 27 + \frac{2}{3 } \ln^4 ( -z ) + 40 \zeta_4 + 8 \zeta_2 \ln^2 ( -z ) + 16 \zeta_3 \ln ( -z ) \nonumber \\ & & \hspace{10 mm } - 2 \frac{1-z}{z } \biggl [ 8 { { \mbox{s}}_{1,3\!}\left(z\right ) } - 10 { { \mbox{s}}_{2,2\!}\left(z\right ) } + 14 { { \mbox{li}}_{4}\left(z\right ) } + \frac{1}{3 } \ln^4 ( 1-z ) + 2 \zeta_2 { { \mbox{li}}_{2}\left(z\right ) } \nonumber \\ & & \hspace{15 mm } + 8 \ln ( 1-z ) { { \mbox{s}}_{1,2\!}\left(z\right ) } - 10 \ln ( 1-z ) { { \mbox{li}}_{3}\left(z\right ) } - 4 \zeta_3 \ln ( 1-z ) \nonumber \\ & & \hspace{15 mm } + 4 \ln^2 ( 1-z ) { { \mbox{li}}_{2}\left(z\right ) } + 2 \zeta_2 \ln^2 ( 1-z ) + \ln^2 ( -z ) \ln^2 ( 1-z ) \nonumber \\ & & \hspace{15 mm } - 2 \ln ( -z ) { { \mbox{s}}_{1,2\!}\left(z\right ) } + \ln ( -z ) { { \mbox{li}}_{3}\left(z\right ) } - \frac{2}{3 } \ln ( -z ) \ln^3 ( 1-z ) \nonumber \\ & & \hspace{15 mm } - 4 \zeta_2 \ln ( -z ) \ln ( 1-z ) - 2 \ln ( -z ) \ln ( 1-z ) { { \mbox{li}}_{2}\left(z\right ) } \nonumber \\ & & \hspace{15 mm } + \ln^2 ( -z ) { { \mbox{li}}_{2}\left(z\right ) } - \frac{2}{3 } \ln^3 ( -z ) \ln ( 1-z ) \biggr ] \biggr \ } + { \cal o } ( { \varepsilon}^5 ) \biggr ] \ ; , \label{v0002:ep}\end{aligned}\ ] ] where @xmath260 . to check eq .
( [ v0002:ep ] ) , we evaluate the first few coefficients of the @xmath230 expansion of the original diagram in the large - mass limit @xcite using the program packages developed in refs .
@xcite to find agreement .
let us consider the 3-loop bubble diagram @xmath261 depicted in fig .
( [ diagrams ] ) with two massive lines and three massless ones : @xmath262^{\alpha_1 } [ k_1 ^ 2]^{\alpha_2 } [ ( k_2\!-\!k_3)^2]^{\beta } [ k_3 ^ 2\!-\ ! m^2]^\lambda [ k_2 ^ 2 \!-\ !
m^2]^\sigma } \;.
\nonumber \\ & & \ ] ] its mellin - barnes representation reads : @xmath263 ^ 3 ( -m^2)^{n - a_{1,2,\lambda } } ( -m^2)^{n/2 - a_{\sigma,\beta } } \nonumber \\ & & \times \frac { \gamma \left ( \frac{n}{2 } \!-\ !
\alpha_1 \right ) \gamma\left ( \frac{n}{2 } \!-\ ! \alpha_2 \right ) \gamma\left ( \frac{n}{2 } \!-\ !
\beta \right ) \gamma\left(a_{1,2 } \!-\ !
\frac{n}{2 } \right ) } { \gamma(\alpha_1 ) \gamma(\alpha_2 ) \gamma(\sigma ) \gamma(\lambda ) \gamma(\beta ) \gamma \left ( \frac{n}{2 } \right ) \gamma(n \!-\ !
a_{1,2 } ) } \nonumber \\ & & \times \int dt \left ( \frac{m^2}{m^2 } \right)^t \frac { \gamma(-t ) \gamma(\beta \!+\ !
t ) \gamma\left(a_{\beta , \sigma } \!-\ ! \tfrac{n}{2 } \!+\ ! t \right ) \gamma\left(n \!-\ !
a_{1,2 } \ ! + \ !
t \right ) \gamma\left(a_{1,2,\lambda } \!-\ !
t \right ) } { \gamma\left ( t \!+\ ! \tfrac{n}{2 } \right ) } \ ; , \label{mb : e10002}\end{aligned}\ ] ] where @xmath264 closing the contour of integration in eq .
( [ mb : e10002 ] ) on the right , we obtain @xmath265 ^ 3 ( -m^2)^{n - a_{1,2,\lambda } } ( -m^2)^{n/2 - a_{\sigma,\beta } } \nonumber \\ & & \times \frac { \gamma \left ( \frac{n}{2 } \!-\ ! \alpha_1
\right ) \gamma\left ( \frac{n}{2 } \!-\ !
\alpha_2 \right ) \gamma\left ( \frac{n}{2 } \!-\ !
\beta \right ) \gamma\left(a_{1,2 } \!-\ !
\frac{n}{2 } \right ) } { \gamma(\alpha_1 ) \gamma(\alpha_2 ) \gamma(\sigma ) \gamma(\lambda ) \gamma(\beta ) \gamma \left ( \frac{n}{2 } \right ) \gamma(n \!-\ !
a_{1,2 } ) } \nonumber \\ & & \biggl [ \frac{\gamma(\beta ) \gamma \left ( a_{\sigma,\beta } \!-\ !
\frac{n}{2 } \right ) \gamma(n \!-\ ! a_{1,2 } ) \gamma(a_{1,2,\lambda } \!-\ !
n ) } { \gamma \left(\frac{n}{2 } \right ) } { } _ { 3}f_2\left(\begin{array}{c| } \beta , a_{\beta,\sigma } \!-\ !
\tfrac{n}{2 } , n \!-\ ! a_{1,2 } \\ \tfrac{n}{2 } , 1 \!+\ !
n - a_{1,2,\lambda } \end{array } ~ \frac{m^2}{m^2 } \right ) \nonumber \\ & & + \left ( \frac{m^2}{m^2 } \right)^{a_{1,2,\lambda } - n } \frac { \gamma(\lambda ) \gamma \left(a_{1,2,\lambda,\beta } \!-\ !
n \right ) \gamma \left(a_{1,2,\beta,\lambda,\sigma } \!-\ !
\tfrac{3n}{2 } \right ) \gamma ( n \!-\ !
a_{1,2,\lambda } ) } { \gamma \left(a_{1,2,\lambda } \!-\ !
\tfrac{n}{2 } \right ) } \nonumber \\ & & \hspace{10 mm } \times { } _ { 3}f_2\left(\begin{array}{c| } \lambda , a_{1,2,\lambda,\beta } \!-\ !
n , a_{1,2,\lambda,\beta,\sigma } \!-\ !
\tfrac{3n}{2 } \\ a_{1,2,\lambda } \!-\ !
\tfrac{n}{2 } , 1 \!+\ ! a_{1,2,\lambda } - n \end{array } ~ \frac{m^2}{m^2 } \right ) \biggr ] \;.
\label{e10002}\end{aligned}\ ] ] both hypergeometric functions in eq . ( [ e10002 ] )
are reducible to the @xmath247 functions with one unity upper parameter , namely , @xmath266 respectively , plus rational functions of @xmath81 . for this diagram
there is only one nontrivial master - integral in the framework of integration - by - parts relations and an integral expressible in terms of the product of @xmath249-functions . for completeness
, we will present a few first terms of the @xmath230-expansion of the master - integral . for @xmath250 , where @xmath230 is the parameter of dimensional regularization @xcite and @xmath267 , we have @xmath268 ^ 3 ( m^2)^{1 - 2 { \varepsilon } } ( m^2)^{-{\varepsilon } } \nonumber \\ & & \times \frac { \gamma(1-{\varepsilon } ) \gamma^2(1+{\varepsilon } ) \gamma(1 + 2{\varepsilon } ) } { 2{\varepsilon}^3 ( 1-{\varepsilon})^2 ( 1 - 2{\varepsilon } ) } \biggl [ { } _ { 2}f_1\left(\begin{array}{c| } 1 , { \varepsilon}\\ 2-{\varepsilon}\end{array } ~ \frac{m^2}{m^2 } \right ) \nonumber \\ & & + \left ( \frac{m^2}{m^2 } \right)^{-1 + 2{\varepsilon } } \frac { \gamma ( 2-{\varepsilon } ) \gamma ( 1 + 3{\varepsilon } ) } { 3 ( 1 - 3{\varepsilon } ) \gamma^2(1+{\varepsilon } ) } { } _ { 2}f_1\left(\begin{array}{c| } 1 , -1 \!+\ !
3{\varepsilon}\\ 1 \!+\ ! { \varepsilon}\end{array } ~ \frac{m^2}{m^2 } \right ) \biggr ] \;.\end{aligned}\ ] ] the all - order @xmath230-expansion for the first hypergeometric function is given by eq .
( [ all - order ] ) and for the second hypergeometric function we apply the following relation : @xmath269 the first coefficients of the @xmath230 expansion read : @xmath270 ^ 3 \gamma^3(1+{\varepsilon } ) } ( m^2)^ { 2 { \varepsilon } } ( m^2)^{-1+{\varepsilon } } = \frac { \gamma(1-{\varepsilon } ) \gamma(1 + 2{\varepsilon } ) z } { ( 1-{\varepsilon } ) ( 1 - 2{\varepsilon})^2 \gamma(1+{\varepsilon } ) } \nonumber \\ & & \times \biggl [ \frac{1}{6 { \varepsilon}^3 } ( 1 + 2 z ) + \frac{1}{6 { \varepsilon}^2 } \left[1 + 2 ( 1-z ) \ln z \right ] \nonumber \\ & & + \frac{1}{3 { \varepsilon } } \left [ \frac{3}{2 } + \ln z + ( 1 - 4 z ) \zeta_2 + ( 1-z ) \left ( \ln^2 z - 3 \ln z \ln ( 1-z ) - 3 { { \mbox{li}}_{2}\left(z\right ) } \right ) \right ] \nonumber \\ & & + \frac{3}{2 } \!+\ ! \frac{1}{3
\!+\ ! 3 \zeta_3 \!+\ ! \ln z \!+\ ! \frac{1}{3 } \ln^2 z - ( 1\!-\!z ) \biggl\ { \frac{10}{3 } \zeta_3 \!-\ ! \frac{2}{9 } \ln^3 z \!-\ ! 2 { { \mbox{s}}_{1,2\!}\left(z\right ) } \!-\ ! \frac{2}{3 } \zeta_2 \ln z \nonumber \\ & & \hspace{5 mm } + \ln(1-z ) \left [ 2 \zeta_2 - 2 { { \mbox{li}}_{2}\left(z\right ) } + \ln^2 z \right ] - \ln z
\ln^2(1-z ) + \ln z { { \mbox{li}}_{2}\left(z\right ) } \biggr\ } \nonumber \\ & & + { \varepsilon}\biggl\ { \frac{9}{2 } + \zeta_2 - \frac{1}{3 } \zeta_3 - \frac{11}{2 } \zeta_4 + \frac{2}{9 } \ln^3 z + \ln^2 z + \frac{1}{3 } ( 9 + 2 \zeta_2 ) \ln z \nonumber \\ & & \hspace{5 mm } + ( 1\!-\!z ) \biggl\ { 7 \zeta_4 - { { \mbox{li}}_{4}\left(z\right ) } + \ln z \biggl [ 2 { { \mbox{s}}_{1,2\!}\left(z\right ) } + { { \mbox{li}}_{3}\left(z\right ) } - \frac{2}{3 } \zeta_3 \biggr ] \nonumber \\ & & \hspace{20 mm } + \frac{1}{9 } \ln^4 z - 4 { { \mbox{s}}_{1,3\!}\left(z\right ) } + \ln ( 1-z ) \biggl [ 4 \zeta_3 - 4 { { \mbox{s}}_{1,2\!}\left(z\right ) } - \frac{2}{3 } \ln^3 z \biggr ] \nonumber \\ & & \hspace{20 mm } + 2 \ln(1-z ) \biggr [ \ln ( 1-z ) - \ln z \biggr ] \biggl [ \zeta_2 - { { \mbox{li}}_{2}\left(z\right ) } \biggr ] + \ln^2 z \ln^2 ( 1-z ) \nonumber \\ & & \hspace{20 mm } - \frac{2}{3 } \ln z \ln^3 ( 1-z ) - \ln^2 z \biggl [ { { \mbox{li}}_{2}\left(z\right ) } - \frac{2}{3 } \zeta_2 \biggr ] \biggr\ } \biggr\ } \biggr ] \ ; , \label{e10002:expansion}\end{aligned}\ ] ] where @xmath271 for @xmath272 , eq.([e10002:expansion ] ) coincides with the results of @xcite .
let us consider the @xmath277-loop bubble type diagram @xmath278 ( see fig .
[ diagrams2 ] ) defined as @xmath279^{\rho_1 } \cdots [ k_{r-1}^2]^{\rho_{r-1 } } [ ( k_1 \!+\ ! \cdots \!+\ !
} \!+\ ! p)^2]^{\rho_r } } \nonumber \\ & & \hspace{-5 mm } \times \frac{1 } { [ k_r^2]^{\sigma_1 } \cdots [ k_{r+x-1}^2]^{\sigma_x } [ k_{r+x}^2-m^2]^{\alpha_1 } [ ( k_r \!+\ ! \cdots \!+\ !
} \!+\ ! p)^2-m^2]^{\alpha_2 } } \nonumber \\ & & \hspace{-5 mm } \times \frac{1 } { [ k_{r+x+1}^2]^{\nu_1 } \cdots [ k_{r+x+s}^2]^{\nu_s } [ k_{q-1}^2-m^2]^{\beta_1 } [ k_{q}^2-m^2]^{\beta_2 } } \;,\end{aligned}\ ] ] where @xmath280 and @xmath281 and @xmath135 are the number of massless lines .
the mellin - barnes representation for diagram @xmath273 is : @xmath282^q ( -m^2)^{\tfrac{n}{2}\!-\!\alpha_{1,2 } } ( -m^2)^{\tfrac{n}{2}(q-1)\!-\!a_{\rho,\sigma,\nu,\beta_{1,2 } } } } { \gamma(\alpha_1 ) \gamma(\alpha_2 ) \gamma(\beta_1 ) \gamma(\beta_2 ) \gamma\left ( \frac{n}{2 } \right ) } \nonumber \\ & & \hspace{-5 mm } \times \left\ { \pi_{i=1}^{x } \frac{\gamma(\tfrac{n}{2}-\sigma_i)}{\gamma(\sigma_i ) } \right\ } \left\ { \pi_{j=1}^{r } \frac{\gamma(\tfrac{n}{2}-\rho_j)}{\gamma(\rho_j ) } \right\ } \left\ { \pi_{k=1}^{s } \frac{\gamma(\tfrac{n}{2}-\nu_k)}{\gamma(\nu_k ) } \right\ } \frac { \gamma\left(\rho-\tfrac{n}{2}(r-1)\right ) } { \gamma\left(\tfrac{n}{2}r-\rho \right ) } \nonumber \\ & & \hspace{-5 mm } \times \int dt \left
( \frac{m^2}{m^2}\right)^t \gamma\left(\sigma\!-\!\tfrac{n}{2}x \!-\!t\right ) \gamma\left(a_{\beta_1,\beta_2,\nu,\rho,\sigma}\!-\!\tfrac{n}{2}(q-1)\!-\!t\right ) \nonumber \\ & & \hspace{-5 mm } \times \frac{\gamma(\alpha_1\!+\!t ) \gamma(\alpha_2\!+\!t ) \gamma(\alpha_{1,2}\!-\!\tfrac{n}{2}\!+\!t ) \gamma(\tfrac{n}{2}\!+\!t ) \gamma(\tfrac{n}{2}(r\!+\!x ) \!-\ !
a_{\rho,\sigma}\!+\!t ) } { \gamma(\alpha_{1,2}\!+\!2 t ) \gamma(\tfrac{n}{2}(x\!+\!1)\!-\!\sigma\!+\!t ) } \nonumber \\ & & \hspace{-5 mm } \times \frac { \gamma(a_{\nu,\rho,\sigma}\!-\!\tfrac{n}{2}(q\!-\!3)\!-\!t ) \gamma(a_{\beta_1,\nu,\rho,\sigma}\!-\!\tfrac{n}{2}(q-2)\!-\!t ) \gamma(a_{\beta_2,\nu,\rho,\sigma}\!-\!\tfrac{n}{2}(q-2)\!-\!t ) } { \gamma(a_{\rho,\sigma } \!-\ !
\tfrac{n}{2}(r\!+\!x\!-\!1 ) \!-\ !
t ) \gamma(\beta_{1,2}\!+\!2 a_{\sigma,\rho,\nu } \!-\ !
n(q\!-\!2 ) \!-\ ! 2 t ) } \ ; , \label{b1122000mb}\end{aligned}\ ] ] where @xmath283 and we have introduced the short notation , @xmath284 the result could be written as a sum of @xmath285 hypergeometric functions of type @xmath286 of argument @xmath287 or @xmath288 .
the expression is really long to be considered here .
more interesting particular cases could be derived when @xmath289 take special values .
diagram @xmath290 in fig .
[ diagrams2 ] emerges from @xmath278 with @xmath291 , @xmath292 and @xmath293 . in this case ,
@xmath294 and eq . ( [ b1122000 mb ] ) is simplified due to @xmath295 the detailed analysis of this diagram was presented in @xcite . the result for diagram @xmath290
can be expressed as a sum of four hypergeometric functions of type @xmath296 or of lower order function .
the general result again is to be published .
we present only the results of differential reduction . in the case when the powers of the propagators are all integers , the appearing hypergeometric functions are reducible to the following hypergeometric functions and their derivatives : @xmath297 where @xmath298 are arbitrary integers . in the last expression , some rational functions are also generated . in accordance with section [ count ] , this integral has four master - integrals for arbitrary values of @xmath135 and @xmath277 .
further simplification is possible for ( i ) @xmath299 and ( ii ) @xmath300 . in these cases we have : * for @xmath299 , @xmath301 and in accordance with section [ count ] , this integral has two master - integrals only . * for @xmath300 @xmath302 and in accordance with section [ count ] , this integral has three nontrivial master - integrals only . for a next diagram , @xmath303 ( see fig . [ diagrams ]
) , the mellin - barnes representation follows from eq .
( [ b1122000 mb ] ) putting @xmath304 and @xmath305 . for simplicity
, we will present it an explicit form : @xmath306^q ( -m^2)^{\tfrac{n}{2}\!-\!\alpha_{1,2 } } ( -m^2)^{\tfrac{n}{2}(q-1)\!-\!a_{\sigma,\beta_{1,2 } } } } { \gamma(\alpha_1 )
\gamma(\alpha_2 ) \gamma(\beta_1 ) \gamma(\beta_2 ) \gamma\left ( \frac{n}{2 }
\right ) } \left\ { \pi_{i=1}^{q-3 } \frac{\gamma(\tfrac{n}{2}-\sigma_i)}{\gamma(\sigma_i ) } \right\ } \nonumber \\ & & \hspace{-5 mm } \times \int dt \left
( \frac{m^2}{m^2}\right)^t \gamma\left(\sigma\!-\!\tfrac{n}{2}(q\!-\!3 ) \!-\!t \right ) \gamma\left(a_{\beta_1,\beta_2,\sigma}\!-\!\tfrac{n}{2}(q\!-\!1)\!-\!t \right ) \gamma\left(\tfrac{n}{2}\!+\!t \right ) \nonumber \\ & & \hspace{-5 mm } \times \frac{\gamma(\alpha_1\!+\!t ) \gamma(\alpha_2\!+\!t ) \gamma\left(\alpha_{1,2}\!-\!\tfrac{n}{2}\!+\!t \right ) \gamma\left(a_{\beta_1,\sigma}\!-\!\tfrac{n}{2}(q\!-\!2)\!-\!t \right ) \gamma\left(a_{\beta_2,\sigma}\!-\!\tfrac{n}{2}(q\!-\!2)\!-\!t \right ) } { \gamma(\alpha_{1,2}\!+\!2 t ) \gamma\left(\beta_{1,2}\!+\!2 \sigma \!-\ !
n(q\!-\!2 ) \!-\ !
2 t \right ) } \;. \label{b11220mb}\end{aligned}\ ] ] and in agreement with eq .
( [ coefficients : b ] ) we have introduced the short notation , @xmath307 the result could be written as a sum of four hypergeometric functions of type @xmath308 of the argument @xmath287 or @xmath288 .
again result is to long to be published .
the results of differential reduction follows from eq .
( [ basis : b112200 ] ) by substitution @xmath305 : @xmath309 further simplification is possible for ( i ) @xmath310 and ( ii ) @xmath311 . in these cases
, we have : * for @xmath310 , @xmath312 and in accordance with section [ count ] , this integral has two master - integrals only * for @xmath311 @xmath313 and in accordance with section [ count ] , this integral has three nontrivial master - integrals only ( see @xcite ) . the three - loop diagram @xmath314 depicted in fig .
( [ diagrams2 ] ) corresponds to a particular case of diagram @xmath303 for @xmath315 and @xmath310 . in this case ,
@xmath316 and closing the contour of integration on the left we obtain :
@xmath317 ^ 3 ( -m^2)^{\tfrac{n}{2}\!-\!\alpha_{1,2 } } ( -m^2)^{n\!-\!\beta_{1,2 } } } { \gamma(\alpha_1 ) \gamma(\alpha_2 ) \gamma(\beta_1 ) \gamma(\beta_2 ) \gamma\left ( \frac{n}{2 } \right ) } \nonumber \\ & & \hspace{-3 mm } \biggl [ \frac { \gamma\left(\beta_1\!-\!\frac{n}{2 } \right ) \gamma\left(\beta_2\!-\!\frac{n}{2 } \right ) \gamma\left(\alpha_{1,2}\!-\!\frac{n}{2 } \right ) \gamma\left(\frac{n}{2 } \right ) \gamma(\alpha_1 ) \gamma(\alpha_2 ) } { \gamma(\alpha_{1,2 } ) } \nonumber \\ & & \hspace{-5 mm } \times ~_{6}f_5\left(\begin{array}{c| } \alpha_1 , \alpha_2 , \alpha_{1,2}\!-\!\frac{n}{2 } , \frac{n}{2 } , \frac{1\!-\!\beta_{1,2}+n}{2 } , \frac{2\!-\!\beta_{1,2}+n}{2 } \\ 1\!-\!\beta_1\!+\!\frac{n}{2 } , 1\!-\!\beta_2\!+\!\frac{n}{2 } , 1\!-\!\beta_{1,2}\!+\!n , \frac{\alpha_{1,2}}{2 } , \frac{\alpha_{1,2}+1}{2 } \end{array } ~\frac{m^2}{m^2 } \right ) \nonumber \\ & & \hspace{-6 mm } + \frac { \gamma\left(\frac{n}{2}\!-\!\beta_1 \right ) \gamma\left(\frac{n}{2}\!-\!\beta_2 \right ) \gamma\left(\beta_{1,2}\!-\!\frac{n}{2 }
\right ) \gamma\left(a_{\alpha_1,\beta_{1,2}}\!-\!n \right ) \gamma\left(a_{\alpha_2,\beta_{1,2}}\!-\!n \right ) \gamma\left(a_{\alpha_{1,2},\beta_{1,2}}\!-\!\frac{3n}{2 } \right ) } { \gamma(\alpha_{1,2}\!+\!2\beta_{1,2}\!-\!2n ) } \nonumber \\ & & \hspace{-5 mm } \times \left ( \frac{m^2}{m^2 } \right)^{\beta_{1,2}-n } ~_{6}f_5\left(\begin{array}{c| } \beta_{1,2}\!-\!\frac{n}{2 } , a_{\alpha_1,\beta_{1,2}}\!-\!n , a_{\alpha_2,\beta_{1,2}}\!-\!n , a_{\alpha_{1,2},\beta_{1,2}}\!-\!\frac{3n}{2 } , \frac{1\!+\!\beta_{1,2}-n}{2 } , \frac{2\!+\!\beta_{1,2}-n}{2 } \\ 1\!+\!\beta_1\!-\!\frac{n}{2 } , 1\!+\!\beta_2\!-\!\frac{n}{2 } , 1\!+\!\beta_{1,2}\!-\!n , \frac{\alpha_{1,2}}{2}\!+\!\beta_{1,2}\!-\!n , \frac{\alpha_{1,2}+1}{2}\!+\!\beta_{1,2}\!-\!n \end{array } ~\frac{m^2}{m^2 } \right ) \nonumber \\ & & \hspace{-6 mm } + \frac { \gamma(\beta_1 ) \gamma \left(\beta_2 \!-\ ! \frac{n}{2 } \right ) \gamma \left(\frac{n}{2 } \!-\ !
\beta_1 \right ) \gamma \left(\alpha_1 \!+\ ! \beta_1 \!-\ !
\frac{n}{2 } \right ) \gamma \left(\alpha_2 \!+\ ! \beta_1 \!-\ !
\frac{n}{2 } \right ) \gamma \left(\alpha_{1,2 } \!+\ ! \beta_1 \!-\ !
n \right ) } { \gamma(\alpha_{1,2 } \!+\ ! 2 \beta_1 \!-\ !
n ) } \nonumber \\ & & \hspace{-5 mm } \left ( \frac{m^2}{m^2 } \right)^{\beta_1-n/2 } { } _ { 6}f_5\left(\begin{array}{c| } \beta_1 , \alpha_1 \!+\ ! \beta_1 \!-\ !
\frac{n}{2 } , \alpha_2 \!+\ ! \beta_1 \!-\ !
\frac{n}{2 } , \alpha_{1,2 } \!+\ ! \beta_1 \!-\ !
n , 1 \!+\ ! \frac{\beta_1 \!-\ !
\beta_2}{2 } , \frac{1 \!+\ ! \beta_1 \!-\ !
\beta_2}{2 } \\ 1 \!+\ ! \beta_1 \!-\
! \beta_2 , 1 \!+\ ! \beta_1 \!-\ !
\tfrac{n}{2 } , 1 \!-\
! \beta_2 \!+\ !
\tfrac{n}{2 } , \tfrac{\alpha_{1,2}}{2 } \!+\ ! \beta_1 \!-\ !
n , \tfrac{\alpha_{1,2}+1}{2 } \!+\ ! \beta_1 \!-\!n \end{array } ~ \frac{m^2}{m^2 } \right ) \nonumber \\ & & \hspace{-5 mm } + \frac { \gamma(\beta_2 ) \gamma \left(\beta_1 \!-\ !
\frac{n}{2 } \right ) \gamma \left(\frac{n}{2 } \!-\ !
\beta_2 \right ) \gamma \left(\alpha_1 \!+\ ! \beta_2 \!-\ !
\frac{n}{2 } \right ) \gamma \left(\alpha_2 \!+\ ! \beta_2 \!-\ !
\frac{n}{2 } \right ) \gamma \left(\alpha_{1,2}\!+\ ! \beta_2 \!-\ !
n \right ) } { \gamma(\alpha_{1,2 } \!+\ ! 2 \beta_2 \!-\ !
n ) } \nonumber \\ & & \hspace{-5 mm } \left ( \frac{m^2}{m^2 } \right)^{\beta_2-n/2 } { } _ { 6}f_5\left(\begin{array}{c| } \beta_2 , \alpha_1 \!+\ ! \beta_2 \!-\ !
\tfrac{n}{2 } , \alpha_2 \!+\ ! \beta_2 \!-\ !
\tfrac{n}{2 } , \alpha_{1,2 } \!+\ ! \beta_2 \!-\ !
n , \frac{1 \!+\ ! \beta_2 \!-\ !
\beta_1}{2 } , \frac{2 \!+\ ! \beta_2 \!-\ !
\beta_1}{2 } \\ 1 \!+\ ! \beta_2 \!-\
! \beta_1 , 1 \!+\ ! \beta_2 \!-\ !
\tfrac{n}{2 } , 1 \!-\
! \beta_1 \!+\ !
\tfrac{n}{2 } , \tfrac{\alpha_{1,2 } \!-\ !
n}{2 } \!+\ !
\beta_2 , \tfrac{\alpha_{1,2 } \!+\ !
n}{2 } \!+\ !
\beta_2 \end{array } ~ \frac{m^2}{m^2 } \right ) \biggr ] \;.
\label{b1122}\end{aligned}\ ] ] let us consider the @xmath277-loop propagator type diagram @xmath320 ( see fig . [ diagrams3 ] )
defined as @xmath321^{\alpha_1 } [ k_2 ^ 2-m_2 ^ 2]^{\alpha_2 } [ k_3 ^ 2]^{\sigma_1 } \cdots [ ( p - k_1-\cdots -k_q)^2]^{\sigma_{q-1 } } } \;. \label{j023}\end{aligned}\ ] ] the mellin - barnes representation of eq.([j023 ] ) is @xmath322^{q } } { \gamma(\alpha_1)\gamma(\alpha_2 ) } \biggl\ { \pi_{i=1}^{q-1 } \frac{\gamma\left(\frac{n}{2}-\sigma_i \right)}{\gamma(\sigma_i ) } \biggr\ } \nonumber \\ & & ( p^2)^{\tfrac{n}{2}q - a_{\alpha_1,\alpha_2,\sigma } } \int \int dt_1 dt_2 \gamma(-t_1 ) \gamma(-t_2 ) \left ( - \frac{m_1 ^ 2}{p^2 } \right)^{t_1 } \left ( - \frac{m_2 ^ 2}{p^2 } \right)^{t_2 } \nonumber \\ & & \hspace{10 mm } \times \frac { \gamma\left ( \frac{n}{2 } \!-\ ! \alpha_1 \!-\ !
t_1 \right ) \gamma\left ( \frac{n}{2 } \!-\
! \alpha_2 \!-\ !
t_2 \right ) \gamma\left ( a_{\alpha_1 , \alpha_2 , \sigma } \!-\ !
t_2 \right ) } { \gamma\left ( \frac{3}{2}n - a_{\alpha_1,\alpha_2,\sigma } \!-\ ! t_1 \!-\ !
t_2 \right ) } \ ; , \label{j023:mellin - barnes}\end{aligned}\ ] ] where @xmath323 closing the contour of integration in eq.([j023:mellin - barnes ] ) on the left , we obtain @xmath322^{q } ( -m_1 ^ 2)^{\tfrac{n}{2}q - a_{\alpha_1,\alpha_2,\sigma } } } { \gamma(\alpha_1)\gamma(\alpha_2 ) \gamma\left(\frac{n}{2}\right ) } \biggl\ { \pi_{i=1}^{q-1 } \frac{\gamma\left(\frac{n}{2}-\sigma_i \right)}{\gamma(\sigma_i ) } \biggr\ } \nonumber \\ & & \biggl [ \gamma\left ( \frac{n}{2 } \!-\ !
\alpha_2 \right ) \gamma\left ( a_{\alpha_1 , \alpha_2 , \sigma } \!-\ ! \frac{n}{2 } q \right ) \gamma\left ( a_{\alpha_2 , \sigma } \!-\
! \frac{n}{2 } ( q-1 )
\right ) \nonumber \\ & & \hspace{5 mm } \times \left .
f_4\left ( a_{\alpha_1 , \alpha_2 , \sigma } - \frac{n}{2 } q , a_{\alpha_2 , \sigma } - \frac{n}{2 } ( q\!-\!1 ) , \frac{n}{2 } , 1 \!+\ ! \alpha_2 \!-\
! \frac{n}{2 }
\right| \frac{p^2}{m_1 ^ 2 } , \frac{m_2 ^ 2}{m_1 ^ 2 } \right ) \nonumber \\ & & + \left ( \frac{m_2 ^ 2}{m_1 ^ 2 } \right)^{\tfrac{n}{2}-\alpha_2 } \gamma\left ( \alpha_2 - \frac{n}{2 } \right ) \gamma\left ( a_{\alpha_1 , \sigma } \!-\
! \frac{n}{2 } ( q-1 )
\right ) \gamma\left(\sigma \!-\ ! \frac{n}{2 } ( q\!-\!2 ) \right ) \nonumber \\ & & \hspace{5 mm } \times \left .
f_4\left ( a_{\alpha_1 , \sigma } - \frac{n}{2 } ( q-1 ) , \sigma - \frac{n}{2 } ( q\!-\!2 ) , \frac{n}{2 } , 1 \!-\ ! \alpha_2 \!+\
! \frac{n}{2 }
\right| \frac{p^2}{m_1 ^ 2 } , \frac{m_2 ^ 2}{m_1 ^ 2 } \right ) \biggr ] \;.
\label{j023:hyper}\end{aligned}\ ] ] let us analyze the result of the differential reduction of hypergeometric functions in eq .
( [ j023:hyper ] ) assuming that all powers of propagators are integer . for @xmath324 ,
both hypergeometric functions are reducible to function and their first derivative only ( in accordance with the results of section [ exception ] ) @xmath325 where @xmath326 is a set of integers , and @xmath327 , @xmath328 are rational functions
. for @xmath329 the result of differential reduction coincides with eq .
( [ reductionf234 ] ) . in accordance with section [ count ] , there are three nontrivial master integrals for @xmath324 and four ones for @xmath329 . for completeness , we also present the explicit results for the diagram @xmath330 ( see fig . [ diagrams3 ] ) ( when external momenta on mass shell of one of the internal masses ) @xmath331^{\alpha } [ k_2 ^ 2-m^2]^{\lambda } [ k_3 ^ 2]^{\sigma_1 } \cdots [ ( p - k_1-\cdots -k_q)^2]^{\sigma_{q-1 } } } \right|_{p^2=m^2 } \;. \label{j012}\end{aligned}\ ] ] the hypergeometric representation for this diagram is @xmath332^{q } ( -m^2)^{\tfrac{n}{2}q - a_{\alpha,\lambda,\sigma } } } { \gamma(\alpha)\gamma(\lambda ) \gamma\left(\frac{n}{2}\right ) } \biggl\ { \pi_{i=1}^{q-1 } \frac{\gamma\left(\frac{n}{2}-\sigma_i \right)}{\gamma(\sigma_i ) } \biggr\ } \nonumber \\ & & \biggl [ \gamma\left ( \frac{n}{2 } \!-\ ! \alpha \right ) \gamma\left ( a_{\alpha , \lambda , \sigma } \!-\ !
\frac{n}{2 } q \right ) \gamma\left ( a_{\alpha , \sigma } \!-\ ! \frac{n}{2 } ( q-1 )
\right ) \nonumber \\ & & \hspace{5 mm } \times ~_{4}f_3\left(\begin{array}{c| } a_{\alpha , \lambda , \sigma } - \frac{n}{2 } q , a_{\alpha , \sigma } - \frac{n}{2 } ( q\!-\!1 ) , \tfrac{\alpha}{2 } , \tfrac{\alpha+1}{2 }
\\ 1 + \alpha - \tfrac{n}{2 } , \alpha , \frac{n}{2 } , \end{array } ~\frac{4m^2}{m^2 } \right ) \nonumber \\ & & + \left ( \frac{m^2}{m^2 } \right)^{\tfrac{n}{2}-\alpha } \gamma\left ( \alpha - \frac{n}{2 } \right ) \gamma\left ( a_{\lambda , \sigma } \!-\ ! \frac{n}{2 } ( q-1 )
\right ) \gamma\left(\sigma \!-\ ! \frac{n}{2 } ( q\!-\!2 ) \right ) \nonumber \\ & & \hspace{5 mm } \times ~_{4}f_3\left(\begin{array}{c| } a_{\lambda , \sigma } - \frac{n}{2 } ( q-1 ) , \sigma - \frac{n}{2 } ( q\!-\!2 ) , \frac{n-\alpha}{2 } , \frac{n-\alpha+1}{2 } \\
\frac{n}{2 } , n - \alpha , 1 \!-\ ! \alpha \!+\ !
\frac{n}{2 } \end{array } ~\frac{4m^2}{m^2 } \right ) \biggr ] \ ; , \label{j012:hyper}\end{aligned}\ ] ] where @xmath333 in the case when the powers of propagators are all integer , the result of the differential reduction of the hypergeometric functions in eq .
( [ j012:hyper ] ) are the following ( see also discussion in @xcite ) : * for @xmath324 @xmath334 in accordance with eq.([count ] ) , there are two nontrivial master - integrals only . * for @xmath329 @xmath335 in accordance with eq.([count ] ) , there are three nontrivial master - integrals only .
the differential reduction algorithm @xcite allows one to compare horn - type hypergeometric functions with parameters whose values differ by integer . in this paper , we have presented the mathematica based package * hyperdire * for differential reduction of the generalized hypergeometric functions @xmath336 and appell functions @xmath124 and @xmath125 to a set of basis functions , defined by eqs .
( [ decomposition ] ) , ( [ reductionf1 ] ) , ( [ reductionf234 ] ) , respectively .
these functions are closely related with a large class of feynman diagrams @xcite .
in contrast to existing packages , our package performs reduction of hypergeometric functions before @xmath230-expansion and works with arbitrary values of parameters
. the package could be easy extended to other horn - type hypergeometric functions by adding new modules . as an illustration of our approach
, we considered a few examples , namely the feynman diagrams shown in figs . [ diagrams]-[diagrams3 ] with arbitrary powers of propagators and space - time dimension .
* acknowledgments .
* + we are grateful to a. davydychev , a. grozin , a. isaev , a. kotikov , o. krotikova , g. somogyi , v. spiridonov , o. tarasov , o. veretin , b.f.l ward , s. yost for useful discussions , and g. sandukovskaya for careful reading of the manuscript . m.y.k.s research was supported in part by bmbf , grant no .
05h09gue , dfg , grant no .
kn 365/32 , and hgf , grant no .
ha 101 .
bailey , _ generalized hypergeometric series _ , cambridge university press , 1935 ; + a. erdelyi ( ed . ) , _ higher transcendental functions _ , vol.1 , mcgraw - hill , new york , 1953 ; + l.j .
slater , _ generalized hypergeometric functions _
, cambridge university press , cambridge 1966 ; + h. exton , _ multiple hypergeometric functions and applications _ john wiley & sons , inc .
, new york - london - sydney , 1976 .
srivastava , p.w .
karlsson , _ multiple gaussian hypergeometric series _ ,
+ john wiley & sons , inc . , new york , 1985 . c.g .
bollini , j.j .
giambiagi , phys .
b * 40 * ( 1972 ) 566 ; r. gastmans , r. meuldermans , nucl .
b * 63 * ( 1973 ) 277 ; d.s .
kershaw , phys .
d * 8 * ( 1973 ) 2708 ; a.c.t .
wu , phys .
d * 9 * ( 1974 ) 370 ; v.i .
kucheryavyi , teor . mat .
* 20 * ( 1974 ) 29 ; k. mano , phys
. rev .
d * 11 * ( 1975 ) 452 ; r.j .
gonsalves , phys .
d * 28 * ( 1983 ) 1542 ; g. kramer , b. lampe , j. math .
* 28 * ( 1987 ) 945 ; e.e .
boos , a.i .
davydychev , theor . math
. phys . 89 ( 1991 ) 1052 ; a.i .
davydychev , j. math .
* 32 * ( 1991 ) 1052 ; a.i .
davydychev , j. math . phys .
* 33 * ( 1992 ) 358 ; d.j .
broadhurst , z. phys .
c * 54 * , 599 ( 1992 ) ; d.j .
broadhurst , j. fleischer , o.v .
tarasov , z. phys .
c * 60 * ( 1993 ) 287 ; a.i .
davydychev , j.b .
tausk , nucl .
b * 397 * ( 1993 ) 123 ; a.i .
davydychev , j.b .
tausk , phys .
d * 53 * ( 1996 ) 7381 ; f.a .
berends et al .
, z. phys .
c * 63 * ( 1994 ) 227 ; s. bauberger et al .
b * 434 * ( 1995 ) 383 ; f.a .
berends et all .
b * 439 * ( 1995 ) 536 ; a.v .
kotikov , phys .
b * 375 * ( 1996 ) 240 ; j. fleischer et al .
, nucl . phys .
b * 539 * ( 1999 ) 671 [ erratum - ibid .
b * 571 * ( 2000 ) 511 ] ; a.i .
davydychev , a.g .
grozin , phys .
d * 59 * ( 1999 ) 054023 ; t. gehrmann , e. remiddi , nucl .
b * 580 * ( 2000 ) 485 ; a.i .
davydychev , phys .
d * 61 * ( 2000 ) 087701 ; c. anastasiou , e.w.n .
glover , c. oleari , nucl .
b * 575 * ( 2000 ) 416 [ erratum - ibid .
b * 585 * ( 2000 ) 763 ] ; a.i .
davydychev , m.yu .
kalmykov , nucl .
proc . suppl .
* 89 * ( 2000 ) 283 ; o.v .
tarasov , nucl .
* 89 * ( 2000 ) 237 ; l.g .
cabral - rosetti , m.a .
sanchis - lozano , j. comput .
* 115 * ( 2000 ) 93 [ arxiv : hep - ph/9809213 ] ; a.i .
davydychev , m.yu .
kalmykov , nucl .
b * 605 * ( 2001 ) 266 ; a.t .
suzuki , e.s .
santos , a.g.m .
schmidt , j. phys .
a * 36 * ( 2003 ) 4465 ; j. fleischer , f. jegerlehner , o.v .
tarasov , nucl .
b * 672 * ( 2003 ) 303 ; f. jegerlehner , m.yu .
kalmykov , o. veretin , nucl .
b * 658 * ( 2003 ) 49 ; f. jegerlehner , m.yu .
kalmykov , nucl . phys .
b * 676 * ( 2004 ) 365 ; t. gehrmann , t. huber , d. matre , phys .
b * 622 * ( 2005 ) 295 ; a.i .
davydychev , nucl .
instrum .
a * 559 * ( 2006 ) 293 ; e. bejdakic , y. schrder , nucl .
b ( proc . suppl . )
* 160 * ( 2006 ) 155 ; f. jegerlehner , o.v .
tarasov , phys .
b * 639 * ( 2006 ) 299 ; o.v .
tarasov , phys .
b * 638 * ( 2006 ) 195 ; f. jegerlehner , o.v .
tarasov , nucl .
* 157 * ( 2006 ) 125 ; l.g .
cabral - rosetti , m.a .
sanchis - lozano , j. phys .
* 37 * ( 2006 ) 82 [ arxiv : hep - ph/0206081 ] ; a.g .
grozin , t. huber , d. matre ; jhep * 0707 * ( 2007 ) 033 ; j. gluza , k. kajda , t. riemann , comput .
commun . 177 ( 2007 ) 879 ; m.a .
shpot , j. math .
* 48 * ( 2007 ) 123512 ; o.v .
tarasov , phys .
b * 670 * ( 2008 ) 67 ; t. huber , nucl .
b ( proc . suppl . ) 183 ( 2008 ) 238 ; v. del duca , c. duhr , e.w .
nigel glover , v.a .
smirnov , jhep * 1001 * , 042 ( 2010 ) ; b.a .
kniehl , o.v .
tarasov , nucl .
b * 820 * ( 2009 ) 178 ; b.a .
kniehl , o.v .
tarasov , nucl .
b * 833 * ( 2010 ) 298 ; l.v .
bork , d.i .
kazakov , g.s .
vartanov , arxiv:1011.2440 ; g. somogyi , arxiv:1101.3557 ; + t. kaneko , arxiv : 1105.2080 .
rainville , bull .
* 51 * ( 1945 ) 714 ; + j. wimp , math .
* 22 * ( 1968 ) 363 ; + w.miller,jr . , siam j. math .
, * 3 * ( 1972 ) 31 ; _ ibid . _
* 4 * ( 1973 ) 638 ; _ ibid . _
* 5 * ( 1974 ) 309 ; j. math .
* 13 * ( 1972 ) 1393 ; siam j. appl .
, * 25 * ( 1973 ) 226 ; + takeshi sasaki , siam j. math
. anal . *
22 * ( 1991 ) 821 ; + e. horikawa , j. math .
tokyo * 1 * ( 1994 ) 181 ; + a. adolphson , b. dwork , trans .
* 347 * ( 1995 ) 615 ; + k. roach , proc . of issac96 , 301 - 308 , acm , new york ; + k. roach , proc .
of issac99 , 205 - 211 , acm , new york ; + r.g .
buschman , j. comput .
math . * 107 * ( 1999 ) 127 .
b. sturmfels , n. takayama , _ grbner bases and hypergeometric functions _ , p. 246258
, london math .
lecture note ser .
251 , cambridge univ . press , cambridge , 1998 ; + m. saito , b. sturmfels , n. takayama , compositio math .
* 115 * ( 1999 ) 185204 ; + m. saito , b. sturmfels , n. takayama , _ grbner deformations of hypergeometric differential equations _ , ( springer - verlag , berlin , 2000 ) . m. bronstein , m. petkovek , theoret . comput .
* 157 * ( 1996 ) 333 ; + f. chyzak , b. salvy , j. symbolic comput * 26 * ( 1998 ) 187 - 227 ; + f. chyzak , _ grbner bases , symbolic summation and symbolic integration _ ,
p.3260 , london math .
lecture note ser .
251 , cambridge univ .
press , cambridge , 1998 .
kalmykov , jhep * 0604 * ( 2006 ) 056 ; + m.yu .
kalmykov et al . , proc . of acat-08 [ arxiv:0901.4716 ] ; + v.v .
bytev et al . , [ arxiv:0902.1352 ] .
bytev , m.yu .
kalmykov , b.a .
kniehl , nucl .
b * 836 * ( 2010 ) 129 .
kalmykov , b.f.l .
ward , s. yost , jhep * 0702 * ( 2007 ) 040 ; jhep * 0711 * ( 2007 ) 009 ; + s.a .
yost , m.yu .
kalmykov , b.f.l .
ward , proc . of ichep 2008 [ arxiv:0808.2605 ]
; + m.yu .
kalmykov et al .
, proc . of quarks-08 , [ arxiv:0810.3238 ] ; + m.yu .
kalmykov , b. kniehl , nucl . phys .
b * 809 * ( 2009 ) 365 ; phys . part .
nucl . * 41 * ( 2010 ) 942 .
s. weinzierl , comput .
commun .
* 145 * ( 2002 ) 357 ; + s. moch , p. uwer , comput .
commun .
* 174 * ( 2006 ) 759 .
t. huber , d. matre , comput .
commun .
* 175 * ( 2006 ) 122 ; ibid * 178 * ( 2008 ) 755 . j. blumlein , s. klein , c. schneider f. stan , arxiv:1011.2656 . s. moch , p. uwer , s. weinzierl , j. math . phys .
* 43 * ( 2002 ) 3363 ; + s. weinzierl , j. math .
* 45 * ( 2004 ) 2656 .
davydychev , m.yu .
kalmykov , nucl .
b * 699 * ( 2004 ) 3 ; + m.yu .
kalmykov , nucl .
* 135 * ( 2004 ) 280 ; + m.yu .
kalmykov , b.f.l .
ward , s. yost , jhep * 0710 * ( 2007 ) 048 .
tkachov , phys .
b 100 ( 1981 ) 65 ; + k.g .
chetyrkin , f.v .
tkachov , nucl .
b 192 ( 1981 ) 159 .
s. laporta , int .
j. mod .
phys . a * 15
* ( 2000 ) 5087 ; + c. anastasiou , a. lazopoulos , jhep * 0407 * ( 2004 ) 046 ; + a.v .
smirnov , jhep * 0810 * ( 2008 ) 107 ; + c. studerus , comput .
commun .
* 181 * ( 2010 ) 1293 .
yost et al .
, pos * ichep2010 * ( 2010 ) 135 [ arxiv:1101.2348 ] .
tarasov , nucl .
b * 502 * ( 1997 ) 455 .
vladimirov , theor .
43 ( 1980 ) 417 [ teor
. mat .
fiz . 43 ( 1980 )
g. thooft and m. veltman , nucl .
b * 44 * ( 1972 ) 189 ; + c.g .
bollini and j.j .
giambiagi , nuovo cim .
b * 12 * ( 1972 ) 20 ; + j.f .
ashmore , lett .
nuovo cim .
* 4 * ( 1972 ) 289 ; + g.m .
cicuta and e. montaldi , lett .
nuovo cim .
* 4 * ( 1972 ) 329 .
smirnov , commun .
* 134 * ( 1990 ) 109 ; + f.v .
tkachov , int .
phys . a * 8 * ( 1993 ) 2047 ; + g.b .
pivovarov , f.v .
tkachov , int .
phys . a * 8 * ( 1993 ) 2241 .
j. fleischer m.yu .
kalmykov , comput .
commun .
* 128 * ( 2000 ) 531 ; + f. jegerlehner , m.yu .
kalmykov , o. veretin , nucl .
b * 641 * ( 2002 ) 285 . | the hyperdire is a project devoted to creation of a set of mathematica based programs for differential reduction of hypergeometric functions .
the current version includes two parts : one , * pfq * , is relevant to manipulation with hypergeometric functions @xmath0 , and second , * appellf1f4 * , to manipulations with appell s hypergeometric functions of two variables @xmath1 . as application ,
a few multiloop feynman diagrams of propagator type are considered . * hyperdire + hypergeometric functions differential reduction + mathematica based packages for differential reduction of generalized hypergeometric functions : + now with @xmath2 * vladimir v. bytev@xmath3 , , + + luruper chaussee 149 , 22761 hamburg , germany + @xmath4 joint institute for nuclear research , + @xmath5 dubna ( moscow region ) , russia * program summary * _ title of program _ : hyperdire + _ version _ : 1.0.0 _ release _ : 1.0.0 _ catalogue number _ : + _ program obtained from https://sites.google.com/site/loopcalculations/home_ : + _ e - mail : [email protected]_ + _ licensing terms _ : gnu general public licence + _ computers _ : all computers running mathematica + _ operating systems _ : operation systems running mathematica + _ programming language _ : mathematica + _ keywords _ : generalized hypergeometric functions , appell functions , feynman integrals .
+ _ nature of the problem _ : reduction of hypergeometric functions : @xmath6 to set of basis functions .
+ _ method of solution _ : differential reduction reduction + _ restriction on the complexity of the problem _ : none + _ typical running time _ : depending on the complexity of problem .
* long write - up * |
nucleons are the most common form of hadronic matter on the earth and probably in the whole universe .
understanding their internal structure will give us insight into how the real world works .
an important source of information about the nucleon internal structure is the nucleon excitation spectrum .
our present knowledge on this aspect came almost entirely from partial - wave analyses of @xmath15 total , elastic , and charge - exchange scattering data of more than twenty years ago@xcite . since the late 1970 s , very little has happened in experimental @xmath16 baryon spectroscopy .
considering its importance for the understanding of the baryon structure and for distinguishing various pictures @xcite of the nonperturbative regime of qcd , a new generation of experiments on n * physics with electromagnetic probes have recently been started at new facilities , such as cebaf at jlab , elsa at bonn , graal at grenoble and so on .
one of us , zou suggested @xcite that we can also study @xmath16 baryon in j@xmath17 decay to baryon - antibaryon final states , which provide a new laboratory for study of @xmath16 baryon , especially in the mass range of 1 - 2 gev .
for example , the @xmath18 is an excellent channel to study the @xmath19 state which has a very large decay branching ratio to the @xmath20 @xcite , while other baryon resonances below 2.0 gev do not have a large branching ratio to decay in this channel@xcite , a fact noted very early in the development of the quark shell model @xcite . in this paper ,
based on 7.8 million @xmath0 events collected at bepc , the events for @xmath1 and @xmath2 have been selected and reconstructed .
we perform a partial wave analysis(pwa ) on @xmath21 data in the full mass region of @xmath22(@xmath23 ) .
this is the first pwa study of @xmath16 baryon in @xmath0 hadronic decay in the world .
two s - wave @xmath16 baryon , namely @xmath19 and @xmath24 , are found in their @xmath25 decay mode .
the new information on @xmath26 couplings provides a new source for studying baryon structure .
the @xmath4 and @xmath27 are detected in their @xmath28 decay modes .
each candidate event is required to have two oppositely signed charged tracks with a good helix fit in the polar angle range @xmath29 in mdc and at least 2 reconstructed @xmath30 s in bsc .
a vertex is required within an interaction region @xmath31 cm longitudinally and 2 cm radially . a minimum energy cut of 60 mev is imposed on the photons .
showers associated with charged tracks are also removed .
after previous selection , we use tof information to identify the @xmath32 pairs , and at least one track with unambiguous tof information is required .
the open angle of two charged tracks smaller than @xmath33 is required in order to remove back to back events ; to remove radiative bhabha events , we require @xmath34 , where @xmath35 , @xmath36 ( @xmath37 , @xmath38 ) are the energy deposited in bsc and momentum of positron ( electron ) respectively .
events are fitted kinematically to the 4c hypotheses @xmath39 .
figure [ fig : gg ] shows the invariance mass spactrum of the @xmath40 , we can see the clear @xmath27 and @xmath4 signals .
meanwhile , the events are also fitted to @xmath41 and @xmath42 .
we require @xmath43 to reject the @xmath44 and @xmath45 backgrounds . in order to improve the mass resolution ,
5c fits are performed on the selected events , the extra constraints are those of the @xmath4 and @xmath27 masses for @xmath6 and @xmath46 decays , respectively .
@xmath47 ( @xmath48 ) is required .
figure [ fig : ppi0 ] and figure [ fig : peta ] show the @xmath49 and @xmath22 mass distributions from the decays @xmath1 , @xmath2 respectively .
clear peaks are observed around 1480 mev in @xmath49 invariant mass spectrum .
the @xmath22 events peak strongly in the neighborhood of @xmath4-production threshold , and we shall show that the data require a strong @xmath50 peak near the threshold . there is an additional obvious bump around 1600 - 1700mev , it favors @xmath7 in our @xmath51 ( @xmath52 ) partial wave analysis .
a pwa analysis is performed for the @xmath6 channel@xcite with the amplitudes constructed from lorentz - invariant combinations of the momenta and the photon polarization 4-vectors for @xmath0 initial states with helicity @xmath53 .
the relative magnitudes and phases of the amplitudes are determined by a maximum likelihood fit to the data .
based on the study of @xmath32 and @xmath22 invariant mass distributions in our data , the decay chain @xmath6 is analyzed taking into account two @xmath22(@xmath54 ) s - waves(@xmath51 ) and one @xmath22(@xmath54 ) p - wave(@xmath55 ) intermediate processes .
the two s - waves amplitudes are only used to fit the data in the low mass region near the @xmath4 production threshold . while the p - wave is used to fit the data in the high mass region .
the background from multi-@xmath27 is @xmath56 in the 5c fit , we have included a phase space background in the pwa fit to allow for this ; the @xmath22 mass projection fitted to the real data is shown in figure [ fig : peta ] .
we now discuss the features of the data and the outcome of fits .
a peak at @xmath57 1535 mev near the @xmath4 threshold optimises @xmath58 mev as shown in figure [ fig : fit](a ) , the data favour @xmath59 over @xmath60 .
a fit with @xmath61 instead gives @xmath62 worse by 16.0 than for @xmath63 assignment(our definition of @xmath64 is such that it increases by 0.5 for a one standard deviation change in one parameter ) . with our 4 fitted parameters , the statistical significance of the peak is above 6.0@xmath65 .
for the width scan as shown in figure [ fig : fit](b ) , our data require a width , @xmath66 mev .
our results for @xmath19 are consistent with the resonance parameters measured by krusche et al.@xcite at the mami acceletator in mainz on the @xmath4 photoproduction . at @xmath67 1650 mev ,
there is a further peak .
we fit it with a @xmath68 resonance .
figure [ fig : fit](c ) are the scan of mass .
its mass optimise at @xmath69 mev with @xmath70 mev fixed to pdg value .
we have tried fits to this peak with resonances having quantum numbers @xmath71 .
we find that log likelihood is better for @xmath63 than @xmath71 by 9.0 . with our 4 fitted parameters ,
the statistical significance of the peak is @xmath72 .
our results for @xmath24 are consistent with the parameters proposed by pdg .
a small improvement to the fit is given by including a @xmath73 resonance , which optimises at @xmath74 mev and @xmath75 mev fixed .
the statistical significance of the peak is only @xmath76 .
we have tried @xmath68 instead @xmath77 , but the fit is much worse .
in summary , we have studied the @xmath21 decay channel , and a pwa analysis is performed on the data .
there is a definite requirement for a @xmath78 component at @xmath79 mev with @xmath80 mev near the threshold for @xmath4 production .
in addition , there is an obvious @xmath7 resonance , with @xmath81 mev and @xmath82 mev fixed to pdg data . in the higher @xmath22 mass region , there is an evidence of @xmath83 signal around 1800 mev , we can not get any conclusion for this state due to the low statistics .
all above analysis is the first step for us to probe @xmath16 baryons at bes .
we will perform detail studies of @xmath16 baryons in the following @xmath0 decay channels : @xmath84 , @xmath85 , @xmath86 , @xmath87 , @xmath88 and so on .
one of the authors , h.b .
li , is grateful to prof .
leonard kisslinger for the helpful discussions .
this work is supported in part by chinese national science foundation under contract no .
19290401 and 19605007 . | based on 7.8 million @xmath0 events collected at bepc , the events for @xmath1 and @xmath2 have been selected and reconstructed .
clear peaks are observed around 1480 mev in @xmath3 invariant mass spectrum and near the threshold for @xmath4 production in @xmath5 invariant mass spectrum .
a partial wave analysis of @xmath6 data has been performed .
two @xmath7 resonances are observed with mass and width ( m , @xmath8 ) at ( @xmath9 , @xmath10 ) mev and ( @xmath11 , @xmath12)mev , and are considered to be the nucleon resonances @xmath13 and @xmath14 respectively . |
due to the accelerated expansion of the universe , scalar field dark matter models emerged to account for the astronomic observations ( see @xcite for a review ) .
the models describe dark matter as a scalar .
these models are special issues of more general theories known as scalar - tensor theories .
in contrast with pure gravity , four - dimensional scalar - tensor gravity admits a variety of _ static _ , spherically symmetric , neutral or charged configurations including schwarzschild black holes ( bh ) , phantom bhs , naked singularities ( ns ) , particle - like ( p - l ) solutions and wormholes @xcite .
we consider the four - dimensional action of general relativity coupled to a self - gravitating real scalar field with positive kinetic energy ( einstein - scalar gravity ) @xmath0\,,\ ] ] where @xmath1 is the scalar curvature and @xmath2 is a potential for @xmath3 .
the field equations derived from are @xcite @xmath4 ( where @xmath5 ) , with @xmath6 being the conserved energy - momentum tensor of @xmath3 @xmath7 for static , spherically symmetric configurations , the metric can be put in the form @xmath8 we fix the behavior of @xmath9 at spatial infinity such that @xmath10 as @xmath11 ( this is no loss of generality ) . with the ansatz ,
the first equation in reduces to eq and the second group of equations in lead to eqs to @xcite @xmath12 only three equations of this set are independent ; when the last three equations , to , are satisfied the first one is automatically satisfied .
theorem 1 of ref .
@xcite and its generalization to @xmath13-dimension @xcite rules out wormhole , horn and flux tube type solutions to eqs to . a consequence of that ,
the range of @xmath14 is from @xmath15 at the center to @xmath16 ; correspondingly , the range of @xmath17 is from @xmath18 at the center of configuration ( @xmath19 ) to @xmath16 .
since @xmath20 ( real scalar field ) , leads to @xmath21 . hence , the function @xmath9 is concave down on its range .
[ fig1 ] shows a generic typical behavior of @xmath9 where @xmath22 is the location of the center . in this paper @xmath22
is treated as a(n arbitrary ) scaling constant ( which can be set to 1 ) to which we compare the masses of the solutions and the whole theory is scale invariant is the same as the rescaling @xmath23 , @xmath24 , @xmath25 , @xmath26 , @xmath27 . ] .
@xmath22 is also a measure of the strength of the field .
we also make it such that the value of @xmath17 at @xmath15 is precisely @xmath22 ; this is no loss of generality as justified in appendix a. for the case of four - dimensional field equations to , bronnikov @xcite established some restrictive results , which have been latter generalized to @xmath28-dimensional field equations @xcite .
we are concerned with the following conclusions @xcite . *
all horizons are simple . *
all possible global structures for @xmath29 and for @xmath30 are the same ; * possible non - singular solutions have either minkowski / ads or de sitter structures .
p - l asymptotically flat solutions with mass @xmath31 and @xmath32 having both signs are not excluded ; however , if @xmath33 no such p - l solutions ( @xmath31 ) exist .
* for all functions @xmath34 , eqs to do not admit static , spherically symmetric regular bh and wormhole solutions with any asymptotic .
* for @xmath33 , eqs to do not admit any hairy bh solutions ( no scalar hair theorems @xcite ) . an asymptotically flat bh solution to eqs to is characterized by the presence of event horizons . since all horizons are simple , there is a change in the sign of @xmath35 at each event horizon .
thus for a bh , @xmath35 has both signs in the domain @xmath36 . for an asymptotically flat p - l or ns solution
there are no event horizons , thus @xmath35 has a constant sign for @xmath37 and because @xmath38 at spatial infinity so it is for @xmath37 . only for p - l solutions ,
the functions @xmath35 , @xmath34 & @xmath39 are finite everywhere for @xmath37 .
some static , spherically symmetric , asymptotically flat solutions to eqs to have been derived in @xcite , some of which represent different configurations of _ singular _ bhs . in ref @xcite , bronnikov and shikin constructed bh , ns and p - l solutions .
for the latter solution , their function @xmath40 ( eq ( b.3 ) of ref @xcite ) does not approach 1 at spatial infinity ( @xmath41 ) ; rather , it approaches the limit @xmath42 , where @xmath22 is their parametric constant .
thus , their
p - l " solution is asymptotically flat only in the limit @xmath43 , corresponding to the trivial case @xmath44 with zero mass and constant functions @xmath45 : this is minkowski space time .
recently , bronnikov and chernakova @xcite constructed a two - parameter asymptotically flat bh characterized by its negative potential @xmath32 and where @xmath46 diverge near the center and converge as @xmath47 , while @xmath40 is finite everywhere .
very recently , nikonov _ et al _ @xcite have obtained a two - parameter bhs and nss as well as a one - parameter p - l solutions . for the latter , all the three functions @xmath48
are finite everywhere , @xmath32 has both signs and @xmath3 is an incomplete elliptic integral of the first kind .
most of the solutions constructed previously @xcite were derived using the inverse problem method .
the method consists in choosing a suitable function @xmath9 then solving eqs to for @xmath49 ( eq is automatically satisfied ) .
there are physically many interesting exact solutions to eqs to depending on the particular form of @xmath9 .
the behavior of the solutions at spatial infinity is somehow defined by the boundary condition requiring that the solutions be asymptotically flat @xmath50 where @xmath51 is the schwarzschild mass . at the center
there is no specific boundary condition , this is why there is a _ variety _ of possible asymptotically flat solutions , however , grouped into three categories : different ( singular ) bh solutions , ns and p - l solutions . to be more specific , the bh set ( @xmath40 having both signs ) include solutions where @xmath32 is finite everywhere ( this paper ) as well as solutions with divergent @xmath32 @xcite ; the same applies to the ns set where @xmath52 ( this paper ) .
the solutions discussed in @xcite and those discussed in this paper have different boundary conditions at the center .
it is an interesting issue to investigate the behavior of the generic solutions near the center to determine the criteria by which the solutions emerge .
the solutions discussed in @xcite have been derived by choosing a special function has been chosen such that @xmath53 , where @xmath54 is their parametric constant . ]
@xmath9 , however , they lacked a systematic treatment and many questions remained unanswered .
for instance , we do not know a ) whether it is possible to obtain a bh ( resp .
ns ) where both @xmath32 & @xmath3 are finite everywhere , b ) what conditions are required for obtaining , say , a bh or a p - l solution , c ) under which conditions the potential @xmath32 has both signs , etc . in section [ sec2 ]
, we derive new integral formulas for @xmath49 and proof five corollaries and a theorem on selection criteria for obtaining predefined static , spherically symmetric , asymptotically flat solutions to eqs to
. first we deal with the most generic case then focus on a constrained case with minimum conditions . in section [ sec3 ] ,
the behaviors of @xmath49 near the center and at spatial infinity are discussed , instances of solutions of particular interest are provided and two energy conditions are analyzed leading to two further corollaries . in section [ sec4 ] , we derive further solutions by the inverse problem method .
two appendices have been added for completeness .
eq is readily brought to the form @xmath55 ' = 2 $ ] which is integrated by @xmath56 where @xmath57 is constant of integration . integrating leads to @xmath58 where @xmath59 is a second constant of integration . with the boundary conditions
, we must have @xmath60 and @xmath61 .
hence , @xmath62 using @xmath55 ' = 2 $ ] into , we bring @xmath32 to the form @xmath63 substituting into the one or the other expression of @xmath32 , one obtains @xmath64 finally , without loss of generality we take the negative sign in ( @xmath65 ) and introduce two arbitrary constants @xmath66 & @xmath67 , we obtain @xmath68 throughout the remaining part of the paper , when the limits of @xmath49 at the center @xmath22 exist , we define @xmath69 by @xmath70 @xmath71 & @xmath72 , respectively .
this definition extends to their derivatives at the center .
notice that eqs , & have been derived without referring to any extra condition ( except the boundary conditions and , in order to integrate , we have assumed implicitly that @xmath73 exists ( @xmath9 differentiable ) ) , thus the following first two corollaries are generic .
consider the case where @xmath74 .
since @xmath75 we have @xmath76 , thus the integral in is positive leading to @xmath52 .
hence , if the mass of the spherically symmetric , asymptotically flat solution to eqs to is such that @xmath74 where @xmath22 is the scaling constant , then the configuration is either a ns or a p - l solution . a negative - masse , spherically symmetric , asymptotically flat solution is either a ns or a p - l solution .
now , let us look for the conditions that make ( the limit of ) @xmath40 finite at @xmath22 .
writing @xmath77 in , where @xmath78 is any fixed number in the open domain ( @xmath79 ) .
the improper integral @xmath80 converges by the limit comparison test .
in fact , let @xmath81 & @xmath82 , since @xmath83 and since @xmath84 converges for @xmath85 , then the integral of @xmath86 is convergent . with that said ,
the value of @xmath40 at @xmath22 is the following limit @xmath87 the conditions @xmath21 ( eq ) and @xmath88 ( eq ) imply @xmath89 for @xmath37 .
now , with @xmath19 , the limit in exists only is finite with @xmath90 arbitrary is @xmath91 .
however , for this function @xmath92 & @xmath93 are divergent at @xmath22 , consequently @xmath32 too diverges at @xmath22 , thus there is no p - l solution . ]
if @xmath94 . in this case @xmath95
if @xmath96 , @xmath97 and the solution is a ns by corollary 1 . if @xmath98 , @xmath99 and the solution is a bh since @xmath35 has both signs . in this case
( @xmath98 ) , @xmath35 is manifestly positive for @xmath100 .
for @xmath101 , we can write the integral in as @xmath102 where @xmath103 and @xmath104 . as @xmath17 decreases from @xmath105 to @xmath22 , the second integral @xmath106 decreases monotonically from @xmath107 to @xmath108 ( since we found @xmath99 ) . consequently , the sum of the two integrals and @xmath35 vanish at one and only one point @xmath109 ( event horizon ) between @xmath22 and @xmath105 : @xmath110 .
thus for @xmath98 , @xmath109 is a simple root of @xmath111 ( all horizons are simple ) .
if @xmath9 is differentiable at @xmath22 , a necessary condition to find a spherically symmetric , asymptotically flat p - l solution to eqs to is @xmath112 and in this case @xmath113 . in the other cases ,
a ) @xmath96 including negative masses , the solution is a spherically symmetric , asymptotically flat ns with @xmath97 or b ) @xmath98 , the solution is a spherically symmetric , asymptotically flat bh with @xmath99 and the event horizon @xmath114 .
the condition @xmath112 in the above statement for obtaining p - l solutions being necessary , one may derive solutions with @xmath112 which are not p - l solutions . by corollary 1 , these solutions must be nss .
thus , corollary 2 does not rule out nss with @xmath112 .
solutions with @xmath112 and @xmath115 divergent are indeed nss .
we shall see in the next subsection that , for the conditions on which corollary 2 rests , @xmath35 is finite for @xmath116 . since the condition @xmath112 ensures finiteness of @xmath40 at @xmath22 , thus by corollary 2 we have solved the eigenvalue problem consisting in determining an everywhere finite solution @xmath35 to eq with the eigenvalue " @xmath117 . repeating the steps
& for @xmath118 , we obtain @xmath119 = 2\lim_{\rho\to c^+}\frac{\int_{u}^\rho\frac{3m - z}{r^4(z)}\,\text{d}z}{1/r^3(\rho)}= -\frac{2}{3r'(c)}\,\lim_{\rho\to c^+}(3m-\rho)=\frac{2(c-3m)}{3r'(c)}\,,\ ] ] hence for @xmath120 , @xmath40 behaves near the center as @xmath121 in addition to the boundary conditions and differentiability condition of @xmath9 , we impose further ( minimum set of ) conditions on @xmath9 and its first derivatives in order to derive further selection criteria .
we will do that without losing much generality .
thus , throughout the rest of this section , we assume that @xmath9 is four times differentiable in the domain @xmath36 , which means that there @xmath9 and the derivatives @xmath73 to @xmath122 are smooth functions ( continuous ) and finite @xmath123 notice that @xmath124 ; particularly , we have @xmath125 . in part 1 of appendix
a , however , we will see that the third and fourth derivatives of @xmath9 need not exist on the whole domain @xmath126 ; rather , on a small half - open interval containing @xmath22 .
the conditions & do not ensure existence of taylor series expansions or others for @xmath9 at the center @xmath22 since the fifth and higher order derivatives of @xmath9 are not assumed to exist there and elsewhere .
the same is true for @xmath127 , which are related to @xmath9 via eqs , & .
for this reason the method we are considering in this subsection ( as was the case in the previous one ) and appendix a is more general , different from and even advantageous than series expansion methods .
we start by proving the finiteness of @xmath128 at spatial infinity ( that of @xmath35 is imposed in ) .
using in & one obtains @xmath129 & @xmath130 .
the result for @xmath32 is obvious and to derive that for @xmath3 we take the limit as @xmath47 of both sides of @xmath131 we can always bring @xmath9 to the form @xmath132 such that @xmath133 , @xmath134 & @xmath135 .
more precisely , as @xmath47 , @xmath136 where @xmath137 is some positive number behaves as @xmath138 ( @xmath139 ) .
thus as @xmath47 , @xmath140 ( @xmath141 ) . ] . hence , as @xmath47 , @xmath142 and the integral in the r.h.s of converges to some constant . for @xmath143 ( @xmath144 ) ,
the improper integral in converges by the limit comparison test , which means that @xmath40 is finite and continuous for @xmath143 .
@xmath35 , @xmath73 & @xmath145 being finite functions in the domain @xmath143 , eq implies that @xmath32 remains finite for @xmath143 . finally , since the function @xmath146 is continuous on any closed interval included in the open domain @xmath147 , the integral in exists for @xmath66 , which means that @xmath3 is finite and continuous for @xmath143 .
we have thus shown that the three functions @xmath49 are finite and continuous for @xmath143 ; the unique point where the functions may diverge is the center @xmath22 .
notice that the continuity and regularity of the functions @xmath49 for @xmath143 is independent of the nature of the solution ( bh , ns or p - l solution ) . thus if @xmath9 is constrained by & , no singularity at finite @xmath14 ( resp .
@xmath17 ) occurs in the open domain @xmath148 ( resp .
@xmath147 ) whatever the nature of the solution is .
now , we look for the conditions that make the ( limits of the ) functions @xmath46 finite at @xmath22 ( that of @xmath40 has been established in corollary 2 ) . the behavior of @xmath32 near the center is derived from as follows .
@xmath149 & = & 1 - \lim_{\rho\to c^+}[(ra)r '' ] + \lim_{\rho\to c^+}[r'\,t(\rho ) ] \\ \label{ben3}&= & 1 - \frac{2(c-3m)r''(c)}{3r'(c ) } + r'(c)\lim_{\rho\to c^+}t(\rho)\,,\end{aligned}\ ] ] where we have used to evaluate the first limit in . in
, @xmath150 is defined by @xmath151 which is by an indeterminate form 0/0 as @xmath152 .
using eq with @xmath61 , it is straightforward to show that @xmath153 .
thus @xmath154 solving for @xmath155 then substituting back into yields @xmath156
. thus , for @xmath120 we have @xmath157 if @xmath9 is two times differentiable with @xmath125 ( @xmath9 need not satisfy all the conditions & ) and if @xmath158 , then as @xmath152 a ) the function @xmath32 ( resp .
@xmath40 ) for a ns solution with @xmath96 ( corollary 2 ) approaches @xmath108 as @xmath159 ( resp .
@xmath160 as @xmath161 ) and b ) the function @xmath32 ( resp .
@xmath40 ) for a bh solution ( @xmath98 by corollary 2 ) approaches @xmath160 as @xmath162 ( resp . @xmath108 as @xmath163 ) . concerning extreme nss with @xmath112
, we have the following result .
if @xmath9 is two times differentiable with @xmath125 ( @xmath9 need not satisfy all the conditions & ) and if @xmath158 , then as @xmath152 the function @xmath32 ( resp .
@xmath40 ) for the ns with @xmath112 approaches @xmath108 as @xmath163 ( resp . converges to @xmath164 by corollary 2 ) .
now , we will look for conditions to have a p - l solution and provide a proof for corollary 5 ( eqs , & ) . by corollary 2 ,
p - l solutions exist only if @xmath112 .
let @xmath112 . in part 1 of appendix
a , we will prove the general result that if @xmath112 then @xmath165 converges . with @xmath166 ,
eq reads @xmath167/{r^2}$ ] . taking the limit of the previous formula as @xmath168 , replacing @xmath165 by @xmath169 and using @xmath170 we obtain @xmath171 dropping @xmath169 from both sides of
, the remaining equation reads @xmath172 solving the last equation for @xmath169 we obtain @xmath173 eq reads @xmath174 using and @xmath170 in we obtain @xmath175 the first equation in reads @xmath176 notice that by & , @xmath177 , thus if @xmath158 ( @xmath178 ) then @xmath179 @xmath34 approaches @xmath108 as @xmath163 ; this establishes a proof for corollary 5 .
in this case @xmath32 & @xmath40 behave near the center as @xmath180 now , we further assume that @xmath181 ( with @xmath112 ) , the limit of both sides of as @xmath168 leads to @xmath182 since @xmath181 this implies @xmath183 by making thus the first limit in the r.h.s of an indeterminate form 0/0 .
next , consider the second term in @xmath184 where the second limit in vanishes by the imposed condition : @xmath181 .
the first and third limits are indeterminate forms 0/0 , leading to upon further differentiating @xmath185 eqs & result in @xmath186 in part 1 of appendix a we will prove the general result that if @xmath112 & @xmath187 then @xmath188 ( @xmath189 ) converges and provide a formula for @xmath190 . in our case , since @xmath181 & @xmath191 , the latter condition is satisfied and the desired formula for @xmath190 will have the form ( see eq ) @xmath192 notice that is derived from after dropping the first term , which is zero : @xmath193 .
eq implies @xmath194 @xmath195 @xmath196 @xmath195 @xmath197 + @xmath198^{1/2}$ ] @xmath195 @xmath199^{1/2}=0 $ ] ( by ) , thus @xmath3 is differentiable at @xmath22 and consequently it is continuous there , which means that @xmath200 exists and is finite .
finally , we managed to make the three function @xmath49 finite at @xmath22 , thus finite everywhere in the domain @xmath36 under the following conditions .
suppose @xmath9 is four times differentiable at every point of the domain @xmath36 and satisfies the conditions & ( @xmath201 & @xmath202 need not exist on the whole domain @xmath126 ; rather , on a small half - open interval containing @xmath22 ) . if @xmath181 , @xmath191 & @xmath112 , any spherically symmetric , asymptotically flat solution to eqs to is a one parameter p - l configuration with @xmath113 , @xmath183 & @xmath203 .
table [ tab1 ] summarizes the results of the above discussions and includes some results from the following section .
1.011|l|l|l|l| corollary 2 : & @xmath112 ( nc ) : & a ) @xmath204 ( sc ) : & @xmath205 ( sc ) : + @xmath73 exists & @xmath206 & @xmath207 & @xmath207 + & & b ) @xmath112 : possible & + corollaries 4 & 5 : & & a ) @xmath204 ( sc ) : & @xmath205 ( sc ) : + @xmath73 & @xmath145 exist , & & @xmath207 & @xmath207 + @xmath158 & & @xmath208 & @xmath208 + & & b ) @xmath112 : & + & & @xmath206 & + & & @xmath209 & + theorem : & @xmath112 ( sc ) : & @xmath204 ( sc ) : & @xmath205 ( sc ) : + @xmath73 , @xmath145 , @xmath210 & @xmath206 & @xmath183 & @xmath207 & @xmath207 + & @xmath122 exist , & @xmath211 & @xmath212 & @xmath212 + @xmath181 , @xmath191 & @xmath213 & @xmath213 & @xmath213 + most of the solutions constructed previously @xcite were derived using the inverse problem method , which we are aiming to use in section [ sec4 ] .
the method consists in choosing a suitable function @xmath9 then using eqs , & to obtain @xmath49 . from the above discussions ( five corollaries and a theorem )
, we have learnt how to set conditions on @xmath9 and to fix the sign of @xmath214 in order to obtain a predefined final solution .
for instance , for @xmath9 given by @xmath215 ( where @xmath216 ) if we choose @xmath217 , then @xmath9 satisfies all the requirements of the theorem but one ; if we further choose @xmath112 in , then eqs , & lead to a one - parameter p - l solution for each value of @xmath217 . if , instead , we choose @xmath96 or @xmath98 the solution is by corollary 2 a two - parameter ns or a bh , respectively , for each value of @xmath217 .
notice how crucial is the condition @xmath191 for obtaining p - l solutions for the set of the continuous everywhere four times differentiable functions @xmath9 satisfying the conditions & . for @xmath218 , this condition fails ( @xmath181 & @xmath219 ) and the above formula gives rise to no p - l solution , whatever the value of @xmath90 is ; even when @xmath112 , the potential @xmath32 still diverges at @xmath22 as @xmath220 ( see eq ) .
similar conclusion holds for @xmath221 . in the following section
, we will derive further properties of the solutions at both the center and spatial infinity for more restrictive conditions on @xmath9 .
we will also discuss the null and weak energy conditions .
here we study the properties of the asymptotically flat solutions ( in their generic forms ) by investigating their behavior near the center and at spatial infinity .
we also examine the weak energy condition ( wec ) at these end - points and the null energy condition ( nec ) on the whole range of @xmath17 . in this section
, we rather assume that @xmath9 is five times differentiable : @xmath222 the behavior of @xmath40 near the center is discussed in 1 ) part 1 of appendix a for @xmath9 satisfying & with @xmath112 and 2 ) part 2 of appendix a for @xmath9 satisfying & with @xmath120 .
if @xmath9 satisfies & , @xmath49 take the following forms near the center @xmath18 ( @xmath15 ) , where @xmath190 is determined in part 2 of appendix a by @xmath223 @xmath224}{3r'^{\,4}(c)}\,\frac{1}{x}\nonumber \\ \label{v11 } & & -\frac{3a''(c)}{2}-\frac{14(c-3m)r''^{\,3}(c)}{r'^{\,5}(c ) } - \frac{10r'''(c)}{3r'^{\,3}(c ) } + \frac{157(c-3m)r''(c)r'''(c)}{18r'^{\,4}(c ) } \\ & & - \frac{3(c-3m)r^{(4)}(c)}{4r'^{\,3}(c ) } + \frac{14r''^{\,2}(c)}{r'^{\,4}(c ) } - \frac{3g}{r'^{\,5}(c)}\ln x + \frac{h}{r'^{\,6}(c)}\,x\ln x+\cdots\,,\nonumber\end{aligned}\ ] ] where @xmath225 and @xmath226 , @xmath227 & @xmath228 are polynomials of @xmath229 .
the polynomials @xmath226 & @xmath227 are given by @xmath230}{18}\,,\\ g & = & \frac{4r'''(c)r'^{\,2}(c)-15r''^{\,2}(c)r'(c)}{3 } + 5(c-3m)r''^{\,3}(c ) \\ & & - \frac{10(c-3m)r''(c)r'''(c)r'(c)}{3 } + \frac{(c-3m)r^{(4)}(c)r'^{\,2}(c)}{3}\,.\end{aligned}\ ] ] from now on and throughout the rest of this paper , we choose @xmath231 and @xmath232 ( precisely @xmath233 ) in .
the function @xmath9 being concave down on its range , the first nonzero derivative @xmath234 , @xmath235 , is negative ( @xmath236 always ) .
the behavior of @xmath39 depends of the first non - zero @xmath234 : @xmath237 where @xmath238 ( @xmath239 and we have chosen @xmath233 ) .
the behavior at spatial infinity is somehow defined by the boundary condition .
this condition does not completely fix the behavior of @xmath49 at spatial infinity .
we fix that by requiring @xmath9 of the form @xmath240 .
the constant @xmath241 can be absorbed in the first term upon performing the translation at @xmath15 , which is now @xmath242 , without modifying the scaling constant ( @xmath243 ) . in appendix
a , we perform a similar translation and work with a radial coordinate @xmath244 whose value at @xmath15 is @xmath107 instead of @xmath22 . such a translation is no trouble , only some equations , like , change the form ( compare with ) , however , the derivations and conclusions are the same . ] @xmath245 .
thus , at spatial infinity we assume @xmath9 of the form @xmath246 this is the simplest form to which one may bring @xmath9 to .
the solutions discussed in @xcite have the same expansion as in . upon substituting into then into , one obtains @xmath247 notice that , since the function @xmath9 is concave down on its range , the first nonzero constant @xmath248 in is negative .
accordingly , the series expansions for @xmath3 , which are derived from or , read @xmath249 one draws the following conclusions concerning the behavior of @xmath250 as @xmath47 .
although @xmath251 as @xmath47 , it may approach zero from above or from below by adjusting the parameters @xmath252 .
for instance if @xmath253 , @xmath254 & @xmath255 , the potential is positive at spatial infinity , which means the weak energy condition ( wec ) is satisfied . since in this case @xmath256 ,
the previous condition on @xmath51 is satisfied if @xmath257 : the mass of the solution is bounded from above by @xmath258 .
particular solutions where the mass is bounded from above have been obtained in @xcite .
conversely , if the mass is bounded from blow @xmath259 , the potential is negative at spatial infinity .
we will show in the following subsection that in this case the wec is also satisfied . if , otherwise @xmath260 but @xmath261 , the potential is always positive at spatial infinity ( @xmath262 ) since in this case @xmath263 an again the wec is satisfied .
the scalar @xmath3 , which also vanishes at spatial infinity , has always the same sign as @xmath47 . finally , for @xmath31
, @xmath40 always approaches its limit 1 from below . at spatial infinity
, we assume that @xmath264 admits an expansion in powers of @xmath265 given by . then if @xmath253 ( @xmath256 ) & @xmath254 , there is a critical mass @xmath266 beyond which ( @xmath267 ) the potential is negative at spatial infinity ; if , instead , @xmath268 the potential is positive . for @xmath269 the potential may have any sign . in all cases , the wec is satisfied at spatial infinity . in the paragraph preceding corollary 3
we have mentioned that the finiteness of @xmath49 for @xmath143 is independent of the nature of the solution . a first important and yet general similar result from the set of eqs , & is that @xmath3 is always finite at the center , provided conditions & are satisfied .
although @xmath270 is finite for @xmath112 , as seen from eq , it is obvious from eq that the condition @xmath112 alone does not ensure existence of p - l solutions for , in this case , @xmath32 diverges near the center as @xmath271\to -\infty$ ] and the solution is a ns .
similarly , the conditions @xmath272 do not ensure existence of p - l solutions for @xmath32 diverges near the center as @xmath273 and the solution is again a ns .
other important conclusions derived from eqs & , not discussed yet in the literature , follow . under constraints & : + a ) the field eqs to do not admit bh solutions ( @xmath98 ) where the metric @xmath40 is finite everywhere .
however , one can derive a bh solution where both @xmath46 are _ finite _ everywhere by choosing @xmath274 ; in this case , @xmath40 diverges at the center as @xmath275\to -\infty$ ] ( or ) .
one may call such a solution a
semi - regular " bh since all matter fields are finite .
we have a set of such two - parametric solutions given by provided we take @xmath276 to ensure that @xmath274 and in we choose @xmath98 .
the simplest of these solutions is @xmath277 in ref @xcite , another semi - regular " black hole solution with @xmath40 finite everywhere has been constructed , however , by choosing a function @xmath9 not differentiable at the center for which @xmath278 ( in our solutions @xmath125 ) .
this choice of @xmath9 caused @xmath46 to diverge at the center .
+ b ) similarly , if @xmath96 the field eqs to do not admit ns solutions where @xmath40 is finite everywhere . however ,
if @xmath96 one can obtain a ns where both @xmath46 are _ finite _ everywhere by choosing @xmath274 ; in this case , @xmath40 diverges at the center as @xmath275\to + \infty$ ] ( or ) .
we have a set of such two - parametric solutions given by provided we take @xmath276 to ensure that @xmath274 and in we choose @xmath96 .
the simplest of these solutions is @xmath279 + c ) for any ns with @xmath112 , @xmath40 is finite everywhere and @xmath32 diverges at the center as @xmath273 if ( @xmath280 ) or as @xmath271\to -\infty$ ] if @xmath158 .
a solution of the case ( @xmath280 ) is ( eq with @xmath218 ) @xmath281 and a solution of the case @xmath158 is ( eq with @xmath221 ) is @xmath282 for the case of self - gravitating real scalar field , the nec and wec are the mostly discussed conditions in the literature @xcite . if @xmath283 are _ any _ future - directed null ( @xmath284 ) and timelike ( @xmath285 ) four - vectors , the nec and wec state @xmath286 it was established in @xcite that the conditions which rule out wormholes , horns and flux tubes rest on the validity of the nec . in this section we will show that (
p - l , ns or bh ) solutions to eqs to do not violate the nec .
hence , the consideration of the nec does not allow for further selection criteria .
let @xmath287 denotes the sign of @xmath40 such that @xmath288 ( @xmath289 only inside a bh ) .
a tetrad frame associated with the metric is given by @xmath290 future - directed null ( @xmath291 ) and four - velocity ( @xmath292 ) vectors take the forms @xmath293 for the static , spherically symmetric solutions to eqs to , the fields @xmath49 depend only on @xmath17 .
eqs & lead then to @xmath294 , @xmath295 and @xmath296 if @xmath297 .
hence @xmath298 since @xmath299 ( the r.h.s is zero if @xmath300 ( with @xmath301 ) or / and @xmath181 ) , solutions to eqs to do not violate the nec .
notice that the nec is fulfilled even if the field eqs to are not satisfied .
similarly , we obtain @xmath302\nonumber\\ \label{ec5 } = & n^2\{[\epsilon ( 1-s_1 ^ 2)-s_2 ^ 2-s_3 ^ 2]v + [ 1+s_1 ^ 2-\epsilon s_2 ^ 2-\epsilon s_3 ^ 2]|a|\phi'^{\,2}/2\}\,.\end{aligned}\ ] ] the number in the first square parenthesis of is positive by ( @xmath303 real ) .
this leads to @xmath304 if @xmath301 or @xmath305 if @xmath289 .
thus , at the points where @xmath33 , the r.h.s of is positive and the wec is satisfied . the sign of the r.h.s of can not be determined without knowing the solution to eqs to .
since the functions @xmath306 are related to @xmath9 and its derivatives via eqs , & , it suffices to know the shape of @xmath9 to completely determine the sign of the r.h.s of . in this paper
we have chosen @xmath9 by imposing the conditions & at spatial infinity and the conditions & at the center .
consider @xmath9 of the form and let @xmath307 be the mass density and radial pressure due to the field @xmath3 .
expressing @xmath306 in powers of @xmath265 , we obtain @xmath308a_\ell}{(\ell+3)\rho^{\ell+3 } } + \cdots\,,\;\;(\ell\geq 1)\\ & \rho^\phi = t^{t}{}_{t}=-\frac{2\ell(\ell+1)^2a_\ell}{(\ell+3)\rho^{\ell+3 } } + \cdots\ , , \;\ ; p^\phi_\rho = -t^{\rho}{}_{\rho}=-\frac{4\ell(\ell+1)a_\ell}{(\ell+3)\rho^{\ell+3 } } + \cdots\,,\nonumber\end{aligned}\ ] ] ( at spatial infinity @xmath301 & @xmath309 ) . here
@xmath248 is the first non - zero constant @xmath252 in : @xmath310 .
thus , the dominant term in the r.h.s of is positive .
\a ) the spherically symmetric , asymptotically flat solutions to eqs to do not violate the nec .
b ) at the points where @xmath33 , the wec is satisfied and the mass density is positive .
c ) if the function @xmath264 admits an expansion in powers of @xmath265 , then at spatial infinity the spherically symmetric , asymptotically flat solutions to eqs to , whether @xmath311 or @xmath312 , do not violate the wec and the mass density and radial pressure are positive .
the consideration of the wec does not allow for further selection criteria at least at spatial infinity ; this is not the case near the center where it is generally possible to select within the same family ( p - l , ns or bh solution ) those solutions which do not violate the wec , as will be established shortly below . for that purpose , we focus on the case of a bh solution then tabulate the results for the other solutions .
we assume the conditions & hold and use eqs & . for a bh solution @xmath313 , @xmath314 & @xmath315 .
if @xmath158 ( @xmath178 ) , in the limit @xmath316 eq leads to @xmath317 , so the wec is satisfied . if @xmath181 , @xmath318 so that the wec is satisfied if @xmath319
now , if @xmath181 & @xmath191 , we have @xmath320 and the wec is satisfied if @xmath321 . finally , if @xmath181 , @xmath191 & @xmath322 , @xmath323 and the wec is satisfied if @xmath324 .
table [ tab2 ] shows the conditions for which the wec is satisfied near the center .
there is no case of validity of the wec for a ns with @xmath325 & @xmath181 ( case 1b ) in table [ tab2 ] ) : the wec is violated @xmath326 .
in fact , under these conditions we have @xmath327 .
the wec is violated if @xmath328 & @xmath329 have opposite signs . since for a ns @xmath301 , @xmath330 & @xmath331
, this implies @xmath332 , thus @xmath328 may have both signs .
for instance , for a four - velocity vector or an observer with @xmath333 & @xmath334 , @xmath335 and the wec is violated if @xmath319 ; for an other observer with @xmath336 & @xmath334 , @xmath337 and the wec is violated if @xmath338 . similarly , there is no case of validity of the wec for a ns with @xmath339 & @xmath178 ( case 2a ) in table [ tab2 ] ) .
.static , spherically symmetric , asymptotically flat solutions to eqs to .
validity of the wec near the center for p - l , ns & bh solutions .
@xmath9 is subject to conditions & .
[ cols= " < , < , < " , ]
the purpose of this section is not to discuss in detail the physical properties of the derived solutions ; rather it is to show how to construct predefined solutions using the first five corollaries .
the existence of a critical mass @xmath340 and the validity the wec at spatial infinity and that of the nec will be the cases since for all solutions derived in this section and those discussed earlier , eqs & to , the function @xmath264 admits an expansion in powers of @xmath265 at spatial infinity .
we will use the inverse problem method by which we select a function @xmath9 and integrate eqs & to obtain @xmath35 & @xmath39 , respectively , and derive @xmath34 from eq .
for the three solutions presented below , the expressions for @xmath127 , if they are available , are given in the corresponding part of appendix b. let @xmath9 be defined by @xmath341 eqs , & lead to eqs for @xmath342 . by corollaries 4 & 5 , the solution is either a bh or a ns . using eq with @xmath343 & @xmath344
we conclude that @xmath32 is always negative at spatial infinity for positive masses ( @xmath345 ) whatever the solution is . at the center @xmath22 , @xmath40 & @xmath32
have opposite signs : @xmath346 & @xmath347 .
[ fig2 ] depicts a bh solution for @xmath348 .
the horizon is at @xmath349 ( @xmath110 by corollary 2 ) , the potential @xmath32 crosses the axis at @xmath350 and reaches its minimum value at @xmath351 . the behavior of the solution is as stated in corollary 4 .
+ fig .
[ fig3 ] depicts a ns solution for @xmath352 the potential @xmath32 of which is always negative .
the behavior of the solution is as stated in corollary 5 .
+ in the previous example it was not possible to derive a bh solution with @xmath32 positive at spatial infinity or a ns with @xmath32 positive near the center .
we realize that by making another choice for @xmath9 @xmath353 where @xmath354 is any positive number and the center @xmath355 . at spatial infinity
we have @xmath356 . by corollary 6 , the potential is positive as @xmath47 if @xmath357 , regardless of the nature of the solution .
a bh with @xmath358 or a ns with @xmath359 has its potential positive as @xmath47 . at the center @xmath22
, on can show using eqs & , with @xmath9 given by , that @xmath40 & @xmath32 have always opposite signs for all @xmath31 & @xmath360 .
eqs , & lead to eqs to for @xmath49 . by corollaries 4 & 5 ,
the solution is either a bh or a ns .
taking @xmath361 & @xmath362 ( @xmath363 ) , one obtains the bh solution depicted in fig .
[ fig4 ] where the horizon is at @xmath364 ( @xmath110 by corollary 2 ) .
the potential crosses the axis at @xmath365 & @xmath366 and reaches its minimum and maximum values at @xmath367 & @xmath368 , respectively .
the solution behaves as stated in corollary 4 .
+ taking @xmath369 & @xmath370 , one obtains the extreme ns depicted in fig .
the function @xmath40 is first concave up then concave down , which is obvious from the two plots of @xmath40 in fig .
the solution behaves as stated in corollary 5 .
+ in the two previous example & it was not possible to derive a p - l solution .
we realize that by making a third choice for @xmath9 . we choose @xmath9 of the form with @xmath371 ( the simplest case ) . in order to fulfill the requirements of the theorem we take @xmath112 . since the einstein - scalar theory is scale invariant ,
we restrict ourselves to the case @xmath372 .
once the expression of @xmath35 is obtained for @xmath372 , one generates the general expression of @xmath40 for @xmath373 by replacing @xmath17 by @xmath374 and @xmath51 by @xmath375 in the r.h.s of @xmath35 .
hence , we start with @xmath376 eq leads to eq for @xmath40 .
the solution behaves as stated in the theorem with @xmath377 @xmath378 ( eq ) . by corollary 6 , positive masses ( @xmath379 )
have negative potential at spatial infinity .
+ + the p - l solutions derived from for @xmath217 have similar behaviors with @xmath32 negative everywhere .
plots of @xmath40 & @xmath32 for @xmath380 and the same mass and center are shown in fig .
as @xmath381 , @xmath40 & @xmath32 approach their minkowskian limits .
such a behavior is desirable in theories where the effects of the scalar field are confined .
the selection criteria we have established in the first five corollaries and theorem have been used to help discussing and constructing explicit solutions of particular interest . in section [ sec2 ]
we provided a unified formula for @xmath9 ( eq ) which generates different p - l , ns & bh solutions ( behavior of @xmath32 and horizon location ( for bhs ) depend on @xmath137 ) . in section [ sec3 ]
we have shown the existence of new kind of bhs ( eq ) where the metric @xmath40 is singular at the center only and the matter fields @xmath32 & @xmath3 are regular everywhere .
a parallel conclusion applies to the existence of nss ( eq ) .
we have shown that the nec ( any @xmath32 ) and the wec ( only @xmath33 ) are always satisfied . particularly , if @xmath382 admits an expansion in terms of @xmath265 , then 1 ) there exists a sort of critical mass ( if @xmath253 ) such that @xmath32 at spatial infinity & @xmath383 have the same sign , regardless the nature of the solution .
2 ) the solutions to the field equations obey the wec at spatial infinity ( any @xmath32 ) .
if the derivatives of @xmath14 at the center and @xmath384 have specific signs , the wec is satisfied there .
we have constructed bhs , fig .
[ fig2 ] ( resp . fig .
[ fig4 ] ) , the potential function of which has one local minimum value ( resp .
one local minimum and one local maximum values ) .
the two bhs behave differently at spatial infinity .
we have also discussed nss , fig .
[ fig3 ] ( resp . fig .
[ fig5 ] ) , the potential function of which is negative everywhere ( resp . has both signs ) .
we have constructed families ( a family for each value of @xmath217 in ) of one - parameter p - l solutions , figures [ fig6 ] & [ fig7 ] , which have negative potential everywhere .
each family of p - l solution is different from that constructed in ref @xcite .
appendix a is divided into two parts and its purpose is three - fold . in part 1 , where @xmath112 and @xmath9 satisfies the conditions & , we prove that a ) @xmath385 is finite and b ) if @xmath187 ( particularly if @xmath181 & @xmath191 ) then @xmath386 is finite and provide a value for the latter . in part 2 , where @xmath120 and @xmath9 satisfies the conditions & , we derive formulas , & for @xmath49 .
we introduce the variable @xmath387 .
while @xmath388 is supposed to be four times differentiable by hypotheses & , @xmath389 need be defined on a small half - open domain including the center @xmath390 . by taylor s theorem
there exists a number @xmath391 such that @xmath392 now , we introduce the continuous function @xmath393\ , , & \text{if}\ ; x>0 \\ 1\ , , & \text{if}\ ; x=0 \end{array}\,.\right .\ ] ] since @xmath388 has only one zero : @xmath19 , the function @xmath394 has no zero in the half - open domain @xmath395 .
the derivatives of @xmath396 up to order @xmath397 exist and are finite for @xmath395 . particularly , the first derivative of @xmath396 depends on @xmath93 and its second derivative depends on @xmath201 : @xmath398 . for
all @xmath399 we have @xmath400 next , consider the continuous function @xmath401 defined by @xmath402\ , , & \text{if}\ ; x>0 \\ 1/r'^{\,4}(c)\ , ,
& \text{if}\ ; x=0 \end{array}\,.\right .\ ] ] the function @xmath403 and its first three derivatives are finite and continuous on the domain @xmath395 and have no poles in it . by taylor s theorem
there exists a number @xmath404 such that @xmath405 notice how the derivative of order @xmath406 of @xmath403 depends on the derivative of order @xmath137 of @xmath14 .
we introduce an arbitrary constant @xmath407 : @xmath408 and let @xmath409 , @xmath410 , @xmath411 , @xmath412 & @xmath413 be the constants @xmath414 with @xmath112 we rewrite using the new variable @xmath244 and the constant @xmath409 @xmath415 for @xmath416 ( @xmath417 ) , @xmath418 . thus using & in leads to @xmath419 since @xmath420 has no pole in the domain @xmath395 , @xmath421 is finite and continuous on it . taking the derivative of with respect to @xmath244 then the limit as @xmath422 we obtain eq .
the non - vanishing contributions come from the second term in the r.h.s of , which has a derivative @xmath423 as @xmath422 , and the third one the derivative of which approaches @xmath424 as @xmath422 .
we have thus proven that @xmath425 is finite whenever @xmath112 .
it is obvious from that a necessary condition for @xmath386 to be finite is @xmath426 we assume that holds .
now , @xmath427 is provided by @xmath428\,.\end{aligned}\ ] ] all functions in the r.h.s of are continuous and finite for @xmath395 ( @xmath429 depends on @xmath201 which exists and is finite by hypothesis ) . using ,
& in , one sees that the only non - finite contributions to @xmath386 may come from the second and third terms in .
these two contributions to @xmath430 are @xmath431 thus , we have proven that @xmath386 is finite whenever condition is satisfied . as we have seen earlier , the potential @xmath32 is finite at the center only if @xmath181 & @xmath191 ( eqs & ) . in this case
eq is satisfied .
we postpone the derivation of a formula for @xmath190 to part 2 .
we will see that for @xmath112 , @xmath181 & @xmath191 , @xmath190 reduces to @xmath384 @xmath432 @xmath388 is supposed to be five times differentiable by hypotheses & . as before @xmath433 need be defined on a small half - open domain including the center @xmath390 . replace & , respectively , by @xmath434 ( @xmath435 depends on @xmath436 ) .
@xmath40 ( eq ) takes the form @xmath437r^2(x+c ) - 2r^2(x+c)\int_{x_0}^{x}\frac{\sigma(y)\,\text{d}y}{y^3 } - 2(c-3m)r^2(x+c)\int_{x_0}^{x}\frac{\sigma(y)\,\text{d}y}{y^4}\,.\end{aligned}\ ] ] define the new continuous functions @xmath438 & @xmath439 by @xmath440 replacing @xmath441 in by the r.h.s of and using , we arrive at @xmath442r^2(x+c ) - 2r^2(x+c)\bigg[i_1(x)-\frac{1}{2r'^{\,4}(c)x^2 } + \frac{2r''(c)}{r'^{\,5}(c)x}+\delta\ln x+\lambda x - ( x_0\rightarrow x)\bigg]\nonumber\\ \label{nw6 } & -2(c-3m)r^2(x+c)\bigg[i_2(x)-\frac{1}{3r'^{\,4}(c)x^3 } + \frac{r''(c)}{r'^{\,5}(c)x^2}-\frac{\delta}{x}+\lambda\ln x - ( x_0\rightarrow x)\bigg]\,,\end{aligned}\ ] ] where ( @xmath443 ) substitutes @xmath407 for @xmath244 in the terms within the square parentheses . in order to obtain the behavior of @xmath40 & @xmath32 near the center ( for small @xmath244 ) , we replace @xmath388 , @xmath438 & @xmath439 in by the r.h.s of , @xmath444 & @xmath445 , respectively , then extract the coefficients of @xmath446 , @xmath244 , @xmath447 , @xmath448 and the independent term , we arrive first at , then using , we obtain .
the derivation of is straightforward .
the coefficient of @xmath448 in reads @xmath449+\frac{r'''(c)}{3r'^{\,3}(c ) } -\frac{15r''^{\,2}(c)}{4r'^{\,4}(c)}\\ & + ( c-3m)\bigg[-2r'^{\,2}(c)\int_{\infty}^{x_0}\frac{\sigma(y)\,\text{d}y}{y^4 } -\frac{2}{3r'^{\,2}(c)x_0 ^ 3}+\frac{2r''(c)}{r'^{\,3}(c)x_0 ^ 2}-\frac{2r'^{\,2}(c)\delta}{x_0}\nonumber\\ & 2r'^{\,2}(c)\lambda \ln x_0 + 2r'^{\,2}(c)[i_2(x_0)-i_2(0)]+2\delta r'(c)r''(c)+\frac{r^{(4)}(c)}{18r'^{\,3}(c ) } -\frac{r''^{\,3}(c)}{2r'^{\,5}(c)}-\frac{5r''(c)r'''(c)}{9r'^{\,4}(c)}\bigg]\,.\nonumber\end{aligned}\ ] ] @xmath190 depends apparently on @xmath407 , however , using eqs & it is straightforward to obtain @xmath450/\text{d}x_0 $ ] @xmath451 .
hence , @xmath190 is independent of @xmath407 and is finite since every function or constant in the r.h.s of is finite . in the limit @xmath452 , many terms drop from the r.h.s of and
the remaining terms provide the desired expression for @xmath190 @xmath453\\ & + ( c-3m)\lim_{x_0\to 0^+}\bigg[-2r'^{\,2}(c)\int_{\infty}^{x_0}\frac{\sigma(y)\,\text{d}y}{y^4 } -\frac{2}{3r'^{\,2}(c)x_0 ^ 3}+\frac{2r''(c)}{r'^{\,3}(c)x_0 ^ 2}-\frac{2r'^{\,2}(c)\delta}{x_0}+ 2r'^{\,2}(c)\lambda \ln x_0\bigg]\nonumber\\ & + \frac{r'''(c)}{3r'^{\,3}(c)}-\frac{15r''^{\,2}(c)}{4r'^{\,4}(c ) } + ( c-3m)\bigg[2\delta r'(c)r''(c)+\frac{r^{(4)}(c)}{18r'^{\,3}(c ) } -\frac{r''^{\,3}(c)}{2r'^{\,5}(c)}-\frac{5r''(c)r'''(c)}{9r'^{\,4}(c)}\bigg]\,,\nonumber\end{aligned}\ ] ] where the limit in the r.h.s depends on the particular form of @xmath388 and can not be obtained in closed form . if @xmath339 , @xmath181 & @xmath191 , eq reduces to .
this section is divided into three parts . in part 1
to part 3 , we show the expressions of @xmath127 , if they are available , corresponding to subsections [ suse1 ] to [ suse3 ] , respectively .
the functions @xmath127 of subsections [ suse2 ] take the forms @xmath455}{3 ( 4 \rho ^2 - 4 b \rho - b^2)^2 } + \frac{30 b m-13 b^2 - 2(11 b -18 m ) \rho } { 4 \rho ^2 - 4 b \rho - b^2}\nonumber\\ & & -\frac{3 ( 11 b-18 m ) \ell(\rho)}{32 \sqrt{2 } b^3}+\frac{109 b-318 m-3\sqrt{2}(11 b+18 m ) \ell(\rho)}{192 b ( \rho -b ) ^2}\\ & & + \frac{89 b-6 m+3\sqrt{2}(11 b-18 m ) \ell(\rho)}{384 ( \rho -b)^3}-\frac{5 [ ( 264 b m-76 b^2 + 3b\sqrt{2}(11 b-18 m ) \ell(\rho)]}{3072 ( \rho -b)^4}\,,\nonumber\\ \label{su22}\phi & = & \frac{3 \pi } { 2 } - 4 \arctan\left((\sqrt{2}+1 ) \sqrt{\frac{2
\rho -(\sqrt{2}+1 ) b}{2 \rho + ( \sqrt{2}-1 ) b}}\right)\,,\end{aligned}\ ] ] where @xmath456 .
the function @xmath35 of subsections [ suse3 ] is derived here . with our choice , the integrand in is a rational function of the form @xmath457 $ ] , @xmath458 , @xmath459 , @xmath460 & @xmath461 . with @xmath462 , @xmath463 & @xmath464
we obtain @xmath465 + \frac{3\pi k\beta^2\gamma}{\delta^5}\right .
\nonumber\\ & -\frac{10\pi\beta^4(2j\beta^2-i\gamma)}{\delta^7 } - \frac{\pi(2q\alpha+h\gamma)}{2\alpha\delta } - \frac{2\pi \nu\beta^2}{\delta^3 } + \frac{\pi \mu\gamma}{\delta^3 } - \frac{6\pi l\beta^4}{\delta^5 } - \frac{\mu(2+\gamma \rho)}{\delta^2\omega } + \frac{10j\beta^4\omega'}{\delta^6\omega } + \frac{l\omega'}{2\delta^2\omega^2}\nonumber\\ & -\frac{k(2+\gamma \rho)}{2\delta^2\omega^2 } + \frac{5j\beta^2\omega'}{3\delta^4\omega^2 } - \frac{5i\gamma\omega'}{6\delta^4\omega^2 } -\frac{5i\beta^2\gamma\omega'}{\delta^6\omega } + \frac{3l\beta^2\omega'}{\delta^4\omega } - \frac{3k\gamma\omega'}{2\delta^4\omega } + \frac{\nu\omega'}{\delta^2\omega } + \frac{j\omega'}{3\delta^2\omega^3 } - \frac{i\gamma\omega'}{6\beta^2\delta^2\omega^3 } \nonumber\\ \label{su3 } & \left .
+ \left(\frac{20\beta^4(2j\beta^2-i\gamma)}{\delta^7 } + \frac{6\beta^2(2l\beta^2-k\gamma)}{\delta^5 } + \frac{2(2\nu\beta^2-\mu\gamma)}{\delta^3 } + \frac{2q\alpha+h\gamma}{\alpha\delta}\right ) \arctan\left(\frac{\omega'}{\delta}\right)\right\}2r^2(\rho)\,,\end{aligned}\ ] ] where @xmath466 is the derivative of @xmath467 with respect to @xmath17 and @xmath54 , @xmath468 , @xmath86 , @xmath469 , @xmath470 , @xmath471 , @xmath472 , @xmath473 , @xmath474 , @xmath475 , @xmath476 & @xmath477 are constants : @xmath478 @xmath32 is derived from and @xmath3 from .
a computer - aided algebra leads to a sizable analytic expression for @xmath3 depending on the incomplete elliptic integral of the first kind and the complete elliptic integral of the third kind . though the plot of @xmath3 shown in fig .
[ fig6 ] has been drawn using the analytic expression of @xmath3 , a numerical evaluation of the integral in is as fast as the numerical evaluations of the two non - elementary elliptic functions of which @xmath3 depends .
99 matos , t. , guzmn , s.,urea - lpez , l.a.,nez , d. : scalar field dark matter , ( 2001 ) , arxiv : astro - ph/0102419 bronnikov , k.a .
, chernakova , m.s .
: charged black holes and unusual wormholes in scalar - tensor gravity , grav . cosmol .
* 13 * , 51 - 55 ( 2007 ) , arxiv : gr - qc/0703107 bronnikov , k.a.,fabris , j.c . : regular phantom black holes , phys .
96 * , 251101 ( 2006 ) , arxiv : gr - qc/0511109 nikonov , v.v . ,
tchemarina , j.v . ,
tsirulev , a.n . : a two - parameter family of exact asymptotically flat solutions to the einstein - scalar field equations , class .
quantum grav . * 25 * , 138001 ( 2008 ) bronnikov , k.a . :
spherically symmetric false vacuum : no - go theorems and global structure , phys .
d 64 * , 064013 ( 2001 ) , arxiv : gr - qc/0104092 bronnikov , k.a . , shikin , g.n . :
spherically symmetric scalar vacuum : no - go theorems , black holes and solitons , grav .
* 8 * , 107 - 116 ( 2002 ) , arxiv : gr - qc/0109027 bechmann , o. , lechtenfeld , o. : exact black - hole solution with self - interacting scalar field , class .
quantum grav .
* 12 * , 1473 - 1482 ( 1995 ) , arxiv : gr - qc/9502011 chan , k.c.k . ,
horne , j.h . ,
mann , r.b . :
charged dilaton black holes with unusual asymptotics , nucl . phys .
* b 447 * , 441 - 461 ( 1995 ) , arxiv : gr - qc/9502042 bekenstein , j.d . : black holes : classical properties , thermodynamics , and heuristic quantization , ( 1998 ) , arxiv : gr - qc/9808028 | we make a systematic investigation of the generic properties of static , spherically symmetric , asymptotically flat solutions to the field equations describing gravity minimally coupled to a nonlinear self - gravitating real scalar field .
seven corollaries and a theorem on selection criteria for two- and one - parametric solutions are proven and conditions for obtaining particle - like solution , black holes or naked singularities are derived .
a series of exact solutions in closed forms describing different black holes , naked singularities and particle - like solutions are provided .
* keywords . * einstein - scalar gravity ; black holes ; naked singularities ; particle - like solutions ; exact solutions ( 1,0)431 |
the low - mass x - ray binary ( lmxb ) gs 1826238 was discovered with _ ginga _ in 1988 ( makino et al . 1988 ) . due to its temporal and spectral similarities
to cyg x1 and gx 3394 , the source was originally tentatively classified as a black hole candidate ( bhc ) ( tanaka 1989 ) .
later optical studies lead to the identification of a @xmath6 mag ( and therefore low - mass ) optical counterpart ( barret et al .
the companion was subsequently found to have a 2.1 hr modulation ( and implied orbital period ) and a refined position of @xmath7 and @xmath8 ( j2000 ) ( homer et al .
1998 ) .
the spectral and temporal characteristics that tanaka ( 1989 ) initially used to associate gs 1826238 with a black hole system were later found to be present in other x - ray bursters ( e.g. , 4u 1608522 , yoshida et al . 1993 ) .
moreover , the photon index of its energy spectrum was measured to have a relatively low cut - off energy ( @xmath1 58 kev ) for a bhc , and was perhaps more indicative of the typically cooler neutron star ( ns ) hard x - ray spectra ( strickman et al . 1996 ) .
x - ray bursts from this source were first conclusively observed with _
bepposax _ by ubertini et al .
( 1997 ) , firmly establishing the source as a ns and strongly suggesting it to be weakly magnetized ( @xmath9 g ) .
the periodicity of the type i bursts from gs 1826238 has been remarkably stable over the span of years ( ubertini et al .
although quasi - periodic bursting is not unique among lmxbs , such consistency over long durations is indeed unusual .
the regular intervals between bursts suggest that the accretion rate is stable , that the accreted matter is completely consumed during the bursts , and that the fraction of the stellar surface covered prior to each burst is approximately constant .
investigations of burst recurrence rates and energetics have lead to convincing arguments that type i bursts stem from unstable thermonuclear burning of accreted hydrogen and helium ( e.g. , strohmayer & bildsten 2003 ) . as freshly accreted material falls onto the ns surface
, it is hydrostatically compressed by new material at a rate per unit area @xmath10 10@xmath11 g @xmath12 s@xmath13 , assuming isotropic accretion and a ns radius of 10 km .
the thermal energy deposited by the infalling matter causes temperatures in most of the thin ns `` atmosphere '' to exceed @xmath14 k , so that during the accumulation phase hydrogen burns via the hot cno cycle at a rate that is limited only by the mass fraction @xmath15 and not the temperature ( bildsten 2000 ) . within hours to days ,
the extreme gravity on the ns surface ( @xmath1 10@xmath16 cm s@xmath17 ) compresses the accumulated matter to densities high enough to trigger unstable thermonuclear ignition .
gs 1826238 , in particular , has near limit - cycle behavior with stable hydrogen burning during the accumulation phase followed by mixed hydrogen and helium burning triggered by thermally unstable helium ignition ( bildsten 2000 ) .
the @xmath4-parameter the ratio of the integrated persistent fluence between bursts to the burst fluence and the long burst duration ( @xmath1 150 s ) , imply that after thermonuclear ignition the hydrogen burns via the rapid - proton ( @xmath18 ) process where energy is released through successive proton captures and @xmath19 decays ( wallace & woosley 1981 ) . the measured @xmath4-value for gs 1826238 of @xmath1 42 ( galloway et al .
2004 ; hereafter g04 ) is remarkably consistent with theoretical predictions : the gravitational energy released during accretion onto a 1.4 m@xmath20 ns is about 200 mev per nucleon , while the energy released through thermonuclear fusion is about 5 mev per nucleon for a solar mix going to iron group elements , giving an expected value of 40 .
moreover , the variation in @xmath4 with the global accretion rate @xmath21 implies solar metallicity in the accreted layer ( g04 ) , although recent work using an adaptive nuclear reaction network shows that the critical mass required for a burst is independent of the composition of the accreted material ( woosley et al .
2004 ) , and so attempts to infer the metallicity of the fuel from burst properties is complicated . during accretion the gravitational potential energy released per nucleon is deposited partly in the disk , and partly in the boundary layer or on the ns surface . in low accretion rate ( atoll )
lmxbs , the dominant persistent emission spectral component is clearly inverse comptonization , where soft `` seed '' photons emitted from a cold layer in the accretion disk gain energy in successive scatterings with hot electrons in the disk or surrounding accretion disk corona ( adc ) ( church 2001 ) .
stable hydrogen burning between bursts and the thermal energy released on impact are observable as blackbody emission , assuming there is an optical path through which the radiation can escape .
the addition of a blackbody component in spectral models is not always necessary ( as we find in [ spec ] ) , which suggests this form of radiation is ( up)scattered or absorbed by the obscuring disk or adc .
another possibility is that the blackbody component is simply too weak to be observed relative to the comptonized emission .
mitsuda et al .
( 1989 ) proposed that most lmxb spectra can be described by a multi - temperature disk blackbody from the inner accretion disk plus blackbody emission from the ns which is comptonized in the local region of the star .
more recently , the dipping class of lmxbs has constrained the types of emission models that are most likely .
the constraints are provided by the fact that the models must fit the persistent emission at various levels of decreased flux resulting from partial occultation and absorption of the ns , accretion disk , and corona by the bulge in the outer disk where accretion flow from the companion star impacts .
the evolution of these spectra at various levels of flux are well - described by a model consisting of point - like blackbody emission plus comptonized emission from an extended accretion disk corona ( e.g. , church et al .
1997 , 1998 ; bauciska - church et al . 2001 ; smale et al .
2001 ) . during dipping ,
the blackbody component is observed to disappear rapidly , indicating that the emission comes from a point and not from an extended region . as expected ,
the comptonized component is observed to gradually decrease . by measuring the ingress times of the dip , church ( 2001 ) was able to estimate the size of the comptonizing region for several sources and found the radius of the region to be typically @xmath1 50,000 km .
this measurement is consistent with the value obtained much earlier by canizares ( 1976 ) for 3u 182030 .
the large extent of the emission region is evidence for a comptonized component that is emitted from an adc and not the disk itself .
moreover , smale et al .
( 2001 ) observed complete covering of this region in many sources , suggesting that the corona must be geometrically thin , as it would be unlikely that a corona with spherical geometry would be completely occulted . on the contrary ,
studies of emission and absorption features in dipping sources by boirin et al .
( 2005 ) show that the spectra are also consistent with a less strongly ionized absorber along the line of sight rather than a simple increase in absorption .
progressive covering of the comptonized component in their models is not required .
the persistent emission spectrum of gs 1826238 has been discussed by previous authors ( e.g. , barret et al . 1995 , ubertini et al . 1999 , del sordo et al .
1999 , in t zand et al .
1999a , barret et al .
2000 , kong et al .
the spectral model that has consistently produced an acceptable fit is a blackbody plus comptonized emission , the latter being modeled in xspec with either a cut - off power law ( cutoffpl ) or the more explicit comptt .
broadband _ bepposax _
spectra in the 0.1200 kev range found the spectrum to be consistent with the comptonization of a 0.6 kev wien spectrum by a plasma with @xmath22 kev , plus a 3.8 kev blackbody ( in t zand et al .
1999a ) . del sordo et al .
( 1999 ) found the spectrum to be well - fitted with a blackbody plus cut - off power law , with @xmath23 kev , @xmath24 , and cut - off energy @xmath25 kev .
strickman et al .
( 1996 ) and in t zand et al .
( 1999a ) also obtained a fit with this model and the results of each group are fairly consistent .
kong et al .
( 2000 ) fit the persistent emission spectrum from 0.510 kev with a blackbody ( @xmath26 kev ) plus power law ( @xmath27 ) .
the burst emission spectrum has been studied a few times previously . in t zand et al .
( 1999a ) studied burst spectra and found significant flux up to 60 kev , indicating that the burst emission may be comptonized in a similar manner as the persistent emission .
the peak blackbody temperature was measured to be @xmath1 2 kev , which cooled to about 1.3 kev over 100 s. ubertini et al .
( 1999 ) studied the spectrum for two bursts , though with only two separate integration bins of 13 s at burst peak and 43 s through the decay .
the peak spectrum was fit with a blackbody temperature of @xmath1 2.12.3 kev and the decay with temperature @xmath1 1.61.9 kev .
kong et al .
( 2000 ) fit the burst spectrum by fixing the power law component to its persistent value and adding a blackbody ; the peak temperature was @xmath28 kev .
marshall et al .
( 2003 ) searched the present data for absorption features during the burst peak and decay to search for evidence for the gravitational redshift to the surface of the ns .
no features were observed .
we begin in 2 by presenting the observations of gs 1826238 that were used in this study , and we specifically discuss the data preparation prior to spectral analysis . in [ bp ] we measure the burst periodicity and compare to previous measurements .
we use the burst recurrence time and energetics to discuss the type of burning occurring on the ns surface .
we conclude the section by discussing a probable detection of burst oscillations during the decaying phases of the burst . in
[ spec ] we examine the evolution of the broadband persistent emission spectrum between bursts to see whether or not any significant changes in the parameters occur during the accumulation phase .
we find that the persistent emission data from 0.5200 kev are best fit with a model characterized by absorbed emission from two distinct comptonizing regions , plus iron line emission .
we also report the detection of interstellar absorption features .
we study the evolution of the burst spectrum in [ burspec ] by adding a blackbody to the persistent emission spectrum .
the disk comptonization parameters are fixed at their persistent emission best - fit values , and the temperature of the boundary layer seed photons is tied to the blackbody temperature . from the blackbody parameters we derive a blackbody radius , and by accounting for the accretion disk covering factor , we constrain the source distance assuming a 1012 km ns radius . in [ geo ] we attempt to obtain information on the geometry of the burning by comparing the persistent and burst spectra .
we conclude by summarizing the major results or our work .
in this paper we utilize two simultaneous observations from 2002 july 29 , taken with _
chandra _ and _ rossi x - ray timing explorer _
( _ rxte _ ) . the 68.2 ks _ chandra _
observation was made using the advanced ccd imaging spectrometer ( acis ; garmire et al .
2003 ) and is sensitive to photons from 0.310 kev .
the source was focused onto one of the back - illuminated acis chips ( s3 ) . to minimize the adverse effects due to pile - up , and to gain better time resolution on the rise of the bursts , we used half - frame readout on the inner 4 acis chips with 1.74 s frame - time .
we use the `` level 2 '' event lists from the standard data processing , and apply a standard correction to destreak the acis - s4 chip caused by a flaw in the chip readout . to obtain high resolution spectra , we had the high energy transmission grating ( hetg ; canizares et al .
1992 ) inserted into the optical path . the 36 ks _ rxte _
observation uses the proportional counter array ( pca ; jahoda et al .
1996 ) , and the high energy x - ray timing experiment ( hexte ; rothschild et al .
the pca is made up of five proportional counter units ( pcus ) and is sensitive to photons from 260 kev .
the hexte instrument comprises two clusters , each of which contains four nai / csi scintillation detectors , and is sensitive to photons from 15250 kev .
both instruments have large effective areas ( @xmath1 6000 @xmath29 and 1400 @xmath29 , respectively ) and microsecond timing .
table 1 provides a summary of the gs 1826238 observations used in the spectral analysis .
clcr 2739 & _ chandra_/acis - s & 0.58.3 & 68.2 + 01,02,000 & _ rxte_/pca & 323 & 24.6 + 01,02,000 & _ rxte_/hexte & 17200 & 19.4 + the _ chandra _ hetg spectrometer is composed of the medium energy grating ( meg ) and the high energy grating ( heg ) . the two gratings are slightly offset and appear as `` whiskers '' traversing all acis chips .
the meg is calibrated from 0.35.0 kev and the heg is calibrated from 1.08.3 kev . in this analysis , we only made use of the 1@xmath30 order meg / heg data .
all of the _ chandra _ analysis made extensive use of the chandra interactive analysis of observations ( ciao ) version 3.01 software with calibration version ( caldb ) 2.26 .
we used the ciao routine _ tgextract _ for the extraction of grating spectra , _ mkgrmf _ to produce response matrices , and _ fullgarf _ to create the auxiliary response matrices .
we added the positive and negative diffraction orders and their corresponding auxiliary response files using the script _
systematic errors of 10% were derived from calibration observations of the crab pulsar and mrk 421 , and were added to both the persistent and burst emission fits .
we binned the persistent and burst emission data in the largest possible bins to obtain maximum statistical quality , while maintaining sufficient resolution to observe prominent absorption or emission lines . for the persistent emission analysis we bin the data in 1000 count pha bins , and in the burst analysis we use 500 count bins .
this generally led to bins with width @xmath1 50100 ev .
we excluded the meg 1@xmath30 order data from 0.80.9 kev because these data fall on a gap between the acis chips .
after preliminary analysis , we concluded that the data near 2.1 kev are affected by a residual calibration uncertainty in the response due to an iridium m - edge . to correct for this
, we included an inverse edge ( e.g. , miller et al .
2002 ) , frozen at 2.065 kev though allowing the optical depth to vary .
the best - fit optical depth was consistently @xmath31 .
in addition , the data from 2.02.5 kev are possibly affected by pile - up , which we estimate to be about 10% at the maximum .
spectral distortion due to pile - up is accounted for by the hetg systematics .
the extraction of pca and hexte energy spectra used scripts developed at the university of california , san diego and the university of tbingen that incorporate the standard software for _ rxte _ data reduction ( ftools ) .
the large effective area of the pca provides excellent statistics without rebinning the data , however the first three channels and the data above 23 kev were not used .
we placed systematic uncertainties of 0.5% on the pca data to account for the likely uncertainty in the response ( kreykenbohm et al .
the hexte data below 17 kev and above 200 kev are not included in the analysis .
we used successively larger energy bins at higher energies to account for the decreasing flux density . during the burst intervals hexte
did not yield a significant number of counts above 40 kev .
we included no systematics to the hexte data .
for analysis during the burst time intervals we added the data for the 3 bursts ( corresponding to the 2@xmath32 , 3@xmath33 , and 6@xmath34 _ chandra _ bursts ) viewed when pcus 0 and 2 were taking data ; pcus 0 and 2 had the greatest coverage in this observation .
the pha bins were added together using the ftools task _
as discussed in 1 , gs 1826238 has exhibited a nearly periodic bursting behavior since type i bursts from this source were first discovered in 1997 .
the remarkable consistency of the burst cycle can easily be seen in fig .
chandra _
light curve uses all hetg data from 0.38 kev , but does not include any counts from the zeroth order source region as these data are affected by severe pile - up . not only is the burst recurrence time approximately constant , but the light curve of any particular burst is virtually indistinguishable from any other ( see fig . 1 ( _ c _ ) ) . although the time between sequential bursts has been rather constant , the average burst recurrence time has been observed to be steadily decreasing over long time scales ( cocchi et al .
2000 , g04 ) . using a total of 44 _ rxte _ observations from 1997 november to 2002
july , g04 measured the burst interval to be 5.74 @xmath0 0.13 hr initially , 4.10 @xmath0 0.08 hr in 2000 , and 3.56 @xmath0 0.03 hr in 2002 .
we measured the average burst recurrence time of the five complete _ chandra _ intervals to be 3.54 @xmath0 0.03 hr ( 12750 @xmath0 102 s ; also in 2002 ) .
the individual intervals were measured from burst peak to peak and are listed in table 2 .
the 2002 recurrence time measurement by g04 was measured using the same pca data that we use in this paper , while our measurement is based on the _ chandra _ data .
[ fig1 ] cc[!b ] 1@xmath352 & 12861 + 2@xmath353 & 12884 + 3@xmath354 & 12510 + 4@xmath355 & 12994 + 5@xmath356 & 12502 + the 40% decrease in the burst recurrence time from 1997 to 2002 has been coupled with a 66% increase in the mean persistent flux ( g04 ) .
such behavior is expected since the increase in persistent luminosity is assumed to be due to an increase in the global accretion rate , and therefore less time is required to reach the critical amount of fuel . while the observations of g04 are consistent with this trend , bildsten ( 2000 ) found exactly the opposite behavior in many low accretion rate lmxbs .
for example , bursts from 4u 170544 became less frequent as @xmath21 increased .
to explain the conundrum , bildsten ( 2000 ) suggested that a greater fraction of the stellar surface of 4u 170544 may be covered prior to ignition , so that the accretion rate per unit area actually decreased
. this may also be the case for gs 1826238 : if the relation between the rise in the persistent flux and the decrease in the burst recurrence time were linear , we would expect an even greater decrease in burst recurrence time than has been observed by g04 .
another possible explanation for the longer - than - expected recurrence time regards the fact that at smaller recurrence times , less helium will be made prior to ignition , and so the column depth required to produce the instability will increase ( cumming & bildsten 2000 ) . the location of the burning during bursts is not assumed to be spherically symmetrical .
it is likely that ignition begins near the stellar equator since matter is preferentially deposited there during accretion ( spitkovsky et al .
the anisotropic burning caused by the hot spot at the point of ignition may be revealed through the observation of burst oscillations , the frequencies of which correspond to the ns rotation frequencies ( e.g. , muno 2004 ) .
we searched for burst oscillations by computing and averaging the power spectra for the 3 bursts observed with 125 @xmath36s resolution event mode pca data using a sampling rate of 2048 hz ( nyquist frequency : 1024 hz ) . to improve statistics , we averaged power spectra for sequential 0.25 , 0.5 and 1 s sections of the light curve ( 4 , 2 , and 1 hz resolution ) , with total segment lengths varying from 3 to 30 s. we searched in the @xmath1 10 s rise , around the burst peak , and at different times during the burst decay until 65 s from the burst peak .
pca deadtime effects cause a @xmath1 1% decrease in the mean value of the poisson noise to 1.98 ( from 2.0 ) in the leahy normalization ( see leahy et al . 1983 ) .
we initially included the entire 260 kev energy band and made no significant detections at any point during the burst .
we then tried varying the energy ranges included in the pca light curves . with the 1030 kev data
, we made a detection 1530 s from the burst peak around 612 hz ( figure 2 ) .
the significance of the signal was maximized with 0.25 s segment light curves ; therefore , 180 separate power spectra were averaged .
the signal could be observed in each burst separately though with less significance .
we then tried expanding the energy range that was included in the pca light curves ; the signal weakened with the inclusion of either higher and lower energy photons .
the 101024 hz leahy - normalized power spectrum is presented in figure 2 , along with a power spectrum centered on the signal and a histogram of the difference in power from the mean .
the power at 612 hz deviates from the poisson level by 4.7@xmath37 , which with 256 bins means that there is a 0.033% chance that the detection is spurious . using a power spectrum with 1 hz resolution
, we fit a lorentzian function to the peak , giving a frequency @xmath38 611.2 hz with @xmath39 hz .
the average rms amplitude of the peak is 4.8% , which is consistent with the 5% rms amplitude of typical burst oscillations observed in lmxbs ( muno 2004 ) .
along with burst energetics and timing , we can learn much from gs 1826238 through its energy spectra .
we began our investigation by creating spectra for 1 ks intervals following the onset of a burst in order to see if there were significant changes in the best - fit parameters between bursts . since the average burst recurrence time for the five complete burst intervals in the 68 ks _
chandra _
observation was measured to be 12750 @xmath0 102 s , we were able to create 1 ks intervals up to 1112 ks after the burst . to improve statistics we stacked the datasets from all five _ chandra _ burst intervals .
we also constructed nearly two complete burst intervals using the simultaneously acquired _ rxte _ data .
although these data have time gaps , we determined the time since the previous burst for any individual _
_ dataset by comparing it to the _ chandra _ burst times . with the benefit of the simultaneous observations we obtained statistically significant measurements from 0.5200 kev for each 1 ks interval ( except 910 ks ) . to each dataset
, we fit a model composed of comptonized emission plus a broken power law .
this preliminary model was only used to examine whether or not there were significant changes to the spectrum with time , and ultimately , a different model was determined to be more physically plausible ( see below ) .
nevertheless , this exercise showed us that the best - fit parameters remain approximately constant after 1 ks , although with slightly elevated soft x - ray flux through 3 ks from the peak of the burst . in the following fits ,
we therefore define the persistent spectrum to be for @xmath40 ks . in the integrated persistent emission analysis
below , we found that the inclusion of data from 13 ks only minimally distorted most spectral parameters , however the inferred @xmath41 was underestimated by @xmath1 50% due to the excess of soft photons .
we are fortunate to have _ chandra _ data along with _ rxte _ since observations of bursters with low - earth orbit telescopes may miss some bursts if they occur during a gap in coverage when the source is occulted by the earth or the telescope is passing through the south atlantic anomaly .
( _ b _ ) provides an illustration of the potential ambiguity involved with the selection of persistent emission datasets . for example , it is clear that any persistent emission analysis would be slightly distorted if the first 3 ks of the second pca observation ( corresponding to the decay of the first burst observed with _ chandra _ ) were included . by summing all pha datasets that fall within the 312 ks interval
, we further constrain the statistics and explore standard models for lmxbs in addition to other models .
lmxbs are typically modeled with a blackbody or disk blackbody to represent the ns surface or the inner accretion disk , plus comptonized emission which can be modeled in xspec with comptt or a cut - off power law ( cutoffpl ) .
although these models have similar profiles , they are not necessarily interchangeable .
while comptonized emission can be empirically described with a cut - off power law , the model of titarchuk ( 1994 ) is derived from analytical equations that are founded on the real analytical theory of comptonization of soft photons , and the output parameters are physically explicit which allow direct interpretation . on the contrary , the index of a cut - off power law can not be interpreted in physical terms , while in comptt it is related to the combination of the optical depth and plasma temperature .
albeit , the cut - off energy of cutoffpl is related to the plasma temperature as @xmath42 . for these reasons we chose the titarchuk ( 1994 ) model to represent comptonized emission . with this description ,
the spectrum is governed entirely by the plasma temperature and the @xmath19 parameter , which characterizes the distribution law of the number of scatterings ( i.e. , @xmath43 is the probability that a seed photon undergoes @xmath44 scatterings before escaping the plasma ) .
the optical depth of the plasma is calculated from the @xmath19 parameter and depends on the input geometry : disk or spherical . generally , for a given @xmath19 parameter
, a larger optical depth is inferred from a spherical geometry than a disk geometry .
this is because the final spectral shape is determined by photons which undergo many more scatterings than the mean , and the longest dimension of a spherical plasma cloud is clearly shorter than that of a disk ( titarchuk 1994 ) .
we initially tried to fit the data by employing a single comptt component , but the resulting fit was unacceptable due to excesses above 70 kev and below 1.5 kev .
the addition of a blackbody component reduced the soft excess , and allowed the plasma temperature and thus the up - scattering efficiency to increase ( since @xmath45 , where @xmath46 is the comptonization parameter ) to try to match the hard excess .
although the fit was better ( @xmath47 ) , it was still unacceptable .
this is contrary to the results of in t zand et al .
( 1999a ) , who successfully used this model , although the 3.8 kev best - fit blackbody temperature is probably non - physical .
the best - fit blackbody temperature that we obtained with this model was @xmath1 0.5 kev , which seems more reasonable .
after trying various models in conjunction with comptt , we found the fit to be most improved by an additional comptt component , resulting in a model that can be interpreted as the comptonization of soft photons from two distinct emission regions .
one of the emission regions contributes 62% of the total luminosity , so we label it as the primary " emission region , and the other as the secondary " emission region .
the primary emission region is characterized by @xmath48 kev , and @xmath49 19.722.1 kev ( depending on the chosen geometry ) , while the secondary emission region has @xmath50 kev and @xmath49 6.16.4 kev .
the primary emission accounts for about 70% of the flux from 0.53 kev , and 80% of the flux from 20200 kev , while each comptonization component contributes nearly an equal amount of flux for 320 kev . based on the inferred @xmath19 parameter , which is not a direct input in xspec
, the primary emission region has @xmath51 if the region is disk - shaped , or @xmath52 if it is spherical .
the secondary emission region is characterized by @xmath53 and @xmath54 for disk and spherical geometries , respectively .
the physical interpretation of each comptonizing region must be inferred from the best - fit parameters and the corresponding flux contributions . for one , we expect the comptonizing region closer to the stellar surface to have hotter seed photons , which suggests that the secondary component is emitted nearer to the ns .
the secondary component may therefore result from boundary layer or surface emission .
conversely , the primary emission possibly stems from the accretion disk or adc , and a disk geometry implies @xmath51 . from the relative flux contributions , we see that this interpretation requires about 3/5 of the total flux to be emitted from the cooler disk region .
accretion theory predicts that half of the total gravitational binding energy per unit mass is released in the disk , and the other half in the boundary layer or on the stellar surface ( frank et al .
this is not inconsistent since the accretion disk and the boundary layer are not physically distinct ; rather , there may be a smooth transition from one to the other .
secondly , the optical depth of the secondary emission ( independent of the true geometry ) is large enough to naturally explain the lack of a visible blackbody component in the gs 1826238 spectrum . according to inogamov & sunyaev ( 1999 )
, the spreading of accretion flow from the equator to the poles leads to two bright rings of enhanced emission that are symmetric about the equator .
the latitude of these rings increases with accretion rate , and in the accretion regime of gs 1826238 , i.e. @xmath1 10@xmath55 g s@xmath13 ( inferred from the total flux ) , these rings lie @xmath1 0.51.5 km from the equator . therefore , assuming the blackbody emission comes from these rings , the secondary comptonized emission region only has to be large enough to cover the inner @xmath01.5 km above and below the equator . following in t zand et al .
( 1999b ) , an approximate effective radius for the spherical emission area of the wien seed photons is given by @xmath56 km , where @xmath57 is in kpc , the flux is measured in erg s@xmath13 @xmath12 , and the seed photons are measured in kev .
this approach yielded a soft photon emitting region with radius @xmath58 km , which is only consistent with the seed photons being generated at the boundary layer if the region is confined to a half - thickness of @xmath1 2.5 km , depending on the inclination of the system , and assuming that any emission from below " the accretion disk is not observable .
finally , since the flux of photons passing through the boundary layer relative to the disk is larger per unit area , we expect compton cooling to maintain a lower plasma temperature for the secondary comptonized emission region .
this is indeed what is observed ( see table 3 ) .
in addition to a dual comptonization model , we also modeled the spectrum with a single comptt plus a broken power law in order to give an estimate of its shape for possible non - thermal interpretation .
if the secondary emission is indeed non - thermal , perhaps it stems from an adc generated by magnetohydrodynamic turbulence , analogous to what is observed in the solar corona ( crosby et al .
alternatively , the emission may be synchrotron radiation from a relativistic jet escaping the system .
such jets have been found to be rather common among lmxbs and link these systems to active galactic nuclei ( agns ) ( e.g. , fender 2002 ) .
clearly , these possibilities are highly speculative . after obtaining a fit with the two component models , the residuals revealed a line feature around 6.5 kev , which prompted the addition of a gaussian to the models .
the best - fit value for the line was measured to be approximately @xmath59 kev , with a flux which corresponds to an equivalent width ( ew ) of @xmath1 37.2 ev . we interpret this feature as a neutral fe k@xmath4 line .
an f - test showed that the probability for an improvement to the fit occurring by chance is @xmath60 for the dual comptonization model , and @xmath61 for the comptonization plus broken power law model .
however , it should be noted that the use of an f - test to measure the significance of lines may not be valid ( protassov et al .
we present the best - fit parameters for these models in table 3 .
we also present the results of a blackbody plus cut - off power law model , which is described below . to account for the uncertainties in the relative instrumental flux calibrations
, we introduced a multiplicative constant into the spectral models . for each model ,
the normalizations of the _ chandra _ instruments relative to pca were @xmath62 , while for hexte the relative normalization was generally @xmath1 0.850.9 .
figure 3 shows the spectral fit for the dual comptonization model , and the residuals for all three models .
lccc n@xmath63 ( @xmath64 @xmath12 ) & 1.60@xmath65 & 1.50@xmath66 & 4.29@xmath67 + kt@xmath68/kt@xmath69 , kt@xmath70 ( kev ) & 0.42@xmath71/0.82@xmath71 & 0.68@xmath72/0.86@xmath73 & 1.34@xmath74 + r@xmath75 , r@xmath70 ( km ) & 4.02@xmath76 & 5.88@xmath77 & 2.38@xmath78 + kt@xmath68/kt@xmath79 ( kev ) & 20.79@xmath80/6.84@xmath81 & 5.85@xmath82/4.74@xmath83 & + @xmath84/@xmath85 & 2.56@xmath74/9.38@xmath74 & 4.69@xmath86/10.98@xmath87 & + @xmath88/@xmath89 & 1.07@xmath90/4.71@xmath91 & 1.01@xmath92/4.47@xmath93 & + @xmath94 & & 1.19@xmath95 & 1.26@xmath96 + @xmath97 & & 2.43@xmath98 & + e@xmath99 , e@xmath100 ( kev ) & & 35.41@xmath101 & 41.58@xmath102 + e@xmath103 ( kev ) & 6.44@xmath104 & 6.45@xmath105 & 6.37@xmath106 + ew@xmath103 ( ev ) & 37.2@xmath107 & 28.3@xmath108 & 15.2 + @xmath109 ( d.o.f . ) & 0.70 ( 521 ) & 0.75 ( 522 ) & 0.84 ( 527 ) + [ spectra ] to facilitate comparison to the results of previous work , we also fitted the spectrum of gs 1826238 with a blackbody plus cut - off power law model ( model 3 ) , and a blackbody plus power law from 0.510 kev ( not included in the table ) .
these models have been used by others to provide acceptable fits to the gs 1826238 persistent emission , but only the _ bepposax _ observations had similar energy coverage .
for example , the _ asca _ sis instrument used in the kong et al .
( 2000 ) analysis did not provide coverage above 10 kev , and the barret et al .
( 1995 ) _ rosat _ observation was restricted to energies below 2.2 kev . in order to compare to the kong et al .
( 2000 ) results , we used the same blackbody plus power law model and omitted all data above 10 kev . with this approach we obtained a fit with @xmath110 kev and @xmath111 , which are each within @xmath1 15% of their results . extending the energy range to 200 kev ,
however , revealed a higher blackbody temperature , a steeper photon index , and the power law had to be changed to a cut - off power law ( model 3 ) .
this shows that the use of _ rxte _ s extended energy range is crucial to minimize uncertainties in the soft component parameters due to less precise knowledge of the photon index or comptonized component when measured with the lower _ chandra _ energy range . as mentioned in 1 and 4.2 , in t zand et al .
( 1999a ) fit the 0.1200 kev persistent emission spectrum with a blackbody plus comptt , but they also modeled the spectrum with a blackbody plus a cut - off power law .
the best - fit blackbody temperature was @xmath112 kev , @xmath113 , @xmath114 kev , and @xmath115 @xmath12 , which slightly differs from @xmath116 kev , @xmath24 , @xmath117 kev , and @xmath118 @xmath12 of model 3 .
if we set the model 3 parameters to those obtained by in t zand et al .
( 1999a ) , however , we obtain a similar @xmath119 . by examining the residuals
, the slope of the model 3 appears to be too steep from @xmath1 2.55 kev and 1220 kev , and too shallow for @xmath1 512 kev .
a slight excess also remains for energies above about 60 kev ( see fig .
3 _ bottom _ ) .
the 2.4 km blackbody radius derived from this model is obviously smaller than expected for a ns , though the size of the radius may be explained if the blackbody emission is confined to the equatorial region . for a 10 km ns
, a blackbody radius of 2.4 km would correspond to emission from a strip of half - height @xmath1 0.6 km .
we note that since @xmath42 , the @xmath1 42 kev cut - off energy is consistent with the @xmath1 20 kev plasma temperature of the primary comptonized component of model 1 .
however , unlike the dual comptonization model and the comptonized emission plus broken power law model , the addition of a line feature at @xmath1 6.45 kev to model 3 did not significantly improve the fit , though we include it to preserve consistency .
we place an upper limit on the equivalent width at 15.2 ev with 90% confidence .
even though the blackbody plus cut - off power law fit that we obtained is acceptable , we are less confident with this model as compared to model 1 since the inferred @xmath41 is inconsistent with the value obtained through the galactic hydrogen survey of dickey & lockman ( 1990 ) , where the average column density in the direction of gs 1826238 was found to be @xmath1 1.9 @xmath64 @xmath12 .
moreover , such a high column density is inconsistent with the x - ray halo measurements presented by thompson et al .
( 2005 , in preparation ) .
briefly , for the x - ray halo analysis we created an exposure - corrected radial profile of the source and subtracted a normalized model point spread function to obtain the x - ray halo profile . using empirical relations from the dust scattering analysis of draine ( 2003 ) , we found that @xmath120 3% of the source flux is scattered into the x - ray halo .
this result was then used to obtain a scattering optical depth , which can be converted to an approximate hydrogen column density through a linear regression derived using observations of x - ray halos around 29 sources ( cf . predehl & schmitt 1995 , fig .
the hydrogen column density that we obtained in this fashion is consistent with the best - fit value of models 1 and 2 . by freezing @xmath41 in model 3 to the common lower value
, the model fits the data poorly .
we therefore conclude that the common lmxb model of a blackbody plus comptonized emission described by comptt _ or _ cutoffpl is not appropriate for the persistent emission spectrum of gs 1826238 .
the persistent emission spectrum was searched for interstellar absorption features .
this effort turned up a good candidate at about 17.15 .
the feature is broad ( 0.102 @xmath0 0.027 fwhm ) with an optical depth at line center of 0.63 @xmath0 0.09 , when fitted with a gaussian .
the formal significance is about 3.8@xmath37 ; based on about 800 bins searched at 0.03 binning , there is about a 6% chance that one would find such a feature due to random fluctuations . including a narrower feature at about 17.5 with a significance of 2.9@xmath37 , we modeled both features with structure in the fe l edge due to the interstellar medium ( ism ) .
the wavelengths of these features are a good match to those of fe l iii at 17.51 and fe l ii at 17.19 .
the match is not perfect .
figure 4 shows the near edge extended absorption fine structure ( nexafs ) of the fe l edge as measured by kortright and kim ( 2000 ) and the edge structure of butadiene iron tricarbonyl ( c@xmath121feh@xmath122o@xmath123 ) from the corex data base maintained by a. p. hitchcock .
neither model matches the data without a slight energy shift to match the l iii feature .
furthermore , the 17.15 line is somewhat broader than expected and may have an excess of absorption at the short wavelength side .
schulz et al . ( 2002 ) identified the fe l features in the spectrum of cyg x-1 and suspected that there is a mix of fe molecules , citing , in particular , a feature near 17.15 a that might be due to a more pure form of fe . without the corresponding l iii feature in our spectrum , however , we suspect that the extra broadening might result from statistical fluctuation .
the models use a cosmic abundance of fe ( @xmath124 relative to h ) and the observed ism column density of @xmath125 @xmath12 .
thus , these absorption features are consistent with fe in the ism .
we also studied the evolution of the spectrum throughout the burst and its decay by subdividing the first 1 ks from burst peak into 8 datasets with progressively larger integration times ( table 4 ) .
those data above 40 kev were ignored as too few counts were obtained for significance .
all of the datasets were fit with the xspec model phabs*edge(comptt@xmath68 + comptt@xmath75 + bbody ) ( d : disk , s : sphere ) , with the hydrogen column density frozen at the common value of @xmath126 @xmath12 .
we initially allowed the disk comptonization parameters to vary , but the best - fit values for each dataset , including the normalizations , changed by @xmath120 5% , so they were frozen at the persistent emission values . since the blackbody emission during a burst provides a large influx of seed photons in the inner boundary layer , the spherical seed photon temperature was fixed to the blackbody temperature , which is justified since most comptonized emission is emitted within @xmath1 20 km of the surface of the ns ( frank et al .
table [ bbpar ] shows the evolution of the spectral parameters .
cccccccc 010 & 1.76@xmath127 & 12.0@xmath128 & 11.6@xmath129 & 3.1@xmath130 & 9.9@xmath131 & 13.5 & 19.6 + 1030 & 1.72@xmath132 & 10.8@xmath133 & 11.5@xmath134 & 3.8@xmath135 & 7.8@xmath136 & 10.4 & 14.9 + 3065 & 1.50@xmath137 & 5.1@xmath138 & 10.4@xmath139 & 4.3@xmath140 & 8.0@xmath141 & 7.1 & 10.4 + 65100 & 1.24@xmath142 & 3.1@xmath143 & 11.9@xmath144 & 4.5@xmath145 & 10.2@xmath146 & 3.8 & 6.6 + 100150 & 1.01@xmath147 & 0.6@xmath148 & 7.9@xmath149 & 5.0@xmath150 & 11.8@xmath151 & 2.0 & 3.6 + 150200 & 0.86@xmath152 & & & 7.1@xmath153 & 11.2@xmath154 & 1.4 & 2.9 + 200500 & 0.83@xmath155 & & & 6.8@xmath156 & 8.7@xmath157 & 1.4 & 2.9 + 5001000 & 0.81@xmath158 & & & 6.9@xmath159 & 9.2@xmath160 & 1.3 & 2.8 + using the results from table 4 , we can place an upper limit to the source distance assuming the blackbody luminosity is eddington - limited , which it is not since the bursts are significantly sub - eddington . during the first 10 s following the peak of the burst
, the average flux from 0.540.0 kev was measured to be @xmath161 erg @xmath12 s@xmath13 .
we corrected the peak flux by assuming an approximately constant spectrum for this interval ( i.e. constant spectral parameters ) , and by measuring the burst decay scale time ( @xmath1 35 s over the energy band ) .
the _ peak _
flux was found to be @xmath1 12% larger than the 010 s mean , or @xmath162 erg @xmath12 s@xmath13 .
moreover , we must account for radiation that is absorbed or scattered out of the line - of - sight on its way to the telescope .
the hydrogen column density toward the source was assumed to be about @xmath163 @xmath12 , which corresponds to a scattering optical depth of 0.02 ( predehl & schmitt 1995 ) .
thus scattering and absorption along the line of sight reduce the observed flux by @xmath164 . if we assume that the blackbody normalization is also @xmath1 12% larger than the 010 s mean , and that the _ unabsorbed _ normalization is @xmath1 2% larger still , we arrive at a maximum source distance of 9.2 kpc .
we also assume a 1.4 m@xmath20 ns . ] for the bursts being at the eddington limit .
this upper limit on the distance is consistent with the 8 kpc upper limit derived by in t zand et al .
( 1999a ) .
the thermal nature of the burst spectrum with a blackbody temperature of @xmath1 1.8 kev , accompanied by the cooling , is typical for a type i x - ray burst ( lewin et al .
while the blackbody temperature deceases by a factor of about 2.2 during burst decay , the normalization decreases by more than a factor of 30 .
nevertheless , the derived blackbody radius remains approximately constant as long as blackbody emission is observable .
figure 5 shows the derived blackbody radius and the fraction of blackbody flux ( @xmath165 ) from 210 kev during the first 150 s after burst ignition .
clearly these data are consistent with a blackbody radius of @xmath1 10.311.7 km , assuming a source distance of 6 kpc and a ratio @xmath166 ( spectral hardening factor ) of 1.4 .
such a correction must be applied since nss do not radiate as perfect blackbodies during x - ray bursts because the thermalization of photons occurs at scattering optical depths greater than unity , where the temperature is higher than the effective temperature .
rather , the photons are thermalized at optical depths of @xmath1 45 ( e.g. , ebisuzaki et al .
1984 , london et al .
1986 , madej et al .
2004 ) . assuming that the entire stellar surface is involved when blackbody emission is observable , i.e. the burning does not cease in one area sooner than another , the slight decrease in the inferred blackbody radius for the 100150 s interval ( see fig .
5 ) could be explained by a @xmath1 10% increase in the spectral hardening factor .
this is a possibility since at lower effective temperatures ( @xmath167 kev ) the spectral hardening factor increases with decreasing temperature , as the relative contribution of electron scattering to the total opacity decreases ( london et al .
. however , within the error limits , the blackbody radius for this interval is still consistent with the measured values during the first 100 s of the burst .
[ bbr ] the derived blackbody radius can not be interpreted as the physical size of the ns since lmxbs do not have isotropic radiation fields . due to the presence of an optically thick accretion disk , any blackbody emission from the surface of the ns that is in the `` shadow '' of the accretion disk would not be directly observed but would possibly emerge as comptonized emission . the only geometrical arrangement for which we can naively interpret the blackbody radius as
the stellar radius is when the accretion disk is perpendicular to the line of sight ( @xmath168 ) . in all other cases
the inferred blackbody radii underestimate the stellar radius , and so one must account for the covered region in order to extract a stellar radius from a measured blackbody temperature and flux .
the area of the covered region increases nonlinearly with increasing binary inclination , moving from 0% of the projected area when @xmath169 to 50% when @xmath170 .
furthermore , since accretion disks are not infinitely thin there will also be some portion of the surface covered by the disk .
the amount of covering by the disk can be parametrized by the disk half - height @xmath171 . in spite of the nonlinear relation between the amount of covering and the inclination and disk half - height
, the fraction of the stellar surface that is covered can be calculated using straightforward geometrical principles . with these issues in mind
we interpreted @xmath172 as the observable _ projected _ area of the star , and by calculating the fraction of surface covered we derived an effective stellar radius ( @xmath173 ) as a function of inclination and disk half - height for source distances of 6 kpc and 5 kpc , which is presented in fig .
6 . using optical observations of the counterpart of gs 1826238 , mescheryakov et al . ( 2004 ) were able to derive an approximate inclination of @xmath174 @xmath175@xmath176 , and so we only include this range in the figure . [ bm ] from photometric measurements of the optical counterpart , barret et al .
( 1995 ) derived an approximate lower limit to the distance to gs 1826238 of 4 kpc . assuming the range of possible binary inclinations used above is correct , it is apparent that a source distance near 5 kpc gives a ns radius within the commonly assumed range of @xmath1 1012 km ( see the right axis of fig .
6 ) . on the contrary
, a distance of 6 kpc does not give a ns radius within the common range .
even with an infinitely thin accretion disk and an inclination at the lower bound ( @xmath177 ) of mescheryakov et al .
( 2004 ) , the smallest possible stellar radius at this distance within the error limits is @xmath1 11 km . from the results of the table 4 ,
it is clear that the burst photons immediately cool the comptonizing plasma to @xmath1 3.4 kev . over the next 150 s ,
the plasma temperature recovers the persistent emission value of @xmath1 6.8 kev .
this type of plasma cooling during a burst may be an example of compton cooling .
the same inverse compton scattering process that transfers energy to the persistent emission photons also transfers energy to the photons emitted during a burst .
the difference is that during the burst the flux of photons increases by more than an order of magnitude ( for the 010 s interval ) , so that the balance between heating and radiative cooling is disturbed .
it is also evident that the optical depth of the plasma is weakly constrained in the burst spectral analysis .
although the best - fit values for the first few intervals seem to indicate a larger optical depth , the range of uncertainty still includes the persistent emission value .
since blackbody emission is observable during the bursts but not during the quiescent phase , we can speculate as to the geometries of the corresponding emission regions .
to begin , gs 1826238 has a weak magnetic field so infalling matter probably impacts the ns along the equator rather than at the poles .
the release of gravitational binding energy through thermalization of the accreted material may result in a strip of blackbody emission along the equator ( church & bauciska - church 2001 ) , or the accumulation and spreading of accretion flow from the equator toward the poles may cause two bright emission rings that are symmetric about the equator ( inogamov & sunyaev 1999 ) . the inability to see
blackbody emission in the interval between bursts suggests that the equatorial strip or enhanced bright rings must be covered by an optically thick layer .
this is explained naturally by the persistent emission dual comptonization model .
at the onset of a burst , ignition likely starts near the equator and then rapidly spreads to cover the entire stellar surface in a couple seconds or less ( spitkovsky et al .
once the entire accumulated layer is burning , the subsequent blackbody emission indicates that there is an optical path for the radiation outside of the equatorial strip .
it could be argued that the accretion flow is disrupted during a burst , or that a spherical corona surrounding the ns is temporarily blown away , and that this provides a path for the blackbody emission .
however , the flux provided by the inner comptonizing region also increases substantially during a burst , so the boundary layer plasma must still be present to up - scatter the seed photons .
moreover , our measurement of the wien emission radius ( @xmath1 4 km ) is too small for the corona to completely surround the star .
instead , we assume that this emission region is likely confined to the equatorial region .
this explanation supports theories where the comptonizing plasma or adc is geometrically thin .
our investigation of the lmxb gs 1826238 using simultaneous _ rxte _ and _ chandra _ observations has led us to the following conclusions . from the observation of five uninterrupted burst intervals with _
chandra _ in july 2002 , we measured a burst recurrence time of 3.54 @xmath0 0.03 hr , which is consistent with g04 s measurement .
g04 s measured decrease in the burst recurrence time between 1997 and 2002 has been coupled with an even larger percentage increase in the mean persistent flux , possibly indicating that a greater fraction ( as compared to 1997 ) of the stellar surface may be covered prior to ignition .
we detected 611.2 hz burst oscillations with a 0.033% chance that the signal resulted from random fluctuations .
the average rms amplitude of the peak is 4.8% .
the _ rxte / chandra _ 0.5200 kev integrated persistent emission spectrum is best fit with a dual comptonization model , whereby two distinct comptonizing regions exist and are characterized by a different set of parameters .
the extended energy range of _ rxte _ is essential to constraining the photon index or comptonization parameter , and eliminates the need for a blackbody component .
this result is contrary to most spectral models of lmxbs that include a visible blackbody component .
the spectrum also requires a neutral fe k@xmath4 emission line at 6.45 kev with ew @xmath1 37.2 ev .
in addition , we find strong evidence of interstellar fe l absorption features at the about 17.15 and 17.5 , with significance 3.8@xmath37 and 2.9@xmath37 , respectively . during a burst ,
blackbody emission accounts for the majority of the flux though it quickly disappears after @xmath1 150 s. throughout this period , the data are consistent with a blackbody radius between 10.3 km and 11.7 km for a distance of 6 kpc if one assumes the blackbody flux comes from the full @xmath178 of the neutron star surface . by accounting for the fraction of the surface that is obscured by the accretion disk
, however , the source distance must be nearer to 5 kpc to be consistent with a neutron star radius of 1012 km .
we also see compton cooling during the bursts , as the plasma temperature immediately decreases to @xmath1 3 kev and then slowly returns to the persistent emission value of @xmath1 6.8 kev after about 150 s. since blackbody emission is not observed in the persistent spectrum yet dominates the burst spectrum , we conclude that the emission from those regions of the stellar surface along the equator are covered in the persistent phase . during the burst
the entire surface is radiating as a blackbody , and so this emission can be seen outside of the covered region .
we gratefully acknowledge support from nasa grants nas5 - 30702 and go2 - 3060x .
we thank tod strohmayer for identifying a mistake we initially made when creating power spectra in the search for burst oscillations , and lev titarchuk for useful information regarding his comptt model in xspec .
we also appreciate the input given by jrn wilms , and many comments and suggestions by the referee which significantly improved this paper .
this research made use of data the high energy astrophysics science archive research center online service , provided by the nasa / goddard space flight center .
the analysis of the _ chandra _ data made extensive use of the chandra interactive analysis of observations ( ciao ) , http://cxc.harvard.edu/ciao . | using simultaneous observations from the _ chandra x - ray observatory _ and the _ rossi x - ray timing explorer _ , we investigated the low - mass x - ray binary ( lmxb ) and `` clocked burster '' gs 1826238 with the goal of studying its spectral and timing properties . the uninterrupted
_ chandra _ observation captured 6 bursts ( _ rxte _ saw 3 of the 6 ) , yielding a recurrence time of 3.54 @xmath0 0.03 hr . using the proportional counter array on board _ rxte
_ , we made a probable detection of 611 hz burst oscillations in the decaying phases of the bursts with an average rms signal amplitude of 4.8% .
the integrated persistent emission spectrum can be described as the dual comptonization of @xmath1 0.3 kev soft photons by a plasma with @xmath2 kev and @xmath3 ( interpreted as emission from the accretion disk corona ) , plus the comptonization of hotter @xmath1 0.8 kev seed photons by a @xmath1 6.8 kev plasma ( interpreted as emission from or near the boundary layer ) .
we discovered evidence for a neutral fe k@xmath4 emission line , and we found interstellar fe l ii and fe l iii absorption features .
the burst spectrum can be fit by fixing the disk comptonization parameters to the persistent emission best - fit values , and adding a blackbody .
the temperature of the boundary layer seed photons was tied to the blackbody temperature .
the blackbody / seed photon temperature at the peak of the burst is @xmath1 1.8 kev and returns to @xmath1 0.8 kev over 200 s. the blackbody radius is consistent with @xmath5 10.311.7 km assuming a distance of 6 kpc , though this value can not be interpreted as the physical size of the neutron star due to partial covering of the stellar surface by the accretion disk . by accounting for the fraction of the surface that is obscured by the disk as a function of binary inclination
, we determined the source distance must actually be near 5 kpc in order for the stellar radius to lie within the commonly assumed range of 1012 km .
the order of magnitude increase in flux at burst peak is seen to cause compton cooling of the electron plasma surrounding the disk , as the plasma temperature decreases to @xmath1 3 kev at burst onset , and then slowly returns to the persistent emission value after about 150 s. |
the incredible experimental success of the renormalizable @xcite standard model @xcite , well beyond any theoretical expectation , puts severe constraints to any attempt to solve the hierarchy problem introducing new particles or states at the weak scale .
all experiments confirm the standard model .
it is true that all these experiments could have missed new light particles signals for some fortuitous coincidence or for some theoretical and still unknown reason , but we should not neglect the impressive convergence of all experiments towards the same conclusion : the spectrum and the properties of the low energy theory coincide with that of the renormalizable standard model .
we stress the word renormalizable , because it is ruled out not only the real production of such new light particles but also their virtual exchange , that could appear under the form of new effective non renormalizable operators .
limits to these operators can be found in the literature .
they arise from precision electroweak data , from fcnc tests , but also from proton stability and neutrino masses . here
we argue that this experimental scenario discloses a clue that can help us to guess how to extend the standard model . with the ( renormalizable ) standard model hamiltonian , we are able to compute the scattering matrix for processes that involve different type of @xmath0 and @xmath1 states .
all experimental tests prove the amazing accuracy of such predictions .
in particular , experiments rule out the possibility to modify the renormalizable standard model hamiltonian either adding new light particles or introducing new higher dimensional operators with a characteristic scale @xmath2 ( according to the operator choice , @xmath3 can range from the tev scale , up to the grand unification scale ) . on the other hand
the same hamiltonian is not able to give a realistic estimate of the effective potential , explaining the origin of the hierarchy between the weak scale and the unification scale .
the mass parameter in the higgs potential is unstable under radiative corrections and it depends quadratically on the cut - off @xmath2 , the scale of the new physics . theories that introduces this cut - off close to the weak scale directly to the hamiltonian , do not explain why present experiments have not yet found evidence for departures from the standard model .
theorists acknowledge the severe constraints to a relatively light new physics scale , and cope with them , assuming new symmetries that could reduce the impact of such light scale to the experimental tests , and getting used to the idea that a certain amount of fine tuning must be accepted .
this scenario becomes less puzzling , if we look closer at the physical quantities that we are comparing .
from one side we have very accurate predictions of the scattering matrix elements , between @xmath0 and @xmath1 states with is the higgs field . ]
the average @xmath4 very close to @xmath5 , the electroweak vev .
these matrix elements , and the time evolution of these states are very well described by the standard model .
on the other hand , we know that the same standard model hamiltonian is unable to evaluate the energy of states with a very different vev @xmath6 ( i.e. the sm does not give an acceptable effective potential )
. however this set of states is experimentally very different from the previous one , since a state with @xmath7 can not be reproduced in a laboratory . to our opinion
these considerations raise a doubt : can we extrapolate ordinary quantum field theory to include the time evolution of these states , with very different vev , and very far from our experience ? in the following we will try to see if we can give a theoretical basis to this question .
it is useful to exemplify the problem as follows : we show that a non linear schrdinger equation could explain an apparent paradox in a similar situation but in the simpler context of first quantization . to be concrete
we take an explicit mathematical example .
consider the following equation @xmath8 where @xmath9 is an analytic real function with @xmath10 , and @xmath9 going exponentially to zero when @xmath11 the solutions of this equation conserve the probability , and for any @xmath12 we have ( if @xmath13 is normalized , at @xmath14 ) @xmath15 @xmath16this is simple to check once we observe that @xmath17 , and its time derivatives , are real for any @xmath18 .
then @xmath19 where we have used ( [ nonlinear ] ) .
one can also verify that the energy @xmath20 with @xmath21 is conserved .
one can study the class of solutions of this equation that satisfy @xmath22 in this approximation the ( [ nonlinear ] ) is an ordinary schrdinger equation with the potential of an harmonic oscillator ( and with bounded states ) provided that we replace @xmath23 on the other hand , since @xmath9 exponentially goes to zero when @xmath24 , we can also have solutions that are similar to a free gaussian wave packet , centered in @xmath25 and going to infinity when @xmath18 goes to infinity .
this situation is similar to the puzzle that we have discussed in the previous section , where the experimental scenario apparently required two distinct hamiltonians with two different class of solutions : one with small vev and another one with very large vev .
clearly a unique linear equation can not describe the evolution of the physical system in both physical regions , and only a non linear equation can correctly describe the time evolution in all regions . in the following
we would like to see if this _ naive _ idea can be exploited to solve the hierarchy problem ( taking into account all necessary modifications required by the different physical context ) .
to start , we remind the formalism and the basics of the second quantization of a scalar field . the classical action for a free scalar field is @xmath26 from which we get the hamiltonian @xmath27 the rule of second quantization imposes replacing @xmath28 and @xmath29 with field operators @xmath30 and @xmath31 satisfying the equal time commutation relations below @xmath32=[\,\,\dot{\phi}(y),\,\hat{\phi}% ( x)\,]=i\text{% % tcimacro{\unicode{0x127}}% % beginexpansion h\hskip-.2em\llap{\protect\rule[1.1ex]{.325em}{.1ex}}\hskip.2em% % endexpansion } \delta ^{3}(x - y ) .
\label{commutation}\ ] ] we can give an explicit representation of the commutation ( [ commutation ] ) in the space of functionals$ ] ( with square brackets ) denotes a functional of the real function @xmath33 , while @xmath34 denotes an ordinary function @xmath35 of the real number @xmath28 . ]
@xmath36 $ ] with @xmath28 being a real function @xmath33 of the three dimensional space .
the product of two states @xmath37 can be ( formally ) defined in terms of a path integral ( i.e. a functional integration ) as follows .
the functional integration is defined as usual @xcite ; an infinite constant factor is always understood to get a well defined and finite result .
we will come back to this infinite constant after . ]
@xmath38s_{2}[\phi ] .
\label{product}\ ] ] with these definitions , it is easy to guess the action of @xmath39 and @xmath40 onto a state @xmath41 .
namely @xmath42 \label{replace0}\ ] ] and @xmath43 \label{replace}\ ] ] where @xmath44h-.2em.2em@xmath45 is a functional derivative .
one can check that @xmath46{.325em}{.1ex}}\hskip.2em% % endexpansion } \,\frac{\delta } { \delta \phi ( y)},\phi ( x)]=i\text{% % tcimacro{\unicode{0x127}}% % beginexpansion h\hskip-.2em\llap{\protect\rule[1.1ex]{.325em}{.1ex}}\hskip.2em% % endexpansion } \delta ^{3}(x - y ) \label{comm1}\ ] ] as required by the ( [ commutation ] ) . with the replacement ( [ replace0],[replace ] ) ,
the hamiltonian ( [ hami1 ] ) in this functional representation becomes @xmath47{.325em}{.1ex}}\hskip.2em% % endexpansion } ^{2}\frac{\delta ^{2}}{\delta \phi ( x)^{2}}-\phi \,\triangledown ^{2}\phi + m^{2}\phi ^{2})d^{3}x .
\label{hamifunc}\ ] ] this functional operator leads to the schrdinger functional equation @xmath48{.325em}{.1ex}}\hskip.2em% % endexpansion } ^{2}\frac{\delta ^{2}s[\phi ] } { \delta \phi ( x)^{2}}-(\phi \,\triangledown ^{2}\phi ) s[\phi ] + m^{2}\phi ^{2}s[\phi ] ) d^{3}x = e\,s[\phi ] \label{eq : scroe}\ ] ] where @xmath35 is a generic eigenfunctional of @xmath49 with eigenvalue @xmath50 .
the state with minimal @xmath50 can be written is divergent , but this divergence can be removed by a proper subtraction procedure . ]
@xmath51=n\exp ( -\frac{1}{2}\int d^{3}x\,\phi ( x)\sqrt{-\triangledown ^{2}+m^{2}}\phi ( x ) ) \label{vacuum}\ ] ] with a ( divergent ) normalization constant @xmath52 that sets @xmath53s_{0}[\phi ] = 1 .
\label{normalizzazione}\ ] ] the above integration @xmath54 is understood in the functional sense @xcite as in common path integrals .
one can recognizes that @xmath55 $ ] , as given in ( [ vacuum ] ) , is similar to the wave function of the common harmonic oscillator , apart from the obvious change of functions with functionals .
starting from the vacuum @xmath56we can build other states , for instance we can create a new state through the action of the operator @xmath30 at the position @xmath57 , @xmath58 and the corresponding functional becomes @xmath59=n\,\ \phi ( z)\exp ( -\int d^{3}x\,\phi ( x)\sqrt{-\triangledown ^{2}+m^{2}}\phi ( x ) ) \label{oneparticle}\ ] ] we can also compute the product @xmath60s_{y}[\phi ] = n^{2}\dint \mathcal{d}\phi \,\phi ( x)\phi ( y)\exp ( -\int d^{3}x\,\phi ( x)\sqrt{-\triangledown ^{2}+m^{2}}\phi ( x))\ ] ] this is easily evaluated once we realize that this is simply the definition of the two point green function of three dimensional ( euclidean ) scalar theory with inverse propagator @xmath61 this yields @xmath62 and this is the same result that one gets taking the scalar propagator at equal times .
it is important to note that the constant @xmath52 used in ( [ vacuum ] ) and in ( [ oneparticle ] ) is the same : thus the product ( [ product ] ) is well defined once we have chosen a vacuum functional ( [ vacuum ] ) , that sets the normalization @xmath52 ( [ normalizzazione ] ) , once for all .
we stress that until now we have not introduced any new physics or new physical concept . we have only introduced a rather unusual formalism to define second quantization .
our aim , here , is to put in better evidence the parallel between the well known schrdinger equation in first quantization and its ( less popular ) analogue in second quantization ( [ eq : scroe ] ) .
this will become useful after , because it will make more transparent some essential clues of the hierarchy problem that will lead us to naturally extend quantum field theory .
in what follows we will only introduce some basic concepts concerning the quantization of functional fields .
we do not have the intention to be comprehensive , and sometimes we will be not rigorous . even if the way to proceed is not unique , some important conclusions concerning the hierarchy problem can be achieved without entering into difficult theoretical details . in the discussion
below , we will often put in evidence the parallel with first and second quantization , this will help us to guess how some results of these well known theories could be generalized into some sort of third quantization .
namely a theory where the role of fields @xmath63 is now played by functionals @xmath36 $ ] .
note that eq .
( [ eq : scroe ] ) can come from the action @xmath64 below , where the integral @xmath54 is taken in the functional sense @xmath65\left ( i\frac{% \partial } { \partial t}+\frac{1}{2}\left ( -\text{% % tcimacro{\unicode{0x127}}% % beginexpansion h\hskip-.2em\llap{\protect\rule[1.1ex]{.325em}{.1ex}}\hskip.2em% % endexpansion } ^{2}\frac{\delta ^{2}}{\delta \phi ( x)^{2}}-(\phi \,\triangledown ^{2}\phi ) + m^{2}\phi ^{2}\right ) \right ) s[\phi , t]\,d^{3}x\,dt+\text{h.c . }
\label{fund}\ ] ] in fact , if we formally define a functional derivative , such that @xmath66}s[\phi ] = \prod_{x}\delta \left ( \phi ( x)-\phi ^{\prime } ( x)\right ) \equiv \delta ^{\infty } [ \phi -\phi ^{\prime } ] \label{delta1}\ ] ] and @xmath66}\int \mathcal{d}\phi \,s[\phi ] g[\phi ] = g[\phi ^{\prime } ] , \label{delta}\ ] ] the equation @xmath67=0 $ ] is equivalent to the ( [ eq : scroe ] ) .
let us try to quantize the action @xmath64 .
as usual the momentum conjugate of @xmath36 $ ] is obtained taking the derivative of @xmath68 with respect @xmath69 $ ] . thus @xmath36 $ ] and @xmath70 $ ] represent a couple of conjugate momenta , that for the rule of quantization must obey the equal time anti - commutation prescription ( if we consider fermion - like functional operators ) @xmath71,s[\phi ^{\prime } , t]\}=\prod_{i=1,n}\delta ^{k}\left ( \phi ( x_{i})-\phi ^{\prime } ( x_{i})\right ) \equiv
\delta ^{\infty } [ \phi -\phi ^{\prime } ] \label{anticom}\ ] ] as expected , the space - time dirac delta is replaced by a functional dirac delta .
the hamiltonian for these _ free _ _ functional _ fields @xmath35 is given by @xmath72\,h\,s[\phi ] \label{hami}\ ] ] where @xmath49 is a functional operator given by ( [ hamifunc ] ) , in the representation ( [ replace0],[replace ] ) .
we can also define the operator @xmath73\,\ s[\phi ] \ ] ] that commutes with the hamiltonian @xmath74 . from the commutation relation @xmath75\right ] = s[\phi ] $ ] , we recognize that @xmath76 and @xmath35 are creation and annihilation _ functional _ operators .
the eigenstates of the hamiltonian ( [ hami ] ) can be found , following the same arguments used in second quantization , and exploiting the commutativity of the hamiltonian with the number of fields operator .@xmath77 .
namely , we start from the vacuum @xmath78 , defined by @xmath79 from which we get @xmath80 the vacuum is an eigenstate of the hamiltonian with zero energy
. then we proceed building all the other states
. we can apply iteratively and several times the creation _ functional _ operator .
for example , the one _ functional _ creation operator @xmath76 can creates one field states with a generic wave functional @xmath81 , as follows @xmath82s^{\dagger } [ \phi ] |0 > \label{stato1}\ ] ] with @xmath83f_{n}[\phi ] = 1\ ] ] the low index 1 in the label @xmath84 , means that the state ( [ stato1 ] ) is an eigenstate of @xmath77 , with eigenvalue 1 .
it is easy to verify that @xmath85 implies the schrdinger equation ( [ eq : scroe ] ) ( with @xmath35 replaced by @xmath81 ) . for certain choices of the hamiltonian @xmath49 , the state with minimum energy could be of the type ( [ stato1 ] ) , and we can label it @xmath86 .
states with two functional fields @xmath87 would lead to two identical ( and decoupled)equations for two different functionals @xmath88 $ ] and @xmath89 $ ] : they would describe two parallel worlds that do not talk to each other , as long as we restrict to the free includes interaction in the common second quantization sense , but it is free in the third quantization sense .
this will become more clear in the next section , when we will add a third quantization interaction . ]
hamiltonian ( [ hami ] ) . .
a ( higgs ) field that in second quantization is represented by a scalar field operator @xmath30 ( see section 2 ) , in the context of a third quantization can be replaced by a proper combination of the functional operators @xmath90 $ ] and @xmath36 $ ] .
namely , @xmath91\phi ( x)s[\phi ] \label{phi } \\
\hat{\pi}(x ) & \rightarrow & \hat{\pi}(x)\equiv i\int \mathcal{d}\phi \,s^{\dagger } [ \phi ] \,\ \frac{\delta \,s[\phi ] } { \delta \phi ( x)}.\end{aligned}\ ] ] one can verify that the commutation relations ( [ comm1 ] ) are satisfied ( using ( [ anticom ] ) ) .
the operators @xmath92 and @xmath93 are common field operators ( i.e. not functional operators ) , and until now the ( [ phi ] ) can be considered an alternative ) ( bilinear in the functional operators @xmath94 , and with physical observables defined by @xmath95 and @xmath96 we have a theory that is completely equivalent to a quantum field theory .
departures from quantum field theory will become manifest when we will add to ( [ hami ] ) an interaction with the insertion of several functional operators @xmath35 . ] representation of the algebra ( [ comm1 ] ) . in the heisenberg picture the functional field operator @xmath36 $ ] becomes time dependent , and satisfies the quantum mechanical equation @xmath97}{d\,t}=\left [ \mathcal{h},s[\phi , t]\right]\ ] ] from which we get @xmath98 = \hat{\pi}(x , t ) \label{erenfest2 } \\
\frac{d^{2}\,\hat{\phi}(x , t)}{d\,t^{2 } } & = & -i\,\ \frac{d\,\hat{\pi}(x , t)}{% d\,t}=-\left [ \mathcal{h},\hat{\pi}(x , t)\right ] = ( \triangledown ^{2}-m^{2})% \hat{\phi}(x , t ) \label{erenfest3}\end{aligned}\ ] ] in particular we can find that the ordinary two point green function of the scalar field , applying the definition ( [ phi ] ) , @xmath99 is the time ordered product .
@xmath86 is the state with minimum energy , with @xmath100 , and @xmath101 is equal to ( [ vacuum ] ) , namely @xmath102s^{\dagger } [ \phi ] |0 > \label{vuoto}\ ] ] then applying twice ( [ erenfest2 ] ) , we can check the equation @xmath103 this confirms that the state @xmath86,in third quantization , describes a physical system that is equivalent to the one described by a scalar quantum field theory with green functions given by the rules of second quantization .
the concept of green functions can be generalized to third quantization , provided that we replace functions with functionals .
for example , in complete analogy with the second quantization , we can define a two - point green functional @xmath104=<0|t\left\ { s[\phi ^{\prime } , t^{\prime } ] s^{\dagger } [ \phi , t]\right\ } |0 > \label{greenfunc}\ ] ] one can verify that it corresponds to the inverse functional operator appearing in ( [ fund ] ) , once we define a proper time ordering prescription to get rid of the pole singularities ( the @xmath105 feynman prescription ) .
we have seen in the first section that the hierarchy puzzle could be solved if we add a non linear term to equation ( [ eq : scroe ] ) .
one could introduce it by hand modifying ( [ eq : scroe ] ) , as in a non - linear schrdinger equation , and studying the phenomenological consequences .
however here we prefer to stick to quantum theory .
in fact , we notice that adding a non linear interaction reminds us the embedding of first quantization into second quantization , _
i.e. _ the possibility of interaction changing the number of particles , that in the second quantization language corresponds to operators involving more than two fields . in other words
the hierarchy puzzle can be seen as an hint to move forward with some sort of third quantization beyond the second one . unfortunately , until now , we have not many clues to guess which type of interaction to add .
we have not yet any principle as powerful as that of gauge theories leading to unambiguous forms for the interactions .
nevertheless , note that ( [ anticom ] ) implies that the mass dimension of the functional field @xmath36 $ ] is not well defined since the functional delta has the same dimensions of the product of an infinite number of dirac delta .
but we also know that @xmath106,s[\phi ^{\prime } , t]\}=\dint \mathcal{d}\phi \,\
\,\delta ^{\infty } [ \phi -\phi ^{\prime } ] = 1,\ ] ] thus the dimension of the integral @xmath107 is the inverse of a pair of @xmath108 if we want to add a pair of functional fields @xmath109 we are obliged to add an integral @xmath107 , if we want to keep the action with the right mass dimensions . for instance @xmath110s[\phi ] = \mathcal{n}$ ] is a dimensionless operator . in the following , we restrict ourselves to hamiltonians that commutes with @xmath77 , thus the number of fields is conserved .
the phenomenological consequences of additional insertions to the hamiltonian of different powers of @xmath111 , @xmath112 , etc .
are rather trivial .
instead , let us consider the operator can carry internal indices of some internal symmetry .
this justifies why we have inserted the square @xmath113 ( to build an invariant operator ) , instead of @xmath28 .
also @xmath35 can carry an index of an internal symmetry , but to simplify the notation we remove this index . ]
@xmath114\phi ^{2}(x)s[\phi ] $ ] .
@xmath115 commutes with @xmath77 , it has well defined mass dimensions , and can be used to build hamiltonians with much more interesting phenomenology . as example , we will consider the hamiltonian ( [ hami ] ) with an additional interaction of the form @xmath116 where @xmath117 is an arbitrary function , defined by the expansion above with coefficients @xmath118 .
this interaction is local in the space - time dimensions .
it is understood that some renormalization prescription for the @xmath119 are considered ; for our purpose we can also look at the ( [ inter ] ) as an effective hamiltonian , arising from a more fundamental theory involving different functional operators in addition to @xmath35 .
the insertion of ( [ inter ] ) , leads to a new action @xmath64 , given by @xmath120\,\left ( i% \frac{\partial } { \partial t}+h\right ) \,s[\phi , t]+\,\int dt\,d^{3}xv\left ( \,\dint \mathcal{d}\phi ^{\prime } s^{\dagger } [ \phi ^{\prime } , t]\,\phi ^{\prime 2}(x)\,s[\phi ^{\prime } , t]\right ) \label{azion}\ ] ] the green functional ( [ greenfunc ] ) at the zeroth order approximation , as anticipated in the previous section , is given by @xmath121=i\left ( i\frac{\partial } { % \partial t}+h+i\,\varepsilon \right ) ^{-1}[\phi , t;\,\phi ^{\prime } , t^{\prime } ] \label{zeroorder}\ ] ] however the exact two point green functional is affected by the interaction @xmath122 that can change the vacuum of the theory ( in the following we have in mind an analogous example in second quantization , the nambu - jona - lasinio model ) . naively ,
taking the derivative @xmath123 , we deduce the equation @xmath124=\
s[\phi , t]+v^{\prime } \left ( \left\langle 0|\mathcal{o}|0\right\rangle \,\right ) \int d^{3}x\text { } \phi ^{2}(x)\,\,s[\phi , t ] \label{meanfield}\ ] ] where , in the mean field approximation , we have replaced the interacting term with its vacuum expectation value @xmath125 . in this approximation
, we see that we still have a linear schrdinger equation ( [ meanfield ] ) , as in second quantization ( [ eq : scroe ] ) .
the only effect of the interaction @xmath117 , is to modify the bare mass of the field @xmath28 .
in fact , a mass @xmath126 adds directly to the hamiltonian @xmath49 , in ( [ meanfield ] ) , and thus modifies also the green functional ( [ zeroorder ] ) .
apart from this renormalization of the bare mass , the physics described by the approximate equation ( [ meanfield ] ) , is identical to an ordinary quantum field theory .
if we wish to compute the exact green functional , we have to compute all loop contributions induced by the interaction @xmath117 .
these can be formally taken into account in a non perturbative approach for composite operators ( see chapter 8 in @xcite ) : the exact green functions arise from the functional minimization of an effective action @xmath127 $ ] , a functional of the green function @xmath128 .
@xmath129 is the sum of all ( two - particle irreducible ) vacuum loop diagrams .
we will not repeat here the arguments leading to the equations below , instead we assume a straightforward generalization of the discussion in @xcite .
we replace green functions with green functionals , and integrations over space coordinates with functional integrations over the function @xmath33 . as a result ,
the effective action for a composite functional operator @xmath130\,s[\phi ^{\prime } , t^{\prime } ] $ ] is a functional @xmath127 $ ] of the green functional ( [ greenfunc ] ) , that can be written ] ( compare with eq .
8.47 of @xcite ) @xmath131=-itr(\log ( g^{-1}))-itr(g_{0}^{-1}g)+\gamma _ { 2}[g ] \label{eff}\ ] ] where @xmath132 is given by all two - loop ( and higher ) two - particle irreducible vacuum graphs
. 0.6 cm in the following ( see fig .
1 ) , we will consider only vacuum loop diagrams , at first order in @xmath117 , i.e. with only one space - time point ( represented in figure 1 , by the big grey point ) .
@xmath117 contains several powers of the operator @xmath133 ; each @xmath134 contains a functional integration with respect a distinct function @xmath28 .
this explains why we have also depicted few small dark points , each one representing a distinct functional point @xmath135 etc .
for example , the top diagram on the left comes from the interaction term @xmath136 . @xmath136 two functional integrations with respect
the two functions @xmath137 ( depicted by the two distinct black points ) . in the same diagram
there are two lines ( i.e. two green functionals or propagators ) , attached at the same black point .
that is to say each green functional is evaluate at the same functional point ( @xmath138 $ ] ) . from the ( [ eff ] ) , and reminding the definition ( [ delta1 ] ) of functional derivative , we can write the equation @xmath139}{\delta \,g}=ig^{-1}-ig_{0}^{-1}+\frac{\delta \gamma _ { 2}[g]}{\delta \,g}. \label{equazione}\ ] ] this yields the exact functional @xmath128 that we want to compute .
we can distinguish two type of contributions , when we take the derivative of the functional , @xmath140 .
the first is obtained when the derivative @xmath141 $ ] acts on the green functional @xmath142 $ ] , computed at the same function @xmath143 , ( and same time @xmath18 ) .
for example , those coming from the diagrams on the left of figure 1 . after the integration over @xmath143 ,
these yield a term proportional to a functional dirac delta @xmath144 $ ] .
a second one , arises when the derivative acts on a green functional @xmath145 $ ] with different functions @xmath146 and @xmath147 ( but same time @xmath18 ) : the full derivative takes the form @xmath148}{\delta \,g[\phi , t;\phi ^{\prime } , t^{\prime } ] } % = a(v^{2})\,\,\,\,\,\,\delta ( t - t^{\prime } ) \int d^{3}x\,\phi ^{2}[x]\,\delta ^{\infty } ( \phi -\phi ^{\prime } ) + \\ \\ + b(v^{2})\,\,\,\ \ \delta ( t - t^{\prime } ) \int d^{3}x\,\phi ^{2}[x]\,\,g[\phi , t;\,\phi ^{\prime } , t]\phi ^{\prime 2}[x]+ ... \end{array}\ ] ] where the dots @xmath149 stand for all other possible combinations of the green functionals @xmath128 arising from the same interaction @xmath117 . for brevity , we do not list them all , since they are unnecessary for the discussion below .
the functions @xmath150 and @xmath151 have appeared as a result of all integrations of the functionals @xmath128 over @xmath152 .
their explicit dependance on the vacuum expectation value @xmath153 , can be computed only if a specific interaction @xmath117 is chosen , and after the evaluation of all functional integrals . among these , we note for instance @xmath154\,\,\,\phi \lbrack x]^{2}\equiv \left\langle 0\left| \hat{\phi}^{2}(x)\right| 0\right\rangle = v^{2 } \label{vev}\ ] ] in this step we have assumed ) . the choice ( [ bound])corresponds to a particular @xmath155 prescription for the pole singularities in ( [ zeroorder ] ) .
these prescriptions set the boundary conditions for @xmath128 at @xmath156 . ] the boundary condition at @xmath157 @xmath158=\,\ f_{0}(\phi ) f_{0}^{\ast } ( \phi ^{\prime } ) .
\label{bound}\ ] ] where @xmath101 is of the form ( [ vacuum ] ) . with this prescription ,
we get a vacuum with one ( and only one ) higgs field ( see eq .
( [ vuoto ] ) and eq .
( [ greenfunc ] ) ) . as in ordinary quantum field theory , where the evaluation of the composite operator @xmath159 needs a suitable renormalization prescription
, we understand that suitable counterterms have been added to ensure a finite result in the integration ( [ vev ] ) .
note that @xmath5 is the true vev of the theory\rightarrow g[\phi , t;\,\phi ^{\prime } , t^{\prime } ] $ ] when @xmath160 and @xmath161 then the integral @xmath162 in eq.([vev ] ) yields a constant @xmath5 independent of @xmath163 ; i.e. all radiative corrections are taken into account , since @xmath128 is the full green functional of the theory@xmath164 the full equation ( [ equazione ] ) for the green functional @xmath165 $ ] can be written @xmath166^{-1}=i\frac{\partial } { \partial t}% + h+i\,\varepsilon + a[v^{2}]\,\,\,\,\,\,\delta ( t - t^{\prime } ) \int d^{3}x\,\phi ^{2}[x]\,\delta ^{\infty } ( \phi -\phi ^{\prime } ) + \\ \\ + b\left [ v^{2}\right ] \,\,\,\ \ \delta ( t - t^{\prime } ) \int d^{3}x\,\phi ^{2}[x]\,\,g[\phi , t;\,\phi ^{\prime } , t]\phi ^{\prime 2}[x]+ ... \end{array } \label{miransky}\ ] ] we can distinguish two parts in the right - hand side : the one proportional to @xmath167\,\ \delta ^{\infty } ( \phi -\phi ^{\prime } ) $ ] is similar to a mass term in @xmath49 ( see eq .
( [ eq : scroe ] ) ) and can be reabsorbed into a redefinition of the bare mass of the field @xmath168)\,\phi ^{2}$ ] .
one can recognize that in the mean field approximation ( eq .
( [ meanfield ] ) ) @xmath169=v^{\prime } [ v^{2}]$ ] ( and @xmath170 ) . the last term is not proportional to @xmath171 ; for the moment , as a first approximation , we assume @xmath172 $ ] to be very small and negligeable .
the equations ( [ miransky ] ) and ( [ vev ] ) are a system of coupled equations , from which we would like to determine @xmath5 and @xmath128 .
let us forget for a moment ( [ vev ] ) .
we look for solutions @xmath128 of ( [ miransky ] ) , treating @xmath5 as a free parameter .
we recognize that ( [ miransky ] ) , ( with @xmath173 ) is the same equation that we would get in absence of the interaction @xmath174 ( see the section before ) , apart from a redefinition of the bare mass @xmath175)\,\phi ^{2}. \label{bare}\ ] ] this implies that we can compute @xmath128 following the more familiar procedure of second quantization .
we remember that @xmath128 is defined in ( [ greenfunc ] ) .
in particular we can compute the vacuum ( [ vev ] ) of this modified hamiltonian as follows .
we first calculate the full effective potential of an ordinary quantum field theory with hamiltonian @xmath49 , but with modified bare masses ( [ bare ] ) .
the minimization of this non perturbative effective potential would give @xmath176 @xmath177 comes from a function @xmath178 of the bare parameters @xmath179 at the large scale @xmath2 .
the function @xmath180 can be computed in the case of a perturbative quantum field theory ; exploiting the renormalization group equations ( rge ) one usually finds @xmath5 as a simple function of the common higgs potential parameters @xmath181 and @xmath182 . in any case if @xmath183 and the other mass parameters at the high energy scale are of the order of the large scale @xmath2 also @xmath5 from eq .
( [ minimo ] ) is of order @xmath2 .
this is the origin of the hierarchy problem : if any of the mass parameters of the theory entering the @xmath178 is large , then also @xmath5 is large . only with a fine tuning of those large masses
@xmath184 etc .
one can get a small vev @xmath185 now let us see how this paradox can be solved in the context of third quantization . in this case
@xmath184 , the bare mass at the scale @xmath2 , is not a free parameter but it is itself a function of the vev @xmath186 through equation ( [ vev ] ) and ( [ bare ] ) .
using these equations the ( [ minimo ] ) becomes @xmath187),g_{1}^{2}, ...
\right ) \label{minimo2}\ ] ] it is understood that the above equation is exact , i.e. it includes all quantum corrections : the function @xmath178 includes the exact ( all loops ) dependance from the parameters @xmath188 given at the very high energy .
it is clear that the equation ( [ minimo2 ] ) can have solutions with @xmath189 : even if the function @xmath190 is rather simple ( e.g. polynomial ) .
mathematical examples could be easily built if @xmath191 contains logarithms or exponentials of the vev @xmath5 .
the hierarchy arises exclusively from the interactions at the third quantization level , that in the ( [ minimo2 ] ) are represented by the function @xmath191 .
this interplay between second and third quantization is parallel to a similar connection between first and second quantization .
an interesting example , is provided by the hydrogen atom . in this specific physical system ,
one finds the minimum of an hamiltonian ( in first quantization ) , with the fine structure constant @xmath192 , that can be calculated in second quantization .
namely , firstly we solve the schrdinger equation ( first quantization ) , and we find the electron wave function @xmath193 .
@xmath193 contains the characteristic scale @xmath194 of the virtual momentum exchanged between the proton and the electron .
this @xmath194 is clearly a function of the input parameter @xmath192 . but
this input parameter @xmath192 is in turn a function of @xmath194 , since the fine structure constant is obtained ( in second quantization ) from the @xmath195 function and the rge , evaluated at the characteristic scale @xmath196 the interplay between first and second quantization can be summarized as follows : @xmath194 is obtained from @xmath197 applying the common procedure of first quantization . in turn
@xmath198 is a function of the characteristic scale @xmath194 , and can be computed using second quantization and the rge equations .
this connection leads to a non linear equation that is absolutely equivalent to ( [ minimo2 ] ) : in fact second quantization ( and the minimization of the full effective potential)is used to derive the characteristic scale @xmath5 as a function @xmath199 of the bare parameters @xmath200 while third quantization ( through the interaction @xmath117 ) is necessary to calculate @xmath201 ( the bare mass ) as a function of @xmath5 .
this leads to ( [ minimo2 ] ) .
even if the phenomenological consequences of a third quantization need an appropriate study of well defined and realistic models , we can anticipate some generic possible consequences implied by this scenario .
these could lead to new and exotic experimental signatures .
one of these can be inferred by the additional terms proportional to @xmath202 , that appeared in ( [ miransky ] ) ( and that we have neglected until now ) .
if we consider non trivial vacua , then the functional @xmath128 solving ( [ miransky ] ) could be different from that one of an ordinary quantum field theory .
let us consider the modification to the solution @xmath128 , induced by the additional term @xmath202 . if @xmath203 indicates the green functional at @xmath170 , then eq.([miransky ] ) gives us the first order perturbative correction @xmath204=g_{0}^{-1}[\phi , t;\phi ^{\prime } , t^{\prime } ] + b(v^{2})\ \delta ( t - t^{\prime } ) \int d^{3}x\phi ^{2}(x)\phi ^{\prime 2}(x)g_{0}[\phi , t;\phi ^{\prime } , t ] .
\label{greenuno}\ ] ] plugging the ( [ bound ] ) in ( [ greenuno ] ) , we get the green functional @xmath205 describing the time evolution of the quantum state ( the wave functional ) from the initial time @xmath206 to the final time @xmath18 . to the ordinary hamiltonian @xmath49
, we have to add an operator of the form @xmath207 the insertion of the vacuum projector @xmath208 is rather unusual
. however note , that if @xmath209 is the renormalizable hamiltonian of the standard model , it includes the standard model symmetries , like the baryon and the lepton number . thus the vacuum @xmath210 ( [ vacuum ] ) , ( obtained from @xmath211 , eq . ( [ zeroorder ] ) ) is invariant under these symmetries .
thus also the new interaction ( [ newint ] ) looks symmetric .
all accidental symmetries of the renormalizable standard model are preserved . at first glance
, we can also anticipate few possible experimental signatures .
let us write down the green function involved in the scattering of two higgses into two higgses @xmath212 . for simplicity
we consider the case where @xmath49 in ( [ zeroorder ] ) contains only the kinetic term , thus the only interaction in this process , comes from @xmath213 ( [ newint ] ) , and it appears only at the first order in @xmath151 .
the four point green function is @xmath214 with the @xmath215 in the far future and @xmath216 in the far past . at the first non trivial order
we get @xmath217 where @xmath218 is inserted in the middle , given that @xmath219 and we have to respect the time ordering prescription .
@xmath220 is mediated by @xmath213 as if it were an ordinary @xmath221 operator . on the other hand , the same effective @xmath222 interaction ,
added to the original interaction , would also modify a process like @xmath223 ; but this process is not affected by @xmath213 , as it can be easily checked , because we know that @xmath224 . in other words , in such a scenario ,
an anomalous self - interaction of the higgs boson could be proven by the existence of two discrepant precision measurements of the same effective coupling constant but in two different physical processes .
the unexpected experimental success of the renormalizable standard model puts severe constraints to any attempt to solve the hierarchy problem by adding new particles in the low energy spectrum of the theory .
their real ( or virtual ) production is ruled out , at least in the region around the electroweak scale .
we have seen that the main difficulty is due to the theoretical assumption that the time evolution operator of quantum states with very different vacuum expectation values of @xmath28 are described by the same linear operator , the hamiltonian h. while the standard model hamiltonian , seems to be very accurate to describe particle physics phenomenology ( in experimental tests concerning only quantum states very close to the electroweak vacuum ) , it seems to be unable to estimate the energy of vacua with very different expectation values of the field operator @xmath28 .
this paradox becomes more affordable if we accept the possibility that the time evolution of the quantum states is described by a non - linear schrdinger equation .
motivated by this example , we have explored how such a modification can be achieved still in the context of quantum mechanics , but with an embedding of second quantization into some sort of third quantization .
we already know that first quantization arises as an approximation of second quantization : namely , if we neglect the interaction between particles and we consider only two - point green functions .
in fact , two - point green functions satisfy the schrdinger equation of first quantization .
similarly , we can imagine that second quantization ( i.e. quantum field theory ) arises as an approximation of third quantization , where the role of points and fields is now played , respectively , by fields and functional fields . as for first quantization
, second quantization is obtained in the limit when only the two - point green functional is relevant . in this limit , the hamiltonian of quantum field theory is simply the inverse of the full two - point green functional of the third quantization theory .
thus , to derive the hamiltonian of second quantization from the full third quantization theory , we have used the formalism of the effective action of composite operators .
we have found that the interaction of third quantization @xmath191 , can significantly change the equation ( [ minimo2 ] ) without adding any new particles in the low energy spectrum .
this minimal impact in the low energy spectrum is , to our opinion , in better agreement with the experimental evidence . even if we have not yet shown that the idea suggested in this paper can lead to a realistic and well defined theory , we have anticipated few possible experimental signature , from a simple and naive analysis of our mathematical example above .
namely , a new type of operators could appear in the effective hamiltonian of second quantization . beside the common self interaction of the operators @xmath225
, one could also have some exotic operators of the form @xmath226 with the insertion of the vacuum projector .
this interaction would lead to some anomalous ( and non- linear ) interactions in scattering processes involving the higgs particle .
99 g.t hooft and m. veltman , _ nucl .
* b44 * ( 1972)189 .
s. weinberg , _ phys .
* 19 * , 1264 ( 1967 ) .
a. salam , in _ elementary particle physics _
n. svartholm ( almqvist and wiksells , stockholm,1968):p.367 ; s.l .
glashow , j. iliopoulos , and l. maiani , _ phys .
* d2*,1285 ( 1970 ) .
c.itzykson and j.zuber , _ quantum field theory _ , mcgraw - hill book co. , newyork , 1980 .
v. a. miransky , _ dinamical symmetry breaking in quantum field theories _ , world scientific publishing , singapore , 1993 . | while in first and second quantization the fundamental operators are respectively coordinates and fields ( functions ) , an extension of quantum field theory can be achieved if the usual pair of conjugate momenta is represented by functionals . after a brief introduction on the hierarchy problem , we show how the ordinary quantum field theory can arise from a specific limit of this extension .
we will also show how this extension can offer new solutions to the hierarchy problem .
the peculiarity that makes this scenario appealing ( as possible solution of the hierarchy problem ) , is the absence of new light particles at ( or close to ) the electroweak scale .
this is in much better agreement with the experimental observation , since ( until now ) all searches for new physics signals have confirmed the remarkable success of the standard model ( contrary to common expectations ) . |
the study of dusty debris disks around main sequence stars can provide clues to the possible presence of orbiting bodies such as planets , the gravitational effects of which have predictable effects on the orbital dynamics of dust particles ( see , for example , @xcite ) .
one system of considerable interest is fomalhaut ( @xmath5 psa ) , of which submillimeter - wave mapping at 450 and 850 @xmath0 m has shown the presence of structure dominated by a pair of intensity maxima interpreted as the ansae of an inclined ring @xcite .
observations at 24 , 70 , and 160 @xmath0 m with _ spitzer _ @xcite show that the relative intensities of the ansae are wavelength - dependent @xcite .
the spatial variation of dust column density in the disks provides information about the perturbing bodies when interpreted using dynamical models .
however , since the observed ( intensity ) images are sensitive to variations in dust temperature as well as column density , we clearly need a tool to separate these effects . with this in mind , we have developed an inversion procedure ( diskfit ) designed to estimate the 2-dimensional spatial distribution of optical depth ( or relative column density ) of a circumstellar disk in either the sky plane or disk plane , given a set of observed images at multiple wavelengths , the corresponding point source response functions ( psfs ) , and certain model assumptions .
we have applied this procedure to our new 350 @xmath0 m image of fomalhaut in conjunction with _ spitzer
fomalhaut was observed at 350 @xmath0 m with the sharc ii camera @xcite at the nasmyth focus of the caltech submillimeter observatory ( cso ) on 2004 sep 18 ut . of a total period of 4.7 hr on source , three hour - long integrations between 8.5 and 12.3 h ut provided the best signal - to - noise due to low atmospheric opacity ( 0.039 at 225 ghz , 1.05 at 350 @xmath0 m ; both at zenith ) , remarkable atmospheric stability , and relatively high source elevation ( @xmath6 to @xmath7 ) .
our final map therefore consists of a coaddition of the images from that period , with a total integration time of 3.0 hr .
the telescope executed a lissajous scanning pattern with peak - to - peak amplitude @xmath8 and period 20 s in the azimuth direction , and peak - to - peak amplitude @xmath9 and period 14.14 s in the elevation direction .
of the 12 rows of the detector array , only rows 18 were operational for this run ; at a plate scale of 457 @xmath10 , the instantaneous field of view was then @xmath11 .
the bolometer signals were sampled at 28 hz and then reduced with the crush software program ( a. kovacs , in preparationsharc / crush / index.htm ] ) .
we used the -deep " option which takes a more aggressive approach at modeling atmospheric and instrumental effects , but found the results to be relatively insensitive to this choice , perhaps due to the atmospheric stability .
automatic correction was made for the atmospheric opacity and increased gain of the detectors under reduced atmospheric loading , the latter being a @xmath12% effect implemented using a lookup table .
absolute calibration was accomplished with hourly , interspersed observations of uranus and neptune .
uranus ( assumed flux 237 jy @xmath13 , 250 jy total ) was observed at elevations of @xmath14 and @xmath15 , and neptune ( assumed flux 92 jy @xmath13 , 94 jy total ) was observed at elevations of @xmath16 and @xmath17 .
the resultant four scale factors from instrument units to jy @xmath13 had a relative rms of 7% .
we estimate an absolute flux calibration accuracy of 20% and pointing accuracy of @xmath18 .
the psf ( fwhm @xmath19 ) was obtained from observations of neptune , appropriately rotated and coadded so as to recreate accurately the rotational smearing due to the changing parallactic angle during the fomalhaut observations .
the upper panel of figure [ fig1 ] shows the resulting image of fomalhaut at 350 @xmath0 m .
the integrated flux density is @xmath20 jy . in order to enhance the spatial resolution of the image
, we have deconvolved it using the hires implementation @xcite of the maximum correlation method @xcite .
the result is shown in the lower panel of figure [ fig1 ] , the spatial resolution of which is approximately @xmath21 .
it is apparent that the source is strikingly symmetrical , suggestive of a simple ring .
this is , in fact , the first observation which clearly shows the ring - like morphology .
a key question , though , is whether the dust ring is truly uniform or has density variations .
we have addressed this issue by combining our 350 @xmath0 m image with other data using our diskfit inversion procedure .
the principal model assumptions involved in diskfit are : 1 .
the wavelength variation of opacity , @xmath22 , is the same throughout the disk ; we have assumed a power - law variation of the form @xmath23 .
the radial variation of dust temperature is given by the energy balance of individual grains @xcite , based on a grain size - scale parameter , @xmath24 , which represents the wavelength above which the grains radiate inefficiently .
the vertical scale height of the disk follows a power - law variation with respect to radial distance , @xmath25 . in terms of the opening angle , @xmath26 ,
corresponding to 1 scale height above and below midplane , we have assumed @xmath27}/100)^\gamma$ ] , where @xmath28 is the opening angle at @xmath29 au .
there are two main steps in the inversion procedure : * step 1 : parameter estimation . * in this initial step , we obtain estimates of some preliminary parameters necessary for the optical depth mapping .
these parameters consist of @xmath30 , @xmath28 , and @xmath31 ( defined above ) , the disk inclination , @xmath32 , with respect to the sky plane , and the position angle , @xmath33 , of the tilt axis .
we base these estimates on a simple model involving a circular disk of inner and outer radii @xmath34 and @xmath35 , respectively , with a power - law radial dependence of optical depth ( normal to disk plane ) , represented by @xmath36 .
allowance is made for the possibility that the center of the dust ring may be displaced from the star , as an approximate way of taking into account the effects of orbital ellipticity .
a maximum likelihood estimate is then made of the full set of parameters characterizing this model , using the set of observed images and psfs , taking account of the measurement noise in each observed image , and making allowance for the presence of small errors in the position calibration of each observed image .
* step 2 : inversion for optical depth map . * based on the estimated values of @xmath37 , and @xmath30 from the previous step , the full inversion for the 2-dimensional distribution of line - of - sight optical depth ( referred to as a tau map " ) in the plane of the sky is accomplished using an algorithm which is similar mathematically to the richardson - lucy procedure @xcite .
since the psfs at each wavelength are implicitly deconvolved using the prior knowledge of positivity of optical depth , some superresolution is obtained .
the output is a map of the line - of - sight optical depth at an arbitrary reference wavelength ( equivalent to relative column density ) and the associated uncertainty map . as a final step
, we can transform this sky projection into a pole - on view using a linear estimation procedure which takes into account the flared geometry . the input data for the parameter - fit step consisted of our 350 @xmath0 m image and _ spitzer _ images at 160 and 70 @xmath0 m from the multiband imaging photometer for _ spitzer _
( mips)see @xcite .
although a mips image at 24 @xmath0 m was also available , we did not include it in this step since it contains a significant response to the presence of an additional warm dust component @xcite which was not included in the simple ring model . both large ( blackbody ) and intermediate - sized ( @xmath38 @xmath0 m ; @xcite ) grains were considered for the purpose of the dust temperature calculation . although the results were not strongly dependent on the assumed grain properties , the best fit was obtained under the blackbody assumption , consistent with the presence of large grains as inferred by @xcite .
the complete set of estimated parameter values is given in table [ table1 ] , with error bars representing the effect of random measurement noise ; @xmath39 and @xmath40 represent the estimated offsets of the center of the dust ring with respect to the star , in a right - handed coordinate system in which the @xmath41-axis corresponds to the disk tilt axis ( positive = nnw ) .
we have verified that the results are unaffected by the possible effects of imperfect removal of an instrumental near - ir leak in the 160 @xmath0 m image by repeating the fit excluding the 160 @xmath0 m data ; there was no significant difference in the estimated parameters .
figure [ fig2 ] shows the residual images from the uniform - ring model fit as compared to the observed images ; each plot represents a slice along the major axis of the tilted ring .
the residuals are more or less consistent with the estimated background noise and correspond to a @xmath42 .
the fact that the latter value is somewhat larger than unity suggests the possibility of a slight deviation from a uniform ring .
the next step ( tau map estimation ) involved reprocessing the raw images , making use of the above estimates of @xmath37 , and @xmath30 .
the input data were the observed images at all four available wavelengths ( 350 , 160 , 70 , and 24 @xmath0 m ) ; the resulting tau map in the sky - plane projection is shown in the left hand panel of figure [ fig3 ] . in this figure ,
the line - of - sight optical depth variation is expressed at a reference wavelength of 24 @xmath0 m ; the spatial resolution is approximately 22 au .
the peak values of optical depth in the ring are @xmath43 and @xmath44 at the sse and nnw ansae , respectively .
the density - weighted dust temperature in the ring is 42 k. since the uncertainty of the estimated optical depth increases with radial distance , the map has been truncated at 310 au , at which the uncertainty is twice that at ring maximum .
it is apparent that the sse peak on the tau map is slightly closer to the star than the nnw peak this is consistent with the result obtained independently from the parametrized model fit . the inferred displacement between the star and disk center results from the need to preferentially heat the sse portion of the ring in order to reproduce the observed 30% intensity enhancement of the sse ansa at 70 @xmath0 m ( see figure [ fig2 ] and the discussion in @xcite ) .
therefore , the estimation accuracy of the displacement ( @xmath45 au ) is dependent on these intensity ratios rather than the absolute pointing of the telescope .
since the tau map embodies more observational data than the model fit , we can refine our estimates of the other ring parameters ; in particular we obtain 180 au for @xmath35 .
we have deprojected the tau map to produce a pole - on view of the disk , but due to the large inclination , the resulting errors were a factor of three higher than for the sky projection .
the dynamic range was consequently insufficient for any further elucidation of the density structure and so we have not included the result here .
our 350 @xmath0 m image is qualitatively different from the 450 @xmath0 m image published by @xcite in that it has the morphology of a symmetrical ring as opposed to the horseshoe shape at 450 @xmath0 m ( ansae connected by a single arc ) .
the latter was interpreted in terms of a dust clump " suggestive of material trapped in the mean motion resonance of a large planet .
since our 350 @xmath0 m image is of similar signal to noise ratio and resolution , we should have detected such a feature ( at the @xmath46 level in this case ) .
more detailed examination shows that the arclike bridge at 450 @xmath0 m has about the same relative intensity with respect to the peak ( 0.5 ) as that of either arm of the ring at 350 @xmath0 m , whereas the western arm at 450 @xmath0 m is substantially weaker ( @xmath47 ) .
thus rather than the eastern arm at 450 @xmath0 m being an enhancement , the western arm is actually a depression .
this argues against the presence of a dust clump in the former location .
our estimated disk parameters agree essentially with previous values @xcite , including the opacity index ( @xmath48 ) which is consistent with relatively large ( @xmath49 @xmath0 m ) grains .
one possible discrepancy is that our opening angle of @xmath50 implies a smaller vertical disk thickness ( @xmath51 au ) than previous estimates ( @xmath52 and 120 au at 450 and 850 @xmath0 m , respectively@xcite ) .
however , the estimated value is very sensitive to psf width , suggesting that at least some of the differences may be resolution - related .
our optical depth mapping indicated an @xmath53 au displacement of the ring center from the star .
such a displacement was suggested by @xcite as a consequence of gravitational perturbation by orbiting bodies , observable by the resulting pericenter glow ; the latter was invoked by @xcite to explain the brightness asymmetry of the ansae .
our estimated displacement corresponds to a forced eccentricity of approximately 0.06 , consistent with the findings of stapelfeldt et al . in this scenario
, one would expect an excess of material at apocenter due to the slower orbital motion there . in order to see whether our data show such an effect , we have simulated the tau map using the results of dynamical modeling for a forced eccentricity of 0.06 , and the result is shown in the right hand panel of figure [ fig3 ] .
although the model tau map shows two equal peaks ( in agreement with the estimated tau map ) , the nnw ansa ( apocenter ) is more radially extended than the ssw ansa since the orbits of individual grains are more dispersed there , and hence the integrated column density at apocenter is greater .
a comparison of integrated column density in the two halves of the ring ( apocenter and pericenter ) , as inferred from the model and estimated tau maps ( truncated at a dust temperature of 35 k ) , indicates a predicted 12% apocentric excess from the model , as compared to a value of @xmath54% estimated from the observations .
the agreement between these values supports the interpretation in terms of forced eccentricity .
if the orbit of the perturbing body ( presumably a planet ) is within the inner boundary of the ring , our models indicate that a forced eccentricity of 0.06 would require an orbital eccentricity of @xmath55}$ ] for the planet , over a wide range of the planet s semi - major axis , @xmath56 .
for example , if the inner edge of the ring is associated with the planet s 2:3 mean motion resonance ( analagous to neptune and the kuiper belt in our own solar system ) , then @xmath57 au and @xmath58 . a constraint on the planet s
mass can be derived from the timescale required to produce the observed perturbation ; the estimated age of fomalhaut ( @xmath59 myr ) then implies a value of @xmath60 earth mass .
an additional asymmetry is apparent in the estimated tau map of figure [ fig3 ] , in that the ring has a slight teardrop " shape , being more pointed at pericenter .
this effect is apparent in the 350 @xmath0 m data alone ( see figure [ fig1 ] ) and also in a tau map made independently using only the 24 , 70 , & 160 @xmath0 m _ spitzer
_ images ( not reproduced here ) .
a possible explanation is warping of the disk , whereby the sse portion of the ring has greater inclination to the line of sight than the nnw portion . to maintain such
a warp would require at least two planets , since perturbations by a single planet would result simply in the disk adjusting to that planet s orbital plane .
stronger constraints on the properties of the orbiting body ( or bodies ) could be obtained from a more detailed knowledge of the azimuthal structure of column density around the ring , which would require a more accurate pole - on view than we were able to obtain with these data .
such information could , in principle , be obtained by supplementing the present data using additional observations including , for example , the sed mode of mips , and submillimeter mapping at higher signal to noise ratio .
we thank dr .
matt bradford for his assistance during the observations .
this work was performed by the jet propulsion laboratory , california institute of technology , under contract with the national aeronautics and space administration .
research at the caltech submillimeter observatory is supported by nsf grant ast-0229008 .
aumann , h. h. , fowler , j. w. & melnyk , m. 1990 , aj , 99 , 1674 backman , d. e. & paresce , f. 1993 , in protostars and planets iii ( tucson : univ .
arizona press ) , 1253 backus , c. , velusamy , t. , thompson , t.j .
, arballo , j. , 2005 , in adass xiv , eds .
shopbell , m.c .
britton , & r. ebert , aps conf .
( in press ) chini , r. , krgel & kreysa , e. 1990 , a&a , 227 , l5 dent , w. r. f. , walker , h. j. , holland , w. s. , & greaves , j. s. 2000 , mnras , 314 , 702 dowell , c. d. , et al .
2003 , in millimeter and submillimeter detectors for astronomy " , ed . t. g. phillips & j. zmuidzinas , proc .
spie , 4855 , 73 holland , w. s. et al .
2003 , apj , 582 , 1141 lucy , l. b. 1974 , aj , 79 , 745 quillen , a. c. & thorndike , s. 2002 , apj , 578 , l149 richardson , w. h. 1972 , j. opt .
am . , 62 , 55 rieke , g. h. et al .
2004 , apjs , 154 , 25 stapelfeldt , k. r. et al .
2004 , apjs , 154 , 458 werner , m. w. et al .
2004 , apjs , 154 , 1 wilner , d. j. , holman , m. j. , kuchner , m. j. & ho , p. t. p. 2002
, apj , 569 , l115 wyatt , m. c. et al . 1999 ,
apj , 527 , 918 ccl @xmath32 & @xmath61 & inclination with respect to the plane of the sky + @xmath33 & @xmath62 & position angle of tilt axis + @xmath34 & @xmath63 au & inner radius + @xmath35 & @xmath64 au & outer radius + @xmath28 & @xmath65 & opening angle for the flared - disk geometry + @xmath31 & @xmath66 & power - law index for radial variation of @xmath26 + @xmath67 & @xmath68 & optical depth ( normal to disk plane ) at @xmath69 ; @xmath70 m + @xmath5 & @xmath71 & power - law index for radial variation of column density + @xmath30 & @xmath72 & power - law index for wavelength variation of opacity + @xmath73 & @xmath74 & offsets of dust ring from star in au ( see text ) + | we have imaged the circumstellar disk of fomalhaut at 350 @xmath0 m wavelength , using sharc ii at the caltech submillimeter observatory .
the spatial resolution of the raw images ( @xmath1 ) has been enhanced by a factor of three using the hires deconvolution procedure .
we find that at this wavelength and signal to noise ratio ( @xmath2 ) , the observed morphology is that of a simple inclined ring ( @xmath3 ) , with little or no other apparent structure this is the first observation that shows clearly the ring morphology of the disk .
we have combined our 350 @xmath0 m data with _
spitzer space telescope _
images at 24 , 70 , and 160 @xmath0 m in order to estimate the 2-dimensional spatial variation of relative column density ( tau map " ) using our diskfit procedure .
the tau map is based on the following physical assumptions : ( 1 ) the wavelength variation of opacity is the same throughout the disk , ( 2 ) the radial variation of dust temperature is dictated by the energy balance of individual grains in the stellar radiation field , and ( 3 ) the vertical scale height of the disk follows a power - law radial variation .
the results confirm the ring - like morphology , but also show that the geometric center is displaced from the star by about 8 au and that the ring has an apocentric enhancement of approximately 14% in integrated column density .
if we interpret the displacement in terms of elliptical orbital motion due to gravitational perturbation by an unseen planet , then the implied forced eccentricity is @xmath4 ; dynamical modeling then predicts an apocentric density enhancement consistent with that inferred from the tau map . |
a graph @xmath9 is an ordered pair @xmath10 consisting of a set @xmath11 of vertices and a set @xmath12 of edges .
when the context is clear , @xmath11 and @xmath12 are simply written as @xmath13 and @xmath14 , respectively .
the _ open neighborhood _ of vertex @xmath15 is the set @xmath16 .
the _ closed neighborhood _ of vertex @xmath15 is the set @xmath17 = n(v)\cup \{v\}$ ] . for a set @xmath18 of vertices , @xmath19=\bigcup_{v\in s } n[v]$ ] .
a set @xmath20 is a _ dominating set _ of @xmath9 if @xmath19=v$ ] @xcite .
the _ domination number _ of @xmath9 , denoted by @xmath21 , is the cardinality of a minimum dominating set . for two natural numbers @xmath22 and @xmath23 with @xmath24 and @xmath25 , the _ generalized petersen graph _
@xmath26 is a graph on @xmath27 vertices with @xmath28 and @xmath29 with subscripts modulo @xmath22 @xcite .
hereafter , all operations on the subscripts of vertices are taken modulo @xmath22 unless stated otherwise . in @xcite ,
behzad , behzad , and praeger showed that @xmath30 for odd @xmath24 and conjectured that @xmath31 is exactly the domination number of @xmath2 . in @xcite ,
ebrahimi , jahanbakht , and mahmoodian ( independently , yan , kang , and xu @xcite and fu , yang , and jiang @xcite ) affirmed that @xmath32 for @xmath24 . in this paper
, we are concerned with @xmath3 when there is a faulty vertex @xmath4 in @xmath2 , where @xmath3 denotes the domination number of @xmath2 with faulty vertex @xmath4 , i.e. , @xmath4 is removed from @xmath2 .
thus a faulty vertex can not be chosen as a vertex in the dominating set .
we shall show that , for @xmath33 , @xmath34 the _ alteration domination number _ of @xmath9 , denoted by @xmath35 , is the minimum number of points whose removal increases or decreases the domination number of @xmath9 @xcite .
the _ bondage number _ of @xmath9 , denoted by @xmath36 , is the minimum number of edges whose removal from @xmath9 results in a graph with larger domination number @xcite .
it can be regarded as the fault tolerance problem when removing vertices or edges from a graph .
fault tolerance is also an important issue on engineering @xcite .
this motivates us to study the the domination number of @xmath2 with a faulty vertex . by our result
, we can find that the lower and upper bounds of @xmath37 are as follows : @xmath38 if @xmath6 or @xmath7 ; otherwise , @xmath39 .
moreover , we can find that @xmath40 if @xmath41 , @xmath42 or @xmath43 .
this paper is organized as follows . in section [ preliminaries ]
we review some preliminaries of dominating sets in generalized petersen graphs . in section [ faulty petersen ] some properties
are introduced when there is a faulty vertex in @xmath2 .
section [ main results ] contains our main results .
we conclude in section [ conclusion ] .
in this paper , we follow the terminology of @xcite . however ,
for clarity , we introduce some of them as follows .
let @xmath44 denote the resulting graph after @xmath45 is removed from @xmath2 . in particular , @xmath46 is denoted by @xmath47 where @xmath4 , for some @xmath48 , is the faulty vertex in @xmath2 .
we also use @xmath49 and @xmath50 to denote adding an element @xmath45 to a set @xmath18 and removing an element @xmath45 from a set @xmath18 , respectively . in the rest of this paper
, @xmath18 always stands for a domination set of @xmath47 .
a minimum dominating set of @xmath9 is called a _
@xmath21-set_. when the graph @xmath9 is clear from the context , @xmath21-set is written as @xmath51-set .
block _ of @xmath0 is an induced subgraph of 5 consecutive pairs of vertices ( see figure [ fig1 ] ) .
denote by @xmath52 if a block of @xmath0 is centered at @xmath53 and @xmath54 .
when there is no possible ambiguity , @xmath52 and @xmath55 are used interchangeably .
the vertices of @xmath56 can be partitioned into @xmath57 , @xmath58 , and @xmath59 .
let @xmath60\setminus \mathcal{b}_i=\{v_{i+3},u_{i+3},v_{i+4}\}$ ] , @xmath61\setminus \mathcal{b}_i=\{v_{i-3},u_{i-3},v_{i-4}\}$ ] , and @xmath62 .
when the context is clear , @xmath63 is written as @xmath64 .
let @xmath65 which contains the indices in @xmath66 .
( 360,120)(0,0 ) ( 70,-10 ) ( 160,103)@xmath52 [ pre : gammapn2 ] for @xmath67 , @xmath68 .
[ coro : gammapn2 ] for @xmath67 , @xmath69 .
let @xmath18 be a @xmath1-set . if @xmath70 , then @xmath18 is also a dominating set of @xmath47 and @xmath71 .
for the case where @xmath72 , since @xmath73 , by symmetry , we can relabel the subscripts of the vertices in @xmath2 but not @xmath18 such that @xmath70 .
thus , in this case , @xmath74 . to prove that @xmath75 , suppose to the contrary that there exists a dominating set @xmath18 of @xmath76 with @xmath77 .
it is clear that @xmath78 is also a dominating set of @xmath2 whose cardinality is @xmath79 , a contradiction .
this completes the proof .
[ p1 ] let @xmath18 be a minimum dominating set in @xmath47 and assume that the vertex @xmath4 in the corresponding graph @xmath2 has at least one neighbor in @xmath18
. then @xmath80 . since @xmath81
, @xmath18 is also a dominating set of @xmath2 .
this implies that @xmath82 . by theorem [ pre : gammapn2 ]
, there exists a @xmath83-set , say @xmath84 , with @xmath85 .
clearly , @xmath84 is also a dominating set of @xmath76 .
thus @xmath86 .
this further implies that @xmath87 . by theorem [ pre : gammapn2 ]
, the lemma follows .
in this section , we introduce some properties of @xmath66 in @xmath2 , where @xmath4 is a faulty vertex . by lemma [ p1 ]
, it remains to consider the case where @xmath88 .
thus , in the rest of this paper , we assume that @xmath18 is a minimum dominating set under the condition that @xmath88 unless stated otherwise .
thus , in this case , @xmath89 which implies @xmath90 . hereafter , when we say that a vertex @xmath91 is dominated with respect to @xmath18 , then @xmath91 is either in @xmath18 or @xmath91 is adjacent to some vertex in @xmath18 .
[ pp : exchange ] assume that @xmath18 is a dominating set of graph @xmath9 and @xmath92 , where @xmath93 and @xmath94 .
if all vertices in @xmath95 $ ] are dominated by @xmath96 , then @xmath96 is also a dominating set of @xmath9 with @xmath97 .
[ lmm : p_i=3 ] if there exists @xmath98 and @xmath99 for some @xmath100 , then @xmath101 . since @xmath102=\mathcal{b}_i$ ] , the vertices in @xmath103 can only be dominated by some vertices in @xmath52 .
note that any two vertices in @xmath103 have no neighbor in common except @xmath53 .
however , @xmath98 and @xmath99 .
this results in @xmath104\cap s|\geqslant 3 $ ] . thus @xmath101 and the lemma follows .
[ lmm : p_j>1 ] assume that there exists a minimum dominating set @xmath18 such that @xmath105 .
then there exists a minimum dominating set @xmath18 such that @xmath106 and @xmath107 for all @xmath108 . if @xmath109 for @xmath100 in @xmath18 , then we are done .
thus we assume that there exists @xmath110 for some @xmath111 . by lemma
[ lmm : p_i=3 ] , @xmath112 ; otherwise , @xmath113 .
we may assume that none of @xmath114 and @xmath115 is @xmath4 ; otherwise , reverse @xmath52 so that @xmath116 and @xmath117 are interchanged .
this further implies that all vertices in @xmath117 must be dominated by some vertices in @xmath118 .
since any two vertices in @xmath117 have no common neighbor in @xmath118 , all vertices in @xmath118 must be in @xmath18 so that the vertices in @xmath117 can be dominated . by proposition [ pp :
exchange ] , the set @xmath119 is also a minimum dominating set under the condition that @xmath88 since the vertices in @xmath120 $ ] are still dominated by the vertices in @xmath96 . note that @xmath121 for @xmath122 except @xmath123 .
it is easy to verify that @xmath124 and @xmath125 .
moreover , @xmath96 has one less elements in @xmath126 than that of @xmath18 . by applying the above process repeatedly until the set @xmath126 becomes empty , this results in a minimum dominating set with @xmath127 for all @xmath100
this completes the proof .
[ type i set ] a minimum dominating set @xmath18 is called a type i set if @xmath127 for @xmath100 . for a type i set @xmath18 , the cardinality of the set @xmath128
is called its couple number .
[ pp : l_iandr_i ] assume that @xmath18 is a type i set .
if @xmath129 and @xmath130 , then either @xmath131 or @xmath132 .
since @xmath129 and @xmath130 , @xmath133 . to show that either @xmath131 or @xmath132 , it is equivalent to showing that @xmath134 or @xmath135 is impossible .
suppose to the contrary that @xmath136 ( or @xmath137 ) .
it can be found that vertex @xmath138 ( or @xmath139 ) is not dominated , a contradiction . by proposition
[ pp : l_iandr_i ] , in the rest of this paper , we only consider the case where @xmath131 and @xmath140 when @xmath18 is a type i set with @xmath141 . for the case
where @xmath142 and @xmath132 , we can reverse the generalized petersen graph so that it yields @xmath131 and @xmath143 . in total
, there are nine possible cases for the vertices in @xmath144 when @xmath88 , @xmath141 , and @xmath131 . however , only four of them are feasible ( see figures [ figb](a)-(d ) ) . for example , if @xmath131 and @xmath145 , then @xmath138 is not dominated and it is an infeasible case . for the case in figure [ figb](b ) , by proposition [ pp : exchange ] ,
@xmath146 is still a type i set .
note that the pattern of @xmath147 is exactly the case in figure [ figb](a ) . for the case in figure [ figb](c ) , by proposition [ pp :
exchange ] , @xmath148 is also a type i set .
furthermore , the pattern of @xmath149 is also exactly the case in figure [ figb](a ) . for the case in figure [ figb](d ) , by proposition [ pp : exchange ] , @xmath150 is also a type i set while @xmath151 .
we shall define a type iii set later for this case .
henceforth , if @xmath18 is a type i set with @xmath130 and @xmath129 , then we may assume that it is a type ii set which is defined as follows . [ type ii set ] a type i set @xmath18 with @xmath141 is called a type ii set if @xmath152 @xmath153see figure @xmath154 .
now we consider the case where @xmath88 and @xmath155 . by symmetry , we only need to consider the cases where @xmath156 and @xmath142 .
there are only seven feasible combinations ( see figures [ fig8](a)-[fig8](g ) ) .
every minimum dominating set @xmath18 in the cases of figures [ fig8](e)-[fig8](g ) can be transformed to a type ii set .
that is , set @xmath157 in figure [ fig8](e ) , set @xmath158 in figure [ fig8](f ) , and set @xmath159 in figure [ fig8](g ) .
note that @xmath160 , n[v_{f-2}]$ ] , and @xmath161 $ ] are still dominated by the vertices in @xmath162 , and @xmath163 , respectively .
thus , by proposition [ pp : exchange ] , @xmath162 , and @xmath163 are type ii sets .
[ type ii set ] a type i set @xmath18 with @xmath151 is called a type iii set if @xmath144 is equal to one of the following four sets : @xmath164 , @xmath165 , @xmath166 , and @xmath167 , @xmath153see figures @xmath168-@xmath169 , or precisely , are called _ type iii@xmath170-iii@xmath171 sets _ , respectively . henceforth , if @xmath18 is a type i set with @xmath172 and @xmath173 , then we may assume that it is a type iii set .
[ lmm : gamma=2 ] assume that @xmath18 is a type ii @xmath153or iii@xmath174 set with couple number @xmath175 .
if there exists @xmath176 for @xmath177 and @xmath178 , then there exists a type ii @xmath153or iii@xmath174 set with couple number @xmath179 . clearly , by lemma [ lmm : p_i=3 ] , @xmath112 . since @xmath180 ,
the other vertex of @xmath52 in @xmath18 , say @xmath181 , must be in @xmath182 .
we only consider the case where @xmath183 ( see figure [ fig2 ] ) .
the other case is similar .
since @xmath184 , by using a similar argument as in lemma [ lmm : p_j>1 ] , @xmath185 . clearly , at least one vertex in @xmath186 must be in @xmath18 so that @xmath187 and @xmath188 are dominated .
this results in @xmath189 no matter whether @xmath190 or @xmath191 is equal to @xmath192 or not . now let @xmath119 .
it is easy to check that @xmath121 for @xmath122 except @xmath193 .
however , @xmath194 and @xmath195 .
thus the couple number of @xmath96 is one less than that of @xmath18 .
this completes the proof .
an illustration for lemma [ lmm : gamma=2 ] . ]
note that if @xmath196 @xmath153respectively , @xmath197 in lemma [ lmm : gamma=2 ] , then @xmath198 @xmath153respectively , @xmath199 .
this yields @xmath200 @xmath153respectively , @xmath201 which is removed under our assumption .
thus we can not obtain a type ii @xmath153or iii@xmath174 set @xmath96 by setting @xmath119 @xmath153respectively , @xmath202 .
for the case where @xmath203 , @xmath53 even might not be in @xmath18 since @xmath4 is either @xmath204 or @xmath205 which is removed .
hereafter , we assume that @xmath18 is with the smallest couple number if @xmath18 is a type ii or iii set .
[ pseudogamma2 ] let @xmath18 be a type ii @xmath153or iii@xmath174 set with the smallest couple number .
a vertex @xmath53 for @xmath177 is called a pseudo - couple vertex with respect to @xmath18 if @xmath178 .
notice that if @xmath53 is a pseudo - couple vertex , then @xmath206 when @xmath18 is a type ii @xmath153or iii@xmath174 set with the smallest couple number .
[ lm : gammageq3 ] assume that @xmath18 is a type ii or iii set with the smallest couple number .
if there exists @xmath176 for @xmath177 , then either @xmath207 or @xmath208 . since @xmath176 and @xmath18 is with the smallest couple number , by lemma [ lmm : gamma=2 ] , @xmath53 can not be a pseudo - couple vertex and a vertex @xmath209 must be in @xmath18 .
we consider the following three cases .
* case 1 . * @xmath210 .
it is clear that @xmath211 and @xmath114 must be dominated by @xmath212 and @xmath115 , respectively .
similarly , @xmath213 and @xmath214 must be dominated by @xmath215 and @xmath216 , respectively ( see figure [ fig5](a ) ) . thus both @xmath217 and @xmath218 are greater than or equal to 4 .
* @xmath219 . in this case
, @xmath115 and @xmath220 must be in @xmath18 so that @xmath114 and @xmath221 are dominated .
thus @xmath222 ( see figure [ fig5](b ) ) .
* @xmath223 . by using a similar argument as in case 2 , the case holds .
this completes the proof .
[ lm : j - kneq4 ] assume that @xmath18 is a type ii or iii set with the smallest couple number .
if there exist @xmath224 for distinct @xmath225 , then @xmath226 .
suppose to the contrary that @xmath227 .
that is , @xmath228 is either equal to @xmath229 or @xmath230 .
we only consider the case where @xmath231 .
the other case is similar . by lemma [ lmm : gamma=2 ]
, @xmath53 can not be a pseudo - couple vertex and a vertex @xmath232 must be in @xmath18 .
analogous to lemma [ lm : gammageq3 ] , we consider the following three cases . * case 1 .
* @xmath210 .
we can find that @xmath233 ( see figure [ fig5](a ) ) . if @xmath234 is also in @xmath18 , then @xmath235 , a contradiction . if @xmath234 is not in @xmath18 , then , by lemma [ lmm : p_i=3 ] , @xmath236 , a contradiction too .
* case 2 . *
@xmath219 . in this case , @xmath237 ( see figure [ fig5](b ) ) . by using a similar argument as in case 1 , we can find that @xmath236 .
thus this case is also impossible .
* @xmath223 . in this case , all vertices in @xmath118 must be in @xmath18 .
this contradicts that @xmath238 .
this concludes the proof of this lemma .
by lemma [ lmm : p_i=3 ] , @xmath239 . in the following ,
we investigate the cardinalities of minimum dominating sets of @xmath47 under all possible values of @xmath240 . by lemma [ lm : gammageq3 ] , if there exists @xmath176 for @xmath177 , then either @xmath207 or @xmath241 must hold .
this yields @xmath242 or @xmath243 . by lemma [ lm : j - kneq4 ] , if @xmath176 , then both @xmath244 and @xmath245 are greater than or equal to 3 .
this means that no two distinct @xmath224 use the same @xmath246 to obtain the average number 3 .
this ensures that the average number of @xmath64 is greater than or equal to 3 when @xmath177 .
to prove the lower bound of @xmath247 , our main idea is to count the number of @xmath176 for @xmath248 , say @xmath91 , which can not gain support from any other @xmath249 so that their average is greater than or equal to 3 .
this yields @xmath250 .
[ lm : typeiibcd ] if @xmath18 is a type iii set with the smallest couple number , then @xmath251 .
first we consider the case where @xmath18 is a type iii(a ) , iii(b ) , or iii(c ) set . by inspection ( see figure [ fig8 ] ) , it can be found that @xmath101 for @xmath248 if @xmath18 is a type iii(a ) or iii(b ) set .
for the case where @xmath18 is a type iii(c ) set , one of the elements in @xmath252 ( respectively , @xmath253 ) must be in @xmath18 so that @xmath254 ( respectively , @xmath255 ) is dominated .
thus @xmath101 for @xmath248 if @xmath18 is a type iii(c ) set .
this also implies that , in those three types of dominating sets , if there exists @xmath176 in @xmath18 , then , by lemmas [ lm : gammageq3 ] and [ lm : j - kneq4 ] , either @xmath256 or @xmath257 must hold . as a consequence , @xmath258
this yields @xmath251 . to complete the proof
, it remains to consider the case where @xmath18 is a type iii(d ) set . according to the possible values of @xmath259
, we consider the following two cases . * case 1 .
* @xmath260 . since @xmath260 , vertices @xmath261 and @xmath262 must be in @xmath18 so that @xmath263 and @xmath264 are dominated ( see figure [ fig9](a ) ) .
this further implies that both @xmath265 and @xmath266 are greater than or equal to 4 .
note that if @xmath267 , then @xmath268 and @xmath269 must be in @xmath18 so that @xmath270 and @xmath271 are dominated .
accordingly , @xmath272 .
thus @xmath273 can gain support from @xmath274 such that @xmath275 .
we can find that the minimum values of @xmath276 are @xmath277 , and @xmath278 , respectively .
thus every @xmath176 in @xmath66 can gain support from a vertex @xmath279 .
thus @xmath280 which yields @xmath281 . thus this case holds .
* @xmath282 .
if @xmath283 or @xmath284 , then , after setting @xmath285 and @xmath286 , @xmath287 becomes a type ii set which will be considered in lemma [ lm : ri=3 ] .
we may assume that @xmath288 and either @xmath263 or @xmath289 is in @xmath18 ( see figure [ fig9](b ) ) .
note that , in this case , @xmath262 must be in @xmath18 so that @xmath264 is dominated .
thus the minimum values of @xmath290 are @xmath291 , and @xmath278 , respectively . clearly , their average is greater than or equal to 3 .
note that one of the vertices in @xmath292 $ ] must be in @xmath18 so that @xmath293 is dominated .
thus both @xmath273 and @xmath294 are greater than or equal to 3 .
this ensures that they will not gain support from @xmath266 and @xmath265 on computing the average value 3 .
hence @xmath280 and @xmath281 .
this completes the proof .
[ lm : gamma_i>4 ] if @xmath18 is a minimum dominating set with the smallest couple number under the condition that @xmath88 and @xmath295 , then @xmath281 .
since the case where @xmath18 is a type iii set with the smallest couple number is a special case of @xmath295 , by lemma [ lm : typeiibcd ] , this corollary follows . it remains to investigate lower bounds for type ii sets . by using a similar classification in @xcite ,
we consider the following five classes : @xmath296 , and @xmath43 . [ 5nn3 ] for @xmath41 or @xmath42 , if @xmath18 is a type ii set , then @xmath251 .
let @xmath18 be a type ii set with the smallest couple number .
by inspection on figure [ figb](a ) , only the elements in @xmath297 are possibly equal to 2 and all other @xmath101 after gaining support .
note that @xmath298 since @xmath299\cap s\neq { \varnothing}$ ] .
this yields @xmath300 . for @xmath41 , @xmath301 , and @xmath302 for @xmath303
this completes the proof .
[ repeatedpatten ] let @xmath18 be a dominating set of @xmath47 .
a block @xmath52 is called a self - contained block if @xmath304 @xmath153see figure [ fig : selfcontained]@xmath305 .
[ twoselfcontainedblocks ] if both @xmath52 and @xmath306 are self - contained blocks , then @xmath307 for @xmath308 . by inspection
( see figure [ fig : selfcontained](b ) ) , the proposition follows .
[ lm : ri=3 ] for @xmath309 and @xmath129 , if @xmath18 is a type ii set and any two of @xmath259 , @xmath310 and @xmath311 are greater than 2 , then @xmath312 . analogous to lemma [ 5nn3 ] , only the elements in @xmath297 are possibly equal to 2 and all other @xmath236 after gaining support . if any two of @xmath259 , @xmath310 and @xmath311 are greater than 2 , then @xmath313 . by replacing @xmath22 by @xmath43 , this yields @xmath314 .
clearly , @xmath315 when @xmath309 .
thus @xmath312 .
this completes the proof .
[ tm : ri=3 ] for @xmath309 , if @xmath18 is a type ii set , then @xmath316 .
suppose to the contrary that @xmath317 .
we claim that exactly one of the vertices in @xmath318 is in @xmath18 .
we argue the claim by contradiction and assume that there are two vertices in @xmath319 . in this case , if one of @xmath264 and @xmath254 and one of @xmath289 and @xmath263 are in @xmath18 , then @xmath320 and @xmath321 . by lemma [ lm : ri=3 ] , @xmath316 , a contradiction . if @xmath322 , then @xmath323 so that @xmath264 is dominated .
this results in @xmath324 and @xmath325 .
thus @xmath310 can gain support from @xmath265 , by lemma [ lm : ri=3 ] , @xmath316 , a contradiction .
thus the claim holds and @xmath260 .
accordingly , we have the following four cases to consider . * case 1 .
* @xmath326 . in this case , @xmath327 so that @xmath263 and @xmath289 are dominated ( see figure [ boundof5k4](a ) ) .
this results in @xmath328 , @xmath329 and @xmath330 since @xmath331\cap s \neq { \varnothing}$ ] .
the minimum values of @xmath64 for @xmath332 are @xmath333 , and 3 , respectively , and @xmath259 can gain support from @xmath266 .
this leads to @xmath334 .
by using a similar argument as in lemma [ lm : ri=3 ] , @xmath335 , a contradiction .
thus this case is impossible .
* case 2 . * @xmath336 . since @xmath260 and @xmath336 , both @xmath337
( see figure [ boundof5k4](b ) ) .
furthermore , @xmath338 so that @xmath289 is dominated .
we claim that either @xmath261 or @xmath262 is in @xmath18 . if both @xmath261 and @xmath262 are not in @xmath18 , then @xmath339 .
however , @xmath254 is a pseudo - couple vertex and , by lemma [ lmm : gamma=2 ] , @xmath18 does not have the smallest couple number , a contradiction .
thus this claim holds and @xmath340 . note that @xmath341 .
the reason is that if @xmath342 , then @xmath330 ; otherwise , by lemma [ lmm : p_i=3 ] , @xmath330 .
the minimum values of @xmath64 for @xmath332 are @xmath333 , and 3 , respectively .
thus @xmath259 can gain support from @xmath266 .
thus this case is also impossible . *
* @xmath343 . in this case , @xmath344 so that @xmath264 is dominated ( see figure [ boundof5k4](c ) ) .
we claim that @xmath345 .
suppose to the contrary that @xmath346 .
this yields @xmath347 and @xmath348 .
this results in the minimum values of @xmath64 for @xmath332 to be @xmath349 , and 3 , respectively , and @xmath259 and @xmath310 can gain support from @xmath266 and @xmath265 , respectively .
hence @xmath350 , a contradiction .
thus the claim holds .
when @xmath345 , at least one vertex in @xmath351 must be in @xmath18 so that @xmath293 is dominated .
if two vertices in @xmath352 are in @xmath18 , this results in @xmath353 and @xmath354 .
this further implies that @xmath355 , a contradiction .
thus at most one of @xmath356 , and @xmath270 can be in @xmath18 .
we claim that @xmath271 can not be in @xmath18 either .
if @xmath357 , then @xmath358 to ensure that @xmath270 is dominated .
moreover , @xmath359 no matter whether @xmath268 is a pseudo - couple vertex or not .
this results in the minimum values of @xmath64 for @xmath332 to be @xmath360 , and 3 , respectively .
hence @xmath361 , a contradiction .
thus the claim holds and only one of @xmath293 and @xmath270 can be in @xmath18 . if @xmath293 is in @xmath18 , then , after replacing @xmath293 by @xmath270 , all vertices in @xmath292 $ ] are also dominated .
thus we only consider the case where @xmath362 .
note that , in this case , @xmath363 is exactly a self - contained block . by repeating the above procedure on @xmath364 for @xmath365 ,
the only possible result is that all @xmath364 are also self - contained blocks and @xmath366 ( see figure [ boundof5k4](d ) ) . by proposition [ twoselfcontainedblocks ]
, we have @xmath367 for @xmath368 .
finally , we can find that at least one vertex in @xmath369 must be in @xmath18 so that @xmath370 is dominated .
this results in @xmath371 for @xmath372 .
thus @xmath373 and @xmath311 can gain support from those vertices .
this yields @xmath374 , a contradiction .
* @xmath375 . in this case , @xmath262 and @xmath271 are in @xmath18 so that @xmath264 and @xmath289 are dominated . since @xmath376 is still a a type ii set with the smallest couple number which is already considered in case 3
therefore , this case is also impossible .
this concludes the proof of the lemma .
[ 5kn+1 + 2 ] for @xmath6 and @xmath377 , if @xmath18 is a type ii set , then @xmath378 .
by using a similar argument as in case 3 of lemma [ tm : ri=3 ] , we can construct a dominating set @xmath96 of @xmath2 when @xmath6 ( see figure [ fig - kuo-6](a ) ) .
note that all @xmath194 for @xmath100 except @xmath379 .
thus @xmath380 and @xmath381 .
however , @xmath382 .
this yields @xmath383 and @xmath378 .
similarly , when @xmath377 , let @xmath384 ( see figure [ fig - kuo-6](b ) ) .
it can be verified easily that @xmath287 is a dominating set of @xmath46 .
note that all @xmath385 for @xmath100 except @xmath386 .
thus @xmath387 and @xmath388 .
however , @xmath389 .
this yields @xmath390 and @xmath391 .
this completes the proof .
we summarize our results as the following theorem .
[ mainresult ] assume that @xmath4 is a faulty vertex in @xmath2 . then for @xmath33 @xmath34 by corollaries [ coro : gammapn2 ] and [ lm : gamma_i>4 ] and lemmas [ lm : typeiibcd ] and [ 5nn3 ] , @xmath8 when @xmath41 and @xmath42 . by corollaries [ coro : gammapn2 ] and [ lm : gamma_i>4 ] and lemmas [ lm : typeiibcd ] and [ tm : ri=3 ] , @xmath8 when @xmath43 . by corollaries [ coro : gammapn2 ] and [ lm : gamma_i>4 ] and lemmas [ lm : typeiibcd ] and [ 5kn+1 + 2 ] , @xmath392 when @xmath6 or @xmath7 .
this completes the proof .
in this paper , we show that @xmath393 if @xmath6 or @xmath7 ; otherwise , @xmath8 .
our results can be applied to the alteration domination number of @xmath2 . by theorem [ mainresult ]
, we can find the lower and upper bounds for @xmath37 as follows : @xmath38 if @xmath6 or @xmath7 ; otherwise , @xmath39 . as a further study , it is interesting to find out the exact value of @xmath37 . on the bondage problem in @xmath2
, it is clear that the domination number is still @xmath31 after removing any edge from @xmath2 . by theorem [ mainresult ]
, we can find that @xmath394 if @xmath41 , @xmath42 or @xmath43 .
it is also interesting to find out the exact value of @xmath395 . | in this paper , we investigate the domination number of generalized petersen graphs @xmath0 when there is a faulty vertex .
denote by @xmath1 the domination number of @xmath2 and @xmath3 the domination number of @xmath2 with a faulty vertex @xmath4 .
we show that @xmath5 when @xmath6 or @xmath7 and @xmath8 for the other cases . 0.2 in * keywords : * domination ; domination alternation ; generalized petersen graph . |
we start the introduction by considering the following views of great men : @xmath0``although nobody is in doubt today of the validity of the remarkable interpretation of thermodynamics with which statistical mechanics , following the efforts of boltzmann and gibbs , has recently provided us , it still remains extremely difficult to give a completely accurate justification for it '' -louis de broglie@xcite .
@xmath0``the problem [ defining the ensemble distribution depending upon the external conditions imposed on a system ] was solved by gibbs , although a rigorous justification of the distributions obtained is a complicated problem that is still not completely solved at the present time .
it is not even clear to what extent this rigorous justification is possible '' -d .
n. zubarev@xcite .
@xmath0``there is probably no other well - established field of theoretical physical science that is as much plagued by paradox and criticism of its fundamental logic as is statistical mechanics '' -joseph edward mayer and maria goeppert mayer@xcite @xmath0 `` if we are describing only a state of knowledge about a single system , then clearly there can be nothing physically real about frequencies in the ensemble ; and it makes no sense to ask , ' ' which ensemble is the correct one ? `` ... gibbs understood this clearly ; and that , i suggest , is the reason why he does not say a word about ergodic theorems , ... '' e. t. jaynes@xcite @xmath0 there is lot of confusion regarding ergodic hypothesis in the literature ( regarding whether it is necessary for statistical mechanics or not ) and in the usual discussions of people . also the reasons how and why statistical mechanics works are not properly understood .
the problem appears complicated as one has to take into consideration various ingredients of vary different nature ( from the character of experimental measurements , large number of degrees of freedom , and the microscopic dynamical properties ) involved in the statistical approach for dynamical systems . in the present review we try to dis - entangle various ingredients and present the resolution of the ergodic problem in a simpler and clear way .
a detailed study of literature shows that there are mainly three camps at the foundations of statistical mechanics ( 1 ) ergodic school , ( 2 ) non - ergodic school , and ( 3 ) quantum foundations of statistical mechanics .
we critically analyse all the three approaches to reach on a broader understanding of the foundations of statistical mechanics or we attempt to see the `` complete picture '' .
aim of the present review is to show that ergodicity is not relevant for equilibrium statistical mechanics of macroscopic systems ( with tolman and landau ) .
we will see that the solution of the problem is in the very peculiar nature of the macroscopic observables ( they are sum functions ) and with the very large number of the degrees of freedom involved as first pointed out by khinchin .
similar arguments are used by landau based upon the approximate property of `` statistical independence '' .
we critically review these ideas .
we also review the role of chaos ( classical and quantum ) in the foundations of statistical mechanics , i.e. , where it is important and where it is not important . for students and beginners
the purpose of the article is best served if they first understand the first chapter of landau - lifshitz s book on statistical physics and read the little book of khinchin on the mathematical foundations of statistical mechanics .
_ again our motivation is to have a broader view of the subject and a pedagogical presentation . _ in the present section we present a brief summary of the essential topics in the form of questions and answers . in section
ii we review the ergodic approach to the foundations of statistical mechanics by dividing it into further subsections . first defining the ergodic problem and then giving a historical account of how the ergodic problem came up with the works of maxwell and boltzmann . in subsection
ii(c ) the reformulation of the ergodic problem by birkhoff is given and the resolution of the ergodic problem by khinchin is given in subsection d. the role of chaos , integrability , and non - integrability is discussed in subsection e. in section iii we present the non - ergodic approach .
this section consists of two subsections one on landau - lifshitz s approach and another on e. t. jaynes approach with a critique .
we also advance the plausibility arguments for equal - a - priori probability hypothesis .
the last section is devoted to the quantum foundations of statistical mechanics . in this
we review the eigenstate thermalization hypothesis , the quantum ergodic theory of von neumann , and other recent quantum - statistical typicality approaches .
we will see that von neumann s quantum ergodic theorem is a general statement applicable to systems with many degrees of freedom and the eigenstate thermalization hypothesis is a consequence of quantum chaos , other approaches falling under typicality properties are also analyzed .
the issue of compatibility of microscopic time reversibility and macroscopic time irreversibility is not considered here .
this has been clarified in detail in the beautiful papers of jeol lebowitz@xcite(for irreversibility in quasi - isolated systems see@xcite ) . in their famous book@xcite lev
landau wrote : `` in the discussion of the foundations of statistical physics , we consider from the start the distribution of small subsystems ...... this allows the complete avoidance of ergodic or similar hypotheses , which are in fact not essential as regard to these aims [ i.e. , the foundations of physical statistics ] '' @xmath0 as we know from the most used prescription of gibbs we calculate the phase space averages of dynamical quantities using appropriate ensembles and we find that these phase averages agree very well with experiments .
clearly actual experiments are not done on a hypothetical ensemble they are done on the actual system in the laboratory and these experiments take a finite amount of time . thus it is usually argued that actual measurements are time averages and they are equal to phase averages due to ergodicity , @xmath1 now we will give the following reasons against ergodicity .
\(1 ) if one assumes that during the measurement time the system samples all the microstates and @xmath2 limit of equation ( 1 ) is reached during the measurement time , then ones assumption is wrong .
it is a well know fact that time taken by the system to sample all the available phase is fantastically large ( more than the age of the universe ) .
for example@xcite , if we consider a nanoparticle with @xmath3 nuclei each with spin half and consider only the nuclear spin system .
the total number of quantum state are @xmath4 .
consider that these spins continuously flip from up spin to down spin and vice versa and this spin flipping is caused by phonons from the system and bath in which the nanoparicle is present .
at room temperature this frequency is about @xmath5 cycles per second .
let us assume that each spin makes a transition during this period .
thus the total number of transitions per second is @xmath6 .
the time taken by the system to go over all the states is @xmath7 , and the age of universe is @xmath8 . _ thus clearly during the laboratory measurement time the full ensemble is not realized , and only a very tiny fraction of the total number of the members of the ensemble is realized . _ note that the time scale diverges nearly exponentially with the number of atoms in the sample .
\(2 ) if one accepts ergodic hypothesis ( and do not accept explanation ( 1 ) ) , that the laboratory measurements are time averages ( measurement time scale being much large as compared to microscopic dynamical time scale ( time taken to flip a spin @xmath9 ) , then one accepts equation ( 1 ) and accepts @xmath2 ) , then time average value of an observable ( or a result of measurement ) in a series of experiments ( repeating the same experiment again and again ) on the same system will agree with ensemble average .
this is in fundamental contradiction with what we observe i.e. thermodynamical measurements are not strictly reproducible@xcite .
fluctuations in thermodynamic measurements are the unavoidable consequence of our lack of control of microscopic dynamics and thus our lack of control of the initial micro - condition at the start of the measurement .
the above discussion against ergodicity opens the question `` but how does statistical mechanics work ? '' and the theory has been so successful . the answer is : in statistical mechanics we concern with macroscopic systems and with macroscopic observables which are very special ones .
special in the sense that they are `` sum functions '' i.e. , their value for the whole system is equal to the sum of equivalent functions for the small parts of the body ( for example , energy of the whole system is equal to the sum of energies of the small parts of the body ) . due to the property called `` statistical independence '' of small parts of the body ,
the sum functions take almost constant value on the energy hypersurface ( details are given in section ii(d ) ) .
@xmath0the microscopic time taken by the measurement produces that `` same '' value .
thus equality of time average and phase space average is imposed due to the fact of `` temporal constancy of special macroscopic observables in equilibrium '' .
the magnitude of fluctuations goes as @xmath10 .
thus these are very small for macroscopic systems . put differently , for an extremely large number micro - states ( called typical micro - states ) the macro - state of the system
is the same . only very exceptional micro - states ( `` bad states '' with fantastically small probability )
give non - typical behaviour .
to explain these lines let us consider a macroscopic system and measure an observable .
repeat the experiment many many times , each time the starting micro - state is different but important point is that each time we get approximately the same value for the observable ( since huge number of micro - states are equilibrium micro - states ) .
it is extremely rare that we get a value during a measurement which is very different from the value in other measurements .
this `` typicality '' is at the heart of why statistical mechanics explain thermodynamic behaviour . here
the probability theory inters into statistical description .
the typicality explains its high predictability for macroscopic systems ( even though statistical mechanics has probabilistic basis ) .
as the macro - observables are extremely insensitive to micro - condition and microscopic dynamics , the role of microscopic dynamics is very insignificant in equilibrium statistical mechanics .
see section ii .
d. for a detailed explanation , and the role of fermi - pasta - ulam problem in this connection ( section ii .
also it is good idea to abandon 19th century philosophy of explaining everything we see around with microscopic dynamics called `` reductionism '' .
boltzmann initially tried to explain 2nd law of thermodynamics with microscopic molecular dynamics but later on realized the need of of statistical laws .
at each level of complexity new laws emerge not explainable `` purely '' by shroedinger equation@xcite .
in standard practice , for example for a system of volume @xmath11 with @xmath12 particles in equilibrium with very large heat bath at temperature @xmath13 , one use s gibbs canonical ensemble @xmath14 here @xmath15 is the probability to find the system in the infinitesimal phase volume @xmath16 around the phase point @xmath17 [ @xmath18 for a system of @xmath19 degrees of freedom ] .
@xmath20 with @xmath21 called boltzmann constant .
the negative logarithm of partition function @xmath22 is proportional to the free energy of the system @xmath23 from the free energy and its derivatives all the needed information is extracted .
each time we get the gibbs ensemble averaging , for example , pressure is given as @xmath24 with hamiltonian @xmath25 as a function of the volume @xmath11 ( through potential function with boundaries etc . ) .
thus our computational algorithm involves phase space averaging .
but the actual experiments are done on the given system in the laboratory ( not on the hypothetical ensemble ) .
measurements during the experimentation take finite amount of time and thus what we measure in laboratory is the time averages not the ensemble averages .
so the immediate question arises how to justify the replacement of time averages with ensemble average .
this is called the ergodic problem .
there is a considerable confusion regarding the precise meaning of ergodic hypothesis .
thus we will present a brief historical account about the origin and its mis - interpretations by ehrenfests@xcite . in this analysis
we will better understand the ergodic program for justifying statistical mechanics .
boltzmann s first attempt to reduce the second law of thermodynamics to a theorem in mechanics appeared in 1866 and his last papers in columns of nature in 1890s regarding his debates on irreversibility . during his scientific career ( 1866 - 1895 )
boltzmann tried to understand the kinetic theory of gases and thermodynamics with many different approaches based on atomic constitution of matter .
favouring one approach at one time and then rejecting it for the another and then returning back again to the previous one .
this is a very distinctive character of boltzmann@xcite .
roughly , from 1866 - 1871 he was interested in deriving second law of thermodynamics purely from mechanics ( in his 1866 paper no probabilistic argument is present ) , in 1867 he reads maxwell s work and his subsequent papers from 1868 to 1871 uses probabilistic notions after maxwell s velocity distribution function . in these papers
he extends maxwell s results to a gas in an external potential ( maxwell - boltzmann law ) . in the paper@xcite of 1868
he first introduces his ergodic hypothesis .
he was considering an isolated gas in an arbitrary initial state ( see ehrenfests@xcite ) , then he argues based on the empirical fact that systems tend to equilibrium and permanently stay there that average behaviour over a long time interval will the thermal equilibrium behaviour ( note that this long before his h - theorem ) . according to ehrenfests here he introduces the concept of ensembles ( much before gibbs ) . in justifying the equivalence of ensemble averages and temporal averages he introduces a hypothesis ( note that he do not use the word ergodic at this point
, word ergoden appears much later in his 1884 paper in a different context ) .
thus the present day terminology is largely due to ehrenfests .
note also that this is the birth of modern statistical mechanics .
the older treatments in which statistical aspects are related to single molecules are called ( by ehrenfests ) the `` kineto - statistics of the molecule '' and the 1868 paper treats gas model as a whole and ehrenfests call it `` kineto - statistics of the gas model '' . important point to be noted here is that the version of ergodic hypothesis which is attributed to boltzmann is stronger than what the founding father has thought : `` the great irregularity of the thermal motion , and the multiplicity of forces that act on the the body from outside , make it probable that the atoms themselves , by virtue of motion that we call heat , pass through all possible positions and velocities consistent with the equation of kinetic energy , and that we can therefore apply the equations previously developed to the coordinates and velocities of the atoms of warm bodies '' now the present strong form ( trajectory passing through each and every point of the energy surface for an isolated system ) of ergodic hypothesis can not be attributed to boltzmann , it is the mis - interpretation on the ehrenfests side .
the representative phase point of the system passing through every point of the phase space is impossible .
in fact later in 1913 the impossibility of ergodic systems ( stronger form ) was proved by independently by rosenthal and pancherel on measure - theoretic arguments .
ehrenfests after mis - interpreting the boltzmann ergodic hypothesis proposed some what weaker form and they called it the quasi - ergodic hypothesis , which meant that the trajectory of phase point covers the energy surface densely even though not actually passing through every point of it .
an excellent discussion of this mis - interpretation is given in the work of s. g. brush@xcite .
in the period 1872 - 1878 he wrote two most important papers of his life , the 1872 paper contains what we now call `` boltzmann equation '' and h - theorem .
here he claimed that his h - theorem provided the desired theorem from mechanics to explain irreversibility and second law of thermodynamics .
this came under a serious objection due to loschmidt in 1876 who insisted on the incompatibility of the time - asymmetric behaviour ( shown by h - theorem ) and the time - symmetric behaviour of microscopic equations of motion ( these debates are now well understood and documented@xcite ) . as a result boltzmann rethought the basis of his approach and presented a very different approach in 1877 what we now call permutational argument and his famous result @xmath26 ( see appendix a ) written in modern form by max planck .
equilibrium is now conceived as most probable macro - state instead of stationary macro - state ( in time ) .
it is highly probable for a system initially in a non - equilibrium state to move towards equilibrium state ( real space density distribution in accord with external conditions(potentials ) and maxwellian velocity distribution ) as the phase space volume of the equilibrium macro - state is fantastically large as compared to non - equilibrium state .
thus the equilibrium is most probable state but arbitrary deviations from it are also probable , but it turns out that the probability is fantastically small .
this originates the notions of typicality .
third period , in 1880s he left the `` purely '' probabilistic approach and again went back to his mechanical explanation of the laws thermodynamics ( we see here that boltzmann did not stick to one approach ) .
he wrote an important paper in 1884 largely forgotten by now ( see the review by gallavotti@xcite ) .
this paper is a generalization of a paper by helmholtz on the mechanical models of thermodynamics based on monocyclic systems .
boltzmann proves a theorem called heat theorem , i.e. , @xmath27 . here
@xmath13 is the time average of the kinetic energy @xmath28 , @xmath29 is the total energy , and @xmath30 the potential energy which depend on an experimentally controllable parameter @xmath11 .
the motivation here was to find mechanical analogues of thermodynamic entropy .
result was proved for monocyclic systems by helmholtz and boltzmann generalize this for high dimensional systems under the assumption of ergodicity ( note that this is the second time he introduces ergodic considerations ) . in his discrete world - view
he imagined the phase trajectory visiting all the discrete cells of the phase space . with the introduction of some stationary phase space distribution ( he calls it `` monode '' )
, the calculation of difficult time averages can be replaced with much simpler phase averages .
he considers an ensemble of systems in exactly same macroscopic conditions and a family of stationary distributions which he calls `` monode '' ( note that this the introduction of ensembles much before gibbs ) . a monode which verifies the heat theorem is called by him the `` orthode '' . here
( first time ) he considers two orthodes ( 1 ) ergode ( @xmath31 microcanonical ensemble ) and ( 2 ) holode ( @xmath32 canonical ensemble ) , see fig 1 . as we have seen in section i(a ) , in real thermodynamic measurements we do not have infinite time averages , and thus we are not justified in equating the finite time averages with infinite time averages and then using the ergodic hypothesis ( which has never been proved in generality ) .
the answer to this was already clear to boltzmann .
the answer lies in the peculiarities of the thermodynamic observables and the large number of degrees of freedom involved .
we will discuss this in section ii(d ) .
we also note that the original ergodic hypothesis as envisaged by boltzmann was a weaker statement as compared to modern version of that .
ehrenfests 1911 encyclopedia article@xcite generated lot of interest about ergodic systems in the community of mathematicians ( paul ehrenfest was the student of boltzmann ) . @xmath0if one is able to prove the ergodic hypothesis for a given system i.e. , @xmath33 , then 42 one eliminates the necessity of determining initial state of the system and of solving hamilton s equations .
one justifies the dynamical foundation of statistical mechanics .
clearly , then statistical mechanics reduces to a branch of mechanics .
but story is not so simple .
it is a very difficult problem . in 1931
, g. d. brikhoff@xcite reduced the ergodic property of a dynamical system to an equivalent property called `` metric transitivity '' .
he proves the following theorems : 45 @xmath34 exists for almost every @xmath35 .
a necessary and sufficient condition for the system to be ergodic is that the phase space be metrically transitive .
a system is `` metrically in - transitive ''
iff there exists regions @xmath36 and @xmath37 of phase space such that @xmath38 and @xmath39 , which are invariant under the system s dynamics : @xmath40 and @xmath41 for all @xmath42 . in simple words phase point
wanders all the available phase space if and only if the system is metrically transitive .
@xmath0this property again can not be experimentally verifiable .
thus the implications of brikhoff s theorems for physical statistical mechanics are inconclusive .
@xmath0but luckily or unluckily , we have seen that ergodicity is not required for doing statistical mechanics of macroscopic systems .
khinchin@xcite(see also the 4th chapter of the book@xcite ) approached the ergodic problem from physical point of view .
he pointed out the importance of the following points 1 . in statistical mechanical systems of interest
the number of the degrees of freedom is very large .
2 . in statistical mechanical systems
the observables of interest are very special ones they are `` sum functions '' ( @xmath43 ) . with these considerations he proves the following two theorems : @xmath0theorem ( 1 ) :
consider a physical quantity @xmath44 pertaining to a single molecule and define phase coefficient of correlation : @xmath45 . if @xmath46 for @xmath47 the function @xmath48 is ergodic ( @xmath49 ) .
the physical assumption that goes into this theorem is : `` because of the fact that the given system consists of a very large number of molecules it is natural to expect that knowledge of the state of a single molecule at a certain moment does not permit us to predict anything ( or almost anything ) about the state in which this molecule will be found after a sufficiently long time '' khinchin .
@xmath50this is equivalent to the idea of molecular chaos .
next he proves a much more relevant result ( to physical statistics):@xmath50theorem ( 2)@xcite : @xmath51 here @xmath52 is the actual value of the observable ( no time averaging ! ) pertaining to the whole system , and @xmath53 and @xmath54 are @xmath55 .
he proves that the correlation co - efficient between phase functions is actually very small for large @xmath12 .
@xmath0this implies that the physically relevant observable are self averaging !
the above statement is the resolution of the whole problem _ ( for macroscopic systems and sum - function observables ) . _
as far as one is concerned with macroscopic systems ( large @xmath12 ) and sum function observables chaos , integrability , and non - integrability does not play any role in the foundations of statistical mechanics . as usual statistical mechanical systems in the laboratory
are never isolated , the molecular chaos assumption of khinchin seems to be true and the consequences are clear to us . but for `` isolated '' low dimensional systems and non - sum - function observables the case is not so simple . before 1955
there was a general consensus that weak extraneous interactions makes the statistical mechanical systems ergodic ( phase point wandering the whole of phase space ) . in 1955 ,
fermi and his collaborators did a numerical experiment on a chain of non - linearly coupled harmonic oscillators .
their expectation was that the small non - linear coupling will make the system ergodic ( energy stored initially in one of the normal modes will go over in time to all other modes ) .
but the result was quite surprising .
consider the hamiltonain of the system as @xmath56 .
for @xmath57 system is completely integrable and the energy @xmath58 of each normal mode @xmath59 is conserved .
it was expected that when @xmath60 energy stored initially in one normal mode will `` spread '' to all other modes .
but this did not happen , the energy periodically comes back to the original mode showing no sign of equipartition of energy .
this `` paradox '' can be explained with kam theorem . which states that
if @xmath61 is small enough then on the constant energy surface , invariant tori survive in a region whose measure tends to @xmath62 as @xmath63 .
there is a threshold @xmath64 called kam threshold , if ( 1 ) @xmath65 kam tori play a major role and system does not follow equipartition , and ( 2 ) if @xmath66 then kam tori has a minor role and equipartition happens . in 1987 , livi , pettini , ruffo , and vulpiani published an important paper@xcite on the relevance of chaos regarding the predictions of canonical ensemble .
they divided the chain into coupled subsystems containing large number of particles ( see figure 3 ) .
let @xmath67 is the per particle time averaged energy of the subsystem @xmath68 and @xmath69 it s heat capacity @xmath70 .
it was found that the time averages limit ( private communication with roberto livi ) ] agrees well with the predictions of the canonical ensemble , even though the system was having the kam tori .
however the story is not so simple .
they also considered another model of coupled rotators in which only @xmath67 showed canonical behaviour not @xmath69 as the integrability - to- non - integrability parameter was changed ( see for details chapter 4 of @xcite ) .
note also that @xmath69 is not a sum function .
thus one can loosely say that _
`` more coarse your observables , less you depend on chaos '' . _
as expressed at the beginning of section i ( a ) , lev landau did not believe in the need of ergodic or similar hypotheses in justifying the foundations of statistical mechanics@xcite . from the beginning they consider a system always present in a bath , i.e. , subsystem in a given system which is in itself present in bigger system ( as in nature no system is ideally isolated ) .
they argue that the extreme complexity of the interactions of the system with the surrounding bodies makes the system s phase point to wander in the phase space .
let in a sufficiently long interval of time @xmath13 the phase point spends @xmath71 amount of time in a given volume @xmath72 of the phase space , then , @xmath73 is the probability that , if the system is observed at any arbitrary time , its phase point will be found in the volume @xmath72 .
going to the infinitesimals @xmath74 . here
@xmath75 is the density of that temporal probability distribution , i.e. , @xmath76 is the probability to find the system , observed at any time , in the infinitesimal phase volume @xmath16 .
the important point to be noted here is that we have `` only one '' system under consideration ( no statistical ensemble is considered at this point ) .
the distribution defined @xmath75 is a temporal distribution for that system . with the introduction of statistical distribution function
one can calculate the mean value of any dynamical quantity @xmath77 as @xmath78 it is obvious by the definition ( 6 ) of probability that the statistical averaging ( equation ( 7 ) ) is exactly equivalent to the infinite time averaging @xmath79 thus avoiding any ergodic hypothesis , as there is no ad - hoc introduction of any probability distribution .
the @xmath75 is the actual probability distribution in the system s phase space from its temporal behaviour .
but it is non - trivial to prove that @xmath80 this has never been proved in generality ( it turns out that this is a very difficult problem for a system with large number of degrees of freedom ) .
we will see from landau s argument and with the assumption of `` statistical independence '' that @xmath81 when number of degrees of freedom involved becomes very large .
then they argue@xcite that predictions of statistical mechanics are very reliable due to the fact that the relevant observables take almost constant value on the energy surface .
they also consider `` sum '' functions @xmath82 .
their considerations are based on a very general fact of `` statistical independence '' of various parts ( pertain to subsystems ) of the macroscopic observables ( it is not true for long range interacting systems ) . on `` statistical independence '' they prove : @xmath83 .
one can say @xmath0@xmath84 1 .
the distribution function of two sub - systems is equal to the product of individual sub - systems functions ( statistical independence ) , i.e. , @xmath85 ( for simplicity of notation here @xmath86 ) .
thus log of the distribution function is an additive integral - of - motion .
3 . due to liouville s theorem ( hamiltonian dynamics )
distribution function is an integral - of - motion ( temporal invariance of the phase space distribution of an ensemble ) .
we have only seven independent additive integrals of motion @xmath88 , and @xmath89(from mechanics ) .
@xmath90 6 .
from which canonical ensemble ensues @xmath91 , say , for a system in a box .
the important point to be noted is that there are @xmath92 other integrals of motion ( excluding the additive ones , i.e. , energy , three components of linear and three components of angular momentum ) .
these remaining integrals of motion are not additive thus they do not play any role in the landau s argument .
thus we see in lev landau s program for the foundations of statistical mechanics , if one starts from the beginning with a system present in a bigger system and one assumes the property of statistical independence and take note of the additive integrals of motion then one can reach the canonical distribution and no ergodic hypothesis is required as we have not constructed any ensemble .
we observe that the property of statistical independence is equivalent to the hypothesis of equal - a - priori probabilities .
we give two reasons : reason ( 1 ) : there are two views at the foundations of statistical mechanics ( 1 ) is based on the principle of equal - a - priori probabilities ( kubo s book , tolman s etc ) and the ( 2 ) is based on quasi - closed subsystems and the additive integrals of motion ( due to l. d. landau , see their book ) .
if one analyze them carefully one sees that both view - points are consistent with each other , first one comes from the second .
the principle of equal - a - priori probability is just a consequence of the fact that there are mechanical invariants of motion proportional to log of the distribution function for a quasi - closed subsystem @xmath93 ( due to statistical independence ) , which is another integral of motion due to liouville s theorem , thus one has @xmath94 .
the right hand side of this equation is the content of dynamics ( conservation laws ) .
the left hand side consists of a more subtle quantity related to statistical properties and a consequence of the theorem of dynamics and the property of statistical independence . since @xmath95 is constant no matter where is phase point is in the available phase space .
consequently log of @xmath96 , thus @xmath96 is same for all possible @xmath17 .
this is an equal - a - priori probabilities ( see figure 4 ) .
as tolman put it microstate has no preference to be in this or that part of phase space , but with the help of statistical independence it is more clearly seen .
@xmath0reason ( 2 ) : if one assume `` statistical independence '' then @xmath97 .
let @xmath98 , where @xmath99 .
now it is easy to prove that @xmath100 . thus @xmath101 is almost constant on the energy surface @xmath102 for @xmath103 . which again force us to say that @xmath104 take almost constant value on the energy surface a statement of equal - a - priori probabilities .
thus we see that the hypothesis of statistical independence directly gives us equal - a - priori probabilities .
the important point here to be noted is that the above reasons are only qualitative observations based on statistical independence and these are not rigorous " mathematical theorems .
most fundamental ingredient : the property of `` statistical independence '' .
@xmath0if one assume that various parts(macroscopic ) of the body are statistically independent from each other , then one can construct canonical and microcanonical formulation without any ad - hoc and extra hypothesis ( only based on statistical independence and some additive mechanical invariants of motion ) .
consider a macroscopic system in equilibrium , all its parts ( macroscopic ) or subsystems of the whole system are in equilibrium with each other .
let @xmath105 is the energy of the subsystem .
it remains constant when the whole system is in equilibrium and the subsystem although small as compared to the whole system is still macroscopic and the relative fluctuations in @xmath105 are very small @xmath106 .
one can argue that after a sufficiently long time intervals the subsystem can not be considered quasi - closed , as the effect of interaction of subsystems , however weak , will ultimately appear , and this weak interaction ultimately leads to the establishment of statistical equilibrium .
we will see below that in equilibrium this effect does not apply but in non - equilibrium it applies and we do not have any `` non - equilibrium canonical '' formulation .
let us consider the subsystem for @xmath71 amount of time ( much less than the relaxation time of the subsystem under consideration ) .
the dynamics is hamiltonian and due the liouville s theporem @xmath107 is constant of motion . with landau s argument @xmath108 or @xmath109
now let us wait for a time interval sufficiently long as compared to the relaxation time .
after this , the subsystem is no longer quasi - closed and say the distribution is @xmath110 at the end of this long time interval . again consider an interval of time @xmath111 much less than the relaxation time , during this interval system is again closed and landau s argument again applies @xmath112 or @xmath113 . from the above two equations we have @xmath114 . now thermodynamic considerations show that @xmath115 is to be identified with inverse temperature of any subsystem of the whole system . as the whole system is in equilibrium
thus all the subsystems of the whole system are in equilibrium with each other thus @xmath116 .
also normalization conditions demand that @xmath117 .
this gives us @xmath118 and the distribution remains stationary .
thus in equilibrium landau s argument applies at all times .
moreover @xmath119 is constant , as rhs of this equation is constant both in time and phase ( we neglect the fluctuations in @xmath105 ) .
thus @xmath120 , which is microcanonical ensemble . on similar lines with landau
s argument we can proceed to show the canonical distribution @xmath119 . here
we allow fluctuations in @xmath105 but keep @xmath115 same for all subsystems . _
the most important point to be noted here is that we do not `` mentally construct '' an ensemble and we do not assume equal - a - priori probability hypothesis .
all the statistical formulation comes out from the property of statistical independence " and the few additive integrals of motion ( note that statistical independence does not apply to long range interacting systems ) .
all other integrals of motion does not play any role in the landau s argument because they are not additive .
_ ( 1)temporal definition of probability : in a sufficiently long interval of time @xmath13 the phase point spends @xmath71 amount of time in a given volume @xmath72 of the phase space , thus @xmath121 is the probability that , if the system is observed at any arbitrary time , its phase point will be found in the volume @xmath72 . in the above definition , non - stationary case
can not be defined , as the probability to find the system in @xmath122 itself change with time .
thus the above definition excludes ubiquitous non - equilibrium cases .
\(2 ) as landau - lifshitz s whole analysis is based upon quasi - closed subsystems , the relation @xmath123 holds good only for not too long intervals of time `` . as after a sufficiently long time intervals
the subsystem can not be considered quasi - closed . in their own words ' ' over a sufficiently long interval of time ( compared to the relaxation time ) , the effect of interaction of subsystems , however weak , will ultimately appear .
moreover , it is just this relatively weak interaction which leads finally to the establishment of statistical equilibrium " .
we have seen that in equilibrium landau s argument applies at all times , but when the system is not settled to equilibrium ( temperature equilibrium ) we can not apply landau s argument due to the above reason .
there are two schools at the foundations of probability theory ( 1 ) frequentists ( probability from many random experiments ) , and ( 2 ) bernoulli - bayes - laplacians or simply bayesians ( principle of indifference ) .
jaynes standpoint : the ergodic problem is a conceptual problem related to the very concept of ensemble itself which is a byproduct of frequency theory of probability .
ergodic problem becomes devoid of any meaning when the probabilities of various micro - states are interpreted with bernoulli - bayes - laplace theory of probability . in the bayesian theory of probability , probabilities of occurrence of events are independent of the frequency concept
it is a more general viewpoint in which frequency theory is a special case and is based upon the principle of indeference .
thus if we do not visualize the probability of occurrence of a micro - state with frequency ( or construct ensemble ) there is no question of ensemble averaging and ergodic problem is completely bypassed .
then the statistical mechanical theory is attacked with statistical inference ( shannan entropy approach ) . in jaynes
words@xcite : @xmath0 `` we can have our justification for the rules of statistical mechanics , in a way that is incomparably simpler than anyone had thought possible , if we are willing to pay the price . the price is simply that we must loosen the connections between probability and frequency , by returning to the original viewpoint of bernoulli and laplace .
the only new feature is that their principle of insufficient reason is now generalized to the principle of maximum entropy . '' for more details see his detailed account@xcite .
the present author feels that it does not matter ( atleast in the computational problems of statistical mechanics ) that which theory of probability one is accepting .
but there is a disadvantage if we attack statistical mechanical problem with bayesian viewpoint .
if one accepts the bayesian viewpoint , then one abandons the concept of ensembles .
the probabilities of the microstates are obtained by maximizing the shannon s entropy ( with the given amount of macroscopic information ) .
but this also means that one is neglecting the dynamical properties of constituents of matter ( although they are not so important for sum - functions , a - la khinchin ) .
obviously molecules of a gas obey newton s laws or schroedinger s equation .
the phase point of an isolated system has no preference for this region or that region of phase space ( the phase point of an isolated system visits all the accessible regions of phase space , _ with equal frequency _
although time involved are fantastically large ) .
it is also the fact that in macroscopic observables ( which are so coarse on microscopic scale ) the microscopic dynamical properties do not reflect _ in all its detail _ , only very general microscopic dynamical properties reflect .
for example , additive mechanical invariants play the essential role ( in landau s approach ) . thus neglecting the dynamical properties of the constituents
altogether is not a valid starting point .
thus it seems difficult to describe the integrable to non - integrable transition in dynamical systems with jaynes approach . as we have seen that the sum - function macroscopic observables are so weakly depend on the dynamics of the phase point and
thus this is the resolution ( with khinchin ) , not with the bayesian viewpoint , which makes ergodic problem devoid of any meaning .
historically it is to be noted that prof .
g. uhlenbeck objected jaynes ideas@xcite .
since in the quantum case position and momentum do not commute with each other , the classical concept of phase space that helps us to envisage the dynamics of the phase point on the energy surface breaks down .
however an equation analogous to equation ( 1 ) can be written in the quantum case also .
the equivalent of `` phase space energy shell '' in the quantum case is the sub - hilbert space @xmath124 of the total hilbert space @xmath125 .
@xmath126 is defined as the space spanned by all @xmath127 ( eigenstates of the system s hamiltonian ) such that their corresponding eigenvalues @xmath128 are in a specified interval of energy from @xmath129 to @xmath130 .
the quantum equivalent of the classical ergodic hypothesis ( equation ( 1 ) ) is written as @xmath131 for a system in pure quantum state ( for details see equation ( 11 ) below ) .
for a system in mixed quantum state replace @xmath132 with @xmath133 as the diagonal element of the density matix in the energy representation .
_ here again we encounter ( lhs of the above equation ) the problem of infinite time averaging , and all the arguments of section i(a ) are valid , but we will see below that the `` resolving results '' analogous to the khinchin s self averaging theorem hold in quantum case also either based upon quantum chaos or based on von neumann s `` scale separation '' ideas .
we again see the fundamental role of typicality and large dimensional hilbert spaces_. eigenstate thermalization hypothesis(eth ) implies that the thermalization happens at the level of individual eigenstates . consider a system with hamiltonian @xmath25 and eigensystem @xmath134 prepared in some initial state
. if system has unitary evolution , then at any later time @xmath42 @xmath135 quantum mechanical mean of an operator @xmath136 pertaining to the system is : @xmath137 consider infinite time average @xmath138 @xmath139 is also called the diagonal ensemble in@xcite . now the thermodynamic universality demands : long time average @xmath32 average over appropriate statistical ensemble ( microcanonical or canonical etc . )
@xmath140 this is an universality relation with lhs depends on initial conditions but rhs does not !
_ now eigenstate thermalization hypothesis ( eth ) says that there are no eigenstate to eigenstate fluctuations in @xmath141 s for eigenstates close in energy _
@xmath142 this is called the eigenstate thermalization hypothesis ( eth ) by deutsch ( 1991 ) and srednicki ( 1994)@xcite : @xmath0``the quantum expectation value @xmath143 in a large interacting many - body system is equal to the thermal average '' @xmath144 if this is valid in generality , then one explain thermodynamic universality . also if the system obeys berry s conjecture ( eigenfunctions are random gaussian variables for a classically chaotic quantum system ) the off - diagonal elements @xmath145 are negligible and _ then one obtains thermal behaviour without any time averaging at all_@xcite .
also it has been numerically shown@xcite that the magnitude of the off - diagonal elements of the momentum distribution operator are very small as compared to the diagonal elements @xmath146 .
clearly much work is needed to show eth for other operators and _ importantly for sum - function operators _ ( across the integrability - to-(non)integrability transition ) . for detailed discussion see@xcite and the ref@xcite for the behaviour of diagonal and off - diagonal elements of various other observables .
study of literature shows that eth has been proved analytically only in special circumstances ( see figure 5 ) .
_ figure shows that some kind of chaos is necessary for eth to hold_. this is also evident in the numerical experiment of rigol etal@xcite for a small system ( five hard core bosons on a lattice ) that eth hold good in the non - integrable system but fails in the integrable one ( for marginal momentum distribution as the observable ) . clearly systems with very small number of degrees - of - freedom can not be fit into khinchin s program ( which requires macroscopic systems ) .
this implies that the statistical mechanical universality ( irrelevance of chaos ) that we enjoy with macroscopic systems and sum function observables may no longer holds good for small isolated quantum systems .
we present here the qualitative statement of von neumann s quantum ergodic theorem , for a precise definition see@xcite .
our aim here is to understand the physical basis of the theorem , to draw some analogies with the approach by khinchin , and to see the connection with the eth ( eigenstate thermalization hypothesis ) .
setup : consider a system with hamiltonian @xmath25 and total hilbert space @xmath125 .
consider that the total hilbert space is partitioned into mutually orthogonal sub - hilbert - spaces @xmath126 , with family @xmath147 ( @xmath148 ) .
let @xmath149 and @xmath150 .
let total number of partitions is @xmath151 i.e. , @xmath152 .
the physical basis of this partitioning ( according to von neumann ) is that each @xmath126 belongs to a different macro - state ( or macro - observer ) .
this is the crucial insight of von neumann , as in quantum theory all operators corresponding to observables do not commute among themselves .
von neumann take a physical stand point that the measurement process is coarse on microscopic variables , thus operators can be taken `` approximately '' commuting ( measurement errors are large as compared to bounds of uncertainty relations ) .
this is called `` rounding '' of operators .
the aim of the quantum ergodic theorem ( qet ) is to tell us some kind of universality or `` normality '' that the quantum expectation values of operators are close to the microcanonical averaging for most of the time ( under some conditions , see below ) .
it is a kind of quantum h - theorem which says that for all initial states the time evolution take the system from initial non - equilibrium state to the equilibrium state ( provided the system satisfy some conditions ) .
thus the total hilbert space is partitioned in sub - spaces @xmath126 and the macroscopic observables take specific values in each @xmath126 ( values taken in @xmath126 are different from those taken in @xmath153 when @xmath154 ) .
the `` rounded '' operators commute with each other but they do not commute with the hamiltonian ( this is analogous to non - integrals of motion in classical gibbs picture ) . qualitative statement of qet : let @xmath155 be the projection to @xmath126 with macro - state @xmath156 having some values of observables .
@xmath0then qet says : for _ all _ initial wavefunctions @xmath157 ( note that this includes the special initial non - equilibrium states ) , we have for most of the time @xmath158 in words : probability distribution over macro - states is approximately equal to the ratio of the dimension of hilbert space corresponding to that macro - state to the dimension of the total hilbert space . thus the macro - state with largest @xmath159 is most probable ( which is in accord with common sense , equilibrium state has the largest @xmath160 and it is typically observed )
. conditions involved in qet are : 1 .
hamiltonian is free of resonances i.e. , @xmath161 , unless either @xmath162 or @xmath163 .
the quantity @xmath164 is small for all @xmath156 .
the meaning of small here is that the quantity @xmath165 is smaller than @xmath166 . here
@xmath61 and @xmath167 are small positive numbers and @xmath151 is the number of partitions @xmath168 . for more details see@xcite and english translation of the original von neumann s article@xcite .
this is qet .
this property should be better called `` normality '' after gltz .
gltz also give a stronger bound on the deviations from the average@xcite .
the notion of `` normality '' is a special case of a broader notion of `` typicality '' .
there are various kinds of normality properties as discussed in the next subsection .
strict determinism of classical mechanics is replaced by the notion of `` typicality '' of statistical mechanical systems .
the notion of typicality captures the _ almost _ deterministic behaviour of statistical mechanical systems ( although atypical behaviour is possible in principle but highly improbable ) . 1 .
qet is a typicality result ( typicality due to `` scale separation''(see dnt below ) ) valid for large systems ( i.e. , containing large number of particles ) but eth ( again a typicality result
typicality due to quantum chaos ) has been applied even for few boson atoms in an optical lattice .
eth involves quantum chaos and qet does not involve quantum chaos !
eth has been proved ( analytically ) only in special circumstances based on the assumption of quantum chaos@xcite , see figure(5 ) .
general theory is seriously lacking .
recent literature can be organized under the following four notions ( see figure 6 ) : qualitative definition : we know that a small system weakly coupled to a large bath is described by canonical ensemble when the composite system ( system + bath ) is described by microcanonical ensemble .
kinematical canonical typicality says that even if the state of a composite system is pure ( rather than a mixture as in the traditional case ) the reduced density matrix of the system is canonical .
kct has been announced in@xcite extending the works of schrdinger@xcite .
related works appeared in@xcite . in @xcite
the basic assumption is that the probability distribution is uniform over all normalized wavefunctions @xmath169 with energy in the shell @xmath170 $ ] of the composite system .
the system density matrix ( when the composite is in pure state ) @xmath171 is obtained by tracing out bath degrees of freedom . the crucial role in the tracing
is played by the expansion co - efficients @xmath172 in @xmath173 which were assumed gaussian random variables .
it is shown that @xmath174 is approximately canonical ( see for details@xcite ) . a rather general treatment of kct
is given in@xcite .
they prove kct using levy s lemma which is analogous to the law of large numbers .
they prove that @xmath175 , where @xmath176 is the trace distance @xmath177 between @xmath174 ( reduced density matrix of the system when the composite is in the pure state ) and @xmath178 ( the traditional ( when composite is in the mixed state ) reduced density matrix of the system ) .
the average @xmath179 is over the environment states with the standard unitarily invariant measure .
@xmath180 is the dimension of the system s hilbert space and @xmath181 is a measure of the effictive size of the environment @xmath182 ( @xmath183 is the dimension of the environment s hilbert space ) .
thus when bath or environment is much larger than the system i.e. , @xmath184 then @xmath185 and if @xmath186 is very small then @xmath174 and @xmath178 are close to each other , @xmath175 , which is a statement of kct ( for details and related results see@xcite ) . qualitative definition : when the composite system ( system + bath ) is in a pure state , the expectation value of an operator pertaining to the system , _ after sufficiently large time _ , will be almost equal to the canonical expectation value .
dct has been proved in@xcite and there were numerical experiments@xcite for its evidence . in @xcite dct
has been proved with an assumption of weak coupling between system and bath with @xmath187 ( @xmath188 is the minimum spacing between the energy levels of the system , @xmath189 is the magnitude of system - bath coupling , and @xmath190 is the maximum spacing between the energy levels of the bath ) .
also a hypothesis of `` equal weights for eigenstates '' is used which makes the energy expansion co - efficients small .
the statement of dct goes like this .
suppose that the composite - system is in the pure state @xmath191 at any time @xmath42 , the expectation value of an operator of the system is written as @xmath192
. then dct says , after sufficiently long time , @xmath193 . for proof see@xcite .
dct has been proved in a much more general setting in@xcite and also in@xcite using an assumption that matrix elements of the interaction hamiltonian has random phases and then constructing markovian master equations . in @xcite ,
the authors extend their previous kinematical results ( last subsection ) .
they prove that @xmath194{\rangle}_t \leq \frac{1}{2}\sqrt{\frac{d_s}{d^{eff}(\omega_b ) } } \leq \frac{1}{2}\sqrt{\frac{d_s^2}{d^{eff}(\omega ) } } $ ] . here
$ ] is the system density matix ( @xmath196 is the total density matrix ( system + bath ) ) and @xmath197 . in words
, theorem says that the time average of the trace distance ( see previous subsection ) between instantaneous system s density matrix and its time averaged density matrix is small if @xmath198 .
this means that system spends almost all of its time near to @xmath199 .
this is called equilibration ( weakly fluctuating about a steady state this steady state
need not be an equilibrium state ) .
the important point is that this result is completely general ( except the assumption that the hamiltonian is free of resonances condition no ( 1 ) in von neumann s theorem ) .
it does not depend upon the nature of system - bath coupling , condition of bath ( whether it is in equilibrium or not ) .
they also treat the problem of thermalization ( i.e. , initial state independence , as the steady state reached may depend on the initial conditions ) . for a detailed account and further results ,
see their very nice paper@xcite . our purpose here is to introduce various typicality theorems and to give the excitement to the reader .
interested reader should go to the original literature .
qualitative definition : for an isolated quantum system in an arbitrary pure state , the quantum expectation value of an operator in that state _ hardly _ deviates from the ensemble average .
the meaning of `` hardly '' is clarified below .
a clear account of knt is given in@xcite .
the statement of knt theorem there goes like this ; consider that the system is in some pure state @xmath200 , with @xmath201 as eigenstates .
the quantity @xmath202 is small . here
the average @xmath203 , where @xmath204 is the probability distribution over the expansion co - efficients .
the meaning of `` small '' in the above statement is that , @xmath205 , here @xmath206 s are positive numbers typically of order one , and @xmath207 is the difference between the maximum and minimum eigenvalues of @xmath208 . for @xmath28 operators
one has @xmath209 \leq \frac{k ( max_{n } q_n ) ( tr \rho^2)}{{\epsilon}}$ ] . where for @xmath28 operators @xmath210 runs from @xmath62 to @xmath28 , and @xmath211 is the difference between the maximum and minimum eigenvalues of @xmath212 .
the smallness of @xmath213 ( which is must for typicality ) is imposed due to the conditions involved in the above theorem ; 1 .
the expansion co - efficients @xmath214 are statistically independent and @xmath215 are equally likely for arbitrary phase @xmath216 .
thus @xmath217 .
2 . the mixed state @xmath218 , with ( overbar means average over @xmath219 has low purity @xmath220 . the author@xcite justify these assumptions in a qualitative way that these are valid in `` practice '' if not `` in principle '' for a system with large number of degrees of freedom .
the rigorous justification is serious lacking .
the role of quantum chaos in these theorems is not clear at present .
knt is also discussed in @xcite where it is termed as `` thermodynamic normality '' the von neumann s quantum ergodic theorem is the first known theorem of dynamical normal typicality .
this theorem is based upon the observer s inability to further resolve @xmath126 by macroscopic means and the bound on fluctuations of the quantum expectation values of the operators from microcanonical averaging due to the structure imposed on the hilbert space by macroscopic perspective ( the large @xmath221 and large @xmath222 , see subsection b ) .
important point is that _ no _ chaos or disorder assumption was invoked by neumann . in his own words@xcite
`` ... we emphasize that the true state ( about which we do calculations ) is a wavefunction , i.e. , something microscopic to introduce a macroscopic description of the state would mean to introduce disorder assumptions , which is what we definitely want to avoid '' . _ his work is a result of `` scale separation '' ( i.e. , observer s inability to further resolve @xmath126 by macroscopic means , the large @xmath221 and large @xmath222 etc)_. spirit of his work resembles with that of khinchin and more recent works@xcite , in contrast to eth which involves quantum chaos .
this state of affairs ( un - necessity or necessity of disorder ) make the situation complicated and no satisfactory resolution is available at present .
we again consider the role of quantum chaos ( previous example was eth ) . the important work in this direction ( quantum chaos in the foundations of statistical mechanics ) was done by asher peres in 1980s@xcite .
he discusses dnt based upon his definition of quantum chaos that the observables are represented by pseudorandom matrices when the hamiltonian is diagonal .
due to this , expectation values of these observables tend to equilibrium values and fluctuations around these equilibrium values are , on the average , very small .
his argument for quantum ergodicity goes like this ; the quantum expectation value of an observable is @xmath223 for non - degenerate spectrum , time average is @xmath224 ( here @xmath225 is the density matrix in energy representation ) . here comes the peres argument ; consider that the system has very many energy levels in the range @xmath129 to @xmath226 , ( 1 ) distribution of these energy levels is very very different in regular and chaotic case@xcite , ( 2 ) the observable @xmath227 too behave very differently for regular and chaotic case ( regular system have selection rules , most @xmath227 vanish and only a few are large , and in chaotic case @xmath227 is pseudorandom ( his definition of quantum chaos ) and very numerous ) , and ( 3 ) for a chaotic system , the pseudorandom @xmath227 are _ statistically independent _ from @xmath225 in that energy range and _ their average does not appreciably depend on the energy interval @xmath228_. as @xmath229 , one has @xmath230 .
this he calls quantum ergodicity .
the fundamental assumption used is the assumption of statistical independence of @xmath227 from @xmath225 .
then he discuss more relevant property , called mixing in quantum case , that the fluctuations of @xmath231 about its equilibrium value @xmath232 are , on the average , small i.e. , @xmath233 ( @xmath234 is the large number of energy levels in the interval @xmath129 to @xmath130 ) . to prove this he again used the assumption of statistical independence .
thus we see from peres program that dnt is a consequence of quantum chaos ( observables as pseudorandom matrices ) and _ the assumption of statistical independence .
_ this is in sharp contrast with the approach of von neumann ( as von neumann did not use quantum chaos ) , still we do not have a coherent picture that when chaos is important ( peres ) and when chaos is not important ( von neumann ) ?
various typicality approaches are summarized in figure ( 6 ) .
1 . justification of gibbs ensembles is a complicated problem .
2 . ergodic hypothesis is not necessary for the workings of statistical mechanics as far one is concerned with sum - function observables and macroscopic systems .
statistical mechanics explain thermodynamic behaviour because of the property of statistical independence , or , more accurate approach of khinchin which is the cause of typicality and canonical formalism can be obtained from landau s program without using any ad - hoc hypothesis except the property of statistical independence .
the issues of chaos , integrability or non - integrability becomes important when one deals with small isolated systems , for example , @xmath235 or so bosonic or fermionic atoms in an optical lattice .
they are irrelevant for a macroscopic system and sum - function observables .
clearly von neumann s quantum ergodic theorem and other quantum - statistical typicality theorems are analogous to khinchin s self averaging theorem .
these become more accurate for a system with large number of degrees of freedom and these say that for such systems the expectation values of the observables ( sum - functions in khinchin s case ) stay close to the microcanoncal / canonical predictions . in principle , khinchin s approach should come out as a special case of von neumann s theory .
clearly much work is needed in this direction .
one can contemplate _ classical _ kct ( this traditional statistical mechanics ) , _ classical _ dct , and so on . 7 .
the above presented consolidation or amalgamation of various quantum - statistical typicality theorems singles out eth and peres dnt .
eth and peres dnt involves quantum chaos , whereas others like von neumann s do not involve quantum chaos .
still we do not have a coherent picture ; when chaos is important ( peres ) and when chaos is not important ( von neumann ) ?
i would like to end this manuscript with the following aphoristic words of prof .
jeol lebowitz about the ergodic hypothesis ; @xmath0``now it is the real time to recognize that the ergodic hypothesis is not a necessary and is not a sufficient condition for the foundation of statistical mechanics '' author would like to thank to prof .
sheldon goldstein and prof .
marcos rigol for many critical discussions , to prof .
angelo vulpiani and dr .
jitesh bhatt for pointing out many corrections and suggestions .
he is also thankful to prof .
michael kastner for giving him the opportunity to present this work at 2nd stellenbosch workshop on statistical physics from 7th march to 18th march 2011 at stellenbosch , south africa , and also for the partial financial help .
boltzmann introduced the notion of `` komplexions '' of molecules as follows .
consider the kinetic energies of the molecules take the following discrete values @xmath237 , and let @xmath238 number of molecules posses a kinetic energy @xmath239 .
number of komplexions for a given distribution is @xmath241 .
he noted that maximization of @xmath242 with the above constraints leads to maxwellian kinetic energy distribution .
@xmath0boltzmann then defines `` degree of permutability '' @xmath243 , and finds that ( by direct computation ) that @xmath244 is equal to the entropy of an ideal gas in a reversible process up to an additive constant .
@xmath245.\ ] ] it was max planck who gave the general foundations to the fundamental principle @xmath246 , ( using @xmath247 in place of @xmath242 ) introducing first time the constant @xmath21 called the boltzmann constant .
he proposed the following fundamental hypothesis called planck s thermodynamic probability hypothesis .
`` the entropy of physical system in a definite state depends solely on the thermodynamic probability @xmath247 of this state '' , @xmath248 .
@xmath0considering two systems in equilibrium with each other . from second law ,
one has additivity of entropy @xmath249 , thus @xmath250 and then by differentiating one can obtain @xmath251 .
it gives @xmath252 the important point here is that the equilibrium is conceived as the most probable state rather than the temporally constant state .
but as we know that the fluctuations in observables are extremely small for macroscopic systems both definitions are approximately consistent with each other .
goldstein s. , lebowitz j. , tumulka r. , zanghi n. , long - time behaviour of macroscopic quantum systems : commentary accompanying the english translation of john von neumann s 1929 article on the quantum ergodic theorem " john von neumann , `` proof of the ergodic theorem and the h - theorem in quantum mechanics '' , eur .
j. h * 35 * : 201 - 237 ( 2010 ) , arxiv:1003.2133 , [ english translation by roderich tumulka of `` beweis des ergodensatzes und des h - theorems '' in zeitschrift fuer physik * 57 * : 30 - 70 ( 1929 ) ] .
sheldon goldstein , joel l. lebowitz , christian mastrodonato , roderich tumulka , nino zanghi , `` normal typicality and von neumann s quantum ergodic theorem '' , proceedings of the royal society a * 466 * , 3203 ( 2010 ) .
r. v. jensen and r. shanker , `` statistical behavior in deterministic quantum systems with few degrees of freedom '' , phys .
lett . , * 54 * , 1879 ( 1985 ) ; k. saito , s. takesue , and s. miyashita , `` system - size dependence of statistical behavior in quantum system '' , j. phys .
. jpn . * 65 * , 1243 ( 1996 ) ; `` thermal conduction in a quantum system '' , phys .
e * 54 * , 2404 ( 1996 )
. a version of dnt ( equilibration as opposed to thermalization ) has been proved in@xcite .
here it is shown that the time average of the difference in the instantaneous expectation values of an observable and its ensemble average is very small .
ensemble here is the diagonal ensemble ( equation 11 ) . in literature
it is called equilibration@xcite .
the agreement of instantaneous expectation value of an observable with microcanonical or canonical ensemble is called thermalization which is a much more difficult problem and possibly involves chaos .
hal tasaki , `` the approach to thermal equilibrium and ' ' thermodynamic normality``an observation based on the works by goldstein , lebowitz , mastrodonato , tumulka , and zanghi in 2009 , and by von neumann in 1929 '' , arxiv:1003.5424 .
r. balian and c. bloch , `` distribution of eigenfrequencies for the wave equation in a finite domain : iii .
eigenfrequency density oscillations'',ann .
phys.(ny ) * 69 * , 76 ( 1972 ) ; `` solution of the schrdinger equation in terms of classical paths '' , bf 85 , 514 ( 1974 ) . | as the title says we want to answer the question ; how and why does statistical mechanics work ? as we know from the most used prescription of gibbs
we calculate the phase space averages of dynamical quantities and we find that these phase averages agree very well with experiments . clearly actual experiments are not done on a hypothetical ensemble they are done on the actual system in the laboratory and these experiments take a finite amount of time .
thus it is usually argued that actual measurements are time averages and they are equal to phase averages due to ergodicity .
aim of the present review is to show that ergodicity is not relevant for equilibrium statistical mechanics ( with tolman and landau ) .
we will see that the solution of the problem is in the very peculiar nature of the macroscopic observables and with the very large number of the degrees of freedom involved in macroscopic systems as first pointed out by khinchin .
similar arguments are used by landau based upon the approximate property of `` statistical independence '' .
we review these ideas in detail and in some cases present a critique .
we review the role of chaos ( classical and quantum ) where it is important and where it is not important .
we criticise the ideas of e. t. jaynes who says that the ergodic problem is conceptual one and is related to the very concept of ensemble itself which is a by - product of frequency theory of probability , and the ergodic problem becomes irrelevant when the probabilities of various micro - states are interpreted with laplace - bernoulli theory of probability ( bayesian viewpoint ) . in the end
we critically review various quantum approaches ( quantum - statistical typicality approaches ) to the foundations of statistical mechanics .
the literature on quantum - statistical typicality is organized under four notions ( 1 ) kinematical canonical typicality , ( 2 ) dynamical canonical typicality , ( 3 ) kinematical normal typicality , and ( 4 ) dynamical normal typicality .
analogies are seen in the khinchin s classical approach and in the modern quantum - statistical typicality approaches .
`` unless the conceptual problems of a field have been clearly resolved , you can not say which mathematical problems are the relevant ones worth working on , and your efforts are more than likely to be wasted '' -
e. t. jaynes . |
motivated by the magnetic confinement fusion , which is one of the main application in plasma physics , we analyse the dynamics of a population of charged particles , under the action of a strong uniform magnetic field .
the goal of this note is to study the finite larmor radius regime , that is , we assume that the particle distribution fluctuates at the larmor radius scale along the orthogonal directions , with respect to the magnetic field @xcite , @xcite , @xcite . to simplify , we consider the two dimensional setting , _
@xmath0 , with a magnetic field orthogonal to @xmath1 .
we chose a regime such that 1 .
the reference time @xmath2 is much larger than the cyclotronic period ( strong magnetic field ) _
@xmath3 notice that the above hypothesis writes also @xmath4 , where @xmath5 is the reference velocity ( along the orthogonal directions ) , and @xmath6 is the typical larmor radius .
the kinetic energy is much larger than the potential energy @xmath7 where @xmath8 is the particle mass , @xmath9 is the particle charge , and @xmath10 is the reference electric potential .
the larmor radius is of the same order as the debye length _
i.e. , _ @xmath11 here @xmath12 is the electric permittivity of the vacuum and @xmath13 is the charge concentration . accordingly , the presence density @xmath14 and the electric potential @xmath15 satisfy the following vlasov - poisson system , up to some multiplicative constants , of order one @xmath16 @xmath17 @xmath18 here @xmath19 stands for the rescaled cyclotronic frequency ( the real cyclotronic frequency is @xmath20 , and the real cyclotronic period is @xmath21 ) , and it is assumed constant ( uniform magnetic field ) . for any @xmath22 , we denote by @xmath23 the vector @xmath24 .
we study the stability of the family @xmath25 , when @xmath26 becomes small .
the asymptotic behavior follows by filtering out the fast oscillations of the caracteristic equations for .
it is easily seen that the changes over one cyclotronic period of the quantities @xmath27 , @xmath28 , are negligible .
we expect that the family @xmath29 converges , as @xmath26 becomes small , toward some profile @xmath30 .
[ mainresult1 ] let @xmath31 be a non negative , smooth , compactly supported presence density and @xmath25 be the solutions of the vlasov - poisson sistem , , .
we denote by @xmath32 the solution of @xmath33 ( { \tilde{x } } , { \tilde{v } } ) \cdot \nabla _ { { \tilde{x } } } { \tilde{f}}+ { \mathcal a } [ { \tilde{f}}(t ) ] ( { \tilde{x } } , { \tilde{v } } ) \cdot \nabla _ { { \tilde{v } } } { \tilde{f}}= 0,\;\;(t , { \tilde{x } } , { \tilde{v } } ) \in { { \mathbb r}}_+ \times { { \mathbb r}}^2 \times { { \mathbb r}}^2\ ] ] with the initial condition @xmath34 where the velocity and acceleration vector fields @xmath35 are given by @xmath36 ( { \tilde{x } } , { \tilde{v } } ) = - \frac{^\bot \nabla _ { { \tilde{x } } } } { { \omega _ c } } { \tilde{\phi } } [ { \tilde{f}}(t)],\;\;{\mathcal a } [ { \tilde{f}}(t ) ] ( { \tilde{x } } , { \tilde{v } } ) = { \omega _ c}\;^\bot \nabla _ { { \tilde{v } } } { \tilde{\phi } } [ { \tilde{f}}(t)]\ ] ] @xmath37 ( { \tilde{x } } , { \tilde{v } } ) = - \frac{1}{2\pi } { \int_{{{\mathbb r}}^2}\!\ ! \int _ { { { \mathbb r}}^2 } \left \{\ln \frac{| { \tilde{v}}- { \tilde{w}}|}{| { \omega _ c}| } \ ; { { \bf 1}_{\{| { \tilde{x}}- { \tilde{y}}| \leq \frac{| { \tilde{v}}- { \tilde{w}}|}{| { \omega _ c}| } \ } } } + \ln | { \tilde{x}}- { \tilde{y}}|\ ; { { \bf 1}_{\{| { \tilde{x}}- { \tilde{y}}| > \frac{| { \tilde{v}}- { \tilde{w}}|}{| { \omega _ c}| } \}}}\right \ } { \tilde{f}}(t , { \tilde{y } } , { \tilde{w } } ) \;\mathrm{d}\tilde{w}\mathrm{d}\tilde{y}}.\ ] ] therefore @xmath38 when @xmath39 .
ce travail sinscrit dans le cadre de la modlisation des plasmas de fusion .
nous concentrons notre tude au rgime du rayon de larmor fini pour le systme de vlasov - poisson bi - dimensionnel .
sous les hypothses , , , ce rgime est dcrit par , , .
la mthode dveloppe ici consiste exprimer le potentiel lectrique laide de la solution fondamentale de loprateur de laplace dans @xmath40 , puis insrer cette expression dans les trajectoires de lquation de vlasov .
nous obtenons alors , laide des mthodes classiques de gyro - moyenne @xcite , @xcite , @xcite les trajectoires limites et ainsi , les expressions effectives des champs vitesse et acclration de la nouvelle quation de vlasov , dcrivant le rgime asymptotique considr . pour plus de dtails sur les preuves de ces rsultats , nous renvoyons @xcite .
notons @xmath41 la solution fondamentale de loprateur de laplace dans @xmath40 @xmath42 cest - - dire @xmath43 dans @xmath44 .
le potentiel lectrique , solution de lquation de poisson , scrit donc @xmath45 les quations caractristiques de lquation de transport sont donnes par @xmath46 nous cherchons des quantits qui varient peu sur une priode cyclotronique . plus
exactement , tout instant fix @xmath47 , on introduit le changement de coordonnes @xmath48 o @xmath49 dsigne la rotation de @xmath40 dangle @xmath50 . on vrifie aisment que le dterminant jacobien vaut @xmath51 et alors ces transformations prservent la mesure de lebesgue de @xmath52 _ i.e. , _ @xmath53 .
en effet , @xmath54 est le centre du cercle de larmor dcrit par une particule passant par @xmath55 avec la vitesse @xmath56 .
ce centre ne varie pas lchelle du mouvement rapide cyclotronique , correspondant la frquence cyclotronique @xmath57 .
plus exactement on a @xmath58 @xmath59 on souhaite remplacer le potentiel lectrique par lxpression de .
on introduit les densits de prsence en les coordonnes @xmath60 @xmath61 ainsi , conduit @xmath62 et par consquent , , deviennent @xmath63 @xmath64 en prenant la moyenne de sur la priode cyclotronique @xmath65 $ ] , avec @xmath66 , et en introduisant la variable rapide @xmath67 , @xmath68 $ ] , nous obtenons @xmath69 o la fonction @xmath70 est dfinie par @xmath71 nous procdons de la manire identique pour obtenir , partir de @xmath72 en passant la limite dans , , quand @xmath39 , nous obtenons les trajectoires aprs filtration du mouvement rapide cyclotronique @xmath73 o @xmath74 est la distribution limite . par la suite nous dterminons une expression pour @xmath75 .
cela rsulte de la proprit de la moyenne pour les fonctions harmoniques .
en effet , si @xmath76 , la fonction @xmath77 est harmonique dans louvert @xmath78 , contenant le disque ferm , de centre @xmath79 et de rayon @xmath80 , et par consquent nous avons , grce la formule de la moyenne @xmath81 plus exactement , on dmontre cf .
@xcite [ calcint ] pour tout @xmath82 , nous avons @xmath83 @xmath84
nous introduisons le potentiel lectrique @xmath85 ( { \tilde{x } } , { \tilde{v } } ) & = { \int_{{{\mathbb r}}^2}\!\ ! \int _ { { { \mathbb r}}^2 } { { \mathcal e } } ( { \tilde{x}}- { \tilde{y } } , { \tilde{v}}- { \tilde{w } } ) { \tilde{f}}(t , { \tilde{y } } , { \tilde{w } } ) \;\mathrm{d}\tilde{w}\mathrm{d}\tilde{y } } \\ & = { \int_{{{\mathbb r}}^2}\!\ ! \int _ { { { \mathbb r}}^2 } \left \ { e \left ( \frac { { \tilde{v}}- { \tilde{w } } } { { \omega _ c}}\right ) { { \bf 1}_{\{| { \tilde{x}}- { \tilde{y}}| \leq | { \tilde{v}}- { \tilde{w}}|/| { \omega _ c}|\ } } } + e ( { \tilde{x}}- { \tilde{y } } ) \ ; { { \bf 1}_{\{| { \tilde{x}}- { \tilde{y}}| > | { \tilde{v}}- { \tilde{w}}|/| { \omega _ c}|\ } } } \right \ } { \tilde{f}}(t , { \tilde{y } } , { \tilde{w } } ) \;\mathrm{d}\tilde{w}\mathrm{d}\tilde{y } } \nonumber \end{aligned}\ ] ] et les fonctions @xmath86 ( { \tilde{x } } , { \tilde{v } } ) = - \frac{^\bot \nabla _ \xi } { { \omega _ c } } { \int_{{{\mathbb r}}^2}\!\ ! \int _
{ { { \mathbb r}}^2 } { { \mathcal e } } ( { \tilde{x}}- { \tilde{y } } , { \tilde{v}}- { \tilde{w } } ) { \tilde{f}}(t , { \tilde{y } } , { \tilde{w } } ) \;\mathrm{d}\tilde{w}\mathrm{d}\tilde{y } } = - \frac{^\bot \nabla _ { { \tilde{x } } } } { { \omega _ c } } { \tilde{\phi } } [ { \tilde{f}}(t)]\ ] ] @xmath87 ( { \tilde{x } } , { \tilde{v } } ) = { \omega _ c}{^\bot \nabla _ \eta } { \int_{{{\mathbb r}}^2}\!\ ! \int _ { { { \mathbb r}}^2 } { { \mathcal e } } ( { \tilde{x}}- { \tilde{y } } , { \tilde{v}}- { \tilde{w } } ) { \tilde{f}}(t , { \tilde{y } } , { \tilde{w } } ) \;\mathrm{d}\tilde{w}\mathrm{d}\tilde{y } } = { \omega _ c}{^\bot \nabla _ { { \tilde{v } } } } { \tilde{\phi } } [ { \tilde{f}}(t)].\ ] ] en drivant sous le signe intgral , il est galement possible de reprsenter les champs vitesse et acclration sous la forme ( cf .
lemme [ calcint ] ) @xmath88 ( { \tilde{x } } , { \tilde{v } } ) & = - \frac{1 } { { \omega _ c } } { \int_{{{\mathbb r}}^2}\!\ ! \int _ { { { \mathbb r}}^2 } { ^\bot \nabla _ \xi } { { \mathcal e } } ( { \tilde{x}}- { \tilde{y } } , { \tilde{v}}- { \tilde{w } } ) { \tilde{f}}(t , { \tilde{y } } , { \tilde{w } } ) \;\mathrm{d}\tilde{w}\mathrm{d}\tilde{y } } \\ & = - \frac{1 } { { \omega _ c } } { \int_{{{\mathbb r}}^2}\!\ ! \int _ { { { \mathbb r}}^2 } { ^\bot \nabla e } ( { \tilde{x}}- { \tilde{y } } ) \ ; { { \bf 1}_{\{| { \tilde{x}}- { \tilde{y}}| > | { \tilde{v}}- { \tilde{w}}|/| { \omega _ c}|\ } } } { \tilde{f}}(t , { \tilde{y } } , { \tilde{w } } ) \;\mathrm{d}\tilde{w}\mathrm{d}\tilde{y } } \nonumber \end{aligned}\ ] ] @xmath89 ( { \tilde{x } } , { \tilde{v } } ) & = { \omega _ c } { \int_{{{\mathbb r}}^2}\!\ ! \int _ { { { \mathbb r}}^2 } { ^\bot \nabla _ \eta } { { \mathcal e } } ( { \tilde{x}}- { \tilde{y } } , { \tilde{v}}- { \tilde{w } } ) { \tilde{f}}(t , { \tilde{y } } , { \tilde{w } } ) \;\mathrm{d}\tilde{w}\mathrm{d}\tilde{y } } \\ & = { \omega _ c } { \int_{{{\mathbb r}}^2}\!\ ! \int _ { { { \mathbb r}}^2 } { ^\bot \nabla e } \left ( \frac { { \tilde{v}}- { \tilde{w } } } { { \omega _ c}}\right ) { { \bf 1}_{\{| { \tilde{x}}- { \tilde{y}}| \leq | { \tilde{v}}- { \tilde{w}}|/| { \omega _ c}|\ } } } { \tilde{f}}(t , { \tilde{y } } , { \tilde{w } } ) \;\mathrm{d}\tilde{w}\mathrm{d}\tilde{y}}. \nonumber \end{aligned}\ ] ] les trajectoires limites sont dtermines par les champs vitesse et acclration @xmath90 , { \mathcal a } [ { \tilde{f}}]$ ] @xmath91 ( { \tilde{x}}(t ) , { \tilde{v}}(t)),\;\;\frac { { \mathrm{d } } { \tilde{v } } } { { \mathrm{d}}t } = { \mathcal a } [ { \tilde{f}}(t ) ] ( { \tilde{x}}(t ) , { \tilde{v}}(t)).\ ] ] les densits de prsence tant conserves le long des trajectoires , nous obtenons @xmath92 et par consquent la densit limite @xmath93 est solution de , . notons que les quations caractristiques limites forment un systme hamiltonien , en les variables conjugues @xmath94 et @xmath95 @xmath96}{\partial ( { \omega _ c } { \tilde{x}}_1 ) } ( { \tilde{x}}(t ) , { \tilde{v}}(t)),\;\;\frac { { \mathrm{d } } ( { \omega _ c}^{-1 } { \tilde{v}}_1 ) } { { \mathrm{d}}t } = \frac{\partial { \tilde{\phi } } [ { \tilde{f}}(t)]}{\partial { \tilde{v}}_2 } ( { \tilde{x}}(t ) , { \tilde{v}}(t))\ ] ] @xmath97}{\partial { \tilde{x}}_2 } ( { \tilde{x}}(t ) , { \tilde{v}}(t)),\;\;\frac { { \mathrm{d } } { \tilde{v}}_2 } { { \mathrm{d}}t } = -\frac{\partial { \tilde{\phi } } [ { \tilde{f}}(t)]}{\partial ( { \omega _ c}^{-1 } { \tilde{v}}_1 ) } ( { \tilde{x}}(t ) , { \tilde{v}}(t)).\ ] ]
les champs de vitesse et acclration tant divergence nulle @xmath98 = - \frac{1 } { { \omega _ c } } { \mathrm{div } } _ { { \tilde{x } } } \;^\bot \nabla _ { { \tilde{x } } } { \tilde{\phi } } [ { \tilde{f}}(t ) ] = 0,\;\ ; { \mathrm{div } } _ { { \tilde{v } } } { \mathcal a } [ { \tilde{f}}(t ) ] = { \omega _ c } { \mathrm{div } } _ { { \tilde{v } } } \;^\bot \nabla _ { { \tilde{v } } } { \tilde{\phi } } [ { \tilde{f}}(t ) ] = 0\ ] ] lquation scrit aussi sous la forme conservative @xmath99\ } + { \mathrm{div } } _ { { \tilde{v } } } \ { { \tilde{f}}{\mathcal a } [ { \tilde{f}}(t)]\ } = 0.\ ] ] en particulier nous obtenons la conservation de la masse .
plus gnralement , nous dmontrons le rsultat suivant . 1 .
pour tout @xmath103 nous avons @xmath104 \ ; { \mathrm{d } } { \tilde{w } } { \mathrm{d } } { \tilde{y } } { \mathrm{d } } { \tilde{v } } { \mathrm{d } } { \tilde{x}}. \nonumber \end{aligned}\ ] ] 2 . en particulier ,
pour tout @xmath103 nous avons @xmath105 1 .
pour tout @xmath103 , nous obtenons , grce aux formules de reprsentation , @xmath106 + \nabla _ { { \tilde{v}}}\psi \cdot { \mathcal a } [ { \tilde{f}}(t ) ] \right ] { \tilde{f}}(t , { \tilde{x } } , { \tilde{v } } ) \;\mathrm{d}\tilde{v}\mathrm{d}\tilde{x } } \\ & = - \frac{1 } { { \omega _ c } } \int _ { { { \mathbb r}}^2}\!\!\int _ { { { \mathbb r}}^2}\!\!\int _ { { { \mathbb r}}^2}\!\!\int _ { { { \mathbb r}}^2}\!\!\nabla _ { { \tilde{x } } } \psi ( { \tilde{x } } , { \tilde{v } } ) \cdot { ^\bot \nabla } e ( { \tilde{x}}- { \tilde{y } } ) \ ; { { \bf 1}_{\{| { \tilde{x}}- { \tilde{y}}| > | { \tilde{v}}- { \tilde{w}}|/| { \omega _ c}|\ } } } { \tilde{f}}(t , { \tilde{y } } , { \tilde{w } } ) { \tilde{f}}(t , { \tilde{x } } , { \tilde{v } } ) \ ; { \mathrm{d } } { \tilde{w } } { \mathrm{d } } { \tilde{y } } { \mathrm{d } } { \tilde{v } } { \mathrm{d } } { \tilde{x}}\\ & + \;\;\;\;\int _ { { { \mathbb r}}^2}\!\!\int _ { { { \mathbb r}}^2}\!\!\int _ { { { \mathbb r}}^2}\!\!\int _ { { { \mathbb r}}^2}\!\!\nabla _ { { \tilde{v } } } \psi ( { \tilde{x } } , { \tilde{v } } ) \cdot { ^\bot \nabla } e \left(\frac { { \tilde{v}}- { \tilde{w } } } { { \omega _ c}}\right ) { { \bf 1}_{\{| { \tilde{x}}- { \tilde{y}}| < | { \tilde{v}}- { \tilde{w}}|/| { \omega _ c}|\ } } } { \tilde{f}}(t , { \tilde{y } } , { \tilde{w } } ) { \tilde{f}}(t , { \tilde{x } } , { \tilde{v } } ) \ ; { \mathrm{d } } { \tilde{w } } { \mathrm{d } } { \tilde{y } } { \mathrm{d } } { \tilde{v } } { \mathrm{d } } { \tilde{x}}.\end{aligned}\ ] ] la formule rsulte en interchangeant @xmath60 contre @xmath107 , combin fubini .
2 . les conservations rsultent facilement , par applique successivement aux fonctions @xmath108 .
etant donn que lnergie cintique est conserve , et comme on sattend ce que lnergie globale soit conserve , nous devrions retrouver aussi la conservation de lnergie lectrique .
effectivement nous dmontrons lnergie lectrique scrit sous la forme @xmath110 ( { \tilde{x } } , { \tilde{v } } ) { \tilde{f}}(t , { \tilde{x } } , { \tilde{v } } ) \;\mathrm{d}\tilde{v}\mathrm{d}\tilde{x } } = \frac{1}{2}\int _ { { { \mathbb r}}^2}\!\!\int _ { { { \mathbb r}}^2}\!\!\int _ { { { \mathbb r}}^2}\!\!\int _ { { { \mathbb r}}^2}\!\ ! { { \mathcal e } } ( { \tilde{x}}- { \tilde{y } } , { \tilde{v}}- { \tilde{w } } ) { \tilde{f}}(t , { \tilde{x } } , { \tilde{v } } ) { \tilde{f}}(t , { \tilde{y } } , { \tilde{w}})\ ; { \mathrm{d } } { \tilde{w } } { \mathrm{d } } { \tilde{y } } { \mathrm{d } } { \tilde{v } } { \mathrm{d } } { \tilde{x}}\ ] ] et en utilisant la parit de @xmath111 en les variables @xmath79 et @xmath112 , nous obtenons facilement , par fubini , que @xmath113 ( { \tilde{x } } , { \tilde{v } } ) { \tilde{f}}(t , { \tilde{x } } , { \tilde{v } } ) \;\mathrm{d}\tilde{v}\mathrm{d}\tilde{x } } & = { \int_{{{\mathbb r}}^2}\!\ ! \int _ { { { \mathbb r}}^2 } { \tilde{\phi } } [ { \tilde{f}}(t ) ] ( { \tilde{x } } , { \tilde{v } } ) \partial _ t { \tilde{f}}\;\mathrm{d}\tilde{v}\mathrm{d}\tilde{x}}\\ & = { \int_{{{\mathbb r}}^2}\!\ ! \int _ { { { \mathbb r}}^2 } \left [ \nabla _ { { \tilde{x } } } { \tilde{\phi } } [ { \tilde{f}}(t ) ] \cdot { \mathcal v } [ { \tilde{f}}(t ) ] + \nabla _ { { \tilde{v } } } { \tilde{\phi } } [ { \tilde{f}}(t ) ] \cdot { \mathcal a } [ { \tilde{f}}(t ) ] \right ] { \tilde{f}}\;\mathrm{d}\tilde{v}\mathrm{d}\tilde{x } } = 0.\end{aligned}\ ] ] | nous tudions le rgime du rayon de larmor fini pour le systme de vlasov - poisson , dans le cas o la longueur de debye est gale au rayon de larmor .
le champ magntique est suppos uniforme .
nous restreignons ltude de ce problme non linaire au cas bi - dimensionnel .
nous obtenons le modle limite en appliquant les mthodes de gyro - moyenne cf .
@xcite , @xcite .
nous donnons lexpression explicite du champ dadvection effectif de lquation de vlasov , dans laquelle nous avons substitu le champ lectrique auto - consistant , via la rsolution de lquation de poisson moyenne lchelle cyclotronique .
nous mettons en vidence la structure hamiltonienne du modle limite et prsentons ses proprits : conservations de la masse , de lnergie cintique , de lnergie lectrique , etc . 0.5 * abstract * 0.5 we study the finite larmor radius regime for the vlasov - poisson system .
the magnetic field is assumed to be uniform .
we investigate this non linear problem in the two dimensional setting .
we derive the limit model by appealing to gyro - average methods cf .
@xcite , @xcite .
we indicate the explicit expression of the effective advection field , entering the vlasov equation , after substituting the self - consistent electric field , obtained by the resolution of the averaged ( with respect to the cyclotronic time scale ) poisson equation .
we emphasize the hamiltonian structure of the limit model and present its properties : conservationss of the mass , kinetic energy , electric energy , etc . |
it is generally accepted from energy budget considerations that supernova remnants ( snrs ) are the primary source of galactic cosmic rays ( crs ) , which are inferred to span the energy range from about @xmath5 to @xmath6 ev ( blandford & ostriker 1978 ; axford 1994 ; biermann 1995 , and references therein ) . though radio observations of synchrotron radiation from supernova remnants supply bountiful evidence that they are the source of crs at gev energies , evidence regarding higher energy cosmic rays is largely circumstantial . a search for ultra - high energy @xmath7-rays ( @xmath010@xmath8 ev ) coincident with snrs detected nothing significant ( allen , et al .
the best evidence thus far for shock acceleration of @xmath0100 tev crs , comes from x - ray observations of sn1006 ( koyama et al . 1995 ) .
sn1006 appears to be unique in that the non - thermal x - ray flux from its shell dominates the thermal , but current limits on the non - thermal flux from the shells of other young snr do not exclude the presence of a synchrotron component , and hence ongoing cr acceleration to tev energies . in this paper we report the discovery of a site in another snr in which high energy cosmic rays are possibly being accelerated , but by a different mechanism from that in sn1006 . using _ asca _
gis data , we have found a concentration of hard x - ray emission along the southern rim of the middle - aged remnant ic443 , whose spectrum is consistent with a power law , and whose morphology suggests enhanced shock acceleration resulting from the snr shock encountering dense clouds in the ism .
the observations are consistent with the prediction of such an effect by jones & kang ( 1993 , hereafter jk93 ) .
ic443 ( g189.1 + 3.0 ) is a nearby ( @xmath9 kpc ) supernova remnant ( snr ) of intermediate age .
many infrared and radio line observations have demonstrated that ic443 is interacting with a ring of molecular clouds ( e.g. , denoyer 1979 ; burton _ et al . _ 1988 , hereafter bgbw ; burton _ et al . _
1990 ; dickman _ et al . _
1992 , hereafter dszh ; van dishoeck _ et al . _
the region in which the most complex interactions between the molecular cloud and the snr shock is occurring also has an unusually flat radio spectral index for a shell - like snr ( @xmath100.24 green 1986 , hereafter g86 ) . in the x - ray band
, ic443 has been the subject of a number of comprehensive studies , most notably those of petre _ et al . _ ( 1988 , hereafter pssw ) and asaoka & aschenbach ( 1994 , hereafter aa94 ) .
it has an irregular soft x - ray morphology , influenced strongly by its interactions with an h i cloud to the northeast and the foreground molecular cloud and shadowing by the foreground snr g189.6 + 3.3 . using _
, wang et al . ( 1992 , hereafter wahk ) resolved spectrally a hard component which they were able to characterize by either a thermal model with kt@xmath014 kev or a power law with photon index @xmath112.2 .
scanning observations by _
a-2 showed that the hard emission is more centrally located than the bright soft emission .
neither the a-2 nor the _ ginga _ observations provided information regarding the extent of the hard emission .
ic443 is also coincident with a high energy @xmath7-ray source , which has led to speculation that the @xmath7-rays are produced by the interaction of cosmic rays , accelerated by ic443 s shocks , with nearby molecular cloud material ( sturner & dermer 1995 ; esposito _ et al . _ 1996 ,
hereafter ehks ) .
models of broad - band non - thermal spectra have recently been produced by sturner et al .
( 1996 ) ; in these synchrotron radiation dominates from the radio to the soft x - ray , while electron bremsstrahlung and @xmath12 decay dominate the @xmath7-ray emission for a supernova remnant like ic443 .
our paper is organized as follows .
we first present the _ asca _
gis images showing the location of the hard x - ray emission region ( [ asca_maps.sec ] ) .
we compare the _ asca _ image with that from the _ rosat _ hri .
we then discuss our analysis of the gis spectrum and the gis and hri light curves from that region ( [ asca_spectra.sec ] ) .
finally , we discuss possible emission mechanisms and the implications for the acceleration of cosmic rays by supernova remnants ( [ discussion.sec ] ) .
+ + [ hbtp ] .asca observations of the snr ic443 [ obs_tab ] [ cols="<,^,^ " , ] + fits are to the energy range 0.512 kev + letting the scaling ( gis2 & gis3 ) , column density , temperature and line strengths vary + we froze the line strengths and temperature to the off source ( model # 1 ) values , but fit the column density and scaling .
+ the best - fit 5 kev flux density is @xmath13 .
+ we froze the column density and line strengths to the off source ( model # 1 ) values , but fit the temperature and scaling .
+ the hxf and the ridge are embedded within the diffuse emission of ic443 at least in projection .
thus in order to determine the spectral properties of these features , one must first adequately characterize the `` background '' thermal emission .
we therefore extracted a spectrum from the adjacent larger region shown in figure [ spectra_region ] .
we assume , based on the smoothness of the spectral hardness ratio map , that the spectrum of the diffuse emission does not vary appreciably in the neighborhood of the hxf ; based on the previous x - ray observations the spectral variations in the diffuse emission would be most significant below 1 kev , outside the effective gis band . to this `` background '' spectrum
we fit an _ ad hoc _ thermal emission model , comprised of a thermal bremsstrahlung continuum and narrow gaussians to represent the most prominent emission lines ( the he@xmath2 transitions of ne , mg , si , s and ar ) . such _ ad hoc _ models have been used previously to model _ asca _
spectra from other remnants ( e.g. miyata _ et al . _ 1994
; holt _ et al .
the best fit model yields [email protected] kev , comparable with published temperatures for ic443 ( pssw , wahk ) . however , it is important not to over - interpret our thermal results , because we have made no attempt to distinguish the he@xmath2 transitions from weaker nearby spectral lines ( including each element s corresponding h - like ly@xmath2 transition ) or to develop a physically self - consistent thermal model .
we next applied this same model to the gis spectrum of the region containing the hxf ( also shown in figure [ spectra_region ] ) , allowing only the normalization to vary .
as shown in table [ good_fit.tab ] , this yields an unacceptable fit . as indicated in fig.[spectra ] , the `` background '' model provides a good fit to the data below @xmath02 kev , but there is a substantial flux excess at higher energies .
an acceptable fit to the `` on hxf '' spectrum is obtained by adding a second continuum component .
the current data do not allow us to distinguish between a power law with @[email protected] and thermal bremsstrahlung with kt=@xmath15 kev .
we also performed this same analysis using a smaller on - source region , tightly drawn around the hxf .
however , this was less effective : because of the lack of thermal background counts , it is harder to spectrally distinguish the background from the non - thermal source emission .
nevertheless we measure a photon index of @xmath14=@xmath16 , using the technique described above .
we applied the same spectral analysis procedure as described above for the hxf to the hard ridge .
since the ridge is farther from the `` background '' region than the hxf ( see fig .
[ spectra_region ] ) and we fit the thermal model using the ao-1 data , we were uncertain if it was appropriate to use the same background thermal model for both the hxf and the ridge . to test this
, we performed our analysis using both the ao-1 `` background '' region shown in fig .
[ spectra_region ] and one using pv phase data closer to the ridge .
we do not detect any significant difference in the spectral index or flux of the hard ridge region as a function of the chosen background region . because of its larger extent and lower surface brightness , the ridge spectrum has lower signal to noise than the hxf . as with the hxf , fitting the spectrum with the background model yielded a clear excess above 2 kev .
it was not possible , however , to constrain the spectral properties of the ridge s hard component .
while we have no evidence suggesting a different spectral form , we can not eliminate the possibility that the hard emission from the ridge has a different origin .
the gis spectral parameters for this region are similar to those obtained using the non - imaging ( 1field of view ) _ ginga _ lac .
wahk reported the presence of a hard component that was fit by a thermal model with kt=11@xmath17 kev ( from their fig .
3 ) or a power - law with photon index @xmath18 . assuming a uniform surface brightness inside the shell and an absorbing column density of @xmath19 , wahk found the total 220 kev emitted flux from ic443 to be @xmath20 . as the total flux is dominated by the @xmath0 1 kev thermal component , a comparison between the broad band _
ginga _ flux and
the gis flux from the hard features is inappropriate .
a more reasonable quantity to compare is the flux density at 7 kev , where the relative contribution of the @xmath0 1 kev component is negligible . using figure 2 in wahk , and assuming that the _ ginga _ lac has an effective area of 3000 @xmath21 and near unity quantum efficiency at 7 kev
, we estimate the _ ginga _ flux density at 7 kev to be 20 @xmath22 10@xmath23 photons s@xmath24 kev@xmath24 @xmath25 . from the gis data , we find the hxf flux density at 7 kev to be approximately 4 @xmath22 10@xmath23 photons s@xmath24 kev@xmath24 @xmath25 , and the flux density of the portion of the hard ridge within the field of view to be about 2 @xmath22 10@xmath23 photons s@xmath24 kev@xmath24 @xmath25 .
thus the detected flux in the hard regions seems to account for only 30% of the hard flux detected by _
ginga _ , despite their prominence in the gis images .
moreover , integration over the rest of the surface of the remnant contained in the gis fields ( about 90 percent of the total ) yields a flux density of no more than 4 @xmath22 10@xmath23 photons s@xmath24 kev@xmath24 @xmath25 , leaving a factor of two discrepancy between the gis and _ ginga _ 7 kev flux densities .
two possible resolutions of this discrepancy are that there is significant hard flux from the small fraction of ic 443 unobserved by the gis , or there is a calibration discrepancy between the gis and _ ginga_. to check this latter possibility , we compare the broad band flux from all of ic 443 in the common 2 - 10 kev band . the 2 - 20 kev _ ginga _ flux is 9 @xmath22 10@xmath26 ergs @xmath25 s@xmath24 ( wahk ) . the 2 - 10 kev flux from ic 443 as measured by _
a-2 is 7@xmath31 @xmath22 10@xmath26 ergs @xmath25 s@xmath24 ( pssw ) , consistent with the _ ginga _ measurement .
integrating the 2 - 10 kev flux within the field of view of the two gis pointings , we find 5@xmath31 @xmath22 10@xmath26 ergs @xmath25 s@xmath24 ; this represents 90 percent or more of the total flux from the remnant .
the consistency of these numbers indicates that contribution of cross calibration uncertainties to the discrepancy is probably small , but could be as large as a factor of 2 .
thus while it is possible that the gis has observed all the regions in ic 443 producing hard flux , we can not rule out the possibility that the unobserved portion of the ic 443 rim to the south and east of the ridge contribute significantly to the hard flux .
a key conclusion about the spatial distribution of the hard x - ray emission in ic 443 , independent of a resolution of this discrepancy , is that the hard emission is highly localized .
the flux density from the bulk of the remnant surface area is at most 40 percent of the total hard flux .
this provides a strict upper limit on the luminosity of a hard thermal component , distributed throughout the remnant .
wahk inferred that ic443 is a very young ( 1200 y ) remnant based on their interpretation of the hard component as hot gas arising from a high velocity shock , and ic443 s coincidence with the guest star of 837 a.d . despite the restrictive limit we have placed on the flux of a hot thermal component , we nevertheless can not rule out their conclusion about the age of ic443 .
moreover , high shock velocities may also be required to accelerate electrons to the relativistic energies required for the emission of x - ray synchrotron radiation , and wahk s historical evidence is naturally unaffected by our observations . in order to search for periodicity from the unresolved hxf , we performed coherent fast fourier transforms ( ffts ) on the _ rosat _ hri and _ asca _ events recorded from the region used for spectral analysis ( after applying the standard barycenter corrections ) .
all our results are consistent with a non - periodic signal .
the temporal resolution of the gis is 0.0625 s , so we can not use the gis data to search for a frequency faster than 8 hz . on the other hand , the _ rosat _ hri data have millisecond resolution .
we therefore performed summed ffts on the combined data set ( _ rosat _ + _ asca _ ) in the low frequency range 0.058 hz , and only the _ rosat _ data in the high frequency range 101000 hz .
we find 99% confidence pulsed fraction upper limits of @xmath015% and @xmath033% in the low and high frequency ranges respectively .
we have discovered an isolated hard x - ray emitting feature and a ridge of hard emission in the southeast of the snr ic443 using the _ asca _ gis .
the hxf s x - ray spectrum can be characterized by either a power law of energy spectral index @[email protected] or thermal bremsstrahlung with kt=@xmath15 kev ; the ridge spectrum appears similar .
the features can account for most of the hard x - ray flux from ic443 ( [ asca_spectra.sec ] ) .
the core of the hxf is marginally resolved with the rosat hri , but the low level surrounding emission extends about 10 along the radio - bright shell .
( [ rosat_hri_image.sec ] , fig .
[ hri_image ] ) .
it is spatially coincident with the @xmath2720200 cm flat spectral ( @xmath28 ) region ( g86 ; fig .
[ figure_discussion.fig ] ) , and in the general region of , but not uniquely coincident with , regions of cloud / shock interactions ( denoyer , 1979 ; fig .
[ figure_discussion.fig ] ) .
it is located outside the 95 percent confidence error circle of the nearby egret source ( ehks ; fig .
[ figure_discussion.fig ] ) .
no strong periodicity is observed from the feature , with a pulse fraction upper limit of 33% ( [ periodicity.sec ] ) , and there is no known pulsar near it .
we focus here on the origin of the emission from the hxf . because of the uncertainty of the spectral parameters we know substantially less about the ridge .
while some of our discussion below applies to all the emission ( e.g. our conclusion that it is non - thermal ) , a more definitive understanding of the nature of the ridge emission ( which could be different from that of the hxf ) awaits a better quality x - ray spectrum .
the data suggest that the most likely mechanism for producing hard x - rays is synchrotron emission . before addressing whether this emission arises from a plerion or shock accelerated electrons , we first discuss our rationale for ruling out alternative mechanisms , including bremsstrahlung from a thermal and a non - thermal population of electrons , and inverse - compton scattering .
[ [ thermal - bremsstrahlung ] ] thermal bremsstrahlung + + + + + + + + + + + + + + + + + + + + + + a thermal bremsstrahlung model provides an adequate fit to the _ asca _ and the _ ginga _ spectra separately .
the inferred temperatures are discrepant , however ( @xmath29 kev for _ ginga _ versus @xmath30 kev for _ asca _ ) . in contrast
, a power law model yields the same photon index .
if the 220 kev spectrum is better characterized by a power law , then the temperature obtained from a thermal model fit from each instrument would yield a temperature characteristic of the instrumental bandpass , which is exactly the case here .
thus , a thermal bremsstrahlung model is inconsistent with the _ asca _ and _ ginga _ observations taken together .
[ [ inverse - compton ] ] inverse compton + + + + + + + + + + + + + + + one possibility discussed by wahk is that x - rays are produced by inverse - compton scattering of infrared photons off electrons .
gaisser et al .
( 1996 ) found that an inverse - compton component , scattering off the microwave background , would have a photon index of @xmath14=@xmath31 . on the other hand ,
if the scattering photons were locally produced , one would expect the hard x - ray emission to arise from ir - bright regions .
the localization of the hard x - ray emission a restricted region , the brightest part of which has no ir counterpart argues against an inverse - compton origin . [ [ accelerated - bremsstrahlung ] ] accelerated bremsstrahlung + + + + + + + + + + + + + + + + + + + + + + + + + + it has been suggested that electrons accelerated in the shock to mev energies may be responsible for the generation of hard x - rays via bremsstrahlung in supernova remnants ( sturner _ et al . _ 1995
; asvarov _ et al . _ 1990 ) .
the fact that the radio spectral index around the hard x - ray feature is flatter than elsewhere in ic443 is strong evidence against this mechanism .
if we assume that the same acceleration mechanism produces both the mev and the gev electrons ( responsible for the radio emission ) , then there will be fewer mev electrons in the hard region than elsewhere in the remnant .
thus , if this mechanism were operating efficiently , this region would be _ dimmer _ in hard x - rays than everywhere else in the remnant .
[ [ synchrotron - radiation ] ] synchrotron radiation + + + + + + + + + + + + + + + + + + + + + the most straightforward mechanism for producing a power law x - ray spectrum is synchrotron radiation .
none of the observations contradict this interpretation ; all are consistent with it . in particular , the fact that the radio spectrum of this region is flatter than elsewhere argues that at higher energies ( including the x - ray band ) , the synchrotron flux will be enhanced over elsewhere in the snr .
in fact , as we show below , the physical size of the hxf corresponds well with the synchrotron loss time of x - ray producing electrons .
wahk dismissed synchrotron emission as a likely model , because an acceleration mechanism efficient enough to accelerate electrons to @xmath010 tev seemed unlikely and they calculated synchrotron loss times to be less than the age of the snr . their dismissal was based on the assumption that the hard x - ray emission is spatially uniform ; we now know from this _ asca _ observation that the hard emission arises primarily from localized regions .
x - ray synchrotron emission in a snr can be produced by high energy electrons accelerated and interacting with a magnetic field in one of two locations : the magnetosphere of a pulsar ( giving rise to a plerion ) or the forward shock of a snr .
the observations support both interpretations to some extent , but more issues arise from the presence of a plerion than from an isolated region of intense shock acceleration .
the compact size of the hard x - ray feature is the primary evidence in favor of a plerionic interpretation .
the lack of x - ray pulsations is not necessarily a problem : the upper limit on the pulsed fraction is higher than other pulsars like vela ( pravdo et al .
1976 ) . nor is the existence of extended emission without an obvious embedded point source unprecedented : the snr 3c 58 has a small extent , a power law spectrum and shows no pulsations , but is generally regarded as a plerion ( helfand et al . 1995 ) .
difficulties with the interpretation arise when trying to associate a plerion with ic443 .
wilson ( 1986 ) has shown that the median x - ray : radio @xmath32 flux ratio for plerions is about 1 , with a range from about 0.1500 .
the flux ratio of the hxf is about 900 , which would give it the highest known x - ray : radio flux ratio of any plerion .
furthermore , if the feature is a plerion associated with the ic443 shell it requires an extremely high projected velocity of @xmath33 @xmath34 , for distance @xmath35 , angular distance @xmath36 from the explosion center , and age @xmath37 .
ic443 is in a complex region of interstellar space , and the diffuse emission is thought to the product of multiple supernova events ( aa94 ) .
the proper motion problem is circumvented if the hard x - ray feature represents the site of the most recent explosion which we can take from chinese records to have occurred in a.d .
837 ( wahk ) .
if that were the case , then it is surprising that this explosion has apparently not affected the temperature or surface brightness distribution of the diffuse emission .
the former suggests a more centrally located explosion ; the latter an explosion in the northeastern quadrant of the snr . on the other hand ,
the presence around the hard feature of many interesting structures associated with collisions between the shock front and concentrations of material suggests the hard x - ray emission is related to them .
in particular , shock / cloud collisions can locally enhance particle acceleration ( jk93 ) .
we consider first whether shock acceleration is a plausible source of the hard x - rays , and then how the morphology of the feature might arise .
dickel & milne ( 1976 ) found faraday rotation measures in ic443 of about 200 rad m@xmath38 where the field is along the line of sight , and close to zero where it is perpendicular .
taking the density and path length measurements from the x - ray ( pssw ) , we estimate the magnetic field strength @xmath39 to be : @xmath40 the synchrotron photon / electron energy relation therefore can be written : @xmath41 where @xmath42 is the observed photon energy produced as synchrotron radiation from an electron at energy @xmath43 .
so while radio emission ( @xmath44 kev ) is produced by gev electrons , production of 5 kev x - rays requires the presence of @xmath020 tev electrons . in attempting to explain
the non - thermal component dominating the x - ray emission from sn1006 as synchrotron emission from highly relativistic electrons , reynolds ( 1996 ) has shown that it is possible to accelerate electrons ( and ions ) in snr shocks to energies exceeding 100 tev .
while his model is not strictly applicable here , some aspects of it can be used to establish the plausibility of shock accelerated electrons as the source of the hard x - ray emission in ic443 . in the simplest models ,
the synchrotron spectrum is expected to be a broken power law .
the various break frequencies correspond to electron energies where two canonical times scales equate , such as the synchrotron loss time and the acceleration time scale . at each break , the spectral index increases by about 0.5 .
the electron diffusion time is also important in determining the steepening of the spectrum , but its influence is less straightforward to estimate . in the ic443 hard feature ,
the difference between the radio and the x - ray spectral index is approximately 1.1 , suggesting the need for at least two spectral breaks between the bands .
the time it takes for an electron of energy e to lose half its energy via synchrotron radiation in a magnetic field of strength b is given by : @xmath45 which is approximately the light crossing time of the hxf . using equations [ sync_loss.eq ] and [ photon_electron.eq ] , and assuming a snr age between 1000 and 5000 yr ( wahk , aa94 ) , we find that the break frequency where the synchrotron loss time becomes comparable to the age of the snr occurs in the far infrared . additionally , by assuming a shock velocity on the order of 1000 km s@xmath24
, we find the break frequency associated with the equivalence between the acceleration and loss times is in the hard ultraviolet band .
the frequencies at which these two spectral breaks occur supports the idea that the flat radio and the hard x - ray spectra are both produced by shock acceleration . while reynolds ( 1996 )
formalism facilitates the plausibility argument above and is especially suggestive with regard to the ridge region , his models for producing x - rays from shock accelerated electrons from a large segment of the snr shell clearly do not apply to the hxf in ic443 .
a different mechanism must be operating there .
such a mechanism has been suggested by jk93 .
figure [ figure_discussion.fig ] shows bgbw s h@xmath46 image of shocked molecular gas .
this shows that the molecular cloud has a clumpy ring structure exterior to the site of the supernova event .
et al . _ ( 1992 ) found that if the ring is circular , it must be inclined @xmath47 from the line of sight and have a diameter of @xmath48 pc .
doppler shift information shows the southern edge of the ring ( near our hxf ) expanding toward us , and the northern edge away implying that the ring is now expanding at a deprojected velocity of 25 km / s . while , we agree with this picture in principle , we point out that to second order the ring is not circular ; most noticeably there is a section of negative curvature surrounding the hxf ( fig . [ figure_discussion.fig ] ) .
it is this change of curvature that may create the hxf .
jk93 simulated cosmic - ray acceleration as shocks impact dense clouds which is the case here .
they found that after a shock / cloud impact , particle acceleration is significantly enhanced in the `` tail shock '' emanating from the cloud in the direction of shock propagation .
their results suggest that each clump in the molecular ring could produce enhanced shock acceleration in a long tail behind it .
because of the negative curvature of one section of the ring , the flow will be refracted , thereby `` focusing '' the enhanced shock acceleration at the hard x - ray feature where the tail shocks overlap .
this interaction between these tail shocks will both increase the effective shock velocities and provide a complicated structure which will require further study .
the total hard flux from the diffuse emission along the shell ( the ridge ) is about half that from the hxf , but the ridge may extend out of the field of view .
it corresponds to an arc bright in h@xmath49 , and radio continuum . while we can not yet be certain that the emission in non - thermal , the higher ambient density as indicated by the infrared line emission argues that the temperature of shock - heated material should be lower than elsewhere in ic443 not higher .
the arguments made above for the hxf against emission mechanisms other than synchrotron hold for the ridge region as well . while we withhold judgment until a higher quality spectrum can be obtained , we suggest that we are seeing evidence here as well for tev electrons . whether these arise from the jk93 mechanism or are
more closely related to the forward shock is still an open question .
in this paper we have presented evidence for a localized region of particle acceleration to electron energies over 20 tev , within the shell type supernova remnant ic443 .
because of the unique positions of the hard x - ray feature and shocked clouds , we believe that the hxf is caused by the focusing of tail shocks around each molecular cloud , thus enhancing particle acceleration in this location . in addition
we have found a ridge of hard emission which in coincident with both the radio - synchrotron shell and shocked molecular gas , but whose spectral parameters are poorly constrained . more observational and theoretical research is needed to fully understand this phenomenon .
multi - frequency and polarization observations with the vla will allow us to both measure the feature s spectral index and look for other indicators of on - going particle acceleration .
a deep single - dish radio observation of this area is necessary to verify that the hxf is indeed non - plerionic .
the most important future observation will be to observe the full extent of the ridge and the shocked molecular gas with asca to better characterize the spectrum and image the full extent of the hard non - thermal emission .
we thank the following people for contributing to these results : richard mushotzky for bringing the hard feature to our attention ; michael burton for kindly providing his h@xmath49 emission line image in digitized form ; steve reynolds and tom jones for their insights into shock acceleration ; matthew baring and ocker de jager for their thoughtful comments and suggestions .
allen , g.e .
, et al . , 1995 , , 448 , l25 asaoka , i. , & aschenbach , b. 1994 , , 284 , 573 ( aa94 ) asvarov , a.i . , guseinov , o.h . , kasumov , f.k . & dogel , v.a . , 1990 , , 229 , 196a axford , w.i . , 1994 , , 90 , 937 biermann , p.l . , 1995 , , 74 , 385 blandford r.d
& ostriker j.p . , 1978 , , 221 , l29 burton , m.g . ,
geballe , t.r . , brand , j.l . & webster , a.s . , 1988 , , 231 , 617 ( bgbw ) burton , m.g . ,
hollenbach , d.j . ,
haas , m.r . & erickson , e.f .
, 1990 , , 355 , 197 denoyer , l.k . , 1979 , , 232 , l165 dickel , j.r .
& milne , d.k .
, 1976 , aust . j. phys . , 29 , 435 dickman , r.l . , snell , r.l . ,
ziurys , l.m . ,
huang , y.l .
, 1992 , , 400 , 203 esposito , j.a . ,
hunter , s.d . , kanbach , g. & sreekumar , p. , 1996
, , 461 , 820 ( ehks ) gaisser , t.k . , protheroe , r.j .
, stanev , todor , 1996 , , _ submitted _ green , d.a . , 1986 , , 221 , 473 ( g86 ) helfand , d.j . ,
becker , r.h .
& white , r.l . , 1995 , , 453 , 741 holt , s.s . ,
gotthelf , e.v .
, tsunemi , h. & negoro , h. , 1994 , , 46 , l151 jones , t.w . &
kang , h. , 1993 , , 402 , 560 ( jk93 ) koyama , k. , petre , r. , gotthelf , e.v . ,
hwang , u. , matsuura , m. , ozaki , m. & holt , s.s .
, , 378 , 255 lasker , b.m . , et al . , 1990 , , 99 , 2019 miyata , e. , tsunemi , h. , pisarski , r. & kissel , s.e . , 1994 , , 46 , l101 petre , r. , szymkowiak , a.e
. , seward , f.d .
& willingale , r. , 1988 , , 335 , 215 ( pssw ) pravdo , s. h. , becker , r. h. , boldt , e. a. , holt , s. s. , rothschild , r. e. , serlemitsos , p. j. & swank , j. h. , 1976 , , 208 , l67 reynolds , s.p . , 1996 , , 459 .
l13 sturner , s.j .
& dermer , c.d . , 1995 , , 293 , l17 sturner , s.j . ,
skibo , j.g .
, dermer , c.d . & mattox , j.r . , 1995 , baas , 187 , # 58.05 sturner , s.j . , skibo , j.g .
& dermer , c.d .
, 1996 , _ in preparation _ taylor , j.h . ,
manchester , r.n . &
lyne , a.g . , 1993 , , 88 , 529 wang et al .
1992 , pasj , 44 , 303 .
( wahk ) wilson , a.s . , 1986 , , 302 , 718 | we present evidence for shock acceleration of cosmic rays to high energies ( @xmath010 tev ) in the supernova remnant ic443 .
x - ray imaging spectroscopy with _ asca
_ reveals two regions of particularly hard emission : an unresolved source embedded in an extended emission region , and a ridge of emission coincident with the southeastern rim .
both features are located on part of the radio shell where the shock wave is interacting with molecular gas , and together they account for a majority of the emission at 7 kev . though we would not have noticed it _ a priori _ , the unresolved feature is coincident with one resolved by the _ rosat _ hri . because this feature overlaps a unique region of flat radio spectral index ( @xmath1 ) , has about equal light - crossing and synchrotron loss times , and a power law spectrum with a spectral index of @[email protected] ,
we conclude that the hard x - ray feature is synchrotron radiation from a site of enhanced particle acceleration .
evidence against a plerion includes a lack of observed periodicity ( the pulsed fraction upper limit is 33% ) , the spectral similarity with the more extended hard region , the location of the source outside the 95% error circle of the nearby _ egret _ source , the fact that it is nestled in a bend in the molecular cloud ring with which ic443 is interacting , and the requirement of an extremely high transverse velocity ( @xmath45,000 km / s ) .
we conclude that the anomalous feature is most likely tracing enhanced particle acceleration by shocks that are formed as the supernova blast wave impacts the ring of molecular clouds .
february 5 , 1997 accepted by _ the astrophysical journal _ |
grs 1915 + 105 is the prototypical microquasar , a galactic x - ray binary ejecting plasma clouds at [email protected] ( mirabel & rodriguez 1994 ) .
three different sources of infrared ( ir ) radiation are expected from a x - ray binary , namely thermal emission from the outer part of accretion disk , thermal emission from the companion star or synchrotron emission from the jet(s ) .
previous ir observations of grs 1915 + 105 include , besides a few single measurements since 1993 , some few hrs photometry in conjunction with x - ray / radio measurements ( eikenberry 2000 ) , and a 2 month monitoring in 1996 ( bandyopadhyay 1998 ) .
k band observations have been performed with the 1.55-m telescope of the us naval observatory flagstaff equipped with the 256@xmath1256 pixel hgcdte ircam camera during 810 day intervals around full moon . because the main interest was in the long - term behaviour , `` only '' about 13 observations per month were done , consisting of groups of 3 or 4 individual observations per night .
each group represents a set of coadded integrations with an exposure of 400 sec . in total , observations were performed on 34 nights between 05 may 1999 and 15 jun 2000 .
these observations show grs 1915 + 105 to be variable between 12.1 k 13.7 mag ( not extinction - corrected ; fig .
1 ) . our long coverage allows to search for correlations with other emission characteristics , e.g. the rxte all - sky monitor ( asm ) quick - look results provided by the asm / rxte team . after selecting those k band observations which were done within 4 hrs of an asm scan
, we find that both the x - ray flux as well as the x - ray hardness ratio ( flux ratio of 512 kev vs. 35 kev band ) are positively correlated with the k band flux .
a regression analysis suggests that a doubling of the x - ray count rate ( flux ) leads to a 33% increase in the k band flux .
this argues against an explanation due to illumination of the secondary .
instead , it rather could be synchrotron radiation related to ( small ) ejections / flares as seen also at shorter time scales ( eikenberry 2000 ) . in this picture
the correlation of the k band flux with the x - ray hardness ratio is compatible with the correlation of x - ray spectral slope and radio flux found by rau & greiner ( 2001 ) . a large ( small ) value of the hardness ratio corresponds to a flat ( steep ) x - ray power law slope , and therefore a high ( low ) temperature of the electrons in the corona if the general comptonization paradigm is adopted .
these correlations then imply that strong k band flux and weak radio flux occur during times of a flat x - ray power law , i.e. most of the synchrotron power is radiated either in the infrared or radio band depending only on the temperature of electrons in the corona . | we report on long - term k band photometry of the galactic microquasar grs 1915 + 105 since march 1999 .
we see variations of more than 1 mag , which are seemingly correlated with variations in x - ray intensity and spectral slope . |
the interaction of mesons with nuclei is a major source for our understanding of the strong interaction . for long - lived mesons , like charged pions or kaons ,
secondary beams can be used for detailed studies of elastic and inelastic reactions , revealing the relevant potentials .
however , most mesons are short - lived so that their interaction with nucleons can only be studied in indirect ways , making use of final - state interactions ( fsi ) .
the general idea is to produce the mesons with some initial reaction in a nucleus and then study their interaction with the same nucleus .
it is much discussed whether the strong interaction allows the formation of quasi - bound meson - nucleus states .
so far , all known meson - nucleus bound states involve at least partly the electromagnetic interaction .
pionic atoms are well established .
more recently deeply bound pionic states have also been reported @xcite , but in this case the binding results from the superposition of the repulsive s - wave @xmath8-nucleus interaction with the attractive coulomb force .
neutral mesons on the other hand could form quasi - bound states only via the strong interaction .
the meson - nucleus interaction for slow pions is much too weak to produce quasi - bound states , but the situation may be different for @xmath0 , @xmath9 , and @xmath10-mesons .
the case of @xmath0-mesons is special because the threshold region of @xmath0-production is dominated by an @xmath11-wave resonance @xcite , the s@xmath12(1535 ) , which couples very strongly to @xmath13 ( branching ratio @xmath1450% @xcite ) . as a consequence fsi in nuclei
are important ; @xmath0-mesons are absorbed in nuclei with typical cross sections around 30 mb , basically independent of their kinetic energy @xmath15 over a wide range of @xmath15 from a few mev to 1 gev @xcite .
first hints for the possible existence of bound @xmath0-nucleus states came from the analysis of the @xmath16 scattering length that characterizes the potential at low kinetic energies . already in 1985 bhalerao and liu @xcite reported an attractive @xmath11-wave @xmath16 interaction from a coupled channel analysis of pion - induced reactions , yielding scattering lengths @xmath17 with real parts between 0.27 fm and 0.28 fm and imaginary parts between 0.19 fm and 0.22 fm . shortly afterwards ,
liu and haider @xcite pointed out that this interaction might lead to the formation of quasi - bound @xmath0-nucleus states for @xmath18 10 .
experimental evidence for such ` heavy ' @xmath0-mesic nuclei has been searched for in pion - induced reactions on nuclei in the oxygen region @xcite , but those experiments did not produce conclusive evidence .
more recently , sokol and co - workers @xcite claimed evidence for the formation of @xmath0-mesic nuclei from bremsstrahlung - induced reactions on @xmath19c @xmath20 where the @xmath21 pairs were detected in the final state . in this type of experiment , the @xmath0-meson
is produced in quasi - free kinematics on a nucleon ( @xmath22 ) , which takes away the largest part of the momentum and the @xmath0 is almost at rest in the residual @xmath23 nucleus .
if a quasi - bound state is produced , the @xmath0-meson has a large chance to be re - captured by a nucleon into the s@xmath12 excitation , which may then decay into a pion - nucleon back - to - back pair .
such pairs were searched for in the experiment , however , background from quasi - free pion production must be considered .
sokol and pavlyuchenko @xcite claim an enhancement above this background for certain kinematic conditions .
the topic gained much new interest after precise low - energy data for the photoproduction of @xmath0-mesons off the proton @xcite , deuteron @xcite and helium nuclei @xcite became available and refined model analyses of the scattering length were done by many groups ( see @xcite for a summary ) .
the results for the imaginary part are rather stable , most cluster between 0.2 fm and 0.3 fm . the real part , which determines the existence of quasi - bound states , is much less constrained .
it runs all the way from a negative value of @xmath240.15 fm to numbers close to and even above @xmath251 fm .
however , most of the more recent analyses prefer large values above 0.5 fm , which has raised controversial discussions about the possible existence of very light mesic nuclei and prompted theoretical studies of the @xmath0-interaction with @xmath26h , @xmath1h , @xmath1he , and @xmath27he systems @xcite experimental evidence for light @xmath0-mesic nuclei has mostly been searched for in the threshold behavior of @xmath0-production reactions .
the idea is that quasi - bound states in the vicinity of the production threshold will give rise to an enhancement of the cross section relative to the expectation for phase - space dependence .
many different hadron induced reactions have been studied in view of such threshold effects .
already the measurement of pion induced @xmath0-production off @xmath1he in the reaction @xmath28 at lampf @xcite revealed production cross sections significantly larger than dwia predictions .
subsequently , different nucleon - nucleon and nucleon - deuteron reactions , in particular @xmath29 @xcite , @xmath30 @xcite , @xmath31 @xcite , @xmath32 @xcite , @xmath33 @xcite , @xmath34 @xcite , and @xmath35 @xcite have been studied .
interesting threshold effects have been found in most of them , but in particular the @xmath36 @xcite and @xmath37 reactions @xcite show an extremely steep rise at threshold , implying a very large @xmath38 scattering length .
wilkin and collaborators @xcite have argued from an analysis of the angular distributions that , not only the magnitude of the @xmath11-wave amplitude falls rapidly in the threshold region , but that its phase also varies strongly , which supports the idea of a quasi - bound or virtual @xmath39 state very close to the threshold .
if such states do exist , they should show up as threshold enhancements independently of the initial state of the reaction .
photoproduction of @xmath0-mesons from light nuclei is a very clean tool for the preparation of the @xmath0-nucleus final state with small relative momenta but , due to the much smaller electromagnetic cross sections , sensitive threshold measurements have been sparse until now .
a particular problem is that the @xmath0-mesons have to be produced coherently off the target nuclei . as a consequence of the results from photoproduction of @xmath0-mesons off light nuclei ,
it is now well understood that the threshold production is dominated by an isovector spin - flip amplitude exciting the s@xmath12(1535 ) resonance ( see @xcite for a summary ) .
therefore , the coherent cross section is very small for the isoscalar deuteron and practically forbidden for the isoscalar - scalar @xmath27he nucleus , where only higher partial waves could contribute . among the light targets only the ( @xmath40,@xmath41 ) @xmath1h and @xmath1he nuclei have reasonably large cross sections for the @xmath42 reaction .
the @xmath1he system was previously investigated with photon - induced reactions by pfeiffer et al .
possible evidence for the formation of a quasi - bound state was reported from the behavior of two different reactions .
the coherent @xmath0-photoproduction @xmath43 showed a strong threshold enhancement with angular distributions much more isotropic in the threshold region than expected from the nuclear form factor .
these are indications for strong fsi effects .
furthermore , in an approach similar to the sokol experiment @xcite , the excitation function for @xmath44 pairs emitted back - to - back in the @xmath45 center - of - momentum ( c.m . )
system was investigated .
the difference between the excitation functions for opening angles between 170@xmath46 - 180@xmath46 ( almost back - to - back ) and 150@xmath46 - 170@xmath46 ( background from quasi - free pion production ) showed a narrow peak around the @xmath0-production threshold .
such a peak is expected to arise when quasi - bound @xmath0-mesons are re - captured into a nuclear s@xmath12 excitation with subsequent decay into the pion - nucleon channel .
however , the statistical quality of the measurements was limited and it was pointed out by several authors @xcite that the data do not prove the existence of a quasi - bound state . the present experiment aimed at a measurement with significantly improved statistical quality and better control of systematic effects .
this was achieved by using a detector system with much larger solid - angle coverage .
apart from counting statistics this is important for two aspects .
it allows a measurement of not only the @xmath47 decays but also the @xmath4 decay chain .
comparison of the two results helps to estimate systematic effects from the detection efficiency .
even more important , the coincident detection of recoil nucleons from non - coherent production processes becomes much more efficient and can be used to suppress this most important background .
the experiment was performed at the tagged photon beam of the mainz mami accelerator @xcite .
the electron beam of 1508 mev was used to produce bremsstrahlung photons in a copper radiator of 10@xmath48 m thickness , which were tagged with the upgraded glasgow magnetic spectrometer @xcite for photon energies from 0.45 gev to 1.4 gev . the typical bin width for the photon beam energy ( 4 mev )
is defined by the geometrical size of the plastic scintillators in the focal plane detector of the tagger .
the intrinsic resolution of the magnetic spectrometer is better by more than an order of magnitude .
the size of the tagged photon beam spot on the target was restricted by a 3 mm diameter collimator placed downstream from the radiator foil .
the target was a mylar cylinder of 3.0 cm diameter and 5.08 cm length filled with liquid @xmath1he at a temperature of 2.6 k. the corresponding target density was 0.073 nuclei / barn .
the reaction products were detected with an electromagnetic calorimeter combining the crystal ball detector ( cb ) @xcite made of 672 nai crystals with 384 baf@xmath49 crystals from the taps detector @xcite , configured as a forward wall .
the crystal ball was equipped with an additional * p*article * i*dentification * d*etector ( pid ) @xcite for the identification of charged particles and all modules of the taps detector had individual plastic scintillators in front for the same purpose .
the pid , in combination with the crystal ball , and the taps - taps - veto system could be used for a @xmath50 analysis for the separation of different charged particles .
the crystal ball covered the full azimuthal range for polar angles from 20@xmath46 to 160@xmath46 , corresponding to 93% of the full solid angle .
the taps detector , mounted 1.457 m downstream from the target , covered polar angles from @xmath145@xmath46 to 21@xmath46 .
this setup is similar to the one described in detail in @xcite .
the only difference is that , in the earlier setup , the taps forward wall consisted of 510 modules and was placed 1.75 m from the target . for the present experiment the main trigger condition was a multiplicity of two separated hits in the combined cb / taps system and an integrated energy deposition of at least 300 mev in the crystal ball .
details for the basis of the data analysis including calibration procedures , identification of photons and recoil nucleons and the absolute normalization of cross sections are discussed in @xcite . here , we will only outline the specific steps for the identification of coherent @xmath0-production .
results for quasi - free production of @xmath0-mesons off nucleons bound in @xmath1he in view of the structure recently observed for @xmath0-photoproduction off the neutron @xcite will be reported elsewhere .
the analysis was based on the @xmath3 and @xmath4 decay channels .
the measurement of both channels allowed an additional control of systematic uncertainties .
events were analyzed with either two or six photon candidates and no further hit in the detector system .
the latter condition reduces the incoherent background since events with recoil nucleons were suppressed .
for the two - photon decay , @xmath0-mesons were identified with a standard invariant - mass analysis .
residual background under the @xmath0-peak was subtracted by fitting the spectra with simulated line shapes and a background polynomial .
this procedure was applied individually for each combination of incident photon energy and meson c.m .- polar angle .
for the six photon decay , the photons were first combined via a @xmath51-test to three pairs , which were the best solution for three @xmath6 invariant masses .
a cut between 110 mev - 150 mev was made on these invariant masses .
subsequently , the six - photon invariant mass was constructed .
the corresponding spectra were basically background free under the @xmath0-peak ( direct triple-@xmath6 production has a very small cross section in the energy range of interest ) .
the most important step in the analysis is the separation of the coherent reaction from breakup where nucleons are removed from the nucleus .
since the @xmath1he recoil nuclei can not be detected ( they are mostly stopped in the target ) only the overdetermined reaction kinematics can be used .
the kinetic energy of the @xmath0-mesons in the @xmath52-@xmath1he c.m .- system is fixed not only by the incident photon energy but also from the measured laboratory energy and polar angle of the meson .
the difference , the missing energy @xmath53 , is shown for both data samples in fig .
[ fig : misse ] and compared to the simulated line shapes of the coherent and breakup reactions .
the simulations were done with the geant4 program package @xcite , taking into account all details of the target and detector setup .
the simulation of the coherent part is straightforward due to the two - body final - state kinematics . in case of the breakup part
, the momentum distribution of the bound nucleons has to be considered , and this was taken from the work of mccarthy , sick , and whitney @xcite .
several final states , such as @xmath54 , @xmath55 with different participant - spectator combinations , contribute .
good agreement between data and simulations was achieved , at all the incident photon energies investigated , with three different reaction components : the coherent process , the breakup process with a quasi - free participant nucleon , and the breakup process where the recoil is taken by a ` participant ' deuteron .
the relative contributions of these processes have been determined by fitting the data with a superposition of their line shapes ( cf fig .
[ fig : misse ] ) . since in principle the detection efficiency for
the different components can be different for the @xmath47 and @xmath56 decays , no constraints relating the two channels were imposed on the fits . as a consequence , for some energy bins the fitted relative contributions of the breakup processes are different for the two decay channels , but the fit of the coherent contribution is quite stable .
this is so because the breakup background makes almost no contribution at positive missing energies , so that the coherent contribution is only weakly dependent on the exact shape of the background .
the lowest range of photon energies ( 600 mev @xmath57 608 mev ) lies between the kinematic thresholds of coherent and breakup reactions .
consequently , only the coherent reaction can contribute .
in fact , a clean signal at zero missing energy is seen , which agrees with the simulated coherent signal shape . at even lower photon energies ,
the spectra show no signal above statistical noise .
the count rates are roughly a factor of 2.75 larger for the @xmath3 data than for the @xmath4 data ( note the different scales for the left and right parts of fig .
[ fig : misse ] ) , which is due to the respective detection efficiencies ( @xmath1480% for @xmath58 , @xmath1435% for @xmath59 ) and decay branching ratios ( @xmath60=([email protected])% , @xmath62=([email protected])% @xcite ) . absolutely normalized total and differential cross sections for both decay channels
have been extracted from the yields , the target density , the decay branching ratios @xcite , the simulated detection efficiency , the electron beam flux measured in the focal - plane detector of the tagging spectrometer , and the tagging efficiency @xmath63 , i.e. the number of correlated photons that pass through the collimator .
the latter was measured in special low - intensity runs with a lead - glass detector in the photon beam ( see @xcite for details ) and ranged from 60% to 75% . for the systematic uncertainties we estimate 10% for the overall normalization including target density ( measurement of target temperature and possible deformation of the cold target cell , @xmath147 - 8 % ) , photon flux ( measurement of tagging efficiency and electron flux , @xmath645% ) and decay branching ratios ( almost negligible @xmath641 % ) . for the reaction dependent simulations of the detection efficiency ,
5% uncertainties are estimated . finally , the energy dependent uncertainty for the reaction identification ( in particular the separation in the missing energy spectra ) ranges from 2% at threshold to 20% at the highest incident photon energies . for the analysis of the excitation function of @xmath6-p back - to - back
pairs in the @xmath52-@xmath1he c.m .- system , monte carlo simulations were done for the signal and the background coming from quasi - free production of @xmath6 mesons . for the signal ,
the decay of bound s@xmath12 resonances into @xmath65 was modelled with a momentum distribution corresponding to the nuclear fermi - motion .
these simulations were used to select the kinematical conditions best suited for the signal . in the spectra of the pion - proton opening angle , signal events appear roughly between 150@xmath46 - 180@xmath46 , while the background is distributed between 100@xmath46 and 180@xmath46 .
the total cross section data for coherent @xmath0-production from the two @xmath0 decay branches is compared to model predictions @xcite in fig . [
fig : total_raw ] .
the agreement between the two data sets is quite good , but none of the existing models reproduce the data , even when their systematic uncertainties ( shown as shaded band in the figure ) are considered .
the most prominent feature of the data is the extremely sharp rise at threshold . here
one should note that the binning of the data ( 4 mev for the 2@xmath52-channel , 8 mev for the 6@xmath52-channel ) is very large compared to the energy resolution of the magnetic tagging spectrometer ( roughly 200 kev ) .
the work by tiator , bennhold and kamalov @xcite is based on a coupled - channel analysis of pion- and photon - induced pion- and eta - production off the nucleon , which parameterizes the resonance contributions with an isobar model and the background terms with effective lagrangians . for the coherent production off @xmath1he ,
realistic nuclear wave functions were used , but the model was based on the plane - wave impulse approximation ( pwia ) .
the result strongly underestimates the magnitude of the measured cross section and does not reproduce the energy dependence .
fix and arenhvel @xcite have modelled coherent @xmath0-photoproduction off @xmath1he and @xmath1h in pwia , in a distorted - wave impulse approximation ( dwia ) , using an optical potential , and in a full four - body scattering model .
they find strong fsi effects , which amplify the threshold cross section for the full four - body model with respect to pwia and dwia ( which give similar results ) .
nevertheless their cross section underestimates the data by roughly a factor of two .
shevchenko and collaborators @xcite have also calculated coherent @xmath0-production off the three - nucleon system in a microscopic few - body description .
they find a strong dependence of the result on the elastic @xmath16 rescattering , which is not sufficiently constrained by experiment .
two examples for different fsi modelling are shown in fig .
[ fig : total_raw ] .
they exhibit strong threshold effects , but do not reproduce the measurements above the breakup threshold .
the average of the two data sets is compared to the previous result from pfeiffer et al .
@xcite in fig .
[ fig : total ] .
the two data sets are in reasonable agreement ( within uncertainties ) except for two points around 620 mev . here
, one should note that , in this range , the systematic uncertainty in the previous data was large because the coherent signal had to be extracted by fitting a small coherent contribution in the missing energy spectra , which were dominated by the quasi - free reaction .
the quasi - free contribution is less important at lower incident photon energies , and is better separated from the coherent component at higher incident photon energies .
the present experiment profited in this range from the almost @xmath66 coverage of the detector , which allowed some suppression of the quasi - free background by the detection of associated recoil nucleons .
also shown in this figure are the results from a simple pwia model .
it uses the effective photon energy @xmath67 given by @xmath68 where @xmath69 is the laboratory energy of the incident photon and @xmath70 is the polar angle of the @xmath0 meson in the photon - nucleus c.m .
system with @xmath71 .
the effective photon energy corresponds to the @xmath72 of the incident photon ( four - momentum @xmath73 ) and a nucleon ( four - momentum @xmath74 ) with three - momentum @xmath75 from the nucleon motion inside the nucleus .
the nucleon momentum is related in the factorization approximation to the momentum @xmath76 transferred to the nucleus by @xcite : @xmath77 where all momenta are in the laboratory system
. the coherent cross section is then composed of three factors : @xmath78 the elementary cross section @xmath79 , a kinematical factor , and the nuclear form factor for point - like nucleons .
the kinematical factor accounts for the change of phase - space between the different c.m .
systems and is derived from the photon and @xmath0 three - momenta in the photon - nucleon ( @xmath80 , @xmath81 ) , and photon - nucleus ( @xmath82 , @xmath83 ) c.m . systems .
the nuclear form - factor @xmath84 was taken from @xcite .
it includes the spatial distribution of the charge ( respectively magnetic moment ) of the nucleon , while here the distribution of point - like nucleons is relevant .
we have therefore used the ratio of the nuclear form factor @xmath85 and the nucleon dipole form factor @xmath86 to approximate the distribution of point - like nucleons in the @xmath1he nucleus . since @xmath0-photoproduction in this energy range
is almost only due to the @xmath87 excitation , the dominant electromagnetic multipole is the @xmath88 spin - flip . neglecting small contributions from other multipoles and higher components in the @xmath1he wave function , only the unpaired neutron contributes to coherent production because a spin - flip of a proton is pauli - blocked .
therefore , we use for the elementary cross section the experimental results for the @xmath89 reaction from @xcite .
since the angular dependence up to the s@xmath12 peak is almost isotropic , one can approximate the c.m .
differential cross section by : @xmath90 where the total neutron cross section @xmath91 is parameterized as a breit - wigner curve with an energy dependent width using the numerical values from @xcite ( @xmath92=1546 mev , @xmath93=176 mev , @xmath94=90@xmath9510@xmath96 gev@xmath97 , @xmath98=0.5 , @xmath99=0.4 , @xmath100=0.1 ) .
variation of the parameters within their uncertainties changes the predictions for the coherent cross section typically on the 10% level .
the results from the pwia model are compared to the data in figs .
[ fig : total ] and [ fig : ang ] .
they are shown for the approximation of eq .
[ eq : isotropic ] assuming isotropic angular distributions for the elementary @xmath89 reaction , and also for realistic angular distributions taken from @xcite .
the difference is , as expected , negligible . despite the simplicity of the model
, the predicted cross section agrees with experiment within statistical and systematic uncertainties for the highest incident photon energies . in the threshold region
it is in excellent agreement with the pwia calculation by fix and arenhvel .
however , in this region it underestimates the data by nearly one order of magnitude , which is evidence for strong fsi effects at threshold .
the behavior of the angular distributions of the pwia is dominated by the nuclear form factor , which is responsible for the strong forward peaking . at high incident photon energies
the data show the same tendency , though the rise to forward angles is not as steep as in the model .
this indicates significant fsi effects even for those energies , since the more trivial approximations in the pwia model can not explain this discrepancy . the use of realistic angular distributions for the elementary cross section instead of eq .
[ eq : isotropic ] ( dashed curves in fig .
[ fig : ang ] ) has basically no impact . the use of more realistic nuclear wave functions instead of the simplified scalar form factor in eq .
[ eq : main ] leads only to comparatively small changes @xcite .
finally , the experimental angular resolution is only responsible for minor effects .
this is demonstrated by the dotted curves in fig .
[ fig : ang ] , which have been folded with the detector response . towards the threshold ,
the measured angular distributions become almost isotropic . between coherent and breakup threshold ,
they develop a rise to backward angles , while in pwia the form factor still causes a forward peaking .
together with the disagreement in the absolute scale of the cross section in the threshold region , this comparison clearly demonstrates the large influence of fsi effects . the excitation function of @xmath6-proton pairs for different ranges of their opening angle in the @xmath52-@xmath1he c.m .
system was analyzed in a manner similar to that of pfeiffer et al . @xcite .
the idea was , that when @xmath0-mesons are produced into an @xmath0-@xmath1he resonant state , overlapping with the coherent production threshold , those produced at photon energies below the threshold are off - shell and can not be emitted without violating energy conservation .
they can , however , be captured by a nucleon which is then excited to the s@xmath12-resonance .
since this resonance has a @xmath1450% decay branching ratio into @xmath65 , the expected signal would be pion - nucleon pairs emitted back - to - back in the @xmath52-@xmath1he c.m .
the results are shown in fig .
[ fig : pi0 ] .
the figure shows the difference of the excitation functions for opening angles from 165@xmath46 - 180@xmath46 ( back - to - back range taking into account effects of fermi motion ) and the data from 150@xmath46 - 165@xmath46 .
these ranges were optimized with monte carlo simulations of signal and background and are slightly different from those used by pfeiffer et al .
( 170@xmath46 - 180@xmath46 and 160@xmath46 - 170@xmath46 ) , but this is a minor effect .
the result shows the peak at the coherent @xmath0 threshold that had been observed previously , but now with much higher statistical significance .
however , the data also show some structure at higher incident photon energies which was not previously seen because the earlier experiment covered only incident photon energies up to 800 mev .
the much higher statistical quality of the new data allowed a detailed analysis of these structures .
this is summarized in the insert of fig .
[ fig : pi0 ] .
it shows the excitation functions for different ranges of opening angle , when the overall energy dependence of the data ( @xmath101 ) is removed .
the broad , peak - like structures are due to the second and third resonance regions of the nucleon .
the position of these signals shifts to higher incident photon energies for smaller opening angles .
this is a purely kinematical effect ; small opening angles correspond to decays of nucleon resonances moving forward in the photon - nucleus c.m
therefore , for a given mass @xmath102 of the resonance , small opening angles correspond to large incident photon energies ( and nucleon fermi momenta parallel to the photon momentum ) . unfortunately , subtraction of the normalized excitation functions for the 165@xmath46 - 180@xmath46 and 150@xmath46 - 165@xmath46 opening - angle ranges produces a narrow peak exactly at the @xmath0-production threshold .
this peak does , however , not originate from a special structure in the back - to - back data , but it is due to the subtraction of the shifted low energy tails of the second resonance region in single pion production . here
, we have only presented the results from an analysis similar to the previous work by pfeiffer et al . @xcite .
the much better statistical quality of the present data allows also more refined analyses with additional cuts to enhance the signal - to - background ratio . for this purpose
we have done detailed monte carlco simulations of the kinematical correlations for the signal and the background from quasi - free pion production and analyzed the excitation functions with optimized cuts . however , also with these analyses no structures that could be uniquely related to the decay of an @xmath0-mesic state into @xmath65 could be identified . the @xmath103 background before very specific cuts is at least three orders of magnitude larger than any expected signal from an @xmath0-mesic state . in consequence
, it seems to be impossible to identify in this reaction channel a small signal from the decay of bound s@xmath12 resonances on top of these complicated background structures , even when the signal exists .
coherent photoproduction of @xmath0-mesons off @xmath1he has been measured with much improved statistical quality compared to the pilot experiment of pfeiffer et al . @xcite .
the total cross section rises sharply between the coherent and breakup thresholds . compared to a pwia , which is in fair agreement with the data in the s@xmath12 peak ,
the threshold values are enhanced by nearly one order of magnitude .
this is very different e.g. from the behavior of coherent @xmath0 photoproduction off the deuteron , which is in reasonable agreement with pwia @xcite .
the angular distributions at threshold are almost isotropic , and are unlike the forward peaked distributions expected to result from the form factor behavior .
this result is similar to that previously observed for the hadron induced reactions @xmath36 @xcite and @xmath37 @xcite .
this independence from the initial state is strong evidence for dominant @xmath0-nucleus interaction effects , related to a resonant state in the vicinity of the @xmath0 production threshold .
the excitation function of @xmath44 back - to - back pairs was investigated as an independent signal for the formation of a quasi - bound or virtual @xmath0-nucleus state , as originally suggested in the work by sokol et al .
a peak - like structure at the @xmath0-production threshold had been observed in a previous measurement of photoproduction from @xmath1he by pfeiffer et al .
@xcite . this signal was reproduced with much improved statistical significance , but it could be attributed to an artifact arising from the complicated background structure of quasi - free pion production .
we wish to acknowledge the outstanding support of the accelerator group and operators of mami .
we gratefully thank c. wilkin for many stimulating discussions , instructive suggestions , and for careful and critical reading of the manuscript .
we thank a. fix and l. tiator for very helpful discussions about the pwia calculations .
this work was supported by schweizerischer nationalfonds , deutsche forschungsgemeinschaft ( sfb 443 , sfb / tr 16 ) , dfg - rfbr ( grant no .
05 - 02 - 04014 ) , uk science and technology facilities council , stfc , european community - research infrastructure activity ( fp6 ) , the us doe , us nsf and nserc ( canada ) .
99 h. geissel , et al .
88 ( 2002 ) 122301 .
b. krusche et al .
74 ( 1995 ) 3736 .
b. krusche et al . , phys . lett .
b 397 ( 1997 ) 171 . k. nakamura et al . , j. phys . g 37 ( 2010 ) 075021 .
m. rbig - landau et al .
b 373 ( 1996 ) 45 .
t. mertens et al .
j. a 38 ( 2008 ) 195 .
r.s.bhalerao and l.c.liu , phys .
54 ( 1985 ) 865 .
l.c.liu and q.haider , phys .
c 34 ( 1986 ) 1845 .
r.e.chrien et al . , phys . rev .
lett . 60 ( 1988 ) 2595 .
j.d.johnson et al . , phys .
c 47 ( 1993 ) 2571 .
sokol et al . , fizika b 8 ( 1999 ) 85 .
sokol and l.n .
pavlyuchenko , phys . of .
atomic nuclei 71 ( 2008 ) 509 .
b. krusche at al .
b 358 ( 1995 ) 40 . p. hoffmann - rothe et al . , phys .
( 1997 ) 4697 .
j. weiss et al . , eur .
phys . j. a 11 ( 2001 ) 371 .
j. weiss et al . , eur .
j. a 16 ( 2003 ) 275 . v. hejny et al .
j. a 6 ( 1999 ) 83 . v. hejny et al .
j. a 13 ( 2002 ) 493 .
arndt et al . , phys .
c 72 ( 2005 ) 045202 .
t.ueda , phys . rev .
66 ( 1991 ) 297 .
t.ueda , phys .
b 291 ( 1992 ) 228 .
c.wilkin , phys .
c 47 ( 1993 ) r938 .
s.a.rakityanski et al . , phys .
b 359 ( 1995 ) 33 . s.a .
rakityanski et al . , phys .
c 53 ( 1996 ) r2043 .
a.m. green and s. wycech , phys .
c 54 ( 1996 ) 1970 .
scoccola , d.o .
riska , phys .
b 444 ( 1998 ) 21 . n.v .
shevchenko et al . , eur .
j. a 9 ( 2000 ) 143 . v. yu .
grishina et al . , phys .
b 475 ( 2000 ) 9 .
h. garcilazo and m.t .
pea , phys .
c 63 ( 2001 ) 021001(r ) .
peng et al .
58 ( 1987 ) 2027 .
peng et al . , phys .
63 ( 1989 ) 2353 .
h. caln et al . , phys .
b 366 ( 1996 ) 366 .
j. smyrski et al . , phys .
b 474 ( 2000 ) 182 .
p. moskal et al . , phys .
c 69 ( 2004 ) 025203 . f. plouin , p.fleury , and c. wilkin , phys
65 ( 1990 ) 690 . h. caln et al . , phys .
80 ( 1998 ) 2069 .
b. mayer et al .
c 53 ( 1996 ) 2068 .
j. smyrski et al . , phys .
b 649 ( 2007 ) 258 .
t. mersmann et al . , phys .
98 ( 2007 ) 242301 .
t. rausmann et al . , phys .
c 80 ( 2009 ) 017001 .
a. wronska et al . , eur .
phys . j a 26 ( 2005 ) 421 .
n. willis et al . , phys .
b 406 ( 1997 ) 14 .
a. budzanowski et al . , nucl .
a 821 ( 2009 ) 193 .
f. hibou et al . , eur .
j. a 7 ( 2000 ) 537 .
c. wilkin et al .
b 654 ( 2007 ) 92 .
b. krusche and s. schadmand , prog .
( 2003 ) 399 .
m. pfeiffer et al .
92 ( 2004 ) 252001 .
a. sibirtsev et al . , phys .
c 70 ( 2004 ) 047001 . c. hanhart , phys .
( 2005 ) 049101 .
h. herminghaus et al .
, ieee trans . on nucl . science . 30 ( 1983 ) 3274 .
h . kaiser et al . , nucl .
instr . meth . a 593 ( 2008 ) 159 .
i. anthony et al .
instr . meth . a 301 ( 1991 ) 230 .
hall , g.j .
miller , r. beck , p.jennewein , nucl .
a 368 ( 1996 ) 698 .
mcgeorge et al . , eur .
j. a 37 ( 2008 ) 129 .
a. starostin et al . , phys .
c 64 ( 2001 ) 055205 .
r. novotny , ieee trans .
science 38 ( 1991 ) 379 .
gabler et al . , nucl .
instr . meth . a 346 ( 1994 )
d. watts , in _ calorimetry in particle physics , proceedings of the 11th international conference , perugia , italy 2004 _ , edited by c. cecchi , p. cenci , p. lubrano , and m. pepe ( world scientific , singapore , 2005 , p. 560
s. schumann et al . ,
j. a 43 ( 2010 ) 269 .
i. jaegle et al . , phys .
( 2008 ) 252002 .
i. jaegle et al . , eur .
j. a 47 ( 2011 ) 89 .
s. agostinelli et al . , nucl .
instr . meth . a 506 ( 2003 ) 250 .
j. s. mccarthy , i. sick , r. r. whitney , phys . rev .
c 15 ( 1977 ) 1396 .
l. tiator , c. bennhold , and s.s .
kamalov nucl .
phys . a 580 ( 1994 ) 455 . ; and s.s .
kamalov , priv . com .
a. fix and h. arenhvel , phys .
c 68 ( 2003 ) 044002 . ; and a. fix , priv . com .
shevchenko et al .
phys . a 714 ( 2002 ) 277 .
chumbalov , r.a .
eramzhyan , and s.s .
kamalov , z. phys . a 328 ( 1987 ) 195 . ; and l. tiator , priv . com . | coherent photoproduction of @xmath0-mesons off @xmath1he , i.e. the reaction @xmath2 , has been investigated in the near - threshold region .
the experiment was performed at the glasgow tagged photon facility of the mainz mami accelerator with the combined crystal ball - taps detector .
angular distributions and the total cross section were measured using the @xmath3 and @xmath4 decay channels .
the observed extremely sharp rise of the cross section at threshold and the behavior of the angular distributions are evidence for a strong @xmath5 final state interaction , pointing to the existence of a resonant state .
the search for further evidence of this state in the excitation function of @xmath6-proton back - to - back emission in the @xmath7 reaction revealed a very complicated structure of the background and could not support previous conclusions . , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , |
apart from convective phenomena like the granulation most features we observe in the solar atmosphere are due to the presence of magnetic fields .
the dynamics of this magnetic field lie at the root of solar activity , giving rise to variable amounts of uv , x - ray and particle radiation that influence not only earth climate , but also our daily lives in a society that depends more and more on technology .
thus , there is a real incentive to understand , and eventually be able to predict , the behavior of the solar magnetic field .
while larger scale magnetic field concentrations like sunspot are more evident , longer lived , and easier to observe , the field on the solar surface seems to be distributed predominantly in small - scale magnetic elements like micropores , network bright points and low contrast internetwork fields @xcite .
because of their small size , characteristically @xmath8 @xcite , at or below the resolution limit of current solar instrumentation , and their rapid time scale of evolution of 10 - 100 sec , near sustainable cadence of multi - wavelength observations , the morphology , thermal structure and dynamics of small - scale flux concentrations is difficult to study . despite a characteristic field strength in the kilo gauss range @xcite , the small sizes of magnetic elements result in low values of the magnetic flux and make the field difficult to observe .
therefore , long integration times are needed to reach the required magnetograph sensitivity if the elements remain unresolved , and this results in smeared images due to the short dynamical time scales of the elements .
fortunately , indirect methods which allow much shorter exposure times exist to examine the structure of magnetic elements , and to follow their evolution . in particular , these methods involve imaging the sun in wideband ( of order 1 nm ) filters centered on molecular band heads towards the blue end of the spectrum .
examples are the band head at 388.3 nm due to b@xmath9x@xmath9 electronic transitions of cn @xcite , and the so - called g band ( originally designated by * ? ? ?
* ) around 430.5 nm due to a@xmath0x@xmath1 transitions of the ch molecule @xcite . in these molecular wideband images
the magnetic elements appear as subarcsecond sized bright points with contrasts of typically 30% , compared to the average photosphere @xcite .
@xcite find that g - band bright points are cospatial and comorphous with magnetic elements in intergranular lanes to within @xmath10 .
also in strong photospheric lines @xcite and the wings of chromospheric lines like h@xmath11 @xcite , and k @xcite bright points seem to coincide with magnetic field concentrations . however , the cores of these lines are considerably darker than the integrated g - band intensity and require much narrower bandpasses , making imaging in them much less practical . until now measurement of magnetic field strength in molecular bright points has had to rely on cospatial magnetograms in magnetically sensitive atomic lines .
this is an arduous process , which involves careful alignment of images taken at different wavelengths and a relative `` destretching '' of images taken at different times to prevent contamination of the magnetogram signal due to seeing induced image distortions .
in addition , not every magnetic element has an associated bright point or vice versa , at a given time , although both @xcite and @xcite conclude that there is a one - to - one correspondence when the temporal evolution of the magnetic field and bright grains is followed . an important point in the diagnostic use of bright point imaging for the study of magnetic element dynamics and morphology is the question of why the associated bright points are particularly bright in the molecular band heads .
model calculations of g - band brightness in semi - empirical fluxtube atmospheres provide reasonable values for the bright point contrast @xcite .
flux - tube models are hotter than the quiet sun in the layers producing the observed light .
they are evacuated owing to the presence of a magnetic field , thus allowing us to observe deep and therefore hot photospheric layers . for this very reason , @xcite concluded that g - band contrast in the bright points is enhanced compared to the surrounding photosphere because the opacity in the ch lines is less affected by the higher temperatures present in magnetic elements than the continuum opacity , which is mostly due to h@xmath12 . in the continuum
a rise in temperature leads to a higher formation height of intensity at consequently lower temperatures . on the other hand , in the g band
the formation height remains more or less constant ( because of the increase in h@xmath12 opacity and decrease in ch line opacity through dissociation ) so that the emergent intensity reflects the higher temperatures .
similarly , @xcite emphasized that the evacuation of magnetic elements due to the requirement of pressure balance with the non - magnetic surroundings leads to excess ch dissociation making the g band relatively more transparent , thereby exposing hotter layers below .
this has been clearly illustrated by @xcite by modeling the g - band intensity in a snapshot from a three - dimensional simulation of solar magneto - convection . that the ch concentration is strongly reduced in evacuated flux - tube regions
is logical because the molecular concentration depends quadratically on density _ if _ the molecular dissociation is not controlled by radiation , which seems to be indeed the case as pointed out by @xcite .
such a quadratic dependence on density would explain why ch lines show stronger weakening than atomic lines in the g band @xcite , and cause higher bright - point contrast in the ch and cn band heads than in atomic line cores , because in the solar photosphere atomic ionization is mostly dominated by the radiation field , making the opacity in atomic lines only linearly dependent on the density in the flux concentrations . in this paper
we propose a technique that hopefully will allow us to shed further light on the relationship between the magnetic field in the bright points and their increased contrast , namely the measurement of the field through analysis of the polarization signal due to the zeeman effect in the ch lines that make up the bulk of the opacity in the g band . over the last few years
, we have witnessed an increasing interest in using molecular line polarization as a tool for magnetic field diagnostics , concerning both the molecular zeeman effect ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) and the hanle effect in molecular lines ( e.g. , * ? ? ?
* ; * ? ? ?
the quantum theoretical background for treating the molecular zeeman effect has been mostly in place since 1930 ( see * ? ? ? * for a description and references ) , but was first worked out practically in terms of splitting patterns and intensity of polarization components for hund s case ( a ) transitions of any multiplicity and hund s case ( b ) transitions of doublets by @xcite .
his results were extended by @xcite to general multiplicity in hund s case ( b ) and transitions of intermediate ( a - b ) coupling , i.e. , with partial spin decoupling .
here we assume the upper and lower states of the a@xmath0x@xmath1 transitions of the ch molecule to be in pure hund s case ( b ) , where the electron spin is decoupled from the molecular rotation ( e.g. , * ? ? ? * ) .
this is a good approximation for these states because the of small values of the spin coupling constants in both , as discussed by @xcite .
these authors explicitly show the difference in effective land factor between case ( b ) and the intermediate case ( a - b , their fig .
even for very small rotational quantum numbers the perturbations caused by incomplete spin decoupling are very small , justifying our assumption .
the structure of this paper is as follows . in section [ sec : zeeman ] we briefly review the zeeman effect in ch lines and show the effective land factors of satellite branch transitions , which behave differently than those in the main branch .
we discuss our model calculations in section [ sec : modelcalculations ] , and give conclusions in section [ sec : conclusions ] .
radiative transitions in a molecule are susceptible to the zeeman effect in the presence of an external magnetic field and may produce polarized radiation that can be analyzed to infer the properties of the field , much like in the atomic case . if a molecule has a non - zero magnetic moment , the interaction of that moment with the magnetic field will split the molecular energy levels into a pattern that is in general , however , more complex than in the atomic case because of the additional degrees of freedom inherent in molecular structure . in a diatomic molecule both vibration along the internuclear axis and rotation around the center of mass influence level energies , but only rotation affects the molecule s magnetic moment because it alters the way the electronic spin and orbital angular momenta add to the total angular momentum @xmath13 of the molecule , which in turn couples to the magnetic field .
three factors can contribute to the magnetic moment of a diatomic molecule : the orbital and spin angular momenta of the electrons , the rotation of the molecule , and the magnetic moment associated with the nuclear spins .
the first contribution is of the order of a bohr magneton @xmath14 ( in si units ) , while the second and third contributions are smaller by a factor of the order @xmath15 .
since the ch lines we consider in this paper have non - zero orbital and spin angular momenta in both their upper and lower levels , we can safely ignore the contributions of molecular rotation and nuclear spin here .
let @xmath16 and @xmath17 be the total electronic orbital and spin angular momenta in the diatomic molecule , and @xmath18 the component of @xmath16 along the internuclear axis . if @xmath18 is non - zero it is coupled to the electric field along the internuclear axis , and is space quantized with quantum number @xmath19 . in spectroscopic notation
these states are denoted by the upper - case greek letters @xmath20 , @xmath21 , @xmath22 , @xmath23 , etc .
, which are similar to the @xmath24 , @xmath6 , @xmath25 , @xmath26 , designations in atomic systems .
the total angular momentum @xmath13 is always the resultant of @xmath16 , @xmath17 , the angular momentum of rotation @xmath27 , and the total spin of the nuclei .
different simplifying cases can be distinguished depending on how strong the relative coupling between these angular momenta is , as first put forward by hund ( see * ? ? ?
if both spin and orbital angular momenta are strongly coupled to the internuclear axis and only very weakly to the rotation , hund s case ( a ) applies . if on the other hand @xmath17 is only coupled weakly to the internuclear axis , whereas @xmath18
is strongly coupled we have hund s case ( b ) .
this is the case that applies to both the upper and lower level of the ch lines in the g - band region and is discussed further here .
figure [ fig : hunds_b ] shows a vector diagram of the angular momenta in hund s case ( b ) . in this case the orbital angular momentum @xmath18 and rotational angular momentum @xmath27 combine to form @xmath28 , with quantum number @xmath29 @xmath28 in turn combines with spin angular momentum @xmath17 to form @xmath13 , which can have quantum numbers @xmath30 according to the standard addition rule for quantized angular momentum vectors ( e.g. , * ? ? ?
* ) . in a magnetic field
the total angular momentum @xmath13 is space quantized such that the component in the field direction is @xmath31 , where @xmath32 in case the molecule has a non - zero magnetic moment a precession of @xmath13 around the field will take place , and states with different @xmath33 will have different energies : @xmath34 where @xmath35 is the mean value of the component of the molecular magnetic moment along the field direction , @xmath36 is the mean magnetic moment along the direction of @xmath13 , and @xmath37 is the field strength .
the time average of @xmath36 is composed of a contribution @xmath38 due to the orbital angular momentum which is coupled to the internuclear axis in hund s case ( b ) , and a contribution @xmath39 due to the spin coupled to the rotation axis . since we are considering field strengths outside the paschen - back regime the precession of @xmath13 around the field can be considered slow compared to that of @xmath28 and @xmath17 about @xmath13 , while in hund s case ( b ) the latter is in turn slower than the nutation of @xmath18 about @xmath28 .
therefore , we have @xmath40 \cos(\mathbf{j } , \mathbf{b } ) { \mu_{\mathrm{bohr}}}\\ & = & \left [ \frac{\lambda^2}{\sqrt{n(n+1 ) } } \cos(\mathbf{n } , \mathbf{j } ) + 2 \sqrt{s(s+1 ) } \cos(\mathbf{s } , \mathbf{j } ) \right ] \frac{m}{\sqrt{j(j+1 ) } } { \mu_{\mathrm{bohr } } } , \label{eq : meanmub}\end{aligned}\ ] ] where we used @xmath41 and @xmath42 @xcite .
according to equation ( [ eq : meanmub ] ) the magnetic field splits the molecular level ( for given @xmath28 and @xmath17 ) into @xmath43 equidistant components with a splitting that only depends on the relevant quantum numbers , not on the molecular constants .
closer inspection also reveals that when @xmath44 is large and the first term in the bracket vanishes , the result becomes independent of @xmath2 for large @xmath2 and the maximum splitting ( i.e. , @xmath45 ) is of the order of the normal zeeman splitting .
this is the case , for instance , when the rotational quantum number @xmath5 is large so that @xmath18 is almost perpendicular to @xmath13 ( see figure [ fig : hunds_b ] ) , resulting in a contribution to the magnetic moment due to the orbital angular momentum that averages to nearly zero . for a given @xmath13 and
given orientation of @xmath17 with respect to @xmath13 , each spin level is split into @xmath46 components with a splitting that reduces approximately linearly with @xmath44 for large values of @xmath44 ( according to the first term in eq .
[ [ eq : meanmub ] ] ; see also fig . 145 in @xcite @xcite )
. using vector rules @xmath47 the quantum rule @xmath48 , and similar expressions for @xmath49 and @xmath50 it follows that @xmath51 and similarly @xmath52 ( * ? ? ?
* but note the minor error in the denominator of the latter expression there ) . combining equations ( [ eq : energydiff ] , [ eq : meanmub ] , [ eq : cosnj ] , and [ eq : cossj ] ) we obtain @xmath53 with @xmath54 the land factor for energy - level splitting in hund s case ( b ) given by @xmath55}{2n(n+1 ) } + \\
\nonumber & & j(j+1 ) - n(n+1 ) + s(s+1)\biggr\ } \frac{1}{j(j+1)}. \label{eq : hundsbsplitting}\end{aligned}\ ] ] electronic transitions in the molecule obey selection rules in the same way as atomic transitions .
rules for molecular transitions can be split up into two categories : general selection rules that hold true for all electronic transitions and selection rules that only hold true for specific coupling cases . for all electric dipole transitions in molecules the selection rule for total angular momentum quantum number @xmath2 holds rigorously : @xmath56
if both the upper and lower levels of a transition are of hund s case ( a ) or ( b ) , then the quantum numbers for the spin angular momentum quantum number @xmath24 and orbital angular momentum @xmath57 are defined for both levels and obey the selection rules : @xmath58 when both upper and lower level are strictly of case ( b ) , i.e. , with the spin coupled to the molecular rotation , the quantum number @xmath44 is defined and obeys the selection rule @xmath59 with the added restriction that @xmath60 is not allowed for @xmath20@xmath20 transitions @xcite .
the set of transitions that have @xmath61 are called main branch transitions .
these are termed @xmath62 , @xmath63 , and @xmath64 branches for @xmath65 , respectively , with @xmath66 referring to the upper level total angular momentum quantum number and @xmath67 to the lower level , and @xmath68 depending on whether @xmath69 down to @xmath70 ( * ? ? ? * and 222 ) .
line strengths ( defined as @xmath71 , where @xmath72 and @xmath73 are the statistical weight and energy of the lower level , @xmath74 is the oscillator strength , @xmath75 the boltzmann constant , and @xmath76 the temperature ) of ch a@xmath0x@xmath1 main branch lines in the g - band region are plotted in figure [ fig : strength_main ] as a function of wavelength for a typical photospheric temperature of @xmath77 k. for @xmath78 , the lines form what are known as the satellite branches . in case ( b ) , the satellite branches lie near the main branches with the same values of @xmath79 , but the line strengths in the satellite branches are always considerably smaller than those in the main branches and fall rapidly with increasing @xmath44 @xcite .
satellite branches are termed @xmath80 , @xmath81 , etc . with the left superscript referring to the value of @xmath82 in the same way as the main branches are denoted .
strengths of the ch a@xmath0x@xmath1 satellite lines in the g - band region are plotted in figure [ fig : strength_sat ] .
comparison with figure [ fig : strength_main ] shows that the satellite branch lines are generally much weaker than those in the main branch .
note that figure [ fig : strength_sat ] also contains values for @xmath83 and @xmath84 branches , in which the superscripts @xmath24 and @xmath85 imply @xmath86 , respectively .
these values violate selection rule ( [ eq : selectn ] ) , and indicate that the associated `` forbidden '' transitions have small but non - zero transition probabilities . in the presence of a magnetic field
radiative transitions between shifted levels are allowed with the following selection rule for the component @xmath33 of the total angular momentum along the magnetic field : @xmath87 these transitions have characteristic polarizations that follow the same rules as for the atomic case . in the case of @xmath88 the observed line component is polarized parallel to the field when viewed perpendicular to the field ( @xmath89-component ) , and in the case of @xmath90 , the components are linearly polarized perpendicular to the field when viewed perpendicular to the field and left- and right hand circularly polarized when viewed in the field direction ( @xmath91-components ) . the shift in energy of each individual line component with respect to its zero - field value
is : @xmath92 as for atomic transitions the strength for the individual components with @xmath93 within each radiative transition from a lower level with quantum numbers @xmath94 to an upper level with @xmath95 is obtained from the wigner - eckart theorem @xcite .
table 1 lists the unnormalized strengths @xmath96 for each allowed @xmath97 and @xmath98 .
these strengths have to be normalized for each given @xmath99 according to : @xmath100 so that the sum of strengths over all zeeman split components with a particular polarization within one transition yields unity .
table 1 + the unnormalized strengths @xmath96 + [ cols="<,^,^,^",options="header " , ] [ tab : componentstrength ] a convenient quantity to describe the characteristics of an anomalous zeeman multiplet is the effective land @xmath101 factor for the transition , which corresponds to the @xmath101-factor of a zeeman triplet whose @xmath102-components lie at the wavelengths of the centers of gravity of the @xmath102-components of the anomalously split line .
thus the effective land factor for a general zeeman multiplet is calculated as the normalized and weighted sum of the shift @xmath103 over all transitions in the multiplet with @xmath104 : @xmath105 equation [ eq : effectivelande ] was used to calculate the effective land factors of the satellite branches of the ch a@xmath0x@xmath1 system in the g - band region as a function of the lower - level angular momentum quantum number @xmath67 shown in figure [ fig : satellite ] . a similar graph is shown by ( * ? ? ?
* their fig .
9 ) for the main branch transitions of the ch molecule .
note that while @xmath106 for all the main branches tends to zero with increasing @xmath67 , this is not the case for the those of the @xmath107 and @xmath108 , and @xmath84 and @xmath109 type satellite branches , which tend to @xmath110 and @xmath111 , respectively
. only effective land factors of the @xmath112 type satellite branches follow the behavior of the main branch transitions . despite the tendency of @xmath106 of the @xmath6 and @xmath5 satellite branches to @xmath7
, we found in the calculations described below that the satellite branch lines contribute little to the polarization signal in the g band because they are generally much weaker than the main branch lines as indicated by comparison of figures [ fig : strength_main ] and [ fig : strength_sat ] .
to get a first impression of the feasibility of bright point magnetic field measurement through the zeeman effect in the ch lines of the g band we calculated the polarization at these wavelengths due to a 0.1 t constant vertical field in a one - dimensional plane - parallel model of the quiet solar atmosphere .
in addition , we modeled the polarization signal through a two - dimensional cross section of a snapshot from a three - dimensional solar magneto - convection simulation .
we implemented the rules for energy level splitting in hund s case ( b ) ( see eq .
[ [ eq : hundsbsplitting ] ] ) and line component strength ( table 1 ) in a numerical radiative transfer code to calculate the splitting patterns of the ch lines in the g - band region .
the code accounts for all components in each of the ch lines and does not rely on the calculated effective land factors .
concentration of the ch molecule in each location of the models was calculated by solving the coupled chemical equilibrium equations of a set of 12 molecules consisting of h@xmath113 , h@xmath114 , c@xmath113 , n@xmath113 , o@xmath113 , ch , co , cn , nh , no , oh , and h@xmath113o , and their atomic constituents . the population of individual ch levels was calculated according to the saha boltzmann equations , so that the molecular line opacity and source function were assumed to be in local thermodynamic equilibrium ( lte ) . values of wavelengths and line strengths were taken from file hydrides.asc of cd - rom 18 by kurucz[multiblock footnote omitted ] , which results in a similar emergent spectrum as the list provided by @xcite , although the latter includes many more weaker lines .
in addition , opacities due to background atomic lines were included from files gf0430.10 and gf0440.10 of cd - rom 1 by kurucz@xmath115 .
the opacities and source functions of these lines were assumed to be in lte as well , and for each of the electric dipole transitions among them zeeman splitting patterns ( e.g. , * ? ? ?
* ) were calculated as input for the polarized radiative transfer calculation . for each wavelength for which the emergent intensity was to be calculated the contributions from molecular and atomic lines to the @xmath116 propagation matrix @xmath117 and
the 4-element stokes emissivity @xmath118 @xcite were added as well as the ( unpolarized ) contributions from relevant background continuum processes .
given the propagation matrix and emission vector the stokes transfer equation was solved using the quasi - parabolic delo method proposed by ( * ? ? ?
* see @xcite 2000 for a first application of this method ) .
the delo method works analogous to the short characteristics method @xcite for unpolarized radiation and is easily generalized to multi - dimensional geometry . to investigate whether the intensity and polarization spectra anywhere in the g band
are dominated solely by ch lines we solved the full stokes radiative transfer equations over a wavelength range of 3 nm centered around 430.50 nm through a one - dimensional hydrostatic model of the average quiet solar atmosphere ( falc , model c of * ? ? ?
* ) with an imposed vertical magnetic field of 0.1 t ( 1000 g ) .
the resulting stokes @xmath4 and @xmath3 spectra are given in figure [ fig : gbandspectrum ] , where the thick solid curves represent the calculated @xmath4 and @xmath3 spectra taking account of both the atomic and ch molecular line contributions , and the dash - dotted curves represent the calculated spectra with the contribution from atomic lines omitted .
also plotted in this figure are the transmission curve of a typical g - band filter ( dotted curve ) with a fwhm of 1 nm and the disk centre solar intensity spectrum ( thin solid curve , * ? ? ?
overall , there is good agreement between the calculated intensity spectrum that includes all lines ( thick solid curve ) and the observed spatially averaged atlas spectrum ( thin solid line ) , indicating that the linelists we use are adequate for our purposes .
comparison between the solid curve and dash - dotted curves in @xmath4 and @xmath3 allows us to locate regions where the spectra are dominated by ch lines alone .
we identified the locations at 429.74 , 429.79 , 430.35 through 430.41 , 430.89 , 431.19 , and 431.35 nm as the possible regions of interest for observations if we want to measure bright point magnetic fields through the zeeman effect in the ch lines .
typical circular polarization signals at the wavelengths where ch lines alone dominate the spectrum are 25% for the 0.1 t constant vertical field compared to 20% for the strongest atomic lines in this region .
these predicted values are well within the range of what could be observed with sufficient spatial resolution .
none of these regions is dominated by a single ch line , rather the spectrum in each is the result of overlapping lines of varying degrees of magnetic sensitivity . as a result
, none of the profiles in stokes @xmath3 has a regular double - lobed antisymmetric shape .
in particular , the profile around 430.40 nm is notable .
the one - lobed profile is the result of two @xmath119 main branch lines in the @xmath120 band at 430.3925 and 430.3932 nm with @xmath121 and @xmath122 , and @xmath123 that overlap with lines that are much less magnetically sensitive .
it raises the interesting possibility of recovering the bright point magnetic field with narrowband imaging in only one polarization , avoiding the difficulties of reconstructing the polarization map from two different exposures . at the low rotational quantum number of the two @xmath119 lines
the electron spin is not completely decoupled from the inter - nuclear axis and calculation of the effective land factors for these transitions has to be treated with the intermediate case ( a - b ) formalism @xcite .
future polarimetry employing these lines would benefit from applying a more precise value of @xmath106 , which according to the graph in fig .
9 of @xcite is about 0.1 higher than the value for pure case ( b ) .
in fact , we have verified that performing the synthesis using the zeeman patterns calculated in the intermediate case ( a - b ) by means of the formalism developed by @xcite gives similar results . for a prediction of
the amount of zeeman induced polarization under more realistic conditions in the solar photosphere we solved the stokes transfer equations in a two - dimensional cross section through a simulation of magneto - convection @xcite . in this simulation
the physical domain extends 12 mm in both horizontal directions with a resolution of 96 km , extends 2.5 mm below the surface , and spans 3 mm vertically .
periodic boundary conditions were used in the horizontal directions , and transmitting vertical boundary conditions were specified .
radiative contributions to the energy balance were calculated by solving radiative transfer in lte using a four bin opacity distribution function . for the stokes transfer calculations a two - dimensional cut through a single three - dimensional simulation snapshot was selected .
the upper part of this slice , starting at 300 km below the photosphere , was interpolated onto a finer vertical grid , logarithmically in densities , and linearly in temperature and velocities . through this interpolated cross section
the stokes transfer equations were solved in two - dimensional geometry over a subset of wavelengths centered on 430.4 nm . the snapshot employed for this paper had an average vertical field of 0.003 t. the emergent stokes @xmath4 and @xmath3 spectra from the slice in the wavelength region around 430.40 nm are shown in figure [ fig : stokes_2d ] .
the polarization signal at wavelengths 430.36 and 430.40 is due to the zeeman effect in ch lines alone .
it is most prominent at positions 1.5 , 3.0 , and 4.5 mm .
all three of these locations are associated with downflows as deduced from the redshifts in the intensity spectrum , and are probably positioned in intergranular lanes . at positions 1.5 and 4.5 mm along the slice the line - of - sight component ( in the vertical direction ) of the field is strongest and results in a circular polarization signal of about 1% of the local continuum intensity .
given the size of these magnetic elements of order 100 km , and if the element is isolated without oposite magnetic polarities within the resolution element , they should produce a polarization signal of order at least @xmath124 even when the ( seeing induced ) telescope resolution is only of order 1 mm ( corresponding to slightly more than @xmath125 ) .
this is within the capabilities of current polarimetric instrumentation .
it is worth noting that the spectral feature at 430.40 does not consistently show a one - lobed profile as it does in the spectrum produced by the one - dimensional hydrostatic model . in particular , at position 1.5 mm the blue lobe is visible as well , altough much weaker than its red counterpart .
the difference with the hydrostatic model most likely derives through an interplay between the magnetic field and velocity gradients in the dynamic case , and the different formation heights of the magnetically sensitive and non - sensitive lines .
this inconsistency would make magnetic field mapping through narrow band imaging in one polarity a less reliable method , but its downside would have to be weighed against the improved spatial resolution this technique promises , including the full benefit it would have from image improving techniques like adaptive optics and post facto image reconstruction .
we have solved the full set of stokes transfer equations in two different models of the solar atmosphere in the wavelength region of the g band around 430 nm .
we include polarization due to the zeeman effect in atomic transitions , and most importantly in the molecular ch lines that are the major contributors to opacity in the g band .
judging from the calculated spectrum through a one - dimensional plane parallel model of the solar atmosphere ( falc ) we find several wavelengths regions at which both the intensity and the polarization signals are dominated by ch lines , without significant contribution from atomic lines .
the amplitude of the calculated polarization signal in these locations , of order 25% in the hydrostatic falc model with 0.1 t constant vertical field and of the order of 1% of the continuum intensity in stokes @xmath3 through the magneto - convection slice , leads us to predict that measurement of field strength in the ch lines that provide small - scale magnetic elements with their characteristic brightness is a realistic prospect . because of the property of the lines of the strong ch main branch to become less magnetically sensitive with increasing rotational quantum number ( as evidenced by the behavior of their effective land factors as shown by @xcite ) the polarization spectrum of the g band is significantly less crowded than its intensity spectrum
nevertheless , overlap of lines still occurs and none of the isolated ch features shows a regular symmetric @xmath3-profile even in the absence of any flows . in this paper we show that the effective land factors of the @xmath6 and @xmath5 satellite branches go asymptotically @xmath7 with increasing molecular rotation ( figure [ fig : satellite ] ) .
however , these satellite branch transitions are much weaker than their counter parts in the main branches and do not contribute significantly to the polarization signal in the g band . within the 3 nm wavelength interval we investigated we found one location where several ch lines overlap and dominate the spectrum to produce a single - lobed @xmath3-profile at 430.40 nm .
in particular , the spectrum calculated through the hydrostatic falc model shows this behavior , which prompts us to suggest that measurement of bright - point line - of - sight magnetic field strength could be achieved through narrowband ( of order 20 pm fwhm or 200 m ) imaging in one direction of circular polarization . such a method would have the full benifit of image quality improving techniques like adaptive optics and phase diversity restoration for optimal spatial resolution .
unfortunately , the @xmath3 spectrum calculated through the two - dimensional magneto - convection slice shows a very asymmetric two - lobed profile at 430.40 nm in some locations rather than a uniformly single - lobed one .
nonetheless , the benefit of optimal spatial resolution may outweigh the polarimetric accuracy in some cases .
proof that such curious single - lobed @xmath3-profiles indeed exist in the solar spectrum comes from recent observations of sunspots obtained with the zimpol polarimeter attached to the gregory coud telescope of irsol ( locarno ) , which will be discussed in an upcoming paper .
we conclude that the technique of magnetic field measurement through stokes polarimetry of the ch lines finally holds the promise for the development of a method that will allow us to unambiguously determine what the role of the field is in the appearance of g - band bright points . if such measurements are successful we will no longer have to deal with the difficulty of interpreting magnetic field measurements in atomic lines together with g - band intensity measurements introduced by the different formation properties of these species .
we would like to thank the referee , svetlana berdyugina , for her careful reading of the manuscript .
we are grateful to bob stein for providing the snapshot from which we extracted the two - dimensional magneto - convection slice used here .
the work of aar and jtb has been partially funded by the spanish ministry of science and technology through project aya2001 - 1649 .
this research has made use of nasa s astrophysics data system ( ads ) . , k. , schmidt , w. , rimmele , t. , & sigwarth , m. 2001 , in m. sigwarth ( ed . ) , advanced solar polarimetry theory , observation , and instrumentation , asp conf .
ser . , vol .
236 , the astronomical society of the pacific , san francisco , ca , p. 439
, r. j. , kiselman , d. , rouppe van der voort , l. , & plez , b. 2001 , in m. sigwarth ( ed . ) , advanced solar polarimetry theory , observation , and instrumentation , asp conf .
236 , the astronomical society of the pacific , san francisco , ca , p. 445
, r. f. , bercik , d. , & nordlund , a. 2003 , in a. a. pevtsov , h. uitenbroek ( eds . ) , current theoretical models and high resolution solar observations : preparing for atst , asp conf .
286 , the astronomical society of the pacific , san francisco , ca , p. 121
uitenbroek , h. 2003 , in a. a. pevtsov , h. uitenbroek ( eds . ) , current theoretical models and high resolution solar observations : preparing for atst , asp conf . ser
. , vol .
286 , the astronomical society of the pacific , san francisco , ca , p. 404 | we investigate the possibility of measuring magnetic field strength in g - band bright points through the analysis of zeeman polarization in molecular ch lines . to this end
we solve the equations of polarized radiative transfer in the g band through a standard plane - parallel model of the solar atmosphere with an imposed magnetic field , and through a more realistic snapshot from a simulation of solar magneto - convection .
this region of the spectrum is crowded with many atomic and molecular lines .
nevertheless , we find several instances of isolated groups of ch lines that are predicted to produce a measurable stokes v signal in the presence of magnetic fields . in part
this is possible because the effective land factors of lines in the stronger main branch of the ch a@xmath0x@xmath1 transition tend to zero rather quickly for increasing total angular momentum @xmath2 , resulting in a stokes @xmath3 spectrum of the g band that is less crowded than the corresponding stokes @xmath4 spectrum .
we indicate that , by contrast , the effective land factors of the @xmath5 and @xmath6 satellite sub - branches of this transition tend to @xmath7 for increasing @xmath2 .
however , these lines are in general considerably weaker , and do not contribute significantly to the polarization signal . in one wavelength location near 430.4 nm the overlap of several magnetically sensitive and non - sensitive ch lines is predicted to result in a single - lobed stokes @xmath3 profile , raising the possibility of high spatial - resolution narrow - band polarimetric imaging . in the magneto - convection snapshot
we find circular polarization signals of the order of 1% prompting us to conclude that measuring magnetic field strength in small - scale elements through the zeeman effect in ch lines is a realistic prospect . |
one of the most important observations in high energy collisions is the @xmath0 spectra for different species of final state particles . from the spectra , we can learn a lot about the regularities of the particle productions .
in many studies , searching for a scaling behaviour of the @xmath0 spectra is helpful to reveal these regularities . in @xcite , the authors showed a scaling behaviour in the pion @xmath0 spectra with different collision centralities at midrapidity in au+au collisions at the relativistic heavy ion collider ( rhic ) .
this scaling behaviour was extended to noncentral regions in au+au and d+au collisions @xcite .
similar scaling behaviours were found in the proton and anti - proton @xmath0 spectra with different collision centralities at midrapidity in au+au collisions at rhic @xcite .
recently , we observed a scaling behaviour in the @xmath0 spectra of inclusive charged hadrons in proton - proton ( proton - antiproton ) collisions at @xmath2 = 0.9 , 2.76 and 7 ( 0.63 , 1.8 and 1.96 ) tev when a linear transformation , @xmath5 , was applied on these spectra @xcite .
here the scaling parameter @xmath4 depends on the collision energy @xmath2 .
this scaling behaviour was explained by the colour string percolation model in a qualitative way . in this paper
, we will extend this scaling behaviour to the @xmath0 spectra of pions , kaons and protons produced in @xmath1 collisions at 0.9 , 2.76 and 7 tev . a new method , the quality factor method @xcite , will be adopted in the searching for the scaling parameter @xmath4 . compared with the method utilized in @xcite , this method does not rely on the shape of the scaling function ,
thus it is more robust .
we will argue that the pions , kaons and protons originate from different distributions of clusters which are formed by strings overlapping , and the colour string percolation model can describe the scaling behaviours of the identified @xmath0 spectra not only in a qualitative way but also in a quantitative way simultaneously . as a similar geometrical scaling behaviour
was suggested when the @xmath0 spectra of identified particles in @xmath1 collisions at 0.9 , 2.76 and 7 tev were presented in terms of a scaling variable @xmath6 , where @xmath7 gev / c and @xmath8 is around 0.27 @xcite , we would like to compare the scaling behaviours presented in @xmath9 and @xmath10 .
the organization of the paper is as follows . in section [ sec : method ] , we will illustrate the procedure to search for the scaling behaviours in the @xmath0 spectra of identified particles . in section [ sec : scaling_behaviour ] , the scaling behaviours of pions , kaons and protons will be described .
section [ sec : comp ] shows the comparison between the scaling behaviours presented in the variables @xmath9 and @xmath10 . in section [ sec : mechanism ] , we will use the colour string percolation model to explain the scaling behaviours of the identified particles in a quantitative way .
finally , the conclusion is made in section [ sec : conclusion ] .
as done in @xcite , we will search for the scaling behaviours of the identified pion , kaon and proton @xmath0 spectra with the following steps .
let us take the pion @xmath0 spectra as an example , we define a scaling variable , @xmath11 , and a scaled @xmath0 spectra , @xmath12 . here
the parameters @xmath4 and @xmath13 depend on the collision energy . by choosing proper @xmath4 and @xmath13
, the data points of the pion @xmath0 spectra in @xmath1 collisions at 0.9 , 2.76 and 7 tev will migrate into one curve . as a convention
, we usually set both @xmath4 and @xmath13 of the highest energy collisions to be 1 .
however , the @xmath0 coverages of the pion spectra at 0.9 and 7 tev are much smaller than the coverage at 2.76 tev . in order to make the scaling function @xmath14 to describe the pion spectra in the large @xmath0 region faithfully , we choose both @xmath4 and @xmath13 for the collisions at 2.76 tev as 1 . with different choices of @xmath4 and @xmath13 , the scaling functions @xmath14 are different . in order
to get rid of this arbitrariness , we introduce another scaling variable , @xmath15 , and the corresponding normalized scaling function @xmath16 . here
the ways to search for the scaling behaviours in the @xmath0 spectra of kaons and protons are identical to the one for pions .
in this paper , the @xmath0 spectra of identified pions , kaons and protons in @xmath1 collisions at 0.9 , 2.76 and 7 tev were published by the alice collaboration @xcite . here
the identified pion , kaon and proton @xmath0 spectra refer to the spectra of @xmath18 , @xmath19 and @xmath20 .
the data at 2.76 tev cover a pion @xmath0 range of 0.1 - 20.0 gev / c , which is much wider than the ones at 0.9 and 7 tev , 0.1 - 2.6 gev / c and 0.1 - 3.0 gev / c .
similar results are obtained in the comparison among the kaon or proton @xmath0 ranges at 0.9 , 2.76 and 7 tev . as described in section [
sec : method ] , both the parameters @xmath4 and @xmath13 at 2.76 tev are set as 1 , thus the scaling function @xmath14 is nothing but the @xmath0 spectra of pions , kaons or protons at this energy . in @xcite , @xmath14 is written in the tsallis form , in which the mass effect of the charged hadron is ignored . in this work , since the threshold of the @xmath0 range is 0.1 ( 0.1 and 0.1 ) gev / c at 2.76 tev @xcite , which is below the mass of pions ( kaons and protons ) 0.14 ( 0.494 and 0.938 ) gev / c@xmath21 @xcite , we write the scaling function @xmath14 for pions , kaons or protons with the modified tsallis form @xmath22^{\frac{1}{1-q } } , \label{eq : phi_z_pt_spectrum_pp}\end{aligned}\ ] ] where @xmath23 , @xmath24 and @xmath25 are free parameters , @xmath26 is the mass of the particle species , and @xmath27 is a measure of the non - extensivity .
these free parameters are given by fitting equation to the @xmath0 spectra of identified species at 2.76 tev with the least @xmath28s method .
the square root of the sum of the statistical and systematic uncertainties of the data points has been taken into account in these fits .
lists the parameters @xmath23 , @xmath24 , @xmath25 as well as their uncertainties returned by the fits .
the last line of this table shows the @xmath29s per degrees of freedom ( dof ) , named reduced @xmath28s , for these fits .
.[tab : id_particles_fit_parameters ] the parameters @xmath23 , @xmath24 and @xmath25 of the scaling functions @xmath14 for pions , kaons and protons .
the uncertainties quoted are due to the statistical plus systematic errors of the data points added in quadrature .
the last line shows the reduced @xmath28s for the fits on the @xmath0 spectra of identified particles at 2.76 tev . [ cols="^,^,^,^ " , ] upper panel : the scaling behaviour of the pion @xmath0 spectra presented in @xmath9 for @xmath1 collisions at 0.9 , 2.76 and 7 tev .
the solid curve is described by the csp fit in equation with parameters in the second column of .
the data points are taken from @xcite .
lower panel : the @xmath30 distribution as a function of @xmath9 for pions . ]
now we would like to seek for the reason why the csp model can describe the scaling behaviour of the identified @xmath0 spectra . under the transformation @xmath31 , @xmath32 and @xmath33 , both @xmath34 and @xmath35 are invariant .
so the @xmath0 spectra of identified particles in equation are also invariant .
this invariance is the scaling behaviour we are looking for . comparing the @xmath36 transformation in the csp model @xmath37 with the one used to search for
the scaling behaviour @xmath38 , we know that @xmath4 is proportional to @xmath39 . as described in @xcite , @xmath40 , where the average is taken over all clusters for identified particles , thus @xmath4 should be proportional to @xmath41 .
since the degree of string overlap @xmath42 grows with energy , the scaling parameter @xmath4 should also increase with energy .
that s indeed what we see for the identified particles in the third column of .
therefore the csp model can explain the scaling behaviours presented in @xmath9 for pions , kaons and protons separately .
however , the rate of the @xmath4 values increasing with energy for pions is different with the one for kaons or protons .
this could also be explained by the csp model .
as @xmath43 , and the values of @xmath44 are the same for pions ( kaons or protons ) at 0.9 , 2.76 and 7 tev , the ratio between the values of @xmath4 at different energies should be equal to the ratio between the values of @xmath45 at different energies .
@xmath45 is evaluated in terms of the csp model as @xcite @xmath46 the cluster fragmentation functions @xmath47 are the same for pions ( kaons or protons ) at different energies .
however , the cluster s size distributions @xmath34 are different for different particles , which can be seen from the comparison among the @xmath48 values for pions , kaons and protons in . as a result , for different species of particles , the values of @xmath4 change in a different rate with energy . in a summary , the csp model could be utilized to explain the scaling behaviours presented in @xmath9 for the identified particles simultaneously .
in this paper , we have extended the scaling behaviour in the inclusive charged hadron @xmath0 spectra to the spectra of identified particles at 0.9 , 2.76 and 7 tev .
this scaling behaviour is exhibited when @xmath0 is replaced by @xmath11 .
the scaling parameter @xmath4 is determined by the quality factor method , which does not rely on the shape of the scaling function .
we have argued that the pions , kaons and protons originate from different distributions of clusters formed by strings overlapping , and the scaling behaviours of these identified particles could be explained by the colour string percolation model in a quantitative way at the same time .
the author would like to thank prof . c. b. yang at central china normal university for valuable discussions .
this work was supported by the fundamental research funds for the central universities of china under grant no .
gk201502006 , by the scientific research foundation for the returned overseas chinese scholars , state education ministry , and by the national natural science foundation of china under grant nos .
11447024 and 11505108 .
00 hwa r c and yang c b 2003 _ phys . rev .
lett . _ * 90 * 212301 .
zhu l l and yang c b 2007 _ phys .
c _ * 75 * 044904 .
zhang w c , zeng y , nie w x , zhu l l and yang c b 2007 _ phys .
c _ * 76 * 044910 .
zhang w c and yang c b 2014 _ j. phys .
g : nucl . part .
phys . _ * 41 * 105006 .
gelis f et al .
2007 _ phys .
b _ * 647 * 376 - 379 .
beuf g et al .
2008 _ phys .
d _ * 78 * 074004 .
praszalowicz m 2013 _ phys .
b _ * 727 * 461 - 467 .
aamodt k et al .
( alice collaboration ) 2011 _ eur .
j. c _ * 71 * 1655 .
abelev b et al .
( alice collaboration ) 2014 _ phys .
b _ * 736 * 196 - 207 .
adam j et al .
( alice collaboration ) 2015 _ eur .
j. c _ * 75 * 226 .
olive k a et al .
( particle data group ) 2014 _ chin .
c _ * 38 * 090001 .
khandai p k et al .
2013 _ int .
* 28 * 1350066 .
praszalowicz m 2011 _ phys .
lett . _ * 106 * 142002 .
cunqueiro l et al .
2008 _ eur .
phys . j. c _ * 53 * 585 - 589 .
dias de deus j et al .
2005 _ eur .
j. c _ * 41 * 229 - 241 .
schwinger j 1951 _ phys .
rev . _ * 82 * 664 .
braun m a et al .
2015 _ phys .
rep . _ * 599 * 1 - 50 .
dias de deus j and pajares c 2006 _ phys .
b _ * 642 * 455 - 458 . | we extend the scaling behaviour observed in the inclusive charged hadron transverse momentum ( @xmath0 ) distributions to the @xmath0 spectra of pions , kaons and protons produced in proton - proton ( @xmath1 ) collisions with center of mass energies ( @xmath2 ) at 0.9 , 2.76 and 7 tev .
this scaling behaviour arises when a linear transformation , @xmath3 , is applied on the pion , kaon or proton @xmath0 spectra .
the scaling parameter @xmath4 depends on @xmath2 and is determined by a new method , the quality factor method , which does not rely on the shape of the scaling function .
we argue that the pions , kaons and protons originate from different distributions of clusters which are formed by strings overlapping , and the scaling behaviours of these identified particles @xmath0 spectra could be understood with the colour string percolation model in a quantitative way simultaneously . |
due to their importance in si technology , states induced by the dangling bonds ( dbs ) on si(100 ) surface , known as @xmath5 centers , have been a topic of study for decades@xcite .
these are important due to their role in introducing states inside the band gap@xcite , fermi level pinning@xcite , charging issue - consider variation of threshold voltage due to charged dbs , reliability issues - consider negative bias temperature instability@xcite , enhancement of 1/f noise due to trap states resulting in poor dc characteristics@xcite . with advances in nanoscale science , detailed understanding of these surface states is needed to exploit their role in possible technological applications , _
e.g. _ use of db as a template for the growth of molecular nanostructures@xcite .
this realization has been the motivation behind research like desorption of h from si surface forming a db@xcite , electronic and electrostatic effect of db on transport through a styrene wire@xcite , db dynamics on si surface@xcite , jahn - teller distortion in db wires at low temperature@xcite , magnetism in db structures@xcite , passivation of dbs with h and d@xcite , etc .
-4 in watanabe _ _ et al__@xcite predicted that an unpaired db ( si dimer with one hydrogenated atom ) wire on h : si(100)-(2@xmath01 ) surface in the direction along the dimer row would introduce states inside the band gap .
this prediction was later experimentally observed , however the mechanism is still not clear@xcite .
recently , cakmak _ _ et al__@xcite theoretically reproduced their results and found similar trend .
all these calculations were based on density functional theory ( dft ) where band gap is underestimated @xcite .
it is thus hard to quantitatively benchmark the results . in this paper
, we use extended hckel theory ( eht ) for electronic band structure calculations using transferable parameters developed by cerda _ _ et al__@xcite .
notably , eht gives correct si bulk band gap and other features as reported by kienle _
_ et al__@xcite .
it has been applied to si surfaces obtaining satisfactory results@xcite .
the state of the art in electronic structure calculations for si is gw approximation@xcite .
however , computational complexity prohibits its use in transport calculations for large systems like the ones discussed in this paper where the si atoms per unit cell may approach 384 .
eht provides computationally accessible , yet accurate approach .
apart from this , to the best of our knowledge there has not yet been a systematic study of isolated dbs , db wires and db clusters on h : si(100)-(2@xmath01 ) surface using a method that quantitatively addresses the surface states . in a previous study@xcite
, we report that a db electronically affects si atoms up to approximately 10@xmath4 away from it , by introducing a near - midgap state in the local density of states of the neighboring si atoms .
therefore , to simulate dbs in isolated , wired and clustered configurations , we use a large enough unit cell in the directions parallel to and perpendicular to the dimer row , conveniently referred to as @xmath3 $ ] and [ 110 ] directions respectively @xcite ( as shown in fig .
few large unit cells are shown in fig . 2 to demonstrate the approach .
this approach has been used previously in ref .
table i summarizes different surfaces being investigated .
we use electronic band structure principles in @xmath3 $ ] and [ 110 ] directions .
however in [ 100 ] direction , _
i.e. _ direction normal to surface , we simulate a semi - infinite structure by using sixteen layers of si atoms as shown in fig .
1(a - d ) .
the important aspect of this approximation is the use of a large enough unit cell in [ 100 ] direction , which minimizes the quantum confinement effects and the correct si band gap for h : si(100)-(2@xmath01 ) is obtained ( as shown in fig .
3(a ) ) . fig .
3(a ) also shows the @xmath1 and @xmath2 states for unpassivated asymmetric dimer ( ad ) and unpassivated sd surface , referred to as si(100)-(2@xmath01)-ad surface and si(100)-(2@xmath01)-sd surface respectively .
3(b ) shows transmission through the above - mentioned three surfaces .
h : si(100)-(2@xmath01 ) surface has a clean band gap , si(100)-(2@xmath01)-ad surface has 0.62ev of band gap and si(100)-(2@xmath01)-sd surface has no band gap .
we present calculations of isolated unpaired db ( si dimer with one atom hydrogenated ) and paired db ( unpassivated si dimer ) in fig .
4(a ) and ( b ) respectively .
an unpaired db introduces a near - midgap state as we previously report@xcite , whereas a paired db introduces a @xmath1 state close to the valence band maximum ( @xmath6 ) and a @xmath2 state close to the conduction band minimum ( @xmath7 ) .
these states due to a paired db are similar to the @xmath1 and @xmath2 states introduced by si(100)-(2@xmath01)-ad surface as shown in fig .
3(a ) , however dispersion is smaller than that of ad surface states .
we provide details of bandwidths of different states in section iv .
furthermore , we model db wires consisting of unpaired and paired dbs in @xmath3 $ ] and [ 110 ] directions as shown in fig .
4(a ) and ( b ) respectively .
for unpaired db wires , in @xmath3 $ ] and [ 110 ] direction , a near - midgap state similar to that of an isolated unpaired db is introduced .
however , the dispersion in the surface state due to an unpaired db wire in @xmath3 $ ] direction is larger than that of an isolated unpaired db , whereas the dispersion in surface states due to an unpaired db wire in [ 110 ] direction is only slightly larger .
similarly , for wires consisting of paired dbs in @xmath3 $ ] and [ 110 ] direction , @xmath1 and @xmath2 states are introduced close to @xmath6 and @xmath7 respectively , which are similar to those introduced by an isolated paired db .
the dispersion in surface states of these wires follow similar pattern as that of unpaired db wires in @xmath3 $ ] and [ 110 ] directions respectively . finally , we report electronic band structure calculations for isolated clusters of dbs as shown in fig .
5 , reporting that electronic structure of clusters can be interpreted as combinations of those of unpaired and paired dbs , but sufficiently dispersed due to interaction within the dbs .
this paper is organized in four sections .
first we talk about atomic structure and the assumptions made .
then we briefly discuss the theoretical approach .
finally , results and conclusions are presented .
.[tab : table1]details of the unit cells used to simulate different surfaces .
sixteen layer atomic structures shown in fig . 1 are used to construct bigger unit cells .
thus , a ( 4@xmath06 ) unit cell has dimensions equal to 15.36@xmath4 and 23.04@xmath4 in [ 110 ] ( perpendicular to dimer row ) and @xmath3 $ ] ( along dimer row ) directions respectively .
the total number of si atoms ( excluding h atoms on top and back surface ) per unit cell are given . [ cols="<,<,<,<",options="header " , ] -5.5 in
we use atomic structure of reconstructed si(100)-(2@xmath01 ) surface as reported by ramstad _
et al _ @xcite where the top four layers are relaxed due to surface reconstruction .
et al _ @xcite report structure for five layers of si(100)-(2@xmath01)-ad surface and si(100)-(2@xmath01)-sd surface .
we add eleven bulk layers to make a sixteen layer structure and finally passivate the bottom layer with h in dihydride configuration as shown in fig .
1(a ) and ( b ) respectively .
we then add h atoms to each of the si dimer atoms of si(100)-(2@xmath01)-sd structure , resulting in h : si(100)-(2@xmath01 ) structure as shown in fig .
furthermore , we remove one of the h atoms from h : si(100)-(2@xmath01 ) so that the resulting structure has an unpaired db as shown in fig . 1(d ) . fig .
1(e ) shows a h : si(100)-(2@xmath01 ) surface with directions along and perpendicular to the dimer rows referred to as @xmath3 $ ] and [ 110 ] respectively .
apart from this , we add h atoms with h - si bond length = 1.47@xmath4 and h - si dimer angle = @xmath8 .
the bond length and the angle were obtained by structural optimization of a @xmath9 cluster using local spin density approximation with 6 - 311 g * basis set as shown in fig .
1(f ) - calculations are performed using gaussian@xcite .
the atomic co - ordinates for si(100)-(2@xmath01)-ad unit cell ( fig .
1(a ) ) and h : si(100)-(2@xmath01 ) unit cell ( fig .
1(c ) ) are given here@xcite .
these four unit cells , shown in fig .
1(a - d ) , are used as building blocks to construct the bigger unit cell for doing isolated db , db wire and cluster calculations . in @xmath3
$ ] and [ 110 ] direction , the unit cell consists of varying layers of si atoms depending upon the requirement . for example , to study an isolated unpaired db , we need to include enough h : si layers in @xmath3 $ ] and [ 110 ] direction so that the dbs in adjacent unit cells do not influence each other .
we obtain this information from a previous study@xcite that a db affects neighboring atoms up to approximately 10@xmath4 .
table i summarizes the dimensions of the unit cells and the number of atoms used to simulate different surfaces .
three representative examples of the unit cells are shown in fig .
2 . we show ( 4@xmath04 ) , ( 4@xmath01 ) and ( 2@xmath04 ) unit cells in fig .
2(a - c ) , which are used for calculating electronic band structure of an isolated unpaired db and unpaired db wire in @xmath3 $ ] and [ 110 ] directions respectively . since , dbs
interact electronically over a distance of approximately 10@xmath4 , a ( 4@xmath04 ) unit cell makes certain that dbs in consecutive unit cells are electronically isolated .
the calculated electronic structure is thus that of an isolated db .
similarly , db wires and db clusters are also spaced sufficiently apart .
2(d ) shows part of the surface brillouin zone used in the calculation ( with symmetry points ( @xmath10)@xcite ) . for h :
si(100)-(2@xmath01 ) , cakmak _ et al _ @xcite report that the dimer distance is 2.35@xmath4 as compared to 2.23@xmath4 for si(100)-(2@xmath01)-sd as obtained by ramstad _ _ et al__@xcite - suggesting an increase of approximately 5% .
however , surface states after passivation are well below @xmath6 and hence do not affect the band gap . it has also been reported elsewhere@xcite that an unpaired db introduces a structural perturbation in h : si(100)-(2@xmath01 ) structure , which is weak .
we find that using this structure for si dimer , the surface state is shifted by approximately 20mev . at room temperatue
, this surface relaxation is anticipated to have small effect and hence in our calculations , we ignore it for simplicity .
furthermore , db wires could go through peierls distortion .
it has however been reported @xcite that the total energy gain due to such distortion is 14mev and is not anticipated to have a significant effect at room temperature .
moreover , jahn - teller distortion in db clusters at low temperature has been discussed in ref.@xcite .
the effects due to these structural changes are also anticipated to be small at room temperature . apart from this ,
the effect of dopant atoms and other defects ( both surface and bulk ) are ignored .
all the atomic visulalizations in this paper are done using gaussview@xcite .
we model the si surface by using a periodic system of unit cells repeating in @xmath3 $ ] and [ 110 ] directions .
the lattice is defined by @xmath11 , where @xmath12 and @xmath13 are lattice unit vectors in @xmath3 $ ] and [ 110 ] directions ; p and q are indices representing the repeating unit cells in @xmath12 and @xmath13 directions respectively .
the number of neighboring unit cells depend on the size of the unit cell .
hamiltonian ( h ) and overlap ( s ) matrices are computed in eht scheme using parameters developed by cerda _
et al _ @xcite . with periodic boundary conditions in
@xmath3 $ ] and [ 110 ] directions , fourier transform technique is used to transform h and s from the real space to the reciprocal ( k ) space following the scheme as follows , @xmath14 @xmath15 where @xmath16 , @xmath17 and @xmath18 are reciprocal lattice vectors of the surface brillouin zone in @xmath3 $ ] and [ 110 ] direction ; n represents the center unit cell ; m represents the neighboring unit cells ; n is the total number of unit cells ; @xmath19 is the displacement between neighboring ( @xmath20 ) and center ( @xmath21 ) unit cell .
finally , we calculate energy eigen values for different values of k along the surface brillouin zone symmetry points ( @xmath10 ) .
-3.25 in -3.25 in
3(a ) shows the electronic band diagram of h : si(100)-(2@xmath01 ) surface showing the surface and bulk states resulting in a clean band gap of 1.18ev . for an si(100)-(2@xmath01)-ad surface
, it has been experimentally known that the @xmath1 state is 0.2ev below @xmath6 and @xmath2 state is 0.4ev below @xmath7@xcite . these @xmath1 and @xmath2 states for si(100)-(2@xmath01)-ad surface
are shown in fig .
3(a ) , which are consistent with the experimental observation .
the bandwidth of these states is 1.22ev and 1.26ev respectively .
the band gap obtained in this case is 0.62ev .
similarly for si(100)-(2@xmath01)-sd surface , the @xmath1 and @xmath2 states disperse through the whole band gap with bandwidths of 1.39ev and 1.42ev respectively .
3(b ) gives a complementary view of fig .
3(a ) showing transmission through the above - mentioned three surfaces in the unit steps of @xmath22 .
4 shows calculations for an isolated unpaired db and unpaired db wire in @xmath3 $ ] and [ 110 ] direction .
an isolated unpaired db introduces a near - midgap state@xcite with a very small bandwidth of 0.11ev as shown in fig .
we attribute this small dispersion to the energetic isolation of this state from the conduction and the valence band states . for the db wire in @xmath3
$ ] direction , the bandwidth of the near - midgap state is 1.16ev , resulting in a very small band gap as shown in fig .
we interpret this large dispersion due to the dbs being in the close proximity of 3.84@xmath4 and hence interacting strongly .
for the db wire in [ 110 ] direction , the distance between dbs is 7.68@xmath4 . since dbs are farther apart in this case , the bandwidth is 0.26ev .
this bandwidth is much smaller than that of the wire in @xmath3 $ ] direction and larger than an isolated db as shown in fig .
4(b ) presents calculation for an isolated paired db and paired db wire in @xmath3 $ ] and [ 110 ] direction .
the electronic structure of a paired db is significantly different from an isolated unpaired db .
it introduces a bonding ( @xmath1 ) and anti - bonding ( @xmath2 ) state below @xmath6 and up to 0.28ev below @xmath7 respectively .
qualitatively , this behavior is similar to unpassivated reconstructed si surface .
however , dispersion of these states is much smaller .
the bandwidth of this @xmath1 and @xmath2 state is 0.14ev and 0.42ev respectively . the electronic band structure for a paired db wire in @xmath3 $ ] direction shows the similar @xmath1 and @xmath2 states with bandwidths of 1.17ev and 1.08ev respectively .
this large dispersion is due to strong coupling between the dbs in this configuration , being 3.84@xmath4 apart .
as expected , the dispersion of @xmath1 and @xmath2 states for db pair wire in [ 110 ] direction is less than that of @xmath3 $ ] direction and more than that of an isolated db pair .
the bandwidth is 0.29ev and 0.57ev for @xmath1 and @xmath2 state respectively . for an isolated paired db and paired db wire along [ 110 ] direction ,
only the state inside or close to band gap are shown , whereas , for the paired db wire along @xmath3 $ ] direction all the bulk and surface states are shown .
since the dimensions of the unit cell for configurations a , b and c are different , the symmetry points ( @xmath10 ) along the surface brillouin zone are labeled accordingly .
qualitatively , the results reported are similar to those of watanabe _ _ et al__@xcite for unpaired db wire in @xmath3 $ ] direction and unpaired and paired db wire in @xmath3 $ ] and [ 110 ] directions . however , there are some differences in the shape of dispersion for the paired db wire in these directions .
5 presents electronic band structure calculation for db clusters .
the qualitative behavior can be attributed to a combined behavior of individual unpaired and paired dbs , which are 3.84@xmath4 apart .
5(a ) shows states due to a cluster of unpaired and paired dbs , 3.84@xmath4 apart .
the resulting states consist of one near - midgap state with bandwidth of 0.12ev and two @xmath1 and @xmath2 state with bandwidth of 0.26ev and 0.22ev respectively .
qualitatively , we interpret these states as a combination of states due to an unpaired and paired db respectively . however , the dispersion is larger , which we attribute to the mutual couplings of these unpaired and paired dbs .
similarly , states due to a cluster of two paired dbs , 3.84@xmath4 apart , result in two @xmath1 and two @xmath2 states , as shown in fig
these states are similar to the ones due to isolated paired db . however , one of these @xmath1 states ( plotted as green dotted line ) has very small bandwidth of 83mev that needs to be explored further .
the bandwidth of other @xmath1 state is 0.25ev and that of two @xmath2 states is 0.2ev and 0.22ev .
5(b ) extends the cluster size by incorporating two clusters of ( i ) paired - unpaired - paired dbs , each 3.84@xmath4 away , and ( ii ) unpaired - paired - unpaired dbs , each 3.84@xmath4 away .
qualitatively , the behavior can again be interpreted as combination of surface states due to individual paired and unpaired dbs .
however , we again find a @xmath1 state of relatively small bandwidth of 76mev that needs further exploration .
we have reported the electronic band structure calculations showing differences between states induced in si band gap due to isolated db , db wires and clusters .
we extend on previous studies and report that the details of these states inside band gap vary considerably from isolated dbs to db wires and clusters .
an unpaired db behaves completely differently from a paired db .
similarly , isolated dbs are different from db wires and clusters . furthermore , we report that an isolated unpaired db and paired db have a very small dispersion in their induced states .
however , wires introduce a larger dispersion that can be attributed to the additional coupling between the dbs . similarly , db clusters broaden the induced states depending upon the coupling between the individual unpaired and paired dbs .
we have ignored contribution of dopant atoms , dephasing processes , further surface relaxation effects and charge transfer effects in our calculations .
we thank s. datta , g .- c .
liang , t. raza , d. kienle and k. h. bevan for useful discussions .
this work was supported by the nasa institute for nanoelectronics and computing ( inac ) .
computational facilities were provided by the nsf network for computational nanotechnology and nanohub.org@xcite .
m. j. frisch , g. w. trucks , h. b. schlegel , g. e. scuseria , m. a. robb , j. r. cheeseman , j. a. montgomery , jr .
, t. vreven , k. n. kudin , j. c. burant , j. m. millam , s. s. iyengar , j. tomasi , v. barone , b. mennucci , m. cossi , g. scalmani , n. rega , g. a. petersson , h. nakatsuji , m. hada , m. ehara , k. toyota , r. fukuda , j. hasegawa , m. ishida , t. nakajima , y. honda , o. kitao , h. nakai , m. klene , x. li , j. e. knox , h. p. hratchian , j. b. cross , c. adamo , j. jaramillo , r. gomperts , r. e. stratmann , o. yazyev , a. j. austin , r. cammi , c. pomelli , j. w. ochterski , p. y. ayala , k. morokuma , g. a. voth , p. salvador , j. j. dannenberg , v. g. zakrzewski , s. dapprich , a. d. daniels , m. c. strain , o. farkas , d. k. malick , a. d. rabuck , k. raghavachari , j. b. foresman , j. v. ortiz , q. cui , a. g. baboul , s. clifford , j. cioslowski , b. b. stefanov , g. liu , a. liashenko , p. piskorz , i. komaromi , r. l. martin , d. j. fox , t. keith , m. a. al - laham , c. y. peng , a. nanayakkara , m. challacombe , p. m. w. gill , b. johnson , w. chen , m. w. wong , c. gonzalez , and j. a. pople , _ gaussian 03 , revision b.03 _ ( gaussian , inc . , wallingford ct , 2004 ) .
atomic co - ordinates for si(100)-(2@xmath01)-ad unit cell - ( si , 1.16200 , 0.00000 , -0.92100 ) ; ( si , 3.30559 , 0.00000 , -0.21300 ) ; ( si , 0.06600 , 1.91979 , -1.49850 ) ; ( si , 3.74059 , 1.91979 , -1.46950 ) ; ( si , 1.95079 , 1.91979 , -2.95500 ) ; ( si , 5.73438 , 1.91979 , -2.71800 ) ; ( si , 1.90679 , 0.00000 , -4.22750 ) ; ( si , 5.75438 , 0.00000 , -4.07050 ) .
the remaining twenty four si and four h atoms have same co - ordinates as those of the last twenty four si and four h atoms of the h : si(100)-(2@xmath01 ) unit cell@xcite .
atomic co - ordinates for h : si(100)-(2@xmath01 ) unit cell - ( h , 0.31140 , -0.07020 , 0.85890 ) ; ( h , 3.52820 , -0.07020 , 0.85890 ) ; ( si , 0.80500 , 0.00000 , -0.52400 ) ; ( si , 3.03459 , 0.00000 , -0.52400 ) ; ( si , 0.07500 , 1.91979 , -1.49850 ) ; ( si , 3.76459 , 1.91979 , -1.49850 ) ; ( si ,
1.91979 , 1.91979 , -2.93100 ) ; ( si , 5.75938 , 1.91979 , -2.71000 ) ; ( si , 1.91979 , 0.00000 , -4.21150 ) ; ( si , 5.75938 , 0.00000 , -4.07050 ) ; ( si , 0.00000 , 0.00000 , -5.43000 ) ; ( si , 3.83959 , 0.00000 , -5.43000 ) ; ( si , 0.00000 , 1.91979 , -6.78750 ) ; ( si , 3.83959 , 1.91979 , -6.78750 ) ; ( si , 1.91979 , 1.91979 , -8.14500 ) ; ( si , 5.75938 , 1.91979 , -8.14500 ) ; ( si , 1.91979 , 0.00000 , -9.50250 ) ; ( si , 5.75938 , 0.00000 , -9.50250 ) ; ( si , 0.00000 , 0.00000 , -10.86400 ) ; ( si , 3.83959 , 0.00000 , -10.86400 ) ; ( si , 0.00000 , 1.91979 , -12.22150 ) ; ( si , 3.83959 , 1.91979 , -12.22150 ) ; ( si , 1.91979 , 1.91979 , -13.57900 ) ; ( si , 5.75938 , 1.91979 , -13.57900 ) ; ( si , 1.91979 , 0.00000 , -14.93650 ) ; ( si , 5.75938 , 0.00000 , -14.93650 ) ; ( si , 0.00000 , 0.00000 , -16.29400 ) ; ( si , 3.83959 , 0.00000 , -16.29400 ) ; ( si , 0.00000 , 1.91979 , -17.65150 ) ; ( si , 3.83959 , 1.91979 , -17.65150 ) ; ( si , 1.91979 , 1.91979 , -19.00900 ) ; ( si , 5.75938 , 1.91979 , -19.00900 ) ; ( si , 1.91979 , 0.00000 , -20.36650 ) ; ( si , 5.75938 , 0.00000 , -20.36650 ) ; ( h , 0.71959 , 0.00000 , -21.21520 ) ; ( h , 3.11999 , 0.00000 , -21.21520 ) ; ( h , 4.55918 , 0.00000 , -21.21520 ) ; ( h , 6.95958 , 0.00000 , -21.21520 ) . the unit cells for fig .
1(b , d ) can be generated by removing respective h atom(s ) . | we theoretically study the electronic band structure of isolated unpaired and paired dangling bonds ( db ) , db wires and db clusters on h : si(100)-(2@xmath01 ) surface using extended hckel theory ( eht ) and report their effect on the si band gap .
an isolated unpaired db introduces a near - midgap state , whereas a paired db leads to @xmath1 and @xmath2 states , similar to those introduced by an unpassivated asymmetric dimer ( ad ) si(100)-(2@xmath01 ) surface .
such induced states have very small dispersion due to their isolation from the other states , which reside in conduction and valence band . on the other hand , the surface state induced due to an unpaired db wire in the direction along the dimer row ( referred to as @xmath3 $ ] )
, has large dispersion due to the strong coupling between the adjacent dbs , being 3.84@xmath4 apart .
however , in the direction perpendicular to the dimer row ( referred to as [ 110 ] ) , due to the reduced coupling between the dbs being 7.68@xmath4 apart , the dispersion in the surface state is similar to that of an isolated unpaired db .
apart from this , a paired db wire in @xmath3 $ ] direction introduces @xmath1 and @xmath2 states similar to those of an ad surface and a paired db wire in [ 110 ] direction exhibits surface states similar to those of an isolated paired db , as expected . besides this
, we report the electronic structure of different db clusters , which exhibit states inside the band gap that can be interpreted as superpositions of states due to unpaired and paired dbs . |
contemporary scientific investigations frequently encounter a common issue of exploring the relationship between a response and a number of covariates . in machine learning research , the subject is typically addressed through learning a underling rule from the data that accurately predicates future values of the response .
for instance , in banking industry , financial analysts are interested in building a system that helps to judge the risk of a loan request .
such a system is often trained based on the risk assessments from previous loan applications together with the empirical experiences . an incoming loan request
is then viewed as a new input , upon which the corresponding potential risk ( response ) is to be predicted . in such applications ,
the predictive accuracy of a trained rule is of the key importance . in the past decade
, various strategies have been developed to improve the prediction ( generalization ) capability of a learning process , which include @xmath4 regularization as an well - known example @xcite .
the @xmath0 regularization learning prevents over - fitting by shrinking the model coefficients and thereby attains a higher predictive value . to be specific ,
suppose that the data @xmath6 for @xmath7 are collected independently and identically according to an unknown but definite distribution , where @xmath8 is a response of @xmath9th unit and @xmath10 is the corresponding @xmath11-dimensional covariates .
let @xmath12 be a sample dependent space ( sdhs ) with @xmath13 and @xmath14 being a positive definite kernel function .
the coefficient - based @xmath0 regularization strategy ( @xmath0 regularizer ) takes the form of @xmath15 where @xmath16 is a regularization parameter and @xmath17 @xmath18 is defined by @xmath19 with different choices of order @xmath1 , ( [ algorihtm ] ) leads to various specific forms of the @xmath20 regularizer .
in particular , when @xmath21 , @xmath22 corresponds to the ridge regressor @xcite , which smoothly shrinks the coefficients toward zero .
when @xmath23 , @xmath22 leads to the lasso @xcite , which set small coefficients exactly at zero and thereby also serves as a variable selection operator . when @xmath24 , @xmath22 coincides with the bridge estimator @xcite , which tends to produce highly sparse estimates through a non - continuous shrinkage .
the varying forms and properties of @xmath22 make the choice of order @xmath1 crucial in applications . apparently , an optimal @xmath1 may depend on many factors such as the learning algorithms , the purposes of studies and so forth .
these factors make a simple answer to this question infeasible in general . to facilitate the use of @xmath4-regularization , alteratively
, we intend to seek for a modeling strategy where an elaborative selection on @xmath1 is avoidable . specifically , we attempt to reveal some insights for the role of @xmath1 in @xmath4-learning via answering the following question : * problem 1 . *
are there any kernels such that the generalization capability of ( [ algorihtm ] ) is independent of @xmath1 ? in this paper , we provides a positive answer to problem 1 under the framework of statistical learning theory .
specifically , we provide a featured class of positive definite kernels , under which the @xmath20 estimators for @xmath5 attain similar generalization error bounds .
we then show that these estimated bounds are almost essential in the sense that up to a logarithmic factor the upper and lower bounds are asymptotically identical . in the proposed modeling context
, the choice of @xmath1 does not have a strong impact in terms of the generalization capability . from this perspective
, @xmath1 can be arbitrarily specified , or specified merely by other no generalization criteria like smoothness , computational complexity , sparsity , etc .. the reminder of the paper is organized as follows . in section 2 ,
we provide a literature review and explain our motivation of the research .
in section 3 , we present some preliminaries including spherical harmonics , gegenbauer polynomials and so on . in section 4
, we introduce a class of well - localized needlet type kernels of petrushev and xu @xcite and show some crucial properties of them which will play important roles in our analysis . in section 5 , we then study the generalization capabilities of @xmath0-regularizer associated with the constructed kernels for different @xmath1 . in section 6 , we provide the proof of the main results .
we conclude the paper with some useful remarks in the last section .
in practice , the choice of @xmath1 in ( [ algorihtm ] ) is critical , since it embodies certain potential attributions of the anticipated solutions such as sparsity , smoothness , computational complexity , memory requirement and generalization capability of course .
the following simple simulation illustrates that different choice of @xmath1 can lead to different sparsity of the solutions . the samples are identically and independently drawn according to the uniform distribution from the two dimensional sinc function pulsing a gaussian noise @xmath25 with @xmath26 .
there are totally 256 training samples and 256 test samples . in fig . 1
, we show that different choice of @xmath1 may deduce different sparsity of the estimator for the kernel @xmath27 .
it can be found that @xmath0 @xmath28 regularizers can deduce sparse estimator , while it impossible for @xmath29 regularizer .
learning schemes , width=302,height=226 ] therefore , for a given learning task , how to choose @xmath1 is an important and crucial problem for @xmath0 regularization learning . in other words
, which standards should be adopted to measure the quality of @xmath0 regularizers deserves study . as the most important standard of statistical learning theory , the generalization capability of @xmath0 regularization scheme ( [ algorihtm ] ) may depend on the choice of kernel , the size of samples @xmath30 , the regularization parameter @xmath31 , the behavior of priors , and , of course , the choice of @xmath1 . if we take the generalization capability of @xmath0 regularization learning as a function of @xmath1 , we then
automatically wonder how this function behaves when @xmath1 changes for a fixed kernel .
if the generalization capabilities depends heavily on @xmath1 , then it is natural to choose the @xmath1 such that the generalization capability of the corresponding @xmath0 regularizer is the smallest .
if the generalization capabilities is independent of @xmath1 , then @xmath1 can be arbitrarily specified , or specified merely by other no generalization criteria like smoothness , computational complexity , sparsity .
however , the relation between the generalization capability and @xmath1 depends heavily on the kernel selection . to show this
, we compare the generalization capabilities of @xmath29 , @xmath2 , @xmath32 and @xmath33 regularization schemes for two kernels : @xmath34 and @xmath35 in the simulation .
the one case shows that the generalization capabilities of @xmath0 regularization schemes may be independent of @xmath1 and the other case shows that the generalization capability of ( [ algorihtm ] ) depends heavily on @xmath1 . in the left of fig .
2 , we report the relation between the test error and regularization parameter for the kernel @xmath34 . it is shown that when the regularization parameters are appropriately tuned , all of the aforementioned regularization schemes may possess the similar generalization capabilities . in the right of fig .
2 , for the kernel @xmath35 , we see that the generalization capability of @xmath0 regularization depends heavily on the choice of @xmath1 .
regularization schemes with different @xmath1 . ] regularization schemes with different @xmath1 . ] from these simulations , we see that finding kernels such that the generalization capability of ( [ algorihtm ] ) is independent of @xmath1 is of special importance in theoretical and practical applications . in particular , if such kernels exist , with such kernels , @xmath1 can be solely chosen on the basis of algorithmic and practical considerations for @xmath0 regularization . here
we emphasize that all these conclusions can , of course only be made in the premise that the obtained generalization capabilities of all @xmath0 regularizers are ( almost ) optimal .
there have been several papers that focus on the generalization capability analysis of the @xmath0 regularization scheme ( [ algorihtm ] ) .
wu and zhou @xcite were the first , to the best of our knowledge , to show a mathematical foundation of learning algorithms in sdhs .
they claimed that the data dependent nature of the algorithm leads to an extra error term called hypothesis error , which is essentially different form regularization schemes with sample independent hypothesis spaces ( sihss ) .
based on this , the authors proposed a coefficient - based regularization strategy and conducted a theoretical analysis of the strategy by dividing the generalization error into approximation error , sample error and hypothesis error .
following their work , xiao and zhou @xcite derived a learning rate of @xmath2 regularizer via bounding the regularization error , sample error and hypothesis error , respectively .
their result was improved in @xcite by adopting a concentration inequality technique with @xmath29 empirical covering numbers to tackle the sample error . on the other hand , for @xmath0 @xmath36 regularizers , tong et al .
@xcite deduced an upper bound for generalization error by using a different method to cope with the hypothesis error .
later , the learning rate of @xcite was improved further in @xcite by giving a sharper estimation of the sample error . in all those researches , some sharp restrictions on the probability distributions ( priors ) have been imposed , say , both spectrum assumption of the regression function and concentration property of the marginal distribution should be satisfied . noting this , for @xmath29 regularizer , sun and
wu @xcite conducted a generalization capability analysis for @xmath29 regularizer by using the spectrum assumption to the regression function only .
for @xmath2 regularizer , by using a sophisticated functional analysis method , zhang et al .
@xcite and song et al .
@xcite built the regularized least square algorithm on the reproducing kernel banach space ( rkbs ) , and they proved that the regularized least square algorithm in rkbs is equivalent to @xmath2 regularizer if the kernel satisfies some restricted conditions . following this method , song and
zhang @xcite deduced a similar learning rate for the @xmath2 regularizer and eliminated the concentration property assumption on the marginal distribution .
limiting @xmath1 within @xmath37 $ ] is certainly incomplete to judge whether the generalization capability of @xmath0 regularization depends on the choice of @xmath1 .
moreover , in the context of learning theory , to intrinsically characterize the generalization capability of a learning strategy , the essential generalization bound @xcite rather than the upper bound is required , that is , we must deduce a lower and an upper bound simultaneously for the learning strategy and prove that the upper and lower bounds can be asymptotically identical .
we notice , however , that most of the previously known estimations on generalization capability of learning schemes ( [ algorihtm ] ) are only concerned with the upper bound estimation .
thus , their results can not serve the answer to problem 1 .
different from the pervious work , the essential bound estimation of generalization error for @xmath0 regularization schemes ( [ algorihtm ] ) with @xmath38 will be presented in the present paper . as a consequence
, we provide an affirmative answer to problem 1 .
in this section , we introduce some preliminaries on spherical harmonics , gegenbauer polynomial and orthonormal basis construction . , which will be used in the construction of the positive definite needlet kernel .
the gegenbauer polynomials are defined by the generating function @xcite @xmath39 where @xmath40 and @xmath41 .
the coefficients @xmath42 are algebraic polynomials of degree @xmath43 which are called the gegenbauer polynomials associated with @xmath44 .
it is known that the family of polynomials @xmath45 is a complete orthogonal system in the weighted space @xmath46 , @xmath47 $ ] , @xmath48 and there holds @xmath49 where @xmath50 define @xmath51 then it is easy to see that @xmath52 is a complete orthonormal system for the weighted @xmath53 space @xmath46 , where @xmath54 .
let @xmath55 be the unit ball in @xmath56 , @xmath57 be the unit sphere in @xmath56 and @xmath58 be the set of algebraic polynomials of degree not larger than @xmath43 defined on @xmath55 .
denote by @xmath59 the aero element of @xmath57 . then @xmath60 the following important properties of @xmath61
are established in @xcite .
[ lem11 ] let @xmath61 be defined as above .
then for each @xmath62 we have @xmath63 @xmath64 @xmath65 and @xmath66 where @xmath67 , and @xmath68 . for any integer @xmath69 ,
the restriction to @xmath70 of a homogeneous harmonic polynomial with degree @xmath71 is called a spherical harmonic of degree @xmath71 .
the class of all spherical harmonics with degree @xmath71 is denoted by @xmath72 , and the class of all spherical polynomials with total degrees @xmath73 is denoted by @xmath74 .
it is obvious that @xmath75 .
the dimension of @xmath72 is given by @xmath76 and that of @xmath74 is @xmath77 where @xmath78 denotes that there exist absolute constants @xmath79 and @xmath80 such that @xmath81
. the well known addition formula is given by ( see @xcite and @xcite ) @xmath82 where @xmath83 is arbitrary orthonormal basis of @xmath84 .
for @xmath85 and @xmath86 , we say that a finite subset @xmath87 is an @xmath88-covering of @xmath89 if @xmath90 where @xmath91 denotes the cardinality of the set @xmath92 and @xmath93 denotes the spherical cap with the center @xmath94 and the angle @xmath95 .
the following positive cubature formula can be found in @xcite . [
fixed cubature ] there exists a constant @xmath96 depending only on @xmath11 such that for any positive integer @xmath43 and any @xmath97-covering of @xmath89 satisfying @xmath98 .
there exists a set of numbers @xmath99 such that @xmath100 define @xmath101 then it follows from @xcite ( or @xcite ) that @xmath102 consists an orthonormal basis for @xmath58 , where @xmath103 of course , @xmath104 is an orthonormal basis for @xmath105 .
the following lemma [ reproducing kernel ] defines a reproducing kernel of @xmath106 , whose proof will be presented in appendix a. [ reproducing kernel ] the space @xmath107 is a reproducing kernel hilbert space .
the unique reproducing kernel of this space is @xmath108
in this section , we construct a concrete positive definite needlet kernel @xcite and show its properties .
a function @xmath109 is said to be admissible if @xmath110 @xmath111 , and @xmath109 satisfies the following condition @xcite : @xmath112,\eta(t)=1\ \mbox{on}\ [ 0,1],\ \mbox{and}\ 0\leq\eta(t)\leq 1\ \mbox{on}\ [ 1,2].\ ] ] such a function can be easily constructed out of an orthogonal wavelet mask @xcite .
we define a kernel @xmath113 as the following @xmath114 as @xmath115 is admissible , the constructed kernel @xmath116 called the needlet kernel ( or localized polynomial kernel ) @xcite henceforth , is positive definite .
we will show that so defined kernel function @xmath117 , deduces the @xmath0 regularization learning whose learning rate is independent of the choice of @xmath1 . to this end
, we first show several useful properties of the needlet kernel .
the following proposition [ prop1.0 ] which can be deduced directly from lemma [ reproducing kernel ] and the definition of @xmath115 reveals that @xmath118 possesses reproducing property for @xmath58 .
[ prop1.0 ] let @xmath118 be defined as in ( [ best kernel ] ) . for arbitrary @xmath119
, there holds @xmath120 since @xmath115 is an admissible function by definition , it follows that @xmath121 is an algebraic polynomial of degree not larger than @xmath122 for any fixed @xmath123 . at the first glance , as a polynomial kernel
, it may have good frequency localization property while have bad space localization property .
the following proposition [ prop2 ] , which can be found in ( * ? ? ?
* theorem 4.2 ) , however , advocates that @xmath118 is actually a polynomial kernel possessing very good spacial localized properties .
this makes it widely applicable in approximation theory and signal processing @xcite .
[ prop2 ] let @xmath118 be defined as in ( [ best kernel ] ) . for arbitrary @xmath124
, there exists a constant @xmath125 depending only on @xmath126 , @xmath11 and @xmath109 such that @xmath127 let @xmath128 be the best approximation error of @xmath58 .
define @xmath129 it has been shown in ( * ? ? ? * remak 4.8 ) that the integral operator @xmath130 possesses the following compressive property : [ prop3.0 ] if @xmath131 is defined as in ( [ best operator ] ) , then , for arbitrary @xmath132 , there exists a constant @xmath133 depending only on @xmath11 and @xmath134 such that @xmath135 by propositions [ prop1.0 ] , [ prop2 ] and [ prop3.0 ] , a standard method in approximation theory @xcite yields the following best approximation property of @xmath136 .
[ prop4 ] let @xmath137 and @xmath138 be defined in ( [ best operator ] ) , then for arbitrary @xmath139 , there exists a constant @xmath133 depending only on @xmath11 and @xmath134 such that @xmath140
in this section , we conduct a detailed generalization capability analysis of the @xmath0 regularization scheme ( [ algorihtm ] ) when the kernel function @xmath141 is specified as @xmath117 .
our aim is to derive an almost essential learning rate of @xmath0 regularization strategy ( [ algorihtm ] ) .
we first present a quick review of learning theory .
then , we given the main result of this paper , where a @xmath1-independent learning rate of @xmath0 regularization schemes ( [ algorihtm ] ) is deduced .
at last , we present some remarks on the main result .
let @xmath142 be an input space and @xmath143 an output space .
assume that there exists a unknown but definite relationship between @xmath144 and @xmath145 , which is modeled by a probability distribution @xmath146 on @xmath147 .
it is assumed that @xmath146 admits the decomposition @xmath148 let @xmath149 be a set of finite random samples of size @xmath30 , @xmath150 , drawn identically , independently according to @xmath146 from @xmath151 .
the set of examples @xmath152 is called a training set . without loss of generality
, we assume that @xmath153 almost everywhere .
the aim of learning is to learn from a training set a function @xmath154 such that @xmath155 is an effective estimate of @xmath156 when @xmath157 is given .
one natural measurement of the error incurred by using @xmath158 of this purpose is the generalization error , @xmath159 which is minimized by the regression function @xcite defined by @xmath160 we do not know this ideal minimizer @xmath161 , since @xmath146 is unknown , but we have access to random examples from @xmath162 sampled according to @xmath146 .
let @xmath163 be the hilbert space of @xmath164 square integrable functions on @xmath165 , with norm @xmath166 in the setting of @xmath167 , it is well known that , for every @xmath168 , there holds @xmath169 the goal of learning is then to construct a function @xmath170 that approximates @xmath161 , in the norm @xmath171 , using the finite sample @xmath152 .
one of the main points of this paper is to formulate the learning problem in terms of probability estimates rather than expectation estimates . to this end , we present a formal way to measure the performance of learning schemes in probability .
let @xmath172 and @xmath173 be the class of all borel measures @xmath146 on @xmath151 such that @xmath174 .
for each @xmath175 , we enter into a competition over all estimators established in the hypothesis space @xmath176 , @xmath177 and we define the accuracy confidence function by @xcite @xmath178 furthermore , we define the accuracy confidence function for all possible estimators based on @xmath30 samples @xmath179 by @xmath180 from these definitions , it is obvious that @xmath181 for all @xmath176 .
the sample dependent hypothesis space ( sdhs ) associated with @xmath113 is then defined by @xmath182 and the corresponding @xmath0 regularization scheme is defined by @xmath183 where @xmath184 the projection operator @xmath185 from the space of measurable functions @xmath186 to @xmath187 $ ] is defined by @xmath188 as @xmath189 $ ] by assumption , it is easy to check @xcite that @xmath190 also , for arbitrary @xmath191 , we denote @xmath192 .
we also need to introduce the class of priors . for any @xmath193 , denote by @xmath194 or @xmath195 the fourier transformation of @xmath158 , @xmath196 where @xmath197 .
the inverse fourier transformation will be denoted by @xmath198 . in the space @xmath105 , the derivative of @xmath158 with order @xmath199
is defined as @xmath200 where @xmath201 . here , fourier transformation and derivatives are all taken sense in distribution .
let @xmath95 be any positive number .
we consider the sobolev class of functions @xmath202 it follows from the well known sobolev embedding theorem that @xmath203 provided @xmath204 now , we state the main result of this paper , whose proof will be given in the next section .
[ thm1 ] let @xmath205 with @xmath206 , @xmath150 , @xmath175 be any numbers , and @xmath207 .
if @xmath22 is defined as in ( [ algorihtm1 ] ) with @xmath208 and @xmath38 , then there exist positive constants @xmath209 @xmath210 depending only on @xmath211 , @xmath146 , @xmath1 and @xmath11 , @xmath212 and @xmath213 satisfying @xmath214 such that for any @xmath215 , @xmath216 and for any @xmath217 , @xmath218 we explain theorem [ thm1 ] below in more detail . at first , we explain why the accuracy function is used to characterize the generalization capability of the @xmath0 regularization schemes ( [ algorihtm1 ] ) . in applications , we are often faced with the following problem : there are @xmath30 data available , and we are asked to product an estimator with tolerance at most @xmath219 by using these @xmath30 data only .
in such circumstance , we have to know the probability of success .
it is obvious that such probability depends on @xmath30 and @xmath219 .
for example , if @xmath30 is too small , we can not construct an estimator within small tolerance .
this fact is quantitatively verified by theorem [ thm1 ] .
more specifically , ( [ negative ] ) shows that if there are @xmath30 data available and @xmath205 with @xmath220 , then @xmath0 @xmath221 regularization scheme ( [ algorihtm1 ] ) is impossible to yield an estimator with tolerance error smaller than @xmath222 .
this is not a negative result , since we can see in ( [ negative ] ) also that the main reason of impossibility is the lack of data rather than inappropriateness of the learning scheme ( [ algorihtm1 ] ) .
more importantly , theorem [ thm1 ] reveals a quantitive relation between the probability of success and the tolerance error based on @xmath30 samples .
it says in ( [ theorem 1 ] ) that if the tolerance error @xmath219 is relaxed to @xmath223 or larger , then the probability of success of @xmath0 regularization is at least @xmath224 .
the first inequality ( lower bound ) of ( [ theorem 1 ] ) implies that such confidence can not be improved further .
that is , we have presented an optimal confidence estimation for @xmath0 regularization scheme ( [ algorihtm1 ] ) with @xmath38 .
thus , theorem [ thm1 ] basically concludes the following thing : if @xmath225 , then every estimator deduced from @xmath30 samples by @xmath0 regularization can not approximate the regression function with tolerance smaller than @xmath219 , while if @xmath226 , then the @xmath0 regularization schemes with any @xmath38 can definitely yield the estimators that approximate the regression function with tolerance @xmath219 . the values @xmath227 and @xmath223 thus are critical for indicating the generalization error of a learning scheme .
indeed , the upper bound of generalization error of a learning scheme depends heavily on @xmath223 , while the lower bound of generalization error is relative to @xmath227 .
thus , in order to have a tight generalization error estimate of a learning scheme , we naturally wish to make the interval @xmath228 $ ] as short as possible .
theorem [ thm1 ] shows that , for @xmath0 regularization scheme ( [ algorihtm1 ] ) , @xmath229 , and @xmath230 , which shows that the interval @xmath228 $ ] is almost the shortest one in the sense that up to a logarithmic factor , the upper bound and lower bound are asymptotical identical . noting that the learning rate established in theorem [ thm1 ] is independent of @xmath1
, we thus can conclude that the generalization capability of @xmath0 regularization does not depend on the choice of @xmath1 .
this gives an affirmative answer to problem 1 .
the other advantage of using the accuracy confidence function to measure the generalization capability is that it allows to expose some phenomenon that can not be founded if the classical expectation standard is utilized .
for example , theorem [ thm1 ] shows a sharp phase transition phenomenon of @xmath0 regularization learning , that is , the behavior of the accuracy confidence function changes dramatically within the critical interval @xmath228 $ ] .
it drops from a constant @xmath231 to an exponentially small quantity .
we might call @xmath228 $ ] the interval of phase transition for a corresponding learning scheme . to make this more intuitive ,
let us conduct a simulation on the phase transition of the confidence function below . without loss of generality , we implemented the @xmath29 regularization strategy ( [ algorihtm1 ] ) associated with the kernel ( [ best kernel ] ) for @xmath232 and @xmath233 to yield the estimator .
the regularization parameter @xmath31 was chosen as @xmath234 .
the training samples were drawn independently and identically according to the uniform distribution from the well known @xmath235 function , that is @xmath236 .
the number of the training samples @xmath30 was chosen from @xmath237 to @xmath238 and the tolerance @xmath219 was chosen from @xmath239 to @xmath237 with step - length @xmath239 .
then , there were totally 1000 test data @xmath240 drawn i. i. d according to the uniform distribution from @xmath241 .
the test error was defined as @xmath242 we repeated 100 times simulations at each point , and labeled its value as @xmath237 if @xmath243 is smaller than the tolerance error and @xmath244 otherwise .
simulation result is shown in fig.3 .
we can see from fig.3 that in the upper right part , the colors of all points are red , which means that in those setting , the probability that @xmath243 is smaller than the tolerance is approximately @xmath244 .
thus , if the number of samples is small , then @xmath29 regularization schemes can not provide an estimation with very small tolerance .
in the lower left area , the colors of all points are blue , which means that the probability of @xmath243 smaller than the tolerance is approximately @xmath237 .
between these two areas , there exists a band , that could be called the phase transition area , in which the colors of points vary from red to blue dramatically .
it is seen that the length of phase transition interval monotonously decreases with @xmath30 .
all these coincide with the theoretical assertions of theorem [ thm1 ] .
regularization , width=415,height=302 ] for comparison , we also present a generalization error bound result in terms of expectation error .
corollary [ thm2 ] below can be directly deduced from theorem [ thm1 ] and ( * ? ? ?
* chapter 3 ) , if we notice the identity : @xmath245 [ thm2 ] let @xmath205 with @xmath206 , @xmath246 , @xmath150 , and @xmath207 .
if @xmath22 is defined as in ( [ algorihtm1 ] ) with @xmath247 and @xmath38 , then there exist constants @xmath248 and @xmath249 depending only on @xmath211 , @xmath11 , @xmath1 and @xmath146 such that @xmath250 where @xmath251 is the set of all possible estimators based on @xmath30 samples . it is noted that the representation theorem in learning theory @xcite implies that the generalization capability of an optimal learning algorithm in sdhs is not worse than that of learning in rkhs with convex loss function .
corollary [ thm2 ] then shows that if @xmath205 , then the generalization capability of an optimal learning scheme in sdhs associated with @xmath118 is not worse than that of any optimal learning algorithms in the corresponding rkhs .
more specifically , ( [ generalization error ] ) shows that as far as the learning rate is concerned , all @xmath0 regularization schemes ( [ algorihtm1 ] ) for @xmath252 can realize the same almost optimal theoretical rate .
that is to say , the choice of @xmath1 has no influence on the generalization capability of the learning schemes ( [ algorihtm1 ] ) .
this also gives an affirmative answer to problem 1 in the sense of expectation .
here , we emphasize that the independence of generalization of @xmath0 regularization on @xmath1 is based on the understanding of attaining the same almost optimal generalization error .
thus , in application , @xmath1 can be arbitrarily specified , or specified merely by other no generalization criteria ( like complexity , sparsity , etc . ) .
the methodology we adopted in the proof of theorem [ thm1 ] seems of novelty .
traditionally , the generalization error of learning schemes in sdhs is divided into the approximation , hypothesis and sample errors ( three terms ) @xcite .
all of the aforementioned results about coefficient regularization in sdhs falled into this style . according to @xcite
, the hypothesis error has been regarded as the reflection of nature of data dependence of sdhs ( sample dependent hypothesis space ) , and an indispensable part attributed to an essential characteristic of learning algorithms in sdhs , compared with the learning in sihs ( sample independent hypothesis space ) . with the specific kernel function @xmath118
, we will divide the generalization error of @xmath0 regularization in this paper into the approximation and sample errors ( two terms ) only .
both of these two terms are dependent of the samples .
the success in this paper then reveals that for at least some kernels , the hypothesis error is negligible , or can be avoided in estimation when @xmath0 regularization learning are analyzed in sdhs .
we show that such new methodology can bring an important benefit of yielding an almost optimal generalization error bound for a large types of priors .
such benefit may reasonably be expected to beyond the @xmath0 regularization .
we sketch the methodology to be used as follows . due to
the sample dependent property , any estimators constructed in sdhs may be a random approximant . to bound the approximation error
, we first deduce a probabilistic cubature formula for algebraic polynomial .
then we can discretize the near - best approximation operator @xmath130 based on the probabilistic cubature formula .
thus , the well known jackson - type error estimate @xcite can be applied to derive the approximation error . to bound the sample error
, we will use a different method from the tranditional approaches @xcite .
since the constructed approximant in sdhs is a random approximant , the concentration inequality such as bernstein inequality @xcite can not be available . in our approach ,
based on the prominent property of the constructed approximant , we will bound the sample error by using the concentration inequality established in @xcite twice . then the relation between the so - called pseudo - dimension and covering number @xcite yields the sample error estimate for @xmath0 regularization schemes ( [ algorihtm1 ] ) with arbitrary @xmath253 .
hence , we divide the proof into four subsections . the first subsection is devoted to establish the probabilistic cubature formula .
the second subsection is to construct the random approximant and study the approximation error .
the third subsection is to deduce the sample error and the last subsectionis to derive the final learning rate .
we present the details one by one below . in this subsection , we establish a probabilistic cubature formula . at first , we need several lemmas . the weighted @xmath254 norm on the @xmath255-dimensional unit sphere @xmath256 is defined as follows .
let @xmath257 and @xmath258 .
define @xmath259 the following ( * ? ? ?
* lemma 2.3 ) gives a weighted nikolskii inequality for spherical polynomial .
[ weighted nikolskii ] let @xmath260 .
then for any @xmath261 , @xmath262 where @xmath133 is a positive constant depending only on @xmath263 and @xmath1 .
lemma [ relation ball sphere ] establishes a relation between cubature formula on the unit sphere and cubature formula on the unit ball , which can be found in ( * ? ? ?
* theorem 4.2 ) .
[ relation ball sphere ] if there is a cubature formula of degree @xmath43 on @xmath256 given by @xmath264 whose nodes are all located on @xmath256 , then there exists a cubature formula of degree @xmath43 on @xmath55 , that is , @xmath265 where @xmath266 are the first @xmath11 components of @xmath267 .
the following lemma [ bernstein ] is known as the bernstein inequality for random variables , which can be found in @xcite .
[ bernstein ] let @xmath94 be a random variable on a probability space @xmath151 with mean @xmath268 , variance @xmath269 .
if @xmath270 for almost all @xmath271 . then , for all @xmath175 , @xmath272 we also need a lemma showing that if @xmath273 is a set of independent random variables drawn identically according to a distribution @xmath44 , then with high confidence the cubature formula holds .
[ random cubature on sphere ] let @xmath274 , and @xmath275 . if @xmath276 are i.i.d .
random variables drawn according to arbitrary distribution @xmath44 on @xmath256 , then there exits a set of real numbers @xmath277 such that @xmath278 holds with confidence at least @xmath279 subject to @xmath280 * proof .
* for the sake of brevity , we write @xmath281 in the following . since the sampling set @xmath282 consists of a sequence of i.i.d .
random variables on @xmath283 , the sampling points are a sequence of functions @xmath284 on some probability space @xmath285 . without loss of generality , we assume @xmath286 for arbitrary fixed @xmath134 . if we set @xmath287 , then we have @xmath288 where we have used the equality @xmath289 furthermore , @xmath290 it follows from lemma [ weighted nikolskii ] that @xmath291 hence @xmath292 on the other hand , we have @xmath293 then using lemma [ weighted nikolskii ] again , there holds @xmath294 thus it follows from lemma [ bernstein ] that with confidence at least @xmath295 there holds @xmath296 this means that if @xmath282 is a sequence of i.i.d .
random variables , then the marcinkiewicz - zygmund inequality @xmath297 holds with probability at least @xmath298 then , almost same argument as that in ( * ? ? ?
* theorem 4.1 ) or ( * ? ?
* theorem 4.2 ) implies lemma [ random cubature on sphere ] .
@xmath299 by virtue of the above lemmas , we can prove the following proposition [ probabilistic cubature ] .
[ probabilistic cubature ] let @xmath275 and @xmath300 be a set of random variables independently and identically drawn according to arbitrary distribution @xmath44 .
then there exits a set of real numbers @xmath277 and a constant @xmath133 depending only on @xmath11 such that the equality @xmath301 holds with confidence at least @xmath302 subject to @xmath303 to estimate the upper bound of @xmath304 we first introduce an error decomposition strategy .
it follows from the definition of @xmath22 that , for arbitrary @xmath305 , @xmath306 since @xmath205 with @xmath307 , it follows from the sobolev embedding theorem that @xmath308 .
thus , it can be deduced from proposition [ prop3.0 ] and proposition [ prop4 ] that there exists a @xmath309 such that @xmath310}(f_\rho),\ ] ] where @xmath311 $ ] denotes the largest integer not larger than @xmath312 and @xmath313 denotes the uniform norm on @xmath55 .
the above inequalities together with the well known jackson inequality @xcite imply that there exists a @xmath309 such that for all @xmath205 with @xmath307 , there holds @xmath314 let @xmath315 , where @xmath316 is defined as in ( [ definitionc ] ) .
define @xmath317 then we have @xmath318 where @xmath319 and @xmath320 is called the approximation error and sample error , respectively .
[ approximation error ] let @xmath321 , @xmath206 and @xmath205 .
then , with confidence at least @xmath322 there holds @xmath323 where @xmath133 and @xmath316 are constants depending only on @xmath11 and @xmath95 . *
proof . * from proposition [ prop1.0 ] , it is easy to deduce that @xmath324 thus , lemma [ probabilistic cubature ] with @xmath325 yields that with confidence at least @xmath322 there exists a set of real numbers @xmath277 satisfying @xmath326 for @xmath327 such that @xmath328 the above observation together with ( [ app1 ] ) implies that with confidence at least @xmath322 there exists a @xmath329 such that for arbitrary @xmath205 , there holds @xmath330 and @xmath331 where @xmath133 is a constant depending only on @xmath11 and @xmath211 . indeed , if @xmath332 , we have @xmath333 . without loss of generality , we assume @xmath334
. then there holds @xmath335 if @xmath336 , it follows from the hlder inequality that @xmath337 thus , for all @xmath338 , there holds @xmath339 it thus follows from the definition of @xmath340 that the inequalities @xmath341 holds with confidence at least @xmath342 @xmath299 for further use , we also need introducing some quantities to measure the complexity of a space @xcite .
let @xmath343 be a banach space and @xmath344 a compact set in @xmath343 .
the quantity @xmath345 , where @xmath346 is the number of elements in least @xmath219-net of @xmath344 , is called @xmath219-entropy of @xmath344 in @xmath343 .
the quantity @xmath346 is called the @xmath219-covering number of @xmath344 . for any @xmath347 , define @xmath348 if a vector @xmath349 belongs to @xmath350 , then we denote by @xmath351 the vector @xmath352 @xmath353,@xmath354 .
the vc dimension of a set @xmath344 over @xmath55 , denoted as @xmath355 , is the maximal natural number @xmath30 such that there exists a collection @xmath356 in @xmath55 such that the cardinality of the sgn - vectors set @xmath357 equals to @xmath358 , that is , the set @xmath359 coincides with the set of all vertexes of unit cube in @xmath360 .
the quantity @xmath361 is called pseudo - dimension of the set @xmath344 over @xmath55 , where @xmath362 runs all functions defined on @xmath55 and @xmath363 .
mendelson and vershinin @xcite ( see also @xcite ) has established the following important relation between pseudo - dimension and @xmath219-entropy .
[ relation covering number ] let @xmath364 be a class of functions which consists of all functions @xmath365 satisfying @xmath366 for all @xmath367 . then , @xmath368 where @xmath316 is an absolute positive constant .
the following lemma [ covering number ] @xcite further shows that the pseudo - dimension of arbitrary @xmath30-dimensional vector space is @xmath30 .
[ covering number ] let @xmath176 be an @xmath30-dimensional vector space of functions from @xmath55 into @xmath369
. then @xmath370 .
we also need to apply the following concentration inequality @xcite .
[ concentration inequality ] let @xmath371 be a set of functions on @xmath151 such that , for some @xmath372 , @xmath373 almost everywhere and @xmath374 for each @xmath375 . then , for every @xmath175 , @xmath376 the following proposition [ sample error estimate ] give an upper bound of sample error .
[ sample error estimate ] let @xmath321 , @xmath175 , and @xmath377 be defined as in ( [ algorihtm1 ] ) . then with confidence at least @xmath378 there holds @xmath379 * proof .
* if we set @xmath380 and @xmath381 then @xmath382 both of which are random variables .
hence , we can rewrite the sample error as @xmath383 define @xmath384 as @xmath385 , it follows from ( [ algorihtm1 ] ) that @xmath386 which implies @xmath387 .
let @xmath388 then , for any fixed @xmath389 there exists @xmath390 such that @xmath391 .
it is easy to deduce that @xmath392 @xmath393 and @xmath394.\ ] ] since @xmath395 and @xmath396 almost everywhere , we find that @xmath397 of course , we have @xmath398 almost everywhere and @xmath399\\ & \leq & 16m^2\|\pi_mf - f_\rho\|_\rho^2=16m^2e(g),\end{aligned}\ ] ] therefore , we can apply lemma [ concentration inequality ] to the set of functions @xmath400 with @xmath401 , yielding @xmath402 with confidence at least @xmath403 for every @xmath404 , we have @xmath405 thus , a @xmath406-covering of @xmath407 provides an @xmath219-covering of @xmath408 for any @xmath175 .
this implies @xmath409 it is also needed to derive an upper bound estimation for @xmath410 . for @xmath327 , and @xmath411
, it follows from proposition [ prop2 ] and the hlder inequality that @xmath412 for @xmath336 , and @xmath411 , using ( [ kernelbound ] ) again we can obtain @xmath413 consequently , for arbitrary @xmath411 and arbitrary @xmath252 , there holds @xmath414 noting that @xmath415 is a finite dimensional linear space with its dimension not larger than @xmath416 , it follows from lemma [ covering number ] and lemma [ relation covering number ] that @xmath417 accordingly , @xmath418 which together with ( [ tools1 ] ) further yields @xmath419 with confidence at least @xmath420 now , we turn to estimate @xmath421 . by definition of @xmath422 , we have @xmath423 .
let @xmath424 then for any fixed @xmath425 there exists an @xmath426 such that @xmath427 .
similarly , we have @xmath428 since @xmath395 , @xmath396 and @xmath429 almost everywhere , we get @xmath430 almost everywhere . furthermore , @xmath431 then we apply lemma [ concentration inequality ] again to the set of functions @xmath371 with @xmath432 and obtain @xmath433 with confidence at least @xmath434 for every @xmath435 , we have @xmath436 thus , for any @xmath175 , a @xmath437-covering of @xmath438 provides an @xmath219-covering of @xmath371 .
this means @xmath439 by definition of @xmath440 , we then deduce from ( * ? ?
* theorem 5.3 ) that @xmath441 hence , @xmath442 which together with ( [ tools2 ] ) yields @xmath443 with confidence at least @xmath444 this finishes the proof of proposition [ sample error estimate ] .
@xmath299 now we are in a position to deduce the final learning rate of @xmath0 regularization schemes ( [ algorihtm1 ] ) .
firstly , it follows from propositions [ approximation error ] and [ sample error estimate ] that @xmath445 holds with confidence at least @xmath446 then , by setting @xmath447 , @xmath448 $ ] and @xmath208 , it follows from @xmath206 that @xmath449 that is , for @xmath450 @xmath451 holds with confidence at least @xmath452 . the lower bound can be more easily deduced .
actually , it follows from ( * ? ? ?
* equation ( 3.27 ) ) ( see also @xcite ) that for any estimator @xmath453 , there holds @xmath454 where @xmath455 and @xmath456 for some universal constant @xmath316 . with this , the proof of theorem [ thm1 ] is completed .
in studies and applications , regularization is a fundamental skill to improve on performance of a learning machine .
the @xmath0 regularization schemes ( [ algorihtm ] ) with @xmath38 are well known to be central in use . in this paper
, we have studied the dependency problem of the generalization capability of @xmath0 regularization with the choice of @xmath1 . through formulating a new methodology of estimation of generalization error
, we have shown that there is at least a positive definite kernel , say , @xmath457 , such that associated with such a kernel , the learning rate of the @xmath0 regularization schemes is independent of the choice of @xmath1 .
( to be more precise , we verified that with the kernel @xmath118 , all @xmath0 regularization schemes ( [ algorihtm ] ) can attain the same almost optimal learning rate in the following sense : up to a logarithmic factor , the upper and lower bounds of generalization error of the @xmath0 regularization schemes are asymptotically identical ) .
this implies that for some kernels , the generalization capability of @xmath0 regularization may not depend on @xmath1 .
therefore , as far as the generalization capability is concerned , for those kernels , the choice of @xmath1 is not important , which then relaxes the model selection difficulty in applications .
the problem is , however , far complicated .
we have also illustrated in section 2 that there exists a kernel with which the generalization capability of @xmath0 regularization heavily depends on the choice of @xmath1 .
thus , answering completely whether or not the choice of @xmath1 affects the generalization of @xmath0 regularization is by no means easy and completed . though we have constructed a concrete kernel example , the localized polynomial kernel @xmath118 , with which implementing the @xmath0 regularization in sdhs can realize the almost optimal learning rate , and this is independence of the choice of @xmath1 , we have not provided a practically feasible algorithm to implement the learning with the almost optimal generalization capability .
this is because the kernel @xmath118 we have constructed is not easily computed in practice , even though we can use the cubature formula ( lemma [ fixed cubature ] ) to discretize it .
thus , seeking the kernels that possesses the similar property as that of @xmath118 and can be implemented easily deserve study .
this is under our current investigation .
to prove lemma [ reproducing kernel ] , we need the following aronszajn theorem ( see @xcite ) . [ aronszajn ] let @xmath176 be a separable hilbert space of functions over @xmath165 with orthonormal basis @xmath458 .
@xmath176 is a reproducing kernel hilbert space if and only if @xmath459 for all @xmath144 .
the unique reproducing kernel @xmath141 is defined by @xmath460 * proof of lemma [ reproducing kernel ] . *
since @xmath102 is an orthonormal basis for @xmath58 , for arbitrary @xmath461 , there exists a set of real numbers @xmath462 such that @xmath463 where the summation concerning the index @xmath464 is @xmath465 . on the other hand
, it follows from ( [ basis ] ) that @xmath466 thus , the addition formula ( [ addition ] ) yields @xmath467 the above equality together with ( [ 3 ] ) and ( [ 4 ] ) implies @xmath468 therefore , there holds @xmath469 the above equality together with lemma [ aronszajn ] yields lemma [ reproducing kernel ] . | @xmath0-regularization has been demonstrated to be an attractive technique in machine learning and statistical modeling .
it attempts to improve the generalization ( prediction ) capability of a machine ( model ) through appropriately shrinking its coefficients .
the shape of a @xmath0 estimator differs in varying choices of the regularization order @xmath1 .
in particular , @xmath2 leads to the lasso estimate , while @xmath3 corresponds to the smooth ridge regression .
this makes the order @xmath1 a potential tuning parameter in applications . to facilitate the use of @xmath4-regularization
, we intend to seek for a modeling strategy where an elaborative selection on @xmath1 is avoidable . in this spirit
, we place our investigation within a general framework of @xmath4-regularized kernel learning under a sample dependent hypothesis space ( sdhs ) . for a designated class of kernel functions ,
we show that all @xmath4 estimators for @xmath5 attain similar generalization error bounds .
these estimated bounds are almost optimal in the sense that up to a logarithmic factor , the upper and lower bounds are asymptotically identical .
this finding tentatively reveals that , in some modeling contexts , the choice of @xmath1 might not have a strong impact in terms of the generalization capability . from this perspective
, @xmath1 can be arbitrarily specified , or specified merely by other no generalization criteria like smoothness , computational complexity , sparsity , etc .. * keywords : * learning theory , @xmath0 regularization learning , sample dependent hypothesis space , learning rate + * msc 2000 : * 68t05 , 62g07 . \1 .
institute for information and system sciences , school of mathematics and statistics , xian jiaotong university xian 710049 , p r china \2 .
the methodology center , the pennsylvania state university , department of statistics , 204 e. calder way , suite 400 , state college , pa 16801 , usa |
resonance interaction ( ri ) @xcite is a mechanism describing two quantum emitters ( one is excited and the other one is in ground state ) with same transition frequencies interact with each other by exchanging virtual photons .
ri represents the nature of the interaction between vacuum electromagnetic field and atomic system , and also plays important roles in broad application fields . for instance
, ri can contribute to operation of solid state lasers and fluorescent tubes @xcite .
in quantum physics , ri is very useful in creating entanglement entanmole , entanquad , and also plays important roles in laser cooling lasercool , creating cold molecules @xcite and quantum storage @xcite .
the study of ri can contribute in developing precise measurement methods in chemistry and biology for metering distance between molecules @xcite or between domains in one protein which can provide information about protein conformation @xcite , so it attracts the interests of chemists and biologists as well as physicists .
further more , ri between quantum emitters is the essence of frster resonance energy transfer ( fret ) @xcite , while the latter one lies at the core of biophysics cause it represents the nature of intermolecule interaction , provides the foundation of several detection methods @xcite and plays important role in biological phenomena such as photosynthesis . the investigation into photosynthesis can help in increasing the efficiencies of solar cells ( specially the dye - sensitized solar cells @xcite ) , which is in favor of the usage of solar energy . in biological phenomena ,
ri generally takes place in living cells filled with solution of organic compounds , which are optically active in most cases .
similarly , the electrolyte solution used in dye - sensitized solar cells contains volatile organic compounds @xcite and leads to an optically active surroundings .
when the dipoles ( which can be atoms , molecules , or just chromophores and fluorophores ) are immersed in optically active solutions , the interaction between them must be affected by the optically active surroundings .
the study of this effect can help in understanding the processes in natural photosynthesis , while the simulation of which can contribute in improving the efficiency of solar cells .
we think this effect is important as well as interesting , but so far as we all know , there is an absence of detailed theoretical investigation of this effect . in this paper
, we study the ri between two quantum dipoles which are immersed in optically active surroundings .
quantum electrodynamics is employed to deal with dipole - vacuum interaction .
we give out the analytical expressions of interdipole ri in optically active surroundings as well as several numerical examples .
we find that the collective behaviors of the dipole pair ( containing both the collective radiation rates and the interaction induced level shifts ) are affected by optically active surroundings while the single atom properties are not .
when the orientations of two dipoles are orthogonal and respectively perpendicular to interdipole axis , the optical rotation of the surrounding medium can cause great interaction between the dipoles while there is no interaction in vacuum .
in this section , a brief review of the dynamical analysis of dipole - reservoir interaction will be given out .
we calculate the dipole - light interaction problem in schrdinger picture .
the two dipoles are marked by 1 and 2 .
the quantum state of the system can be expressed by @xmath3 in which the basic states @xmath4 are @xmath5@xmath6 and @xmath7 are the quantum states of the dipoles .
@xmath8 represents that the dipole @xmath9 is in ground state and @xmath10 means dipole @xmath11 is in excited state . for example
, the ket @xmath12 represents dipole @xmath13 is excited while dipole @xmath14 is in ground state .
electromagnet field state @xmath15 means a photon with wave vector @xmath16 and polarization @xmath17 exists , while @xmath18 represents there is no photon ( the electromagnetic vacuum state ) . @xmath19 and @xmath20 are probability amplitudes .
the total hamiltonian @xmath21 can be departed into three parts : the electric dipole hamiltonian @xmath22 , the electromagnetic hamiltonian @xmath23 and the atom - light interaction hamiltonian @xmath24 .
the dipole hamiltonian can be written as @xmath25 in which @xmath26 are dipolar transition operators .
the electromagnetic hamiltonian is @xmath27where @xmath28 and @xmath29 are creation and annihilation operators for photon of mode @xmath30 , respectively .
@xmath31 is modified planck constant and @xmath32 is the angular frequency of the photon .
the interaction hamiltonian writes @xmath33where @xmath34 are the electric dipole operators , and @xmath35 are the dipole transition matrix elements .
@xmath36 is the electric field operator which is given by @xmath37where @xmath38 is the normalization coefficients , and the electric wave mode function @xmath39 is decided by the surrounding medium and the boundary conditions . with the initial condition
that dipole 1 is excited and dipole 2 is unexcited , the probability amplitude can be worked out : @xmath40the expressions of coefficients @xmath41 and @xmath42 as well as calculation process are given in appendix a. then we further define dipole - exchanging symmetric ( des ) state @xmath43 and dipole - exchanging anti - symmetric ( deas ) state @xmath44 which is given by @xmath45and
the corresponding probability amplitudes are @xmath46it is obviously that @xmath47 and @xmath48 are the damping rates and the level shifts of des and deas states , respectively .
@xmath41 refers to single dipole mechanism and @xmath42 represents collective behavior .
according to the solution we get above , the dipole - dipole interaction energy can be calculated as @xmath49the interdipole interaction energy is proportional to the collective level shift ( @xmath50 ) times the population difference between des and deas states .
in last section , we have given out a brief review of dipole - vacuum interaction .
the general formulas are calculated , but to obtain the detail expressions , the mode function @xmath51 must be specified . in this section , we will give the function @xmath52 in optically active medium .
one beam of left ( right ) hand circularly polarized light with wave vector @xmath53 can be expressed by jones vector as @xmath54 @xmath55 and @xmath56 are unit vectors which are orthogonal to each other and perpendicular to wave vector @xmath16 .
@xmath9 is the imaginary unit . according to fresnel s theorem ,
the phase velocities of left hand and right hand polarized light is different in optically active medium , then the general expression of light in optically active medium can be written as @xmath57 in which @xmath58 , @xmath59 represent the left and right hand circularly polarized components , respectively , and @xmath60 , @xmath61 .
@xmath62 are the refractive indexes for the two circularly polarized components and @xmath53 is wave vector in free space . using eq .
( [ wge ] ) , the expression of function @xmath63 in eq .
( [ aij ] ) can be derived @xmath64 in which @xmath65.\ ] ]
after defining the functions @xmath66 , eqs .
( [ aij ] ) can be calculated .
we shall go to the continuum limit@xmath67 and make use of eq .
( [ g1112 ] ) to obtain the analytical expressions of @xmath68 and @xmath42 @xmath69 in which @xmath70 \nonumber \\ & & -\sum_{\lambda } \frac{3n_{\lambda } } { 8}\boldsymbol{\hat{d}}_{2}\cdot \boldsymbol{\hat{r}\hat{r}}\cdot \boldsymbol{\hat{d}}_{1}\left [ \frac{\sin n_{\lambda } k_{0}r}{n_{\lambda } k_{0}r}+3\frac{\cos n_{\lambda } k_{0}r}{% \left ( n_{\lambda } k_{0}r\right ) ^{2}}-3\frac{\sin n_{\lambda } k_{0}r}{% \left ( n_{\lambda } k_{0}r\right ) ^{3}}\right ] \nonumber \\ & & + \sum_{\lambda } \frac{3n_{\lambda } } { 8}s_{\lambda } \left ( \boldsymbol{% \hat{d}}_{2}\times \boldsymbol{\hat{d}}_{1}\right ) \cdot \boldsymbol{\hat{r}}% \left [ \frac{\cos n_{\lambda } k_{0}r}{n_{\lambda } k_{0}r}-\frac{\sin n_{\lambda } k_{0}r}{\left ( n_{\lambda } k_{0}r\right ) ^{2}}\right ] , \label{f1}\end{aligned}\]]and @xmath71 \nonumber \\ & & -\sum_{\lambda } \frac{3n_{\lambda } } { 8}\boldsymbol{\hat{d}}_{2}\cdot \boldsymbol{\hat{r}\hat{r}}\cdot \boldsymbol{\hat{d}}_{1 } \left [ \frac{\cos n_{\lambda } k_{0}r}{n_{\lambda } k_{0}r } -3\frac{\sin n_{\lambda } k_{0}r } { \left ( n_{\lambda } k_{0}r\right ) ^{2 } } -3\frac{\cos n_{\lambda } k_{0}r } { \left ( n_{\lambda } k_{0}r\right ) ^{3 } } \right ] \nonumber \\ & & -\sum_{\lambda } \frac{3n_{\lambda } } { 8}s_{\lambda } \left ( \boldsymbol{% \hat{d}}_{2}\times \boldsymbol{\hat{d}}_{1}\right ) \cdot \boldsymbol{\hat{r}}% \left [ \frac{\sin n_{\lambda } k_{0}r}{n_{\lambda } k_{0}r } + \frac{\cos n_{\lambda } k_{0}r}{\left ( n_{\lambda } k_{0}r\right ) ^{2 } } \right ] \nonumber \\ & & -\sum_{\lambda } \frac{3n_{\lambda } } { 8}s_{\lambda } \left ( \boldsymbol{% \hat{d}}_{2}\times \boldsymbol{\hat{d}}_{1}\right ) \cdot \boldsymbol{\hat{r}}% \frac{2}{\pi } \left [ \frac{1}{n_{\lambda } k_{0}r}i_{1 } \left(n_{\lambda } k_{0}r\right)+\frac{1}{\left ( n_{\lambda } k_{0}r\right ) ^{2}}i_{2 } \left(n_{\lambda } k_{0}r\right)\right ] .
\label{f2}\end{aligned}\ ] ] we have defined @xmath72 which is the spontaneous radiation rate of a single dipole in free space . @xmath73 where @xmath74 is the resonance transition angular frequency of the dipoles and @xmath75 is the vacuum light speed .
@xmath76 is the electric transition dipole momentum , and @xmath77 is the vacuum permittivity .
@xmath78 are unit vectors that represent the polarization orientation of dipoles .
the interdipole distance @xmath79 , and the unit vector @xmath80 represents the interdipole axis . the two functions @xmath81 and @xmath82 in eq .
( [ f2 ] ) are improper integrals defined by @xmath83 up to here , the analytical expressions of @xmath41 and @xmath42 have been calculated .
associated with eqs .
( [ cs ] ) - ( [ ei ] ) , the behaviors of the dipole pair can be fully described . according to eq .
( [ cs ] ) , we obtain the damping rates of des and deas states @xmath84 as well as the level shifts due to the interdipole interaction @xmath85
considering a situation that @xmath86 , then @xmath87 and @xmath88 while @xmath89 .
the physical picture of this result is that dipole 1 spontaneously emits photon while dipole 2 can never be excited .
when the two dipole are very far away separated , they do not interact with each other and the population ( @xmath90 ) on the excited state of dipole @xmath13 ( the initially excited dipole ) decays in a rate of @xmath91 .
that is why we say the single dipole mechanism is decided by @xmath41 . by defining the average refractive index @xmath92 we find the spontaneous radiation rate of one single dipole immersed in optically active media is @xmath93 similarly , the imaginary part of @xmath41 is the vacuum lamb shift of a single dipole , which is obviously divergent according to eq .
( [ al ] ) .
this divergent integral can be renormalized by bethe s method @xmath94 in which @xmath95 is the observable mass of an electron . using the expression of eq .
( [ renorm ] ) , we get the vacuum lamb shift of a single dipole in optically active medium @xmath96 @xmath97 is position independent , which means it will not lead to any mechanical force . for this reason
, we will not discuss the single dipole vacuum lamb shift in the follows .
@xmath41 represents the interaction between one single dipole and the vacuum electromagnetic field . according the analysis above ,
the radiation rate and level shift of a single dipole depend on the average rather then the difference of the refractive indexes for the two circularly polarized components , and the latter one decides the specific rotation .
for this reason , we can say that the behaviors of a single dipole are affected by the average refractive indexes rather than the optical rotation of the surroundings . according to eqs .
( [ dr ] ) and ( [ ls ] ) , we have obtained the damping rates @xmath98 as well as the level shifts @xmath99 of des and deas states .
the total damping rates ( and level shifts ) are the single dipole part @xmath41 plus or minus the collective part @xmath42 .
the value of @xmath42 is decided by the interaction between the dipoles via exchanging virtual photons , therefore it describes the collective behaviors of the dipole pair . by checking eqs .
( [ f1 ] ) and ( [ f2 ] ) , the most notable properties is that in optically active surroundings .
the cross product ( @xmath100 ) of dipole 1 and dipole 2 contributes to the interdipole interaction , while this term does not appear in optically inactive case .
this result is not difficult to understand .
imagining one situation that the orientations of the two dipoles are perpendicular to each other . if the dipoles are located in optically inactive surroundings , when one virtual photon are emitted by dipole 1 , the polarization of the photon is also exactly perpendicular to dipole 2 and it is impossible to be absorbed .
for the two dipoles can just interact with each other by exchanging virtual photons , they can not feel " each other in this perpendicular orientation situation .
but when the dipoles are immersed in optically active surroundings , the polarization of the virtual photon emitted by the donor dipole will be rotated in propagation , and it may has a component parallel to the acceptor dipole then the photon can be absorbed .
this effect make term ( @xmath101 ) appears in eqs .
( [ f1 ] ) and ( [ f2 ] ) , which respectively represent two - dipole collective radiation rate and level shift . furthermore , according to eqs .
( [ f1 ] ) and ( [ f2 ] ) , the two circularly polarized components contribute differently to the interdipole interaction because of the different refractive indexes .
to visually show the results given above , we will give out three numerical examples in the follows .
we plot the damping rates and the level shifts as functions of interdipole distance and compare the different behaviors of the dipole pair in optically active and inactive surroundings . to specify the parameters
, we set the average refractive index of the surrounding medium to be @xmath102 , and the specific rotation divided by wave vector @xmath103 to be @xmath104 .
we first consider a situation that the two dipoles are orthogonally polarized , and the orientations of them are perpendicular to the interdipole axis , as shown in fig .
[ chuizhi ] . according to the physical analysis above , two
orthogonally polarized dipoles in vacuum can not interact with each other while an optically active surroundings can help in establishing interdipole ri .
to check this analysis , we plot the level shifts as well as the radiation rates in optically inactive and active surroundings .
the results in the different situations are compared in fig .
[ chuizhi ] . represents the damping rate of the population of des state defined in eq .
( [ symstate ] ) while @xmath105 refers to deas state . in the optically ( a ) inactive case ,
the level shift is a constant which means the two dipoles do not interaction with each other .
when the dipoles are immersed in ( b ) optically active medium , the ri can be built up by exchanges of virtual photons between the two dipoles . furthermore , the dimensionless dipole - dipole ri @xmath106 at time @xmath107 are plotted as functions of interdipole distance in ( c ) .
@xmath108 represents the dimensionless energy level difference between the des and deas states , while these two states are degenerate without interdipole interaction . ,
scaledwidth=70.0% ] in the optically inactive case , according to fig .
[ chuizhi](b ) , the level shift @xmath109 is zero , which means the des and deas states are degenerate and the two dipoles do not interact with each other .
the damping rates of des and deas state are @xmath110 , which indicate that dipole @xmath13 can not feel the existence of dipole @xmath14 and just decays alone . in the optically active case ,
the damping rates and level shifts vary with the interdipole distance in fig .
chuizhi(c ) , which implies an ri between the dipoles .
we plot the interdipole ri in the case of perpendicular dipole orientations as functions of interdipole distance in fig .
[ chuizhi](c ) . for ri decays with time , we choose a time at @xmath107 , and use @xmath111 as the unit of interaction energy .
the results of this numerical experiment agree with the physical analysis above , and typically shows that optically active surroundings can noticeably affect interdipole ri .
now we further consider a situation that the polarizations of the two dipoles are syntropic and perpendicular to the interdipole axis , as shown in fig .
[ pingxin ] .
are noticeably different in optically ( a ) inactive and ( b ) active cases .
the dimensionless dipole - dipole ri @xmath112 at time @xmath107 are plotted as functions of interdipole distance in ( c ) .
ri behaves rather differently in optically active and inactive surroundings .
, scaledwidth=70.0% ] the level shifts and the radiation rates in optically inactive and active cases are compared in fig .
[ pingxin](a ) and ( b ) .
we see that for the optically inactive case , the first point of minimum of the level shift @xmath113 appears in the neighbourhood of @xmath114 while it does near @xmath115 for the optically active case . obviously that the curves in fig .
[ pingxin](a ) and fig .
[ pingxin](b ) are different , which means the resonance interaction of the dipoles are affected by the optical rotation of the medium .
the dimensionless interaction energy @xmath116 in the case of syntropic dipole orientations at time @xmath107 are plotted in fig .
[ pingxin](c ) .
next we consider a situation that the dipoles are isotropically polarized , namely @xmath117 , as shown in fig .
[ isotropic ] . ) as shown in ( d ) .
the variations of collective radiation rates and level shift with respect to interdipole distance @xmath59 are noticeably different in optically ( a ) inactive and ( b ) active cases .
the dimensionless dipole - dipole ri @xmath106 at time @xmath107 are plotted as functions of interdipole distance in ( c ) .
ri behaves rather differently in optically active and inactive surroundings.,scaledwidth=70.0% ] the level shifts and the radiation rates in optically inactive and active cases are plotted in fig .
[ isotropic](a ) and ( b ) , which behave differently . for example , when the interdipole separation is about @xmath118 , the deas state decays faster in optically inactive surroundings while the des state does in optically active case .
the dimensionless interaction energy @xmath116 in the case of isotropic polarization at time @xmath107 are plotted in fig .
[ isotropic](c ) .
these three numerical examples typically show the effect on ri of optically active surroundings . in studying ri or using it to develop measurement methods , the optical rotation of surroundings in which dipoles are immersed can not be ignored .
in this paper , we have calculated the ri between two electric dipoles immersed in absorption free optically active surroundings .
the analytical expressions of the collective radiation rates and level shifts are given out , which satisfy general dipole orientations and interdipole distances .
considering the optical rotation of surrounding medium , the cross product of two electric dipole moment appears while this term does not contribute to interdipole ri in optically inactive case . by comparing the ri in optically inactive and active cases numerically , we obtain that the interdipole ri are noticeably affected by the optical rotation of surroundings .
especially , in the case of that the orientations of two dipoles are orthogonal , the optical rotation of surrounding medium can help in building up interdipole ri while there is no interaction in situation of free space .
our result can help in studying the nature of ri between electric dipoles immersed in optically active medium , and may contribute in developing accurate methods of measurement in chemistry and biology .
this work was supported by the national natural science foundation of china under grant nos . 11474095 , 11274118 , 11234003 , and 11129402 , and the fundamental research funds for the central universities .
to get eq .
( [ c1c2 ] ) , we make use of schrdinger equation @xmath119 to get the equations of motion of the probability amplitudes @xmath19 and @xmath120 @xmath121to solve these equations , we shall first calculate the formal solutions of eqs .
( [ d3 ] ) and ( [ d4 ] ) @xmath122 , \label{d3mar } \\
d_{4,\boldsymbol{k}\lambda } \left ( t\right ) = \frac{i}{\hbar } a_{k}^{\ast } \sum_{a=1,2}\sum_{i^{\prime } = 1,2}\boldsymbol{d}_{4i^{\prime } } ^{a}\cdot \boldsymbol{w}^{\ast } \left ( \boldsymbol{k}\lambda , \boldsymbol{r}% _ { a}\right ) c_{i^{\prime } } \left ( t\right ) \left [ \pi \delta \left ( \omega _ { k}-\omega _ { 0}\right ) -\frac{i}{\left ( \omega _ { k}-\omega _ { 0}\right ) } % \right ] , \label{d4mar}\end{aligned}\]]in which the markov approximation has been used @xmath123.\]]substituting eqs .
( [ d3mar ] ) and ( [ d4mar ] ) into eqs .
( [ c1 ] ) and ( [ c2 ] ) , we can get the equations of @xmath124 and @xmath125 @xmath126and the coefficients @xmath127 is defined by @xmath128 in which @xmath129 .
the @xmath130 functions are defined as @xmath131 .
the dipole transition matrix elements @xmath132 have been denoted as @xmath133 for simplicity .
the initial state is @xmath134 and @xmath135 , we further define @xmath136 and it is not difficult to get solution ( [ c1c2 ] ) by solving eq . | we study the resonance interaction between two quantum electric dipoles immersed in optically active surroundings .
quantum electrodynamics is employed to deal with dipole - vacuum interaction .
our results show that the optical activity of surroundings will not change the single atom behaviors while it can change the collective behaviors of the two dipoles , as well as greatly affect the dipole - dipole resonance interaction . especially ,
if the orientations of two dipoles are orthogonal and respectively perpendicular to the interdipole axis , the interdipole resonance interaction can be established with the help of optically active surroundings while there is no resonance interaction in vacuum .
@xmath0state key laboratory of precision spectroscopy , department of physics , east china normal university , 500 dongchuan road shanghai 200241 , people s republic of china .
@[email protected]
99 d. p. craig , and t. thirunamachandran , _ molecular quantum electrodynamics _ ( academic press , 1984 ) .
h. margenau , and n. r. kestner , _ theory of intermolecular forces _
( elsevier , 1969 ) .
f. t. rabouw , s. a. den hartog , t. senden and a. meijerink , photonic effects on the frster resonance energy transfer efficiency , " nature communications * 5 * , 3610 ( 2014 ) . g. k. brennen , i. h. deutsch , and p. s. jessen , entanglement and quantum phase transition in the one - dimensional anisotropic @xmath2 model , " phys . rev .
a * 61 * , 062309 ( 2000 ) .
h. matsueda , _ coherence and statistics of photons and atoms _ , edited by j. perina ( wiley , new york , 2001 ) .
w. d. phillips , nobel lecture : laser cooling and trapping of neutral atoms , " rev .
mod .
phys . * 70 * , 721 ( 1998 ) .
c.m .
dion , c. drag , o. dulieu , b. laburthe tolra , f. masnou - seeuws , p. pillet , resonant coupling in the formation of ultracold ground state molecules via photoassociation , " phys .
rev .
lett . * 86*(11 ) , 2253 ( 2001 ) .
j. vala , o. dulieu , f. masnou - seeuws , p. pillet , r. kosloff , coherent control of cold - molecule formation through photoassociation using a chirped - pulsed - laser field , " phys .
rev .
a * 63*(1 ) , 013412 ( 2000 ) .
a. kalachev , and s. krll , coherent control of collective spontaneous emission in an extended atomic ensemble and quantum storage , " phys .
rev .
a * 74*(2 ) , 023814 ( 2006 ) .
r. h. dicke , coherence in spontaneous radiation processes , " phys .
rev . * 93 * , 99 ( 1954 ) .
m. o. scully , collective lamb shift in single photon dicke superradiance , " phys .
rev .
lett .
* 102*(14 ) , 143601 ( 2009 ) .
j. zheng , spectroscopy - based quantitative fluorescence resonance energy transfer analysis , " in _ ion channels : methods and protocols .
methods in molecular biology _ , n. gamper , ed .
( humana press , 2006 ) .
k. truong , and m. ikura , the use of fret imaging microscopy to detect protein protein interactions and protein conformational changes in vivo , " current opinion in structural biology * 11*(5 ) , 573 ( 1991 ) .
h. volkhard , fluorescence resonance energy transfer , " in _ principles of computational cell biology _ , h. kiaris , ed .
( wiley - vch , 2008 ) . th .
frster , _ modern quantum chemistry _ , o. sinanoglu , ed .
( academic , 1965 ) , pt .
3 . b. a pollok , and r. heim , using gfp in fret - based applications , " trends in cell biology * 9*(2 ) , 57 ( 1999 ) .
r. b. sekar , and a. periasamy , fluorescence resonance energy transfer ( fret ) microscopy imaging of live cell protein localizations , " the journal of cell biology * 160*(5 ) , 629 - 633 ( 2003 ) .
b. oregan , and m. grtzel , a low - cost , high - efficiency solar cell based on dye - sensitized colloidal tio2 films , " nature * 353 * , 737 - 740 ( 1991 ) .
shu - ming yang , _ dye - sensitized nanocrystalline photovoltaic solar cells _ ( zhengzhou univerdity press , 2007 ) . |
the theory of jacobian elliptic functions , for reference see @xcite , yields a complete calculation for the motion of a pendulum as a function of time .
this is revisited here , to exhibit a few interesting properties of the elliptic integrals .
write the pendulum energy , with inertia moment @xmath0 and gravity constant @xmath1 ( rather than the usual @xmath2 ) , in the canonical coordinates @xmath3 or as -i g^2(1-)h(b,)where the origin in @xmath4 is set at the unstable equilibrium : the definition implies that @xmath2 has dimension of inverse time and the lyapunov exponents of the unstable equilibrium are @xmath5 . @xmath6 and @xmath4 are canonical coordinates for the motions .
it is well known that near the unstable equilibrium of the pendulum @xmath7 it is possible to define a canonical transformation , mapping the origin into itself , introducing new local cordinates @xmath8 such that b = r_c(p , q),=s_c(p , q)with @xmath9 holomorphic in a polidisk @xmath10 with @xmath11 , and in terms of which the motion near @xmath12 is described by a hamiltonian @xmath13 depending on the product @xmath14 only , of the form @xmath15 with @xmath16 .
the purpose of this note is to derive a proof of the latter statement via the theory of elliptic functions : this is not the simplest approach if one is just interested to know the existence of normal hyperbolic coordinates : existence of @xmath9 could be easily established without deriving their `` explicit '' expressions for @xmath17 and @xmath18 in terms of elliptic functions .
here we also correct a few errors in the earlier attempt made in ( * ? ? ?
* appendix 9 ) . the natural correspondence between the hyperbolic fixed point of the pendulum and its elliptic fixed point is briefly reported in appendix b and leads to the construction of the normal canonical coordinates for the small oscillations , hence to the action - angle variables .
motions near the unstable equilibrium have a quite different nature depending on the sign of the total energy @xmath19 : the ones with @xmath20 are `` oscillations '' ( their motions do not encompass the full perimeter of the circle ) while the ones with @xmath21 are `` librations '' .
therefore it will not be possible to find global action - angle coordinates : motions near the separatrix ( which with our conventions has @xmath22 ) require other coordinates to be expressed in a simple way .
introduce the variables that appear in the theory of jacobi s elliptic functions where , @xmath23 .
hence the separatrix has @xmath24 and @xmath22 ; and the data _ above the separatrix _ correspond to @xmath21 ( or @xmath25 ) .
note that @xmath26 because @xmath27 ( as @xmath22 corresponds to @xmath28 and @xmath29 ) ; the following formulae become singular as @xmath30 , but the singularity is only apparent and it will disappear from all relevant formulae derived or used in the following .
other important quantities in the elliptic functions theory are , see the references ( * ? ? ?
* ( 8.198.1 ) ) , ( * ? ? ?
* ( 8.198.2),(8.146 ) ) , where @xmath31 denotes here what in @xcite would be @xmath32 ( ( * ? ? ?
* ( 8.146.1),(8.194.2 ) ) ) . in terms of the above conventions we have ,
directly from the definitions of @xmath33 ( jacobi s elliptic functions , ( * ? ? ?
* ( 8.14 ) ) ) , and from the equations of motion : ( t)=2(u , ik ) , u=,b(t)=i = k(u , ik ) ( ( * ? ? ?
* ( 8.143),(8.141 ) ) ) . so that the action @xmath34 is given as a function of time b(t)= = 2ig , ( ( * ? ? ?
* ( 8.153.9),(8.153.3 ) ) ) assuming that initial data are assigned with @xmath35 .
the @xmath36 dependence of @xmath37 is naturally expressed via the argument @xmath38 , if the second of eq .
( ) is used , since @xmath39 , see eq .
( ) . this explains the role that g_0(x ) g_02=2 eq .
( ) , will play in the following analysis an important role , and it admits a rather simple product expansion , ( * ? ? ? * ( 8.197.1),(8.197.4 ) ) , g_0(x)=g _ n=1^()^2 and its logarithmic derivative is @xmath40 so that @xmath41 is @xmath42 , where @xmath43 , @xcite , is a jacobi s theta function .
it is also convenient to remark that in a motion with energy @xmath44 it will be h(b(t),(t))u=2 g^2 i
from the theory of elliptic functions the evolution @xmath45 with any initial data _ above the separatrix _ ( _ i.e. _ with @xmath46 and @xmath47 corresponding to a given value of @xmath48 , with @xmath21 ) , can be expressed as @xmath49 and @xmath50 with @xmath51 and , taking into account ( * ? ? ?
* ( 8.146.11 ) ) , with @xmath52 .
definitions in eq .
( ) , ( ) yield g_0()=g2 1(h)=g(1 + 14 h^2+ )which is analytic in @xmath53 by ( * ? ? ?
* ( 8.113.1 ) ) near @xmath54 .
eq . ( ) implies that @xmath55 is analytic in @xmath56 near @xmath57 so that @xmath58 . therefore g_0=(1 + 4 + 12 ^ 2+ )gis
analytic in @xmath59 near @xmath60 .
the evolution of @xmath61 is then a consequence of eq .
( ) which leads to an expression for @xmath62 by the remark that @xmath63 ( just expressing that @xmath34 is @xmath0 times the derivative of @xmath4 ) : namely s(,)=-4_n=1^ and , after developing in powers of @xmath59 the denominators and resumming , the first formula reminds of one found by jacobi which he commented by saying that `` _ inter formulas elegantissimas censeri debet _ '' , @xcite ( _ i.e. _ _ it should be counted among the most elegant formulae _ . note that @xmath64 depends only on @xmath59 , see eq .
( ) , which would be surprising if the mechanical interpretation was not taken into account . the eq .
( ) exhibits the convergence of the map @xmath65 , since @xmath66 in the region above the separatrix : in the latter region eq .
( ) provides a convergent expansion of the solution .
motions with initial coordinate @xmath67 also admit a rather simple representation .
remark that all pendulum motions with @xmath68 ( hence different from the two equilibria ) pass at some time through a phase space point with @xmath35 .
if @xmath69 is their velocity at that moment we can find a quantity @xmath59 such that @xmath70 are given by eq .
( ) with @xmath71 therefore they can be represented , at least as long as @xmath72 by introducing the _ dimensionless _ variables @xmath73 and allowing @xmath74 to be arbitrary .
then the motions will be @xmath75 showing that the motion can be represented by the following two functions , the motions @xmath76 solve the equations of motion if @xmath77 ( _ i.e. _ @xmath78 ) are positive .
but the equations of motion are analytic , hence the formulae eq .
( ) together with @xmath76 , with @xmath79 , give solutions of the pendulum equations independently of the sign of @xmath77 , provided the series converge .
the convergence requires @xmath80 : which represents many data , in particular those in the vicinity of the separatrix .
the coordinates can be called `` hyperbolic '' being suitable to describe motions near the separatrix ( where @xmath81 ) .
we also see that time evolution preserves both volume elements @xmath82 and @xmath83 ; which means that the jacobian determinant @xmath84 must be a function constant over the trajectories , hence a function @xmath85 of @xmath86 .
note that @xmath85 has dimension of an action .
it is then possible to change coordinates setting @xmath87 and choose @xmath88 so that the jacobian determinant for @xmath89 is @xmath90 .
a brief calculation shows that this is achieved by fixing a^2(x)=1x_0^x d(y)dy , which is possible for @xmath91 small because , from eq .
( ) and ( ) , it is @xmath92
. therefore the variables , which will have the dimension of @xmath93 , hence of a square root of an action , p = p a(x),q = q a(x),have jacobian determinant @xmath94 with respect to @xmath3 and the map @xmath95 @xmath8 is area preserving , hence _ canonical_. the hamiltonian eq .
( ) becomes a function @xmath96 of @xmath97 and the derivative of the energy with respect to @xmath91 has to be @xmath98 ( because the @xmath99 are canonically conjugated to @xmath100 ) .
note that @xmath91 has the dimension of an action , while @xmath99 are , dimensionally , square roots of action .
this allows us to find @xmath85 : by imposing that the equations of motion for the @xmath8 canonical variables have to be the hamilton s equations with hamiltonian @xmath101 it follows that @xmath102 , _
i.e. _ @xmath103 or @xmath104 by the above expression for @xmath105 .
the just obtained relation , together with eq . ( ) , gives d(x)=g_0(x)^-1 u(x)which is an explicit expression for the jacobian @xmath106 ( note that the jacobian between @xmath3 and @xmath8 is identically @xmath94 by construction ) .
( ) is dimensionally correct because @xmath107 is dimensionless so that @xmath108 has the correct dimension ( _ i.e. _ energy ) .
the function @xmath108 is in eq .
( ) where @xmath109 , by eq .
( ) , is related to @xmath110 by eq .
( ) , so that ( * ? ? ?
* ( 8.197.3),(8.197.4 ) ) , u(x)=2 g^2 i1k^2=2 g^2 i=32
n=1^()^8 to complete the determination of the canonical hyperbolic coordinates it remains to find an expression for @xmath111 in terms of the elliptic functions to obtain the canonical variable and the hamiltonian in closed form ( rather than as power series as done so far ) .
it is remarkable is that the function @xmath112 defined above , hence such that @xmath113 , seems to be simply a^2(z)=8 i g_0(z),in a common holomorphy domain , for both sides , around @xmath114 .
this is suggested by the agreement of the first @xmath115 coefficients of the expansion of the two sides in powers of @xmath116 : however this is not a proof and the relation eq .
( ) holds because it can be seen to be equivalent to an identity on elliptic functions , as discussed in appendix a below .
\(1 ) the expansion of @xmath85 in powers of @xmath107 can be derived from eq .
( ) , ( ) , while that of @xmath117 is obtained from eq .
( ) and , again , eq . ( ) . + ( 2 ) it is perhaps natural to guess that the function @xmath118 should be closely related to @xmath98 ; this is a guide to its determination as it becomes , then , natural to look for it among the derivatives of @xmath64 with respect to @xmath107 . by dimensional analysis
all @xmath107-derivatives of @xmath119 have the same dimension as @xmath112 .
looking _ also _ at the derivatives of @xmath64 as candidates for @xmath112 is an idea due to one of us ( pg ) .
this follows a similar line of thought which led to a conjecture on the canonical integrability of the `` calogero lattice '' , @xcite , whose proof was discovered in two subsequent works @xcite and @xcite . the relation eq .
( ) is equivalent to a notable identity between elliptic functions , as discussed in appendix a below .
other peculiarities are , setting @xmath120 , \(1 ) the function @xmath121 , hence @xmath122 , have taylor coefficients in powers of @xmath107 which _ are _ all positive integers as it follows from the relations eq .
( ) and eq .
( ) , while @xmath123 seems to have alternating sign taylor coefficients : ( x)-x=2x^2 - 4x^3 + 20x^4 -132 x^5 + 1008 x^6+ where @xmath96 is obtained by power series inversion of @xmath124 and from @xmath125 together with eq .
\(2 ) the function @xmath108 , energy of the pendulum expressed as a function of @xmath107 , has also the form u(x)= 32 i g_0 ^ 2 x f(x)which , remarkably , has by eq .
( ) to depend only upon @xmath107 , and have the form @xmath126 for some @xmath127 .
this is not _ a priori _ evident , unless the mechanical interpretation is kept in mind , from the expressions found for @xmath128 , namely u_x(z)=_=0^,v_x(z)=_=0^ , \(3 ) existence of an analytic canonical map integrating , near the hyperbolic point , the system with energy eq .
( ) into one with hamiltonian @xmath129 is well known : it can be established without an explicit calculation by perturbation analysis , see ( * ? ? ?
* appendix a3 ) , for instance .
calling @xmath130 it is @xmath131 , see ( * ? ? ?
* ( 8.123.2 ) , and @xmath132 , see ( * ? ? ?
* ( 8.122 ) ; the latter `` legendre s relation '' , @xcite , combined with @xmath133 yields the identity h h^2((h)- ( h))= 2.this can be used to obtain an expression for @xmath134 : keeping in mind @xmath135 it is @xmath136 which is transformed into @xmath137 .
( ) it follows , therefore , x=2 x=-2.and the corresponding derivatives with respect to @xmath48 are obtained by multiplying both sides by @xmath138 . to establish eq .
( ) consider the relation , \{h h^2}-h ( h)=0see ( * ? ? ? * ( 8.124.1 ) .
this implies by simple algebra , and keeping in mind that @xmath139 , the following identity h(h)= h^3 ( h h^2(-1h ( h ) + ( h)))which is a known linear equation , solved by @xmath140 .
this can be rewritten , since @xmath139 , as remarking that @xmath141 , eq .
( ) implies , multiplying both sides by @xmath142 , and by the first of eq .
( ) multiplied by @xmath133 this is , using @xmath143 and multiplying by @xmath144 it follows 2ig^2 1k^2= 8i 2 1h(h)(x 1h(h ) ) and setting @xmath145 the last relation is + @xmath146 so that eq .
( ) and eq .
( ) imply eq .
from the above results it is straightforward to find the canonical transformation that converts the pendulum hamiltonian in its normal form around the stable equilibrium point .
the hamiltonian is now given by eq .
( ) with the substitution : @xmath147 .
it is natural to define @xmath148 in order to use the same set of equations from the unstable case .
the system energy is then @xmath149 and large values of @xmath150 correspond now to small oscillations around the equilibrium point . with these conventions , and going through computations similar to the ones performed to study the unstable point ,
the relations found for the latter can be converted into the corresponding ones for the equilibrium point .
in particular , by choosing @xmath151 and @xmath152 the transformation given by eq .
( ) is now : the transformation @xmath154 is not canonical .
the canonical variables , @xmath8 , can be found by looking for a function @xmath155 ( which depends on the constant of motion @xmath156 ) such that @xmath157 and the jacobian of the transformation is one .
it is , as in the hyperbolic case , | revisiting canonical integration of the classical pendulum around its unstable equilibrium , normal hyperbolic canonical coordinates are constructed . |
20 filippone b w and ji x d 2001 _ adv .
phys . _ * 26 * 1 gehrmann t and stirling w j 1996 _ phys .
_ d * 53 * 6100 ackermann k h _ et al . _
( star collaboration ) 2003 _ nucl .
meth . _ a * 499 * 624 adams j _ et al . _
( star collaboration ) 2004 _ phys .
lett . _ * 92 * 171801 adams d l _ et al .
_ 1991 _ phys .
_ b * 264 * 462 nogach l ( star collaboration ) 2006 _ preprint _ hep - ex/0612030 anselmino m , dalesio u , melis s and murgia f 2006 _ phys .
_ d * 74 * 094011 ; dalesio u and murgia f _ private communications _ kouvaris c , qiu j , vogelsang w and yuan f 2006 _ preprint _ hep - ph/0609238 balewski j ( star collaboration ) 2006 _ preprint _ hep - ex/0612036 boer d and vogelsang w 2004 _ phys .
_ d * 69 * 094025 vogelsang w and yuan f 2005 _ phys .
_ d * 72 * 054028 dienfenthaler m ( hermes collaboration ) 2006 _ preprint _ hep - ex/0507013 jager b _ et al .
_ 2004 _ phys .
_ d * 70 * 034010 abelev b i _ et al . _ ( star collaboration ) 2006 _ phys .
lett . _ * 97 * 252001 .
simon f ( star collaboration ) 2006 _ preprint _ hep - ex/0612004 jager b , schafer a , stratmann m and vogelsang w 2003 _ phys . rev .
_ d * 67 * 054005 ; pumplin j _ et al .
_ 2002 _ j. high energy phys . _
* 0207 * 012 kocoloski a ( star collaboration ) 2006 _ preprint _ hep - ex/0612005 xu q h ( star collaboration ) 2006 _ preprint _ hep - ex/0612035 | the star experiment uses polarized p+p collisions at rhic to determine the contributions to the spin of the proton from gluon spin and from orbital angular momentum of the quarks and gluons .
selective star measurements of the longitudinal double spin asymmetry for inclusive jet and inclusive hadron production are presented here . in addition
, we report measurements of the transverse spin asymmetry for di - jet production at mid - rapidity and the transverse single - spin asymmetry for forward @xmath0 production .
the quark and anti - quark contributions to the spin of the proton measured with polarized deep - inelastic scattering ( dis ) fixed target experiments amount to only @xmath1 @xcite .
consequently , the rest of the spin of the proton must come from the gluons and the orbital angular momentum of both quarks and gluons .
however , the gluon polarization , @xmath2 , is poorly constrained from scaling violations in the dis data @xcite , and more precise and direct measurements are needed .
polarized p+p collisions at the relativistic heavy ion collider ( rhic ) provide a very suitable environment , rich with strongly interacting probes , to measure @xmath2 directly and precisely .
there are many processes where the gluon participates directly .
in addition , the high c.m . energy , @xmath3 = 200 gev , and high @xmath4 make next - to - leading order ( nlo ) perturbative quantum chromodynamics ( pqcd ) analysis more reliable .
the spin program of the solenoid tracker at rhic ( star ) experiment @xcite aims , in the short term , to utilize these advantages to measure @xmath2 using inclusive jet and inclusive hadron production measurements in longitudinally polarized p+p collisions .
in addition , the star experiment uses transversely polarized p+p collisions to gain insights into the orbital angular momentum of the partons .
the star collaboration previously observed a sizable analyzing power , @xmath5 , for forward @xmath0 production at large @xmath6 @xcite , similar to previous observations at lower energies @xcite .
the previous star results were in qualitative agreement with several different model predictions @xcite and could not differentiate between them , so higher precision measurements of @xmath5 as a function of both @xmath6 and @xmath4 were needed .
new @xmath5 measurements with higher luminosity and polarization were performed in 2006 @xcite .
figure [ fig : anvxfnpt ] shows preliminary results for @xmath5 as a function of @xmath6 ( left panel ) and as a function of @xmath4 for different @xmath6 bins ( right panel ) , along with model predictions .
the results in the left panel are consistent with the previous results @xcite in having large @xmath5 for high @xmath6 , and provide sufficient accuracy to discriminate between different dynamics .
the right panel shows that the @xmath4 dependence of @xmath5 is not well described by sivers based calculations @xcite , which predict that @xmath5 should decrease as you go higher in @xmath4 for all @xmath6 bins .
the star experiment also probed the sivers effect by measuring the di - jet opening angle in transversely polarized p+p collisions @xcite since boer and vogelsang @xcite suggested that @xmath5 for the distribution in relative azimuthal angle , @xmath7 , between the di - jet pairs could provide r95 mm access to the sivers effect @xcite .
figure [ fig : dijetsivers ] shows preliminary results from 2006 data for @xmath5 , for both beams , as a function of @xmath8 , where @xmath9 and @xmath10 are the psuedorapidities for the two jets .
the measurement was done over @xmath11 $ ] and for @xmath12 .
@xmath5 is extracted from the cross ratio of the spin - sorted @xmath7 distribution in which `` left '' and `` right '' yields were deduced from @xmath13 and @xmath14 , respectively .
@xmath15 , shown in figure [ fig : dijetsivers ] , are consistent with zero - much smaller than the sivers asymetry that was observed in semi - inclusive deep inelastic scattering ( sidis ) by hermes @xcite .
+ the star collaboration uses the measurement of the longitudinal double spin asymmetry , @xmath16 , for inclusive jet and inclusive hadron production to constrain @xmath2 .
@xmath16 is directly sensitive to the polarized gluon distribution function in the proton through gluon - gluon and gluon - quark sub - processes @xcite .
we reported the unpolarized inclusive jet cross section from data taken in 2003 and 2004 @xcite , and preliminary results for the unpolarized inclusive @xmath17 cross section @xcite , and both show satisfactory agreement with nlo pqcd calculations over many orders of magnitude .
the agreement suggests the applicability of pqcd to describe the polarization observables in this kinematic region . left panel : @xmath16 for inclusive @xmath0 production with statistical errors and the systematic band .
right panel : mean momentum fraction of the @xmath0 in their associated jet as a function of @xmath4 and the corresponding mean jet @xmath4 .
the data points are not corrected for jet acceptance or reconstruction efficiency . ]
left panel : @xmath16 for inclusive @xmath0 production with statistical errors and the systematic band .
right panel : mean momentum fraction of the @xmath0 in their associated jet as a function of @xmath4 and the corresponding mean jet @xmath4 .
the data points are not corrected for jet acceptance or reconstruction efficiency . ] in figure [ fig : allpi0nz ] , the left panel , we present preliminary results of the first star @xmath16 measurement of inclusive @xmath0 production at mid - rapidity with statistical errors and the systematics band @xcite .
the systematics band does not include the normalization uncertainty due to the polarization measurement .
in addition to the grsv standard curve , which is based on the best fit to dis data , different @xmath2 scenarios from maximally positive to maximally negative and passing zero @xmath2 are also shown in this and the following figures @xcite .
this measurement is consistent with previously reported @xmath16 for inclusive jets @xcite in disfavoring the maximum positive @xmath2 .
the right panel shows the mean momentum fraction carried by the @xmath0 in its associated jet .
this plot shows that the @xmath0 carries a very large fraction , 70% , of its associated jet momentum .
corrections for jet acceptance and reconstruction efficiency may reduce this fraction by @xmath18 .
preliminary results from 2005 of very promising @xmath16 measurements of inclusive @xmath19 and @xmath20 are shown in figure [ fig : allplusmin ] @xcite .
the error bars are statistical errors , and the associated bands indicate the systematics , which do nt include the scale uncertainty due to the polarization measurement .
several @xmath2 scenarios for these processes are also shown .
the ordering of the measurements of @xmath16 between @xmath19 and @xmath20 is sensitive to the sign of @xmath2 . for now , this measurement is limited by statistics .
we are currently working on 2006 data where we have higher luminosity and polarization .
fig .
[ fig : alljets ] shows @xmath16 for inclusive jet production versus jet @xmath4 .
the solid points are preliminary results from 2005 data and the open circles are from 2003 + 2004 data analysis @xcite , both with statistical error bars .
the dark and hatched bands show the systematic uncertainties for 2005 and combined 2003 and 2004 data , respectively .
these bands do not include scale uncertainties due to the beam polarization .
fig .
[ fig : alljets ] also shows several grsv nlo pqcd scenarios for inclusive jets .
the data are consistent with zero or small gluon polarization and exclude the maximally positive gluon polarization scenario .
in fact , these 2005 results go further in excluding much of the region in between maximally positive gluon polarization and dis based fit scenarios , and thus put a tighter constraint on @xmath2 . in addition
, the star collaboration studied longitudinal spin transfer of @xmath21 in polarized p+p collisions@xcite and inclusive @xmath0 production using the star endcap calorimeter as proof of principle measurements in 2005 .
in summary , we reported on preliminary di - jet sivers @xmath5 measurement from 2006 , and all measured @xmath5 are statistically consistent with zero , several factors smaller than expected from sidis . in addition , we reported on 2006 @xmath5 measurements of forward @xmath0 with higher statistics and polarization , which show inconsistency with twist-3 / sivers calculations ; @xmath5 does not decrease with @xmath4 in all @xmath6 bins . in the longitudinal spin program , we reported on several preliminary @xmath16 measurements of inclusive hadrons and jets at mid - rapidity from 2005 data .
the results are consistent in the region of kinematic overlap and provide improved constraints on the gluon spin contribution to the proton spin .
they exclude large and positive gluon spin contributions . |
solar flares are generally believed to be a result of magnetic reconnection in the corona .
after reconnection , the connectivity of the magnetic field changes and the field relaxes to a lower energy state .
the energy released during reconnection is transported along reconnection - formed flare loops to the lower atmosphere , by conductive flux or non - thermal particles , giving rise to enhanced emissions in optical and ultraviolet ( uv ) wavelengths .
as the chromosphere is impulsively heated , a pressure front is formed that pushes plasmas both upward ( chromospheric evaporation ) and downward ( chromospheric condensation ) .
the upflow fills post - flare loops , where energy is radiated away in soft x - rays ( sxrs ) via thermal bremsstrahlung at temperatures of up to a few tens of million kelvin ( mk ) .
when these loops cool down to a few mk , they begin to be visible in extreme ultraviolet ( euv ) .
( see the comprehensive review by @xcite . ) in this qualitative picture , several fundamental questions remain unanswered in terms of the quantity of energy .
first , we do not know how much energy is released by reconnection , and how much energy in total is radiated by flare plasmas ( and carried away by coronal mass ejections , if any ) ; in principle , the former should equal the latter .
second , we do not fully understand how , and by how much , flare plasmas are heated and particles are accelerated by reconnection - released energy , generating observed radiation signatures across the full spectral range . a variety of hydrodynamic simulations have been performed to understand energy transport in flare loops and the atmospheric response to energy deposition at the feet of these loops in the transition region and chromosphere ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
since plasmas are frozen in magnetic loops , most of these simulations use one - dimensional ( 1d ) models .
some of these hydrodynamic simulations concentrate on the lower atmosphere by using full non - lte formulation of radiation transfer ( i.e. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , while others focus on the upper atmosphere dominated by optically thin emissions and treat the lower atmosphere as boundary conditions ( i.e. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the radiative hydrodynamic simulations have succeeded in reproducing optically thick emissions in solar flares ; however , the corona could only reach a few mk , which is much less than observed temperatures of tens of mk . on the other hand , the hydrodynamic models with optically thin emissions ,
though unable to address emissions from the chromosphere , work better at reproducing high coronal temperatures and densities , and enhancements of uv emissions from the transition region . with high resolution euv images obtained by the _ transition region and coronal explorer _
( _ trace _ ; * ? ? ? * ) , it is found that the post - flare arcade is composed of at least a few hundred flaring loops formed successively ( e.g. , * ? ? ?
it is further confirmed by @xcite that the flare uv ribbons are made of small kernels outlining the feet of flare loops . to simulate sequentially formed loops in solar flares , @xcite developed a
pseudo - two - dimensional model with multiple 1d loops heated progressively from the innermost loop to the outermost .
the multi-1d model was further used to reproduce x - ray and euv observations of flares ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . the hydrodynamic evolution of plasma inside the flare loop appears to be governed by how the loop is heated
. specifically , we need to know when , for how long , by how much , and by what physical mechanism a flare loop is heated . among existing models ,
many start with an unspecified _ ad hoc _ heating source in the corona , and the energy is transported through conductive flux ( e.g. , * ? ? ?
* ; * ? ? ?
? * ; * ? ? ?
some other models specify the source of the energy and/or mechanism of heating .
for example , @xcite , @xcite , @xcite , @xcite , and @xcite assumed that energy is carried by non - thermal particles and the loop is heated by coulomb collisions of electrons with ambient plasmas , either in the corona ( thin target ) or at the lower atmosphere ( thick target ) , using an analytic expression of electron energy derived by @xcite or @xcite .
@xcite and @xcite used poynting flux derived from numerical solutions of the loss - of - equilibrium flare model @xcite as the energy input in the loop heating model , and could relate the resultant plasma radiation time profile with the kinematic evolution of the flux rope ejection .
effort has also been made to constrain the energy input with observations . @xcite and @xcite derived the heating rate empirically by matching the loop heating model results with the sxr fluxes observed by the x - ray sensor ( xrs ) on the _ geostationary operational environmental satellite _ ( _ goes _ ) . in their model ,
the heating events are assumed to have a triangular time profile with a fixed duration ( 60 s in @xcite , and 400 s in @xcite ) and their magnitudes and distributions during the flare are adjusted to match the observed sxr flux .
@xcite modeled and analyzed a flare observed by _ trace _ , using as the heating rates the reconnection - released energy , which is calculated from a patchy reconnection model .
the model takes the observed photospheric magnetic field as the boundary , and uses the observationally measured reconnection rate and time distribution of reconnection - formed flare loops counted in euv images obtained by _
trace_. the modeled coronal radiation compares favorably with sxr observations from _ goes _ xrs and the _ reuven ramaty high energy solar spectroscopic imager _
( _ rhessi _ ; * ? ? ? * ) . along this avenue , recently , @xcite have proposed that the heating rates in individual flare loops could be directly inferred from the time profiles of uv emission at the feet of flare loops . the spatially resolved uv light curves
exhibit an impulsive rise , which indicates , and therefore constrains , when , for how long , and by how much a flare loop is heated .
accordingly , @xcite constructed from these light curves the heating rates of 1600 loops , _ each _ with cross - sectional area of 1 by 1 , anchored at impulsively brightened uv pixels in a c3.2 flare observed by the atmospheric imaging assembly ( aia ; * ? ? ? * ) onboard the _ solar dynamic observatory _ ( _ sdo _ ; * ? ? ? * ) . applying these heating rates to a zero - dimensional loop heating model , the enthalpy - based thermal evolution of loops ( ebtel ; *
* ; * ? ? ?
* ; * ? ? ?
* ) model , they computed plasma evolution in these 1600 flare loops and the synthetic coronal radiation in sxr and euv passbands , which are compared with observations by _ goes _ and aia to verify the empirically determined heating rates .
this method uses all available observations to constrain the heating rates from the input ( the impulsive uv emission from the foot points ) to the output ( the x - ray and euv emission from the coronal loops ) .
the flare studied in their paper is primarily a thermal flare with little hard x - ray ( hxr ) emission , and _ ad hoc _ volumetric heating rates are used in the model . in this paper
, we will improve the method of @xcite and apply the analysis to an m8.0 flare on 2005 may 13 observed by _ trace _ , _ goes _ , and _ rhessi_. the flare exhibits significant thick - target non - thermal hxr emission observed by _ rhessi _
, suggestive of strong direct heating of the lower atmosphere that would result in chromospheric evaporation that sends energy back to the corona .
we also include in the loop heating model this energy flux of non - thermal origin , using observed uv and thick - target hxr emissions as constraints .
as the output of the model , we calculate the time - dependent 320 kev sxr spectrum and compare this with the observations by _
rhessi_. finally , whereas the impulsive rise of the uv foot - point emission is considered to directly relate to the energy release process , the observed gradual decay of the uv emission from the same foot - point is governed by evolution of the overlying coronal plasma in the loop , which therefore provides diagnostics of the loop cooling . in this paper
, we use a simplified model to compute the uv emission during the decay of flare loops to further compare with observations .
these new steps help improve determination of the heating rates , and enhance our understanding of flare energy release , partition , and plasma heating and evolution . in the following text , we present in section [ sec2 ] observations of the flare , especially the uv and hxr observations , which will be used to construct the heating functions , and
apply them to model coronal plasma evolution in section [ sec3 ] . in section [ sec4 ] , we compute time - dependent sxr and uv fluxes from the model output and compare them with observations .
conclusions and discussions are given in section [ sec5 ] .
in this paper , we study a _
goes_-class m8.0 flare that occurred on 2005 may 13 in noaa active region 10759 located at n12e05 at the time of the flare .
figure [ fig1 ] summarizes observations of the flare in a few wavelengths .
high cadence ( 3 s ) uv images obtained by _ trace _ , given in the bottom panel , reveal that the m8.0 flare is a typical two ribbon flare .
the flare uv ribbons expand away from the magnetic polarity inversion line for half an hour , indicating that magnetic reconnection continues to form new loops and release energy in them , as depicted by the standard flare model ( the cshkp model ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
significant x - ray emissions are observed in this flare by _
rhessi_. previous morphology studies of this same flare show that , during the impulsive phase , the sources of non - thermal hxr emissions of @xmath325 kev are located within the uv flare ribbons ( see fig . 2 and fig . 3 of *
* for images of hxr kernels in 2550 kev and 50100 kev energy bands overlaid on the uv contours ) .
these observations confirm that uv and hxr emissions are both produced by heating of the lower atmosphere ( the transition region and chromosphere ) by energetic particles precipitating at the feet of flare loops during the impulsive phase ( see , e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the top panel in figure [ fig1 ] shows normalized and background - subtracted x - ray and uv light curves observed by _ goes _ , _ rhessi _ , and _ trace _ , respectively .
the uv light curve is the total count rate ( in units of dn s@xmath4 ) derived from the semi - calibrated uv images ( see * ? ? ?
* section 3.1 for calibration techniques ) , with pre - flare counts subtracted .
it follows closely the @xmath325 kev hxr light curve during the rise of the flare and the impulsive phase , but continues to rise after the hxr peak at 16:42 ut , and reaches maximum five minutes later .
also plotted is the time derivative of the _ goes _
18 sxr light curve , and it is seen that the uv light curve peaks at the same time ( 16:47 ut ) as the time derivative of the sxr flux . the temporal correlation between hxr and sxr time derivative has been known as the neupert effect @xcite , with the basic idea that non - thermal electrons precipitate at the chromosphere , losing their energy by coulomb collision to give rise to hxr emission , and at the same time , driving chromospheric evaporation to fill post - flare loops that are subsequently observed in sxrs .
this flare , however , exhibits a similar neupert effect between the sxr derivative and uv , instead of hxr , light curves .
this is evidence of continuous energy deposition in the lower atmosphere , most likely by thermal conduction , after the impulsive phase when thick - target hxr emission is no longer significant .
the time sequence of uv images further shows that the continuous rise of the uv emission after the hxrs is produced by newly brightened uv ribbons ( bottom panel in figure [ fig1 ] ) .
the spread of the flare uv ribbons across the longitudinal magnetic field ( lower right panel ) provides a measurement of magnetic reconnection flux @xcite , plotted in the top panel .
the reconnection flux starts to grow at 16:30 ut at the beginning of both uv and x - ray light curves .
the continuous increase of reconnection flux and dramatic decay of non - thermal emission after 16:50 ut ( the post - impulsive phase ) confirm that reconnection continues to form new loops and release energy in them , and the thermal process is dominant in this late phase of the flare .
the post - impulsive reconnection flux amounts to 2@xmath010@xmath5 mx , about one third of the total reconnection flux measured for this flare .
comparison of the uv and x - ray observations of the flare suggests that imaging uv observations provide information of energy release , which is then transported to the lower atmosphere by either non - thermal electrons or thermal conduction , in newly formed flare loops anchored at newly brightened uv ribbons .
based on this idea , we have analyzed spatially resolved uv light curves , using them to construct heating rates of flare loops in a c3.2 flare @xcite . the analysis is applied to this flare , as well , in the following sections .
examples of the uv light curve in a flaring pixel ( 1 by 1 ) , shown in the top panel of figure [ fig2 ] , typically exhibits a rapid rise , peaks within several minutes , and then decays slowly with a characteristic cooling time of more than 10 minutes .
such characteristics have been found in uv observations of other flares as well @xcite . since the lower atmosphere ( the transition region and chromosphere ) responds to energy deposition on very short timescales , the rise time of uv emission reflects the timescale of energy release in the flare loop anchored at this pixel .
the gradual decay , on the other hand , is coupled with subsequent cooling processes in the overlying corona .
to characterize the rapid rise of the spatially resolved uv light curves , we fit the rise of the uv count rate light curve ( in units of dn s@xmath4 ) to a half gaussian , @xmath6 , ( t \leqslant t_0 ) \label{eq_halfgauss}\ ] ] where @xmath7 is the background - subtracted peak count rate , @xmath8 is the peak time , and @xmath9 is the characteristic rise time . for this flare ,
5127 flaring pixels ( each of size 1 by 1 ) are identified from uv images with 3 s cadence .
the lightcurve of each pixel is smoothed to 10 s and its rise phase is fitted to a half gaussian .
examples of the fit to the observed light curve are plotted on the top panel of figure [ fig2 ] .
histograms of the rise times from fits to all the flaring pixels are given in the bottom panel of figure [ fig2 ] , showing that most flaring pixels rise over timescales from a few tens of seconds to a few minutes .
these rise times are systematically shorter than those derived by @xcite in another c3.2 flare
. this may be due to the higher cadence ( 3 s ) of the observations in this flare compared with the 30 s cadence of the other data , or more likely , the more impulsive nature of this flare that has significant non - thermal emission .
observations also show that uv pixels peaking after 16:52 ut , when the hxr emission has been reduced significantly , tend to evolve more slowly than those peaking earlier .
this suggests different evolutionary timescales in both the heating and cooling phases when energy is transported through flare loops by different mechanisms , by non - thermal electrons or by thermal conduction . to examine the distinction , we compare uv light curves of pixels peaking in three stages relative to the hxr evolution .
pixels in the first stage ( uv1 ) peak before 16:45 ut when hxr emission is significant .
these pixels are marked in dark black to light blue colors in figure [ fig1 ] , and the sum of their light curves ( purple curve in figure [ fig2]d ) has a very good temporal correlation with the hxr light curve .
uv pixels categorized in the second stage ( uv2 ) peak during the hxr decay from 16:45 ut to 16:52 ut .
the sum of their light curves ( orange in figure [ fig2]d ) contributes mostly to the peak of the total uv light curve .
uv pixels in the third stage ( uv3 ; red in figure [ fig2]d ) peak after 16:52 ut , when @xmath325 kev hxr emission has nearly ended . compared with uv1 and uv2
, the uv3 light curves evidently rise more slowly and also decay more gradually . in figure
[ fig2]e , we plot the histograms of the rise times for pixels at the three stages , which shows that uv3 pixels rise more slowly on average .
the differences of the uv light curves in the three stages indicates the different effects of beam heating and conductive heating on uv emissions , and in turn , the uv evolution could give us a clue to the nature of energy release , thermal or non - thermal .
hydrodynamic and radiative transfer models of the lower atmosphere during flares have shown that the lower atmosphere responds within a few seconds to the onset of impulsive energy injection @xcite .
therefore , the rapid rise of uv emission from the upper atmosphere or transition region may be considered to scale with the impulsive energy release in reconnection - formed flare loops . with this idea ,
we have implemented a method to construct heating functions and used them to compute plasma evolution in flare loops for a c3.2 flare , which is primarily a thermal flare @xcite . in this study , we apply the same method , using the rising uv emission at the foot - point as an indicator of energy injection into the newly formed flare loop ( or flux tube ) rooted at the foot point . we take the start - time of the uv brightening as the onset of the reconnection event forming the new flux tube .
the rise time of the uv brightness gives the duration of the impulsive energy release in the newly formed tube , and the maximum brightness of the pixel reflects the magnitude of the energy release ( or heating rate ) in the flux tube ; this simply assumes that a brighter pixel is more strongly heated .
these observationally measured quantities may then be used to construct heating rates and study subsequent plasma evolution inside flare loops that are formed and heated sequentially during the flare . in constructing the heating functions , significant progress in the present study is inclusion of the heating term by non - thermal particles that are evident during the impulsive phase .
we have identified over 5000 brightened pixels of size 1 by 1 in _ trace _ 1600 images , and assume that a half loop ( or flux tube ) of constant cross - section ( 1 by 1 ) is anchored at each of the pixels .
these are half loops , because we do not identify connectivity between positive and negative foot - points .
for each half loop , we compute the time - evolution of plasma density , temperature , and pressure averaged along the loop using the ebtel model .
the energy input term in the model , the heating rate , is constructed from observed uv count rates at the foot point . the time - dependent differential emission measure ( dem )
is then derived from these 5000 half loops , which will be used to compute sxr flux and compare with observations by _ rhessi _ and _ goes _ in section [ sec4 ] .
the 0d ebtel model calculates mean properties of loop plasmas , which have been shown to reasonably agree with mean values from simulations using the 1d hydrodynamic code called adaptively refined godunov solver .
the 0d model is highly efficient at computing plasma evolution for over 5000 half loops in our study of the m8.0 flare .
the ebtel model solves two equations .
the energy ( or pressure ) equation takes into account the prescribed heating rates ( @xmath10 ) as the energy input term and coronal radiative loss ( @xmath11 ) as well as the total loss ( @xmath12 ) through the base of the loop ( so called transition region ) as energy loss terms .
the mass equation is governed by mass flow between the transition region and the corona . in the ebtel model , this flow is a result of the difference between the energy input , including the beam heating and conductive flux ( @xmath13 ) from the corona , and the total loss ( @xmath12 ) at the base ( transition region ) . during the heating phase ,
the energy input to the transition region dominates the loss term , driving upflows known as chromospheric evaporation . during the decay
, the coronal plasma in the loop is cooling through thermal conduction and radiation ; meanwhile , the loss through the transition region exceeds the conductive flux into it , which drives downflows , called coronal condensation .
the mean coronal electron density ( @xmath14 ) and pressure ( @xmath15 ) in each loop evolve according to the ebtel equations , @xmath16\label{ebtel_eq2}%\\ % t & = & \frac{p}{2k_{\rm b}n}\label{ebtel_eq3}\end{aligned}\ ] ] the mean temperature @xmath17 is determined by the ideal gas law , @xmath18 ( including both electrons and ions ) . in the equations
, @xmath19 is boltzmann s constant , @xmath20 is the ratio of the loss through the transition region to the coronal radiation ( @xmath21 ) , @xmath22 is the ratio of the average coronal temperature to the apex temperature , and @xmath23 is the ratio of the coronal base temperature to the apex temperature . with the symmetry assumption ,
ebtel only models heating of a half loop , with @xmath24 being the length of the half - loop .
we prescribe two heating terms in the equation : the _ ad hoc _ volumetric heating rate , @xmath25 , and the energy flux carried by beam - driven upflows , @xmath26 .
the coronal radiative loss is given by @xmath27 , @xmath28 being the empirically - determined radiative loss function for optically - thin plasmas ( see equation ( 3 ) of * ? ? ?
* for details ) .
@xmath13 is the conductive flux at the base of the corona , which is defined by @xcite as the location where thermal conduction changes from being an energy loss above to an energy source below .
the classical form of conductive flux is used @xcite , @xmath29 where @xmath30 is the thermal conductivity coefficient , taken to be @xmath31 in cgs units . in the ebtel model , @xmath13 is saturated for large temperature gradients ; in the latest version of ebtel @xcite , the gravity is included in calculating of @xmath20 for semi - circular loops , while the dependence of @xmath22 and @xmath23 on gravity is negligible .
there are three parameters in the ebtel models , @xmath20 , @xmath22 , and @xmath23 . in the latest version of ebtel @xcite , which is used in the this study ,
@xmath20 is self - consistently determined by plasma properties inside the loop .
its value varies about the mean value 2.1 during flux tube evolution , which is not very sensitive to different heating rates in different flux tubes .
fixed values of @xmath32 and @xmath33 are used in the study .
these are mean values determined from 1d simulations . in the simulations , @xmath22 and @xmath23 usually change during the loop evolution , but only within a small range .
for simplicity , in this study , we use fixed mean values of @xmath22 and @xmath23 for 5000 loops throughout their evolution , considering that differences produced by using varying values of these parameters will become insignificant when we sum up contributions from 5000 loops . the critical input to the ebtel model , also a focus of our present study , is the heating rate . in general
, it includes two parts .
the first contribution , denoted by an _
ad hoc _ volumetric heating rate @xmath34 ( in units of erg @xmath35 s@xmath4 ) , is by _ in situ _ heating in the corona .
it may result from current dissipation , shocks ( e.g. , * ? ? ? * ; * ? ? ? * ) , electrons trapped and scattering in the corona(e.g .
, * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , or even return current ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . in the present study ,
the exact mechanisms for this terms are not discussed .
the second contribution , denoted as @xmath36 ( in units of erg @xmath37 s@xmath4 ) , is coronal heating _ from the lower atmosphere _ due to evaporation driven by non - thermal electrons that precipitate at the lower atmosphere during the impulsive phase . of the total flux carried by a non - thermal beam , @xmath36 is a fraction of that sent back upward into the corona .
these two terms are distinguished as they heat the corona in different ways , @xmath34 by primarily raising the temperature of the coronal plasma , and @xmath36 by primarily raising the density - therefore the @xmath36 term also enters the density equation .
predictably , the different heating styles result in different evolutionary patterns .
furthermore , these two terms play different roles in different stages of the flare .
specifically , the @xmath36 term as driven by beams is included only during the impulsive phase when thick - target hxr emission is evident .
we must note that thermal conduction alone is able to produce chromospheric evaporation , which , in the ebtel model , is the consequence of coronal evolution when the conductive flux exceeds the loss through the lower atmosphere during the heating phase , and is therefore not treated as an additional coronal heating term .
the @xmath36 term , on the other hand , is considered in this study to be produced by particles that instantaneously heat the lower atmosphere , and drive upflow _ independent of _ the coronal situation .
therefore , this term contributes to both energy and density of coronal plasmas . for the 2005 may 13 flare , _ rhessi _ observations
show that there is significant hxr emission with energy up to 300 kev during the first ten minutes of the flare .
so both _ in situ _ heating in the corona and heating by beam driven upflow are considered in this paper .
until now it has proven difficult in general to distinguish these two contributions , either theoretically or observationally . in this study , we use an empirical method to introduce these two terms as scaled with the uv and hxr light curves
. determination of these two heating terms and the effect of the partition between the two will be discussed in the following text .
following @xcite , we assume that the heating rate in a flare loop ( or a flux tube ) is proportional to the rise of the uv light curve at its foot - point , given as : @xmath38 } { \text{~erg~s}^{-1}\text{~cm}^{-2}};\ ( 0 < t < \infty)\label{eq_heating}\ ] ] @xmath39 is the total heating flux in the loop anchored at @xmath40th pixel , which is composed of the _ ad hoc _ heating rate ( @xmath41 ) and the beam - driven flux ( @xmath42 ) .
although we only fit the rise of the uv light curve to a half gaussian , we consider the heating function to be symmetric , or a full gaussian . assuming semi - circular post - flare loops
, we estimate the length of the @xmath40th half loop by @xmath43 , where @xmath44 is the distance of the foot - point to the polarity inversion line . as observed , flare ribbons expand away from the polarity inversion line in a rather organized manner ; therefore
, @xmath45 approximately grows linearly with the time of flare brightening ( @xmath46 ) .
for the 5000 half loops , @xmath45 ranges from 35 to 55 mm .
although loops are in general not semi - circular , our experiments have shown that variations of the loop lengths within a factor of 1.5 do not significantly change the synthetic total emission . to relate the total heating rate to uv emission ,
we employ a scalar @xmath47 that converts the count rate ( dn s@xmath4 ) to a heating rate ( erg s@xmath4 ) .
the value of this parameter depends critically on the lower atmospheric response to beam heating or conductive heating and on mechanisms of uv emission . as a rule of thumb
, we consider that @xmath47 takes a larger value when conduction heating dominates than when beam heating dominates , for the simple reason that , with the same amount of energy , beam heating occurring in the lower atmosphere would generate stronger uv emission than conductive flux ( e.g. , * ? ? ?
in this study , @xmath47 takes the value of 1.9@xmath48 ergs dn@xmath4 for loops whose foot - point uv light curves peak before 16:48 ut , then linearly increases until 16:52 ut ; afterwards , when hxr emission has finished , @xmath47 stays constant at 2.5@xmath48 ergs dn@xmath4 .
note that in this study , we do not model the lower atmospheric heating and dynamics , but instead use this simple empirical model to minimize the number of free parameters . the @xmath47 values quoted here
are determined by best matching the model - computed time - dependent sxr emissions with those observed by
_ goes _ , as will be described in the next section .
these @xmath47 values , combined with the peak count rates , correspond to the peak heating flux ( @xmath10 ) of order @xmath49 to 10@xmath50 ergs s@xmath4 @xmath37 for the few thousand half loops in our model . in the heating term , the coronal _ in situ _
heating @xmath51 is present in all flare loops ; @xmath36 is present only for loops brightened during the impulsive phase ( uv1 and uv2 pixels ) , and is gradually switched off after 16:52 ut ( uv3 pixels ) . for one loop , the _ ad hoc _ heating rate and beam driven energy flux both have the same gaussian time profile of @xmath52}$ ] , and the partition of the beam driven energy flux , i.e. , @xmath53 , is constant . for different loops heated at different times represented by @xmath46 ,
the partition @xmath53 is different ; whereas the net heating flux @xmath39 in a loop is proportional to the uv count rate at the foot - point , the beam driven flux @xmath42 is assumed to be proportional to the @xmath325 kev hxr light curve .
this rather simplified treatment can be justified by the hydrodynamic simulations of chromospheric evaporation showing that the evaporation upflow roughly increases as the heating energy flux increases @xcite in the range of 10@xmath54 - 10@xmath50 erg s@xmath4 @xmath37 .
the estimated heating flux of the majority of flux tubes in the impulsive phase in our study is within this range . during the impulsive phase ,
the hxr spectral index does not vary significantly ( not shown in this paper ) , so we can approximate the total beam energy flux , and subsequently the energy flux in the beam driven upflow , as proportional to the hxr count rate light curve .
the partition @xmath53 , therefore , is time - dependent , and is empirically given by @xmath55 , where @xmath56 is a time - dependent function that tracks the hxr light curve , @xmath57 is the maximum partition of the non - thermal energy flux used in this study , and this maximum partition occurs when hxr emission peaks at about 16:42 ut . in addition to the impulsive flare heating , a constant background heating rate of order @xmath58 , which is a few thousandths of the maximum heating rate constructed from uv observations , is imposed on each loop to produce an initial equilibrium of an average coronal temperature of @xmath59 mk and density of @xmath60 @xmath35 .
@xcite have shown that neither the initial state of the loop nor the background heating will affect the plasma evolution as soon as impulsive heating occurs , since the flare heating rate is a few orders of magnitude larger than the background heating rate .
the so devised heating function in a single flux tube is mostly constrained by foot - point uv and hxr observations , but also depends on two free parameters @xmath47 and @xmath61 .
@xmath47 determines the total amount of energy used to heat the corona , and @xmath61 determines the non - thermal partition . since heating via upflows or via direct heating leads to different plasma evolution pattern , the resultant coronal radiation spectrum will differ .
our recent experiments with another flare have shown that model - predicted coronal radiation by high temperature plasmas is very sensitive to parameters defining the heating function but rather insensitive to other model parameters like @xmath20 @xcite .
therefore , these parameters will be eventually constrained by comparing model - predicted and observed x - ray emission . in this study , we start the model with an initial guess of the parameters , and gradually adjust them so that the synthetic sxr spectrum and light curves best match those observed by _
rhessi _ and _
goes_. with methods described above , we compute plasma evolution in each flux tube anchored at the uv flaring pixel using the ebtel model . figure [ fig3 ] shows , in the left panel , time profiles of the mean temperature ( solid ) and density ( dashed ) of a single flux tube rooted at its uv foot point with constant cross - sectional area 1 by 1 .
it is heated impulsively by a gaussian - profile heating rate ( red dot - dashed line in the right panel ) constructed from the uv light curve ( black solid line in the right panel ) with @xmath62 s , and @xmath63 , of which @xmath64% is carried by the beam - driven upflow .
the @xmath47 and @xmath61 values used here are determined from comparing synthetic sxr light curves and spectra with observations , which will be described in the following sections .
coronal temperature rises as the flux tube in the corona is heated by _ ad hoc _ heating . at the same time , the lower atmosphere is heated by non - thermal electron beams and conductive flux , giving rise to enhanced uv emission as observed by _ trace_. chromospheric evaporation is driven to fill the coronal loop .
an increase in coronal density leads to enhancement of coronal radiation ( red dashed line in the right panel ) , which then cools the plasma in the flux tube .
the coronal temperature begins to decline immediately after the peak of the heating , but the density continues to grow until the energy loss in the transition region exceeds the conduction flux , causing a downward flux , or coronal condensation .
the coronal density and radiation then decline , and the flux tube experiences a long decay .
notably , the observed uv 1600 emission decays on the same timescale of coronal evolution .
this has been previously known as optically - thin transition region lines , such as the resonant c iv line that dominates the 1600 broadband uv emission , behave as a coronal pressure gauge " @xcite . to determine @xmath61
, we investigate the effect of beam driven flux in the ebtel model .
we examine the evolution of coronal plasma in the same flux tube shown in figure [ fig3 ] with the same impulsive heating rate @xmath65 but with varying partition , i.e. , @xmath66 values .
the time profiles of coronal - averaged temperature , density , and pressure with different @xmath66 values are plotted in figure [ fig4 ] .
it is shown that , in general , a greater amount of beam driven flux results in lower temperature and higher density .
if the beam driven flux does not dominate , i.e. , @xmath67 , the average temperature decays to @xmath68 mk in a few minutes , and then continues to decay toward the pre - flare value .
from about ten minutes after the heating , the decay of coronal temperature and density is nearly identical for different @xmath66 values . the same evolution pattern of the apex density and temperature is also displayed in 1d hydrodynamic simulations by @xcite , who modeled loop heating with thermal conduction .
their study showed that a flare loop impulsively heated by the same amount of the total energy , but with varying magnitude and duration and at different locations along the loop , would reach the same equilibrium point when radiation and conduction are comparable , and the evolution of the apex density and temperature is identical thereafter . on the other hand ,
when the beam driven flux dominates the energy budget with @xmath69 , the flux tube will attain much higher density in a short time due to strong upflow .
meanwhile , less direct heating @xmath34 results in a lower peak temperature .
the two effects , high density and low temperature , lead to faster radiative cooling in the decay phase .
therefore , the flux tube evolves on much shorter timescales .
a similar evolution pattern has also been produced in the 1d numerical simulation that takes into account beam heating with different pitch angles @xcite .
the experiment above suggests that varying the partition parameter @xmath66 will lead to different temperature and density in the heating phase , and for very large @xmath66 values , evolution of the flux tube in the decay phase is also modified significantly .
therefore , the experiment provides us with two ways to estimate the optimal @xmath66 value in this flare .
first , we may compare the synthetic sxr spectrum using the dem from modeling over 5000 flux tubes with the x - ray spectrum observed by _ rhessi _ during the heating phase .
figure [ fig_dem_nth ] shows the dem constructed using different @xmath61 values for the 5000 tubes .
it is seen that , given the same amount of total energy , the larger value of @xmath61 , i.e. , the larger amount of energy carried by beam driven upflow , will lead to greater dem increase at lower temperatures , whereas the dem at the high temperature end would be reduced .
such variation will affect the synthetic sxr spectrum formed at the temperature range of a few to a few tens of mk .
therefore , the value of @xmath61 can be estimated by comparing the synthetic sxr spectra with _
rhessi _ observations , as will be presented in section [ sec4.2 ] .
second , we will compare the observed and modeled timescales of the foot - point uv emission during the decay phase of the flare , based on the principle that uv line emission behaves as a coronal pressure gauge in the decay phase @xcite .
these will be elaborated on in section [ sec4.3 ] . as a quick look
, the right panel of figure [ fig4 ] shows the observed uv light curve versus pressure ; it appears that the uv emission decays on the same timescale as the coronal pressure for @xmath67 .
our methods yield the best estimate of @xmath47 and @xmath61 , with values given in section [ sec3.2 ] , with which we can determine heating functions of over 5000 flux tubes .
the left panel of figure [ fig6 ] shows the time profiles of the total _ in situ _ heating rate ( @xmath70 ) and total beam heating rate ( @xmath71 ) summed for all the flux tubes , in comparison with the observed uv total counts light curve and hxr 25 - 50 kev counts light curve .
the total heating rate ( @xmath72 ) is shown nearly scaled with the total uv count rate light curve from the rise to the peak note that the decay of the uv light curve is governed by coronal evolution , which is not part of the heating .
the beam driven energy flux is proportional to the observed @xmath325 kev hxr light curve .
the time integral of the total heating rate and beam heating rate yields the estimate of the total energy used to heat the corona and the total energy carried by beam - driven upflows in this flare , which are @xmath73 ergs and @xmath74 ergs , respectively .
the figure shows that the largest non - thermal partition amounts to about 40% , which occurs around the peak of hxr emission at 16:42 ut .
note that although the total heating flux @xmath39 in a single flux tube is considered to be proportional to the uv light curve at the foot - point , the sum of the total heating fluxes in all flux tubes , @xmath75 ) , is scaled with the total uv light curve by different values of @xmath47 during different phases of the flare , as can be seen from the figure .
@xmath47 is time - dependent in our method ; given the same amount of energy , beam heating produces greater uv emission compared with conduction heating @xcite . with these heating rates ,
we compute plasma evolution in over 5000 flare loops formed by reconnection and heated at different times and by different amounts of energy .
the distribution of the peak temperatures and densities of these flux tubes is shown in the right panels of figure [ fig6 ] .
for this flare , the peak temperature of most flux tubes varies from 7 to 26 mk , and the peak temperature distribution appears bi - modal .
the second peak in the distribution at around 22 mk is mostly contributed by flare loops formed around 16:45 ut , and is close to the effective temperature derived from the ratio of the two - channel ( 18 and 0.54 ) emission measured by _ goes _ with the isothermal assumption @xcite .
the peak density of the flux tubes ranges from 530@xmath76 @xmath35 .
the dem of the coronal plasma is readily calculated from the temperatures and densities of these loops .
figure [ fig7 ] shows the time evolution of the coronal plasma dem averaged every 20 seconds for the m8.0 flare .
note that the higher dem at about 2 mk at the beginning of the flare is due to background heating .
it is seen that in the rise phase of the flare , the dem increases toward higher temperature , suggesting that more flux tubes are heated to higher temperatures , while during the decay phase , the dem rises at lower temperature , reflecting cooling of the flux tubes .
the time evolution of the coronal dem will be used to calculate time - dependent sxr emission to verify the model results , as discussed in the next section .
the time - dependent coronal dem computed from over 5000 flux tubes is convolved with instrument response functions to compute synthetic sxr light curves and spectra , which are compared with those observed by _ goes _ and _ rhessi _ ( sections [ sec4.1 ] and [ sec4.2 ] , respectively ) .
the transition region dem during the decay phase is also derived from the coronal pressure @xcite , with which we calculate the the c iv line emission using the chianti atomic database @xcite , and compare it with uv 1600 observations ( section [ sec4.3 ] ) .
the two x - ray sensor ( xrs ) photometers onboard _ goes _ measure full - disk integrated sxr emissions in two energy bands , 18 and 0.54 , with 3-second cadence .
the solar 18 flux , which is dominated by continuum emission @xcite , is believed to originate from hot ( @xmath77 k ) plasma in coronal loops in active regions .
the full - disk sxr images observed by the solar x - ray imager ( sxi ; * ? ? ?
* ; * ? ? ?
* ) onboard _ goes _ indicate that noaa active region 10759 , where the m8.0 flare occurred , is the dominant sxr source on 2005 may 13 from 16:00 ut to 20:00 ut ( images not shown here ) .
therefore , the background - subtracted _ goes _ light curves reflect sxr emissions by the m8.0 flare .
the comparison of the synthetic sxr light curves from the modeled flare plasma with the observations by _ goes _ xrs yields the best - fit model parameter @xmath47 .
figure [ fig8 ] shows the synthetic sxr emissions in two channels plotted against the observed light curves . in both band - passes ,
the synthetic light curves follow the observed ones very well from the rise until 16:50 ut . in the 0.54 channel , the modeled and observed fluxes start to decline at around the same time , but the modeled flux decays more rapidly than the observed . in the 18 channel ,
the modeled radiation flux begins to drop while the observed radiation continues to rise for another 5 minutes .
we also note that at the start of the flare before 16:40 ut , the observed _ goes _
18 flux starts to rise earlier than the modeled flux . as discussed by @xcite , simply changing model parameters can not compensate for the flux deficiency ; instead , these discrepancies are most likely caused by weak heating events in the very early and late phases of the flare , which might not be identified in the uv foot - point emission .
_ rhessi _ is designed to observe solar high - energy emissions from sxrs to gamma - rays ( 3 kev up to 17 mev ) with an unprecedented combination of high time , spatial , and spectral resolutions @xcite .
whereas the thick - target hxr ( @xmath7820 kev ) observations provide us with a guide for constructing heating functions , we will further compare the model computed sxr ( @xmath7920 kev ) spectrum with that observed by _
rhessi_. the 3 to @xmath7920 kev x - ray emission is usually thermal bremsstrahlung radiation produced by plasmas in the flare loops heated to temperatures of a few to a few tens of mk .
_ rhessi _ images confirm that this emission is generated in the corona .
we calculate the time - dependent sxr spectrum using the dem distribution obtained from our model ( see figure [ fig7 ] ) . at each time
( 20 s cadence ) , we first calculate the x - ray spectrum observed at earth ( in units of photons @xmath37 s@xmath4 kev@xmath4 ) from the optically - thin thermal bremsstrahlung radiation using chianti @xcite , including both the line and continuum contributions and assuming solar coronal abundances @xcite and ionization equilibrium @xcite . the calculated spectrum is then convolved with the _ rhessi _ detector response @xcite , obtained from pre - flight and in - flight instrument calibration and modeling , to convert the photon spectrum to the count spectrum that would be directly observed by _ rhessi_. the detector response accounts for instrumental effects due to pulse pileup , escape of k - shell fluorescence photons from the detectors , attenuation by the aluminum shutters , and the energy - dependent detector effective area , and includes the additional in - flight corrections from @xcite . to get the best comparison of the spectra , only detector g4 is used since it had the best in - flight resolution at the time of the flare @xcite . by integrating the synthetic spectrum over different photon energy ranges
, we also obtain synthetic sxr light curves in units of counts rate .
figure [ fig9 ] shows the synthetic sxr light curves in the 36 , 612 , and 1225 kev bands , respectively , in comparison with _ rhessi _ observations . for most of the impulsive phase ,
the synthetic light curves agree very well with the observed ones .
the synthetic fluxes are lower than observed at the start of the flare , for a few minutes after 16:37 ut , when _ rhessi _ emerges from eclipse .
they then rise rapidly to catch up with the observed fluxes as more flare loops formed and heated by reconnection are identified in the uv foot - point observation .
the computed and observed fluxes overlap for 15 minutes from 16:44 ut until 16:59
ut , well after the impulsive phase .
afterwards , the computed fluxes start to drop faster than observed .
this is consistent with the model - observation comparison of the _ goes _ light curves , which usually yield slightly lower plasma temperatures than rhessi light curves ( e.g. , * ? ? ?
* ; * ? ? ?
more details can be learned by comparing the sxr spectra .
figure [ fig10 ] illustrates the synthetic and observed spectra at a few different times , indicated in the top panel of figure [ fig9 ] , during the flare .
( for a complete view , a movie comparing the modeled and observed sxr spectra , from 16:37 to 17:15 ut with 20 s cadence , is available http://solar.physics.montana.edu/wjliu/20050513/hsi_cmp_ebtel2.gif[online ] . )
figure [ fig11 ] summarizes the comparison of the time - dependent spectra with two quantities , the ratio of the synthetic to the observed counts , @xmath80 , averaged over the photon energy range @xmath81 from 6 to 15 kev , and the slope @xmath82 of this energy dependent ratio versus the photon energy , which is obtained by fitting the ratio to a linear function of the photon energy @xmath83 . if the synthetic spectrum is the same as the observations , @xmath84 and @xmath85 .
@xmath86 is a measure of magnitude comparison , and we regard empirically that the synthetic and observed spectra agree with each other when @xmath86 ranges between 0.7 and 1.3 .
@xmath82 is an indication of the plasma temperature distribution , since the thermal bremsstrahlung spectrum is temperature dependent . when @xmath87 , the model has produced more high temperature plasma than observed , and when @xmath88 , there is a lack of hot plasma in the modeled dem . in these plots ,
we only compare the spectrum up to 20 kev , beyond which the spectrum is likely dominated by thick - target bremsstrahlung produced by non - thermal electrons colliding at the lower atmosphere .
for most of the impulsive phase , from 16:44 ut to 16:59 ut , the synthetic spectra agree with the observations ( see panels b , c , and d in figure [ fig10 ] ) , with mean @xmath86 value between 0.7 and 1.3 and @xmath82 close to zero . from 16:50 to 16:55
ut , around the maxima of the _ goes _ sxr light curves , the modeled thermal emission dominates up to 20 kev ( see panels c and d ) , indicating strong heating in flare loops although the non - thermal hxr emission has become insignificant .
these plots showing good agreement between the model and observations suggest that the model has rather accurately reproduced the dem responsible for the thermal bremsstrahlung x - ray radiation in the observed range during this period . from the start of the flare until 16:42
ut , the synthetic spectrum is lower than observed by nearly the same fraction across the energy range up to 15 kev ( panel a in figure [ fig10 ] ) .
the synthetic flux at @xmath8915 kev is still smaller than observed , which may indicate a lack of high - temperature plasma in the model , but may also be due to predominant non - thermal bremsstrahlung radiation down to 15 kev .
_ rhessi _ images in the 1520 kev energy range during this period ( not shown ) show that the x - ray source is rather extended , likely including emissions from both the foot - point and loop - top , which can not be resolved with good photometric accuracy . therefore , both scenarios are plausible . in the late phase of the flare after 17:00
ut , the modeled flux drops below the observed flux again .
in this phase , thick - target hxr emission has ended , so the low flux above 15 kev is likely caused by a lack of high temperature plasma in the model .
we also note that in the early and late phases of the flare , when the average ratio @xmath86 is far below 1 , the slope @xmath82 becomes negative .
these indicate that the model does not generate enough high temperature plasma during these periods , maybe due to a lack of heating events identified in our method .
the bottom panel of the figure shows the number of heating events identified during the flare .
it is seen that the number of heating events drops rather significantly both before 16:39 ut and after 16:55 ut .
it appears that the shortage of heating events leads to insufficient x - ray flux in the calculation about 3 to 5 minutes later , which is roughly the timescale for temperature and density increases in newly formed flux tubes .
the lack of heating events identified from lower atmospheric signatures ( brightened uv or hxr foot - points ) could be explained by the fact that there is neither strong conduction flux nor significant non - thermal beam to deposit enough energy at the lower atmosphere to generate observable signatures . in the rise phase , it is plausible despite the significant non - thermal hxr emission , if the non - thermal electrons are partially or largely trapped in the corona and do not reach the chromosphere to generate hxr foot points ; the trapped electrons would then be a source of _ in situ _ heating in the corona .
such a mechanism has been proposed for the `` pre - impulsive '' phase of a few large flares ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) where non - thermal emission is observed in the corona with no identifiable hxr foot points , and may also contribute here .
comparison between the observed and synthetic spectra , as characterized by @xmath86 and @xmath82 , provides us with observational constraints to the heating function parameters @xmath47 and @xmath61 , in addition to the observed uv and hxr signatures at the flare foot - points that constrain the time profiles of the heating rate .
the long decay of _ trace _ 1600 emission in individual flaring pixels is reported by @xcite and @xcite .
flare emission in this broadband channel includes both the enhanced uv continuum and the c iv line emission . the c iv resonance doublet at 1548 and 1551
could only be emitted over a very narrow temperature range around 10@xmath90 k , and therefore the c iv intensities rise when the transition region is heated with more plasma raised to 10@xmath90 k during the flare . the c iv irradiance was observed to be enhanced by three to four orders of magnitude over the pre - flare emission during the impulsive phase of a large flare @xcite . during the decay phase , the c iv emission is dominated by the evolution of coronal pressure @xcite .
the solar uv continuum radiation in the 10002000 wavelength range is also enhanced during the flare ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
in particular , the continuum emission below 1682 is primarily contributed by bound - free transitions of si i , which is primarily excited by the c iv doublet @xcite .
therefore , the continuum intensity at @xmath91 1682 is approximately proportional to the c iv line intensity @xcite .
though there is no direct c iv measurement for the m8.0 flare studied in this paper , observations of stellar flares show that the time profile of c iv is similar to the uv 1600 emission in solar flares observed by _ trace _ , both exhibiting a fast rise and a long decay @xcite . in this study
, we have used the rapid rise of uv emission to constrain the time , duration , and magnitude of the heating rate in individual flare loops .
the gradual decay of uv emission from the same foot - points reflect the evolution of the over - lying corona that has been heated impulsively and then cools down on a much longer timescale . using the coronal plasma properties from the model
, we further calculate the c iv line emission during the decay phase to compare with the _ trace _ uv observation . to calculate the c iv line emission on the solar surface , it is important to know the opacity .
@xcite and @xcite measured the intensity ratio of two c iv lines at @xmath471548 and @xmath471550 observed on the disk , and found the ratio very close to 2:1 , which is expected for optically thin lines . @xcite also estimated the opacity of c iv to be 0.099 in the active - region spectrum .
we therefore consider that the c iv line is optically thin . the c iv photon flux at the surface
is then calculated through @xmath92 where @xmath93 is the abundance of carbon relative to hydrogen . in this paper the solar coronal abundances of @xcite
are used ; if photospheric abundances ( i.e. , * ? ? ?
* ) are used , the c iv line emission will be reduced by 15% .
@xmath94 is the contribution function for the c iv line and is obtained from chianti @xcite with the ionization equilibrium of @xcite .
it does not change significantly when the electron density @xmath14 varies from @xmath95 @xmath35 to @xmath96 @xmath35 .
dem@xmath97 is the transition region dem , where @xmath98 measures distance along the flux tube from the base of the transition region . for each flux tube , we assume that the transition region is nearly in static equilibrium in the decay phase when the heating has finished , which is roughly when the coronal density peaks . in this case ,
thermal conduction is balanced by radiation , and pressure is uniform from the corona through the transition region when gravity is ignored , and the transition region dem can be calculated as @xcite : @xmath99 where @xmath100 , @xmath101 is the temperature at the base of the transition region , taken to be a nominal 10@xmath102 k , and @xmath28 is the optically - thin radiative loss function . the c iv line emission of one flux tube in its decay phase is calculated with equations ( [ eq_civ ] ) and ( [ eq_dem ] ) and convolved with the _ trace _ 1600 band response function to get synthetic uv flux in units of dn s@xmath4 . the left panel of figure [ fig12 ] shows an example of the computed c iv line emission in one flux tube .
the computed c iv flux appears to evolve along with the observed decay , though the synthetic flux is smaller than the observed by a factor of 2 .
the right panel of the figure shows the sum of the synthetic uv flux in all flaring pixels in comparison with the observed total uv emission in the decay phase .
the observed uv decay in a pixel is obtained by subtracting from the uv light curve the full gaussian derived by fitting the uv rise so that the transition region response to the heating is excluded .
similar to the single - pixel comparison , the computed total c iv emission decays on the same timescale as observed , and the computed flux is lower than observed by a factor of 2 to 3 .
we find this comparison satisfactory for the following reasons .
first , the transition region is not exactly in static equilibrium . during the decay phase , the transition region loss usually exceeds the conductive flux , which leads to downflows from the corona , and therefore the coronal density gradually decreases .
the downflow is literally an energy input into the transition region and gives rise to a greater transition region dem than in the case of static equilibrium .
second , the uv continuum in the decay phase may be also enhanced due to irradiation from the c iv line .
the observed uv emission is the sum of both the line and continuum emissions , though the continuum intensity is considered to be proportional to the line intensity .
all these reasons would explain the observed uv emission being larger than the computed c iv line emission with the static equilibrium assumption .
it is , nevertheless , very encouraging to find that the computed line emission decays at the same rate as the observed uv emission .
in contrast , the computed x - ray emission decays faster than observed , as shown in the previous section .
since the uv decay is derived from the same pixels whose rise phase is used to construct the loop heating rate , such comparison indirectly confirms that the rapid decay of the computed sxr flux most likely indicates the presence of additional heating events in the decay phase , which are not found in foot - point uv radiation signatures .
this is consistent with decades of observations which have shown that coronal plasma temperatures decrease more slowly than would be expected from simple calculations of radiative and conductive cooling ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , implying that additional heating must be taking place despite an apparent lack of accelerated particles .
these additional heating events may thus result from continuous reconnection in the corona that provides direct heating without significantly accelerating non - thermal particles , or which accelerates particles sufficiently weakly such that they thermalize in the ambient plasma at the loop - top ; both scenarios lead to increased sxr coronal emission without associated chromospheric emission that could be identified in _ trace _ broadband uv or _ rhessi _ hxr observations .
a similar mechanism has been proposed for the so - called `` euv late phase '' emission from eruptive ( and often cme - associated ) flares @xcite , where reconnection high in the corona heats post - flare loops , resulting in increased euv coronal emission without associated foot - point brightenings .
we have analyzed and modeled an m8.0 flare on 2005 may 13 observed by _ trace _ , _ rhessi _ , and _ goes _ , in order to determine heating rates in a few thousand flare loops formed by reconnection and subsequently heated .
the rapid rise of spatially resolved uv emission from the lower atmosphere is considered to be the instantaneous response to heating in the flare loops rooted at the foot points , and therefore is used to construct heating rates in individual flare loops to describe when , for how long , and by how much flare loops are heated @xcite . with these heating rates ,
we compute the coronal plasma evolution in these loops using the 0d ebtel model , and compare , for the first time , the computed time - dependent coronal radiation with observations by _ rhessi _ and _ goes _ , and as well the computed uv line emission with _ trace _ observations during the decay phase of flare loops .
the comparison constrains parameters of the loop heating rates , which define the amount of total energy flux and the fraction of energy flux carried by chromospheric upflows driven by non - thermal beams .
the m8.0 flare studied in this paper is observed to have significant thick - target hxr emission , suggesting that the lower atmosphere is also heated by electron beams during the impulsive phase . in this process ,
chromospheric evaporation is driven carrying mass and energy flux back to the corona . in this study
, we experimentally scale the beam - driven energy flux with the observed hxr light curve , and examine how varying the fraction of beam heating changes the modeled coronal temperature and density evolution . in general , for a given total amount of heating flux , a larger fraction of energy carried by beam driven upflow leads to higher density , lower temperature , and more rapid evolution of the coronal plasma .
these properties affect the computed sxr spectrum and its time evolution , and can be compared against the spectrum observed by _
rhessi_. our experiment yields an optimal set of parameters , with which , the model computed sxr light curves and spectra agree very well with observations during the rise phase and peak of the flare for more than 10 minutes .
apart from comparison with the sxr observations , we also compute the c iv line emission at the foot of the flare loop during the decay , whose rise phase is used to construct the heating rate in the same loop .
emission in this optically - thin line during the decay phase is governed by plasma evolution in the overlying coronal loop and contributes significantly to the uv emission observed in the _ trace _ 1600 bandpass .
it is shown that the computed c iv flux is about one third of the observed uv flux , which also includes the continuum radiation , and decays on the same timescale as observed .
this experiment shows an avenue to model heating and cooling of spatially resolved flare loops with self - consistent constraints by high - resolution uv observations from the input ( uv rise ) to the output ( uv decay ) in the same flare loop .
these results suggest that our method , which employs all available observations to constrain flare loop heating model , is able to capture the distribution of impulsive heating rates in numerous flare loops formed and heated sequentially by magnetic reconnection .
the optimal heating function parameters determined for this flare yield the peak heating flux in the range of 10@xmath10310@xmath50 ergs @xmath37 s@xmath4 for over 5000 flare loops of cross - sectional area 1 by 1 .
notably , only @xmath7940% of this energy is carried by beam - driven upflows during the impulsive phase , with the remaining @xmath7860% contributed by _ in situ _ heating in the corona .
the beam heating occurs mostly during the first ten minutes , when the flare exhibits significant thick - target hxr emission ; meanwhile , the _ in situ _ heating energy in the impulsive phase amounts to 6.2@xmath010@xmath2 ergs , over 3 times greater than the beam - driven contribution . in the post - impulsive phase ( after 16:50 ut ) ,
the continuously expanding uv ribbons and the observed neupert effect between the sxr derivative and uv ( rather than hxr ) light curves indicate that new flare loops are formed by continuous reconnection , and _ in situ _ heating in the corona is predominant , amounting to 4.1@xmath010@xmath2 ergs out of the total , with negligible beam heating .
the total energy used in coronal heating in this flare amounts to 1.22@xmath010@xmath1 ergs , of which the total energy carried by beam - driven upflows is estimated to be only 1.9@xmath010@xmath2 ergs ( @xmath10416% ) .
therefore , it appears that _ in situ _ coronal heating and thermal conduction play an important role in the energetics of this m8.0 flare , in contrast to commonly accepted models that consider the coronal thermal plasma to be largely a by - product of beam - driven heating of the lower atmosphere ( chromosphere and transition region ) ( see also discussions by * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
a survey of 37 m- and x - class flares @xcite suggests that _ in situ _ heating may be significant even down to mid - c class , and therefore this heating mechanism must be considered in future studies of flare energetics .
we note that these energies set lower limits for this m8.0 flare .
first , despite the excellent agreement between the model and observation in the rise phase of the flare , the model computed x - ray flux decays more quickly than observed by both _ goes _ and _ rhessi_. it is most likely that magnetic reconnection and energy release continue in the high corona during the decay of the flare .
these heating events might not produce significant uv foot - point emission , and are therefore not identified in our method . for comparison , @xcite analyzed the same flare , and estimated the total energy from _ goes _ ( including energy loss from radiation , conductive cooling and enthalpy flux ) to be 3.1@xmath010@xmath1 ergs .
second , in our experiment , we do not model heating and dynamics of the lower atmosphere , but use a simple scaling relationship to estimate the energy flux carried by non - thermal beam - driven upflows back to the corona .
this amount of energy is a fraction of the total energy carried by non - thermal beams ; the remainder is lost in the lower atmosphere . as a reference
, we fit the hxr spectrum to a power - law distribution with the standard _ rhessi _ software package ( not shown in the paper ) , and , with the low - energy cutoff and spectral index from the fit , estimate the total non - thermal energy to be 7.6@xmath105 ergs , about 4 times the energy carried in the upflow but , interestingly , still somewhat smaller than the energy required to heat the coronal plasma though we note that the non - thermal energy estimate is only a loose lower bound since it depends critically on the fit low - energy cutoff , which is unbounded from below due to obscuration by the highly - dominant thermal emission .
our experiment provides a novel method for investigating energy release in solar flares , which is governed by reconnection and substantiated in formation and heating of numerous flare loops .
this approach has a few advantages .
first , by analyzing the foot - point uv signatures combined with hxr observations , for the first time , we are able to identify and characterize the energy release process in a few thousand flare loops down to 1 by 1 scale , which is thought to be close to the basic scale of flux tubes formed in patchy reconnection . using the _ same _ uv data , we also measure the rate of magnetic reconnection and hence are able to establish the relationship between reconnection and energy release in temporally and spatially resolved manner .
second , using the 0d ebtel model , we are able to efficiently compute plasma properties of this large number of flux tubes , and naturally generate a time - dependent differential emission measure ( dem ) from these flare loops formed and heated at different times and evolving independent of one another .
subsequently , we can predict x - ray and uv radiation signatures directly comparable with observations , which provides further constraints to our determination of energy release rates .
the method utilizes emission signatures at all available wavelengths in both the foot - points and overlying coronal loops .
the dynamics of the lower atmosphere and corona are strongly coupled ; they are governed by different yet coherent physics , and hence should be studied coherently . in this study , based on the insight from previous theoretical and numerical research , we employ some empirical laws to prescribe the relation between heating rates and radiation signatures .
the optimal parameters obtained from the analysis can provide a reference for further investigation based on physical models , such as models of lower atmosphere heating and dynamics .
we also recognize that the 0d ebtel model has limitations in accurately describing coronal plasma properties along the loop .
nevertheless , the model is highly efficient in dealing with a few thousand flare loops and yields the first - order heating rates as useful inputs for more sophisticated loop heating models such as those by @xcite , @xcite , or @xcite .
our study of this event shows very good agreement between observed and modeled high temperature plasma radiations ( e.g. , by _
rhessi _ and _ goes _ ) , but lacks observations of and therefore comparison with low temperature plasma signatures , which are crucial in understanding the flare loop evolution beyond the conduction - dominant regime .
the recently launched _
sdo _ has observed many flares with aia , the high - resolution imaging telescope , in a number of uv and euv bands @xcite , and as well with the broad - band , high resolution spectrometers on the extreme ultraviolet variability experiment ( eve ; * ? ? ? * ) .
these new observations , which cover coronal temperatures from @xmath791 mk up to @xmath10420 mk , can be combined with rhessi observations that provide information of hot ( @xmath781050 mk ) plasmas , as well as information about the non - thermal emission , to fully observationally characterize the temperature distribution and its evolution ( caspi et al .
2013b , in prep . ) , which can be compared with and used to constrain our modeling .
these new facilities will allow us to analyze and model flare plasmas following the entire process of heating and cooling , and also to study resolved individual flare loops in multiple wavelengths in more details , which will improve determination of the energy release rate during magnetic reconnection .
finally , the poor comparison between model and observations during the decay phase of the flare reflects some limitations of the present method .
these can be addressed by more sophisticated one - dimensional modeling that also includes the lower atmosphere , to help determine the amount of lower atmosphere radiation , such as the uv 1600 emission used in this study , as dependent on the heating mechanism and the amount of heating .
such effort will help clarify whether the shortage of computed flux is produced by the method missing weak heating events , or by other significant effects related to imperfect model assumptions .
we thank drs .
j. klimchuk and g. fisher for insightful discussion and the anonymous referee for thoughtful and valuable comments .
the work is supported by nsf grant atm-0748428 .
a. caspi was supported by nasa contracts nas5 - 98033 and nas5 - 02140 , and nasa grants nnx08aj18 g and nnx12ah48 g .
chianti is a collaborative project involving the university of cambridge ( uk ) , george mason university , and the university of michigan ( usa ) .
, n. w. , fisher , g. h. , & siegmund , o. h. w. 1998 , in astronomical society of the pacific conference series , vol .
154 , cool stars , stellar systems , and the sun , ed .
r. a. donahue & j. a. bookbinder , 621 , k. k. , winter , h. d. , & larson , n. l. 2012 , in astronomical society of the pacific conference series , vol . 455 , 4th hinode science meeting : unsolved problems and recent insights , ed . l. bellot rubio , f. reale , & m. carlsson , 199 | we analyze and model an m8.0 flare on 2005 may 13 observed by the _ transition region and coronal explorer _ ( _ trace _ ) and _ reuven ramaty high energy solar spectroscopic imager _ ( _ rhessi _ ) to determine the energy release rate from magnetic reconnection that forms and heats numerous flare loops .
the flare exhibits two ribbons in uv 1600 emission .
analysis shows that the uv light curve at each flaring pixel rises impulsively within a few minutes , and decays slowly with a timescale longer than 10 minutes . since the lower atmosphere ( the transition region and chromosphere ) responds to energy deposit nearly instantaneously , the rapid uv brightening is thought to reflect the energy release process in the newly formed flare loop rooted at the foot point . in this paper
, we utilize the spatially resolved ( down to 1 ) uv light curves and the thick - target hard x - ray emission to construct heating functions of a few thousand flare loops anchored at the uv foot points , and compute plasma evolution in these loops using the enthalpy - based thermal evolution of loops ( ebtel ) model . the modeled coronal temperatures and densities of these flare loops
are then used to calculate coronal radiation .
the computed soft x - ray spectra and light curves compare favorably with those observed by _
rhessi _ and by the _ geostationary operational environmental satellite _ ( _ goes _ ) x - ray sensor ( xrs ) .
the time - dependent transition region differential emission measure ( dem ) for each loop during its decay phase is also computed with a simplified model and used to calculate the optically - thin c iv line emission , which dominates the uv 1600 bandpass during the flare .
the computed c iv line emission decays at the same rate as observed .
this study presents a method to constrain heating of reconnection - formed flare loops using all available observables independently , and provides insight into the physics of energy release and plasma heating during the flare . with this method ,
the lower limit of the total energy used to heat the flare loops in this event is estimated to be 1.22@xmath010@xmath1 ergs , of which only 1.9@xmath010@xmath2 ergs is carried by beam - driven upflows during the impulsive phase , suggesting that the coronal plasmas are predominantly heated _ in situ_. |
in 1898 , heegaard @xcite introduced a heegaard splitting for a closed , orientable , triangulated 3-manifold , i.e. , there is a closed , orientable surface cutting this manifold into two handlebodies . later , moise @xcite proved that every closed , orientable 3-manifold admits a triangulation .
so each closed orientable 3-manifold admits a heegaard splitting .
this makes studying 3-manifolds through heegaard splittings possible .
one astonishing result proved by haken @xcite is that if all heegaard splitting of a 3-manifold are reducible , i.e. , there is an essential simple closed curve in heegaard surface bounding essential disks on both sides , then this manifold is reducible .
later , casson and gordon @xcite defined a weakly reducible heegaard splitting and proved that if a 3-manifold has a weakly reducible and irreducible heegaard splitting , then it contains an embedded closed incompressible surface , i.e. , it is haken .
both of these two phenomenons drive people to think how heegaard splittings reflect 3-manifolds . for classifying 3-manifolds ,
thurston @xcite introduced the geometrization conjecture ( haken version proved by thurston @xcite and full version proved by perelman @xcite ) as follows : for any closed , irreducible , orientable 3-manifold , there are finitely many disjoint , non isotopy essential tori so that after cutting the manifold along those tori , each piece is one of eight geometries . among all of these eight geometries ,
one is hyperbolic , another one is solvable and the left six pieces are seifert . in these eight geometries
, it is known that seifert 3-manifolds have been completely classified .
moreover all of their irreducible heegaard splittings are either vertical or horizontal , see @xcite .
cooper and scharlemann @xcite studied all irreducible heegaard splittings of a solvmanifold . and
there are series of works on heegaard splittings of some typical 3-manifolds , such as lens space , surface @xmath0 etc .
with the curve complex defined by harvey @xcite , hempel @xcite introduced an index - heegaard distance for studying heegaard splitting .
basically , this index- heegaard distance is defined to be the length of a shortest geodesic in curve complex which connects these two boundaries of essential disks from different sides .
then he proved that all heegaard splittings of a seifert 3-manifold have distance at most two ; if a 3-manifold contains an essential torus , then all heegaard splittings of it have distance at most two , where this result is also proved by hartshorn @xcite and scharlemann @xcite . combined with the geometrization conjecture , if a 3-manifold admits a heegaard splitting with heegaard distance at least three , then it is hyperbolic .
so it seems that if we fully understand all of distance two heegaard splittings , then we can fully answer the question that how heegaard splittings reflect 3-manifolds .
so this question is reduced to what dose a 3-manifold look like if it only admits distance at most two heegaard splittings ?
[ ques:1 ] since the hyperbolic 3-manifolds are the most concerned , given a distance two heegaard splitting , it is interesting to know whether the corresponding manifold is hyperbolic or not . by the definition of a distance two heegaard splitting
, there is an essential simple closed curve and a pair of essential disks from different handlebodies so that this curve is disjoint from those two essential disks boundaries .
it seems that this heegaard splitting is simple and hence whether the manifold is hyperbolic or not should not be hard to answer .
however , things for distance two heegaard splittings are complicated because there are examples showing that a 3-manifold admitting a distance two heegaard splitting could be seifert , hyperbolic or contains an essential torus , see @xcite .
thompson @xcite studied all distance two genus two heegaard splittings and found that even for genus two heegaard splittings , those manifolds could be very complicated .
later , rubinstein and thompson @xcite extended this result to genus at least three cases .
but their results give no sufficient conditions to determine whether it is hyperbolic or not . according to geometrization conjecture , except small seifert 3-manifolds and hyperbolic 3-manifolds ,
all of 3-manifolds contain essential tori .
it is known that the heegaard splittings of small seifert 3-manifolds are well understood .
so to answer the question [ ques:1 ] , the first step is to understand possible essential tori in a 3-manifold . in @xcite , the authors studied the curve complex and introduced the definition of a locally large geodesic .
then they constructed infinitely many arbitrary large distance heegaard splitting . in the proof of theorem 1.3 in @xcite
, they found that the locally large property of geodesics forces any geodesic realizing heegaard distance to share some vertex @xmath1 in common .
so if the resulted manifold contains an essential torus @xmath2 , then @xmath2 intersects this heegaard surface in some essential simple closed curves , which are all isotopic to @xmath1 .
thus @xmath2 intersects that heegaard surface in fixed essential simple closed curves . under this circumstance
, it seems the corresponding 3-manifold is not hard to understand . for this reason
, we introduce the definition of a locally large distance two heegaard splitting . a length two geodesic realizing heegaard distance is @xmath3 , where @xmath4 and @xmath5 bound essential disks from two sides of the heegaard surface and @xmath1 is disjoint from both @xmath4 and @xmath5 . as we know , there is a length two geodesic contains a non separating essential simple closed curve as its middle vertex realizing heegaard distance .
so we assume that @xmath1 is represented by a non separating essential simple closed curve .
let @xmath6 be this heegaard surface .
we say the geodesic @xmath7 is locally large if for the surface @xmath8 , @xmath9 , for any pair of @xmath10 and @xmath11 disjoint from @xmath1 , where both @xmath10 and @xmath11 bound essential disks in different sides of @xmath6 respectively .
moreover , by the definition of a locally large geodesic , we say a heegaard splitting is locally large if there is a locally large geodesic realizing heegaard distance .
the main result is if a closed orientable manifold @xmath12 admits a locally large distance two heegaard splitting @xmath13 , then @xmath12 is a hyperbolic manifolds or an amalgamation of a hyperbolic manifold and a small seifert manifolds along an incompressible torus .
moreover there is only one essential torus in the non - hyperbolic case up to isotopy .
[ thm1 ] a 3-manifold is almost hyperbolic if either it is hyperbolic or it is an amalgamation of a hyperbolic 3-manifold and a small non hyperbolic 3-manifold along an incompressible torus
. then the conclusion of theorem [ thm1 ] says that a 3-manifold admitting a locally large distance two heegaard splitting is almost hyperbolic .
the geometrization conjecture indicates that \(1 ) a seifert 3-manifold does not admits a complete hyperbolic structure ; \(2 ) a solvmanifold does not admits a complete hyperbolic structure ; \(3 ) an amalgamation of a complete hyperbolic 3-manifold and a small seifert 3-manifold along a torus is not one of those eight geometries , see @xcite .
thus combined with the result of theorem [ thm1 ] , neither a solvmanifold nor a seifert 3-manifold admits a locally large distance two heegaard splitting . according to geometrization conjecture ,
the most important piece of those eight geometries is hyperbolic . thus giving a sufficient condition for a hyperbolic 3-manifold is critical in studying heegaard splittings .
but the example in section [ sec3 ] shows that the manifold @xmath12 in theorem [ thm1 ] could be a non hyperbolic manifold .
so to give a sufficient condition for a hyperbolic 3-manifold , we need to eliminate the possible essential torus in theorem [ thm1 ] . for this purpose ,
we introduce some definitions .
an essential simple closed curve in heegaard surface is a co - core for a handlebody if there is an essential disk in this handlebody so that its boundary intersects this curve in one point .
the definition of a domain for an essential simple closed curve is in section [ sec5 ] .
the proof of theorem [ thm1 ] implies that under the locally large condition of @xmath1 , the non hyperbolic case happens if two copies of @xmath1 bounds essential tori in both sides of @xmath6 and @xmath1 is not co - core on either side of @xmath6 .
so , suppose that a closed orientable manifold @xmath12 has a locally large distance two heegaard splitting @xmath13 .
let @xmath1 be an essential simple closed curve disjoint from a pair of essential disks from different sides of @xmath6 .
then @xmath12 is hyperbolic if and only if either all domains of @xmath1 have euler characteristic number less than -1 or @xmath1 is co - core for one side of @xmath6 .
[ thm2 ] we introduce the definition of a geodesic of curve complex in section [ sec2 ] , construct a non hyperbolic 3-manifold in section [ sec3 ] and prove theorem [ thm1 ] and [ thm2 ] in section [ sec4 ] and [ sec5 ] .
* acknowledgements .
* the authors would like to thank tao li for careful checking the earlier version of this manuscript , pointing out the genus two case of theorem [ thm1 ] and many helpful conversations .
let @xmath6 be a compact surface of genus at least 1 , and @xmath14 be the curve complex of @xmath6 .
we call a simple closed curve @xmath15 in @xmath6 is essential if @xmath15 bounds no disk in @xmath6 and is not parallel to @xmath16
. it is known that each vertex of @xmath17 is represented by the isotopy class of an essential simple closed curve in @xmath6 . for simplicity
, we do not distinguish the essential simple closed curve @xmath15 and its isotopy class @xmath15 without any further notation .
harvey @xcite defined the curve complex @xmath18 as follows : the vertices of @xmath18 are the isotopy classes of essential simple closed curves on @xmath6 , and @xmath19 distinct vertices @xmath20 determine a k - simplex of @xmath18 if and only if they are represented by pairwise disjoint essential simple closed curves .
for any two vertices @xmath21 and @xmath22 of @xmath18 , the distance of @xmath21 and @xmath22 , denoted by @xmath23 , is defined to be the minimal number of 1-simplexes in a simplicial path joining x to y. in other words , @xmath24 is the smallest integer @xmath25 such that there is a sequence of vertices @xmath26 such that @xmath27 and @xmath28 are represented by two disjoint essential simple closed curves on @xmath6 for each @xmath29 .
for any two sets of vertices in @xmath18 , say @xmath30 and @xmath31 , @xmath32 is defined to be @xmath33 .
for the torus or once punctured torus case , masur and minsky @xcite define @xmath14 as follows : the vertices of @xmath34 are the isotopy classes of essential simple closed curves on @xmath6 , and @xmath19 distinct vertices @xmath20 determine a k - simplex of @xmath18 if and only if @xmath28 and @xmath35 are represented by two simple closed curves @xmath36 and @xmath37 on @xmath6 such that @xmath36 intersects @xmath37 in just one point for each @xmath38 .
the following lemma is well known , see @xcite .
@xmath14 is connected , and the diameter of @xmath14 is infinite .
[ lem2.1 ] a collection @xmath39 is a geodesic in @xmath17 if @xmath40 and @xmath41 for any @xmath42 . and the length of @xmath7 denoted by @xmath43 is defined to be @xmath44 . by lemma [ lem2.1 ] ,
there is a shortest path in @xmath45 connecting any two vertices of @xmath17 .
thus for any two distance @xmath44 vertices @xmath4 and @xmath46 , a geodesic @xmath7 connects @xmath4 and @xmath46 if @xmath47 . now for any two sub - simplicial complex @xmath48 , a geodesic @xmath7 realizing the distance of @xmath30 and @xmath31 if @xmath7 connects an element @xmath49 and an element @xmath50 so that @xmath51 .
let @xmath52 be a compact surface of genus at least 1 with non - empty boundary .
similar to the definition of the curve complex @xmath53 , we can define the arc and curve complex @xmath54 as follows : each vertex of @xmath54 is the isotopy class of an essential simple closed curve or an essential properly embedded arc in @xmath52 , and a set of vertices form a simplex of @xmath54 if these vertices are represented by pairwise disjoint arcs or curves in @xmath52 . for any two disjoint vertices , we place an edge between them .
all the vertices and edges form 1-skeleton of @xmath54 , denoted by @xmath55 . and for each edge , we assign it length 1 .
thus for any two vertices @xmath4 and @xmath5 in @xmath55 , the distance @xmath56 is defined to be the minimal length of paths in @xmath55 connecting @xmath4 and @xmath5 .
similarly , we can define the geodesic in @xmath54 . when @xmath52 is a subsurface of @xmath6 , we call @xmath52 is essential in @xmath6 if the induced map of the inclusion from @xmath57 to @xmath58 is injective
. furthermore , we call @xmath52 is a proper essential subsurface of @xmath6 if @xmath52 is essential in @xmath6 and at least one boundary component of @xmath52 is essential in @xmath6 . for more details ,
see @xcite .
so if @xmath52 is an essential subsurface of @xmath6 , there is some connection between the @xmath54 and @xmath17 . for any @xmath59
, there is a representative essential simple closed curve @xmath60 such that the intersection number @xmath61 is minimal .
hence each component of @xmath62 is essential in @xmath52 or @xmath63 .
now for @xmath64 , let @xmath65 be isotopy classes of the essential components of @xmath62 . for any @xmath66 , @xmath67 if and only if @xmath68 is the essential boundary component of a closed regular neighborhood of @xmath69
. now let @xmath70 .
then the map @xmath71 is the subsurface projection defined in @xcite .
we say @xmath59 cuts @xmath52 if @xmath72 . if @xmath4 , @xmath73 both cut @xmath52 , we write @xmath74 . and if @xmath75 , then @xmath76 and @xmath77 .
the following is immediately followed from the above observation .
let @xmath52 and @xmath6 be as above , @xmath78 be a geodesic of @xmath17 such that @xmath79 cuts @xmath52 for each @xmath80
. then @xmath81 .
[ lem : subsurface projection ] for essential curves @xmath82 in @xmath6 , let @xmath83 be the minimal geometric intersection number up to isotopy .
we call @xmath4 and @xmath5 intersect efficiently if the number of @xmath84 is equal to @xmath83 . one tool for studying the intersection between essential simple closed curves and arcs in @xmath6 is bigon criterion .
@xcite let surface @xmath6 be as above .
then for any two essential curves @xmath85 in @xmath6 , @xmath4 and @xmath46 intersects efficiently if and only if @xmath86 bounds no bigon or half - bigon in @xmath6 .
[ lem : bigon criterion ] assume that @xmath87 is a non - trivial compression body , i.e. , not the product i - bundle of a closed surface .
then there is an essential simple closed curve in @xmath88 bounding an essential disk in @xmath87 .
let @xmath6 be an essential subsurface in @xmath89 .
we call @xmath6 is a hole for @xmath87 if for any essential disk @xmath90 , @xmath91 .
furthermore , we call an essential subsurface @xmath92 is an incompressible hole for @xmath87 if @xmath6 is a hole for @xmath87 and is incompressible in @xmath87 .
otherwise , @xmath6 is a compressible hole for @xmath87 .
masur and schleimer@xcite studied the subsurface projection of an essential disk , and proved that : let @xmath87 be a non - trivial compression body and @xmath6 be a compressible hole for @xmath87 . then for an essential disk @xmath93 in @xmath87 , there are essential disks @xmath94 and @xmath95 satisfying : @xmath96 for @xmath97 , @xmath98 and @xmath16 , they intersect efficiently ; @xmath96 @xmath99 ; @xmath96 there is one component of @xmath100 is disjoint from an component of @xmath101 and @xmath102 .
furthermore , @xmath103 .
[ lem : compressible hole ] see the proof of lemma 11.5 and lemma 11.7 @xcite .
let @xmath104 be a collection of some different points in @xmath6 .
for the manifold @xmath105 , @xmath106 is a collection of essential simple closed curves .
a closed orientbale 3-manifold is seifert if it is obtained by doing dehn surgeries along @xmath107 as follows .
we remove a regular neighborhood of @xmath108 and glue back a solid torus where the meridian curve coincides with some @xmath109 slope . if @xmath110 , then this fiber is called exceptional . as we know , if all of these exceptional fiber are removed , then @xmath111 is @xmath52 @xmath112 @xmath113 .
somehow , @xmath12 is represented as follows : @xmath114 , where @xmath6 is called the base surface .
it is known that this representation is unique with some permutations in order . in studying irreducible heegaard splittings of a seifert manifold @xmath12
, there are two standard ones named as vertical and horizontal heegaard splittings .
to be clear , for a seifert manifold @xmath114 with projection @xmath115 , let @xmath116 be a cell decomposition where each component of @xmath93 or @xmath52 contains at most one singular point in its interior and each component of @xmath117 is a square with one pair of opposite edges in @xmath93 and the other one in @xmath52 , where both @xmath118 and @xmath119 are connected .
then the union of @xmath120 $ ] is a handlebody and @xmath121 , the complement of @xmath122 in @xmath12 , is also a handlebody which is homeomorphic to @xmath123 $ ] , where @xmath124/\sim$ ] .
so @xmath125 is a heegaard splitting of @xmath12 , called a vertical heegaard splitting .
the construction of a vertical heegaard splitting shows that for each seifert manifold , it admits a vertical heegaard splitting .
knowing that fact , people wonder that whether there is another type heegaard splitting for a seifert manifold in general or not .
however , there is no other type of heegaard splitting for a seifert manifold in general . in @xcite , moriah and schultens
proved that except some special cases , almost all of seifert 3-manifolds admit only vertical heegaard splittings .
they @xcite also showed that there are horizontal heegaard splittings for some seifert 3-manifolds as follows . taking a surface bundle @xmath126 , where @xmath127 with one boundary component and @xmath128 is a periodic homeomorphism and fixes @xmath129 point by point .
let @xmath12 be a dehn filling of @xmath130 , where the longitude goes to @xmath129 .
then @xmath131 bounds an annulus @xmath132 in @xmath133 which cuts out an i - bundle of @xmath132 .
it is not hard to see that @xmath134 cuts @xmath12 into two handlebodies , where both of these two handlebodies are compact surfaces product i - bundles .
so it gives a heegaard splitting of @xmath12 , called a horizontal heegaard splitting .
a result in @xcite said that a seifert manifold admits a horizontal heegaard splitting if and only if its euler number is zero .
moreover , they @xcite proved that all irreducible heegaard splittings of a seifert manifold is either vertical or horizontal . from the definition of a seifert 3-manifold , if @xmath12 has a genus at least 1 base surface @xmath6 or @xmath135 but with at least 4 exceptional fibers , @xmath12 contains an essential torus @xmath2 . then for any strongly irreducible heegaard splitting @xmath136 , there are two essential annuli @xmath137 and @xmath138 with @xmath139
. if @xmath12 has @xmath135 as its base surface with at most three exceptional fibers , then \(1 ) for a vertical heegaard splitting , there are essential disks @xmath140 and @xmath141 from two sides of heegaard surface so that their boundaries intersects in at most two points ; \(2 ) for a horizontal heegaard splitting , there is an essential simple closed curve @xmath142 and two essential annuli @xmath143 $ ] in @xmath122 and @xmath144 $ ] in @xmath121 so that @xmath145 contains at most one point .
[ lem : seifert1 ] the proof of second part is contained in the proof of theorem 3.5 in @xcite .
so all we need to prove is the first part . since a weakly reducible
heegaard splitting satisfies the conclusion , we only consider all strongly irreducible vertical heegaard splittings of it .
if this vertical heegaard splitting has genus at least 3 , corollary 3.3 in @xcite says that it has distance at most 1 .
hence there are two essential disks satisfying the conclusion of lemma [ lem : seifert1 ] .
if this vertical heegaard splitting has genus 2 , by the definition , one handlebody @xmath122 is the union of two closed neighborhood of exceptional fibers and a rectangle @xmath146 $ ] and the other one @xmath121 is homeomorphic to an i bundle of one - holed torus with a non trivial dehn surgery .
it is not hard to see that removing two exceptional fibers reduces @xmath12 into a torus @xmath112 i with a non trivial dehn surgery , where we do the dehn surgery along an simple closed curve @xmath142 and the union of a longitude and @xmath142 bounds an embedded annulus .
the rectangle @xmath146 $ ] is isotopic to the closed neighborhood of an properly embedded unknotted arc which connects these two boundaries . after removing a rectangle
@xmath146 $ ] , it is changed into the handlebody @xmath121 .
let @xmath10 be an properly embedded arc in this rectangle where it connects a pair of opposite edges .
then @xmath147 $ ] bounds an essential disk @xmath140 in @xmath121 .
let @xmath11 be an properly embedded essential arc in once punctured torus which intersects the longitude empty .
then @xmath148 $ ] bounds an essential disk @xmath141 in @xmath121 .
it is not hard to see that @xmath149 intersects @xmath150 in two points .
hempel @xcite showed that for a vertical heegaard splitting , its genus equal to the sum of the number of rectangles and 1 .
so it means that only a seifert 3-manifold with base surface @xmath135 with at most three exceptional fibers admits a genus 2 vertical heegaard splitting . by the proof of lemma [ lem : seifert1 ] , for a strongly irreducible vertical heegaard splitting of genus 2
, there are two essential disks @xmath140 and @xmath141 of two sides so that there are two non isotopy essential simple closed curve @xmath151 and @xmath152 in heegaard surface disjoint from both of them .
[ cor:1 ] let @xmath140 and @xmath141 be as in lemma [ lem : seifert1 ] . it is known that @xmath121 is an once punctured torus i - bundle with a non trivial dehn surgery .
let @xmath151 be a longitude in upper boundary and @xmath152 be a longitude in lower boundary .
then these two curves satisfy the conclusion .
let @xmath12 be a compression body with genus @xmath153 , where @xmath154 , and @xmath155 be a torus .
then there are two non - separating spanning annuli @xmath156 and @xmath157 , i.e. , the boundary of an essential annulus lies in different components of @xmath158 , such that @xmath159 are two handlebodies @xmath160 and @xmath161 with same genera , see figure [ fig : spanning annulus ] . ] from figure [ fig : spanning annulus ] , @xmath162 ( resp .
@xmath163 ) consists of @xmath164 and an annulus @xmath165 ( resp . @xmath166 and an annulus @xmath167 ) .
since @xmath164 and @xmath166 are homeomorphic , there is a orientation reversing homeomorphism @xmath168 such that @xmath169 .
since @xmath164 is a genus @xmath170 surface with two boundary , the projective measured lamination space of @xmath164 @xmath171 is not empty .
it is known that the isotopy class of the boundary @xmath172 of an essential disk in @xmath160 is an element of @xmath173 .
then the collection of all essential simple closed curves bounding disks is a subset of @xmath173 .
it is known that the intersection function on @xmath174 defined a weak@xmath175-topology on @xmath174 , see @xcite .
then there is a topology defined on @xmath176 induced by the projection @xmath177 . under this topology ,
let @xmath178 be the closure of all essential simple closed curves in @xmath164 which bound disks in @xmath160 .
so is @xmath179 . by the symmetry of these two handlebodies @xmath87 and @xmath161
, there is an automorphism of @xmath180 such that @xmath181 .
@xmath182 is nowhere dense .
[ fact:3.1 ] the proof is based on and contained in the proof of theorem 1.2 @xcite . for the integrity of this paper ,
we use the theory of measured lamination space and rewrite it here . before proving fact
[ fact:3.1 ] , we introduce a definition as follows . for any essential simple closed curve @xmath183 bounding an essential disk in @xmath160 , there is an disk system @xmath184 in @xmath164 such that \(1 ) one of its vertices is @xmath4 ; \(2 ) all of its vertices are the isotopy classes of the boundaries of pairwise disjoint non - isotopic essential disks in @xmath160 ; \(3 ) it splits @xmath164 into a collection of pairs of pants .
all we need to prove is @xmath182 contains no open set in @xmath173 .
choosing an element @xmath185 represented by an essential non - separating simple closed curve in @xmath164 , by above argument , there is a disk system @xmath184 in @xmath164 . for any element @xmath186 represented by an essential simple
closed curve in @xmath164 , by lemma [ lem : bigon criterion ] , we can isotope @xmath5 such that the intersection number @xmath187 is minimal .
if @xmath5 intersects @xmath184 nonempty , then there is an wave @xmath188 corresponding to the outermost disk component in the complement of @xmath184 in @xmath164 . since the boundary of @xmath189 bounds no essential disk in @xmath160 , the wave @xmath188 is contained in a pair of pants bounded by the boundaries of essential disks .
if @xmath5 intersects @xmath184 empty , then @xmath190 . from penner and harer
@xcite , there is always a birecurrent maximal train track @xmath191 in @xmath164 such that it intersects all the wave like @xmath188 for the disk system @xmath184 . then there is a minimal measured lamination @xmath192 carried by @xmath191 intersecting all the wave like @xmath188 such that the complement of it in @xmath164 is a disk or a one - holed disk with a finite points removed from its boundary , where the one holed disk contains one boundary of @xmath164 .
then @xmath192 is not in @xmath182 because it intersects each element in @xmath182 non empty .
it is known that the collection of essential simple closed curves in @xmath164 is dense in @xmath193 .
then there is a sequence @xmath194 converging to @xmath192 in @xmath193 , where @xmath36 is represented by an essential simple closed curve .
hence there is a number @xmath195 such that @xmath196 intersects all the waves like @xmath188 for the disk system @xmath184 .
so there is a neighborhood @xmath197 of @xmath196 in @xmath193 disjoint from @xmath182 in @xmath193 .
now suppose that there is an open set @xmath198 .
then there is an automorphism @xmath199 , where @xmath200 , and a non separating essential curve @xmath201 bounding disk in @xmath160 such that @xmath202 and @xmath203 is an open neighborhood of @xmath4 in @xmath193
. for each essential simple closed curve @xmath204 which intersects @xmath4 nonempty , let @xmath205 be the dehn twist along @xmath4 in @xmath164 .
it is known that @xmath206 is closed to @xmath4 in @xmath193 .
then @xmath207 for @xmath44 large enough .
hence there is an open subset @xmath208 such that @xmath209 .
it means that @xmath210 .
then @xmath211 .
but since @xmath4 bounds an essential disk in @xmath160 , both of these two maps @xmath205 and @xmath212 map @xmath213 into @xmath182 . hence @xmath214 maps @xmath182 into @xmath182 . a contradiction
since the collection of those stable and unstable laminations of all pseudo anosov automorphisms in @xmath164 is dense in @xmath193 , there is a pseudo anosov map @xmath215 in @xmath164 such that the stable lamination are not in @xmath182 . by the proof of theorem 2.7
@xcite , if @xmath44 is large enough , then @xmath216 for any @xmath4 and @xmath5 bounding essential disks in @xmath160 and @xmath161 respectively . for constructing a non hyperbolic 3-manifold ,
we set @xmath217 be @xmath218 along @xmath166 and @xmath219 in @xmath217 . after pushing @xmath6 a little into the interior of @xmath12 ,
@xmath6 splits @xmath217 into a handlebody @xmath87 and a compression body @xmath220 , see figure [ fig : heegaardsurface2 ] . ]
note : @xmath6 is colored in green and @xmath164 is colored in red , where @xmath6 is parallel to the union of @xmath164 and an annulus @xmath221 .
a heegaard splitting is weakly reducible if there are a pair of essential disks from different sides of the heegaard surface so that their boundaries intersect empty . otherwise , the heegaard splitting is strongly irreducible .
the heegaard splitting @xmath13 is strongly irreducible .
[ fact3.2 ] suppose not .
then the heegaard splitting is weakly reducible .
so there are a pair of essential disks @xmath90 and @xmath222 so that @xmath223 . from the construction of @xmath217 , @xmath224 is incompressible in @xmath217 .
let @xmath225 be @xmath226 , where @xmath227 is a regular neighborhood of @xmath189 in @xmath164 .
after pushing the closure of @xmath228 a little into @xmath217 so that it is disjoint from @xmath229 , @xmath230 is turned into an embedded annulus @xmath231 . then @xmath6 is isotopic to @xmath232 .
it is not hard to see that every essential disk of @xmath87 ( resp .
@xmath220 ) has the property that its boundary cuts @xmath229 .
it means that @xmath229 is a hole for both of @xmath87 and @xmath220 . by the construction of @xmath13 ,
there is an essential disk in @xmath87 ( resp .
@xmath220 ) with its boundary in @xmath229 . then @xmath229 is a compressible hole for both of @xmath87 and @xmath220 . by lemma [ lem : compressible hole ] , for the essential disk @xmath93
, there is an essential disk @xmath233 such that \(1 ) @xmath234 ; \(2 ) there is one component @xmath235 such that @xmath236 ; similarly for the essential disk @xmath117 , there is an essential disk @xmath237 such that \(1 ) @xmath238 ; \(2 ) there is one component @xmath239 such that @xmath240 ; since @xmath223 , by lemma [ lem : subsurface projection ] , @xmath241 . by triangle inequality
, @xmath242 since @xmath229 is an essential subsurface of @xmath164 , every essential simple closed curve in @xmath229 is an essential simple closed curve in @xmath164 .
then @xmath243 since @xmath244 , it is not hard to see that @xmath140 is also an essential disk in @xmath160 .
so is the disk @xmath245 .
then the inequality above implies that @xmath246 for some pair of @xmath4 and @xmath5 bounding essential disks in @xmath160 and @xmath161 respectively .
it contradicts the assumption of @xmath217 .
it is known that every heegaard splitting of a boundary reducible 3-manifold is weakly reducible .
then the torus boundary @xmath247 of @xmath248 is incompressible .
let @xmath249 and @xmath250 be two solid torus .
let @xmath251 be an incompressible annulus so that the core circle of @xmath252 intersects the meridian circle in at least two points up to isotopy .
similarly , choose an annulus @xmath253 in the boundary @xmath250 .
after gluing @xmath249 and @xmath250 together along a homeomorphism between @xmath252 and @xmath253 , the resulted manifold @xmath254 is a small seifert 3-manifold with only one torus boundary @xmath255 , where @xmath255 is incompressible
. let @xmath256 be a homeomorphism such that @xmath257
. then @xmath258 is closed and @xmath255 is incompressible in @xmath259 .
let @xmath260 .
then @xmath261 splits @xmath259 into two 3-manifolds , denoted by @xmath262 and @xmath263 respectively . in this case
, @xmath262 is an amalgamation of @xmath160 and a solid torus @xmath249 along the annulus @xmath264 . then there are disjoint essential disks cutting @xmath262 into some 3-balls .
so @xmath262 is a genus @xmath215 handlebody .
similarly , @xmath263 is a genus @xmath215 handlebody too .
hence @xmath265 is a heegaard splitting of @xmath259 .
@xmath265 is a distance 2 genus @xmath215 heegaard splitting .
[ fact:3.4 ] since @xmath189 is essential in @xmath261 , every compression disk of @xmath164 in @xmath217 is an essential disk of @xmath261 .
since there are compression disks of @xmath164 in two sides , there are essential disks in @xmath262 and @xmath263 disjoint from @xmath266 in @xmath261 .
hence the heegaard splitting @xmath265 has the distance less than or equal to two .
suppose the heegaard splitting @xmath265 has distance less or equal to one .
then there are two essential disks @xmath267 and @xmath268 so that @xmath97 is disjoint from @xmath269 .
it is not hard to see @xmath164 is a compressible hole for both @xmath262 and @xmath263 . by lemma [ lem : compressible hole ] , for the essential disk @xmath93
, there is an essential disk @xmath270 such that \(1 ) @xmath271 ; \(2 ) there is one component @xmath272 such that @xmath273 ; since @xmath189 bounds an essential annulus in @xmath262 , after some isotopy , @xmath140 is a compression disk for @xmath164 in @xmath217 .
similarly for the essential disk @xmath117 , there is an essential disk @xmath274 such that \(1 ) @xmath275 ; \(2 ) there is one component @xmath239 such that @xmath276 ; \(3 ) @xmath245 is a compression disk for @xmath164 in @xmath217 . since @xmath223 , by lemma [ lem : subsurface projection ] , @xmath277 . by triangle inequality , @xmath243
it contradicts the assumption of @xmath217 .
by fact [ fact:3.4 ] , @xmath259 admits a distance 2 , genus @xmath215 heegaard splitting .
furthermore , it contains an essential torus . then there is a free abelian subgroup @xmath278 in its fundamental group .
so @xmath259 is not hyperbolic .
by the definition of heegaard distance , for a distance 2 , genus at least 2 heegaard splitting @xmath265 , there are three essential simple closed curves @xmath279 so that @xmath280 and @xmath4 ( resp .
@xmath5 ) bounds an essential disk in @xmath87 ( resp .
@xmath220 ) . set @xmath281 then it is a geodesic in @xmath17 and realizes the heegaard distance . for the heegaard splitting @xmath265 , there maybe many geodesics in @xmath282 realizing the heegaard distance .
moreover it is unknown that whether all the geodesics have a common vertex or not .
but for the non hyperbolic example @xmath283 in section [ sec3 ] , there is an essential non - separating simple closed curve @xmath284 such that for any pair of essential simple closed curves @xmath4 and @xmath5 disjoint from @xmath284 bounding essential disks in @xmath262 and @xmath263 respectively , @xmath285 .
let @xmath289 be the closure of the complement of @xmath284 in @xmath261 .
since @xmath284 bounds essential disks in neither @xmath262 nor @xmath263 and is non separating , @xmath289 is a compressible hole for both of these two disk complexes of @xmath262 and @xmath263 . by lemma [ lem : compressible hole ] , for @xmath290 ( resp .
@xmath291 ) , there is an essential disk @xmath93 ( @xmath117 ) so that then by lemma [ lem : subsurface projection ] , @xmath297 since each component of @xmath298 is disjoint from @xmath10 and @xmath11 and not isotopic to @xmath284 , by lemma [ lem : subsurface projection ] , @xmath299 then by triangle inequality , @xmath300 .
it contradicts the assumption of @xmath284 . for the closed orientable irreducible 3-manifold @xmath259
, geometrization conjecture says that there are finitely many essential tori so that after cutting @xmath259 along these tori , each piece is either hyperbolic , seifert or solvable . thus to understand the geometry of @xmath259 , the first thing is to check the possible embedded essential tori in it .
it is known that for any possible essential torus @xmath2 in @xmath259 , by schultens lemma @xcite , they can be isotoped to a general position that @xmath301 consists of essential simple closed curves in both @xmath261 and @xmath2 . after pushing the possible boundary parallel annulus to the other side ,
we assume that each component of @xmath302 ( resp .
@xmath303 ) is an essential annulus in @xmath262 ( resp .
@xmath263 ) . on one side ,
a boundary compression on an essential annulus produces an essential disk .
so for each component of @xmath304 , there is a geodesic containing it as its one vertex , which realizes heegaard distance . on the other side , by fact [ fact:4.1 ] , each geodesic realizing distance of heegaard splitting @xmath265 shares the same vertex @xmath284 .
hence each component of @xmath301 is isotopic to @xmath284 .
the reason for case ( 2 ) happening is that we choose a boundary compression disk for @xmath302 so that its interior intersects them empty .
then after doing boundary compression along this disk , the resulted disk @xmath93 is disjoint from all components of @xmath302 .
since these two boundaries of this annulus are isotopic , @xmath93 cuts out a solid torus from @xmath262 .
then all components of @xmath306 are contained in this solid torus after isotopy .
as all components of @xmath302 are pairwise disjoint , all these components of @xmath302 are parallel , i.e. , any two components of @xmath302 cuts out an i - bundle of annulus .
so are @xmath303
. since the union of all these annuli is @xmath2 , suppose not .
then there are at least two essential annulus in @xmath262 .
and there is an essential disk @xmath267 such that @xmath93 cuts out a solid torus containing @xmath302 . for @xmath303
, there is also an essential disk @xmath268 such that @xmath117 cuts out a solid torus containing @xmath303 , see figure [ fig : nested annuli ] . since the distance of heegaard splitting @xmath265 is 2 , @xmath307 .
it means that the red circles coincide with the blue circles in figure [ fig : nested annuli ] .
then the essential annulus bounded by the red circles in @xmath262 and the essential annulus bounded by the blue circles in @xmath263 are patched together in @xmath2 .
and the resulted manifold is a torus or a kleinian bottle . but
@xmath2 contains no kleinian bottle as its subset .
so the resulted manifold is a torus which is the @xmath2 , where it intersects @xmath261 in only two simple closed curves .
a contradiction .
since @xmath12 admits a distance 2 heegaard splitting , by haken s lemma , @xmath12 is irreducible .
it is known that every irreducible closed orientable 3-manifold @xmath12 either contains an essential torus or not .
in the later case , by geometrization conjecture , @xmath12 is either a small seifert 3-manifold or a hyperbolic 3-manifold .
suppose not .
then @xmath12 is a small seifert 3-manifold .
hence it has @xmath135 as its base surface with at most three exceptional fibers .
if @xmath12 has only one or two exceptional fibers , then @xmath12 is a lens space . but
all genus at least 2 heegaard splitting of a lens space is stabilized , reducible , i.e. , they all have distance 0 .
so @xmath12 contains three exceptional fibers .
moriah and schultens @xcite proved that each irreducible heegaard splitting of @xmath12 is either vertical or horizontal .
for the heegaard splitting @xmath13 , if it is vertical , then it has genus 2 . by corollary [ cor:1 ] , there are two essential disks @xmath140 and @xmath141 from two sides of @xmath6 and two non isotopy disjoint essential simple closed curves @xmath151 and @xmath152 so that both @xmath151 and @xmath152 are disjoint from @xmath149 and @xmath309 .
but under the condition that @xmath308 , by the proof of fact [ fact:4.1 ] , @xmath151 is isotopic to @xmath152 .
so it is impossible .
hence it is a horizontal heegaard splitting .
recall that for a horizontal heegaard splitting , @xmath310 , where @xmath129 is connected and @xmath311 , and @xmath312 . and @xmath313 $ ] ( resp .
@xmath220 is homeomorphic to @xmath314 $ ] ) . by lemma [ lem : seifert1 ]
, there is an essential simple closed curve @xmath315 so that @xmath316 $ ] and @xmath317 $ ] so that @xmath318 intersects @xmath319 in at most one point .
it is not hard to see that there are a pair of essential disks of two sides of @xmath6 so that their boundary disjoint from @xmath320 .
by the proof of fact [ fact:4.1 ] , @xmath321 is isotopic to @xmath1 .
let @xmath10 be an arc in @xmath52 disjoint from @xmath142 .
then there is an essential disk @xmath322 $ ] ( resp .
$ ] ) disjoint from @xmath320 .
thus @xmath324 ( resp .
@xmath325 ) . hence @xmath326 it contradict the choice of @xmath1 .
let @xmath224 be an annulus bounds by @xmath327 in @xmath6 and @xmath330 .
let @xmath217 be the amalgamation of these two solid tori along @xmath224 .
it is not hard to see that @xmath217 is a small seifert manifold with a disk as its base surface .
let @xmath331 . in the manifold @xmath254
, @xmath332 consists of two isotopic essential simple closed curves in @xmath333 .
and @xmath334 cuts @xmath254 into two handlebodies .
let @xmath166 be the union of @xmath334 and an annulus @xmath335 bounded by @xmath332 in @xmath336 , see figure [ fig : heegaard surface 3 ] .
after pushing @xmath166 a little into the interior of @xmath337 , @xmath166 cuts @xmath254 into a handlebody and a compression body .
then there is a heegaard splitting @xmath338 for @xmath254 .
similar to the proof of fact [ fact3.2 ] , @xmath166 is a strongly irreducible heegaard surface . from figure
[ fig : heegaard surface 3 ] , every essential disk in @xmath161 or @xmath339 with its boundary disjoint from @xmath332 is a compression disks of @xmath334 in @xmath87 or @xmath220 respectively . since @xmath12 is irreducible and @xmath2 is incompressible , @xmath254 is irreducible and boundary irreducible .
by fact [ fact:4.10 ] , @xmath12 contains only one essential torus @xmath2 up to isotopy .
then @xmath254 is atoroidal .
now suppose @xmath254 contains an essential annulus @xmath156 . by schultens lemma @xcite ,
@xmath340 are all essential simple closed curves in both @xmath156 and @xmath166 . after pushing all the boundary parallel annuli to the different side of @xmath166 , @xmath341 ( resp .
@xmath342 ) are essential annuli .
we say at least one component @xmath343 is not isotopic to @xmath1 . for if not
, then there is an i - bundle of @xmath333 containing @xmath156 after some isotopy , which means that @xmath156 is inessential .
then there is an essential disk @xmath344 ( resp .
@xmath345 ) disjoint from @xmath288 .
since @xmath166 cuts @xmath254 into a handlebody and a compression body , let @xmath161 be the handlebody . from figure
[ fig : heegaard surface 3 ] , @xmath166 is the union of @xmath334 and annulus @xmath335 , where @xmath161 is a disk sum of a handlebody and i - bundle of the annulus @xmath335 . then the boundary of each essential disk in @xmath161 intersects @xmath334 nonempty .
so @xmath334 is a compressible hole . by a similar argument
, @xmath334 is also a compressible hole for @xmath339 .
then by the proof of fact [ fact:4.1 ] , there is a pair of essential disks @xmath346 for @xmath140 and @xmath347 for @xmath245 so that @xmath97 and @xmath269 are both disjoint from @xmath332 and @xmath348 remember that each essential disk in @xmath161 or @xmath339 disjoint from @xmath332 is still an essential disk in @xmath87 or @xmath220 respectively and @xmath349 .
then it contradicts the choice of heegaard splitting @xmath13 .
the main result ( theorem 1.1 ) of johnson , minsky and moriah s paper @xcite says that for a heegaard splitting @xmath350 , if there is an essential subsurface @xmath351 such that the distance of these two projections of disk complexes @xmath352 and @xmath353 into @xmath52 , denoted by @xmath354 , satisfies that @xmath355 , then up to an ambient isotopy , any heegaard splitting of @xmath12 with genus less than or equal to @xmath356 has the subsurface @xmath52 in common . for the heegaard splitting in theorem [ thm:4.1 ] , if condition ( 3 ) is updated into @xmath357 , then any heegaard splitting @xmath358 of it with genus less than or equal to @xmath359 has @xmath164 in common up to an ambient isotopy . since @xmath189 bounds no disk in @xmath12 , @xmath360 is essential . by the calculation of the euler characteristic number , @xmath189 bounds an annulus @xmath224 in @xmath358 .
the proof of theorem [ thm1 ] implies that @xmath224 is parallel to an annulus in @xmath6 .
so the heegaard @xmath6 is the unique minimal heegaard surface up to isotopy .
let @xmath12 , @xmath350 and @xmath1 be the same as in theorem [ thm:4.1 ] . on one side , since @xmath1 is the middle vertex of a geodesic realizing heegaard distance , there are two essential compression disks for @xmath6 disjoint from @xmath1 from two sides of @xmath6 . on the other side ,
as @xmath1 is incompressible on both of these two sides of @xmath6 , there is an essential subsurface @xmath351 containing @xmath1 so that each essential simple closed curve in @xmath52 bounds no essential disk on either side of @xmath6 .
then there is a maximal ( defined later ) essential surface @xmath52 containing @xmath1 so that there is no essential , i.e. , incompressible and non peripheral , simple closed curve in @xmath52 so that it bounds an essential disk on either side of @xmath6 .
it is possible that there are many essential surfaces satisfying the property above .
thus we shall introduce some definitions for distinguishing all those surfaces .
we call two subsurface @xmath361 and @xmath362 are same if @xmath361 is isotopic to @xmath362 in @xmath6 . for a collection of different subsurfaces
, we define a partial order as follows . for any two essential subsurface @xmath361 and @xmath362 of @xmath6 , @xmath363 if @xmath361 can be isotopied into @xmath362 and @xmath364 .
since there is a lowest bound for all euler characteristic numbers of those subsurface , there is a maximal essential subsurface for any sequence of subsurfaces in order . for convenience , for each one of these maximal essential subsurfaces , we call it a domain of @xmath1 . throughout the proof of theorem [ thm:4.1 ] ,
the case that @xmath12 contains an essential torus means that ( 1 ) two copies of @xmath1 bounds an essential annulus in both @xmath87 and @xmath220 , namely , one domain of @xmath1 is an once punctured torus in @xmath6 ; ( 2 ) @xmath1 is not a co - core in either of these two sides of @xmath6 .
so to eliminate the possible essential tori in @xmath12 , it is sufficient to add some conditions related to these two cases ( 1 ) and ( 2 )
. let @xmath12 , @xmath13 and @xmath1 be the same as in theorem [ thm:4.1 ] .
if either each domain of @xmath1 has the euler characteristic number less than -1 or @xmath1 is a co - core for one side of @xmath6 , then @xmath12 is hyperbolic .
[ pro:2 ] the proof of theorem [ thm:4.1 ] suggests that @xmath2 intersects @xmath6 in two copies of @xmath1 .
it means that two copies of @xmath1 bounds an essential annulus @xmath156 ( resp .
@xmath157 ) in @xmath87 ( resp .
@xmath220 ) . then there are two one hole tori of @xmath6 containing @xmath1 in its interior from two sides of @xmath6 .
thus for either side of @xmath6 , there is one domain of @xmath1 with euler characteristic number equal to -1 .
suppose not .
without loss of genericity , @xmath1 is a co - core of the handlebody @xmath87 .
then there is an essential disk @xmath93 so that @xmath365 in one point
. then @xmath366 bounds an essential disk @xmath140 , which cuts @xmath87 into a solid torus @xmath305 and a small genus handlebody .
since the annulus @xmath156 is essential in @xmath87 and @xmath87 is irreducible , by standard innermost disk surgery , @xmath367 . then @xmath368 . as @xmath1 is a co - core ,
the disk @xmath93 intersects @xmath156 in one essential arc .
then there is a boundary compression disk @xmath369 for @xmath156 in @xmath305 so that after doing a boundary compression along @xmath370 , @xmath156 is changed into a trivial disk in @xmath87 . a contradiction .
let @xmath12 , @xmath13 and @xmath1 be the same as in proposition [ pro:2 ] .
then @xmath12 is hyperbolic if and only if either each domain of @xmath1 has the euler characteristic number less than -1 or @xmath1 is a co - core for one side of @xmath6 .
[ thm:5.3 ] for the forward direction .
suppose that ( 1 ) there are two domains @xmath361 and @xmath362 of @xmath1 , where both @xmath361 and @xmath362 are one hole disks and @xmath371 ( resp .
@xmath372 ) bounds an essential disk in @xmath87 ( resp .
@xmath220 ) , and ( 2 ) @xmath1 is not a co - core for both sides of @xmath6
. then @xmath371 ( resp .
@xmath373 ) cuts out a solid torus in @xmath87 ( resp .
@xmath220 ) containing @xmath1 .
let @xmath224 be closed regular neighborhood of @xmath1 . since @xmath1 is not a co - core for either side of @xmath6 , by the standard combinatorial techniques , @xmath374 bounds two essential annuli @xmath156 and @xmath157 in both of @xmath87 and @xmath220 respectively , see figure [ fig:5 ] . by figure [ fig:5 ] , @xmath156 ( resp .
@xmath157 ) cuts out a solid torus @xmath249 ( resp .
@xmath250 ) , where both of these two solid tori have the annulus @xmath224 as their common boundaries surface . since @xmath1 , the core curve of @xmath224 , is not a co - core of either of these two handlebodies @xmath87 and @xmath220
, @xmath377 is a small seifert space .
since @xmath224 is incompressible in both @xmath249 and @xmath250 , @xmath224 is incompressible in @xmath254 . for
if @xmath378 is compressible in @xmath254 , then the compression disk @xmath93 intersects @xmath224 nonempty up to isotopy . otherwise , either @xmath156 or @xmath157 is compressible in @xmath249 or @xmath250 respectively .
then there is an outermost disk @xmath369 for @xmath224 . without loss of genericity
, we assume that @xmath379 in @xmath87 .
it is not hard to see that @xmath370 is a boundary compression disk of @xmath156 in @xmath87 .
after dong boundary compression on @xmath156 along @xmath370 , @xmath156 is changed into a trivial disk in @xmath87 , which is impossible .
the conclusion of theorem [ thm2 ] says that although these 3-manifolds which admit distance two heegaard splittings are complicated , we can still get some kind of classification of them suggested by geometrization conjecture as we consider all those locally large distance two heegaard splittings .
d. cooper and m. scharlemann , _ the structure of a solvmanifold s heegaard splittings _ , proceedings of the gkova geometry - topology conference 1998 ( also in turkish journal of mathematics 23 ( 1999 ) ) 1 - 18 .
h. rubinstein and a. thompson , _ 3-manifolds with heegaard splittings of distance two _ , in proceedings of the conference , geometry and topology downunder , contemporary mathematics , 597 , ( 2013 ) , 341 - 345 .
ruifeng qiu + department of mathematics + shanghai key laboratory of pmmp + east china normal university + e - mail : [email protected] + yanqing zou + department of mathematics + dalian nationalities university , dalian minzu university .
+ e - mail : [email protected] , [email protected] . | from the view of heegaard splitting , it is known that if a closed orientable 3-manifold admits a distance at least three heegaard splitting , then it is hyperbolic . however , for a closed orientable 3-manifold admitting only distance at most two heegaard splittings , there are examples showing that it could be reducible , seifert , toroidal or hyperbolic . according to thurston s geometrization conjecture , the most important piece in eight geometries is hyperbolic .
so for a 3-manifold admitting a distance two heegaard splittings , it is critical to determine the hyperbolicity of it in studying heegaard splittings . inspired by the construction of hyperbolic 3-manifolds admitting distance two heegaard splittings in [ qiu , zou and guo , pacific j. math .
275 ( 2015 ) , no . 1 , 231 - 255 ]
, we introduce the definition of a locally large geodesic in curve complex and also a locally large distance two heegaard splitting .
then we prove that if a 3-manifold admits a locally large distance two heegaard splitting , then it is a hyperbolic manifold or an amalgamation of a hyperbolic manifold and a seifert manifold along an incompressible torus , while the example in section [ sec3 ] shows that there is a non hyperbolic 3-manifold in this case . after examining those non hyperbolic cases , we give a sufficient and necessary condition for a hyperbolic 3-manifold when it admits a locally large distance two heegaard splitting
. [ section ] [ thm]lemma [ thm]corollary [ thm]proposition [ thm]example [ section ] [ section ] [ clm]fact [ section ] [ rmk]note [ rmk]question : hyperbolic 3-manifold , heegaard distance , curve complex , locally large geodesic .
ams classification : 57m27 |
let @xmath2 be an admissible representation of @xmath12 , where @xmath13 is a local non - archimedean field .
let @xmath14 and @xmath15 be a pair of exceptional representations of a double cover @xmath16 of @xmath17 , in the sense of kazhdan and patterson @xcite .
we say that @xmath2 is distinguished if @xmath18 here @xmath19 is the representation contragradient to @xmath2 .
equivalently , the space of @xmath12-invariant trilinear forms on @xmath20 is nonzero .
this space first appeared in a global context .
let @xmath21 be a unitary cuspidal automorphic representation of @xmath22 , where @xmath23 is the adles ring of a global field .
assume that @xmath21 has a trivial central character .
bump and ginzburg @xcite proved that if the partial symmetric square @xmath10-function @xmath24 has a pole at @xmath25 , the following period integral is nonvanishing @xmath26 here @xmath27 is a subgroup of finite index in the center @xmath28 of @xmath22 ( @xmath29 when @xmath30 is odd ) ; @xmath31 is a cusp form in the space of @xmath21 ; @xmath32 and @xmath33 are automorphic forms in the space of a global exceptional representation of @xmath34 . for @xmath35 ,
an earlier work by patterson and piatetski - shapiro @xcite showed that a similar integral characterizes the pole at @xmath25 , for a global field of odd characteristic .
periods of automorphic forms are often related to poles of @xmath10-functions and to questions of functoriality .
ginzburg , jiang and soudry @xcite described such relations in a general setup and also considered several examples .
let @xmath36 be the odd general spin group of rank @xmath37 .
let @xmath38 be the eisenstein series corresponding to an element @xmath39 in the space of the representation of @xmath40 induced from @xmath41 ( @xmath42 , @xmath43 ) . the residual representation @xmath44 is the space spanned by the residues @xmath45 of @xmath38 at @xmath46 .
the following result formulated in @xcite ( theorem 3.3 ) was proved in a series of works ( @xcite ) : the following conditions are equivalent . 1 .
@xmath24 has a pole at @xmath25 .
[ item : period ] the period integral is nonzero .
[ item : residual ] the residual representation @xmath44 is nonzero .
@xmath21 is the langlands functorial transfer of an irreducible generic cuspidal automorphic representation of the split @xmath47 ( if @xmath30 is even ) or @xmath48 ( @xmath30 is odd ) .
note that this was stated in @xcite with @xmath49 replaced by @xmath50 ( see below ) . to prove that the nonvanishing of implies the nontriviality of @xmath44
, one can relate the period to the following co - period integral @xmath51 here @xmath32 and @xmath33 belong to the exceptional , or small , representation of bump , friedberg and ginzburg @xcite , whose analog for @xmath49 was described in @xcite .
this integral was studied in @xcite for @xmath50 ( extended in @xcite to @xmath49 ) and we proved the implication @xmath52 . in the setting of @xmath49 , one can also study the twisted symmetric square @xmath10-function .
the rankin - selberg integral representation for this function has recently been developed by takeda @xcite .
the global unfolding of @xmath53 ( in @xcite ) has a local counterpart .
assume that @xmath2 is a distinguished representation .
let @xmath54 , where @xmath55 is the siegel parabolic subgroup .
using an integral over @xmath56 , we show ( proposition [ proposition : upper heredity of dist 1 ] ) that for a certain pair @xmath57 and @xmath58 of exceptional representations of @xmath59 ( @xcite , see below ) , @xmath60 in other words , @xmath4 is a distinguished representation of @xmath61 . in fact , depending on the central character of @xmath2 , we may need to replace @xmath62 with @xmath63 where @xmath64 is a character of @xmath65 .
( see the proposition for details . )
the local problem is to prove that the langlands quotient @xmath66 of @xmath4 is also distinguished . in the present study we consider an irreducible unitary supercuspidal @xmath2 . in this case
@xmath4 is either irreducible and generic , or is of length two , has a unique irreducible generic subrepresentation and @xmath66 is non - generic .
here is our main result , implying that if @xmath2 is distinguished , so is @xmath66 . for more details
see corollary [ corollary : upper hered dist lq ] .
[ theorem : tensor of small is non generic ] the space of @xmath67 as a representation of @xmath61 does not afford a whittaker functional . a similar inflation " phenomena has already been observed by ginzburg , rallis and soudry @xcite ( theorem 2 ) .
assume that @xmath2 is a supercuspidal and self - dual representation , such that the exterior square @xmath10-function @xmath68 has a pole at @xmath11 .
this implies that @xmath2 has a shalika model and then according to jacquet and rallis @xcite , @xmath2 admits a ( nontrivial ) @xmath69 invariant functional . in turn , the representation parabolically induced from @xmath70 to @xmath71 has an @xmath72 invariant functional .
ginzburg , rallis and soudry @xcite ( theorems 16 - 17 ) showed that an irreducible generic representation of @xmath71 does not admit such a functional .
it follows that the @xmath72 functional factors through the langlands quotient .
this inflation was one of the ingredients used by lapid and mao for the proof of their conjecture on whittaker - fourier coefficients , in the case of the metaplectic group ( @xcite ) .
note that their conjecture actually applies to any quasi - split group as well as the metaplectic group .
our results here and their extension , in a forthcoming work , to an arbitrary irreducible generic distinguished representation , are expected to be used in a proof of this conjecture for even orthogonal groups .
here we have a similar relation between distinguished representations and the symmetric square @xmath10-function . [ theorem : supercuspidal dist gln and pole ] let @xmath2 be an irreducible unitary supercuspidal representation of @xmath73 . then @xmath2 is distinguished if and only if @xmath74 has a pole at @xmath11 .
refer to shahidi @xcite ( theorem 6.2 ) for a description of the poles of @xmath74 in this setting .
the proof of theorem [ theorem : tensor of small is non generic ] is essentially the local analog of the global computation of the co - period in @xcite .
write @xmath75 and let @xmath76 be the center of the unipotent radical @xmath77 .
the global unfolding argument involves a fourier expansion of @xmath32 along @xmath78 .
consider a non - generic character of @xmath78 , this means that its stabilizer in @xmath79 contains a unipotent radical @xmath80 of a parabolic subgroup of @xmath73 .
the corresponding fourier coefficient is constant on @xmath80 , then the cuspidality of @xmath21 is used to prove that the integral vanishes . when @xmath30 is odd ( and @xmath81 ) , all characters of @xmath78 are non - generic . in the even case
there is one generic orbit of characters .
its stabilizer is almost " a jacobi group , its reductive part is @xmath82 , where @xmath83 is a symplectic group in @xmath30 variables .
one might attempt to prove invariance of the product of fourier coefficients under @xmath84 , then use the fact that @xmath21 does not admit nontrivial symplectic periods ( @xcite ) . albeit the fourier coefficients do not enjoy this invariance , a certain convolution against weil theta functions , introduced by ideka @xcite , can be used instead .
the local argument involves the computation of the twisted jacquet modules @xmath85 of @xmath57 with respect to @xmath78 and a representative @xmath86 of some orbit of characters .
in contrast with the global setting , the generic character is the crux of the proof .
roughly , this is because @xmath85 is a tensor of an exceptional representation of @xmath87 and the jacquet module of an exceptional representation of @xmath88 along @xmath89 and a generic character .
when the character is generic , the twisted jacquet module is ( by restriction ) a representation of a jacobi group .
the local theory of smooth representations of jacobi groups @xcite describes such a representation as a tensor @xmath90 , where @xmath91 is the weil representation .
the heisenberg group acts trivially on the space of @xmath92 , while the action of the reductive part separates into an action on the space of @xmath92 , and one on the space of @xmath91 .
we prove that @xmath92 is a trivial representation of @xmath93 .
[ theorem : small rep is weakly minimal ] assume @xmath30 is even and let @xmath94 be a generic character of @xmath95 . as a jacobi representation
@xmath96 is the direct sum of ( possibly infinitely many ) copies of @xmath91 .
note that the action of the @xmath97 part of the stabilizer on @xmath96 is given simply by the central character of @xmath57 .
this result underlies theorem [ theorem : tensor of small is non generic ] .
furthermore , it implies the following multiplicity property .
assume that @xmath2 is irreducible and tempered .
we prove that @xmath2 is distinguished if and only if @xmath4 is .
moreover , the dimensions of and are equal . see proposition [ proposition : one - dim ] for the precise statement .
note that kable @xcite conjectured , and under a certain homogeneity assumption proved ( @xcite corollary 6.1 ) , that enjoys multiplicity one .
these results motivate the introduction of exceptional models " over @xmath73 and @xmath49 .
theorem [ theorem : small rep is weakly minimal ] may have additional applications .
explicit descriptions of jacquet modules of exceptional representations have had numerous applications ( see below ) .
note that when @xmath35 , @xmath55 is the heisenberg parabolic subgroup in the notation of gan and savin @xcite . in their terminology ,
theorem [ theorem : small rep is weakly minimal ] shows that @xmath57 is weakly minimal " . in this case
it is the minimal representation and @xmath98 ( @xcite section 3 ) .
bump , friedberg and ginzburg @xcite constructed the small representation @xmath99 for the special odd orthogonal group .
this is a representation of a
double cover " @xmath100 , obtained by restricting the @xmath101-fold cover of @xmath102 of matsumoto @xcite . for the low rank cases @xmath103 , it is the minimal representation .
in fact for @xmath104 , this representation was already developed by roskies @xcite , sabourin @xcite and torasso @xcite . for @xmath105
, there is no minimal representation for a group of type @xmath106 ( @xcite ) .
bump , friedberg and ginzburg @xcite showed that when @xmath105 , @xmath99 is attached to one of the possible coadjoint orbits , which is smallest next to the minimal one .
this translates into the vanishing of a large class of fourier coefficients , called generic in @xcite ( see also @xcite ) .
locally , this means that a large class of twisted jacquet modules vanish .
the representation @xmath99 was used by these authors to construct a lift with certain functorial properties , between covers of orthogonal groups @xcite .
the representation @xmath107 was used to construct an integral representation ( @xcite ) .
there is a technical issue when working with @xmath100 : the underlying field must contain all @xmath101 @xmath101-th roots of unity .
this can be remedied using @xmath49 .
indeed , one obtains a nontrivial double cover of @xmath61 by restricting the @xmath108-fold cover of @xmath109 of matsumoto @xcite .
the theory of bump , friedberg and ginzburg @xcite can be extended to @xmath59 , mainly because both groups are of type @xmath106 and in particular , the unipotent subgroups are isomorphic .
the details were carried out in @xcite .
our results here are stated for @xmath49 , but apply similarly to @xmath50 and @xmath99 ( see section [ section : relevance to so(2n+1 ) ] ) .
minimal representations have been studied and used extensively , by numerous authors .
knowledge of fourier coefficients , or jacquet modules , has proved very useful for applications @xcite .
they have played a fundamental role in the theta correspondence , the descent method and rankin - selberg integrals .
see , for example @xcite .
we mention that bump and ginzburg @xcite also developed a local theory , where they considered a similar space of equivariant trilinear forms , except that @xmath15 was replaced by a certain induced representation . for @xmath104 , savin @xcite determined the dimension of for an arbitrary irreducible quotient of a principal series representation .
he also conjectured ( and proved for @xmath104 ) , that the class of distinguished spherical representations of @xmath12 is precisely those representations , which are lifts from a certain prescribed classical group .
kable @xcite proved that these lifts are distinguished , the other direction was proved in @xcite .
the group @xmath49 has been the focus of study of a few recent works , among which are the works of asgari @xcite on local @xmath10-functions , asgari and shahidi @xcite on functoriality and hundley and sayag @xcite on the descent construction .
the rest of this work is organized as follows . in section [ section : preliminaries ]
we provide notation and definitions .
in particular we describe the construction of exceptional representations .
section [ section : heisenberg jacquet modules ] contains the proof of theorem [ theorem : small rep is weakly minimal ] .
theorems [ theorem : tensor of small is non generic ] and [ theorem : supercuspidal dist gln and pole ] are proved in section [ section : distinguished representations ] .
section [ section : relevance to so(2n+1 ) ] provides the formulation of our results for @xmath50 .
i wish to express my gratitude to erez lapid for suggesting this project to me .
i would like to thank eitan sayag , for explaining to me how to use his results with offen ( @xcite , proposition 1 ) .
lastly , i thank jim cogdell for his kind encouragement and useful remarks .
let @xmath13 be a local non - archimedean field of characteristic different from @xmath108 .
for an integer @xmath110 , let @xmath111 be the subgroup of @xmath112-th roots of unity in @xmath13 .
set @xmath113 .
we usually fix a nontrivial additive character @xmath114 of @xmath13 . then the normalized weil factor ( @xcite section 14 )
is denoted by @xmath115 ( @xmath116 is @xmath117 of @xcite , @xmath118 ) .
the hilbert symbol of order @xmath112 is @xmath119 .
if @xmath5 is a group , @xmath120 denotes its center . for @xmath121 and @xmath122 , @xmath123 and @xmath124 .
hereby we omit references to the field , e.g. , @xmath125 .
fix the borel subgroup @xmath127 of upper triangular matrices , where @xmath128 is the diagonal torus . for
any @xmath129 such that @xmath130 , denote by @xmath131 the maximal parabolic subgroup whose levi part is isomorphic to @xmath132 .
its unipotent radical is @xmath133 .
mirabolic " subgroup @xmath134 is the subgroup of @xmath73 of matrices with the last row @xmath135 .
let @xmath136 be the identity matrix of @xmath73 and @xmath137 be the matrix with @xmath138 on the anti - diagonal and @xmath139 elsewhere . for @xmath140 , @xmath141 is the transpose of @xmath142 .
we will use the metaplectic double cover @xmath143 of @xmath73 , as constructed by kazhdan and patterson @xcite .
let @xmath144 be the double cover of @xmath145 of matsumoto @xcite and let @xmath146 be the corresponding cocycle of banks , levi and sepanski @xcite ( section 3 ) .
we define a @xmath108-cocycle @xmath147 of @xmath126 via @xmath148 this cocycle is related to the cocycle @xmath149 of @xmath73 defined in @xcite by @xmath150 in particular @xmath151 .
we start by defining the special odd orthogonal group @xmath153 select its borel subgroup @xmath154 .
let @xmath155 be the simple split simply - connected algebraic group of type @xmath156 .
it is the algebraic double cover of @xmath50 .
we will take the borel subgroup , which is the preimage of @xmath157 .
the set of simple roots of @xmath155 is @xmath158 , where @xmath159 for @xmath160 and @xmath161 .
the group @xmath36 is an @xmath13-split connected reductive algebraic group , which can be defined using a based root datum as in @xcite .
it is also embedded in @xmath162 as the levi part of the parabolic subgroup corresponding to @xmath163 ( see @xcite ) .
we adapt this identification , which is more natural for the purpose of cover groups , because we will obtain a cover of @xmath49 by restricting the cover of @xmath164 .
let @xmath165 be the modulus character of a parabolic subgroup @xmath166 . in the degenerate case @xmath167 .
the borel subgroup of @xmath49 is denoted @xmath168 , where @xmath169 is the unipotent radical ( the rank of the torus is @xmath37 ) . for @xmath170 ,
denote by @xmath171 the standard maximal parabolic subgroup of @xmath172 with @xmath173 .
this isomorphism is not canonical .
we describe the choice used in @xcite , which is convenient for certain computations ( see below ) . the derived group @xmath174 of @xmath175 is generated by the root subgroups of @xmath176 .
let @xmath177 be the standard cocharacters of @xmath178 and map @xmath179 for @xmath180 .
regarding @xmath181 , the set @xmath182 identifies @xmath183 and if @xmath184 are the characters of @xmath185 , define @xmath186 and for @xmath187 , @xmath188 . the projection @xmath189 is an isomorphism between unipotent subgroups , hence we can identify unipotent subgroups of @xmath172 with those of @xmath50 .
we use the character @xmath190 to define a canonical " character @xmath191 of @xmath172 .
namely @xmath191 is the extension of @xmath192 to @xmath172 ( the only other choice would be to use @xmath190 ) .
the aforementioned embedding of @xmath193 in @xmath194 has a few properties , suitable for computations . to compute the conjugation of @xmath195 on @xmath196
, we can simply look at this action in @xmath50 , where @xmath197 takes the form @xmath198 .
the image of @xmath97 is @xmath199 .
the restriction of @xmath191 to @xmath200 is @xmath201 .
let @xmath202 be the double cover of @xmath164 , constructed by matsumoto @xcite using @xmath203 as the steinberg symbol . restricting the cover to @xmath172
, we obtain the exact sequence @xmath204 then @xmath205 is a double cover of @xmath172 . for a subset @xmath206 , @xmath207 .
let @xmath208 be the block - compatible section constructed by banks , levi and sepanski @xcite and @xmath209 be the corresponding cocycle . denote the restriction of @xmath209 to @xmath210 by @xmath211 .
in @xcite we proved that @xmath211 satisfies the following block compatibility property : if @xmath212 , @xmath213 we also mention that @xmath214 .
let @xmath5 be an @xmath215-group ( @xcite 1.1 ) .
representations of @xmath5 will be complex and smooth . for a representation @xmath21 of @xmath5 , @xmath216 is the representation contragradient to @xmath21 .
we say that @xmath21 is glued from representations @xmath217 , if @xmath21 has a filtration , whose quotients ( which may be isomorphic or zero ) are , after a permutation , @xmath217 .
regular induction is denoted @xmath218 and @xmath219 is the compact induction .
induction is not normalized .
let @xmath21 be as above and let @xmath220 be a unipotent subgroup , exhausted by its compact subgroups ( always the case here ) .
let @xmath114 be a character of @xmath221 .
the jacquet module of @xmath21 with respect to @xmath221 and @xmath114 is denoted @xmath222 .
it is a representation of the stabilizer of @xmath114 ( and normalizer of @xmath221 ) .
the action is not normalized .
we have an exact sequence @xmath223 the kernel @xmath224 can be characterized by the jacquet - langlands lemma : [ lemma : jacquet kernel as integral](see e.g. @xcite 2.33 ) a vector @xmath225 in the space of @xmath21 belongs to @xmath224 if and only if @xmath226 for some compact subgroup @xmath227 . when @xmath228 , we simply write @xmath229 and @xmath230 .
assume that @xmath231 is a given @xmath112-th cover of @xmath5 .
let @xmath232 be a faithful character .
a representation @xmath21 of @xmath231 is called @xmath233-genuine if it restricts to @xmath233 on @xmath111 . when @xmath234 , such a representation is simply called genuine .
let @xmath235 be a section and assume @xmath21 and @xmath236 are representations of @xmath231 , such that @xmath21 is @xmath233-genuine and @xmath236 is @xmath237-genuine .
then @xmath238 ( outer tensor product ) is a representation of @xmath5 via @xmath239 .
the actual choice of @xmath240 does not matter , whence we omit it .
levi subgroups of classical groups are direct products .
the tensor of representations of the direct factors is usually used , to describe their representations . in passing to a cover group , these factors
do not necessarily commute and then the tensor construction fails .
except for the case of @xmath241 , the preimages of @xmath200 and @xmath181 in @xmath242 do not commute .
the same phenomena occurs in @xmath73 .
the following discussion describes a replacement for the usual tensor product . for more details
see @xcite .
the metaplectic tensor product in the context of @xmath126 has been studied by various authors @xcite .
our definitions were motivated by the construction of kable @xcite ( see remark [ remark : diff kable tensor ] below ) . for any @xmath243 ,
put @xmath244 .
the subgroup @xmath245 is normal in @xmath1 and the quotient is a finite abelian group . if @xmath246 is a representation of @xmath247 , let @xmath248 .
assume @xmath249 .
according to , the preimages of @xmath250 and @xmath251 are commuting in the cover . then if @xmath39 and @xmath21 are genuine representations of @xmath252 and @xmath253 , the representation @xmath254 is a genuine representation of @xmath255 put @xmath256 and define @xmath257 when @xmath241 , @xmath143 and @xmath258 are commuting , then the metaplectic tensor is defined as usual . [
remark : diff kable tensor ] the arguments in @xcite do not readily extend to @xmath49 , mainly because @xmath259 for all @xmath30 , and then @xmath199 does not play a role similar to that of @xmath260 in the cover .
we will use mackey theory to relate this induced representation to @xmath39 and @xmath21 .
we reproduce the following result from @xcite , which mimics @xcite ( theorem 3.1 ) in our context .
[ lemma : induced representation composition factors ] the representation @xmath261 is a direct sum of @xmath262 $ ] copies of @xmath263 since @xmath264 is normal of finite index in @xmath265 and @xmath266 modulo @xmath264 is abelian , @xmath267 here @xmath268 ranges over the finite set of characters of @xmath269 , which are trivial on @xmath270 . by if @xmath271 and @xmath272 , @xmath273 .
hence @xmath274 and the result follows .
we introduce the weil representation , which plays an important role in this work .
let @xmath275 and @xmath276 be the symplectic bilinear form on @xmath277 defined by @xmath278 , where @xmath279 and @xmath225 are regarded as rows .
let @xmath280 be the @xmath281-dimensional heisenberg group , with the group operation given by @xmath282 where @xmath283 and @xmath284 .
let @xmath285 be the symplectic group defined with respect to @xmath276 , i.e. , the group of @xmath286 such that @xmath287 for all @xmath288 .
let @xmath289 be the metaplectic double cover of @xmath285 , realized using the normalized rao cocycle @xcite .
the group @xmath285 acts on @xmath290 via @xmath291 .
fix a nontrivial additive character @xmath114 of @xmath13 .
let @xmath91 be the weil representation of @xmath292 , realized on the space @xmath293 of schwartz - bruhat functions on @xmath294 .
recall the following formulas for @xmath91 ( see @xcite ) : for @xmath295 , @xmath296 let @xmath297 and @xmath298 . since @xmath221 normalizes ( in fact , commutes with ) @xmath299 , @xmath300 is a @xmath221-module .
we will use the following simple observation .
[ claim : jacquet of weil 1 dim ] the vector spaces @xmath300 and @xmath301 are one dimensional . according to lemma [ lemma : jacquet kernel as integral ] and , the space of @xmath302 is @xmath303 .
hence @xmath300 is one dimensional .
then @xmath304 is nonzero , because by lemma [ lemma : jacquet kernel as integral ] and , a function @xmath305 such that @xmath306 does not belong to @xmath307 .
we will also encounter the tensor @xmath308 of two weil representations .
this is a representation of @xmath290 , trivial on @xmath309 , and a representation of @xmath285 . regarding it as a representation of @xmath310
, we can compute its twisted jacquet modules .
the group @xmath285 acts transitively on the nontrivial characters of @xmath310 , hence we can consider only one nontrivial character . [ claim : jacquet modules along hn of double weil reps ] let @xmath311 be a character of @xmath290 , which is trivial on @xmath309 . 1 . if @xmath312 , @xmath313 is the trivial one - dimensional representation of @xmath285 .
2 . if @xmath314 , @xmath313 is the trivial one - dimensional representation of @xmath315 .
first assume @xmath312 .
equality implies that elements @xmath316 in the space of @xmath308 , such that the supports of @xmath240 and @xmath317 ( as functions in @xmath318 ) are different , vanish under the jacquet module along @xmath299 ( use lemma [ lemma : jacquet kernel as integral ] ) .
hence the space of @xmath319 is isomorphic to @xmath318 . since the action of @xmath309 is trivial on @xmath308 , @xmath320 .
now @xmath321 is a normal subgroup of @xmath290 , whence @xmath322 is a quotient of @xmath323 .
it then follows from that the action of @xmath221 is trivial on @xmath323 and in particular , on @xmath322 .
the latter is a representation of @xmath285 , and because @xmath285 is generated ( as an abstract group ) by the subgroups @xmath324 , @xmath325 ( it is enough to take weyl elements @xmath326 ) , @xmath322 must be a trivial representation of @xmath285 .
moreover @xmath322 is one - dimensional .
this can be seen as follows .
replace a function @xmath327 with its fourier transform @xmath328 .
this changes the module structure , but the action of @xmath285 remains trivial .
the action of @xmath329 is now given by @xmath330 ( instead of ) .
next apply lemma [ lemma : jacquet kernel as integral ] , the space of @xmath331 is @xmath332 .
we conclude that @xmath322 is the trivial one - dimensional representation of @xmath285 . now consider the case of the nontrivial @xmath311 , given in the statement of the claim .
since @xmath333 , @xmath313 is a quotient of @xmath323 and in particular , a trivial representation of @xmath221 . in coordinates , @xmath334 ( in particular it stabilizes @xmath311 . )
we see that @xmath335 and using conjugations of @xmath221 by elements from @xmath336 , it follows that @xmath313 is the trivial representation of @xmath315 .
to show this is a one - dimensional representation , argue as above using the fourier transform , the space of @xmath337 is @xmath338 .
we describe the exceptional representations that appear in this work .
kazhdan and patterson @xcite introduced and studied these representations for @xmath73 . for @xmath49 these are essentially the small representations of bump , friedberg and ginzburg @xcite , who developed them using a cover of @xmath50 ( see section [ section : relevance to so(2n+1 ) ] ) . their construction was extended to @xmath49 in @xcite .
let @xmath5 be either @xmath73 or @xmath49 .
let @xmath339 be the borel subgroup of @xmath5 , @xmath340 be the maximal torus and @xmath341 be the subset of simple roots .
for @xmath342 , if @xmath343 is a long root and @xmath81 , put @xmath344 , otherwise @xmath345 .
denote by @xmath346 the coroot of @xmath343 .
let @xmath246 be a genuine character of @xmath347 .
we say that @xmath246 is exceptional if @xmath348 the character @xmath246 corresponds to a genuine irreducible representation @xmath349 of @xmath350 ( when @xmath81 , @xmath350 is a @xmath108-step nilpotent subgroup ) .
since @xmath246 is exceptional , the representation @xmath351 has a unique irreducible quotient @xmath57 , called an exceptional representation .
note that @xmath57 is admissible .
occasionally , we use the notation @xmath352 to record the group .
we appeal to the following explicit description of @xmath353 : @xmath354 with @xmath355 , @xmath356 , where @xmath357 is @xmath138 if @xmath30 is odd , otherwise @xmath358 ; @xmath359 , and if @xmath360 , @xmath361 ( see @xcite p. 57 and
@xcite section 2.1.6 ) .
note that @xmath362 .
furthermore , the cocycle @xmath363 satisfies @xmath364 .
the exceptional characters can be parameterized in the following way .
start by fixing a character @xmath365 of @xmath366 , trivial on @xmath367 .
this does not determine @xmath365 uniquely .
for @xmath73 , take @xmath368 . for @xmath49 take @xmath365 whose restriction to @xmath369 is @xmath370 .
given that @xmath365 is trivial on @xmath371 , this determines @xmath365 uniquely for @xmath49 ( see @xcite section 2.3.3 for an explicit formula ) .
now for any given @xmath246 , there is a character @xmath372 of @xmath65 ( called determinantal character " in @xcite ) and a nontrivial additive character @xmath114 of @xmath13 such that @xmath373 the corresponding exceptional representation will be denoted @xmath374 .
[ claim : taking characters out of exceptional ] we have @xmath375 , where on the right - hand side we pull back @xmath372 to a non - genuine character of @xmath231 via @xmath376 . additionally , if @xmath377 is another additive character of @xmath13 , @xmath378 for some square trivial character @xmath64 of @xmath65 .
the first assertion is clear .
write @xmath379 for some @xmath380
. then @xmath381 , where @xmath382 .
since @xmath383 is trivial on @xmath384 and @xmath385 ( the last equality is trivial if @xmath30 is even , because then @xmath386 ) , we obtain @xmath387 hence @xmath388 .
exceptional representations have a useful inductive property .
let @xmath57 be an exceptional representation of @xmath389 .
following the arguments of bump , friedberg and ginzburg @xcite ( theorem 2.3 ) , we computed @xmath390 ( @xcite ) . for @xmath249 , @xmath391 if @xmath392 ( resp .
@xmath393 ) is odd , the exceptional representation of @xmath394 ( resp .
@xmath395 ) is unique only up to varying the character @xmath114 , or multiplying @xmath372 by a square trivial character of @xmath65 . in any case
, the space on the right - hand side of is unique , because by claim [ claim : taking characters out of exceptional ] , the exceptional representation obtained by such a change to @xmath114 or @xmath372 , has the same restriction to @xmath396 ( resp .
@xmath397 ) as the original representation .
if @xmath241 , @xmath398 in @xcite we did not find the precise exceptional representations appearing on the right - hand side of , but this is simple to obtain : since @xmath359 , there is an exceptional character @xmath399 of @xmath400 such that @xmath401 , where @xmath402 and @xmath403 are the longest weyl elements in the weyl groups of @xmath49 and @xmath73 ( see @xcite claim 2.18 for details ) .
it remains to write @xmath399 using .
the reason for the imprecise result when @xmath404 is the lack of a definition for a tensor product .
these results are sufficient for our applications . for @xmath73 ,
kable proved a result similar to for all standard unipotent radicals , with the tensor replaced by his metaplectic tensor ( @xcite theorem 5.1 ( 4 ) ) .
exceptional representations enjoy the vanishing of a large class of twisted jacquet modules .
the following result is the extension of theorem 2.6 of @xcite and proposition 3 of @xcite to @xmath49 ( this extension appeared in @xcite ) .
it was used in @xcite ( for @xmath50 ) to deduce all vanishing properties .
the unipotent radical @xmath405 is abelian .
a character @xmath406 of @xmath405 takes the form @xmath407 where @xmath408 is a column .
the length of @xmath409 is defined to be @xmath410 .
when @xmath114 is fixed and clear from the context , we also refer to @xmath411 as the length of @xmath406 .
[ theorem : main vanishing jacquet of small rep ] if the length of @xmath406 is nonzero , @xmath412 .
see @xcite ( lemma 2.25 ) for the details .
in this section we describe the twisted jacquet modules of @xmath414 with respect to the center of @xmath77 .
these modules appear in a filtration of @xmath57 as a @xmath415-module and will be used in section [ section : distinguished representations ] .
the group @xmath73 , embedded in @xmath79 , acts on the set of characters of @xmath76 with finitely many orbits .
let @xmath114 be a nontrivial additive character of @xmath13 .
for any @xmath416 , define a character of @xmath78 by @xmath417 ( @xmath418 is regarded as a @xmath419 unipotent matrix in @xmath50 ) .
the stabilizer of @xmath86 in @xmath420 is @xmath421 , with @xmath422 here @xmath423 is the symplectic group in @xmath424 variables , corresponding to a symplectic bilinear form defined according to @xmath425 .
then @xmath85 is a representation of @xmath426 . regarding @xmath427 and @xmath285 as subgroups of @xmath428 , @xmath429 ( @xmath430
was given in section [ subsection : gl_n and its cover ] . )
we turn to the proof of theorem [ theorem : small rep is weakly minimal ] .
namely , for @xmath275 , @xmath431 is the direct sum of copies of the weil representation @xmath91 .
here is an outline of the proof .
the theory of smooth representations of jacobi groups ( @xcite ) implies that any such representation , with a central character @xmath114 , takes the form @xmath90 , where the heisenberg group acts trivially on the space of @xmath92 , and the action of the symplectic group separates into an action on the space of @xmath92 , and one on the space of @xmath91 .
the vanishing properties of @xmath57 - theorem [ theorem : main vanishing jacquet of small rep ] , will show that @xmath92 is trivial .
put @xmath432 .
the image of @xmath97 in @xmath49 is @xmath199 .
hence @xmath433 is the direct product of a jacobi group and @xmath97 .
moreover , @xmath434 ( see section [ subsection : metaplectic gspin ] ) and because @xmath57 is an irreducible representation , @xmath258 acts by the central character of @xmath57 .
therefore in the proof we ignore the @xmath97 part of @xmath428 .
we may replace @xmath86 by any character of @xmath78 in the same @xmath73-orbit , since the jacquet module will be isomorphic . for convenience ,
we redefine @xmath435 .
we use the notation of section [ subsection : the weil representation ] .
the stabilizer @xmath428 is now the symplectic group defined with respect to the form @xmath276 .
the cover @xmath421 is a nontrivial double cover .
we have an epimorphism @xmath436 : @xmath437 the kernel of @xmath438 is contained in the kernel of @xmath86 .
therefore we can regard @xmath85 as a genuine representation of @xmath439 .
as such , it is isomorphic to @xmath90 ( see e.g. @xcite p. 28 ) , where @xmath92 is a non - genuine representation on a space @xmath440 , and the action is given by @xmath441 we must show that @xmath92 is trivial .
consider the subgroup @xmath442 since @xmath285 is generated by the conjugates @xmath443 where @xmath444 , it is enough to prove invariance under @xmath445 .
that is , we show @xmath446 a character of @xmath445 takes the form @xmath447 for some @xmath448 ( on the left - hand side @xmath449 is regarded as a matrix , on the right - hand side as an element of @xmath13 ) .
the action of the torus of @xmath285 on the nontrivial characters of @xmath445 has finitely many orbits , namely the different square classes in @xmath65 .
each of these orbits is open .
therefore the kernel of the jacquet functor @xmath450 is filtered by representations induced from @xmath451 , where @xmath343 ranges over the square classes ( @xcite 5.9 - 5.12 ) .
hence follows if we prove @xmath452 for any @xmath453 .
consider the subgroup @xmath454 ( a direct product ) .
the epimorphism @xmath438 splits over @xmath455 , hence there is a subgroup @xmath456 isomorphic to @xmath455 .
in fact , @xmath457 the pullback of @xmath458 to @xmath221 is @xmath459 .
observe that @xmath460 .
indeed , @xmath461 , which is a quotient of @xmath462 . since for @xmath463 , @xmath464 , any extension of @xmath465 to a character of @xmath405 is a character of nonzero length .
thus theorem [ theorem : main vanishing jacquet of small rep ] yields @xmath466 whence @xmath460 .
therefore by lemma [ lemma : jacquet kernel as integral ] , for any @xmath467 there is a compact @xmath468 ( @xmath469 , see section [ subsection : the weil representation ] ) such that @xmath470 according to claim [ claim : jacquet of weil 1 dim ] and lemma [ lemma : jacquet kernel as integral ] there is @xmath471 such that for all compact subgroups @xmath472 and @xmath473 , @xmath474 take @xmath475 .
we will show that for large enough @xmath476 and @xmath477 , @xmath478 this along with imply that @xmath479 belongs to the space of @xmath480 . plugging into and
changing variables leads to @xmath481 we will show that the inner @xmath482-integration vanishes for all @xmath483 . fix @xmath484 . since @xmath485 is trivial , this inner integration equals @xmath486 again resorting to claim [ claim : jacquet of weil 1 dim ] , @xmath487 where @xmath488 and @xmath489 belongs to the space of @xmath302 .
since @xmath484 varies in a compact subgroup , there is a large enough @xmath477 such that @xmath490 furthermore implies that for large @xmath476 and @xmath477 , @xmath491 and then for any @xmath492 , @xmath493 combining - we conclude that for sufficiently large @xmath477 and @xmath476 , the inner @xmath482-integration vanishes . note the order of selecting the compact subgroups : first , choose @xmath477 and @xmath476 which ensure , they depend only on @xmath479 and @xmath240 . then , increase @xmath477 to have , it will depend on @xmath240 and @xmath476 .
this completes the proof of and thereby .
we conclude that @xmath92 is trivial . in the more general case , for arbitrary @xmath392 , we are less precise . [
proposition : jacquet module c and k ] there are exceptional representations @xmath494 and @xmath495 such that @xmath85 is embedded in a finite direct sum of copies of the representation @xmath496 here @xmath497 is regarded as a representation of @xmath426 by extending it trivially on @xmath498 .
the representations @xmath494 and @xmath495 are related to @xmath57 via and . if @xmath499 , the embedding is in fact an isomorphism and there is only one summand , otherwise there are @xmath500 $ ] summands . for @xmath501 , we have @xmath502 and @xmath499 , whence @xmath503 and the result follows immediately from .
assume @xmath81 .
further assume @xmath504 , otherwise there is nothing to prove . the main part of the proof is to show that @xmath498 acts trivially on @xmath85 .
of course , this holds for @xmath505 .
let @xmath506 and note that @xmath507 .
clearly @xmath508 .
the following claims imply that the action of @xmath498 is trivial .
[ claim : v_k acts trivially]@xmath509 .
[ claim : z_n-2k , k acts trivially]@xmath510 . before proving the claims ,
let us deduce the proposition .
clearly @xmath511 .
assume @xmath512 .
then by , @xmath513 according to lemma [ lemma : induced representation composition factors ] , @xmath514 is the finite direct sum of @xmath500 $ ] copies of @xmath515 let @xmath516 be the stabilizer of @xmath86 in @xmath517 , when @xmath86 is regarded as a character of @xmath89 and @xmath518 .
the double coset space @xmath519 is trivial .
then by virtue of the geometric lemma of bernstein and zelevinsky @xcite ( theorem 5.2 ) , @xmath520 equality implies the subgroups @xmath521 and @xmath289 commute .
therefore @xmath522 thus @xmath523 is the direct sum of representations @xmath497 .
the proposition follows from this .
note that for @xmath499 , @xmath524 and we can apply instead of , then becomes @xmath525 .
a character @xmath526 of @xmath527 extending @xmath86 is defined by its restriction to the nontrivial coordinates on the @xmath528-th column of @xmath529 .
we call @xmath526 nontrivial if this restriction is nontrivial .
we prove that the jacquet module of @xmath85 with respect to @xmath530 and @xmath531 vanishes for any nontrivial @xmath531 .
the group @xmath532 acts transitively on these characters .
therefore , it is enough to show @xmath533 this follows immediately from theorem [ theorem : main vanishing jacquet of small rep ] , because any character of @xmath405 extending @xmath534 has a nonzero length .
thus @xmath509 . for @xmath499
there is nothing to prove ( @xmath535 .
assume @xmath512 .
the claim follows once we show that for any nontrivial character @xmath311 of @xmath430 , @xmath536 the group @xmath430 is abelian and @xmath537 acts on the characters of @xmath430 .
write an element @xmath538 in the form @xmath539 we may assume that @xmath311 does not depend on the coordinates of @xmath540 and @xmath541 , and depends on @xmath542 . for simplicity ,
also assume @xmath543 .
we use the local analog of exchanging roots " , proved by ginzburg , rallis and soudry @xcite ( lemma 2.2 ) .
( for the global setting see @xcite . )
let @xmath544 be the subgroup of elements @xmath545 and @xmath546 be the subgroup consisting of elements @xmath547 .
clearly @xmath548 ( a direct product ) .
also consider the subgroup @xmath549 in general , if @xmath21 is a smooth representation of @xmath550 , then by @xcite ( lemma 2.2 ) , as @xmath551-modules @xmath552 indeed , it is simple to check that the list of properties stated in the lemma are satisfied in this setting ( in the notation of @xcite , @xmath553 , @xmath554 and @xmath555 ) .
lemma 2.2 of @xcite was stated for unipotent subgroups of symplectic groups , but the arguments are general and hold in our setting .
see also section 2.3 of @xcite .
it follows that as @xmath551-modules @xmath556 conjugating the right - hand side by a weyl element of @xmath49 , whose action on @xmath557 is given by the action of @xmath558 we obtain that @xmath559 is a quotient of @xmath560 here @xmath561 , @xmath562 is a certain subgroup of @xmath563 ( obtained from the conjugation of @xmath78 ) and @xmath564 .
note that @xmath565 corresponds to @xmath311 and the coordinate @xmath542 while @xmath566 corresponds to the character @xmath86 of @xmath78 and the @xmath567-th coordinate of @xmath568 .
the character @xmath566 is nontrivial on @xmath562 . finally , by proposition 4 of bump , friedberg and
ginzburg @xcite ( which is easily extended to @xmath49 , given the analog of theorem [ theorem : main vanishing jacquet of small rep ] in @xcite ) , @xmath569 is a quotient of @xmath570 ( this is valid for @xmath571 , here @xmath572 whence @xmath573 ) .
hence , @xmath563 acts trivially on @xmath569 and therefore @xmath574 and follows .
let @xmath5 be either @xmath73 or @xmath49 .
let @xmath2 be an admissible representation of @xmath5 with a central character @xmath575 .
assume that @xmath57 and @xmath58 are a pair of exceptional representations of @xmath231 .
we say that @xmath2 is @xmath576-distinguished if @xmath577 the following result describes the upper hereditary property of a distinguished representation of @xmath73 , when induced to a representation of @xmath49 . following the notation of section [ subsection : the exceptional representations ]
, we denote the exceptional representation of @xmath231 corresponding to @xmath372 and @xmath114 by @xmath578 . for any representation @xmath579 of @xmath73 , @xmath42 and a character @xmath311 of @xmath65 ,
one forms a representation @xmath580 of @xmath79 .
put @xmath581 [ proposition : upper heredity of dist 1 ] let @xmath2 be a @xmath582-distinguished representation of @xmath73 and set @xmath583 . then @xmath584 is a @xmath585-distinguished representation of @xmath49 . by definition
the space @xmath586 of @xmath73-equivariant trilinear forms on @xmath587 is nonzero .
therefore @xmath588 according to , @xmath589 applying frobenius reciprocity we see that @xmath590 is a subrepresentation of @xmath591 a similar result holds for @xmath592 .
we define @xmath593 and prove it is nonzero .
let @xmath240 belong to the space @xmath594 , regarded as an element of , and similarly let @xmath317 belong to the space of @xmath592 .
also take @xmath595 in the space of @xmath584 .
now if @xmath596 belongs to , @xmath597 thus the following integral is ( formally ) well defined ( see e.g. @xcite 1.21 ) , @xmath598 it is absolutely convergent according to the iwasawa decomposition . since @xmath340 satisfies the necessary equivariance properties , it remains to show @xmath599 .
assume @xmath600 for suitable data .
take @xmath595 supported on @xmath601 , for a small compact open neighborhood @xmath602 of the identity in @xmath49 , and such that @xmath603 we may assume @xmath604 ( because @xmath605 is irreducible ) and @xmath606 . using the iwasawa decomposition and @xmath607 then yields @xmath608 since @xmath10 is invariant with respect to @xmath609 , taking a sufficiently small @xmath602 ( with respect to @xmath240 and @xmath317 ) , the @xmath610-integration reduces to a nonzero constant multiple of @xmath611 , which is nonzero .
we conclude that @xmath584 is @xmath585-distinguished .
let @xmath2 be a representation of @xmath5 as above .
write @xmath612 and @xmath613 . since @xmath375
, we may assume @xmath614 , perhaps twisting @xmath2 by a character .
for simplicity , we then say that @xmath2 is @xmath615-distinguished .
if @xmath30 is even , the characters @xmath114 and @xmath616 can be ignored , because @xmath617 does not depend on @xmath114 .
if @xmath30 is odd and @xmath2 is @xmath618-distinguished , then for any @xmath114 there is @xmath616 such that @xmath2 is @xmath615-distinguished .
indeed , write @xmath379 for some @xmath380 and put @xmath619 , then by claim [ claim : taking characters out of exceptional ] and its proof , @xmath620 . in light of these observations
, we say that @xmath2 is distinguished if for any @xmath114 there is @xmath616 such that @xmath2 is @xmath615-distinguished .
proposition [ proposition : upper heredity of dist 1 ] implies , [ corollary : upper hered dist ] let @xmath2 be a distinguished representation of @xmath73
. then @xmath621 is distinguished .
now we prove theorem [ theorem : tensor of small is non generic ] .
namely , for any pair @xmath57 and @xmath58 of exceptional representations of @xmath622 , and a generic character @xmath114 of @xmath557 , @xmath623 we consider the filtrations of @xmath57 and @xmath58 corresponding to the jacquet functor along @xmath76 . the kernel of this functor is glued from representations induced from the jacquet modules described in section [ section : heisenberg jacquet modules ] . taking the twisted jacquet functor along @xmath557 truncates some of these quotients and ,
essentially , reduces the problem to a representation induced from @xmath624 .
theorem [ theorem : small rep is weakly minimal ] then enables us to further reduce the problem , to the vanishing of @xmath625 , which essentially follows from the results of offen and sayag on klyachko models ( @xcite , see also @xcite ) . by an analog of the geometric lemma of bernstein and zelevinsky ( @xcite theorem 5.2 and @xcite 5.9 - 5.12 ) , as a @xmath415-module , @xmath57
is glued from @xmath626 ( see section [ section : heisenberg jacquet modules ] for the notation . )
then as a @xmath55-module @xmath67 is glued from @xmath627 we prove that the jacquet functor with respect to @xmath557 and @xmath114 vanishes on each of these representations .
since functions in @xmath628 are compactly supported modulo @xmath629 , and @xmath78 is normal in @xmath55 , by lemma [ lemma : jacquet kernel as integral ] , @xmath630 to see this consider @xmath595 in the space of @xmath628 and @xmath631 in the space of @xmath632 , and look at the jacquet - langlands integral @xmath633 where @xmath634 is a compact subgroup . since @xmath635 is a non - genuine representation of @xmath421 , we can replace the representation on the right - hand side of with @xmath636 define @xmath637 with respect to @xmath494 as in proposition [ proposition : jacquet module c and k ] and similarly , define @xmath638 with respect to @xmath639 . according to the proposition
, @xmath85 is embedded in a finite direct sum of representations @xmath497 , which are trivial on @xmath498 .
put @xmath640 it is enough to prove that for all @xmath641 , @xmath642 this holds for @xmath499 , simply because @xmath77 is normal in @xmath55 , @xmath0 is trivial on @xmath77 while @xmath114 is not .
the case of @xmath432 is handled by the following claim , whose proof is deferred to below . [ claim : induced from double weil claim 0 ] equality holds for @xmath432 .
lastly , assume @xmath572 .
set @xmath643 . by transitivity of induction @xmath644
the representation @xmath645 is trivial on @xmath498 . by virtue of the geometric lemma of bernstein and zelevinsky ( @xcite theorem 5.2 ) ,
the representation on the right - hand side is glued from jacquet modules of @xmath645 .
note that in general , the quotients are representations induced from jacquet modules , here the induction is trivial because the stabilizer of the character @xmath114 of @xmath557 is @xmath646 .
let @xmath647 be a set of representatives to the double cosets @xmath648 .
that is , @xmath649 .
we can take the elements @xmath326 to be weyl elements of @xmath73 .
when @xmath650 , the quotient corresponding to @xmath326 vanishes .
this implies there is only one quotient , corresponding to @xmath651 , namely @xmath652 here @xmath653 is some modulus character , hereby ignored , and @xmath654 . as a @xmath97-module ,
this representation is isomorphic to @xmath655 here @xmath114 is regarded as a generic character of @xmath656 and @xmath657 . since the case @xmath432 implies @xmath658 equality follows . for @xmath432 , @xmath659 .
now we apply theorem [ theorem : small rep is weakly minimal ] . for simplicity of computations
, we can replace @xmath86 with the character defined in the proof of the theorem , then use the epimorphism @xmath436 given there . by theorem [ theorem :
small rep is weakly minimal ] , @xmath85 is isomorphic to the direct sum of copies of @xmath91 .
pull @xmath91 back to a representation of @xmath426 .
note that the @xmath97 part of @xmath428 was ignored in the proof of theorem [ theorem : small rep is weakly minimal ] , since it acts by a character , so we can ignore this here as well .
equality will follow from @xmath660 we need some notation .
put @xmath661 note that @xmath662 ( in fact , @xmath663 but @xmath97 was ignored ) , this isomorphism restricts to an isomorphism @xmath664 . in this manner
@xmath114 is also a character of @xmath80 , @xmath665 .
since @xmath666 is trivial on @xmath78 , we can regard it as a representation of @xmath667 and if @xmath668 , @xmath669 as @xmath5-modules .
we apply the theory of @xmath215-sheafs of bernstein and zelevinsky ( @xcite 1.13 and section 6 ) . in the following ,
we freely use their terminology and definitions .
let @xmath670 be the @xmath215-sheaf corresponding to @xmath21 ( @xcite 2.23 ) .
the group @xmath557 acts on @xmath455 by @xmath671 and on @xmath672 by @xmath673 .
an @xmath557-invariant @xmath672-distribution on @xmath455 is an element of @xmath674 . since @xmath674 is the algebraic dual of @xmath675 , to prove we will show that there are no nonzero @xmath557-invariant @xmath672-distributions on @xmath455 .
the action of @xmath557 on @xmath455 is constructive ( @xcite theorem a ) .
if @xmath676 , let @xmath677 be the stabilizer of @xmath678 in @xmath557 .
the orbit of @xmath678 is @xmath679 .
the mapping @xmath680 induces a homeomorphism @xmath681 ( @xcite 1.6 ) .
the restriction of @xmath672 to the orbit of @xmath678 ( this restriction is an @xmath215-sheaf , because the action is constructive ) is isomorphic to @xmath682 , where @xmath683 is the representation of @xmath684 acting in the space of @xmath666 by @xmath685 . by virtue of theorem 6.9 of @xcite , to show there are no nonzero @xmath557-invariant @xmath672-distributions on @xmath455 , it is enough to prove that for each representative @xmath678 , @xmath686 we can take representatives @xmath687 .
then @xmath688 , where @xmath689 .
in addition , because @xmath134 fixes @xmath690 , @xmath691 . hence @xmath692 note that @xmath693 for the nontrivial @xmath311 given in claim [ claim : jacquet modules along hn of double weil reps ] . according to that claim
@xmath694 is the trivial one - dimensional representation of @xmath695 . since @xmath696 for any @xmath697 , @xmath698 is trivial on @xmath699 ( the epimorphism @xmath438 is easily seen to be harmless here ) .
but offen and sayag ( @xcite proposition 2 , we use @xmath700 with @xmath701 and @xmath702 , in their notation ) proved that @xmath703 for any @xmath704 .
this implies @xmath705 , as required .
[ corollary : upper hered dist lq ] let @xmath2 be an irreducible unitary supercuspidal @xmath582-distinguished representation of @xmath73 . then the langlands quotient of @xmath706 is @xmath585-distinguished . in particular , if @xmath2 is distinguished , so is the langlands quotient of @xmath621 .
according to proposition [ proposition : upper heredity of dist 1 ] , @xmath707 theorem [ theorem : tensor of small is non generic ] implies @xmath706 is reducible ( because @xmath2 is generic ) , then by casselman and shahidi @xcite ( theorem 1 ) , the langlands quotient @xmath708 is non - generic , and the unique irreducible subspace of @xmath706 is generic .
also note that the length of @xmath706 is @xmath108 ( because it is reducible and @xmath2 is supercuspidal , see @xcite 2.8 ) .
now the result follows from theorem [ theorem : tensor of small is non generic ] and the left exactness of the @xmath709 functor . as a corollary
, we can now prove theorem [ theorem : supercuspidal dist gln and pole ] .
namely , for an irreducible unitary supercuspidal @xmath2 , being distinguished is equivalent to the occurrence of a pole at @xmath11 of @xmath74 .
if @xmath2 is distinguished , as in the proof of corollary [ corollary : upper hered dist lq ] we see that @xmath621 is reducible , then according to casselman and shahidi @xcite ( proposition 5.3 , their conjecture 1.1 was proved for @xmath49 in @xcite ) @xmath74 has a pole at @xmath11 . in the other direction ,
assume @xmath74 has a pole at @xmath11 . as an application of the descent method of ginzburg , rallis and soudry ( see e.g. , @xcite ) , one can globalize @xmath2 to a cuspidal automorphic representation @xmath21 of @xmath22 , such that @xmath24 has a pole at @xmath25 ( see the appendix of @xcite ) .
therefore the nonvanishing of implies @xmath2 is distinguished ( @xcite theorem 7.6 ) .
if @xmath2 is an irreducible unitary supercuspidal distinguished representation , it must be self - dual .
[ remark : applicability to twisted ] the analogous result for an irreducible supercuspidal @xmath582-distinguished @xmath2 should also hold .
one needs to verify the applicability of the globalization argument ( see @xcite ) .
given exceptional representations @xmath57 and @xmath58 of @xmath231 ( as in the beginning of this section ) , we can consider the space of @xmath67 as a model for representations of @xmath5 .
we refer to the dimension of @xmath710 as the multiplicity of @xmath2 .
the next proposition relates the multiplicities of @xmath2 and @xmath584 . in the case of @xmath73 ,
kable @xcite conjectured that the multiplicity of an irreducible representation is at most one .
he proved this for @xmath711 , and for arbitrary @xmath30 under a certain homogeneity condition ( @xcite corollary 6.1 ) .
it is reasonable to believe multiplicity one also holds in the context of @xmath49 ( see @xcite remark 4.2 ) . [ proposition : one - dim ]
let @xmath2 be an irreducible tempered representation of @xmath73 , put @xmath583 and assume @xmath712 .
then @xmath713 in particular , the representation @xmath2 is @xmath582-distinguished if and only if @xmath584 is @xmath585-distinguished .
[ remark : uniqueness current state ] if @xmath2 is @xmath582-distinguished , in particular @xmath714 for all @xmath715 ( @xmath575 - the central character of @xmath2 ) .
it follows that @xmath712 , because a tempered representation is unitary . in this case by @xmath2 enjoys multiplicity one if and only if @xmath584 does .
set @xmath716 , @xmath717 and similarly @xmath15 and @xmath58 ( with @xmath718 and @xmath616 ) . for brevity ,
put @xmath719 .
applying frobenius reciprocity , @xmath720 using the notation of section [ section : heisenberg jacquet modules ] , the representation @xmath721 is filtered by representations @xmath722 the representation @xmath723 is a quotient of @xmath721 and the kernel of the mapping @xmath724 is filtered by @xmath725 with @xmath726 .
regard @xmath725 as a representation of @xmath73 ( by restriction from @xmath79 ) . according to proposition [ proposition : jacquet module c and k ] , @xmath727 .
then @xmath728 thus will follow , once we prove that for @xmath512 , @xmath729 if @xmath30 is even , we claim the following . [
claim : induced from double weil claim 1 ] for any irreducible generic representation @xmath579 of @xmath73 , @xmath730 . in particular equality holds for @xmath432 .
the proof will be given below .
assume @xmath572 .
the representation @xmath85 is embedded in a finite direct sum of representations @xmath497 ( proposition [ proposition : jacquet module c and k ] ) .
the representation @xmath637 is semisimple , and by theorem [ theorem : small rep is weakly minimal ] the representation @xmath731 is also semisimple . hence @xmath635 is embedded in a semisimple representation , which is the finite direct sum of @xmath732 ( @xmath733 ) .
therefore , using the exactness of induction and jacquet functors , @xmath725 is a quotient of @xmath734 .
we conclude that we may replace @xmath725 in with @xmath735 .
an application of @xcite ( theorem 5.2 ) yields @xmath736 where @xmath737 note that the relevant double coset space ( for @xcite theorem 5.2 ) is @xmath738 , there is only one representative to consider . hence to deduce , it is enough to prove @xmath739 the left - hand side equals @xmath740 here @xmath741 ( the normalized jacquet functor ) . it suffices to show that the last space vanishes , when @xmath742 is replaced by any of its irreducible subquotients .
let @xmath743 be one such subquotient . then vanishes if either @xmath744 are zero . as explained in the proof of proposition 4.1 of lapid and mao @xcite , since @xmath2 is tempered , either @xmath745 is generic , or the central character @xmath746 of @xmath747 satisfies @xmath748 for some @xmath749
. in the former case vanishes by claim [ claim : induced from double weil claim 1 ] . in the latter case vanishes . indeed
, @xmath750 acts on @xmath751 by @xmath752 and since @xmath712 , this action is unitary on @xmath753 , but @xmath746 is not unitary . put @xmath432 and @xmath754 . we need to prove @xmath755 we will show that @xmath756 has a filtration , whose quotients are all isomorphic to @xmath757
. then since @xmath579 is irreducible and generic , @xmath758 ( @xcite proposition 1 , take @xmath759 ) and the result follows .
we turn to prove the filtration of @xmath756 .
as in the proof of claim [ claim : induced from double weil claim 0 ] , theorem [ theorem : small rep is weakly minimal ] implies that @xmath756 is filtered by copies of the representation @xmath760 .
we prove @xmath761 ( recall that the left - hand side is regarded as a representation of @xmath73 . )
since @xmath77 is normal in @xmath55 , lemma [ lemma : jacquet kernel as integral ] implies @xmath762 in more detail , if @xmath595 belongs to the space of @xmath763 , the jacquet - langlands integral takes the form @xmath764 for a compact subgroup @xmath765 .
because @xmath595 is compactly supported modulo @xmath766 , this integral vanishes for all @xmath767 if and only if @xmath768 belongs to the space of @xmath769 for all @xmath142 .
it remains to use the exactness of induction .
according to claim [ claim : jacquet modules along hn of double weil reps ] , @xmath322 is the trivial one - dimensional representation of @xmath285 and holds
. we can now improve corollary [ corollary : upper hered dist ] for tempered representations : [ corollary : upper hered dist for tempered ] let @xmath2 be an irreducible tempered representation of @xmath73
. then @xmath2 is distinguished if and only if @xmath621 is distinguished .
bump , friedberg and ginzburg @xcite constructed and studied the small representations for @xmath50 . in this section we briefly recall their results and formulate our results for @xmath50 .
the cover @xmath770 was obtained by restricting the @xmath101-fold cover of @xmath771 of matsumoto @xcite .
it is a double cover " in the sense that the square of the cocycle is trivial on the kernel of the spinor norm .
we use the same notation of @xmath49 , e.g. , @xmath772 ( see section [ subsection : gspin ] ) , @xmath773 is the diagonal torus and @xmath774 is a standard maximal parabolic subgroup . if @xmath775 , @xmath776 is embedded in @xmath194 as @xmath777 .
the block compatibility formula now reads ( see @xcite ( 2.20 ) ) @xmath778 the benefit of this cover is that the preimages of @xmath200 and @xmath779 commute , thus the tensor can be used to describe representations of levi subgroups .
restriction of the cover to @xmath200 is a double cover .
the small representation @xmath780 is unique ; it is the representation of @xmath770 corresponding to the exceptional character @xmath781 according to @xcite ( theorem 2.3 ) , @xmath782 , where @xmath783 was explicitly given , and by @xcite ( theorem 2.6 ) and @xcite ( proposition 3 ) , @xmath412 if the length of @xmath406 is nonzero ( using the notation of section [ subsection : the exceptional representations ] ) .
theorems [ theorem : tensor of small is non generic ] and [ theorem : small rep is weakly minimal ] remain valid as stated .
proposition [ proposition : jacquet module c and k ] now takes the form @xmath784 here @xmath494 is uniquely determined .
indeed , this equality replaces because @xmath785 is a tensor of representations . regarding proposition [ proposition : upper heredity of dist 1 ] ,
assume @xmath2 is @xmath615-distinguished .
twisting @xmath2 by some square trivial character , we obtain a @xmath786-distinguished representation . then , perhaps after using another twist of @xmath2 , it becomes @xmath787-distinguished for the exceptional representation @xmath788 such that @xmath789 .
now the proof of the proposition proceeds as in the case of @xmath49 .
so , in this case one must start with a @xmath787-distinguished representation of @xmath73 , in order to obtain a distinguished representation @xmath621 of @xmath50 .
corollary [ corollary : upper hered dist lq ] is applicable for @xmath2 such that proposition [ proposition : upper heredity of dist 1 ] is valid .
proposition [ proposition : one - dim ] remains valid .
d. jiang and d. soudry .
generic representations and local langlands reciprocity law for @xmath797-adic @xmath805 . in h.
hida , d. ramakrishnan , and f. shahidi , editors , _ contributions to automorphic forms , geometry , and number theory : a volume in honor of joseph shalika _ ,
pages 457519 , baltimore , md , 2004 .
johns hopkins univ . press .
d. kazhdan . the minimal representation of @xmath809 . in _
operator algebras , unitary representations , enveloping algebras , and invariant theory ( paris , 1989 ) _ , volume 92 of _ progr .
_ , pages 125158 .
birkhuser boston , boston , ma , 1990 .
d. kazhdan and g. savin . the smallest representation of simply laced groups . in _ festschrift in honor of i. i. piatetski - shapiro on the occasion of his sixtieth birthday , part i ( ramat aviv , 1989 ) _ , volume 2 of _ israel math
_ , pages 209223 .
weizmann , jerusalem , 1990 . g. mui .
reducibility of standard representations for classical @xmath797-adic groups . in
_ functional analysis viii ( dubrovnik , 2003 ) _ , volume 47 of _ various publ .
_ , pages 132145 .
aarhus , aarhus , 2004 .
p. perrin .
reprsentations de schrdinger , indice de maslov et groupe metaplectique . in _ non commutative harmonic analysis and lie groups ( proc .
marseille - luminy 1980 ) _ , volume 880 of _ lecture notes in math .
_ , pages 370407 .
springer - verlag , 1981 .
d. prasad .
weil representation , howe duality , and the theta correspondence . in _ theta functions : from the classical to the modern _ , volume 1 of _ crm proc .
lecture notes _ , pages 105127 .
soc . , providence , ri , 1993 .
d. soudry .
on langlands functoriality from classical groups to @xmath821 . in j.
tilouine , h. carayol , m. harris , and m .- f .
vignras , editors , _ formes automorphes ( i ) ; actes du semestre du ceb , printemps 2000 _ , volume 298 of _ astrisque _ , pages 335390 , 2005 .
d. a. vogan , jr . singular unitary representations . in _ noncommutative harmonic analysis and lie groups ( marseille , 1980 ) _ , volume 880 of _ lecture notes in math .
_ , pages 506535 .
springer , berlin , 1981 . | let @xmath0 be a representation of a group @xmath1 .
we say that a representation @xmath2 is @xmath3-distinguished , if it is a quotient of @xmath0 .
it is natural to ask whether this notion inflates " to larger groups , in the sense that a representation @xmath4 induced from @xmath2 and @xmath1 to a group @xmath5 , is @xmath6-distinguished .
we study representations distinguished by theta representations : @xmath7 , @xmath0 is a pair of the exceptional representations of kazhdan and patterson , @xmath8 and @xmath9 is a pair of the small representations of bump , friedberg and ginzburg .
we prove a rodier - type hereditary property : a tempered representation @xmath2 is distinguished if and only if @xmath4 is distinguished , and the multiplicity in each model is the same .
if @xmath2 is supercuspidal and distinguished , we prove that the langlands quotient of @xmath4 is distinguished . as a corollary , we characterize supercuspidal distinguished representations , in terms of the pole of the local symmetric square @xmath10-function at @xmath11 . |
to address the effects of anharmonicity , we consider the model potential @xmath95 for a 1d optical lattice , which is a good approximation near the trap center .
this potential introduces three leading - order anharmonic corrections to ( 3 ) , which are the quartic potentials @xmath96 for the initial and likewise for the final lattice , where @xmath97 .
far from the axial trap center , a finite rayleigh length introduces additional ( e.g. , cubic ) leading - order corrections . for a transition between a pair of known trap states in the initial and final lattices
we may approximate the light shift of each @xmath98 by its first - order perturbation , @xmath99 where primes denote final - lattice values .
these shifts introduce the corrections @xmath100 to the total light shift @xmath51 of ( 7 ) , where the brackets denote a thermal average over the allowed pairs of initial and final trap states .
for axial sideband ( sb ) transitions , the effects of @xmath66-dependent excitation rates must be included in this average , as described in the next section . in the lamb - dicke ( ld ) and resolved - sideband regimes with suppressed transverse sb transitions , the trap state pairs satisfy @xmath101 where the integer @xmath102 is introduced to distinguish between axial carrier ( @xmath103 ) and first - order axial sb transitions ( @xmath104 ) .
surprisingly , carrier transitions are nearly unchanged by the leading - order corrections ( [ vxx][vrr ] ) , because the first - order light shifts ( [ deij],[wij ] ) are zero : @xmath105 before evaluating these quantities explicitly in the next section , we can explain this general result as follows .
first , note that any form for @xmath106 must be proportional to the polarizability @xmath107 .
thus , any anharmonic corrections to ( 3 ) , such as ( [ vxx][vrr ] ) , must also be proportional to @xmath107 .
next , note that the expectations @xmath108 and @xmath109 for harmonic oscillator states .
the matrix elements in ( [ deij ] ) for the @xmath110 of ( [ vxx]-[vrr ] ) are therefore independent of @xmath111 and @xmath107 , respectively , and must be equal , thus producing no differential shift .
this general insensitivity of the carrier light shift to quartic anharmonicities also applies to the model potentials @xmath112 and @xmath113 for 2d and 3d optical lattices . for axial
sb transitions with @xmath114 , the total light shift @xmath51 of ( 7 ) due to the harmonic potential ( 3 ) must be modified as follows .
first , there is an `` axial - sb shift '' from the final lattice , @xmath115 which must be added as a fourth part to @xmath51 .
next , if @xmath116 , the populations of the initial lattice with @xmath117 will not participate in the transition , so the expectation @xmath118 must be computed accordingly .
this asymmetry also leads to the relation ( 2 ) between temperature and the ratio of sb areas . additionally , for sb transitions the excitation rates depend on @xmath66 .
the expectation @xmath118 is no longer solely thermal , but must account for this inhomogeneous excitation by weighting each value of @xmath66 with the square of its rabi frequency for the transition , @xmath119 where the ld parameter @xmath120 and the axial wavenumber @xmath121 . as before , we assume the trap states are approximately orthonormal , @xmath122 , which may need to be modified if @xmath1 is far from unity . after normalizing the probabilities for each @xmath66 , the weighted expectations are @xmath123/2 & d = 0 \\ \text{coth}\left [ \delta_x/2 \right]+1/2 & d = -1 \\ \text{coth}\left [ \delta_x/2 \right]-1/2 & d = + 1 \end{cases}\end{aligned}\ ] ] where as before @xmath124 . hence ,
although the form of @xmath125 given by ( 10 ) , @xmath126 will be unchanged for sbs , the value of @xmath53 will depend on @xmath102 following ( [ nxwithd ] ) .
note that the form and value of @xmath127 given by ( 11 ) , @xmath128 is the same for sbs as for carriers .
the anharmonic corrections ( [ vxx][vrr ] ) are important for sbs , unlike carriers , especially for state - insensitive ` magic ' traps with @xmath14 . the contributions ( [ wij ] ) to the shift @xmath51 from these corrections are @xmath129 for both carrier and sb transitions , as derived below .
importantly , note that all these contributions are zero for carriers as argued above , since @xmath130 if @xmath103 .
the expression ( [ wxx ] ) for @xmath131 follows from the expectation @xmath132 @xcite , which gives the matrix elements @xmath133 and from noting that @xmath134 and @xmath135 .
the expression for @xmath136 follows from expectations of the form @xmath137 , which give the matrix elements @xmath138 and from noting that @xmath139 .
the shift @xmath140 of ( [ wrr ] ) is then zero because the condition ( [ ortho ] ) includes only radial carrier transitions .
that is , @xmath141 of ( [ vrr ] ) contributes no shift for the same reasons that @xmath142 if @xmath143 . to demonstrate the effects of anharmonicity on the lineshape of sb transitions ,
let us treat the case of a magic lattice with @xmath14 where there is no thermal broadening of the carrier . in this case
, broadening comes only from the thermal distribution of the anharmonic shifts @xmath144 and @xmath145 . using ( [ vxrelement ] ) with ( [ deij ] )
, we find @xmath146 similarly , using ( [ vxxelement ] ) with ( [ deij ] ) and ( [ ws ] ) , @xmath147/(8 u_0 ) \nonumber \\ & \approx -d ( n_x + 1/2 ) ( \hbar \omega_x)^2 / ( 4 u_0),\end{aligned}\ ] ] where the second line follows from neglecting a constant offset ( half the lattice - photon recoil energy ) that contributes no broadening .
note that ( [ wxx],[wxr ] ) are related to ( [ dexx],[dexr ] ) via ( [ wij ] ) with @xmath14 .
together , the shifts ( [ dexr],[dexx ] ) lead to similar lineshapes as derived for carrier transitions . as before ,
we introduce a function to replace boltzmann exponents , @xmath148 \geq 0.\end{aligned}\ ] ] the discrete step size of @xmath149 is @xmath124 and of @xmath150 is @xmath151 .
since the probability distribution for @xmath67 is unchanged , the probability for the discrete variable @xmath145 follows from the @xmath152 of ( 14 ) , @xmath153 likewise , for @xmath154 the probability @xmath155 of ( 12 ) for @xmath66 is unchanged .
however , we now need to account for inhomogeneous excitation , so the probability for the discrete variable @xmath144 is @xmath156 for @xmath157 , using ( [ rabi ] ) and normalizing , this evaluates to @xmath158 likewise , for @xmath159 where only @xmath160 participate , @xmath161 in the continuum limit , these probabilities simplify to @xmath162 = \lim_{\delta_i \rightarrow 0 } \frac{p_{xi}(\delta e_{xi})}{\delta_i } = v \ , e^{-v}\end{aligned}\ ] ] for both @xmath163 and @xmath164 . following ( 15 ) , the distribution for the total shift @xmath165 is the convolution @xmath166 over the pairs of @xmath66 and @xmath67 satisfying @xmath167 . in the continuum limit this reduces to a gamma distribution similar to ( 17 ) , @xmath168 = \lim_{\delta_x , \delta_r \rightarrow 0 } \ , \frac{p(\delta e)}{\delta_x \delta_r } = \frac{1}{6 } \ , v^3 \ , e^{-v},\end{aligned}\ ] ] for both @xmath164 sbs .
as expected and demonstrated in fig .
2(a ) , the sharp edge of this lineshape is furthest from the carrier . to extract axial trap frequencies @xmath169 from spectra like fig .
2(a ) , we fit the natural logarithm of the data ( to account for linear probe absorption ) with the lineshape ( [ pv ] ) to determine the spacing between the @xmath170 points of the red and blue sbs . the dimensionless fwhm of ( [ pv ] ) is approximately 4.131 . using this with ( [ vfunction ] ) gives the relation @xmath171 between the fwhm @xmath172 ( in temporal frequency units ) of the lineshape ( [ pv ] ) and the temperature @xmath0 .
equation ( 18 ) then follows from this together with eq .
( 4 ) , @xmath173 , and rewriting @xmath174 .
note that for non - magic lattices , the competition of harmonic and anharmonic shifts will lead to both broadening and narrowing effects for sb transitions . | for atoms or molecules in optical lattices , conventional thermometry methods are often unsuitable due to low particle numbers or a lack of cycling transitions .
however , a differential spectroscopic light shift can map temperature onto the line shape with a low sensitivity to trap anharmonicity .
we study narrow molecular transitions to demonstrate precise frequency - based lattice thermometry , as well as carrier cooling .
this approach should be applicable down to nanokelvin temperatures .
we also discuss how the thermal light shift can affect the accuracy of optical lattice clocks .
pacs numbers : 37.10.jk , 37.10.pq , 33.80.-b , 33.20.kf neutral atoms in tight optical traps have proven to be indispensable for time and frequency metrology @xcite and studies of many - body quantum phenomena @xcite .
recently , molecules have been similarly trapped and used for precision measurements of two - body and many - body physics @xcite .
these state - of - the - art experiments require ultracold temperatures to maximize control over the external degrees of freedom . however , few reliable thermometry techniques exist aside from time - of - flight ( tof ) expansion imaging @xcite .
this technique is unreliable with low particle numbers or a lack of cycling transitions , as is often the case for molecules @xcite .
moreover , existing methods often break down at ultralow temperatures in the nanokelvin regime , and new thermometry tools are needed , particulary those not relying on complex modeling @xcite . in this letter
, we show that if atoms or molecules can be trapped and probed in the lamb - dicke ( ld ) and resolved - sideband ( rsb ) regimes @xcite of an optical lattice , the temperature @xmath0 can be determined from the spectrum of the carrier line ( c `` in fig . [ fig : schematic ] ) with a precision that is roughly an order of magnitude higher than for conventional sideband ( sb ) based thermometry @xcite .
this temperature determination requires only the polarizability ratio @xmath1 for the excited and ground states at the trap conditions , and the full - width - at - half - maximum ( fwhm ) @xmath2 of the carrier line shape .
we show that t _ , [ eq : tresult ] where @xmath3 and @xmath4 are the planck and boltzmann constants , @xmath5 , and @xmath6 refers to carrier - based thermometry .
a light shift measurement can directly yield @xmath7 where @xmath8 is the particle mass , @xmath9 is the lattice wavelength , @xmath10 are the trap frequencies along the axial and transverse directions defined relative to the probe beam , and @xmath11 is the temperature - independent light shift as described below . expression ( [ eq : tresult ] ) is valid
( i ) for boltzmann particle statistics in deep lattices , ( ii ) in the rsb regime @xmath12 , where @xmath13 is the unbroadened carrier linewidth that would be measured in a ' ' magic " trap with @xmath14 , ( iii ) if @xmath15 , and ( iv ) if the transverse sbs are not excited , either by ensuring @xmath16 or by strictly controlling the probe - lattice alignment @xcite .
importantly , @xmath6 is insensitive to leading - order trap anharmonicities .
this carrier thermometry should be valid at very low temperatures as long as @xmath1 can be tuned sufficiently far from unity .
for example , if @xmath17 hz and @xmath18 , the subnanokelvin regime is accessible , provided there is a sufficient signal - to - noise ratio to detect the few particles in excited trap levels .
carrier thermometry is applicable in the radio - frequency ( rf ) , microwave , and optical regimes , which includes experiments with alkali and alkaline - earth metal atoms as well as simple molecules .
we apply the technique to accurately characterize the temperature of molecules created via photoassociation in an optical lattice .
we identify and quantify the heating mechanisms , and furthermore use the temperature dependence of the carrier for reducing the molecule temperature by a factor of @xmath19 .
to demonstrate carrier thermometry , we create sr@xmath20 molecules via photoassociation ( pa ) @xcite from laser - cooled @xmath21sr atoms in a one - dimensional ( 1d ) optical lattice @xcite , as illustrated in fig .
[ fig : schematic ] .
the lattice , which is present during laser cooling , is tunable by tens of nanometers around a central wavelength of 910 nm , and has a beam waist @xmath22 @xmath23 m .
the probing is performed on an optical transition to a subradiant excited state of sr@xmath20 @xcite , with @xmath24 hz limited by the laser linewidth .
figure [ fig : latticesidebands](a ) shows an optical spectrum taken along the lattice axis , including the narrow carrier transition and the first - order axial sbs .
the trap frequency @xmath25 is found from the sb spacing with a state - insensitive lattice [ @xmath26 @xcite . as @xmath1 is tuned via wavelength or polarization ( the latter is possible via tensor shifts , and experimentally simpler ) , a differential light shift moves the line center , and an asymmetric @xmath0-dependent line broadening develops . the carrier spectrum in fig .
[ fig : latticesidebands](b ) corresponds to @xmath27 .
the line is thermally broadened , with the fwhm @xmath2 yielding @xmath28 in eq .
( [ eq : tresult ] ) .
the measured temperatures @xmath28 are plotted in fig .
[ fig : temperature ] for various optical lattice light powers , for sr@xmath20 molecules in the two least - bound vibrational levels of the electronic ground state @xmath29 ( @xmath30 and @xmath31 ) . also shown
are axial temperatures independently determined from the ratios of the blue to red sb areas @xcite t_x()^-1 .
[ eq : tratio ] equation ( [ eq : tratio ] ) holds for boltzmann statistics in deep lattices ( in our experiment , trap depth @xmath32 ) . while the methods are in close agreement , carrier thermometry benefits from the larger signal - to - noise ratio for the carrier relative to the sbs .
its uncertainties are smaller by almost an order of magnitude , reaching the percent level . the smaller number of @xmath31 molecules in fig .
[ fig : temperature](b ) hampers the sb but not the carrier measurements .
moreover , at colder temperatures than are reached here , the carrier method is expected to be superior due to the large relative uncertainty of determining the area of a vanishingly small red sb .
we apply carrier thermometry to systematically characterize the temperature of photoassociated molecules .
since pa is the optimal method to create a variety of dimers and can yield molecules in the absolute ground state @xcite , any heating that occurs during pa must be understood and suppressed if such molecules are to reach a high phase - space density .
pa into an electronically excited molecular level is usually followed by a spontaneous or stimulated decay to the electronic ground state . from kinetic considerations
, the temperature of homonuclear dimers created in this way is expected to be nearly the same as for the precursor atoms @xcite . however , for typical conditions we observe heating by roughly a factor of 2 , as shown in fig .
[ fig : systematics](a ) .
this heating is caused by incoherent photon scattering and can be partially suppressed by using shorter pa pulse durations , as shown in figs .
[ fig : systematics](a ) and [ fig : systematics](b ) .
heating of the molecules by the optical lattice [ fig .
[ fig : systematics](c ) ] @xcite is also present , but occurs at a much slower rate .
while sb cooling is an established technique for reducing axial temperatures @xcite , the method described here suggests a `` carrier cooling '' procedure that can reduce the 3d temperature of a cold gas .
figure [ fig : systematics](d ) illustrates the reduction of the molecule temperature by a factor of @xmath19 via weakly exciting the hotter molecules in the tail of the line shape . in the case of open transitions such as for sr@xmath20 , the lowering of @xmath0 is due to energy selection @xcite and is not accompanied by a phase - space density increase .
the experiments in this work utilize a narrow optical transition , but carrier thermometry can be performed in any frequency regime .
we have confirmed the results for a two - photon @xmath33 ghz vibrational transition in ground - state sr@xmath20 . in microwave and rf regimes ,
the ld and rsb conditions for carrier thermometry may be satisfied in optical traps not utilizing a lattice configuration . to obtain eq .
( [ eq : tresult ] ) , we consider the case of a 1d lattice to reflect our setup as well as that of many lattice - clock type experiments , but note that the method is general and extends to 2d and 3d lattices . near the center of a lattice
well , the potential energy is nearly harmonic @xmath34 where @xmath35 @xmath36 , @xmath37 is the lattice light power , @xmath38 is the speed of light , and @xmath39 is the permittivity of vacuum . if @xmath40 , the potential for the final internal state differs from that for the initial state , @xmath41 .
the light shift ( or differential ac stark shift ) of a carrier transition by the lattice is the difference in final and initial expectations @xmath42 using eq .
( [ eq : ur ] ) , the energy expectation values are of the form @xmath43 where @xmath44 is the single - particle hamiltonian , @xmath45 are harmonic oscillator occupation numbers , and @xmath46 . evaluating the net light shift ( [ eq : w1 ] ) using eq .
( [ eq : hi ] ) , under the assumption that sbs are not excited at an appreciable rate such that @xmath47 @xcite , we find @xmath48 + \hbar \omega_r \ , \text{coth}\left [ \frac { \hbar \omega_r } { 2 k_b t } \right ] \right).\end{aligned}\ ] ] note that if equipartition is valid ( @xmath49 ) , eq .
( [ eq : wparts ] ) simplifies to @xmath50 highlighting the nonthermal and thermal contributions to @xmath51 .
the carrier line shape described below permits a clean extraction of @xmath52 , and therefore of @xmath1 , if the trap depth or axial trap frequency ( [ eq : wxwr ] ) is known .
the measured shifts @xmath51 and @xmath11 are marked in fig .
[ fig : latticesidebands](b ) and plotted versus the lattice light power in fig .
[ fig : latticesidebands](c ) . the @xmath0-dependent light shifts @xmath53 and @xmath54 in eq .
( [ eq : wparts ] ) cause asymmetric line broadening @xcite .
the carrier transition from the trap state @xmath55 experiences a differential light shift @xmath56 where the axial and radial contributions are @xmath57 and @xmath58 .
the boltzmann probability distribution for the discrete variable @xmath59 is @xmath60 the partition function is @xmath61 , and the dimensionless function u(e_i)= 0 [ eq : u ] parametrizes the boltzmann exponent with a discrete step size of @xmath62 for @xmath63 .
similarly , for the radial shift , p_r(e_r)=u(e_r)e^-u(e_r ) .
[ eq : per ] the discrete probability for the energy @xmath64 is then the convolution @xmath65 over the pairs of @xmath66 and @xmath67 satisfying @xmath68 . if @xmath69 , which is the case here , the discrete expression ( [ eq : pdiscrete ] ) may be simplified in the continuum limit of @xmath70 . noting that the probability density @xmath71= \lim_{\delta_i \rightarrow 0 } p_i(\delta e_i)/\delta_i,\end{aligned}\ ] ] we obtain @xmath72 and @xmath73 .
the probability @xmath74 then reduces to a gamma distribution [ u(e ) ] = _ 0^_r(u - u_x ) _
x(u_x ) d u_x= u^2 e^-u .
[ eq : lineshape ] the probability density ( [ eq : lineshape ] ) directly yields the spectroscopic line shape since carrier transition rates are nearly independent of @xmath45 ( this assumption may need to be modified if @xmath1 is far from unity ) . the line shape has the form of a boltzmann distribution in a 3d harmonic trap @xcite , as can be expected from ( [ eq : de1])-([der ] ) . in a special case where only the ground axial state is occupied ( @xmath75 ) ,
the line shape ( [ eq : lineshape ] ) is replaced by the 2d boltzmann result @xmath76 = ue^{-u}$ ] .
we have assumed @xmath77 @xcite , but the analysis can be adapted to other situations , including non - boltzmann distributions .
the fwhm of the function ( [ eq : lineshape ] ) is nearly 3.395 , which together with ( [ eq : u ] ) yields eq . ( [ eq : tresult ] ) .
figure [ fig : latticesidebands](b ) shows a fit of line shape ( [ eq : lineshape ] ) to a carrier spectrum , with its fwhm @xmath2 directly yielding temperatures in figs .
[ fig : temperature ] and [ fig : systematics ] .
note that the width of the carrier is much smaller than its light shift ; hence , it was necessary to stabilize @xmath37 .
this was done with minimal error from interference of the forward and retroreflected lattice beams by using a pellicle beam splitter to sample the forward beam produced by an optical fiber with an angled output face .
it is a property of harmonic oscillator eigenstates that the dominant anharmonic corrections ( proportional to @xmath78 , @xmath79 , @xmath80 ) shift their energy by an amount that is independent of trap depth @xcite .
hence eqs .
( [ eq : lineshape ] ) and ( [ eq : tresult ] ) are largely unaffected by anharmonic corrections . however , these corrections allow a determination of temperature from the shape of first - order axial sbs in a 1d lattice , using spectra with @xmath14 as in fig . [
fig : latticesidebands](a ) @xcite .
as for carriers , this approach yields 3d temperatures [ versus 1d temperatures for eq .
( [ eq : tratio ] ) ] , where sensitivity to @xmath0 now arises from the anharmonicity of the lattice trap . introducing leading - order corrections to the harmonic approximation of a sinusoidal potential , and adapting the approach used to derive the carrier line shape @xcite
, we find the temperature in a state - insensitive lattice t_0.484m^2f_x_/k_b , [ eq : sbwidth ] where @xmath81 is the fwhm of the sb line shape given by @xmath82 = ( 1/6)v^3 e^{-v}$ ] , and @xmath83 is a function similar to @xmath84 @xcite . using this approach on the data of fig .
[ fig : temperature](a ) yields @xmath85 that is too high by @xmath86 @xmath23k compared to @xmath28 and @xmath87 .
there are several reasons for eq .
( [ eq : sbwidth ] ) to be less reliable than eqs .
( [ eq : tresult ] ) and ( [ eq : tratio ] ) .
first , sbs are more sensitive than the carrier to distortion by other broadening mechanisms , since there is no tunability of @xmath81 , unlike for @xmath88 .
additionally , any axial displacement from the gaussian lattice beam waist produces new leading - order anharmonic corrections to eqs .
( [ eq : ur ] ) and ( [ eq : sbwidth ] ) that could strongly affect the sb result . more generally , @xmath28 in eq .
( [ eq : tresult ] ) depends only on the polarizability ratio @xmath1 , a fundamental property of the atom or molecule that can be measured with a high accuracy .
in contrast , @xmath85 depends on @xmath25 , which varies slightly across the sample .
note that the thermal contribution @xmath89 in fig .
[ fig : latticesidebands](c ) and eqs .
( [ eq : wparts ] ) and ( [ eq : wlinear ] ) is a significant fraction of the lattice light shift .
this could affect optical lattice clocks @xcite if the atom temperature versus trap depth [ fig .
[ fig : systematics](a ) , stars ] does not linearly extrapolate to exactly zero at @xmath90 @xcite .
for example , if it extrapolates to just @xmath91 @xmath23k , then for @xmath92 as in ref .
@xcite , the residual thermal line pulling from eq .
( [ eq : wlinear ] ) is @xmath93 , comparable to the total uncertainty budget .
furthermore , for typical clock experiment conditions , the full clock shift ( [ eq : wparts ] ) must be used , where @xmath51 is not linear in @xmath0 .
this nonlinearity leads to an effective offset @xmath94 @xmath23k in an experiment with our parameters , again leading to line pulling .
counterintuitively , if the temperature is kept fixed at all lattice depths , this thermal pulling is even more problematic . in general , the variation of temperature with lattice depth depends on the cooling and trapping procedure . in conclusion , we have shown that narrow spectral lines of atoms or molecules tightly trapped in optical lattices allow highly precise 3d temperature determinations , and are not limited by low particle numbers , a lack of cycling transitions , or ultralow temperatures .
the method is purely frequency based , requiring only measurements of the carrier linewidth , light shift , and axial trap frequency , and is mostly immune to trap anharmonicities .
we experimentally demonstrate complete control over molecular external and internal degrees of freedom in the ld and rsb regimes of a weakly state - sensitive optical lattice , use carrier cooling to reduce the temperature of the ultracold molecules , and identify the significant heating processes of photoassociated molecules .
furthermore , the result ( [ eq : tresult ] ) can be inverted to accurately predict light - shift - induced thermal dephasing .
we thank m. g. tarallo , a. t. grier , and s. rolston for discussions , and acknowledge onr grant no .
n00014 - 14 - 1 - 0802 , nist grant no .
60nanb13d163 , and aro grant no .
w911nf-09 - 1 - 0504 for partial support of this work . m. m. and g. z. i. acknowledge nsf igert grant no . dge-1069260 . |
we present an analysis of self - imaging in a regime beyond the paraxial , where deviation from simple paraxial propagation causes apparent self - imaging aberrations .
the resulting structures are examples of aberration without rays and are described analytically using post - paraxial theory .
they are shown to relate to , but surprisingly do not precisely replicate , a standard integral representation of a diffraction cusp . beyond the talbot effect , this result is significant as it illustrates that the effect of aberration as manifested in the replacement of a perfect focus with a cusp - like pattern can occur as a consequence of improving the paraxial approximation , rather than due to imperfections in the optical system . the talbot effect @xcite is a well - known self - imaging phenomenon in classical optics , where a transversely periodic field reproduces itself at periodic distances in the propagation direction .
as a natural consequence of paraxial wave propagation , it has also been observed in the diffraction of matter waves @xcite and has an analogue in the revivals of quantum wavepackets @xcite .
recently , the talbot effect for surface plasmons @xcite has attracted significant attention because of applications in areas such as nanolithography @xcite .
the perfect image revival of the classical talbot effect can be understood in terms of fresnel diffraction @xcite and interpreted as an artefact of the corresponding paraxial approximation .
refining the propagation beyond the paraxial approach leads to aberration - like distortion of perfect self - imaging @xcite .
nonparaxial self - imaging has been studied previously for binary gratings using both scalar theory @xcite and full electromagnetic treatment @xcite , and has particular importance in plasmonics where the transverse periodicity is comparable to the plasmon wavelength @xcite .
such nonparaxial self - images are not perfect revivals as they are in the paraxial case , and are only approximately periodic in the direction of propagation . in this paper
, we investigate a nonparaxial talbot effect and explain that the resulting apparent aberration the maximum intensity shifts and distorts to a cusp - like pattern is not caused by imperfection in a lens or optical system ( since the talbot effect is lensless imaging ) , but rather as an actual improvement of the approximation , appropriate to our diffraction regime . specifically , we invoke the fourth - order post - paraxial approximation @xcite to investigate the talbot - based focusing of scalar waves in a regime which is nonparaxial , but not extremely so : the transverse periodicity is of the order of some tens of wavelengths .
this explains the appearance of diffraction cusps in the intensity patterns of surface plasmons calculated in ref .
@xcite .
we show that these aberration - like effects are not unique to plasmons and are connected to a standard integral representation of the diffraction pattern about a cusped caustic @xcite despite the absence here of any actual ray caustic .
furthermore , we obtain a simple expression for the approximate location of the post - paraxial foci , which are shifted to be short of the paraxial talbot distance .
to demonstrate the physics of the post - paraxial regime , we employ a general two - dimensional ( 2d ) formalism ( transverse in @xmath0 and propagating in @xmath1 ) which displays properties of both paraxial and nonparaxial regimes in different limits . specifically , it consists of a transverse array ( with period @xmath2 in the @xmath0-direction ) of monochromatic 2d point sources of wavelength @xmath3 .
the interference of the propagating waves in the @xmath0-@xmath1 plane weaves a diffraction carpet , which is similar to that studied in the plasmon talbot effect of @xcite .
the paraxial approximation applies when @xmath4 is asymptotically large . in this case
the talbot distance @xmath5 is the spacing between planes of image revival . ) for @xmath6 the paraxial field of eq .
[ eq : parax ] ; @xmath7 exact field of eq .
[ eq : helm ] ; @xmath8 numerically simulated 1d binary gold grating ( of refractive index @xmath9 , thickness @xmath10 nm and slit width to grating period ratio of @xmath11 ) illuminated by a normally incident , te - polarized , monochromatic ( @xmath12 nm ) plane wave ; and @xmath13 the post - paraxial field as given by eq .
[ eq : ppp ] .
a value of @xmath14 is outside the region of validity for the paraxial approximation , since @xmath6 is fundamentally different to @xmath7 .
the fields @xmath7 , @xmath8 and @xmath13 exhibit foci ( indicated by the white box ) which are aberrated compared to @xmath6.,width=566 ] the initial pattern also revives at odd multiples of @xmath15 , albeit transversely shifted by @xmath16 , leading to an alternate definition of the talbot distance @xcite of @xmath17 ( used in this paper ) . in
what follows , we use the dimensionless variables @xmath18 and @xmath19 . in general , for any @xmath20 , the field at a distance @xmath21 , propagating according to the 2d helmholtz equation , is given by ( cf .
eq .
( 1 ) of @xcite ) @xmath22\label{eq : helm},\end{aligned}\ ] ] where we assume negligible attenuation and consider only propagating waves , i.e. , @xmath23 $ ] with square brackets denoting the integer part .
. [ eq : helm ] can be approximated by expanding @xmath24 in the paraxial approximation to eq .
[ eq : helm ] , the quartic term in eq .
[ eq : expansion ] is negligibly small , so @xmath25\label{eq : parax},\end{aligned}\ ] ] where we have neglected the overall phase factor of @xmath26 .
, @xmath27 ) with colours corresponding to fig .
[ fig : regcomp ] ; @xmath6 paraxial focus of eq .
[ eq : parax ] ; @xmath7 the exact field of eq .
[ eq : helm ] ; @xmath8 numerically simulated focus according to the parameters of fig .
[ fig : regcomp]@xmath8 ; @xmath13 post - paraxial focus of eq .
[ eq : psfresult ] ( @xmath28 ) ; @xmath29 the pearcey function eq .
[ eq : pe ] ; and @xmath30 the incomplete pearcey function ( @xmath31 term of eq . [ eq : psfresult]).,width=566 ] as shown in fig .
[ fig : regcomp]@xmath6 , the initial paraxial field revives for integer @xmath21 , since the second term in eq .
[ eq : parax ] will introduce a factor of @xmath32 , and we recover the same field as for @xmath33 .
the factor of @xmath32 determines whether or not the self - image is shifted by half a period .
a post - paraxial refinement is made by retaining the quartic term in eq .
[ eq : expansion ] @xcite , giving @xmath34.\end{aligned}\ ] ] comparison between eq .
[ eq : ppp ] and eq .
[ eq : parax ] shows that the perfect @xmath21-periodicity of the talbot effect is an artefact of the paraxial approximation , and is destroyed in the post - paraxial regime by the inclusion of the quartic ( and possibly higher terms ) @xcite .
figs .
[ fig : regcomp]@xmath7 and [ fig : regcomp]@xmath13 show the exact and post - paraxial intensity distributions respectively , and it is immediately apparent that while the foci of these fields are similar to one another , they are very different , and less defined , from those of the paraxial case , ( @xmath2 ) . comparing the foci of each regime in fig .
[ fig : regcomp ] it is clear that the perfect foci of the paraxial approximation of @xmath6 have been perturbed in the post - paraxial regime to cusp - like distributions , evident in @xmath7 and @xmath13 . fig .
[ fig : regcomp]@xmath8 shows a separate simulation for a binary grating , which has been calculated numerically using rigorous electromagnetic theory @xcite ( parameters given in the caption ) , which very closely resembles the exact and post - paraxial fields in @xmath7 and @xmath13 , including the cusp - like foci .
while such a pattern is typical for aberration , its appearance here is because of a refinement of the usual fresnel paraxial approximation .
in fact , the inclusion of a quartic term destroying a perfect focus is exactly analogous to the spherical aberration of a parabolic mirror , though remarkably , in fig .
[ fig : regcomp ] it is _ improvement _ of the approximation and not any aberration in the system that causes the effect ! the fields in fig .
[ fig : regcomp ] are those for an infinite array , and applying the paraxial approximation to such a system is not fully consistent @xcite , since at all @xmath21 there would be oblique contributions to the field from arbitrarily transversely distant sources a fact that is at odds with the concept of paraxiality .
instead we model a finite array of @xmath35 sources by truncating the sums of eqs .
[ eq : helm ] , [ eq : parax ] and [ eq : ppp ] at @xmath36 , where @xmath37 $ ] @xcite , embodying the walk - off effects of finite arrays @xcite . fig .
[ fig : cusps ] shows the various foci according to this truncation . comparing the paraxial focus fig .
[ fig : cusps]@xmath6 to the exact fig .
[ fig : cusps]@xmath7 , the aberration effect is clear .
fig .
[ fig : cusps]@xmath8 shows the cusp pattern from the diffraction grating of fig .
[ fig : regcomp]@xmath8 , but includes only diffraction orders up to @xmath36 .
there is a slight discrepancy between @xmath7 and @xmath8 because of the equal amplitudes of the planes waves in the superposition of eq .
[ eq : helm ] compared to varying amplitudes of the diffraction orders of the simulation of @xmath8 .
the fields are nevertheless very similar and the cusp reveals the breakdown of the paraxial approximation .
below , we develop an analytic post - paraxial approximation to these fields , shown in figs .
[ fig : regcomp]@xmath38 [ fig : cusps]@xmath38 from inclusion of the quartic term ( and no higher ) , which is enough to capture the structure of the exact and simulated foci . including higher terms improves the quantitative agreement , but without further qualitative changes . to derive a simple representation of the complicated interference pattern , we follow a similar technique to ref .
@xcite but employ a _
finite _ poisson sum formula @xcite to account for the truncation at @xmath36 of the spectrum . applying this to @xmath39 yields the main result of this paper , @xmath40 where @xmath41\right\}$ ] , and @xmath42\right\};\nonumber\\ h(\zeta)\equiv\sqrt{2\gamma}/[\pi(t+\zeta)]^{\frac{1}{4}};\nonumber\\ \tau(\zeta ) \equiv m / h(\zeta);\nonumber\\ \sigma_x(\xi,\mu,\zeta)\equiv 2\pi(\xi-\mu ) h(\zeta);\nonumber\\ \sigma_z(\zeta)\equiv\pi h^2(\zeta)\zeta.\label{eq : scalings}\end{aligned}\ ] ] in the above
, we have set @xmath43 , where @xmath44 is a small shift from the @xmath45 talbot distance , and @xmath46 .
the terms @xmath47 are end - point contributions from the truncation of the spectrum .
we recognise that the integral in eq .
[ eq : psfresult ] is an incomplete version @xcite of the pearcey function @xcite ( i.e. with a finite integration domain ) .
the pearcey function is given by @xmath48,\end{aligned}\ ] ] which describes the generic diffraction pattern decorating a cusp caustic @xcite .
the sum in eq .
[ eq : psfresult ] represents an array of cusp - like foci in the @xmath0-direction , with transverse and longitudinal scalings given by @xmath49 and @xmath50 .
although in optics it is usual to see the pearcey function as decoration around a cusp caustic of rays that arise from a geometric optics approach @xcite , this is not necessary @xcite .
since here we are outside the applicability of geometrical optics , the propagating function eq .
[ eq : helm ] can not be approximated in terms of rays , and instead the pearcey function has come from the quartic post - paraxial term in eq .
[ eq : ppp ] .
importantly , since the maximum of the pearcey function is below the geometric focus , the maxima of the post - paraxial self - images are below the paraxial talbot distance .
this differs from the nonparaxial talbot distance @xcite which is accurate when only a small number of diffraction orders contribute .
fig .
[ fig : cusps]@xmath13 shows the diffraction pattern near the talbot distance ( @xmath51 ) according to eq .
[ eq : psfresult ] for @xmath28 .
the intensity of the pearcey function , scaled to the size set by @xmath49 and @xmath50 , is shown in fig .
[ fig : cusps]@xmath29 .
fig .
[ fig : cusps]@xmath30 is the incomplete pearcey function ( the @xmath31 term of eq . [ eq : psfresult ] ) . comparing @xmath7 and @xmath13 , we see that although the sum in eq .
[ eq : psfresult ] has infinitely many terms , only the first few are needed to accurately describe the fine details of the field about the focus .
the transverse fringes come from the truncation of the integral in eq .
[ eq : psfresult ] , and subsequently are absent from the pearcey function , @xmath29 , and are just evident in @xmath30 . in all these images , it is clear that the maximum intensity is short of the paraxial talbot distance .
we now obtain an expression for the approximate location of the maximum intensity . as a first approximation
we take only the @xmath31 term of eq .
[ eq : psfresult ] and let @xmath52 go to infinity .
this is equivalent to approximating the focal region near the first talbot distance with the pearcey function , eq .
[ eq : pe ] . by using the numerically calculated , constant position of maximum intensity of the unscaled pearcey function , @xmath53
, we can employ the scalings of eq .
[ eq : scalings ] to calculate the approximate location of the maximum intensity of the post - paraxial focus as a shift , @xmath54 , from the @xmath55 talbot distance . by setting @xmath56
, we obtain @xmath57 fig .
[ fig : feather](@xmath2 ) shows the intensity of the exact field along the symmetry axis of the cusp @xmath58 for @xmath59 as a function of increasing @xmath20 .
asymptotically large @xmath20 gives rise to maximum intensity at the talbot distance ( @xmath60 in the paraxial limit ) , but for smaller @xmath20 it is located short of this .
fig .
[ fig : feather]@xmath7 shows the post - paraxial approximation of eq .
[ eq : psfresult ] including only the terms @xmath28 .
the dashed curve is the location of the maximum intensity according to the approximation of eq .
[ eq : peak ] .
note that this depends on @xmath61 , such that consecutive foci will be shifted from their paraxial counterparts by an increasing amount .
the dots indicate the maxima of the exact field ( eq . [ eq : helm ] ) for integer @xmath20 , though they do not always correspond to well localised single peaks , as in the case of @xmath62 and @xmath63 .
the intensity fringes that we observe in fig .
[ fig : feather ] are a result of interference between the @xmath31 and the neighbouring @xmath64 terms .
inclusion of more @xmath65-terms reconstructs the discontinuities of fig .
[ fig : feather]@xmath6 ( which arise from the limits of the sum in eq .
[ eq : helm ] changing discretely with @xmath20 ) . the exact field , eq .
[ eq : helm ] ( normalised by @xmath66 + 1)$ ] ) and @xmath7 the post - paraxial field ( eq . [ eq : psfresult ] , summing @xmath28 ) as a function of @xmath20 .
the dashed line is the approximate location of maximum intensity according to eq .
[ eq : peak ] .
the dots are the actual maxima of the exact field for integer @xmath20 .
the solid line is the talbot distance.,width=415 ] to conclude , we have applied scalar post - paraxial theory to explain the observed cusp - like foci in the talbot effect for the post - paraxial regime , which have replaced the point foci of paraxial theory .
while the analytic form of these post - paraxial foci is related to the standard description of diffraction about a cusp , there is a subtle distinction since it is the _ incomplete _ pearcey function that occurs and not the standard pearcey function .
the finite integration domain of the incomplete pearcey is of significant asymptotic importance when compared with the infinite domain of the complete pearcey function of eq .
[ eq : pe ] .
this is clear when comparing intensity patterns - while both are cusp - like distributions , there is additional complex detail accounted for in the post - paraxial talbot effect that is absent from the normal pearcey function .
another important difference is that there is no ray caustic underpinning the structure of the post - paraxial cusp foci as there is with the complete pearcey function . despite their differences
, the location of the maximum of intensity of the post - paraxial foci can be fairly well approximated using a scaled pearcey function .
it is curious to see that the effect of post - paraxiality on a focus appears to be that of aberration , when compared to the paraxial case .
however , while the analytic form mimics that for aberration ( hence the inclusion of the pearcey function , albeit incomplete ) , it actually embodies an improvement of the paraxial approximation for the post - paraxial regime .
the phenomenon is a universal effect that occurs as the parameters of a system ( regardless of its exact realisation ) move away from paraxiality
. the results in this paper may plausibly find application in plasmonics , since it is a field in which the pertinent wavelengths and periodicities are comparable , and so nonparaxial effects are important . while the lengthscales considered here are slightly greater than those experimentally investigated so far , it would be both interesting and challenging to confirm the aberration - like effects of post - paraxiality on the talbot effect .
we gratefully acknowledge the funding of the epsrc .
mrd is a royal society university research fellow . |
|
in ( * ? ? ? * lwb12 hereafter ) , we carried out source type classifications of 4330 sources from the 2xmmi - dr3 catalog , which is the largest x - ray source catalog ever produced , containing 353,191 x - ray source detections for 262,902 unique x - ray sources from 4953 pointed observations made by @xcite .
for about one third of these 4330 sources we obtained reliable source types from the literature .
they mostly correspond to various types of stars , active galactic nuclei ( agns ) and compact object ( co ) systems containing white dwarfs ( wds ) , neutron stars ( nss ) , and stellar - mass black holes ( bhs ) .
these sources show different source properties in terms of the x - ray spectral shape , x - ray variability , and the multi - wavelength cross - correlation , based on which we came up with a method to classify the rest of the sources into the above three main categories , i.e. , stars , agns , and compact object systems . in brief , we first identified star candidates as those with low x - ray - to - ir flux ratios and/or frequent x - ray flares and then identified compact object system candidates from the remaining sources as those with properties hardly seen in agns , such as soft x - ray spectra and large long - term variability .
the compact object system candidates identified in this way could inevitably include exotic objects with high science value , such as very soft agns like @xcite and tidal disruption events ( tdes ) , in which stars approaching a supermassive bh ( smbh ) is tidally disrupted and subsequently accreted @xcite .
about 200 strong compact object system candidates were found .
some of them were poorly studied in the literature .
because these objects are important targets for studies of dense matter and accretion physics in extreme environment and they are also a treasure for discovery of objects with high science value , we have planned a series of companion papers on these poorly studied compact object system candidates in order to shed new light on their nature . while the other companion papers focus on single sources ( e.g. ,
* ; * ? ? ?
* ) , here we present a relatively large sample ( 18 in total ) , which were selected because they exhibited interesting properties of periodicity , very soft spectra , and/or large long - term variability .
some of their general properties from lwb12 are given in table [ tbl : dercat ] , including the number of source detections in the 2xmmi - dr3 catalog , the number of the observations , the 0.24.5 kev flux variation ( @xmath8 ) , etc . throughout the paper ,
we generally use the unique source index from the 2xmmi - dr3 catalog ( column srcid ) to refer to the sources ( table [ tbl : dercat ] ) . in this paper
, we concentrate primarily on x - ray observations
. we only present those that best demonstrate the source properties based on considerations of data quality , sampling of spectral states , etc .
some sources are also observed by other x - ray observatories , especially the _ chandra _ x - ray observatory ( cxo ) .
we mostly refer to the _ chandra _ source catalog ( csc , release 1.1 , * ? ? ? * ) for their behavior as observed with _
chandra _ , except for one observation , which was analyzed to show the period of source # 150140 . in section
[ sec : reduction ] , we describe the analysis of the selected and _ chandra _ observations of the sources . in section [ sec : res ] , we present the results and discuss the possible nature of each source .
the conclusions of our study are given in section [ sec : conclusion ] . throughout the paper , we assume a flat universe with the hubble constant @xmath9=73 km s@xmath5 mpc@xmath5 and the matter density @xmath10=0.27 .
rlcccccccccccccccccc 7204 & 2xmm j004705.4 - 251942 & 0.4 & @xmath11 & 9 & 2047.70 & 1 & 0.00 & co & s , v , e & & & @xmath12 & & & @xmath13 & @xmath14 & 22.68 & ngc253 & 0.74 + 8085 & 2xmm j005412.9 - 373309 & 0.3 & @xmath15 & 4 & 1794.51 & 4 & 1794.51 & co & s , v , e & & & @xmath16 & & & @xmath17 & @xmath18 & 36.33 & ngc300 & 1.11 + 29251 & 2xmm j031820.8 - 663035 & 0.3 & @xmath19 & 14 & 2189.90 & 2 & 76.19 & co & s , v , e & & & @xmath16 & & & @xmath17 & @xmath18 & 29.50 & ngc1313 & 0.18 + 65459 & 2xmm j080703.5 - 763248 & 0.5 & @xmath20 & 2 & 650.17 & 1 & 0.00 & co & s & & & @xmath21 & & & @xmath22 & @xmath23 & 3.19 & & + 101335 & 2xmm j121028.9 + 391748 & 0.2 & @xmath24 & 8 & 2169.02 & 8 & 2169.02 & co & s & & & @xmath25 & & & @xmath26 & @xmath27 & 2.63 & & + 101369 & 2xmm j121035.0 + 393123 & 0.3 & @xmath28 & 8 & 2169.00 & 3 & 0.78 & co & s & 0.5 & 19.33 & @xmath29 & & & @xmath30 & @xmath31 & 5.09 & & + 105130 & 2xmm j122628.8 + 333626 & 0.3 & @xmath32 & 5 & 1080.48 & 3 & 1080.48 & co & s , v , e & & & @xmath33 & & & @xmath34 & @xmath35 & 64.92 & ngc4395 & 1.65 + 107102 & 2xmm j123103.2 + 110648 & 0.2 & @xmath36 & 3 & 887.86 & 3 & 887.86 & co & s & 3.6 & 19.72 & @xmath37 & & & @xmath38 & @xmath39 & 2.11 & & + 122979 & 2xmm j133135.2 - 315541 & 0.5 & @xmath40 & 2 & 178.26 & 2 & 178.26 & co & h & & & @xmath41 & & & @xmath42 & @xmath43 & 1.67 & & + 142052 & 2xmm j154951.7 - 541630 & 0.2 & @xmath44 & 4 & 1821.84 & 4 & 1821.84 & co & v , g & 0.7 & 16.82 & @xmath45 & 1.3 & 13.76 & @xmath46 & @xmath47 & 23.22 & & + 150140 & 2xmm j165334.4 - 414423 & 0.2 & @xmath48 & 6 & 4.51 & 6 & 4.51 & co & h , g & & & @xmath49 & & & @xmath50 & @xmath51 & 1.58 & & + 165335 & 2xmm j181920.1 - 204541 & 0.4 & @xmath52 & 2 & 155.58 & 1 & 0.00 & co & v , g & 0.6 & 20.24 & @xmath53 & 0.6 & 12.87 & @xmath54 & @xmath55 & 24.46 & & + 174295 & 2xmm j203353.6 + 410717 & 0.4 & @xmath56 & 6 & 915.50 & 2 & 9.97 & co & v , g & 0.7 & 19.70 & @xmath57 & 0.2 & 14.01 & @xmath58 & @xmath59 & 52.24 & & + 183876 & 2xmm j223545.1 - 260451 & 0.3 & @xmath60 & 2 & 1827.13 & 2 & 1827.13 & ulx & s , e & & & @xmath61 & & & @xmath62 & @xmath63 & 1.16 & ngc7314 & 0.82 + 187831 & 2xmm j231818.7 - 422237 & 0.4 & @xmath64 & 3 & 2166.30 & 1 & 0.00 & ulx & s , e & & & @xmath65 & & & @xmath66 & @xmath67 & 5.01 & ngc7582 & 0.92 + 222148 & 2xmmi j004211.2 + 410429 & 0.3 & @xmath68 & 18 & 2782.71 & 5 & 39.63 & co & v , e & & & @xmath69 & & & @xmath70 & @xmath71 & 152.20 & ngc224 & 0.15 + 222243 & 2xmmi j004246.0 + 411736 & 0.3 & @xmath72 & 19 & 2944.80 & 3 & 35.56 & xgs & e & & & @xmath73 & 0.8 & 14.43 & @xmath74 & @xmath75 & 6.00 & ngc224 & 0.02 + 241580 & 2xmmi j123047.0 + 413651 & 0.4 & @xmath76 & 3 & 2217.86 & 2 & 33.89 & co & v , e & & & @xmath77 & & & @xmath78 & @xmath79 & 16.10 & ngc4490 & 0.87 rccccccccc 7204 & 0152020101 & 2003 - 06 - 20 & 8.1/7.3/6.6 & 58.4/77.7/78.2 & 15 + & 0305860401 & 2005 - 05 - 22 & -/-/11.6 & -/-/34.7 & 25 + & 0112800201 & 2000 - 12 - 26 & 11.5/11.6/12.6 & 27.1/33.4/33.6 & 15 + 29251 & 0205230501 & 2004 - 11 - 23 & 6.4/5.4/5.3 & 12.5/15.5/15.5 & 12 + 65459 & 0306560301 & 2005 - 09 - 30 & 2.7/1.7/1.8 & 63.1/86.9/86.6 & 12 + & 0112830201 & 2000 - 12 - 22 & 6.8/6.7/7.8 & 50.8/58.9/58.9 & 13 + & 0402660201 & 2006 - 11 - 29 & -/6.7/7.8 & -/29.3/28.8 & 10 + 101369 & 0112830201 & 2000 - 12 - 22 & 7.0/6.9/5.9 & 50.9/58.9/58.9 & 14 + 105130 & 0200340101 & 2004 - 06 - 03 & 6.9/7.9/8.0 & 60.4/75.7/76.7 & 20 + & 0145800101 & 2003 - 07 - 13 & 7.5/7.1/8.1 & 45.4/58.0/61.4 & 25 + & 0306630101 & 2005 - 12 - 13 & 6.8/-/5.9 & 54.8/-/68.7 & 25 + & 0306630201 & 2005 - 12 - 17 & 6.8/-/5.9 & 80.8/-/92.2 & 25 + & 0105261401 & 2002 - 07 - 27 & 8.0/8.3/- & 10.2/17.4/- & 15 + & 0105261701 & 2003 - 01 - 21 & 17.4/-/- & 15.3/-/- & 25 + & 0203910101 & 2004 - 02 - 08 & 10.1/9.1/8.6 & 5.9/8.7/8.7 & 25 + & 0402910101 & 2006 - 08 - 21 & 8.1/9.1/9.5 & 38.8/45.6/45.5 & 25 + & 0410581901 & 2007 - 08 - 09 & 8.4/9.3/9.9 & 13.2/16.3/16.3 & 13 + & 0560181101 & 2009 - 02 - 04 & 10.2/9.3/8.8 & 48.5/57.1/57.2 & 25 + & 0109490101 & 2001 - 09 - 05 & 7.8/8.4/9.3 & 27.4/32.9/32.9 & 15 + & 0109490201 & 2001 - 09 - 06 & 7.8/8.4/9.3 & 15.9/20.1/20.4 & 15 + & 0109490301 & 2001 - 09 - 07 & 7.8/8.4/9.3 & 28.4/34.0/34.1 & 15 + & 0109490401 & 2001 - 09 - 08 & 7.8/8.5/9.3 & 17.9/30.5/30.7 & 15 + & 0109490501 & 2001 - 09 - 09 & 7.8/8.4/9.3 & 25.4/30.7/30.7 & 15 + & 0109490601 & 2001 - 09 - 10 & 7.8/8.4/9.3 & 27.1/32.5/32.6 & 15 + & 6291(cxo ) & 2005 - 07 - 16 & 9.2 & 44 & 10 + 165335 & 0149610401 & 2003 - 09 - 14 & 5.6/4.7/5.4 & 27.3/32.0/32.0 & 13 + 174295 & 0200450301 & 2004 - 11 - 09 & 13.4/13.5/12.4 & -/21.4/21.4 & 25 + 183876 & 0311190101 & 2006 - 05 - 04 & 8.2/8.2/9.2 & 67.7/82.6/82.7 & 12 + 187831 & 0204610101 & 2005 - 04 - 30 & 2.0/1.0/1.0 & 63.0/77.7/78.0 & 12 + & 0505720201 & 2007 - 12 - 29 & -/13.1/12.7 & -/26.8/26.9 & 20 + & 0505720301 & 2008 - 01 - 08 & -/13.1/12.9 & -/26.5/26.5 & 20 + & 0505720401 & 2008 - 01 - 18 & -/13.1/13.0 & -/21.6/21.7 & 20 + & 0505720501 & 2008 - 01 - 27 & -/13.1/- & -/20.0/- & 20 + & 0505720601 & 2008 - 02 - 07 & -/13.1/13.2 & -/21.2/21.2 & 20 + & 0551690201 & 2008 - 12 - 30 & -/13.2/12.8 & -/21.2/21.2 & 20 + & 0551690301 & 2009 - 01 - 09 & -/13.1/12.9 & -/21.1/21.1 & 20 + & 0551690401 & 2009 - 01 - 15 & -/13.2/13.0 & -/6.1/5.8 & 20 + & 0551690501 & 2009 - 01 - 27 & -/13.2/- & -/20.6/- & 20 + & 0551690601 & 2009 - 02 - 04 & -/13.1/13.1 & -/10.9/10.6 & 20 + & 0600660201 & 2009 - 12 - 28 & -/13.1/12.7 & -/18.1/18.2 & 20 + & 0600660301 & 2010 - 01 - 07 & -/13.1/12.9 & -/16.7/16.8 & 20 + & 0600660401 & 2010 - 01 - 15 & -/13.1/13.0 & -/16.7/16.7 & 20 + & 0600660501 & 2010 - 01 - 25 & -/13.1/- & -/19.1/- & 20 + & 0600660601 & 2010 - 02 - 02 & -/13.1/13.1 & -/16.8/16.8 & 20 + & 0650560201 & 2010 - 12 - 26 & -/13.1/12.7 & -/24.1/24.3 & 20 + & 0650560301 & 2011 - 01 - 04 & -/13.1/12.9 & -/32.4/32.7 & 20 + & 0650560401 & 2011 - 01 - 14 & -/13.2/13.0 & -/18.8/19.4 & 20 + & 0650560501 & 2011 - 01 - 25 & -/13.2/- & -/23.3/- & 20 + & 0650560601 & 2011 - 02 - 03 & -/13.1/13.2 & -/23.3/23.2 & 20 + & 0672130101 & 2011 - 06 - 27 & -/-/12.8 & -/-/99.3 & 20 + & 0672130601 & 2011 - 07 - 05 & -/-/12.8 & -/-/91.5 & 20 + & 0672130701 & 2011 - 07 - 07 & -/-/12.8 & -/-/89.9 & 20 + & 0672130501 & 2011 - 07 - 13 & -/-/12.9 & -/-/42.2 & 20 + 222243 & 0405320801 & 2007 - 01 - 16 & 2.6/1.7/2.4 & 9.9/13.4/13.4 & 13 + 241580 & 0556300101 & 2008 - 05 - 19 & 3.8/2.8/2.7 & 18.7/30.0/30.6 & 13 the details of the observations that we analyzed are given in table [ tbl : obslog ] .
we used sas 11.0.0 and the calibration files of 2011 june for reprocessing the x - ray event files and follow - up analysis .
the data in strong background flare intervals are excluded following the sas thread for filtering against high background .
we extracted the source spectrum for all available cameras from a circular region centered on the source , with the radius @xmath80 for each observation given in table [ tbl : obslog ] .
the background spectrum was extracted from a large circular region , typically with a radius of 50@xmath81100@xmath81 , near the source in each camera .
the event selection criteria followed the default values in the pipeline ( see table 5 in * ? ? ? * ) .
we rebinned the spectra to have at least 20 counts in each bin so as to adopt the @xmath82 statistic for the spectral fits , but for some very faint detections , the spectra were not rebinned , and the fits were carried out using the c statistic . because most of our data have low statistics , we fitted the spectra only with simple models : a powerlaw ( pl , _ powerlaw _ in xspec @xcite ) , a multicolor disk ( mcd , _ diskbb _ in xspec ) , a single temperature blackbody ( bb , _ bbodyrad _ in xspec ) , or their combinations . sometimes , a gaussian emission line ( described by the _
gaussian _ model in xspec ) is added when needed .
all models included an interstellar medium absorption , described by the _ wabs _ model in xspec . for sources that we found to show periodic modulations in the light curves from visual inspection
, we created the leahy power @xcite for each observation to confirm the presence of periodicity .
we used pn light curves extracted from the source region and binned at the frame readout time ( mos1 was used for the observation 0105261401 of source # 122979 due to strong background flares in pn ) . to constrain the values of their periods and examine the profiles of the modulation , we employed the epoch folding search technique .
the folded background - subtracted light curve for each camera is obtained by subtracting the folded light curve from the source region by the folded light curve from the background region ( after being rescaled to the size of the source region ) .
the total folded light curve for several cameras / observations are obtained by summing the folded background - subtracted light curves for individual cameras weighted by their total exposures ( summed over all phase bins ) .
all light curves were barycenter - corrected .
we also analyzed one _ chandra _ observation of source # 150140 ( table [ tbl : obslog ] ) to check the persistence of its period .
it was carried out using the imaging array of the axaf ccd imaging spectrometer ( acis ; * ? ? ?
we used the ciao 4.4 package .
the data were reprocessed to apply the latest calibration ( caldb 4.4.8 ) .
we used radii of 10@xmath81 ( table [ tbl : obslog ] ) and 50@xmath81 for the circular source and background regions , respectively .
the long - term flux curves , light curves of bright detections , and sample spectral fits of our sources are shown in figures [ fig : srcltlc][fig : srcspec ] , respectively .
the spectral fit results are given in table [ tbl : spfit ] .
we discover signs of coherent pulsations from three hard x - ray sources , # 254026 , # 150140 and # 122979 .
source # 254026 was presented in @xcite , and in the following we focus on the latter two sources .
their x - ray spectra can be fitted with an absorbed pl , with photon indices @xmath83 ( table [ tbl : spfit ] ) .
neither of them have optical or ir counterparts found in the usno - b1.0 catalog @xcite or the 2mass point source catalog ( 2mass psc , * ? ? ?
* ) within 4@xmath81 .
source # 150140 appears on the outskirts of the young open cluster ngc 6231 ( about 1.64 kpc away and a few myr old , * ? ? ?
there are six observations spanning only five days in 2001 september , with @xmath8=1.6 .
the 0.310 kev absorbed luminosity is about 7@xmath84@xmath85 erg s@xmath5 if the source is in this cluster or about 2@xmath84@xmath86 erg s@xmath5 if a source distance of 8.5 kpc is assumed .
there is a _
observation taken nearly four years later , with a similar spectrum ( table [ tbl : spfit ] ) .
the spectral fits indicate fairly strong absorption ( @xmath87 cm@xmath5 ) toward the source .
figure [ fig:150140foldcurve ] , which plots the timing properties of the source , shows that there are x - ray modulations at a period around 5845 s ( i.e. , 1.62 hr ) in all the and _ chandra _ observations . in the search for the period using the epoch folding technique , we used only the observations , as their time gap with _
observation is too large and adding the _ chandra _ observation only results in many similar local maxima of the total @xmath82 near the period .
the pulse profile shows only a single peak .
the low phase has count rates consistent with zero but seems to last longer in the _ chandra _ observation ( for about 25% of the period ) than in the observations ( for about 10% of the period ) .
there is no significant variation of the pulse profile with energy .
considering its low x - ray luminosity and hard x - ray spectra , this source is probably a magnetic cataclysmic variable ( cv ) , i.e. , a polar or a intermediate polar . with the period and the pulse profile resembling those of some short - period one - pole polars ( e.g. , v347 pav and gg leo , * ? ? ?
* ) , source # 150140 is most likely one such system .
if it is really a cv , it is probably not in ngc 6231 , which seems too young to form a cv .
this is also supported by the larger column density inferred from our spectral fitting than that of ngc 6231 ( ( 23)@xmath89 @xmath90 , * ? ? ?
source # 122979 has a galactic latitude of 30.2@xmath91 and thus is probably nearby if it is in our galaxy .
it has two observations separated by about half a year , with @xmath8=1.7 .
the second observation has a very large off - axis angle of 17.4@xmath92 , and the source is only in the fov of pn .
the maximum 0.310 kev absorbed luminosity , from the first observation , is about 3@xmath84@xmath85 erg s@xmath5 , assuming a source distance of 1 kpc .
we found a possible @xmath02.1 hr periodicity ( figure [ fig:122979foldcurve ] ) , which , however , needs to be confirmed with longer observations .
the source might be a magnetic cv , considering its probably low luminosity and hard x - ray spectra .
we also found a handful of poorly studied very soft sources .
such sources have been observed in our milky way , the magellanic clouds , and nearby galaxies ( for a recent review , see * ? ? ?
they are often referred to as supersoft ( bb temperature @xmath93 kev ) or quasisoft ( @xmath94 kev ) x - ray sources .
they most probably represent a diverse group . if they are coincident with galactic nuclei , they could be ultrasoft agns or tdes ( e.g. , * ? ? ?
* ; * ? ? ?
if they are off - nuclear and have luminosities of @xmath95@xmath3 erg s@xmath5 , they could be nuclear burning of material accreted by a wd @xcite , supernova remnants ( e.g. , * ? ? ?
* ) , central stars in planetary nebulae ( e.g. , * ? ? ?
* ) , or hot stellar core remnants from tidal stripping of giants by smbhs @xcite . those with luminosities below @xmath7 erg s@xmath5 could be thermally cooling isolated nss ( e.g. , * ? ? ?
* ) , ns x - ray binaries in quiescence ( e.g. , * ? ? ?
* ) , or wds ( e.g. , , which could be a binary @xcite ) .
very soft sources with luminosities above the eddington limit of wds ( @xmath02@xmath84@xmath3 erg s@xmath5 ) , such as ultraluminous off - nuclear x - ray sources ( ulxs , @xmath96@xmath97 erg s@xmath5 ) , have two popular explanations : cool thermal disk emission from accreting intermediate - mass bhs ( imbhs , @xmath0@xmath98@xmath99 , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) and optically thick outflows from supercritical accretion viewed at high inclinations @xcite . the spectral behavior in these two scenarios is expected to be different :
the former should have the disk luminosity following the @xmath100 relation , and the latter should have the observed outflow luminosity decreasing slightly with increasing temperature @xcite .
however , neither trend has been generally observed , and sometimes there is strong debate ( e.g. , * ? ? ?
* ; * ? ? ?
we note that the supercritical accretion explanation for ultralumineous very soft sources might also need to involve imbhs , instead of stellar - mass bhs .
this is because edge - on viewing is required to have very soft spectra , by viewing the cool outflow only ( not the central hot region ) , but at high inclinations , it is hard to achieve super - eddington luminosities due to self - absorption and lack of beaming effect @xcite .
we include seven very soft sources here , all with @xmath101 kev ( see below ) and at high galactic latitudes ( @xmath102 for source # 65459 and @xmath103 for the others ) .
lwb12 selected out these sources for their low hardness ratios unseen in most agns ( table [ tbl : dercat ] ) . to compare more with agns , we also calculated the strength of soft excess @xmath104 , defined as the ratio of unabsorbed 0.32 kev fluxes in the cool and hard components @xcite , using the mcd+pl model ( the bb+pl model gave comparable results ) . because our sources have no significant hard x - ray emission , the pl component is generally not needed , in which case we fixed @xmath105 at a value of 2 often seen in agns ( lwb12 ) and calculated the 90% lower limit of @xmath104 .
our sources have @xmath106 ( see below ) , while most agns have @xmath107 @xcite .
we first look at two sources , # 107102 and # 65459 , which are not associated with any galaxy in the third reference catalog ( rc3 , * ? ? ?
source # 107102 is coincident with an optical source from the sloan digital sky survey ( sdss , release 8 , table [ tbl : sdssmat ] ) , which appears slightly extended . its magellan optical spectrum taken by @xcite showed only narrow emission lines ( no broad h@xmath108 or h@xmath109 lines ) that are consistent with a low - mass ( @xmath99 ) type 2 seyfert nucleus at a redshift of 0.11871 ( a luminosity distance of 532 mpc ) .
the source was detected in all its three observations , with @xmath8=2.1 .
their spectra are all very soft ( @[email protected] kev or @[email protected] kev , table [ tbl : spfit ] , figure [ fig : srcspec ] , @xmath112 ) .
the 0.310 kev absorbed luminosity reached a peak of 3.6@xmath8410@xmath113 erg s@xmath5 in the second observation ( 0306630101 ) .
remarkably , the source showed large coherent oscillations at a period about 3.8 hr in the two observations in 2005 december but not in the observation in 2003 july ( figure [ fig:107102foldcurve ] ) .
however , the minima and maxima of the light curves in the two observations in 2005 do not seem to be well in phase ( figure [ fig:107102foldcurve ] ) .
therefore we have probably detected a quasi - periodic oscialtion ( qpo ) .
the folded light curves in two energy bands ( 0.20.8 kev and 0.82 kev ) in figure [ fig:107102foldcurve ] indicate that the modulation seems larger at higher energies . considering the strong fast variability and little neutral absorption in x - rays , the lack of h@xmath108 and h@xmath109 lines in the magellan optical spectrum taken by @xcite should indicate the real absence of the broad line region , instead of being hidden ( i.e. , not a standard type 2 agn ) , similar to @xcite .
we have devoted a separate work @xcite to carry out detailed studies of the source , including formal calculation of the qpo significance by carefully modeling the red noise ( the 99.9% upper limits ( dashed lines ) in figure [ fig:107102foldcurve ] account for only the poisson noise but not the red noise ) and the spectral fits with more physical models .
we refer to this work for more discussion on the nature of the source .
we note that the source was also discovered independently by @xcite , who reported similar spectral and timing properties of the source as we show above .
one main difference is that we show that the qpo is strong in both observations in 2005 december , not just in the first one as claimed by @xcite .
source # 65459 was detected in one observation on 2005 september 30 , but not in an earlier observation on 2003 december 19 , with @[email protected] ( figure [ fig : srcltlc ] ) .
the detection has @[email protected] kev or @[email protected] kev ( table [ tbl : spfit ] , figure [ fig : srcspec ] , @xmath118 ) .
it shows some short - term variability ( figure [ fig : srclc ] ) .
the source has no counterpart from the usno - b1.0 catalog or the 2mass psc ( table [ tbl : dercat ] ) .
it was not observed by the sdss .
future deep optical imaging is needed to check whether it is associated with any galaxy to help to pin down its nature .
we next look at two sources , # 101369 and # 101335 .
they have angular separations from the center of the galaxy ngc 4151 of @xmath119=7.1@xmath92 and 6.6@xmath92 , respectively .
they were not considered to be associated with this galaxy by lwb12 , because their ratios of the angular separation to the @xmath120 isophote elliptical radius @xmath121 are 2.8 and 2.5 , respectively , larger than the threshold value of 2 used to determine the association .
considering that ngc 4151 in fact has outer arms observed to extend out to 6@xmath92 @xcite , it is still possible that both sources are in this galaxy .
alternatively , they could be associated with distant galaxies , considering that both of them have sdss optical counterpart candidates ( table [ tbl : sdssmat ] ) .
source # 101369 has eight observations spanning nearly six years but was only detected in the first three observations which are continuous in time around 2000 december 22 and have very similar spectra , resulting in @xmath8@xmath965 ( figure [ fig : srcltlc ] ) .
the x - ray spectra are supersoft ( @[email protected] kev , table [ tbl : spfit ] , figure [ fig : srcspec ] , @xmath122 ) .
the maximum 0.310 kev absorbed luminosity is 1.3@xmath8410@xmath123 erg s@xmath5 if it is in ngc 4151 ( 20.3 mpc , * ? ? ?
the sdss counterpart candidate of the source is slightly extended ( the @xmath124-band petrosian radius is @xmath015 ) and has a photometric redshift of [email protected] , corresponding to a luminosity distance of 790 mpc .
the maximum 0.310 kev absorbed luminosity of the source is 2.0@xmath8410@xmath113 erg s@xmath5 using this distance . source # 101335 was detected in all its eight observations spanning nearly six years , with @[email protected] ( figure [ fig : srcltlc ] ) .
we note that the source was at the ccd edge in some observations , which could affect the calculation of the long - term variability .
the source was also detected in one _ chandra _ detection in the csc ( figure [ fig : srcltlc ] ) and in several occasions by _ rosat _ hri and pspc in the early 1990s with similar fluxes @xcite , indicating the persistence of the source over nearly two decades .
the fits with an absorbed pl to band count rates in observations indicate that the source spectra were generally soft except the latest observation 0402660201 on 2006 november 29 ( lwb12 ) .
we analyzed this observation and another observation 0112830201 on 2000 december 22 in detail ( tables [ tbl : obslog ] and [ tbl : spfit ] , figures [ fig : srclc ] and [ fig : srcspec ] ) .
the second observation has a higher s / n , and it can not be fitted well with a bb , while the fit with a mcd is acceptable , with @xmath126 kev ( table [ tbl : spfit ] , @xmath127 ) .
the observation 0112830201 is fainter and shorter and has a low s / n , but the fit with a mcd requires a significantly higher temperature @xmath128 kev , and a pl fit gives @xmath105=2.1 , which probably indicates a state transition . if the source is in ngc 4151 , the 0.310 kev absorbed luminosity reached 2.3@xmath8410@xmath123 erg s@xmath5 and 1.7@xmath8410@xmath123 erg s@xmath5 in the soft observation 0112830201 and the hard observation 0402660201 .
the sdss counterpart candidate of the source is faint and appears blue and point - like ( table [ tbl : sdssmat ] ) .
therefore , if they are in ngc 4151 , both sources # 101369 and # 101335 would be very soft ulxs and could be due to cool thermal disk emission from accreting imbhs or outflows from supercritical accretion .
for source # 101335 , there is possible state transition without large variation in luminosity , which could possibly be explained more easily with the former model .
if the sdss sources are their counterparts in this galaxy , these counterparts would have absolute @xmath124-band magnitudes about @xmath129 and @xmath130 , respectively , which are very bright and could be very massive star clusters . alternatively , these sdss sources are still their counterparts but are distant galaxies
. then sources # 101369 and # 101335 could be caused by nuclear activity .
considering its possible transient nature , sources # 101369 could be a tde .
for source # 101335 , whose flux was relatively constant over nearly two decades , it could be an ultrasoft agn similar to @xcite .
the other three very soft sources that we analyzed are sources # 8085 , # 183876 , and # 187831 , which lwb12 found to be likely within some rc3 galaxies .
source # 8085 was detected on the outskirts of ngc 300 ( @xmath121=1.1 ( table [ tbl : dercat ] ) ) in all its four observations spaning 4.9 years , but exhibiting large long - term variability ( @xmath8=36 , figure [ fig : srcltlc ] ) .
clear spectral change was observed , with @xmath110 varying from 0.07 to 0.12 kev ( table [ tbl : spfit ] , @xmath131 ) .
its maximum 0.310 kev absorbed luminosity is 1.2@xmath8410@xmath132 erg s@xmath5 , assuming a distance of 2 mpc @xcite .
source # 183876 was detected toward seyfert 1 galaxy ngc 7314 ( @xmath121=0.8 ) in two observations separated by 6 years , with no significant spectral variation observed ( @[email protected] , figure [ fig : srcltlc ] ) .
the spectral fitting results of one observation ( on 2006 may 4 ) are given in table [ tbl : spfit ] and shown in figure [ fig : srcspec ] .
the spectrum has @[email protected] kev or @[email protected] kev ( @xmath122 ) . although we added a weak pl in the fit , this component can be due to contamination from the nucleus .
there are also two detections in the csc about one month after the second observation , with comparable fluxes ( figure [ fig : srcltlc ] ) . assuming a distance of 12.9 mpc @xcite , its 0.310 kev absorbed luminosity is about 8.6@xmath8410@xmath132 erg s@xmath5 .
source # 187831 was detected in the second of its three observations of seyfert 2 galaxy ngc 7582 ( @xmath133=0.9 , @[email protected] , figure [ fig : srcltlc ] ) .
the detection has @[email protected] kev or @[email protected] kev ( table [ tbl : spfit ] , figure [ fig : srcspec ] , @xmath134 ) .
the light curve exhibits significant variability ( figure [ fig : srclc ] ) . assuming a distance of 17.6 mpc @xcite , we obtained a 0.310 kev absorbed luminosity of 1.1@xmath8410@xmath123 erg s@xmath5 , making it a ulx .
we note that this source is also close to another galaxy at a redshift of 0.05814 @xcite , at a separation of 46 ( corresponding to 5 kpc ) .
thus it is possible that source # 187831 is in this galaxy instead , which would increase its 0.310 kev absorbed luminosity to be 2.3@xmath8410@xmath135 erg s@xmath5 ( assuming a luminosity distance of 249.6 mpc ) .
sources # 183876 and # 187831 have maximum luminosities well above the eddington limit of wds ( @xmath02@xmath84@xmath3 erg s@xmath5 ) and are probably due to cool thermal disk emission from imbhs or outflows from supercritical accretion .
source # 8085 could have a similar origin , instead of nuclear burning of material accreted by a wd , for its relatively high luminosity and bb temperature ( 0.12 kev ) in one observation . among these three sources ,
only source # 8085 showed clear spectral variation , which could be used to differentiate between the thermal disk and outflow models .
both the bb and mcd fits suggest significant temperature variation at a relatively constant emission area ( table [ tbl : spfit ] ) , which seems to favor the thermal disk model , but this should be confirmed with future high - quality observations .
we analyzed six sources that appear in nearby galaxies and show large long - term variability .
because of their high peak luminosities ( @xmath136 erg s@xmath5 , see below ) and the lack of very hard spectra expected for accretion - powered x - ray pulsars ( lwb12 ) , they are most likely accreting bhs or weakly magnetized nss ( they are mostly low - mass x - ray binaries ) if they are indeed extragalactic ( cvs have luminosities @xmath137 erg s@xmath5 when they are not in the supersoft phase @xcite ) .
the differentiation between accreting bhs and accreting weakly magnetized nss is generally difficult , but hints can still be obtained in some cases .
observation of luminosities much higher than the eddington limit for accreting nss ( 1.8@xmath8410@xmath132 erg s@xmath5 assuming a 1.4- ns and the pure hydrogen accreting material ) and/or observation of soft spectra of @xmath138 kev at high luminosities will favor the accreting bh scenario @xcite , because accreting weakly magnetized nss , even when they are in the soft state , seldom exhibit so soft spectra due to presence of the hot emission from the impact of accreting materials onto the ns surface @xcite .
ns x - ray binaries can appear very soft only near quiescence ( @xmath13910@xmath140 erg s@xmath5 , * ? ? ?
we first look at four sources ( # 7204 , # 29251 , # 105130 and # 241580 ) that are probably bh x - ray binaries .
source # 7204 was detected in one of nine observations of ngc 253 ( @xmath121=0.74 , @xmath114@xmath9623 , figure [ fig : srcltlc ] ) .
the spectrum was fitted with an absorbed pl with @xmath105=2.9 by @xcite .
we reanalyzed the data and obtained a consistent fit ( table [ tbl : spfit ] ) .
considering the high value of @xmath105 , we also fitted the spectra with a mcd plus a weak pl ( @xmath105 was fixed at a value of 2 due to its large uncertainty ) .
the best - fitting disk temperature is about 0.5 kev , and the 0.310 kev absorbed luminosity is 0.86@xmath8410@xmath132 erg s@xmath5 if a distance of 3.0 mpc is assumed @xcite .
source # 29251 was detected in two of fourteen observations of ngc 1313 ( @xmath121=0.18 , @xmath8@xmath9630 , figure [ fig : srcltlc ] ) .
these two detections , separated by only 76 days , are probably from the same outburst .
the fluxes and spectral shapes are similar in these two detections .
we fitted the spectrum of one observation using the mcd model ( table [ tbl : spfit ] , figure [ fig : srcspec ] ) and obtained @xmath117=0.8 kev and a 0.310 kev absorbed luminosity of 4.0@xmath8410@xmath132 erg s@xmath5 , assuming a distance of 3.7 mpc @xcite .
we note that an absorbed pl with @xmath141 can also fit the spectrum ( table [ tbl : spfit ] ) .
the source was also detected by _
rosat _ , with a slightly higher luminosity than we report here @xcite , indicating the recurrent nature of the outbursts .
source # 105130 was detected in three of five observations on the outskirts of ngc 4395 ( @xmath121=1.65 , @xmath114@xmath9665 ) .
we analyzed the brightest observation , taken on 2004 july 3 , and found that the mcd model ( @xmath117=0.6 kev ) fitted the spectrum better than the pl and bb models , with the @xmath82 values decreased by 39 and 181 for 138 degrees of freedom , respectively ( table [ tbl : spfit ] , figure [ fig : srcspec ] ) .
no significant short - term variation is seen in this observation ( figure [ fig : srclc ] ) . assuming the source to be in ngc 4395 at a distance of 3.6 mpc @xcite
, we obtained the maximum 0.310 kev absorbed luminosity of 1.9@xmath8410@xmath132 erg s@xmath5 .
the source has two detections in the csc ( figure [ fig : srcltlc ] ) , which , in addition to observations , seem to suggest at least three outbursts within four years .
the source was also detected in one _ rosat _
hri observation 1rh702725n00 in 1996 ( 2.3@xmath142 erg s@xmath5 , * ? ? ?
source # 241580 was detected in two observations in 2008 may june in the direction of ngc 4490 ( @xmath121=0.87 ) , but not in another observation in 2002 may ( @xmath8@xmath9616 ) .
the spectrum of the brightest observation can be fitted with an absorbed pl ( table [ tbl : spfit ] , figure [ fig : srcspec ] ) , with the corresponding 0.310 kev absorbed luminosity of 7.6@xmath8410@xmath132 erg s@xmath5 if a distance of 7.8 mpc is assumed @xcite .
we see that sources # 7204 , # 29251 , and # 105130 have soft spectra with @xmath143 kev and that sources # 29251 , # 105130 and # 241580 have maximum 0.310 kev absorbed luminosities above the eddington limit for accreting nss .
therefore , these four sources are probably bh x - ray binaries .
we next look at two sources ( # 222148 and # 222243 ) in m 31 .
source # 222148 was detected in five of eighteen observations of m 31 ( @xmath121=0.15 , @xmath8@xmath96152 ) .
there are also eleven detections in the csc .
all these and _ chandra _ detections spanned about 407 days from 2007 november to 2008 december ( figure [ fig : srcltlc ] ) .
the outburst was discovered by @xcite from a 5ks _ chandra _ observation on 2007 july 31 , indicating that the outburst actually began earlier . to check
whether the outburst is still ongoing , we analyzed 19 new observations that were not included in the 2xmmi - dr3 catalog but publicly available as of 2012 august ( so covering data before 2011 july ) .
the spectral fits using an absorbed pl for these 19 observations , in addition to the five detections included in the 2xmmi - dr3 catalog , are given in table [ tbl : spfit ] .
the results show that the source has been x - ray bright for more than four years ( this is also supported by @xcite , who recently analyzed _ chandra _ hrc observations of the source ) .
the fluxes varied strongly at the beginning of the outburst and became fairly constant later .
the spectra are consistently hard , with @[email protected] .
one sample spectral fit is shown in figure [ fig : srcspec ] , using observation 0672130101 , which has the most counts .
the maximum 0.310 kev absorbed luminosity is 6@xmath8410@xmath144 erg s@xmath5 , and the average is around 4@xmath8410@xmath144 erg s@xmath5 ( assuming a distance of 780 kpc , * ? ? ?
although we found no optical and ir counterparts from the usno - b1.0 catalog and 2mass psc , there is an optical source in @xcite at a separation of 18 , which has a @xmath145-band magnitude of [email protected] and colors of @xmath146 , @xmath147 and @xmath148 and is probably a late - type dwarf star .
there have been some x - ray binaries known to remain x - ray bright most of the time since they were discovered to begin the outburst , e.g. , the bh x - ray binary candidates ( since 2005 may , * ? ? ?
* ) and ( since 1992 august , * ? ? ?
* ) and the accretion - powered millisecond x - ray pulsar ( since 2005 june , * ? ? ?
is essentially in the hard state , with 210 kev luminosity mostly around 5@xmath8410@xmath149 erg s@xmath5 @xcite .
has a similar mean luminosity ( 4.4@xmath84@xmath150 erg s@xmath5 in 2.525 kev , * ? ? ?
* ) and is mostly ( but not always ) in the hard state based on the hardness ratios that we calculated from the _ rxte _ monitoring pointed observations ( following a method as described in @xcite ) .
exhibits complicate behavior and has shown several states ( including both the hard and soft / thermal states ) .
if source # 222148 is a bh x - ray binary , it should be in the hard state . to check whether the source could be an accreting weakly magnetized ns in the hard state ,
in which a bb component is normally present and contributes a few percent ( but @xmath13910% ) of the luminosity @xcite , we added a bb component in the fit to the observation 0672130101 .
we found that the spectral fit was not significantly improved ( the @xmath82 decreased by 2.6 from 63.0 while the degrees of freedom decreased by 2 from 74 ) , with the bb flux estimated to be @xmath13920% of the total ( 0.310 kev , absorbed ) at a 90%-confidence level .
thus , although we do not need the bb in the fits , we can not exclude its presence at the level expected for accreting weakly magnetized nss in the hard state .
however , the luminosity of accreting weakly magnetized nss in the hard state is normally lower than 10% of the eddington limit @xcite , while source # 222148 is slightly brighter ( by a factor of 2 or more if the luminosity is estimated over a broader energy band ) .
we note that we can not exclude the possibility that the source is in the soft state of an accreting weakly magnetized ns based on the spectral fit . to demonstrate this
, we also fitted the spectrum of observation 0672130101 with a model of mcd+bb , which are often used to fit the soft state of an accreting weakly magnetized ns @xcite and obtained a similar quality of the fit as that using an absorbed pl ( table [ tbl : spfit ] ) .
this is because the soft - state spectra of accreting weakly magnetized nss can still appear hard below about 7 kev .
future broad band observations with observatories such as _ nustar _ can break such a model degeneracy easily .
source # 222243 was detected in three of nineteen observations of m 31 .
it is only @xmath119=1.5@xmath92 ( @xmath121=0.02 ) away from the center of m 31 .
lwb12 measured @xmath8@xmath966 and @xmath151(0.2 - 12.0 kev)@xmath9610 , but these variation factors were most likely underestimated due to the bright diffuse emission in the galaxy center .
the csc indicates that the source was also detected in nine _ chandra _ observations from 2007 march to 2008 december , with 0.110 kev fluxes about 10@xmath152 erg s@xmath5 @xmath90 .
in fact , there are many more _ chandra _ observations of this source , indicating that the source is generally faint for 12 years @xcite . in comparison ,
the three observations in which the source was detected were made at the beginning of 2007 january , with 0.212.0 kev fluxes in the range of ( 1.412.2)@xmath8410@xmath153 erg s@xmath5 @xmath90 .
their spectra are all hard .
we analyzed the brightest observation 0405320801 and found clear residuals when we fitted the spectrum with a single component model ( i.e. , an absorbed mcd , bb , or pl ) .
we obtained an acceptable fit with a mcd+bb model ( table [ tbl : spfit ] and figure [ fig : srcspec ] ) , indicating that the source might be an accreting weakly magnetized ns in the soft state ( see discussion above for source # 222148 ) .
the 0.310 kev absorbed luminosity is 6.8@xmath8410@xmath144 erg s@xmath5 ( assuming a distance of 780 kpc , * ? ? ?
the source is probably in the globular cluster candidate bh16
( i.e. , pb - in7 , * ? ? ?
* ; * ? ? ?
sources # 142052 , # 165335 and # 174295 are at low galactic latitudes ( @xmath154 ) and have galactic longitudes of about @xmath155 , @xmath156 , and @xmath157 , respectively .
they are highly variable ( @xmath158 , figure [ fig : srcltlc ] , table [ tbl : dercat ] ) .
their x - ray spectra are generally hard and can be fitted with an absorbed pl ( when detected , table [ tbl : spfit ] , figure [ fig : srcspec ] ) .
they all have optical and ir counterparts found within 1.3@xmath81 from the usno - b1.0 catalog and the 2mass psc .
lwb12 classified them as compact object systems based on high x - ray - to - ir flux ratios , lack of flares , and large variability .
source # 142052 was detected in all four observations and one _ chandra _ observation ( obsid : 7287 , from the csc ) spanning five years , with @xmath8=23 ( figure [ fig : srcltlc ] ) .
it is probably , which was detected in the _ asca _ galactic plane survey @xcite .
thus the source is probably persistent but highly variable . in
the spectral fits to two long observations ( 0402910101 and 0560181101 ) , a gaussian fe line was detected above 3@xmath159 ( based on the normalization ) , with an equivalent width ( ew ) of about 1.0 kev ( table [ tbl : spfit ] ) .
the source peak luminosity is in the range between about 2@xmath8410@xmath160 and 10@xmath140 erg s@xmath5 ( 0.310 kev , absorbed ) if the source distance is between 18.5 kpc .
its optical counterpart is probably nsv 20407 , a variable star with the @xmath161-band magnitude ranging from @xmath162 to 16.7 @xcite , supporting to identify it as a cv . further considering that the magnetic cvs often exhibit the fe emission line @xcite ,
source # 142052 is probably one such system .
source # 165335 was detected in one of its two observations ( @xmath163 ) .
it was also detected in a _
observation in 2006 ( obsid : 6405 , from the csc ) and is source 23 in @xcite in the direction of the mrr 32 x - ray cluster .
the flux in this _ chandra _ observation is very low , resulting in a 0.57 kev flux variation factor of 167 , compared with the detection .
the 0.310 kev absorbed luminosity in the detection is within the range of 8@xmath84@xmath85 to 6@xmath84@xmath86 erg s@xmath5 if the source distance is between 18.5 kpc . considering that it appears in a star forming region , the bright observation , which last 32 ks and showed no obvious variability
, might be due to an extremely long ( probably several days ) stellar x - ray flare .
however , stellar x - ray flares with durations of days and flux increasing factors of more than one hundred are rarely seen .
flares lasting for days or more have been detected before from rs cvn stars ( e.g. , * ? ? ?
* ; * ? ? ?
* and references therein ) , but they have flux increasing factors of only a few tens .
lwb12 reported several flares ( e.g. , sources # 64367 and # 146561 ) with flux increasing factors of more than one hundred , but they remained bright for only a few ks . if it is not a star , it is most probably a cv or a very faint x - ray transient ( vfxt ) with a bh or ns as the accretor @xcite , considering the very low peak luminosity of the source .
vfxts are transient sources with peak outburst luminosities between @xmath7 and @xmath164@xmath150 erg s@xmath5 .
the lower limit of @xmath7 erg s@xmath5 is adopted in the definition so that they are not cvs but probably accreting bhs or nss .
it is not clear whether they can in fact have fainter outbursts .
most known vfxts were detected toward the galactic center from the intensive monitoring of the galactic center by _ xmm - newton _ and _ chandra _ @xcite .
our knowledge on this class of objects is still very limited .
source # 174295 was securely detected in an observation on 2004 november 9 , with a hard and highly absorbed spectrum as indicated by the fits with an absorbed pl ( table [ tbl : spfit ] , figure [ fig : srcspec ] ) , but it was hardly detected in the observations ten days before and ten days after it , indicating a short outburst .
the source has a candidate sdss point - like counterpart ( table [ tbl : sdssmat ] ) , which appears very red , probably caused by strong absorption .
the maximum 0.310 kev absorbed luminosity would be between 2@xmath84@xmath165 and @xmath7 erg s@xmath5 if the source distance is within 18.5 kpc . similar to our discussion for source # 165335 above , this source might be a cv , a vfxt with a bh or ns as the accretor , or a star with an extremely energetic stellar flare .
we have studied 18 sources from the 2xmmi - dr3 catalog that were poorly studied in the literature but show interesting properties of periodicity , very soft spectra and/or large long - term variability in x - rays in order to investigate their nature .
our findings can be summarized as follows : 1 .
two sources have been persistently detected with hard spectra ( @xmath83 ) .
one of them ( source # 150140 ) was detected on the outskirts of the young open cluster ngc 6231 and showed a 1.62 hr periodicity in six observations and one _ chandra _ observation separated by nearly four years .
there is also a @xmath02.1 hr period candidate in the other source ( # 122979 ) , but it needs to be confirmed with future long observations .
both sources are good magnetic cv candidates .
2 . seven very soft sources ( @xmath101 kev ) are probably in other galaxies , which would imply the luminosities ranging from about @xmath3 to @xmath4 erg s@xmath5
. some of them might be coincident with galactic nuclei and best explained as ultrasoft agns or tdes , while the others are probably off - nuclear and could be cool thermal disk emission from accreting imbhs or edge - on viewing of optically thick cool outflows from supercritical accretion .
one source ( # 107102 ) is an agn candidate and showed an intermittent qpo at about 3.8 hr , i.e. , in two observations in 2005 december but not in one observation in 2003 july .
3 . six highly variable sources with spectra harder than the above very soft sources appear in nearby rc3 galaxies and probably have luminosities above @xmath6 erg s@xmath5 , making them great candidates for extragalactic x - ray binaries .
one source ( # 222148 , in m 31 ) has remained x - ray bright and hard in 0.310 kev since it was discovered to enter an outburst on 2007 july 31 and thus is probably a new - born persistent source .
three highly variable hard sources appear at low galactic latitudes and have maximum luminosities below about @xmath7 erg s@xmath5 if they are in our galaxy .
they could be cvs , vfxts with a bh or ns as the accretor , or stars with extremely strong flares . although most of our sources still require future optical spectroscopic observations and/or long - term x - ray monitoring to finally pin down their nature , it is clear that our sample contains a variety of objects with all kinds of spectral behavior .
some of them belong to rare classes that still have relatively small numbers of sources known , such as very soft x - ray sources , very faint transient and new - born persistent sources , and our discovery of new objects for these classes is thus important for our understanding of them .
we found their properties mostly consistent with previous studies , but some new results were also obtained .
focusing on the seven very soft x - ray sources , we found the existence of both supersoft and quasisoft x - ray sources and the possibility that some of them are the same class of objects because supersoft and quasisoft spectra can be observed in a single source ( see source # 8085 ) , consistent with previous studies ( e.g. , * ? ? ? * ) .
however , we also observed some evidence that they displayed various kinds of spectral evolution behavior ( being transient or persistent with either small or large long - term variability and displaying state transition ) often seen in x - ray binaries .
it is not clear whether this is completely because our sample might include several classes of objects .
our discovery of some unique objects is also significant , especially source # 107102 , as an agn candidate with very soft spectra and a strong qpo in x - rays @xcite .
there is only one other agn known to show similarly weak hard x - ray emission ( , @xcite ) and only two other galactic nuclei with qpos significantly detected ( , @xcite ; , @xcite ) .
acknowledgments : we thank the anonymous referee for the helpful comments .
we acknowledge the use of public data from the _ chandra _ and data archives , the 2xmm serendipitous source catalog , constructed by the xmm - newton survey science center on behalf of esa , and the _ chandra _ source catalog , provided by the _
chandra _ x - ray center as part of the _ chandra _ data archive . , m. w. , pivovaroff , m. , baganoff , f. , et al . 1998 , in society of photo - optical instrumentation engineers ( spie ) conference series , vol .
3444 , society of photo - optical instrumentation engineers ( spie ) conference series , ed .
r. b. hoover & a. b. walker , 210224 , g. , de vaucouleurs , a. , corwin , jr .
, h. g. , et al .
1991 , third reference catalogue of bright galaxies , ed .
de vaucouleurs , g. , de vaucouleurs , a. , corwin , h. g. , jr . ,
buta , r. j. , paturel , g. , & fouque , p. , d. k. , morgan , e. h. , & chakrabarty , d. 2008 , in american institute of physics conference series , vol . 1068 , american institute of physics conference series , ed .
r. wijnands , d. altamirano , p. soleri , n. degenaar , n. rea , p. casella , a. patruno , & m. linares , 5562 r|c|c|cc|cc|cc|c|cc & & @xmath166 & @xmath167 & @xmath168 & & & @xmath169 & @xmath170 & @xmath171 & @xmath172 & @xmath173 + & & @xmath174 & & & & & @xmath175 & @xmath176 & @xmath177 & @xmath178 & @xmath179 + & & @xmath180 & & & @xmath181 & @xmath182 & & & @xmath183 & @xmath184 & @xmath185 + & & @xmath186 & @xmath187 & @xmath188 & & & & & @xmath189 & @xmath190 & @xmath191 + & & @xmath192 & & & @xmath193 & @xmath194 & & & & @xmath195 & @xmath196 + & & @xmath197 & @xmath198 & @xmath199 & & & & & & @xmath200 & @xmath201 + & & @xmath202 & @xmath203 & @xmath204 & & & & & @xmath205 & @xmath206 & @xmath207 + & & @xmath208 & & & & & @xmath206 & @xmath209 & @xmath210 & @xmath211 & @xmath212 + & & @xmath213 & & & @xmath214 & @xmath215 & & & @xmath216 & @xmath217 & @xmath217 + & & @xmath218 & @xmath219 & @xmath220 & & & & & @xmath221 & @xmath222 & @xmath223 + & & @xmath224 & & & @xmath225 & @xmath226 & & & @xmath227 & @xmath228 & @xmath228 + & & @xmath224 & @xmath229 & @xmath230 & & & & & @xmath231 & @xmath232 & @xmath232 + & & @xmath233 & & & & & @xmath234 & @xmath235 & & @xmath236 & @xmath236 + & & @xmath233 & @xmath237 & @xmath238 & & & & & & @xmath239 & @xmath239 + & & @xmath240 & & & @xmath241 & @xmath242 & @xmath243 & @xmath244 & @xmath245 & @xmath246 & @xmath246 + & & @xmath240 & @xmath247 & @xmath248 & & & @xmath243 & @xmath249 & @xmath250 & @xmath246 & @xmath246 + 105130 & 0200340101 & @xmath224 & @xmath251 & @xmath252 & & & & & @xmath253 & @xmath254 & @xmath254 + & & @xmath255 & & & @xmath256 & @xmath257 & & & @xmath258 & @xmath259 & @xmath259 + & & @xmath260 & @xmath261 & @xmath262 & & & & & @xmath263 & @xmath264 & @xmath265 + & & @xmath255 & & & @xmath266 & @xmath267 & & & @xmath268 & @xmath269 & @xmath269 + & & @xmath270 & @xmath271 & @xmath272 & & & & & @xmath273 & @xmath274 & @xmath275 + & & @xmath276 & & & @xmath277 & @xmath278 & & & @xmath279 & @xmath280 & @xmath280 + & & @xmath270 & @xmath281 & @xmath282 & & & & & @xmath283 & @xmath284 & @xmath285 + & 0105261401 & & & & & & @xmath286 & @xmath287 & @xmath288 & @xmath289 & @xmath289 + & 0105261701 & & & & & & @xmath290 & @xmath291 & @xmath292 & @xmath293 & @xmath293 + & 0203910101 & @xmath294 & & & & & @xmath295 & @xmath296 & @xmath297 & @xmath298 & @xmath299 + & 0402910101 & @xmath300 & & & & & @xmath301 & @xmath302 & @xmath303 & @xmath304 & @xmath305 + & 0410581901 & @xmath306 & & & & & @xmath307 & @xmath308 & & @xmath309 & @xmath310 + & 0560181101 & @xmath311 & & & & & @xmath312 & @xmath313 & @xmath314 & @xmath315 & @xmath316 + & 0109490101 & & & & & & @xmath290 & @xmath317 & @xmath318 & @xmath319 & @xmath320 + & 0109490201 & & & & & & @xmath321 & @xmath322 & @xmath323 & @xmath324 & @xmath325 + & 0109490301 & & & & & & @xmath326 & @xmath327 & @xmath328 & @xmath329 & @xmath330 + & 0109490401 & & & & & & @xmath331 & @xmath332 & @xmath333 & @xmath334 & @xmath335 + & 0109490501 & & & & & & @xmath290 & @xmath336 & @xmath337 & @xmath201 & @xmath338 + & 0109490601 & & & & & & @xmath196 & @xmath339 & @xmath340 & @xmath341 & @xmath342 + & 6291 ( cxo ) & @xmath343 & & & & & @xmath344 & @xmath345 & @xmath346 & @xmath211 & @xmath347 + 165335 & 0149610401 & @xmath348 & & & & & @xmath349 & @xmath350 & @xmath351 & @xmath352 & @xmath353 + 174295 & 0200450301 & @xmath354 & & & & & @xmath355 & @xmath356 & @xmath357 & @xmath287 & @xmath358 + & & @xmath213 & & & @xmath359 & @xmath360 & @xmath361 & @xmath362 & @xmath363 & @xmath364 & @xmath365 + & & @xmath366 & @xmath367 & @xmath368 & & & @xmath369 & @xmath370 & @xmath371 & @xmath364 & @xmath372 + & & @xmath373 & & & @xmath256 & @xmath374 & & & @xmath375 & @xmath376 & @xmath377 + & & @xmath378 & @xmath379 & @xmath380 & & & & & @xmath381 & @xmath376 & @xmath382 + & 0505720201 & @xmath383 & & & & & @xmath307 & @xmath384 & @xmath385 & @xmath386 & @xmath387 + & 0505720301 & @xmath388 & & & & & @xmath389 & @xmath390 & @xmath391 & @xmath392 & @xmath393 + & 0505720401 & @xmath394 & & & & & @xmath395 & @xmath396 & @xmath397 & @xmath398 & @xmath398 + & 0505720501 & @xmath399 & & & & & @xmath400 & @xmath401 & @xmath318 & @xmath402 & @xmath402 + & 0505720601 & @xmath403 & & & & & @xmath404 & @xmath405 & @xmath406 & @xmath407 & @xmath408 + & 0551690201 & @xmath409 & & & & & @xmath410 & @xmath411 & @xmath412 & @xmath413 & @xmath414 + & 0551690301 & @xmath415 & & & & & @xmath416 & @xmath417 & @xmath418 & @xmath419 & @xmath420 + & 0551690401 & @xmath421 & & & & & @xmath422 & @xmath423 & & @xmath424 & @xmath212 + & 0551690501 & @xmath425 & & & & & @xmath426 & @xmath427 & @xmath428 & @xmath429 & @xmath430 + & 0551690601 & @xmath431 & & & & & @xmath432 & @xmath433 & & @xmath434 & @xmath435 + & 0600660201 & @xmath436 & & & & & @xmath437 & @xmath438 & @xmath439 & @xmath440 & @xmath441 + & 0600660301 & @xmath442 & & & & & @xmath443 & @xmath444 & @xmath445 & @xmath446 & @xmath447 + & 0600660401 & @xmath448 & & & & & @xmath449 & @xmath450 & @xmath451 & @xmath452 & @xmath453 + & 0600660501 & @xmath454 & & & & & @xmath455 & @xmath456 & @xmath457 & @xmath458 & @xmath459 + & 0600660601 & @xmath460 & & & & & @xmath461 & @xmath462 & @xmath463 & @xmath464 & @xmath465 + & 0650560201 & @xmath466 & & & & & @xmath467 & @xmath468 & @xmath469 & @xmath470 & @xmath471 + & 0650560301 & @xmath472 & & & & & @xmath473 & @xmath474 & @xmath475 & @xmath476 & @xmath477 + & 0650560401 & @xmath478 & & & & & @xmath479 & @xmath480 & @xmath481 & @xmath482 & @xmath483 + & 0650560501 & @xmath442 & & & & & @xmath389 & @xmath484 & @xmath485 & @xmath486 & @xmath487 + & 0650560601 & @xmath488 & & & & & @xmath437 & @xmath489 & @xmath490 & @xmath491 & @xmath492 + & & @xmath493 & & & & & @xmath494 & @xmath495 & @xmath496 & @xmath497 & @xmath498 + & & @xmath499 & @xmath237 & @xmath500 & @xmath501 & @xmath502 & & & @xmath503 & @xmath504 & @xmath505 + & 0672130601 & @xmath506 & & & & & @xmath494 & @xmath507 & @xmath508 & @xmath509 & @xmath498 + & 0672130701 & @xmath510 & & & & & @xmath301 & @xmath511 & @xmath512 & @xmath513 & @xmath498 + & 0672130501 & @xmath436 & & & & & @xmath514 & @xmath515 & @xmath516 & @xmath517 & @xmath518 + 222243 & 0405320801 & @xmath519 & @xmath520 & @xmath521 & @xmath522 & @xmath523 & & & @xmath524 & @xmath525 & @xmath526 + 241580 & 0556300101 & @xmath527 & & & & & @xmath528 & @xmath529 & @xmath530 & @xmath310 & @xmath531 rlccccccccccccccccc 101335 & j121028.95 + 391748.9 & 0.2 & 6 & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] + 101369 & j121034.99 + 393122.9 & 0.6 & 3 & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] + 107102 & j123103.24 + 110648.6 & 0.5 & 3 & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] + 174295 & j203353.67 + 410717.2 & 0.1 & 6 & [email protected] & [email protected] & [email protected] & [email protected] & [email protected] | we analyze 18 sources that were found to show interesting properties of periodicity , very soft spectra , and/or large long - term variability in x - rays in our project of classification of sources from the 2xmmi - dr3 catalog but were poorly studied in the literature , in order to investigate their nature .
two hard sources show x - ray periodicities of @xmath01.62 hr ( ) and @xmath02.1 hr ( ) and are probably magnetic cataclysmic variables .
one source is an active galactic nucleus ( agn ) candidate showing very soft x - ray spectra ( @xmath1 kev ) and exhibiting an intermittent @xmath03.8 hr quasi - periodic oscillation .
there are six other very soft sources ( with @xmath2 kev ) , which might be in other galaxies with luminosities between @xmath0@xmath3@xmath4 erg s@xmath5 .
they probably represent a diverse group that might include objects such as ultrasoft agns and cool thermal disk emission from accreting intermediate - mass black holes .
six highly variable sources with harder spectra are probably in nearby galaxies with luminosities above @xmath6 erg s@xmath5 and thus are great candidates for extragalactic x - ray binaries .
one of them ( , in m 31 ) is probably a new - born persistent source , having been x - ray bright and hard in 0.310 kev for at least four years since it was discovered to enter an outburst in 2007 .
three highly variable hard sources appear at low galactic latitudes and have maximum luminosities below @xmath0@xmath7 erg s@xmath5 if they are in our galaxy , thus great candidates for cataclysmic variables or very faint x - ray transients harboring a black hole or neutron star .
our interpretations of these sources can be tested with future long - term x - ray monitoring and multi - wavelength observations . |
one of recent developments in the study of galaxy formation is quantitative understanding of the dependence of galaxy properties on environment ( park et al . 2007 ; hwang & park 2009 among others ) .
it has already been noticed since the 1930 s that galaxy luminosity and morphology depend on the local density : the high density regions preferentially harbor more luminous and early morphological type of galaxies ( hubble & humason 1931 ) . in the beginning of the studies on the environmental effects on galaxy properties ( on the simulation side ,
see lemson & kauffmann 1999 ; gao et al .
2005 ; jing et al . 2007 ; hahn et al . 2007 ;
maccio et al .
2007 ; on the observation side , see blanton & berlind 2007 ; park et al .
2007 ; cervantes - sodi et al . 2008 ; skibba et al . 2009
; blanton & moustakas 2009 ) , the environment was distinguished according to the large - scale structure where the galaxies under study are located .
for example , comparative studies for galaxies located within massive galaxy clusters , groups , or voids were carried out ( oemler 1974 ; dressler 1980 ; postman & geller 1984 ; rojas et al .
2004 among many others ) .
another trend of the same kind of study used continuous parameters that measure the local galaxy number density .
various kinds of smoothing kernel were used to estimate the local number density from galaxy positions in redshift space .
the most popular one is the truncated cylindrical cone which is motivated from the fact that massive clusters appear as fingers of god in redshift space ( hogg et al .
2004 ; kauffmann et al . 2004 ; kuehn & ryden 2005 ; reid & spergel 2009 ) .
the gaussian and spline filters are also often used .
these filters can have a fixed size or can vary in size to include a fixed number of galaxies ( park et al .
1994 ; monaco & efstathiou 1999 ; park et al .
2007 ; kim et al . 2008 ) .
even though these two approaches look quite different , they essentially use the local galaxy number density to distinguish among different environments .
extension of the concept of environment beyond the ` local number density ' has been started by several authors .
park , gott & choi ( 2008 ) and park & choi ( 2009 ) used the mass density due to the galaxy plus dark halo systems as a new environmental parameter .
they also divided the ` local ' density into the large - scale background mass density and the small - scale density attributed to the nearest neighbor galaxy .
the mass density is estimated from galaxy luminosity and mass - to - light ratios .
it turned out that the environment set up by the nearest neighbor was critically important in determining galaxy properties .
lee et al .
( 2009 ) used galaxy luminosity density and local color ( difference between luminosity densities in two bands ) as environmental parameters .
lee & lee ( 2008 ) inspected the relation between the ellipticity of the tidal shear and galaxy morphology .
it is expected that some galaxy properties depend on the local galaxy number / mass density sensitively .
color and recent star formation activity may be such properties .
however , the root - mean - square ( rms ) displacement of mass is about @xmath2mpc till the present epoch in the @xmath0cdm model best fit to the wilkinson microwave anisotropy probe ( wmap ) 3-year data ( park & kim 2009 ) , and therefore the local density at the present location of galaxies can not fully represent the environment where galaxies formed and evolved .
this is true particularly for intermediate and high density regions since their sizes are typically only a few mpcs .
therefore , it may be useful to consider environmental parameters other than local density to understand the environmental effects on galaxy formation .
a theoretically motivated environmental parameter is the gravitational potential .
dark matter and baryons are expected to fall into the deep gravitational potential well to form massive objects .
since the gravitational potential field picks up the fluctuation power at scales much larger than those of the density field , the correlation between them will not be perfect . and it will be interesting to see how galaxy properties are related with the ` local ' gravitational potential . furthermore , it is expected that galaxy angular momentum is generated from the large - scale gravitational shear force . according to tidal torque theory ( hoyle 1951 ; peebles 1969 ; doroshkevich 1970 ; white 1984 ; lee & pen 2000 ; vitvitska et al .
2002 ; porciani et al 2002 ) , the origin of galactic angular momentum is originated by the tidal torque operating on primordial gas lump that will form a galaxy .
the torque is given by @xmath3 , where @xmath4 is a shear tensor generated by external material and @xmath5 is a moment of inertial tensor of material being torqued .
navarro et al .
( 2004 ) presented supporting evidence for the theory with direction of galaxy rotation axis .
an accurate estimation of the tidal shear tensor will enable one to verify if the tidal torque theory is really responsible for galaxy spin ( porciani et al .
2002 ; lee & pen 2002 ) .
in this paper we will study how accurately one can estimate the gravitational potential and its functions from a simulated sample of galaxies .
the error sources in this estimation are divided into three categories : \1 .
finite volume of the survey , \2 .
galaxy biasing , and \3 . redshift space distortion .
we then inspect the dependence of dark matter halo properties on various environmental parameters including the ` local ' gravitational potential .
it is hoped that a generalization of environmental parameter beyond the local density allow us to better understand galaxy formation and evolution .
we will examine how accurately one can estimate the gravitational potential and its derivatives when an observational sample is given .
for this purpose we use a set of mock galaxies identified as dark matter halos in an n - body simulation of the universe .
the simulation we use here adopted the cosmological parameters measured from the wmap 3-year data ( spergel et al . 2007 ) , which are @xmath6 , and @xmath7 , where @xmath8 are density parameters due to cosmological constant , matter , and baryon , respectively , @xmath9 is the slope of the primordial power spectrum , and @xmath10 is the rms fluctuation of mass in an @xmath11mpc radius spherical top hat .
it ran @xmath12 cold dark matter ( cdm ) particles whose initial conditions are generated on a @xmath12 mesh in accordance with the @xmath0cdm power spectrum .
the simulation was started at @xmath13 taking 1880 global time steps till the present epoch .
the physical size of the simulation cube is @xmath14mpc .
we used a parallel particle - mesh ( pm ) + tree n - body code ( dubinski et al . 2004 ) to increase the spatial dynamic range .
the gravitational force softening length is set to 50@xmath1kpc which is 0.1 times the mean separation between cdm particles .
the particle mass is @xmath15 .
we identified the gravitationally bound , tidally stable dark matter halos ( physically self - bound or psb halos ) from the cdm particle data at @xmath16 ( kim & park 2006 ) . at the first step the friend - of friend ( fof ) algorithm
is used to search for dark halos adopting the connection length of 0.2 times the mean particle separation .
then subhalos are identified within each fof particle group taking into account the gravitational binding energy with respect to the local density maximum and the tidal force from other more massive subhalos if they exist .
the minimum mass halos contain 30 member particles , and the halo mass function is accurate down to @xmath17 .
the mean separation between dark halos is @xmath18mpc .
the resulting dark matter halo sample consists of three types : isolated halo which does not overlap with other halo ( but it can have close neighbor halos ) , central halo which is the most massive halo in each group of halos , and satellite halo which is not the central one in each group of halos .
galaxies are identified with the halos .
this dark matter halo
galaxy one - to - one correspondence model describes the observed galaxy distribution quite accurately ( kim et al .
2008 ; gott et al . 2008 ; gott et al .
note that our dark halos are not the commonly used fof halos to which the halo occupation distribution prescription is usually applied to statistically distribute galaxies ( zheng et al .
2008 and references therein ) .
our dark halos can be central galaxies or satellites , which are the direct results of the n - body simulation . to match the dark halo sample with observed galaxy samples kim et al .
( 2008 ) adjusted the halo mass threshold , making the mean number densities of halos and galaxies the same .
the resulting halo sample was to be compared with a volume - limited sample of galaxies brighter than an absolute magnitude threshold . in the simulation
we know the positions of all cdm particles within the simulation cube with periodic boundaries .
therefore , the gravitational potential can be calculated through the poisson equation @xmath19 where @xmath20 is the mean density , @xmath21 is the overdensity , and we decomposed the gravitational potential and the density fields into fourier modes .
this true gravitational potential is going to be compared with those obtained by using galaxies ( i.e. dark halos ) . in practice
, we can not observe dark matter particles , but can observe only galaxies .
we adopt a simple method to obtain the gravitational potential and its derivatives from a galaxy redshift sample .
we assume galaxies are locally biased tracers of the underlying mass : @xmath22 , where @xmath23 is the bias factor .
then the poisson equation in the fourier space is @xmath24 the proportionality factor between @xmath25 and @xmath26 is not important in our study because each field will be normalized by its rms value . in section 3.2
we will compare the potential calculated from the galaxy distribution with that from the dark matter distribution . to use equations ( 1 ) and ( 2 ) we first calculate the density field from matter particle or galaxies .
we assign them on a cubical mesh using the triangular - shaped cloud algorithm ( hockney & eastwood 1981 ) , and fourier transform the density array to get @xmath27 or @xmath28
. the gravitational potential in real space is then obtained from equation ( 1 ) or ( 2 ) , and is interpolated at each location of galaxies .
the gradient of @xmath29 and the shear tensor are calculated by finite differencing @xmath29 in real space . near a galaxy at @xmath30 the gravitational potential can be given by the taylor expansion @xmath31 where @xmath32 are the components of local acceleration , and @xmath33 are the eigen values of the shear tensor @xmath34 .
the coordinate axes are assumed aligned with the principal axes of the shear tensor .
the trace of the shear tensor at the galaxy is @xmath35 bardeen et al .
( 1986 ) introduced the parameters @xmath36 characterizing the ellipticity and prolateness of a ( density ) distribution .
lee & lee ( 2008 ) adopted the @xmath37 parameter in their study on tidal shear dependence of galaxy morphology .
however , galaxies typically are not located at extrema of potential field unlike the density field at the smoothing scale we will adopt ( @xmath38mpc ) , and the sum of eigenvalues can be zero , making the above parameters undefined .
this is particularly true for large - scale density and potential fields where the galaxy - scale density peaks are not resolved . to avoid this problem we adopt the asymmetry and prolateness parameters @xmath39 normalized by their rms values .
these parameters are correlated with the local density through equation ( 4 ) .
dependence of galaxy properties on the shear tensor should be then studied at fixed local density . in section 3.3
we will study how the potential and its functions are biased because they are estimated from the galaxy distribution in redshift space .
it is shown that most of the redshift space distortion effects can be removed by making a linear estimation of the peculiar velocity from the redshift space galaxy distribution .
the linear regime peculiar velocity is calculated as follows .
the linearized continuity equation is @xmath40 where @xmath41 is the peculiar velocity in comoving space , @xmath42 is the density contrast @xmath43 , and @xmath44 is the linear growth factor . if we define the velocity potential @xmath45 by @xmath46 , we obtain @xmath47 combining the equations ( 7 ) and ( 8) yields @xmath48 the peculiar velocity is then given by @xmath49 where @xmath50 is the expansion parameter , @xmath51 , and @xmath52 is the hubble parameter . substituting equation ( 9 ) into equation ( 10 ) in fourier space ,
one finally obtains the relation between density and peculiar velocity @xmath53 we actually use a biased tracer of matter field , namely galaxies .
if the galaxy number density fluctuation is related with the matter fluctuation by @xmath54 , the factor @xmath55 in equation ( 11 ) is replaced by @xmath56 .
mpc ( top ) , from a half - size subcube ( middle ) , and from a quarter - size subcube ( bottom ) . shown
are constant shear magnitude contours at four threshold levels in half - size slices . ]
-component of the potential gradient .
( bottom ) same but for the ellipticity of the shear tensor . ] the gravitational potential would be correctly calculated if the mass distribution were given over the infinite space . in practice
, any observational sample covers a finite volume of the universe with complicated boundaries . in this section
we will study the effects of the missing data beyond the sample boundaries on gravitational potential and its derivatives .
we first calculate the reference potential from all dark halos in the whole simulation cube . since we are interested in large - scale environment , we smooth the density field by a @xmath57mpc gaussian .
this scale is also motivated by the fact that the mean separation between @xmath58 galaxies is about @xmath57mpc ( cf . the ` best ' sample of park et al .
2005 ; see also choi et al .
we then calculate a potential using only the dark halos in the half - size cube near a corner of the simulation box .
the remaining part of the simulation cube is replaced by the mean dark halo number density estimated from the half - size cube . to inspect the boundary effects further
, we also calculate a potential from the dark halos in the quarter - size cube at the same corner . in figure 1
we compare the three gravitational potentials in a slice passing through a @xmath59-coordinate of @xmath60mpc of the simulation cube .
the slice includes the center of the quarter cube .
the top panel shows the shear ellipticity field @xmath61 in the slice calculated from all dark halos in the simulation .
we show only the bottom left corner of the slice where a comparison with the half cube result is possible .
the contour levels correspond to @xmath62 high ( red contours ) , mean , @xmath63 low , and @xmath64 low ( black ) shear ellipticity . in the middle panel
the shear ellipticity field calculated from the dark halos in the half cube is shown .
the contour levels are the same .
it is clear that the shear field of the half cube is extraordinarily similar with that of the full cube .
this is particularly true near the center of the half cube , namely at the position ( 256 , 256 ) .
however , large differences are also observed near the boundaries .
the bottom panel shows the shear field of the quarter cube .
it again shows that the shear ellipticity of the quarter cube agree quite well with those of the other ones , particularly near the center of the quarter cube .
it can be seen that there is a weak shear field extending beyond the sample boundary .
the shear field of the half cube also has such leakage outside the boundaries ( truncated in the middle panel ) .
to quantitatively estimate the accuracy of the potential and its derivatives obtained from subcubes we calculate the rms differences of the fields within a cubical shell centered on the center of the subcubes .
the solid lines in the top panel of figure 2 are the difference between the potential from the full cube and that of the half cube within 2 @xmath1mpc - thick cubical shell centered at @xmath65 .
size of a cubical shell is 2(@xmath66 ) , where @xmath67 is the distance of the shell from the surface of the subcube .
for example , @xmath68mpc corresponds the outermost shell of 512@xmath1mpc size .
the dashed line is for the quarter cube .
the top panel shows how the error in the potential varies as one moves into the subcubes .
lines and error bars are the mean and rms differences obtained from eight subcube results .
the difference monotonically decreases as the distance from boundaries increases .
since potentials can be added by a constant without altering physics , the difference of about 0.14@xmath69 near the center can be ignored in the case of half cube .
the figure tells that , if the potential field is to be obtained with error less than say 20% of its rms value , one should stay more than @xmath70mpc from all sample boundaries .
the quarter cube results show a much larger variance that the half cube results .
a similar trend is seen for the potential gradient as shown in the middle panel of figure 2 .
the @xmath71-component of the potential gradient mimics the linear radial peculiar velocity field very far from an observer .
due to lack of data outside subcubes the error in gradient is large near boundaries but monotonically decreases as @xmath67 increases .
the difference is smaller if one ignores the large - scale gradient on the scale of the subcube .
the bottom panel of figure 2 compares among the shear ellipticity from the full , half , and quarter cubes .
it is important to note that the error drops as the distance from boundaries increases essentially in the same way for the half and quarter cube cases .
the error reaches 20% of the rms shear ellipticity at the distance of @xmath72mpc .
the depth of this buffer region must be a function of the amplitude and shape of the power spectrum .
we conclude that one needs to have an observational sample much larger than 60@xmath1mpc to secure the regions where the shear field has error less than 20% of its rms fluctuation .
-axis ) and those from dark halo number ( top ) and dark halo mass ( middle and bottom ) distributions .
dark halos with mass larger than @xmath73 are used in the top and middle panels .
the mass cut is raised to @xmath74 in the bottom panel . ] in the previous section we used the number density of dark halos to calculate the potential field as it is common to use the galaxy number density to define environment of galaxies .
however , it is the matter field that determines the true gravitational potential field , and one needs to understand the relation between the potential from mass and those from mass tracers . in this section we assume that each and every psb dark halo contains one galaxy , and compare the potential calculated from their number density or halo mass density with that from cdm particles in the simulation . in figure 3
we compare the potential from galaxy number density ( @xmath75-axis ) with the correct potential from cdm particles .
dark halos more massive than @xmath76 are used for the number density calculation .
their mean separation is @xmath77mpc , which is equal to that of the sdss galaxies brighter than @xmath78 ( choi et al .
the correlation coefficient between the two potential fields is fairly high ( @xmath79 ) . at @xmath80 the potential from the galaxy number density
has a dispersion of 0.20 times the rms potential value .
figure 4 compares between the magnitudes of the traceless shear tensor @xmath81 , estimated from matter density , halo number density , and halo mass density fields .
top panel shows that the correspondence between the shear fields from the galaxy number density and matter density is not so good .
the correlation coefficient is only 0.858 , and the dispersion of the shear magnitude obtained from galaxy number density is 0.47 times the rms shear magnitude at @xmath82 .
accuracy in potential and shear can be greatly improved if the halo mass is used to weigh galaxies and the halo mass density field , instead of the number density field , is used to calculate the potential .
the second and third panels of figure 4 demonstrate such improvement when the halo mass threshold is set to @xmath76 and @xmath83 , respectively .
the latter objects corresponds to the sdss galaxies brighter than @xmath84 , close to that of the @xmath58 galaxies ( choi et al .
2007 ) , in the sense that their mean separations ( @xmath85mpc ) are the same . in the middle panel , at @xmath86 , the dispersion in @xmath87 is only 0.17 times the rms shear magnitude , an improvement by almost a factor 3 .
the correlation drops as the halo mass threshold increases , but it still remains quite good for massive dark halos with @xmath88 .
figure 5 shows how the correlation changes as a function of the halo mass threshold .
the solid line is the case when the halo mass density field is used , and the dashed line is for the halo number density .
the correlation drops rapidly as the threshold increases above @xmath89m@xmath90 .
we conclude that the gravitational potential can be estimated quite accurately by using the observed galaxy distribution , but accuracy can be greatly improved if the total mass associated with the dark halo plus galaxy system is used to weigh galaxies in the shear tensor calculation .
when one chooses to use the mass field instead of the number density field , one needs to adopt a halo mass estimator .
we suggest to use the red - band optical luminosity together with the morphology - dependent mass - to - light ratios to estimate the relative mass of galaxies .
an example of using @xmath91-band luminosity as the mass estimator for dark halo plus galaxy systems , can be found in park & choi ( 2009 ) , where the method turned out to work quite well in the sense that galaxy properties show interesting dependence on local and global environments at physically meaningful scales .
note that we do nt need to know the absolute value of halo mass if parameters are normalized by their rms values as we do here . using any halo mass estimator monotonically proportional to the actual halo mass
will improve the accuracy of the resulting potential field .
when galaxy distance is obtained from redshift , the galaxy distribution becomes biased in such a way that clusters and groups are stretched , filaments appear more prominent by broadening the interior but compressing the exterior , and voids look elongated along the line of sight ( kaiser 1987 ) .
these redshift space distortion effects generate error when the gravitational potential is estimated directly from the galaxy distribution in redshift space .
we will show here that this error can be almost entirely removed by making a linear correction to the redshift space distribution of galaxies .
we calculate the peculiar velocities of dark halos from the dark halo number density or mass density in redshift space using equation ( 11 ) .
they are compared with the true peculiar velocities in figure 6 .
the upper panel shows the @xmath71-component of the peculiar velocity at each galaxy location calculated from halo number density , and the bottom one shows that from halo mass density .
it can be seen that the relation is a little tighter in the second case .
the top panel of figure 7 shows the correlation between the shear magnitudes estimated from halo number density and from matter density .
the correlation becomes stronger when the halo mass density is used for the reconstruction as shown in the middle panel .
the redshift space distortion effects become tiny when the redshift space positions of dark halos are corrected for the peculiar velocity , which is calculated from the redshift space distribution of dark halos .
this is shown in the bottom panel .
ideally one can iterate this correction process , but we found that one time correction was sufficient .
mpc . ] in previous sections we studied how accurately one can estimate the gravitational potential and its functions when only biased and finite samples are available .
now we will examine how various physical properties of dark matter halos depend on environment including gravitational potential or its functions .
the purpose is to see if their dependence on gravitational potential or its functions is different from that on local density .
so it is necessary to see the correlation between gravitational potential and local density at galaxy positions .
figure 8 shows the potential ( top ) , ellipticity ( middle ) , and prolateness of the shear tensor ( bottom panel ) as a function of overdensity .
each variable is normalized by its standard deviation .
it demonstrates that , even though there exists an overall correlation with the local density , the dispersion of the potential or shear tensor at fixed local density is very large .
this reflects the fact that the gravitational potential picks up large scale power ( see eq .
1 ) , and thus has information on environment independent of local density . in figure 9
we shows four physical parameters of dark matter halos as a function of local density , potential , ellipticity , and prolateness of the shear tensor .
points are the dark halos with mass greater than @xmath92m@xmath90 .
dashed and solid lines are the mean relations for dark halos with @xmath93 and @xmath94m@xmath90 , respectively .
only 1/800 of dark halos in the simulation are plotted , but the mean relations are obtained from 1/8 of dark halos . as expected , massive halos form in high density regions ( top - left panel ) because of high interaction and merger rates ( park , gott , & choi 2008 ; fakhouri & ma 2009 ) .
this plot very much resembles the local density dependence of galaxy luminosity obtained from the cfa survey ( fig .
12 of park et al . 1994 ) and the sdss survey ( fig . 8 of park et al .
it also tells that the maximum mass which a dark halo can acquire is a function of the local density . in low density regions massive halos
simply can not form by the present epoch .
a similar observation can be made from the panel showing the velocity dispersion of dark halos ( the left panel of the third row ) .
the mean halo mass is relatively higher in low potential regions or in high shear ellipticity and prolateness regions , but its dependence on these parameters is weaker than that on local density .
this dependence of halo mass on potential and shear is partly due to the correlation of potential and shear with local density .
the second row of figure 9 shows dependence of the spin parameter @xmath95 on the environmental parameters .
the spin parameter is defined by ( peebles 1969 , gardner 2001 , and for an alternative definition see , bullock et al .
2001 , shaw et al .
2006 , hetznecker & burkert 2006 ) @xmath96 where @xmath97 and @xmath98 are the mass , total angular momentum and energy of a dark halo , respectively .
contrary to other studies ( maccio et al .
2007 ) we detect dependence of @xmath95 on all four environmental parameters . for halos of @xmath99m@xmath90
the spin parameter is about 0.034 in very low density regions , but increases to about 0.039 in high density regions .
dark halos also tend to have higher spin in high shear magnitude regions .
however , it should be noted that dark halo spin is not sensitive to any of these environmental parameters . in particular , dependence of the spin on the tidal
shear is weak , showing only about 4% difference as the environment changes from low to high shear ellipticity regions ( solid line in the third panel of the second row of figure 9 ) .
the panels at the bottom of figure 9 show that dark halos are more spherical in high density , low gravitational potential , or high shear regions .
such a trend is stronger for more massive halos . on average ,
high - mass halos are preferentially more spherical than low - mass ones , and the difference is larger in high density / shear regions .
this is consistent with the observational finding of park et al .
( 2007 ) that early - type galaxies in the sdss survey are rounder as their luminosity increases ( see their figure 15 ) . however , it contrasts with an opposite finding in simulations by many authors ( avila - reese et al .
2005 ; allgood et al .
2006 ; maccio et al .
2007 , 2008 ) who reported that low - mass halos are more spherical than high - mass halos at a given epoch in their simulations .
the main reason for this discrepancy is the difference in the halo definition as shown below . to study the dependence of halo properties on environmental parameters in a more sensible way we divide our halo sample into subsets according to halo mass and type .
we consider three halo types : isolated , central , and satellite halos . since halos can have a different growth history depending on their type
, it will be interesting to see the halo property - environment relation separately .
we are also going to fix the halo mass to subtract the halo mass dependence of a parameter from its environmental dependence .
figure 10 shows the spin parameter and axis ratio for isolated ( top two rows ) , central ( middle two rows ) , and satellite ( bottom two rows ) halos .
dots in the top two rows are the isolated halos with @xmath100 and the long - dashed lines are the mean relations .
only 1/25 of halos are plotted .
the short - dashed lines are for the halos with @xmath101 , and solid lines are for the halos with @xmath102 .
we plot only the spin and @xmath103 parameters here because the velocity dispersion shows almost no dependence on environment in these subsamples with fixed halo mass ranges , where @xmath103 is the ratio of the minor and major radii in a triaxial shape ( chandrasekhar 1969 ; de zeeuw & franx 1991 ; jing & suto 2002 ; smith & watts 2005 ; allgood et al .
2006 ) . in the middle and the bottom two rows of figure 10
the points and lines correspond to the same mass ranges as in the top two rows but they are for central and satellite halos , respectively . a few observations can be made from figure 10 .
first , the dependence of @xmath95 and @xmath103 on the potential and shear is still very weak ( see the second , third , and fourth columns of figure 10 ) . when the parameters are studied as a function of local density , we find diverse relations that depend not only on halo mass but also on halo type .
an interesting finding is that the spin parameter of satellite halos tend to be constant or even decrease as the local density increases while that of isolated or central halos increases .
it seems that the spin of satellites has decreased as they interact more frequently with other halos including the central halo in higher density environment .
an observational evidence for this interpretation is that the spin of late - type galaxies decreases as they approach their neighbors within the virial radius ( cervantes - sodi et al .
there is a clear trend for the spin parameter to decrease as the halo mass increases for all halo types .
the spin and local density tend to be more positively correlated for more massive halos .
it is also found in figure 10 that halos are more spherical in higher density regions for all three types .
furthermore , the dependence of @xmath103 on local density is stronger for more massive halos .
following the same interaction picture , we interpret this phenomenon as a result of tidal interactions between neighboring halos .
halos in high density regions will suffer from the strong tidal force of neighboring halos more frequently , and can become more spherical on average .
our results are consistent with maccio et al .
( 2007 ) who found isolated and central halos are more spherical in high density regions when the density is smoothed by an @xmath104mpc tophat even though their main conclusion was independence of halo properties on large - scale background density .
the dependence of halo shape on halo mass is not monotonic .
figure 10 shows halos are more spherical as mass increases from @xmath105 to @xmath106 , then more elongated as mass increases to @xmath107 .
this result might seem dissonant with some previous reports that more massive halos are less spherical ( avila - reese et al . 2005 ; allgood et al . 2006 ; maccio et al . 2007 ) .
the disagreement may originate from difference in halo type and mass range .
avila - reese et al .
( 2005 ) and maccio et al .
( 2007 ) used the fof or spherical overdensity ( so ) algorithm to identify halos , respectively .
therefore , a halo in these works can be a group of halos connected or embedded with one another and can have multiple density peaks .
the massive halos located at the intersections of dense filaments are more likely to be connected with other halos along the filaments by tenuous bridges of particles and appear more elongated .
on the other hand , the halos in our work ( the psb halos ) have one and the only one density maximum , and are separated into isolated , central , and satellite halos depending on their positions and mass .
the psb halos are to be directly identified with galaxies individually ( kim et al . 2008 ) .
therefore , the shape of the psb halos does not necessarily the same as the fof or so halos as the halo shape depends on halo definition ( bett et al .
2007 ) . the correspondence of our result with the previous ones can be checked by measuring the shape of the fof halos that are identified in our halo finding before the step identifying the psb halos .
figure 11 shows the relations between @xmath103 and halo mass , and those between @xmath95 and @xmath103 for fof ( dots and filled circles ) , isolated ( open circles ) , central ( boxes ) , and satellite ( open triangles ) halos .
the fof halos are most spherical when their mass is near @xmath108 .
note that @xmath103 has no maximum for isolated halos and is increasing as the halo mass increases .
the spin is a monotonically decreasing function of @xmath103 for all types of halo .
figure 11 agrees very well with figure 12 and 13 of bett et al .
( 2007 ) .
the overall message of figure 10 seems that the internal physical properties of dark halos are mainly controlled by small - scale physics .
this is supported by the fact that halo properties depend most sensitively on the local density rather than on the gravitational potential .
it is also consistent with the finding of park & choi ( 2009 ) and park & hwang ( 2009 ) that galaxy properties in the general field and clusters are significantly affected by gravitational and hydrodynamic interactions with the nearest neighbor galaxy .
however , as figure 8 shows , the gravitational potential has information independent of that of the local density , and it is worth to explore if the parameters like the gravitational potential or local shear can tell something more about galaxy formation .
in this paper we demonstrated the gravitational potential and its functions can be reasonably accurately estimated from an observational sample that covers only a finite volume of the universe and suffers from tracer biasing and redshift space distortion .
we found that the error in the gravitational potential and shear tensor decreases rapidly as one moves inside the survey boundaries . in the case of shear tensor
the error becomes less than 20% of its rms value in the regions separated from the survey boundaries by more than about 30@xmath1mpc .
this requires the sample size to be much larger than 60@xmath1mpc for an environment study with accurate estimation of the potential and its functions .
our study also shows that the effects of halo biasing on the gravitational potential estimation can be greatly reduced by weighting dark halos ( or galaxies ) by their mass as was done by park et al .
( 2008 ) and park & choi ( 2009 ) in their studies of small and large - scale density dependence of galaxy properties .
accuracy in the estimation starts to fall down rapidly when the halo mass cut is larger than @xmath89m@xmath90i .
this means that the mass density and potential fields estimated from the distribution of the luminous red galaxies ( lrgs ) will be quite inaccurate because the mean separation of a volume - limited sample of the sdss lrgs is about 20@xmath1mpc ( gott et al .
2009 ) , which corresponds to the halo mass cut of about @xmath109 .
the error due to the redshift space distortion effects can be also reduced by using dark halo mass density in the estimation of potential . but even more reduction can be achieved by correcting the observed ( i.e. redshift space ) distribution of dark halos for the peculiar velocity .
it was sufficient to use the peculiar velocity linearly estimated from the redshift space distribution of dark halos . after making the peculiar velocity correction and using halo mass weight the error due to
the redshift space distortion becomes tiny .
we showed there exists large dispersion in the gravitational potential and the shear at fixed local density .
it demonstrates the potential has large - scale information independent of local density .
we inspected the dependence of dark matter halo properties on local density , gravitational potential , shear ellipticity and prolateness . among these environmental parameters
the local density shows the strongest correlation with the internal physical parameters of dark halos . when halo mass is fixed , the spin and shape parameters are nearly independent of the potential and shear tensor but depend sensitively on the background density in the case of massive halos , in particular . in the following paper we will analyze the main galaxy sample of the sdss dr7 catalog to examine the dependence of various galaxy properties on these environmental parameters
this will extend our understanding on the environmental effects on galaxy formation and evolution .
the authors thank the anonymous referee for valuable comments .
hbp thanks prof .
hyung - mok lee for useful comments .
jhk appreciates the support of the college of applied science , kyung hee university research professor fellowship in cas - rp - kjh2009 .
cbp acknowledges the support of the korea science and engineering foundation ( kosef ) through the astrophysical research center for the structure and evolution of the cosmos ( arcsec ) . , allgood , b. , flores , r. a. , primack , j. r. , kravtsov , a. v. , wechsler , r. h. , faltenbacher , a. , bullock , j. s. 2006 , mnras , 367 , 1781 avila - reese , v. , colin , p. , gottlober , s. , firmani , c. , & maulbertsch , c. 2005 , apj , 634 , 51 bett , p. , eke , v. , frenk , c. s. , jenkins , a. , helly , j. , & navarro , j. 2007 , mnras , 376 , 215 blanton , m. r. & berlind , a. a. 2007 , apj , 664 , 791 blanton , m. r. & moustakas , j. 2009 , 47 , 159 bullock j.s . ,
dekel , a. , kolatt , t.s . ,
kravtsov , a.v . ,
klypin , a.a . ,
porciani , c. , & primack , j. r. 2001 , apj , 555 , 240 cervantes - sodi , b. , hernandez , x. , & park , c. 2010 , mnras , in press cervantes - sodi , b. , hernandez , x. , park , c. , & kim , j. 2008 , mnras , 388 , 863 chandrasekhar , s. 1969 , _ ellipsoidal figures of equilibrium _ , yale univ .
press , new haven choi , y .- y . , park , c. , & vogeley , m. s. 2007 , , 658 , 884 doroshkevich , a. g. 1970 , astrofizika , 6 , 581 dressler , a. 1980 , , 236 , 351 dubinski , j. , kim , j. , park , c. , & humble , r. 2004 , new astronomy , 9 , 111 fakhouri , o. , & ma , c .- p .
2009 , arxiv0906.1196 gao , l. , white , s. d. m. , jenkins , a. , frenk , c. s. , & springel , v. 2005 , mnras , 363 , 379 gardner , j. p. 2001 , apj , 557 , 616 gott , j. r. , choi , y .- y . , park , c. , & kim , j. 2009,apj , 695 , l45 gott , j. r. , hambrick , d. c. , vogeley , m. s. , kim , j. , park , c. , choi , y .- y . , cen , r. , ostriker , j. p. , & nagamine , k. 2008 , , 675 , 16 hahn , o. , porciani , c. , carollo , c. m. , & dekel , a. 2007 , mnras , 375 , 489 hetznecker , h. , & burkert , a. 2006 , mnras , 370 , 1905 hoyle , f. 1951 , in problems of cosmical aerodynamics ( dayton : central air documents office ) , 195 hwang , h. s. & park , c. 2009 , , 700 , 791 hockney , r. w. & eastwood , j. w. 1981 , _ computer simulations using particles _ , mcgraw - hill , new york hogg , d. w. , et al .
2004 , , 601 , l29 hubble , e. & humason , m. l. 1931 , , 74 , 43 jing , y. p. & suto , y. 2002 , apj , 574 , 538 jing , y. p. , suto , y. , & mo , h. j. 2007 , apj , 657 , 664 kaiser , n. 1987 , mnras , 227 , 1 kauffmann , g. , et al .
2004 , mnras , 353 , 713 kim , j. & park , c. 2006 , , 639 , 600 kim , j. , park , c. , & choi , y .- y .
2008 , , 683 , 123 kuehn , f. & ryden , b. s. 2005 , , 634 , 1032 lee , j. , & lee , b. 2008 , apj , 688 , 78 lee , j. h. , lee , m. g. , park , c. , & choi , y .- y .
2010 , mnras in press lee , j. & pen , u .- l . 2000 , apj , 532 , l5 lee , j. & pen , u .- l . 2002 , apj , 567 , l111 lemson , g. & kauffmann , g. 1999 , mnras , 302 , 111 maccio , a. v. , dutton , a. a. , van den bosch , f. c. , & frank , c. 2008 , mnras , 391 , 1940 maccio , a. v. , dutton , a. a. , van den bosch , f. c. , moore , b. , potter , d. , & stadel , j. 2007 , mnras , 378 , 55 monaco , p. & efstathiou , g. 1999 , mnras , 308 , 763 navarro , j. f. , abadi , m. g. , & steinmetz , m. 2004 , , 613 , l41 oemler , a. 1974 , apj , 194 , 1 park , c. & kim , y .- r . 2009 , arxiv:0905.2268 park , c. & choi , y .- y .
2009 , , 691 , 1828 park , c. , gott , j. r. , & choi , y .- y . 2008 , , 674 , 784 park , c. , choi , y .- y . , vogeley , m. s. , gott , j. r. , & blanton , m. r. 2007 , apj , 658 , 898 park , c. , vogeley , m. s. , geller , m. j. , & huchra , j. p. 1994
, , 431 , 569 park , c. , gott , r. j. , & da costa , l. n. 1992 , 392 , l51 peebles , p. j. e. 1969 , apj , 155 , 393 porciani , c. , dekel , a. , & hoffman , y. 2002 , mnras , 332 , 325 postman , m. , & geller , m. 1984 , , 281 , 95 reid , b. & spergel , d. n. 2009 , , 698 , 143 rojas , r. r. , vogeley , m. s. , hoyle , f. , & brinkmann , j. 2004 , , 617 , 50 shaw , l. d. , weller , j. , ostriker , j. p. , & bode , p. 2006 ,
apj , 645 , 815 sheth , r. k. , et al .
2003 , , 594 , 225 skibba et al .
2009 , mnras , 399 , 966 smith , r. e. & watts , p.
i. r. 2005 , mnras , 360 , 203 spergel , d. , et al .
2007 , , 170 , 377 vitvitska , m. , klypin , a. a. , kravtsov , a. v. , wechsler , r. h. , primack , j. r. , & bullock , j. 2002 , apj , 581 , 799 white , s. d. m. 1984 , apj , 286 , 38 de zeeuw , t. & franx , m. 1991 , ara&a , 29 , 239 zheng , z. , zehavi , i. , eisenstein , d. j. , weinberg , d. h. , & jing , y. 2008 , arxiv:0809.1868 | we extend the concept of galaxy environment from the local galaxy number density to the gravitational potential and its functions like the shear tensor . for this purpose
we examine whether or not one can make an accurate estimation of the gravitational potential from an observational sample which is finite in volume , biased due to galaxy biasing , and subject to redshift space distortion .
dark halos in a @xmath0cdm simulation are used in this test .
we find that one needs to stay away from the sample boundaries by more than 30@xmath1mpc to reduce the error within 20% of the root mean square values of the potential or the shear tensor .
the error due to the galaxy biasing can be significantly reduced by using the galaxy mass density field instead of the galaxy number density field .
the error caused by the redshift space distortion can be effectively removed by correcting galaxy positions for the peculiar velocity effects .
we inspect the dependence of dark matter halo properties on four environmental parameters ; local density , gravitational potential , and the ellipticity and prolateness of the shear tensor .
we find the local density has the strongest correlation with halo properties .
this is evidence that the internal physical properties of dark halos are mainly controlled by small - scale physics . in high density regions dark halos
are on average more massive and spherical , and have higher spin parameter and velocity dispersion . in high density regions
dark halos are on average more massive and spherical , and have higher spin parameter and velocity dispersion .
we also study the relation between the environmental parameters and the subtypes of dark halos .
the spin parameter of satellite halos depends only weakly on the local density for all mass ranges studied while that of isloated or central halos depends more sensitively on the local density .
the gravitational potential and the shear tensor have weaker correlations with halo properties , but have environmental information independent of the local density . |
proper understanding of the spin structure of nucleon and associated sum rules is expected to offer an important opportunity to investigate quantum chromodynamics(qcd ) as a theory of strong interaction and hence these observables have been the active frontiers in recent years @xcite .
many successful experimental programs of polarized deep - inelastic lepton - nucleon scattering in combination with remarkable theoretical efforts have been devoted in order to elucidate the internal spin structure of the nucleon .
polarized deep inelastic lepton scattering experiment have been carried out at slac , cern , desy and jefferson laboratory(jlab)@xcite . with the advent of dedicated experimental facilities ,
recent experiments were able to determine the spin structure functions as well as different sum rules over a wide range of @xmath2 and @xmath1 with ever increasing precision .
simultaneously , tremendous progress is observed in the field of theoretical investigation in determining and understanding the shape of quarks and gluon spin distribution functions with perturbative qcd , non - perturbative qcd , chiral perturbation theory@xcite , lattice qcd@xcite , anti - de sitter / conformal field theory ( ads / cft)@xcite , etc .
, along with different reliable theoretical models .
in addition , recently available several dedicated phenomenological works@xcite in extracting polarized parton distribution function(ppdf ) as well as spin structure functions from different experiments within nlo and nnlo qcd analysis have also significant contributions towards the understanding of spin structure of the nucleon . in quantum chromodynamics , the spin structure function @xmath3 is described as mellin convolutions between parton distribution functions ( @xmath4 ) and the wilson coefficients @xmath5 @xcite @xmath6,\label{eq1}\end{aligned}\ ] ] which consists of three parts , non - singlet ( @xmath7 $ ] ) , singlet ( @xmath8 $ ] ) and gluon ( @xmath9 $ ] ) .
the @xmath1 distribution of these spin dependent non - singlet , singlet and gluon distribution functions are governed by a set of integro - differential equations , the dokshitzer - gribov - lipatov - altarelli - parisi ( dglap ) equations which are given by @xcite @xmath10 @xmath11 here @xmath12 are the polarized splitting functions @xcite .
these equations are valid to all orders in the strong coupling constant @xmath13 .
although qcd predicts the @xmath1 dependence of structure functions in accord with the dglap equations but they have limitations on absolute prediction of structure functions .
dglap equations can not predict the initial values from which the evolution starts , they can only predict the evolution of structure functions with @xmath1 , once an initial distribution is given .
further , due to its complicated mathematical structure , an exact analytic determination of the structure functions is currently out of reach and one needs to apply approximated methods to arrive on predictions from the dglap equation .
accordingly several approximate numerical as well as semi - analytical methods for the solution of dglap equation have been discussed considerably over the past years @xcite . in literature
there are essentially two main classes of approaches in order to have solutions of dglap equations : those that solve the equation directly in @xmath2-space and those that solve it for mellin transformations of structure functions and invert the transformation back to @xmath2-space .
the approaches based on mellin transformation method have been achieved much interest because under mellin transformation the integro - differential dglap equation turns into a continuum of independent matrix differential equations , one for each value of moments(@xmath14 ) , which in turn makes the evolution more efficient numerically . however , in this regard as the mellin transformation of both the splitting functions and the initial input is required , which may not be possible for all functions , especially if higher - order corrections are included in the equations , therefore it is not possible to have exact solution to dglap equation in moment space beyond leading order .
in contrast to mellin space , the @xmath2-space method is more flexible , since the inputs are only required in @xmath2-space ; however it is generally considered to less efficient numerically , because of the need to carry out the convolution in dglap equations .
taking into account the advantage of being greater flexibility , despite the difficulty in obtaining high accuracy , the @xmath2-space methods have been serving as the basis of many widely used programs hoppet@xcite , qcdnum@xcite , candia@xcite etc . , and being incorporated by the cteq@xcite , mrst / mstw(see @xcite and references therein ) collaborations .
in addition , several numerical and semi - analytical methods have been developed@xcite and achieved significant phenomenological success . due to the unavailability of exact analytical way of solving the dglap equations , in current analysis this set of equations
are solved numerically by using an initial input distribution for the structure function at a fixed @xmath1 , in terms of some free parameters , the parameters are so adjusted that the parametrization best fit the existing data . however , the consideration of a specific parametrization with large number of parameters is potentially a source of bias , i.e. systematic error which is very difficult to control .
furthermore , when a parametrization is fitted to the data , it is very hard to obtain a determination not only of the best fitting parameters , but also of their errors . therefore , explorations of the possibility of obtaining accurate solutions of dglap evolution equations without an initial input or with initial input , consisting of less number of parameters are always interesting .
nnpdf method is one of the most interesting methods which does not require to assume a functional form and it is largely bias free@xcite . in order to perform a fit
, one must start with a particular ansatz for the structure functions at some reference @xmath15 . in most of the existing fitting analysis , including those in the experimental papers it has been performed by assuming a simple power behavior based on regge theory .
the small-@xmath2 behaviour of @xmath16 in accord with regge theory is given by ( see @xciteand references therein ) @xmath17 and describes the slac@xcite experimental data with @xmath18 and @xmath19@xcite .
although this regge model seems to legitimate as far the early data are concerned , which were mostly taken at moderate @xmath1 ( @xmath20 ) and @xmath2 values of around @xmath21 , but the recent measurement of @xmath16 which are available within the small-@xmath2 interval @xmath22 can be described with a single regge type exchange @xmath23 , in which the intercept has a smooth @xmath1 dependence and @xmath16 varies like @xmath24 with @xmath25 .
similar behaviour was predicted with valon model@xcite and a variation from @xmath26 to @xmath27 was obtained within the interval of @xmath1 from @xmath28 to @xmath29 . on the other hand ,
ref.@xcite predicts a behaviour of the type , @xmath30 , with @xmath31 in which the asymptotic scaling of @xmath16 depends on only one variable @xmath32 .
in addition , there are several studies on the small-@xmath2 behaviour of non - singlet part of spin structure function(see for example @xcite ) . in this manuscript
we have investigated the small-@xmath2 behaviour of @xmath16 structure function based on a simple regge ansatz of the type @xmath33 with @xmath1 dependent intercept .
the underlying idea behind the assumption of this type of model is as follows : hera measurements@xcite suggest that the behavior of @xmath34 structure function at low-@xmath2 is consistent with a dependence @xmath35 , where the coefficient @xmath36 is independent of @xmath1 and the exponent , defined by @xmath37 , is observed to rise linearly with @xmath38 . here
@xmath39 is the qcd cut off parameter and @xmath40 .
thus we see that the rise of the un - polarized structure function ( @xmath41 ) is much steeper than that predicted by regge theory and gets steeper and steeper as @xmath1 increases .
since this observation it has been the challenging issue to resolve whether the regge intercepts for @xmath34 structure function as well as it s non - singlet , singlet and gluon parts , along with the spin structure functions are @xmath1 dependent or not .
further , before the observation at hera , there are several predictions on the @xmath1 dependency of the regge intercept@xcite .
these predictions as well as experimental observations at hera motivated us to consider the possibility that the regge behaved non - singlet part , @xmath42 of @xmath43 structure function is also satisfy a functional behaviour , @xmath44 similar to @xmath41 . again as both the non - singlet structure functions , @xmath42 and @xmath45 are regge behaved@xcite , therefore their @xmath2 dependency will be similar within smaller-@xmath2 region .
further , in qcd the @xmath1 behaviour of these structure functions are governed by the same dglap equation .
therefore the @xmath2 and @xmath1 dependency for both the non - singlet structure functions are similar and in accord with @xmath34 , and hence @xmath46 , here we assume that the @xmath1 dependency of the regge behaved structure function @xmath47 is dominated only by the intercept and it satisfies a relation of the type @xmath48 . in order to investigate the small-@xmath2 behaviour of @xmath47 structure function , in this paper firstly
we have assumed that @xmath49 satisfies the regge like behavior @xmath48 , with the @xmath1 dependent intercept within smaller @xmath2 region and then using the model as the initial input , we have solved the dglap equation analytically in leading order ( lo ) , next - to - leading order ( nlo ) and next - next - to - leading order ( nnlo ) .
we have performed a phenomenological analysis of these solutions in comparison with different experimental measurements@xcite as well as the predictions due to different models @xcite and achieved at a very good phenomenological success .
the phenomenological success achieved in this regard reflects , on one hand the acceptability of the regge ansatz in describing the small @xmath2 behavior of the non - singlet part of spin structure function and on the other hand , the usefulness of the regge ansatz in evolving the spin structure function , @xmath50 in accord with dglap equation with a considerable precision within smaller @xmath2 region .
we have utilized the solutions of the dglap equations in determining the bjorken sum rule ( bsr)@xcite .
bsr relates the difference of proton and neutron structure functions integrated over all possible values of bjorken variable , @xmath2 to the nucleon axial charge @xmath51@xcite .
as bsr holds the flavour non - singlet combination of the spin structure function , it is not marred by the presence of the sea quark and gluon densities about which we have very poor information in particular in the small-@xmath2 region , therefore qcd analysis by means of bsr is less complex and more accurate . at
infinite four - momentum transfer squared , @xmath1 , the sum rule reads @xmath52 however , away from @xmath53 , the polarized bjorken sum rule is given by a sum of two series in powers of the strong coupling constant @xmath54 ( leading twist pqcd correction)and in powers of @xmath55(nonperturbative higher twist corrections ) : @xmath56 + \sum_{i=2}^\infty\frac{\mu_{2i}^{p - n}(q^2)}{q^{2i-2}}\label{eq5 } \end{aligned}\ ] ] here the leading twist term ( bracket term ) consists of pqcd results up to third order of @xmath54@xcite .
recently the calculation of the fourth order contribution to the bsr is also available@xcite . the second term on r.h.s .
is known as higher twist term . the higher order pqcd corrections and higher twist power corrections are significant at low-@xmath1 region(see ref . @xcite and references therein ) . in this paper
we have paid attention to both the parts . as our analysis
is based on the solution of dglap equation and the splitting functions for @xmath50 are not available beyond nnlo , we could not incorporate the available nnnlo corrections to bsr and restricted our study within nnlo perturbative corrections . required higher twist corrections to describe low-@xmath1 data with different perturbative orders are extracted based on a simple parametrisation and by means of fitting with the available experimental data .
bsr represents the area under the curve @xmath50 from from @xmath57 to @xmath58 and in accord with nnlo pqcd , the @xmath1 dependence of this area is given by , @xmath59\label{eq6}. \end{aligned}\ ] ] here @xmath60 is the scale - invariant isovector axial - charge measured in neutron beta - decays@xcite .
the bjorken integral can be resolved as @xmath61 which gives @xmath62 the integral on the left hand side of ( [ eq8 ] ) represents the area under the curve @xmath63 from @xmath64 to @xmath58 . for @xmath65
, this integral will tend to cover the whole area under the curve from @xmath66 to @xmath58 , that is , it will represent the bjorken integral .
again the second part on the right side of ( [ eq8 ] ) represents the part of total area @xmath67 , laying under the curve @xmath63 within smaller @xmath2 region i.e. , from @xmath57 to any smaller value @xmath64 .
thus we see that in order to investigate bsr , we just require the knowledge of @xmath68 structure function within smaller @xmath2 region .
this requirement is fulfilled by using the solutions of dglap equations , which provide well description of the small-@xmath2 behaviour of @xmath0 structure function and determined the @xmath1 behaviour of bsr . when these results for bsr are compared with different experimental data @xcite and the predictions made in ref .
@xcite along with qcd predictions up to nnlo corrections@xcite , we have observed a very good consistency among them .
the paper is organized as follows : in the next section , sec .
[ sec : level2 ] we have discussed about obtaining the small @xmath2 behavior of @xmath50 structure function by solving of the dglap equation for @xmath68 in leading order(lo ) , next - to - leading order(nlo ) and next - next - to - leading order(nnlo ) using the regge like ansatz , @xmath69 as the initial input . also in this section , we have performed a phenomenological analysis of our results in comparison with different experimental data as well as several model predictions .
these expressions as the solutions of dglap equation in lo , nlo and nnlo are then used to determine the bjorken integral with qcd corrections up to nnlo in sec .
[ sec : level3.1 ] , along with a phenomenological analysis of the results . in sec .
[ sec : level3.2 ] , we have extracted the higher twist terms , associated with different perturbative orders based on a simple parametrisation . finally , in sec . [
sec : level4 ] , we have concluded the paper with a brief discussion .
we have considered that the non - singlet part of the spin structure function satisfies the following regge ansatz : @xmath71 and hence we have @xmath72 for simplicity we denote @xmath73 and with this notation , @xmath74 dependence of @xmath75 structure function at a particular value of @xmath76 is givent by @xmath77 dividing ( [ eq252 ] ) by ( [ eq26 ] ) we have the following relation @xmath78 which gives both @xmath74 and @xmath2 dependence of @xmath79 structure function in terms of the @xmath74 dependent function @xmath80 at fixed @xmath76 .
the @xmath74 dependent function , @xmath81 can be obtained from the dglap equation .
the dglap equation for @xmath0 structure function can be written as @xmath82 in terms of the variable @xmath83 . here
the splitting function , @xmath84 is defined up to next - next - to - leading order by @xmath85 , where , @xmath86 , @xmath87 and @xmath88 are the corresponding lo , nlo and nnlo corrections to the splitting functions@xcite .
again , in lo , nlo and nnlo the coupling constant @xmath89 is given by @xmath90 , @xmath91 $ ] and @xmath92\bigg]$ ] respectively , where @xmath93 , @xmath94 and @xmath95 are the corresponding one loop , two loop and three loop corrections of qcd @xmath96 function@xcite . substituting the respective splitting functions along with the corresponding running coupling constant in ( [ eq9 ] ) the dglap evolution equations in lo ,
nlo and nnlo become @xmath97 \label{eq10},\ ] ] @xmath98\nonumber\\+{\bigg(\frac{\alpha ( t)}{2\pi}\bigg)}^2_{nlo } i_{2}(x , t ) \label{eq11},\end{aligned}\ ] ] and @xmath99\nonumber\\+{\bigg(\frac{\alpha ( t)}{2\pi}\bigg)}^2_{nnlo } i_{2}(x , t)+{\bigg(\frac{\alpha ( t)}{2\pi}\bigg)}^3_{nnlo } i_{3}(x , t ) \label{eqnnlo}\end{aligned}\ ] ] respectively . here
the integral functions are given by @xmath100 @xmath101 and @xmath102 we now solve these dglap evolution equations in lo , nlo and nnlo using the @xmath1 dependent regge ansatz , @xmath103 . substituting @xmath104 and @xmath105 in equation ( [ eq10 ] ) and rearranging a bit we can convert the lo dglap equation into an ordinary differential equation which can be easily solved to have @xmath106 . \label{eq16}\ ] ] here @xmath107 and @xmath108 is the constant of integration . at a fixed value of @xmath76 , the @xmath74 dependence of the structure function in lo is given by @xmath109 .
\label{eq20}\ ] ] again the value of the structure function at @xmath76 and @xmath110 in accord with ( [ eq16 ] ) is given by @xmath111\bigg|_{t = t_0}. \label{eq22}\ ] ] dividing ( [ eq20 ] ) by ( [ eq22 ] ) and rearranging a bit we obtain the @xmath74 dependence of @xmath81 in accord with lo dglap evolution equation with respect to the point @xmath112 as @xmath113 . \label{eq23}\ ] ] now substituting @xmath114 from ( [ eq23 ] ) in ( [ eq27 ] ) , we have a relation representing both @xmath2 and @xmath74 dependence of structure function in lo , in terms of the input point @xmath115 given by @xmath116\bigg(\frac{x}{x_0}\bigg)^{(1-bt)}.\label{eq28 } \end{aligned}\ ] ] proceeding in the similar way we can obtain the relation for @xmath79 structure function in nlo and nnlo as @xmath117\bigg(\frac{x}{x_0}\bigg)^{(1-bt)},\label{eq29 } \end{aligned}\ ] ] and @xmath118\bigg(\frac{x}{x_0}\bigg)^{(1-bt)}\label{eqnnlosol } \end{aligned}\ ] ] respectively , where @xmath119 and @xmath120 structure function at @xmath121 compared with the experimental data ref .
@xcite and the results from the analysis presented in ref .
@xcite.,scaledwidth=100.0% ] the expressions , ( [ eq28 ] ) , ( [ eq29 ] ) and ( [ eqnnlosol ] ) represent both @xmath2 and @xmath74 dependence of @xmath122 structure function in lo , nlo and nnlo at small @xmath2 .
they are consisting of the fitting parameter @xmath123 , and a known input point @xmath115 which can be taken from the available experimental data .
if the input point is more accurate and precise , we can expect better fitting .
there are not any specific reason in choosing the input point .
any one of the data points at a certain value of @xmath76 and @xmath110 can be considered as the input point .
of course , the sensitivity of different inputs will be different .
however instead of choosing the input point on the basis of their sensitivity , in our manuscript we have incorporated a suitable condition in determining the input point .
we have considered that particular point from the most recent measurements as the input point in which experimental errors are minimum . under this condition
we have selected the point @xmath124 at @xmath125 and @xmath126 from the recent measurement compass .
the value of the parameter @xmath123 is obtained by fitting the parametrization @xmath127 , which is a reduced form of @xmath128 , with different experimental data taken from compass@xcite , hermes@xcite , smc@xcite and e143@xcite collaborations .
the parametrization in its reduced form consists of only one fitting parameter @xmath123 and hence makes easier the fitting analysis .
as we are interested in the small @xmath2 region , therefore we have restricted our fitting analysis with experimental data sets for @xmath129 and in this analysis the best fitting value for the parameter @xmath123 is found to be @xmath130 with @xmath131 .
using this value for @xmath123 we have calculated the @xmath132 structure function considering the input point @xmath133 , and depicted their @xmath2 evolution for @xmath126 in fig .
[ fig : fig1 ] , in comparison with smc , e143 , hermes and compass experimental results , along with the predictions in ref @xcite based on various models.also we have estimated the uncertainty associated with the fitting parameter @xmath123 and the chosen input point and the respective uncertainty bands are shown in fig .
[ fig : fig1 ] separately . here
the uncertainty due to the fitting parameter is considerably less than that of due to input point .
however both the uncertainties are observed to be decreasing as @xmath2 decreases .
we see that @xmath134 structure functions evolved with respect to the input point are consistent with those of experimental measurements as well as other models .
this implies that the expressions , we have obtained by means of solving the dglap equations analytically , are applicable in describing small @xmath2 behaviour of @xmath0 structure function with a considerable precision and therefore these expressions can be successfully incorporated in @xmath135 for @xmath68 term in order to determine bsr .
the expression ( [ eq8 ] ) suggest that in order to investigate bjorken sum rule ( @xmath136 ) , we require the knowledge of the non - singlet spin structure function @xmath0 within smaller @xmath2 region .
this requirement can be fulfilled by using the solutions of dglap equations obtained above . therefore , substituting ( [ eq28 ] ) , ( [ eq29 ] ) and ( [ eqnnlosol ] ) in ( [ eq8 ] ) and using the corresponding expressions for @xmath137 in lo , nlo and nnlo , we obtain the bjorken integral with lo , nlo and nnlo qcd corrections as @xmath138,\label{eq30 } \end{aligned}\ ] ] @xmath139\label{eq31 } \end{aligned}\ ] ] and @xmath140\label{eq31nnlo } \end{aligned}\ ] ] respectively . considering a known input point @xmath115 from experimental data
, we will be able to calculate the bsr integral up to nnlo corrections using the expressions , ( [ eq30 ] ) , ( [ eq31 ] ) and ( [ eq31nnlo ] ) respectively . in our calculations
we have used @xmath133 as the input point , which is taken from the compass@xcite experimental data . with this input point
we have calculated the bjorken integral and the results in accord with equations ( [ eq30 ] ) , ( [ eq31 ] ) and ( [ eq31nnlo ] ) are depicted in fig .
[ fig : fig2 ] and fig .
[ fig : fig3 ] . in fig .
[ fig : fig2 ] , we have plotted our results for bsr integral in lo , nlo and nnlo as a function of low @xmath2 limit of integration @xmath141 , in comparison with compass and hermes measurements along with the results due to valon model(tsa)@xcite .
the uncertainties due to the parameter , @xmath123 and the input point are estimated only for the nnlo results and as seen from the fig .
[ fig : fig2 ] , they decrease with decrease in @xmath141 . from fig .
[ fig : fig2 ] we observe an overall better description of both compass and hermes data by our results with respect to the predictions due to valon model .
again our approach expects better results for @xmath142 , but there are no compass measurement for @xmath143 and hermes measurement beyond @xmath144 for our comparative analysis .
saturation of the compass data for bsr is observed within @xmath145 , however available hermes results have not saturated within @xmath146 .
thus we may expect to occur saturation within the smaller @xmath2 region and within this region both hermes and compass results might agree with each other and reach an overall compatibility with our measurements . the @xmath1 dependency of bjorken sum rule , as predicted by our expressions ( [ eq30 ] ) , ( [ eq31 ] ) , and ( [ eq31nnlo ] ) is depicted in fig .
[ fig : fig3 ] . here
our results are compared with different experimental data taken from compass @xcite , hermes@xcite , e143@xcite and jlab experiments @xcite and with the theoretical as well as phenomenological analysis , ref .
the results depicted in this figure are calculated using the value of @xmath147 .
here we have also estimated the uncertainty associated with the nnlo results due to the fitting parameter , @xmath123 and the input point , and they are observed to be very small in this regard .
it is also observed that the uncertainty decreases with decrease in @xmath1 .
in addition , the dependence of our results , specifically the nnlo results with the qcd parameter @xmath39 is shown in fig .
[ fig : fig4 ] . fig .
[ fig : fig4 ] reflects that towards low-@xmath1 region , our results are highly sensitive to @xmath39 , however the sensitivity decreases towards the smaller values of @xmath39 .
as seen from this figure our approach has the capability for better description of the low-@xmath1 bsr data with lower values of @xmath39 . in fig .
[ fig : fig5 ] , we have compared our results with theoretical pqcd predictions ( [ eq6 ] ) for bjorken integral up to nnlo . here
our results are calculated with @xmath148 . within the estimated uncertainty
our results show a very good consistency with those of pqcd predictions .
we now incorporate the higher twist contribution to our results presented above . here
we have extracted the effect of first non - leading twist term , ( @xmath149 ) in lo , nlo and nnlo perturbative orders . to extract the first higher twist term we parameterise the bjorken sum rule as @xmath150 here
leading twist(lt ) term corresponds to the pqcd contribution to bsr . incorporating our @xmath1 dependent expressions ( [ eq30 ] ) , ( [ eq31 ] ) and ( [ eq31nnlo ] ) for bsr in lo , nlo and nnlo as the lt terms we have fitted above parametrisation to the low @xmath1 ( @xmath151 ) experimental data taken from compass @xcite , hermes@xcite , e143@xcite and jlab experiments @xcite .
the corresponding values of @xmath149 for which best fitting is obtained in different pt orders are summarised in table [ tab : ht ] , along with the respective @xmath152 values .
the table reflects a substantial change in @xmath149-values with different pt orders , which can be interpreted as the manifestation of duality between higher orders and higher twist@xcite . in fig .
[ fig : fig6 ] we have presented the best fitting results in various orders of pt . here both the results , with ht and without ht are shown .
we observe that our expressions along with the ht corrections provide well description of bsr data .
.higher twist corrections in various perturbative orders [ cols="<,<,<,<",options="header " , ] the calculations , in this paper are made by using mathematica 9@xcite . in order to obtain
sufficient numerical accuracy caution has been taken in evaluation of integrals containing higher order polynomials .
the evaluations have been performed using horner s method and it has improved speed and stability for numeric evaluation of large polynomials present in the splitting functions . ) of bsr to the experimental data taken from compass @xcite , hermes@xcite , e143@xcite and jlab @xcite.,scaledwidth=100.0% ]
in this paper we have obtained some expressions for the non - singlet part of spin structure function , @xmath70 at small-@xmath2 by means of analytical solution of dglap equation in lo , nlo and nnlo using a regge like ansatz with @xmath1 dependent intercept as the initial input .
the solutions are used to calculate the bjorken integral @xmath153 with qcd corrections up to nnlo .
in addition , we have estimated the higher twist correction for the bjorken sum rule .
we have performed phenomenological analyses of these results for @xmath70 and @xmath153 with different experimental data and model predictions . from the phenomenological analysis we have the following observations : i. the regge inspired ansatz in accord with dglap equations provides a very good description of the small-@xmath2 behaviour of @xmath70 , which are consistent with other results taken from ref .
@xcite .
ii . our expressions for bjorken integral , along with considerable higher twist correction provide a very good description of the following experimental measurements : compass @xcite , hermes@xcite , e143@xcite and jlab experiments @xcite , which indicates that the experimental data strongly confirm the bsr prediction .
our results for @xmath1 behaviour of bjorken integral are also consistent with the qcd predictions up to nnlo .
this consistency between our results and theoretical qcd predictions suggests that available data , the regge ansatz and the theoretical framework of pqcd , through this simple method allow us to have a clean test of pqcd predictions on bsr .
the consistency of the results for @xmath0 and @xmath137 due to the regge like model , @xmath154 with different experimental results@xcite and other strong analysis @xcite signifies that the model is applicable in describing the small-@xmath2 behaviour of @xmath0 structure function although it being simple .
moreover , in this method we do not require the knowledge of initial distributions of structure functions at all values of @xmath2 from 0 to 1 . here , we just require one input point at any fixed @xmath2 and @xmath1 and with respect to that point both the @xmath2 and @xmath1 evolution of structure functions can be obtained .
our concluding impression based on all these observations is that the simple but efficient @xmath1 dependent regge ansatz for @xmath0 is capable of evolving successfully the @xmath0 structure function in accord with dglap equation at small-@xmath2 and the regge ansatz and the theoretical framework of pqcd , along with available experimental data lead towards a clean test of pqcd predictions of bjorken sum rule .
blumlein , j. , falcioni , g. , freitas , a. de , nucl .
b * 910 * , 568 ( 2016 ) taghavi - shahri , f. , khanpour , h. , atashbar tehrani , s. , z. alizadeh yazdi , z , a , phys .
d * 93 * , 114024 ( 2016 ) sato , n. , melnitchouk , w. , kuhn , s. e. , ethier , j. j. , accardi , a. , phys .
d * 93 * , 074005 ( 2016 ) ball , r. d. , _ et al .
_ , [ nnpdf collaboration ] , nucl .
b * 877 * , 290 ( 2013 ) jimenez - delgado , p. , accardi , a. , melnitchouk , w. , phys .
d * 89 * , 034025 ( 2014 ) leader , e. , sidorov , a. v. , stamenov , d. b. , phys .
d * 82 * , 114018 ( 2010 ) de florian d. , sassot , r. , stratmann , m. , vogelsang , w. , phys .
* 101 * , 072001 ( 2008 ) ; de florian , d. , sassot , r. , stratmann , m. , vogelsang , w. , phys .
d * 80 * , 034030 ( 2009 ) blumlein , j. , bottcher , h. , nucl
b * 841 * , 205 ( 2010 ) hirai , m. , kumano , s. , nucl .
b * 813 * , 106 ( 2009 ) bourrely , c. , soffer , j. , buccella , f. , eur .
j. c * 23 * , 487 ( 2002 ) khorramian , a. n.,atashbar tehrani , s. , taheri monfared , s. , arbabifar , f. , olness , f. i. , phys .
d * 83 * , 054017 ( 2011 ) arbabifar , f. , khorramian , a. n. , soleymaninia , m. , phys .
d * 89 * , 034006 ( 2014 ) taghavi - shahri , f. , arash f. , phys .
c * 82 * , 035205 ( 2010 ) altarelli , g. , ball , r. d. , forte , s. , ridolfi , g. , nucl.phys .
b * 496 * , 337 ( 1997 ) gribov , v. n. , lipatov , l. n. , sov . j. of nucl . phys .
* 15 * , 438 ( 1997 ) lipatov , l. n. , sov . j. of nucl
. phys . * 20 * , 94 ( 1975 ) dokshitzer , y. l. , sov .
jetp * 46 * , 641 ( 1977 ) altarelli , g. , parisi , g. , nucl .
b * 126 * , 297 ( 1977 ) mertig , r. , van neerven , w. l. , z. phys .
c , * 70 * , 637 ( 1996 ) vogelsang , w. , phys .
d , * 54 * , 2023 ( 1996 ) moch , s. , vermaseren , j. a. m. , a. vogt , arxiv:1409.5131 [ hep - ph ] botje , m. , qcdnum , http://www.nikhef.nl/ h24/qcdnum/ schoeffel , l. , nucl .
instrum . meth .
a * 423 * , 439 ( 1999 ) vogt , a. , comput .
. commun . * 170 * , 65 ( 2005 ) pascaud , c. and zomer , f. , arxiv : hep - ph/0104013 .
weinzierl , s. , comput .
. commun . * 148 * , 314 ( 2002 ) ; roth , m. , weinzierl , s. , phys . lett .
b * 590 * , 190 ( 2004 ) cafarella , a. and coriano , c. , comput .
commun . * 160 * , 213 ( 2004 ) ; cafarella , a. , coriano , c. and guzzi , m. , nucl . phys .
b * 748 * , 253 ( 2006 ) salam , g. p. , rojo , j. , comput .
. commun . *
180 * , 120 ( 2009 ) pumplin , j. , stump , d. r. , tung , w. k. , phys .
d * 65 * , 014011 ( 2002 ) ; stump , d. , et al .
d * 65 * , 014012 ( 2002 ) ; pumplin , j. et al . , phys .
d * 65 * , 014013 ( 2002 ) martin , a. d. , stirling , w. j. , thorne , r. s. , watt , g. , eur .
j. c * 63 * , 189 ( 2009 ) schoeffel , l. , nucl .
instrum . meth .
a * 423 * , 439 ( 1999 ) pascaud , c. , zomer , f. , arxiv : hep - ph/0104013 .
cafarella , a. , coriano , c. , comput .
commun . * 160 * , ( 2004 ) 213 ; cafarella , a. , coriano , c. , guzzi , m. , nucl . phys .
b * 748 * , 253 ( 2006 ) toldra , r. , comput .
commun * 143 * , 287 ( 2002 ) ; cabibbo , n. , petronzio , r. , nucl .
b * 137 * , 395 ( 1978 ) ; gluck , m. , reya , e. , vogt , a. , z. phys .
c * 48 * , 471 ( 1990 ) ; devee , m. , baisya , r. , sarma , j. k. , euro .
j. c * 72 * , 2036(1 - 11 ) ( 2012 ) ; baishya , r. , jamil , u. , sarma , j. k. , phys .
d * 79 * , 0340 , ( 2009 ) ; baishya , r. , sarma , j. k. , phys .
d * 74 * , 107702 ( 2006 ) ; baishya , r. , sarma , j. k. , euro .
j. c * 60 * , 585 ( 2009 ) ; nath , n. m. , baruah , n. , sarma , j. k. , indian j. phys . * 88 * , 745 ( 2014 ) ; baruah , n. , das , m. k. , sarma , j. k. , eur .
j. plus * 129 * , 229 ( 2014)(doi : 10.1140/epjp / i2014 - 14229 - 5 ) ; cooper - sarkar , a. m. , devenish , r. c. e. , deroeck , a. , int . j. mod .
a * 13 * , 3385 ( 1998 ) ; martin , a. d. , stirling , w. j. , thorne , r. s. , watt , g. , eur .
j. c * 63 * , 189 ( 2009 ) ; martin , a. d. , roberts , r. g. , stirling , w. j. , thorne , r. s. , eur .
j. c * 4 * , 463 ( 1998 ) ; gluck , m. , reya , e. , a. vogt , eur .
j. c * 5 * , 461 ( 1998 ) soffer , j. , teryaev , o. v. , phys .
d * 56 * , 1549 ( 1997 ) bass , s.d . ,
brisudova , m. m. , eur .
j a * 4 * , 251 ( 1999 ) ermolaev , b. i. , greco , m. , troyan s. i. , riv .
nuovo cim . * 33 * , 57 ( 2010 ) knauf , a. , hermann , m. m. , soff , g. , phys .
b * 549 * , 109 ( 2002 ) ; kwiecinski , j. , acta .
phys . pol .
b * 29 * , 1201 ( 1998 ) ; kwiecinski , j. , acta phys . pol .
b * 27 * , 893 ( 1996 ) ; bartels , j. , ermolaev , b. i. , ryskin , m. g. , z. phys .
c * 70 * , 273 ( 1996 ) ; bartels , j. , ermolaev , b. i. , ryskin , m. g. , z. phys .
c * 72 * , 627 , ( 1996 ) alekseev , m. g. , ( compass collaboration ) , phys .
b * 690 * , 466 ( 2010 ) ; alekseev , m. g. ( compass collaboration ) , phys .
b * 660 * , 458 ( 2008 ) ; alexakhin , v. yv . _ et al .
_ , ibid . * 647 * , 8 ( 2007 ) ; alekseev , m. _ et al .
_ , ( compass collaboration ) , phys .
b * 680 * , 217 ( 2009 ) ackerstaff , k. , _ et al .
_ , ( hermes collaboration ) , phys .
b * 404 * , 383 ( 1997 ) ; phys .
b * 444 * , 531 ( 1998 ) ; airapetian , a. , _ et al .
_ , , phys . lett .
b * 442 * , 484 ( 1998 ) ; phys .
lett . * 90 * , 092002 ( 2003 ) ; eur .
j. c * 26 * , 527 ( 2003 ) ; phys .
d * 75 * , 012007 ( 2007 ) .
adeva , b. , _ et al .
_ , ( smc collaboration ) , phys . rev .
d * 58 * , ( 1998 ) 112001 ; adeva , b. , _ et al .
_ , ( smc collaboration ) , phys . rev .
d * 60 * , 072004 ( 1999 ) abe , k. , _ et al .
_ , ( e143 collaboration ) , phys .
lett . * 74 * , 346 ( 1995 ) ; phys .
* 75 * , 25 ( 1995 ) ; phys .
lett . * 76 * , 587 ( 1996 ) ; phys .
b * 364 * , 61 ( 1995 ) ; phys .
d * 58 * , 112003 ( 1998 ) goto , y. _ et al .
_ , ( asymmetry analysis collaboration ) , phys . rev .
d * 62 * , 034017 ( 2000 ) ; phys .
d * 69 * , 054021 ( 2004 ) blumlein , j. , bottcher , h. , nucl .
b * 636 * , 225 ( 2002 ) gluck , m. , reya , e. , stratmann , m. , vogelsang , w. , phys .
d * 63 * , 094005 ( 2001 ) bjorken , j. d. , phys
. rev . * 148 * , ( 1996 ) 1467 ; phys .
d * 1 * , 1376 ( 1970 ) nakamura , k. , _ et al .
_ , ( particle data group ) , j. phys .
g * 37 * , 075021 ( 2010 ) larin , s. a. , vermaseren , j. a. m. , phys .
lett b * 259 * , 345 ( 1991 ) baikov , p. a. , chetyrkin , k. g. , kuhn , j. h. , phys .
* 104 * , 132004 ( 2010 ) baikov , p. a. , chetyrkin , k. g. , kuhn , j. h. , phys .
* 101 * , 012002 ( 2008 ) pasechnik , r. s. , shirkov , d. v. , teryaev , o. v. , phys .
d * 78 * , 071902 ( 2008 ) pasechnik , r. s. , shirkov , d. v. , teryaev , o. v. , solovtsova , o. p. , khandramai , v. l. , phys .
d * 81 * , 016010 ( 2010 ) deur , a. , _ et al .
* 93 * , 212001 ( 2004 ) dharmawardane , k. v. , _ et al .
b * 641*,11 ( 2006 ) ; bosted , p. e. , _ et al .
c * 75 * , 36203 ( 2007 ) ; prok , y. , _ et al .
b * 672 * , 12 ( 2009 ) amarian , m. , _ et al .
* 89 * , 242301 ( 2002 ) ; fatemi , r. , _ et al .
_ , phys . rev
* 91 * , 222002 ( 2003 ) ; amarian , m. , _ et al .
lett . * 92 * , 022301 ( 2004 ) burkert , v. d. , ioffe , b. l. , phys .
b * 296 * , 223 ( 1992 ) ; j. exp .
. phys . * 78 * , 619 ( 1994 ) soffer , j. , teryaev , o. v. , phys .
b * 545 * , 323 ( 2002 ) ; phys .
d * 70 * , 116004 ( 2004 ) khandramai , v. l. , _ et al .
b * 706 * , 340 ( 2012 ) kotikov , a. v. , shaikhatdenov , b. g. , phys .
nucl . * 44 * , ( 2013 ) 543 ; phys .
* 45 * , 26 ( 2014 ) | a calculation of the non - singlet part of spin dependent structure function , @xmath0 and associated sum rule , the bjorken sum rule up to next - next - to - leading order(nnlo ) is presented .
we use a unified approach incorporating regge theory and the theoretical framework of perturbative quantum chromodynamics .
using a regge behaved model with @xmath1 dependent intercept as the initial input , we have solved the dokshitzer - gribov - lipatov - altarelli - parisi ( dglap ) evolution equation up to nnlo at small-@xmath2 for @xmath0 and the solutions are utilised to calculate the polarised bjorken sum rule(bsr ) .
we have also extracted the higher twist contribution to bsr based on a simple parametrisation .
these results for both of @xmath0 and bsr , along with higher twist corrections are observed to be consistent with the available data taken from smc , e143 , hermes , compass and jlab experiments .
in addition , our results are also compared with that of other theoretical and phenomenological analysis based on different models and a very good agreement is also observed in this regard .
further a very good consistency between our calculated results and theoretical qcd predictions of bsr is also achieved .
example.eps gsave newpath 20 20 moveto 20 220 lineto 220 220 lineto 220 20 lineto closepath 2 setlinewidth gsave .4 setgray fill grestore stroke grestore |
there is much ongoing experimental effort by the atlas and cms collaborations at the cern large hadron collider ( lhc ) to search for the neutral higgs boson ( @xmath11 ) of the standard model ( sm ) @xcite .
this model of spontaneous symmetry breaking will be tested at the lhc over all of the theoretically preferred mass range , in an experimental programme which is expected to be completed by the end of the 8 tev run of the lhc . at present @xcite
there are only two regions for the mass for @xmath11 which have not been excluded at 95% c.l : i ) a region of light mass , with @xmath12 , and ii ) a region of heavy mass , @xmath13 gev
. however , this simplest model of one fundamental scalar with a vacuum expectation value ( vev ) might not be nature s choice
. there could be additional scalar fields which also contribute to the masses of the fermions and weak bosons , with a more complicated scalar potential which depends on several arbitrary parameters .
importantly , even in the event of no signal for a sm - like higgs boson at the lhc , the search for scalar particles should continue in earnest due to the fact that a non - minimal higgs sector can give rise to different experimental signatures , some of which are challenging to detect .
consequently , it will take much longer for the lhc to probe all of the parameter space of such models .
a commonly - studied extension of the higgs sector of the sm is the `` two higgs doublet model '' ( 2hdm ) , which is composed of two higgs isospin doublets @xcite ( this model has recently been reviewed in @xcite ) .
notably , this structure is necessary in the minimal supersymmetric ( susy ) extension of the sm ( called the `` mssm '' ) .
the extra higgs doublet gives rise to a particle spectrum with multiple higgs bosons ; three are electrically neutral ( two are cp - even , one is cp - odd ) and two are electrically charged ( denoted by @xmath0 ) .
flavour - changing neutral currents ( fcncs ) mediated by scalars at tree level can be eliminated by requiring that the scalar interactions with the fermions are invariant under discrete symmetries ( `` natural flavour conservation '' , nfc ) @xcite .
the discovery of a charged scalar @xmath0 would be unequivocal evidence of a non - minimal higgs sector , and there have been many studies of the prospects of directly observing @xmath0 from a 2hdm or the mssm at the tevatron and lhc @xcite ( for reviews see @xcite ) moreover , the effect of @xmath0 on the decay rates of mesons ( especially @xmath14 mesons ) also plays a major role in constraining @xmath4 and the fermionic couplings of @xmath0 @xcite,@xcite . if @xmath15 , such particles would most copiously ( though not exclusively @xcite ) be produced in the decays of top quarks via @xmath1 @xcite .
searches in this channel have been performed by the tevatron experiments , assuming the decay modes @xmath3 and @xmath2 @xcite . since
no signal has been observed , constraints are obtained on the parameter space of @xmath16 $ ] , where @xmath17 ( i.e. the ratio of the vacuum expectation values of the two scalar doublets ) .
searches in these channels have now been carried out at the lhc : i ) for @xmath3 with 0.035 fb@xmath18 by atlas @xcite , and ii ) for @xmath2 with 4.8 fb@xmath18 by atlas @xcite and with 1 fb@xmath18 by cms @xcite .
these are the first searches for @xmath0 at this collider .
the constraints on @xmath16 $ ] from the lhc searches for @xmath1 are now superior to those obtained from the corresponding tevatron searches .
the phenomenology of @xmath0 in models with three or more higgs doublets ( called multi - higgs doublet models , mhdm ) , was first studied comprehensively in @xcite , with an emphasis on the constraints from low - energy processes ( e.g. the decays of mesons ) .
although the phenomenology of @xmath0 at high - energy colliders in a mhdm and in a 2hdm has many similarities , the possibility of @xmath9 together with an enhanced branching ratio ( br ) for @xmath6 would be a distinctive feature of the mhdm .
this scenario , which was first mentioned in @xcite and studied in more detail in @xcite , is of immediate interest for the ongoing searches for @xmath1 with @xmath3 by the lhc @xcite .
although the current limits on @xmath3 can be applied to the decay @xmath19 ( as discussed in @xcite in the context of the tevatron searches ) , a further improvement in sensitivity to @xmath1 with @xmath19 could be obtained by tagging the @xmath10 quark which originates from @xmath0 @xcite .
we will estimate the increase in sensitivity to br@xmath20 and to the fermionic couplings of @xmath0 in this scenario .
large values of br@xmath20 are also possible in certain 2hdms , such as the `` flipped 2hdm '' with nfc @xcite . however , in this model one would expect @xmath21 due to the constraint from @xmath22 ( @xmath23 gev @xcite ) , and thus @xmath1 with @xmath19 would not proceed unless there were additional new physics beyond that of the 2hdm which contributed to @xmath22 , and weakened the constraint on @xmath4 . in the `` aligned two higgs doublet model '' ( a2hdm )
@xcite there are no fcncs ( as is the case in a 2hdm with nfc ) due to an alignment of the yukawa couplings .
the phenomenology of @xmath0 in the a2hdm @xcite is very similar to that of @xmath0 in a mhdm @xcite , and @xmath24 in the a2hdm is also compatible with constraints from @xmath8 . our numerical results for br(@xmath19 ) in a mhdm apply directly to the a2hdm .
the 2hdm without nfc and without alignment also has a sizeable parameter space for a large br@xmath20 , and a detailed study can be found in @xcite .
our work is organised as follows . in section
ii we describe the fermionic interactions of @xmath0 in the mhdm / a2hdm . in section iii
we quantify the parameter space for a large br@xmath25 in the mhdm / a2hdm .
in section iv we summarise the tevatron / lhc searches for @xmath1 with @xmath3 and discuss how they could be optimised for @xmath6 .
section v contains our numerical results , with conclusions in section vi .
in a general 2hdm each fermion type ( i.e. up - type quarks , down - type quarks and charged leptons ) couples to both of the scalar doublets .
this would lead to fcncs which are mediated by the neutral scalars , and the magnitude of the associated yukawa coupling is constrained by experimental data ( especially meson - antimeson mixing and the decays of mesons ) .
such fcncs can be suppressed by assuming that the flavour - changing yukawa couplings are very small , which can be achieved by invoking a specific structure of the fermion mass matrices @xcite .
an alternative approach is to eliminate the fcncs by requiring that the lagrangian is invariant under a discrete symmetry , which is achieved if each species of fermion couples to at most one scalar doublet ( nfc ) .
this condition leads to four distinct types of 2hdms which differ in their yukawa couplings .
these four models are called model i , model ii , lepton - specific and flipped . and ii@xmath26 in @xcite and @xmath27 and @xmath28 in @xcite .
] models i and ii have received much phenomenological attention , while the study of the lepton - specific and flipped models has been revived recently @xcite , with early studies in @xcite .
we now introduce the fermionic couplings of @xmath0 in a mhdm and the a2hdm , and discuss how these couplings differ from those in the above 2hdms with nfc .
a mhdm is an extension of the 2hdm with @xmath29 scalar doublets , where @xmath30 .
the suppression of fcncs is obtained by imposing nfc . as in the 2hdm , the mhdm has the virtue of predicting @xmath31 at tree level , with finite higher - order corrections which depend on the mass splittings of the scalars . in the mhdm
there are @xmath32 charged scalars , and a detailed study of the phenomenology of the lightest @xmath0 in such models was performed in @xcite , with the assumption that the other @xmath0 are much heavier .
the interaction of the lightest @xmath0 in a mhdm with the fermions is described by the following lagrangian : @xmath33 here @xmath34 and @xmath35 denote up - type quarks and down - type quarks respectively ( for all three generations ) ; @xmath36 is a ckm matrix element ; @xmath37 , @xmath38 and @xmath39 are the masses of the up - type quarks , down - type quarks and charged leptons respectively ; @xmath40 and @xmath41 are chirality projection operators , and @xmath42 gev . in a 2hdm with natural flavour conservation
, the couplings @xmath27 , @xmath28 and @xmath43 are determined solely by @xmath17 .
the values of @xmath27 , @xmath28 and @xmath43 in the four versions of the 2hdm @xcite are given in table [ couplings ] .
it is clear that @xmath44 and @xmath45 are simply related in the 2hdm e.g. one has @xmath46 for the type ii structure .
.the couplings @xmath27,@xmath28 and @xmath43 in the yukawa interactions of @xmath0 in the four versions of the 2hdm with natural flavour conservation . [ cols="^,^,^,^",options="header " , ] in a mhdm the couplings @xmath27 , @xmath28 and @xmath43 are _ arbitrary complex numbers _ , which are defined in terms of the @xmath47 matrix @xmath48 which diagonalises the mass matrix of the charged scalars : @xmath49 we follow the notation of @xcite in which @xmath50 corresponds to the couplings of the charged goldstone boson , and @xmath51 runs from 2 to @xmath29 for the physical charged scalars .
the fermionic couplings of the lightest @xmath0 in a mhdm are taken to be @xmath52 , @xmath53 and @xmath54 . the subscripts @xmath55 take any integer value up to @xmath29 , and specify which of the @xmath29 doublets couples to which fermion type e.g. for a type ii structure one sets @xmath56 and @xmath57 , while for the `` democratic '' 3hdm ( e.g. @xcite ) one has @xmath58 , @xmath57 and @xmath59 . in a 2hdm , @xmath48 is a @xmath60 matrix with elements given by @xmath61 and @xmath62 ( i.e. one free parameter ) . in a 3hdm , @xmath48 is a @xmath63 matrix with four free parameters which can be taken as @xmath64 , @xmath65 , a mixing angle @xmath66 for the two @xmath0 , and a complex phase @xmath67 .
an explicit form of the matrix @xmath48 for the 3hdm is given in @xcite .
due to the unitarity of the matrix @xmath48 one can derive the following identities @xcite : @xmath68 @xmath69
@xmath70 and @xmath71 in a 2hdm these identities reduce to simple trigonometric relations involving @xmath5 . it is evident that the branching ratios of @xmath0 to fermions in the mhdm depend on the three - dimensional parameter space of @xmath72 , @xmath73 and @xmath74 , in contrast to the case in the 2hdm where a single parameter ( @xmath5 ) determines these three couplings .
it is conventional to consider the phenomenology of the lightest @xmath0 , assuming that the other @xmath0 are heavier .
one then drops the @xmath51 subscript on the couplings of the lightest @xmath0 and uses the lagrangian in eq .
( [ lagrangian ] ) .
many experimental observables in flavour physics would receive a contribution from @xmath0 , and thus the magnitudes of @xmath27 , @xmath28 and @xmath43 are constrained .
detailed studies have been performed in @xcite ( in the context of a mhdm ) and more recently in @xcite in the context of the a2hdm ( see also @xcite ) . for @xmath24 these constraints
are roughly as follows : @xmath75 from @xmath76 ( assuming @xmath77 ) , @xmath78 from leptonic @xmath79 decays , and @xmath80 from @xmath81 . in this work
we will derive constraints on @xmath82 from @xmath1 .
a particularly important constraint on the mass and couplings of @xmath0 in a 2hdm / mhdm is the decay @xmath22 @xcite , which has been measured to be in agreement with the sm prediction .
it is the combination of couplings @xmath83 and @xmath84 which enters the decay rate for @xmath22 ( the contribution from @xmath85 can be neglected ) . since @xmath86 in model ii and the flipped 2hdm ( i.e. @xmath87 ) , the stringent bound @xmath88 gev at 95% c.l can be derived for all values of @xmath5 @xcite .
in contrast , in the mhdm with @xmath87 , the combination @xmath83 is only weakly constrained by the sum rule in eq .
( [ sumrulexy ] ) , and can be negative . for @xmath89 ( e.g. model i and the leptonic - specific 2hdm ) the constraint in eq .
( [ sumrulexy ] ) does not apply .
therefore a light @xmath0
( i.e. @xmath24 ) being compatible with @xmath22 is still a possibility .
recent studies of the bounds on @xmath0 of the mhdm from @xmath22 @xcite derive the following approximate @xmath90 intervals for the real part of @xmath83 with @xmath91 gev : @xmath92 in deriving this constraint it is assumed that @xmath93 is not so big ( e.g. @xmath94 , which is required from other low - energy processes such as @xmath76 ) .
there is also a constraint on the imaginary part of @xmath83 from a different process ( @xmath95 ) , but for simplicity we will consider @xmath27 and @xmath28 to be real .
in addition to the above phenomenological constraints , there are constraints on @xmath27 , @xmath28 and @xmath43 from the unitarity of the matrix @xmath48 for the democratic 3hdm , which were studied in @xcite .
there is a non - trivial relationship on the couplings of the lightest @xmath0 given as : @xmath96 this constraint ensures that the magnitudes of @xmath52 , @xmath53 and @xmath54 can not all be simultaneously less than one , or all be simultaneously greater than one .
this is due to the fact that all three vacuum expectation values @xmath97 can not be simultaneously large or small . in our numerical analysis
we will always take @xmath98 ( as discussed above ) , and so the requirement that all three couplings can not be simultaneously greater than unity is automatically satisfied .
we will be concerned with the parameter space of @xmath99 ( which corresponds to large @xmath5 and moderate / small @xmath100 ) , a choice which satisfies the requirement that all three couplings can not be simultaneously less than unity . in fig .
( [ xyzconstraint ] ) we show the region in the plane @xmath101 $ ] allowed by the unitarity constraint in eq .
( [ xyzconstraint ] ) , imposing @xmath102 and the constraint on @xmath83 in eq .
( [ bsy_xy ] ) .
it can be seen that there is no parameter space where both of @xmath82 and @xmath45 are less than unity , and the parameter of interest to us ( i.e. @xmath99 ) is fully compatible with the unitarity constraint in eq .
( [ xyzconstraint ] ) .
this constraint on the couplings is removed in a 4hdm due to the presence of a fourth vacuum expectation value .
$ ] plane allowed by the unitarity constraint of eq .
( [ xyzconstraint ] ) .
the constraint from @xmath8 is shown as @xmath103 for re@xmath104 , and @xmath105 for re@xmath106 . ]
the a2hdm is a 2hdm in which nfc is not imposed @xcite , and both scalar doublets ( @xmath107 and @xmath108 ) couple to all types of fermions .
tree - level fcncs are eliminated by imposing an alignment of the yukawa couplings of @xmath107 and @xmath108 .
the interaction of @xmath0 with the fermions in the a2hdm can be written in the same way as eq .
( [ lagrangian ] ) , but the couplings @xmath27 , @xmath28 , and @xmath43 are determined by five parameters ( instead of the four parameters in the democratic 3hdm ) and the unitarity constraint of eq .
( [ xyzconstraint ] ) does not apply .
apart from these two differences ( which were discussed in @xcite ) , the phenomenology of @xmath0 in the democratic 3hdm and the a2hdm is essentially the same , and our numerical results will apply equally to both models . in particular , the magnitudes of @xmath27,@xmath28 and @xmath43 determine the brs of @xmath0 .
in the a2hdm the extra free parameter can be taken to be a phase in the coupling @xmath28 , and such a phase does not have an effect on the brs of @xmath0 .
moroever , we will be concerned with the parameter space of @xmath99 , which is compatible with the unitarity constraint in a 3hdm .
hereafter , when the text refers to `` @xmath0 of the mhdm '' , the implicit meaning is for an @xmath0 of a 3hdm , the a2hdm , and for a mhdm with more than three scalar doublets .
in a mhdm and in the a2hdm the expressions for the partial widths of the decay modes of @xmath0 are : @xmath110 @xmath111 in @xmath112 the running quark masses should be evaluated at the scale of @xmath4 , and there are qcd vertex corrections which multiply the above partial widths by @xmath113 . in the 2hdm the parameter @xmath5 determines the magnitude of the partial widths .
the branching ratios are well known , and for the case of interest of @xmath9 one finds that the dominant decay mode is either @xmath114 or @xmath115 , depending on the value of @xmath5 . in model
i the brs are independent of @xmath5 , and br(@xmath115 ) is about twice that of br(@xmath114 ) .
the magnitude of br(@xmath19 ) is always less than a few percent in three ( models i , ii and lepton - specific ) of the four versions of the 2hdm with nfc , since the decay rate is suppressed by the small ckm element @xmath116 .
in contrast , a sizeable br(@xmath19 ) can be obtained in the flipped 2hdm for @xmath117 .
this possibility was not stated explicitly in @xcite when the flipped 2hdm was discussed .
the first explicit mention of a large br(@xmath19 ) seems to have been in @xcite , and a quantitative study followed soon afterwards in @xcite . as discussed in section
ii , the condition @xmath9 in the flipped 2hdm would require additional new physics in order to avoid the constraint on @xmath4 from @xmath22 , while this is not the case in the mhdm .
a distinctive signal of @xmath0 from a mhdm for @xmath9 would be a sizeable branching ratio for @xmath19 . for @xmath24 ,
the scenario of @xmath99 in a mhdm gives rise to a `` leptophobic '' @xmath0 with br@xmath118+br@xmath119 .
consequently , br(@xmath115 ) is negligible ( @xmath120 ) .
the other decays of @xmath0 to quarks are subdominant , with br(@xmath121 and br(@xmath122 ) only becomes sizeable for @xmath123 , as can be seen in the numerical analysis in @xcite in the flipped 2hdm .
note that the case of @xmath99 is obtained in the flipped 2hdm for @xmath117 , because @xmath124 in this model . in the scenario of @xmath99
the ratio of the two dominant decays , br@xmath25 and br@xmath118 , approaches a constant value , which is given as follows : @xmath125 the ckm elements are well measured , with @xmath126 ( a direct measurement ) and @xmath127 ( from the assumption that the ckm matrix is unitary ) .
the running quark masses @xmath128 and @xmath129 should be evaluated at the scale @xmath130 , and this constitutes the main uncertainty in the ratio @xmath131 .
there is relatively little uncertainty for @xmath129 , with @xmath132 gev .
there is more uncertainty in the value of @xmath128 , although in recent years there has been much progress in lattice calculations of @xmath128 , and an average of six distinct unquenched calculations @xcite gives @xmath133 mev @xcite at the scale of @xmath134 gev .
a more conservative average of these six calculations , @xmath135 mev , is given in @xcite .
in @xcite the currently preferred range at @xmath134 gev is given as @xmath136 mev .
using @xmath137 mev at the scale of @xmath134 gev ( i.e. roughly the central value of the lattice averages in @xcite ) one obtains @xmath138 mev .
taking @xmath139 mev and @xmath140 mev at @xmath134 gev one obtains @xmath141 mev and @xmath142 mev respectively at @xmath143 gev .
smaller values of @xmath128 will give a larger br@xmath25 , as can be seen from eq .
( [ cbcsratio ] ) . note that the value @xmath144 mev is significantly smaller than the typical values @xmath145 mev which were often used in higgs phenomenology in the past two decades .
we emphasise that the scenario of @xmath99 with @xmath9 has a unique feature that the magnitude of @xmath128 is crucial for determining the relative magnitude of the two dominant decay channels of @xmath0 .
this is not the case for most other non - minimal higgs sectors with @xmath0 that are commonly studied in the literature . in @xcite the magnitude of br@xmath25 in the mhdm
was studied in the plane of @xmath82 and @xmath93 , for @xmath146 and @xmath147 , taking @xmath148 gev and @xmath149 gev . with these quark masses
the maximum value is @xmath150 , which corresponds to br@xmath151 .
however , the values of @xmath148 gev and @xmath149 gev are not realistic ( as was subsequently noted in @xcite ) , and two recent papers @xcite have updated the magnitude of @xmath131 in the flipped 2hdm using realistic running quark masses at the scale of @xmath4 . in @xcite , it appears that @xmath152 gev at the scale of @xmath4 was used , which gives br@xmath153 , in agreement with our results . in @xcite , @xmath154 gev at the scale of @xmath4
was used , with a maximum value for br@xmath25 of @xmath155 .
we note that none of these papers used the precise average @xmath133 mev @xcite of the lattice calculations , which gives @xmath156 mev at the scale of @xmath4 .
this smaller value of @xmath128 leads to a maximum value of br@xmath25 which is larger than that given in @xcite , as discussed below .
we now study the magnitude of @xmath19 as a function of the couplings @xmath157 . in fig .
( [ fig : brcb]a ) we update the numerical study of @xcite for br@xmath25 in the plane @xmath158 $ ] in a mhdm with @xmath159 , using @xmath160 gev and @xmath161 gev at the scale of @xmath162 gev . with these values for
the quark masses the maximum value is br@xmath163 i.e. a significantly larger value than br@xmath151 in @xcite . taking a lower value of @xmath164 gev one has br@xmath165 , and for @xmath166 gev one has br@xmath167 . in fig .
( [ fig : brcb]a ) we also display the bound from @xmath22 ( for @xmath91 gev ) , which is @xmath168 for @xmath83 being real and negative , and @xmath169 for @xmath83 being real and positive .
the parameter space for br@xmath170 roughly corresponds to @xmath171 and @xmath172 for @xmath169 . in fig .
( [ fig : brcb]b ) and fig .
( [ fig : brcb]c ) we show br@xmath118 and br@xmath173 respectively .
as expected , br@xmath118 is maximised for @xmath174 while br@xmath173 is maximised for @xmath175 . in fig .
( [ fig : brcb_xz ] ) we show contours of br@xmath25 in the plane @xmath176 $ ] for @xmath162 gev and @xmath177 . for this value of @xmath93
the constraint from @xmath8 is always satisfied for the displayed range of @xmath178 .
one can see that the largest values of br@xmath25 arise for @xmath179 .
the above discussion has assumed that @xmath0 can not decay into other scalars .
we now briefly discuss the impact of the decay channel @xmath182 , which has been studied in the 2hdm ( type ii ) in @xcite and in other 2hdms with small @xmath44 and @xmath45 in @xcite , with direct searches at lep ( assuming @xmath183 ) performed in @xcite . in a general non - susy 2hdm
the masses of the scalars can be taken as free parameters .
this is in contrast to the mssm in which one expects @xmath184 in most of the parameter space .
the scenarios of @xmath181 and @xmath185 are both possible in a 2hdm , but large mass splittings among the scalars lead to sizeable contributions to electroweak precision observables @xcite , which are parametrised by the @xmath186 , @xmath187 and @xmath48 parameters @xcite .
the case of exact degeneracy ( @xmath188 ) leads to values of @xmath186 , @xmath187 and @xmath48 which are almost identical to those of the sm . a recent analysis in a 2hdm @xcite sets @xmath189 and @xmath190)=1 , and studies the maximum value of the mass splitting @xmath191 ( for earlier studies see @xcite ) . for @xmath192
gev the range @xmath193 is allowed , which corresponds to @xmath194 . for @xmath195 gev
the allowed range is @xmath196 which corresponds to @xmath197 .
consequently , sizeable mass splittings ( of either sign ) of the scalars are possible .
analogous studies in a mhdm have been performed in @xcite , with similar conclusions .
if @xmath181 then the decay channel @xmath182 can compete with the above decays of @xmath0 to fermions , because the coupling @xmath198 is not suppressed by any small parameter . in fig .
( [ fig : braw]a ) we show contours of br(@xmath182 ) in the plane @xmath158 $ ] with @xmath159 , @xmath199 gev and @xmath162 gev .
the contours are essentially vertical in the parameter space of interest ( i.e. @xmath200 and @xmath201 ) because the contribution of the term @xmath202 to the decay widths of @xmath0 to fermions is small . comparing fig .
( [ fig : braw]a ) and fig .
( [ fig : brcb]a ) one can see that for @xmath203 both br@xmath204 ) and br@xmath25 are dominant , with roughly equal brs . for smaller @xmath205 ( e.g. @xmath206 ) the contour of br@xmath204)=@xmath207 would move to higher values of @xmath82 .
since the dominant decay of @xmath208 is expected to be @xmath183 , the detection prospects in this channel should also be promising because there would be more @xmath10 quarks from @xmath1 , @xmath180 , @xmath183 than from @xmath1 with @xmath19 .
we note that there has been a search by the tevatron for the channel @xmath1 , @xmath180 , @xmath209 @xcite , for the case of @xmath210 where @xmath183 is not possible @xcite . at present
there is much speculation about an excess of events around a mass of 125 gev in the search for the sm higgs boson @xcite .
an interpretation of these events as originating from the process @xmath211 has been suggested in @xcite . in fig .
( [ fig : braw]b ) we set @xmath212 gev and @xmath213 gev . since the mass splitting between @xmath0 and @xmath208 is less than in fig .
( [ fig : braw]a ) , the contours move to lower values of @xmath82 , but br@xmath204)=@xmath207 is still possible for @xmath214 .
we note that if the excess of events at 125 gev is attributed to a sm - like higgs , then in the context of a 2hdm a candidate would be the lightest cp - even higgs @xmath11 with a coupling to vector bosons of sm strength ( recent studies of this possibility can be found in @xcite ) .
this scenario would correspond to @xmath215 in a 2hdm , and therefore the coupling @xmath216 ( with a magnitude @xmath217 ) in a 2hdm ) would be close to zero . hence the decay @xmath218 would be suppressed by this small coupling , as well as by the virtuality of @xmath219 .
several recent studies @xcite fit the current data in all the higgs search channels to the case of a neutral higgs boson with arbitrary couplings .
a sm - like higgs boson gives a good fit to the data , although a slight preference for non - sm like couplings is emphasised in @xcite .
if the excess of events at 125 gev turns out to be genuine _ and _ is well described by a non - sm like higgs boson of a 2hdm with a value of @xmath220 which is significantly less than unity , then br(@xmath218 ) could be sizeable , with a magnitude given by fig .
( [ fig : braw]b ) after scaling by @xmath221 . ) , br@xmath222 ) and br@xmath223 ) in the plane [ @xmath27 , @xmath28 ] with @xmath159 .
the constraint from @xmath8 is shown as @xmath103 for re@xmath104 , and @xmath105 for re@xmath106 .
we take @xmath225 gev and @xmath162 gev .
, title="fig : " ] ) , br@xmath222 ) and br@xmath223 ) in the plane [ @xmath27 , @xmath28 ] with @xmath159 .
the constraint from @xmath8 is shown as @xmath103 for re@xmath104 , and @xmath105 for re@xmath106 .
we take @xmath225 gev and @xmath162 gev .
, title="fig : " ] ) , br@xmath222 ) and br@xmath223 ) in the plane [ @xmath27 , @xmath28 ] with @xmath159 .
the constraint from @xmath8 is shown as @xmath103 for re@xmath104 , and @xmath105 for re@xmath106 .
we take @xmath225 gev and @xmath162 gev .
, title="fig : " ] ) in the plane [ @xmath27 , @xmath43 ] with @xmath177 .
we take @xmath225 gev and @xmath162 gev . ] ) in the plane [ @xmath27 , @xmath28 ] with @xmath159 .
the constraint from @xmath8 is shown as @xmath103 for re@xmath104 , and @xmath105 for re@xmath106 . in the left panel ( a )
we take @xmath162 gev and @xmath199 gev , and in the right panel ( b ) we take @xmath213 gev and @xmath212 gev . in both figures
@xmath225 gev.,title="fig : " ] ) in the plane [ @xmath27 , @xmath28 ] with @xmath159 .
the constraint from @xmath8 is shown as @xmath103 for re@xmath104 , and @xmath105 for re@xmath106 . in the left panel ( a )
we take @xmath162 gev and @xmath199 gev , and in the right panel ( b ) we take @xmath213 gev and @xmath212 gev . in both figures
@xmath225 gev.,title="fig : " ]
the case of @xmath227 with a large br(@xmath114 ) can be searched for in the decays of the top quark via @xmath1 @xcite . the first discussion of @xmath1 followed by the decay @xmath19 was given in @xcite .
recently , @xmath1 with decay @xmath19 has been studied in the context of flipped 2hdm @xcite , and in the context of the 2hdm without natural flavor conservation @xcite .
there have been two dedicated searches by the tevatron collaborations @xcite for @xmath1 followed by @xmath114 .
the d0 analysis @xcite with 1 fb@xmath18 performed a search for @xmath1 by studying the effect of the decay @xmath114 on ratios of cross sections for @xmath228 production . in the sm one has br(@xmath229)= 100% , and
the branching ratios of @xmath230 and @xmath231 are known .
the presence of a sizeable br(@xmath1 ) with @xmath114 would change the sm prediction for the ratio of the cross sections for the channels with decay @xmath230 and @xmath231 . for the optimum case of br(@xmath232 , upper bounds on br@xmath233 between 0.19 and 0.22
were obtained for @xmath234 gev .
although the decay @xmath114 was assumed in @xcite the above limits also apply ( to a very good approximation ) to the case of both @xmath114 and @xmath19 having sizeable brs , as discussed in @xcite .
this is because the search strategy merely requires that @xmath0 decays to quark jets .
an alternative strategy was adopted in the cdf analysis @xcite with 2.2 fb@xmath18 .
a direct search for the decay @xmath114 was performed by looking for a peak centered at @xmath235 in the dijet invariant mass distribution , which would be distinct from the peak at @xmath236 from the sm decay @xmath229 with @xmath231 . for the optimum case of br(@xmath237 , upper bounds on br@xmath233 between 0.32 and 0.08
were obtained for @xmath238 gev , with the greatest sensitivity being at @xmath239 gev .
no limits on br@xmath233 were given for the region @xmath240 gev due to the large background from @xmath231 decays . for the region @xmath241 gev , limits on br@xmath233 between 0.09 and 0.12
were derived . as stated in @xcite
, the above limits also apply to other hadronic decays of @xmath0 , although with slight changes in the sensitivity to br@xmath233 because the dijet mass resolution depends mildly on the flavour of the quarks .
the search strategy in @xcite and does not have sensitivity to the region @xmath242 gev due to the large background from @xmath243 .
the combination of the four searches at lep for @xmath244 @xcite derived the limit @xmath245 gev for the scenario of br(@xmath246 , with the following additional small intervals excluded ( at @xmath247 c.l ) : @xmath248 gev and @xmath249 gev .
therefore the region of @xmath242 gev with @xmath0 decaying dominantly to quarks ( e.g. @xmath250 ) has not yet been entirely excluded yet . concerning the prospects at the lhc , there has been a simulation of @xmath1 followed by the decay @xmath114 by the atlas collaboration in @xcite , assuming that one of the top / antitop quarks in the @xmath228 events decays leptonically via @xmath251 .
this strategy is very similar to the cdf analysis of @xcite , and directly looks for a peak centered at @xmath4 in the invariant mass distribution of the jets from @xmath3 .
two @xmath10-tags are applied , and the peak from @xmath3 is obtained by reconstructing the two untagged jets .
the mass resolution of the peak can be further improved by full reconstruction of the @xmath228 event . for @xmath252 tev with 1 fb@xmath18 of luminosity
, values of br@xmath233 as low as 0.04 can be probed for @xmath253 .
this sensitivity is superior to that achieved for the decay @xmath1 followed by @xmath2 with the same integrated luminosity @xcite .
again , as in @xcite there is little or no sensitivity to the region @xmath254 gev .
the first search for @xmath1 followed by the decay @xmath114 at the lhc has been performed by the atlas collaboration with 0.035 fb@xmath18 in @xcite . due to the small amount of integrated luminosity , only one @xmath10-tag was applied .
the limits on br@xmath233 are comparable to those from the tevatron search in @xcite , with limits of br(@xmath255 and @xmath256 for @xmath257 and @xmath258 gev respectively .
if br@xmath25 were the dominant decay channel , as can be the case in the mhdm and the a2hdm , the requirement of tagging the @xmath10 from @xmath19 ( as suggested in @xcite ) would provide sensitivity to br@xmath233 in the problematic region @xmath254 gev , and should improve the sensitivity for @xmath259 gev .
we now estimate the gain in sensitivity using realistic values for the @xmath10-tagging efficiency ( @xmath260 ) , the probability of a @xmath261-quark being misidentified as a @xmath10 quark ( @xmath262 ) and the probability of a light quark being misidentified as a @xmath10-quark ( @xmath263 ) .
the two dominant backgrounds to the peak at @xmath4 in the dijet invariant mass distribution are from @xmath264 and @xmath265 , which we take to be equal in magnitude . for the case of br@xmath20 near @xmath7 , the ratio of the signal to the background ( @xmath266 ) with and without the @xmath10-tag
is given approximately as follows : @xmath267_{\rm btag}}{[s/\sqrt b]_{\rm \not{btag } } } \sim \frac{\epsilon_b\sqrt 2}{\sqrt{(\epsilon_j+\epsilon_c)}}\sim 2.13 \ , .
\label{gain_from_btag}\ ] ] we encourage a detailed simulation by the tevatron and lhc collaborations in order to obtain a more realistic estimate of the increase in sensitivity over the current strategy of not applying a @xmath10-tag to the jets originating from @xmath0 . we note that a recent paper @xcite has performed a simulation for a very similar signature which arises from the decay @xmath268 in a different 2hdm with fcncs .
this signature looks identical to the signature arising from @xmath6 but there are several kinematical differences .
the process @xmath269 would give a peak at @xmath4 in the dijet invariant mass distribution in which only one of jets has originated from a @xmath10 quark , with the other two @xmath270jets coming from the decay of @xmath228 .
in contrast , for @xmath268 both of jets in the dijet invariant mass distribution would originate from @xmath10 quarks , while the third @xmath270jet would come from the decay of @xmath271 or @xmath272 .
the study in @xcite is specifically for @xmath268 , and it was found that the sensitivity to br(@xmath273 ) was significantly superior to that for @xmath1 followed by @xmath3 , which can be attributed to the extra @xmath10-tag i.e. the increase in sensitivity in @xcite compared to that obtained for the lhc simulation without the @xmath10-tag in @xcite is significantly greater than the value of 2.13 in eq .
( [ gain_from_btag ] ) , and could be as large as a factor of six .
we now quantify the magnitude of @xmath19 events produced in the decays of @xmath271 quarks , and compare this with the expected sensitivity at the lhc . for the partial decay widths of @xmath274 and @xmath1 we use the leading - order expressions ( with @xmath275 ) as follows : @xmath276[1-m_w^2/m_t^2]^2 \\
\nonumber \gamma(t\to h^\pm b)=\frac{g_f m_t}{8\sqrt 2 \pi}[m^2_t|y|^2 + m_b^2|x|^2][1-m_{h^\pm}^2/m_t^2]^2\end{aligned}\ ] ] the multiplicative ( vertex ) qcd corrections to both @xmath274 and @xmath1 essentially cancel out in the ratio of partial widths @xcite , and thus they do not affect br@xmath233 significantly . in the phase - space function of both decays we neglect @xmath129 , and in the terms @xmath277 and @xmath278 we use @xmath279 gev and @xmath129 evaluated at the scale of @xmath280 ( i.e. @xmath281 gev ) . in fig .
( [ fig : brcbcs]a ) and fig .
( [ fig : brcbcs]b ) we show contours of the sum of @xmath282 \label{cbcs}\ ] ] in the plane of @xmath158 $ ] for @xmath28380 gev and @xmath283120 gev respectively , setting @xmath159 .
the cross section in eq .
( [ cbcs ] ) is the signature to which the current search strategy at the tevatron and the lhc is sensitive , i.e. one @xmath10-tag ( lhc @xcite ) or two @xmath10-tags ( tevatron @xcite ) are applied to the jets originating from @xmath228 decay , but no @xmath10-tag is applied to the jets originating from @xmath0 . for the case of [ br(@xmath284+br(@xmath285=100%
the current experimental limits for @xmath162 gev are @xmath286 from atlas with 0.035 fb@xmath18 @xcite , @xmath287 from cdf with 2.2 fb@xmath18 @xcite , and @xmath288 from d0 with 1 fb@xmath18 @xcite . in fig .
( [ fig : brcbcs]b ) for @xmath283120 gev these upper limits would exclude the parameter space of @xmath289 and small @xmath93 which is not excluded by the constraint from @xmath8 . for the mass region @xmath242 gev there is only a limit from the d0 search in @xcite , which gives br@xmath290 . from fig .
( [ fig : brcbcs]a ) , for @xmath291 gev one can see that this limit excludes the parameter space of @xmath292 and small @xmath93 . in both fig .
( [ fig : brcbcs]a ) and fig .
( [ fig : brcbcs]b ) we show contours of @xmath293 , which might be reachable in the 8 tev run of the lhc .
simulations by atlas ( with @xmath252 tev ) for @xmath3 @xcite have shown that the lhc should be able to probe values br@xmath294 with 1 fb@xmath18 for @xmath295 gev , with the greatest sensitivity being around @xmath239 gev .
for the operation with @xmath296 tev and an anticipated integrated luminosity of 15 fb@xmath18 one expects increased sensitivity ( e.g. br@xmath297 for @xmath295 gev ) , although the region @xmath298 gev might remain difficult to probe with the strategy of reconstructing the jets from @xmath0 . an alternative way to probe the region
@xmath298 gev is to use the search strategy by d0 in @xcite , and presumably the lhc could improve on the tevatron limit on br@xmath299 for @xmath298 gev . from fig .
( [ fig : brcbcs]b ) ( for @xmath162 gev ) one can see that the region of @xmath300 and @xmath301 , which is not excluded by @xmath22 , would be probed if sensitivity to br@xmath297 were achieved .
however , a large part of the region roughly corresponding to @xmath302 and @xmath303 ( which is also not excluded by @xmath22 ) would require sensitivity to br@xmath304 in order to be probed with the current search strategy for @xmath1 , and this is probably unlikely in the 8 tev run of the lhc .
increased sensitivity to the plane of @xmath158 $ ] can be achieved by requiring a @xmath10-tag on the jets which originate from the decay of @xmath0 . in figs .
( [ fig : brtbcb80 ] ) and ( [ fig : brtbcb120 ] ) , for @xmath291 gev and @xmath162 gev respectively , we show contours of @xmath305 with the extra @xmath10-tag , as described in eq .
( [ gain_from_btag ] ) , the sensitivity should reach @xmath306 , and perhaps as low as @xmath307 . in the latter case , one can see from figs .
( [ fig : brtbcb80]b ) and ( [ fig : brtbcb120]b ) that a large part of the regions of @xmath308 ( for @xmath162 gev ) and @xmath309 ( for @xmath291 gev ) could be probed , even for @xmath302
. therefore there would be sensitivity to a sizeable region of the parameter space of @xmath158 $ ] which is not excluded by @xmath8 , a result which is in contrast to the above case where no @xmath10-tag is applied to the @xmath10-jets originating from @xmath0 .
we encourage a dedicated search for @xmath1 and @xmath19 by the tevatron and lhc collaborations .
such a search would be a well - motivated extension and application of the searches which have already been carried out in @xcite and @xcite , and would offer the possibility of increased sensitivity to the fermionic couplings and mass of @xmath0 in the a2hdm and a mhdm . ) and br@xmath310 ) in the plane [ @xmath27 , @xmath28 ] with @xmath159 , where @xmath291 gev ( left panel ) and @xmath162 gev ( right panel).,title="fig : " ] ) and br@xmath310 ) in the plane [ @xmath27 , @xmath28 ] with @xmath159 , where @xmath291 gev ( left panel ) and @xmath162 gev ( right panel).,title="fig : " ] ) in the plane [ @xmath27 , @xmath28 ] with @xmath159 for @xmath291 gev .
the constraint from @xmath8 is shown as @xmath103 for re@xmath104 , and @xmath105 for re@xmath106 .
we take @xmath225 gev and show the range @xmath311 ( left panel ) and @xmath312 ( right panel ) .
, title="fig : " ] ) in the plane [ @xmath27 , @xmath28 ] with @xmath159 for @xmath291 gev .
the constraint from @xmath8 is shown as @xmath103 for re@xmath104 , and @xmath105 for re@xmath106 .
we take @xmath225 gev and show the range @xmath311 ( left panel ) and @xmath312 ( right panel ) . , title="fig : " ] ) in the plane [ @xmath27 , @xmath28 ] with @xmath159 for @xmath162 gev .
the constraint from @xmath8 is shown as @xmath103 for re@xmath104 , and @xmath105 for re@xmath106 .
we take @xmath225 gev and show the range @xmath311 ( left panel ) and @xmath312 ( right panel ) .
, title="fig : " ] ) in the plane [ @xmath27 , @xmath28 ] with @xmath159 for @xmath162 gev .
the constraint from @xmath8 is shown as @xmath103 for re@xmath104 , and @xmath105 for re@xmath106 .
we take @xmath225 gev and show the range @xmath311 ( left panel ) and @xmath312 ( right panel ) .
, title="fig : " ]
light charged higgs bosons ( @xmath0 ) are being searched for in the decays of top quarks ( @xmath1 ) at the tevatron and at the lhc .
separate searches are being carried out for the decay channels @xmath114 and @xmath115 , with comparable sensitivity to the mass and fermionic couplings of @xmath0 .
the searches for @xmath114 in @xcite and @xcite look for a peak at @xmath4 in the dijet invariant mass distribution , with the assumption that neither of the quarks is a @xmath10 quark . in some models with two or more higgs doublets ( the aligned 2hdm and a mhdm with three or more scalar doublets ) the branching ratio for @xmath19 can be as large as @xmath7 .
moreover , such a @xmath0 could be light enough to be produced via @xmath1 , as well as respect the stringent constraints from @xmath8 on both @xmath4 and the fermionic couplings of @xmath0 .
this is in contrast to @xmath0 in other 2hdms for which a large branching ratio for @xmath19 is possible ( such as the flipped 2hdm for @xmath24 ) , but one expects @xmath21 in order to comply with the measured value of @xmath8 . in the context of the aligned 2hdm and a mhdm we suggested that a dedicated search for @xmath1 and @xmath19 would probe values of the fermionic couplings of @xmath0 which are currently not excluded by measurements of @xmath8 .
such a search would require a @xmath10-tag of one of the jets originating from @xmath0 , and would afford sensitivity to a smaller value of the branching ratio of @xmath1 than that obtained in the ongoing searches , which currently do not make use of this additional @xmath10-tag .
we emphasised that a dedicated search for @xmath1 and @xmath19 at the tevatron and lhc would be a well - motivated and ( possibly ) straightforward extension of the ongoing searches for @xmath1 with decay @xmath114 .
a.g.a was supported by a marie curie incoming international fellowship , fp7-people-2009-iif , contract no . 252263 .
j. h .- s acknowledges the financial support of sni , promep and viep - buap .
s.m is supported in part by the next institute .
j. h .- s .
thanks the university of southampton and the rutherford appleton laboratory for hospitality during his visit to the next institute where part of this work was carried out .
useful comments from a. stuart and a. lytle are gratefully acknowledged .
t. d. lee , phys .
d * 8 * , 1226 ( 1973 ) ; h. e. haber , g. l. kane and t. sterling , nucl .
b * 161 * , 493 ( 1979 ) ; j. f. gunion , h. e. haber , g. l. kane and s. dawson , `` the higgs hunter s guide , '' front . phys . *
80 * , 1 ( 2000 ) .
m. aoki , r. guedes , s. kanemura , s. moretti , r. santos and k. yagyu , phys . rev .
d * 84 * , 055028 ( 2011 ) .
j. f. gunion , h. e. haber , f. e. paige , w. k. tung and s. s. d. willenbrock , nucl .
b * 294 * , 621 ( 1987 ) ; j. l. diaz - cruz and o. a. sampayo , phys . rev .
d * 50 * , 6820 ( 1994 ) ; s. moretti and d. p. roy , phys .
b * 470 * , 209 ( 1999 ) ; d. j. miller , s. moretti , d. p. roy and w. j. stirling , phys . rev .
d * 61 * , 055011 ( 2000 ) .
h. e. logan and d. maclennan , phys .
d * 81 * , 075016 ( 2010 ) .
j. l. diaz - cruz , j. hernandez sanchez , s. moretti , r. noriega - papaqui and a. rosado , phys .
d * 79 * , 095025 ( 2009 ) .
m. aoki , s. kanemura , k. tsumura and k. yagyu , phys . rev .
d * 80 * , 015017 ( 2009 ) .
w. s. hou and r. s. willey , phys .
b * 202 * , 591 ( 1988 ) ; t. g. rizzo , phys .
d * 38 * , 820 ( 1988 ) ; b. grinstein , r. p. springer and m. b. wise , phys .
b * 202 * , 138 ( 1988 ) ; b. grinstein , r. p. springer and m. b. wise , nucl .
b * 339 * , 269 ( 1990 ) .
f. borzumati and c. greub , phys .
d * 58 * , 074004 ( 1998 ) ; f. borzumati and c. greub , phys .
d * 59 * , 057501 ( 1999 ) .
m. misiak , h. m. asatrian , k. bieri , m. czakon , a. czarnecki , t. ewerth , a. ferroglia and p. gambino _ et al .
_ , phys .
lett . * 98 * , 022002 ( 2007 ) .
t. p. cheng and m. sher , phys .
d * 35 * , 3484 ( 1987 ) .
r. m. barnett , g. senjanovic , l. wolfenstein and d. wyler , phys .
b * 136 * , 191 ( 1984 ) ; r. m. barnett , g. senjanovic and d. wyler , phys .
d * 30 * , 1529 ( 1984 ) .
a. g. akeroyd , phys .
b * 377 * , 95 ( 1996 ) ; a. g. akeroyd , j. phys .
g * 24 * , 1983 ( 1998 ) .
s. durr , z. fodor , c. hoelbling , s. d. katz , s. krieg , t. kurth , l. lellouch and t. lippert _ et al .
_ , phys .
b * 701 * , 265 ( 2011 ) ; c. mcneile , c. t. h. davies , e. follana , k. hornbostel and g. p. lepage , phys .
d * 82 * , 034512 ( 2010 ) ; j. laiho and r. s. van de water , arxiv:1112.4861 [ hep - lat ] ; a. bazavov _ et al . _ [ the milc collaboration ] , pos lat * 2009 * , 079 ( 2009 ) ; t. blum , r. zhou , t. doi , m. hayakawa , t. izubuchi , s. uno and n. yamada , phys .
d * 82 * , 094508 ( 2010 ) ; c. kelly , arxiv:1201.0706 [ hep - lat ] .
k. nakamura _ et al .
_ [ particle data group collaboration ] , j. phys .
g * 37 * , 075021 ( 2010 ) .
s. moretti and w. j. stirling , phys .
b * 347 * , 291 ( 1995 ) [ erratum - ibid .
b * 366 * , 451 ( 1996 ) ] ; a. djouadi , j. kalinowski and p. m. zerwas , z. phys .
c * 70 * , 435 ( 1996 ) .
j. abdallah _ et al .
_ [ delphi collaboration ] , eur .
j. c * 34 * , 399 ( 2004 ) ; g. abbiendi _ et al . _ [ opal collaboration ] , arxiv:0812.0267 [ hep - ex ] .
d. toussaint , phys .
d * 18 * , 1626 ( 1978 ) ; s. bertolini , nucl .
b * 272 * , 77 ( 1986 ) .
m. e. peskin and t. takeuchi , phys .
lett . * 65 * , 964 ( 1990 ) ; m. e. peskin and t. takeuchi , phys .
d * 46 * , 381 ( 1992 ) .
r. dermisek and j. f. gunion , phys .
d * 79 * , 055014 ( 2009 ) .
g. burdman , c. haluch and r. matheus , arxiv:1112.3961 [ hep - ph ] .
p. m. ferreira , r. santos , m. sher and j. p. silva , arxiv:1112.3277 [ hep - ph ] ; a. arhrib , r. benbrik and n. gaur , arxiv:1201.2644 [ hep - ph ] .
d. carmi , a. falkowski , e. kuflik and t. volansky , arxiv:1202.3144 [ hep - ph ] ; a. azatov , r. contino and j. galloway , jhep * 1204 * , 127 ( 2012 ) ; j. ellis and t. you , arxiv:1204.0464 [ hep - ph ] .
e. gabrielli , b. mele and m. raidal , arxiv:1202.1796 [ hep - ph ] ; p. p.
giardino , k. kannike , m. raidal and a. strumia , arxiv:1203.4254 [ hep - ph ] .
r. m. barnett , r. cruz , j. f. gunion and b. hubbard , phys .
d * 47 * , 1048 ( 1993 ) .
[ lep higgs working group for higgs boson searches and aleph and delphi and l3 and opal collaborations ] , hep - ex/0107031 . | searches for light charged higgs bosons ( @xmath0 ) in the decay of top quarks , @xmath1 , are being carried out at the lhc and at the tevatron .
it is assumed that the dominant decay channels for such an @xmath0 state are either @xmath2 or @xmath3 , and separate searches are performed with comparable sensitivity to the parameters @xmath4 and @xmath5 of the scalar potential .
the branching ratio for the decay @xmath6 can be as large as @xmath7 in the aligned two higgs doublet model and in models with three or more higgs doublets with natural flavour conservation , while satisfying the constraint from @xmath8 for @xmath9 .
although the current search strategy for @xmath3 is also sensitive to @xmath6 , a considerable gain in sensitivity could be obtained by tagging the @xmath10 quark from the decay @xmath6 .
such an analysis , which could be readily performed at the tevatron and in the 7 tev and 8 tev runs of the lhc , would probe a parameter space of the fermionic couplings of @xmath0 in the above models which at present can not be probed by experimental observables in flavour physics . |
qcd problems with a single heavy quark @xmath0 can be treated in a simpler effective theory hqet , if there exists a 4-velocity @xmath1 such that the heavy - quark momentum is @xmath2 ( @xmath3 is the on - shell mass ) and the characteristic residual momentum is small : @xmath4 .
qcd operators can be written as series in @xmath5 via hqet operators ; the coefficients in these series are determined by matching on - shell matrix elements in both theories . at the tree level , the heavy - quark field @xmath0
is related to the corresponding hqet field @xmath6 ( satisfying @xmath7 ) by @xcite @xmath8 the matrix elements of the bare fields between the on - shell quark with momentum @xmath2 and the vacuum in both theories are given by the on - shell wave - function renormalization constants : @xmath9 ( hqet renormalization constants are denoted by @xmath10 ) .
the dirac spinors are related by the foldy wouthuysen transformation @xmath11 u_v(k)\,.\ ] ] therefore , the bare fields are related by @xmath12\ , , \label{bare}\end{aligned}\ ] ] where the bare matching coefficient is @xmath13 ( we use the covariant gauge : the gauge - fixing term in the lagrangian is @xmath14 , and the free gluon propagator is @xmath15 ; the number of flavours in qcd is @xmath16 ) .
the @xmath17 matching coefficient in ( [ bare ] ) is equal to the leading one , @xmath18 ; this reflexes the reparametrization invariance @xcite .
the @xmath19 renormalized fields are related by the formula similar to ( [ bare ] ) , with the renormalized decoupling coefficient @xmath20 if there are no massive flavours except @xmath0 , then @xmath21 because all loop corrections are scale - free .
the qcd on - shell renormalization constant @xmath22 contains the single scale @xmath3 in this case ; it has been calculated @xcite up to three loops .
the three - loop @xmath19 anomalous dimensions of @xmath6 @xcite and @xmath0 @xcite are also known .
we have to express all three quantities @xmath23 , @xmath24 , @xmath25 via the same variables , say , @xmath26 , @xmath27 , see @xcite . the explicit result for the renormalized matching coefficient @xmath28
can be found in @xcite .
gauge dependence first appears at three loops , as in @xmath22 @xcite .
the requirement of finiteness of the renormalized matching coefficient ( [ renorm ] ) at @xmath29 has allowed the authors of @xcite to extract @xmath30 from their result for @xmath22 . in the large-@xmath31 limit ( see chapter 8 in @xcite for a pedagogical introduction ) : @xmath32 where @xmath33 , @xmath34 ( differences of @xmath35-flavour and @xmath36-flavour quantities can be neglected at the @xmath37 order ) . the difference of the qcd and hqet anomalous dimensions @xmath38 ( it is gauge invariant at this order ) and the borel image @xmath39 are @xcite @xmath40\ , .
\nonumber\end{aligned}\ ] ] this borel image has infrared renormalon poles at each positive half - integer @xmath41 and at @xmath42 .
therefore , the integral in ( [ largeb0 ] ) is not well defined . comparing its residue at the leading pole @xmath43 with the residue of the static - quark self - energy at its ultraviolet pole @xmath43 @xcite
, we can express the renormalon ambiguity of @xmath28 as @xmath44 ( @xmath45 is the ground - state meson residual energy ) .
the matching coefficient is gauge invariant at the order @xmath37 .
expanding @xmath46 and @xmath39 and integrating , we obtain confirm the contributions with the highest power of @xmath35 in each term in our three - loop result , and predict such a contribution at @xmath47 . numerically , in the landau gauge at @xmath48 @xmath49 ( @xmath31 is for @xmath48 flavours ) .
naive nonabelianization @xcite works rather well at two and three loops
. numerical convergence of the series is very poor ; this is related to the infrared renormalon at @xmath43 .
now let us consider the relation between the @xmath19 renormalized electron field in qed and the bloch nordsieck electron field .
the bare matching coefficient @xmath50 is gauge invariant to all orders , see @xcite . in the bloch - nordsieck model , due to exponentiation , @xmath51 ( where the 0-flavour @xmath52 is equal to the on - shell @xmath53 ) . in the full qed , @xmath54 ( see @xcite for the proof ; this has been demonstrated up to four loops by the direct calculation @xcite ) .
the gauge dependence cancels in @xmath55 because of the qed decoupling relation @xmath56 .
therefore , the renormalized matching coefficient @xmath28 in qed is gauge invariant to all orders .
the three - loop result is presented in @xcite .
now we shall consider @xcite @xmath19 renormalized heavy
light qcd quark currents @xmath57 where @xmath58 is a dirac matrix .
they can be expressed via operators in hqet @xmath59 where @xmath60 and @xmath61 are dimension-4 hqet operators with appropriate quantum numbers .
the leading - order matching coefficients @xmath62 have been calculated up to two loops @xcite .
there are 8 dirac structures giving non - vanishing quark currents in 4 dimensions : @xmath63}\,,\quad \gamma_\bot^{[\alpha } \gamma_\bot^{\beta ] } \rlap / v\,,\quad \gamma_\bot^{[\alpha } \gamma_\bot^{\beta } \gamma_\bot^{\gamma]}\,,\quad \gamma_\bot^{[\alpha } \gamma_\bot^{\beta } \gamma_\bot^{\gamma ] } \rlap / v\ , , \nonumber\end{aligned}\ ] ] where @xmath64 .
the last four of them can be obtained from the first four by multiplying by the t hooft
veltman @xmath65 .
we are concerned with flavour non - singlet currents only , therefore , we may also use the anticommuting @xmath66 ( there is no anomaly ) .
the currents renormalized at a scale @xmath67 with different prescriptions for @xmath68 are related by @xcite @xmath69 } q\bigr)_\mu = z_t(\mu ) \bigl(\bar{q } \gamma_5^{\text{hv } } \gamma^{[\alpha } \gamma^{\beta ] } q\bigr)_\mu\ , , \nonumber\end{aligned}\ ] ] where the finite renormalization constants @xmath70 can be reconstructed from the differences of the anomalous dimensions of the currents .
multiplying @xmath58 by @xmath66 does not change the anomalous dimension . in the case of @xmath71}$ ] , multiplying it by @xmath65 just permutes its components , and also does not change the anomalous dimension , therefore , @xmath72 @xmath73 are known up to three loops @xcite .
the anomalous dimension of the hqet current ( [ intro : jtilde ] ) does not depend on the dirac structure @xmath58 .
therefore , there are no factors similar to @xmath74 in hqet . multiplying @xmath58 by @xmath66
does not change the matching coefficient .
therefore , the matching coefficients for the currents in the second row of ( [ match : gamma ] ) can be obtained from those for the first row . in the @xmath1 rest frame @xmath75 in particular , @xmath76}}(\mu)$ ] . in the following
we shall consider only the matching coefficients for the first 4 dirac structures in ( [ match : gamma ] ) . in order to find the coefficients @xmath77 , we equate matrix elements of the left- and right - hand side of ( [ intro : match ] ) from the heavy quark with momentum @xmath2 to the light quark with momentum @xmath78 : @xmath79 the on - shell matrix elements are @xmath80 where @xmath81 and @xmath82 are the bare vertex functions , and @xmath83 differs from @xmath84 because there are no @xmath0 loops in hqet .
the difference between @xmath85 and @xmath86 is of order @xmath87 , and can be neglected .
it is most convenient to use @xmath88 , then the @xmath17 term is absent .
the qcd vertex has two dirac structures : @xmath89 this leads to @xmath90 the hqet vertex has just one dirac structure .
therefore , @xmath91 if all flavours except @xmath0 are massless , all loop corrections to @xmath92 , @xmath30 , and @xmath83 contain no scale and hence vanish : @xmath93 , @xmath94 , @xmath95 .
the quantities @xmath96 , @xmath97 , and @xmath84 contain a single scale @xmath3 ; @xmath97 has been calculated up to 3 loops in @xcite , @xmath84 in @xcite , and @xmath96 in the present work @xcite .
the @xmath19 renormalization constants @xmath98 @xcite and @xmath99 @xcite ( for all @xmath58 ) are also known to 3 loops .
if there is another massive flavour ( @xmath100 in @xmath101-quark hqet ) , then @xmath92 , @xmath30 , and @xmath83 contain a single scale @xmath102 .
the first two quantities have been calculated up to 3 loops in @xcite ; the last one is known from @xcite .
the quantities @xmath96 , @xmath97 , and @xmath84 now contain 2 scales , and are non - trivial functions of @xmath103 .
the renormalization constant @xmath97 has been calculated in this case , up to 3 loops , in @xcite ( the master integrals appearing in this case are discussed in ref .
the other two quantities are found in this work @xcite .
the bare on - shell qcd quantities @xmath104 , @xmath97 , and @xmath84 are expressed via @xmath105 ( and @xmath106 if it is non - zero ; we re - express it via the on - shell mass @xmath102 )
. they do nt contain @xmath67 .
the @xmath19 qcd renormalization constant @xmath99 is expressed via @xmath107 .
the bare on - shell hqet quantities @xmath92 , @xmath30 , and @xmath83 are expressed via @xmath108 and @xmath109 ( they are trivial at @xmath110 ) ; we re - express @xmath109 via the on - shell mass @xmath102 ( which is the same in both theories ) .
these bare quantities also do nt contain @xmath67 .
finally , the @xmath19 hqet renormalization constant @xmath98 is expressed via @xmath111 .
we re - express all the quantities in ( [ match : main ] ) via @xmath111 , see @xcite .
from equation of motion we have @xmath112 where @xmath113 is the @xmath19 mass of the heavy quark @xmath0 . taking the on - shell matrix element between the heavy quark with @xmath114 and the light quark with @xmath115 and re - expressing both qcd matrix elements via the matrix element of the hqet current with @xmath116
, we obtain @xcite @xmath117 the ratio @xmath118 has been calculated at three loops numerically @xcite and then analytically @xcite ( the analytical results @xcite were later independently confirmed in @xcite , and then in several other papers ) ; for @xmath119 , @xmath118 has been found in @xcite .
the matching coefficients have been calculated up to 2 loops in @xcite , and to 3 loops in the present work @xcite .
analytical expressions are long ; numerically , at @xmath110 and @xmath120 we have @xmath121 ( in the middle part of each formula , terms with descending powers of @xmath122 are shown separately ) .
naive nonabelianization @xcite works reasonably well . at @xmath119
, results are expressed via the master integrals depending on @xmath103 @xcite .
their status is summarized in the tables 14 in this paper . in the present work @xcite
, we were able to obtain exact analytical expressions ( via harmonic polylogarithms of @xmath123 ) for @xmath124 terms in the master integrals 5.2 , 5.2a , from the requirement of finiteness of the matching coefficients .
therefore , the table 3 in @xcite should be now replaced with the following table [ tab5 ] ( de means the method of differential equations , and mb the mellin
barnes representation ) .
unfortunately , @xmath125 terms in 4 master integrals are still known only as truncated series in @xmath123 ( the entries @xmath123 in the table ) .
therefore , the @xmath102 corrections to the 3-loop matching coefficients are also known only as truncated series in @xmath123 ( or numerical approximations ) .
we let s apply our results to the matrix elements between a @xmath126 or @xmath127 meson with momentum @xmath128 and the vacuum : @xmath129 the corresponding hqet matrix elements in the @xmath1 rest frame are @xmath130 where the single - meson states are normalized by the non - relativistic condition @xmath131 these two matrix elements are characterized by a single hadronic parameter @xmath132 due to the heavy - quark spin symmetry . from ( [ match : div ] ) we have @xcite @xmath133 where we may replace @xmath134 by the on - shell @xmath101-quark mass @xmath3 , neglecting power corrections .
asymptotics of the perturbative coefficients for the matching coefficients at a large number of loops @xmath141 have been investigated in ref .
@xcite in a model - independent way .
the results contain three unknown normalization constants @xmath142 .
the asymptotics of the perturbative coefficients for @xmath135 contain @xmath143 and @xmath144 ; in the case of @xmath145 it contains only @xmath143 : @xmath146 \frac{n_2}{n_0 } \biggr\ } \nonumber\\ & { } \times\left(\frac{m}{\hat{m}}\right)^{(n+1)}_{l=-5/3}\ , .
\label{asympt}\end{aligned}\ ] ] the coefficient of @xmath147 is about @xmath148 at @xmath149 , and it seems reasonable to neglect this contribution . neglecting also @xmath150 corrections , we obtain @xcite @xmath151 our exact result @xmath152 agrees with this prediction reasonably well .
however , @xmath150 corrections are large and tend to break this agreement .
it is natural to expect that @xmath153 ( and higher ) corrections are also substantial at @xmath149 .
* acknowledgements*. i am grateful to s. bekavac , p. marquard , j.h .
piclum , d. seidel , m. steinhauser for collaboration on @xcite , and to d.j .
broadhurst , k.g .
chetyrkin , d.v .
shirkov , v.a .
smirnov for useful discussions .
larin , j.a.m .
vermaseren , phys .
* b 259 * ( 1991 ) 345 ; s.a .
larin , in _ quarks-92 _ , ed .
grigoriev , v.a .
matveev , v.a .
rubakov , p.g .
tinyakov , world scientific ( 2003 ) p. 201
[ arxiv : hep - ph/9302240 ] . | qcd / hqet matching for the heavy - quark field @xcite and heavy
light quark currents @xcite with three - loop accuracy is discussed . |
in algebraic geometry and number theory , one is frequently interested in abelian varieties with extra endomorphisms , and their moduli spaces .
the study of elliptic curves and their moduli spaces has been extremely influential in the last century ( for instance , see @xcite ) .
higher dimensional abelian varieties have also seen many applications ; however , it has remained quite a challenge to provide explicit computational descriptions for these or their moduli spaces .
there has been quite a lot of work in the direction of explicit approaches to abelian surfaces in the last couple of decades .
this paper is chiefly concerned with moduli spaces of genus @xmath0 curves with decomposable jacobians .
we say that an abelian variety @xmath6 over a field @xmath7 is decomposable or reducible over @xmath7 if it is isogenous over @xmath7 to a product of abelian varieties of smaller dimension .
in particular , we will be interested in principally polarized abelian surfaces @xmath6 which are isogenous to a product of elliptic curves . a natural way to produce
such an abelian surface is as the jacobian of a curve @xmath2 of genus @xmath0 , which has a map of degree @xmath8 to an elliptic curve @xmath3 .
the morphism @xmath9 induces @xmath10 and @xmath11 , with @xmath12 .
the curve @xmath2 also has a map of degree @xmath4 to the complement @xmath13 ( obtained by composing the embedding @xmath14 with the projection map ) and @xmath6 is isogenous to the product @xmath15 with kernel isomorphic to @xmath16 .
we say that @xmath17 is @xmath18-split .
the ( coarse ) moduli space we wish to parametrize is the space @xmath19 of pairs @xmath1 related by an _ optimal _ map of degree @xmath8 .
( here , optimal means that the map does not factor through an unramified cover of @xmath3 . )
it is a double cover of the moduli space @xmath20 of genus @xmath0 curves @xmath2 whose function field has an elliptic subfield of degree @xmath4 ; this latter locus is a hypersurface in @xmath21 , the moduli space of genus @xmath0 curves .
classically , hyperelliptic curves with split jacobians were studied in the guise of the reduction of abelian integrals to elliptic integrals by algebraic transformations , going as far back as legendre and jacobi , who essentially gave a complete description of the moduli space for @xmath22 .
the generic family of curves @xmath2 for @xmath23 was worked out by hermite , goursat , burkhardt , brioschi and bolza , and @xmath24 by bolza .
we refer the reader to @xcite for a summary of this classical literature .
kuhn @xcite revisited this topic , giving a combinatorial description in terms of the branch points of the map @xmath25 as well as the hyperelliptic involution @xmath26 on @xmath2 , and adapting some examples to positive characteristic .
this approach has led to fairly explicit descriptions of @xmath19 for @xmath27 by shaska and others @xcite . from an arithmetic standpoint ,
reducible jacobians for @xmath28 have been studied in @xcite .
frey and kani @xcite have also studied such covers for general @xmath4 by looking at pairs of elliptic curves @xmath3 and @xmath29 such that there is an galois equivariant anti - isometry between their @xmath4-torsion subschemes .
their description connects @xmath19 to several interesting arithmetic questions and applications . in this paper
, we will take a rather different strategy to compute the moduli spaces @xmath19 , which does not focus on the map @xmath9 and its ramification locus .
the key fact upon which our method rests is that a genus @xmath0 curve @xmath2 has a map of degree @xmath4 to an elliptic curve @xmath3 if and only if the point in @xmath30 corresponding to its jacobian @xmath17 lies on the humbert surface @xmath31 ( see @xcite , for instance ) .
in other words , under the birational morphism @xmath32 , the image of @xmath20 is ( birationally ) identified with the humbert surface for discriminant @xmath33 .
recall that for a discriminant @xmath34 , the humbert surface @xmath35 describes principally polarized abelian surfaces which have real multiplication by the quadratic ring of discriminant @xmath34 .
its double cover is the hilbert modular surface @xmath36 which parametrizes pairs @xmath37 , where @xmath6 is a principally polarized abelian surface , and @xmath38 is a ring homomorphism . note that the reason @xmath39 is a double cover is because ( generically ) there are two choices for the action of @xmath40 . in @xcite
, a method was laid out to compute explicit models for these hilbert modular surfaces @xmath36 , relying on the computation of moduli spaces of suitable elliptic k3 surfaces , which are related to the abelian surfaces via a shioda
inose structure . in that paper , equations were given for rational models for the surfaces @xmath36 , for the thirty fundamental discriminants @xmath34 such that @xmath41 .
these hilbert modular surfaces are coarse moduli spaces for principally polarized abelian surfaces with real multiplication by the ring of integers @xmath42 of @xmath43 . in this paper
, we will extend this method to compute the humbert surfaces for _ square _ discriminants @xmath44 , and our desired moduli space @xmath19 ( which is birational to @xmath45 ) , for @xmath5 . to do
so requires several ideas beyond those of @xcite .
first , we generalize the set - up and theorems in @xcite to the case of non - fundamental discriminants .
this allows us to compute @xmath45 as a moduli space of k3 surfaces , but does not elucidate its structure as a moduli space of @xmath1 . in order to do so , we first describe explicitly a normalized tautological family of genus @xmath0 curves over the moduli space , using the method of certifying eigenforms for real multiplication from @xcite .
finally , we describe a new method involving integration of eigenforms to recover the elliptic curves @xmath46 and @xmath47 associated to a given @xmath2 .
our models agree with those in the literature for @xmath48 , and we go well beyond the previous state of the art with the higher values of @xmath4 . in doing so , we exhibit non - rational surfaces @xmath19 : the surfaces @xmath49 and @xmath50 are elliptic k3 , the next three @xmath51 and @xmath52 are honestly elliptic , whereas @xmath53 is a surface of general type .
the geometric classification of these surfaces was studied in @xcite , but explicit algebraic models were not known before this work .
we anticipate that our algebraic models for these moduli spaces will make it easy to carry out explicit arithmetic and geometric investigations .
for instance , one can immediately apply our formulas to produce examples or even families of elliptic curves with anti - isometric @xmath4-torsion galois representations .
the plan of the paper is as follows : in section [ setup ] , we prove the necessary generalizations of the results from @xcite .
each of the following sections treats a separate square discriminant .
we outline the parametrization of the moduli space of elliptic k3 surfaces , and a sequence of `` elliptic hops '' converting to an elliptic fibration with @xmath54 and @xmath55 fibers .
this allows us to compute the map to @xmath30 .
we can then write down the explicit double cover giving a model of @xmath56 .
we carry out some basic geometric analysis of the moduli space , and note any obvious arithmetic curves on it .
we also provide a tautological family of genus @xmath0 curves over @xmath19 , as well as the @xmath57-invariants of both the associated elliptic curves is really the tautological family over the moduli space @xmath19 . ] @xmath46 and @xmath47 .
the auxiliary computer files for this paper , containing various formulas and computations omitted here for lack of space , are available from http://arxiv.org/abs/1412.2849 . to access these , download
the source file for the paper .
this will produce not only the latex file for this paper , but also the computer code . for the computations involved in our work , we made frequent use of the computer algebra systems `
pari - gp ` , ` magma ` , ` maxima ` and ` sage ` . we thank nils bruin , henry cohn , noam elkies , gerard van der geer , david gruenewald , curt mcmullen , ronen mukamel , dan petersen , matthias schtt and the anonymous referee for helpful comments .
this work was supported in part by nsf grant dms-0952486 , and by a grant from the mit solomon buchsbaum research fund .
in this section , we generalize the main theorems of @xcite to the case of non - fundamental discriminant @xmath34 , which is a positive integer congruent to @xmath58 or @xmath59 modulo @xmath60 . in particular , this paper will deal with the case of square discriminant : @xmath44 , in which case the ring in question is an order in a split quadratic algebra ( i.e. @xmath61 ) , rather than a real quadratic field .
we first briefly go over the basic setup for this particular case , since the typical case considered in the literature is that of non - square discriminant . for more background on hilbert modular surfaces see @xcite , and for a quick introduction to the calculations of this section , we refer the reader to @xcite .
the notes @xcite also contain an excellent description of the theory of abelian varieties with extra endomorphisms , and the connection with k3 surfaces which was exploited in @xcite and is also used here .
let @xmath44 , and @xmath42 the quadratic ring of discriminant @xmath34 .
concretely , we can write @xmath62 with componentwise addition and multiplication
. it has a @xmath63-basis @xmath64 given by @xmath65 and @xmath66 .
the two embeddings @xmath67 of @xmath68 into @xmath69 are given by taking the first or second components and composing with @xmath70 .
there are four square roots of @xmath34 in @xmath42 , namely @xmath71 and @xmath72 .
if we label @xmath73 , we see that @xmath59 and @xmath74 also form a @xmath63-basis , just as in the case of non - square discriminant .
the involution which changes @xmath40 to @xmath75 is that which switches the two factors ; i.e. takes @xmath76 to @xmath77 .
the trace map takes @xmath76 to @xmath78 , and the matrix of the trace form with respect to @xmath79 and @xmath80 is @xmath81 which has discriminant @xmath82 .
the dual lattice @xmath83 with respect to the trace form has basis @xmath84 and @xmath85 .
( in fact , for a general discriminant , @xmath83 equals @xmath86 . ) the discriminant group @xmath87 is easily calculated to be @xmath88 if @xmath4 is odd , and @xmath89 if @xmath4 is even .
we now adapt sections @xmath90 and @xmath60 of @xcite to the setting of non - fundamental discriminant .
the outline of the method is as follows . 1 .
the hilbert modular surface @xmath36 is the coarse moduli space of @xmath37 where @xmath6 is a principally polarized abelian surface , and @xmath91 is a homomorphism .
the complex manifold @xmath92 can be obtained by compactifying @xmath93 .
2 . there is a map @xmath94 which simply takes the ppas @xmath6 .
the image @xmath35 is the humbert surface of discriminant @xmath34 in @xmath30 , and the map from the hilbert modular surface has degree @xmath0 .
the key first step is to realize @xmath35 as a moduli space of elliptic k3 surfaces , polarized by a particular lattice @xmath95 of rank @xmath96 and discriminant @xmath97 .
we explicitly parametrize this moduli space @xmath98 , and compute the universal k3 family over it .
we then find a different elliptic fibration on the ( family of ) k3 surfaces , with reducible fibers of type @xmath99 and @xmath100 .
this is accomplished by a sequence of `` elliptic hops '' , which reflect @xmath0- and @xmath90-neighbor steps at the lattice level ( see ( * ? ? ?
* section 5 ) or ( * ? ? ? * appendix a ) ) . by the main result of @xcite , from the final weierstrass equation we may read out the map @xmath101 , which is given in terms of igusa clebsch invariants ( coordinates on @xmath30 @xcite ) .
so far , we have an explicit birational model of @xmath35 . to get to @xmath36
, we must identify the correct double cover .
we first pin down the geometric branch locus as a union of certain modular curves ( by a result of hausmann ) , and interpret these curves in terms of the k3 moduli space ( the picard group jumps in a predictable manner ) .
finally , we find the correct arithmetic twist by point - counting and matching characteristic polynomials of frobenius on the abelian and k3 sides . the first part of ( * ?
* section @xmath90 ) outlines the proof that @xmath93 is the coarse moduli space of principally polarized abelian surfaces with real multiplication by @xmath42 .
the same proof goes through for arbitrary discriminant .
of course , for @xmath44 , usually it is more convenient to understand a ppas with real multiplication by @xmath42 in terms of pairs of elliptic curves which have degree @xmath4 maps from a common genus @xmath0 curve , as described in the introduction .
this is the approach taken in @xcite .
let @xmath6 be a principally polarized abelian surface with @xmath102 .
then @xmath103 .
the lattice @xmath104 has a basis with gram matrix @xmath105 of signature @xmath106 and discriminant @xmath97 .
the proof of @xcite goes through with minor changes , noting that the rosati involution on @xmath107 must be the identity , since the other element of the galois group ( interchange of the factors in the square - discriminant case ) is not a positive involution .
the proof of the next proposition is unchanged from the original .
there is a primitive embedding , unique up to isomorphism , of the lattice @xmath42 into @xmath108 .
let @xmath109 be the orthogonal complement of @xmath42 in @xmath108 .
then there is a primitive embedding , unique up to isomorphism , of @xmath109 into the k3 lattice @xmath110 .
the main theorem of that section identifies the humbert surface with a moduli space of k3 surfaces lattice - polarized by @xmath111 .
let @xmath112 be the lattice @xmath113 .
[ humbert ] let @xmath114 be the moduli space of k3 surfaces that are lattice polarized by @xmath95 .
then the isomorphism @xmath115 of @xcite induces a birational surjective morphism @xmath116 .
its proof relies on a key proposition involving the lattice embedding @xmath117 , where @xmath118 and @xmath119 is the orthogonal complement of @xmath109 in @xmath110 , uniquely described by the proposition above .
the proof of the key proposition and the main theorem go through without any changes .
we must next show that the branch locus of the map @xmath120 corresponds to a divisor @xmath121 in the moduli space of @xmath98 of k3 surfaces lattice polarized by @xmath95 , such that the rank of the k3 surface corresponding to a generic point on @xmath121 jumps to @xmath122 , and the discriminant of the picard group is @xmath123 or @xmath124 . in @xcite ,
the argument used the fact that @xmath34 was fundamental ; we give a more general proof here , valid for arbitrary discriminant .
first , we observe as in @xcite that the branch locus consists of a union of specific modular curves .
let us consider the birational models @xmath125 for @xmath36 ( where @xmath126 ) and @xmath127 for @xmath30 , with @xmath128 being the siegel upper half space for genus @xmath0 .
the two - to - one map @xmath129 factors through @xmath130 , where @xmath131 is the interchange of coordinates on @xmath132 .
the branch locus is the fixed point set of @xmath131 on @xmath125 .
the one - dimensional part of the fixed point set of @xmath131 on @xmath133 is the union of the modular curves @xmath134 , where @xmath135 ranges over @xmath136 if @xmath34 is odd , and over @xmath137 if @xmath34 is even .
next , we analyze what happens to the k3 surfaces along these modular curves .
let @xmath138 be a natural number , and let @xmath2 be any component of the inverse image of a modular curve @xmath139 , under the birational surjective morphism @xmath140 .
the k3 surface corresponding to a generic point of @xmath2 has nron severi group of rank @xmath122 and discriminant @xmath141 . by @xcite , the generic point on any component of a modular curve
@xmath139 corresponds to an abelian surface @xmath6 whose ring of endomorphisms is a quaternionic order @xmath142 of discriminant @xmath143 . by @xcite , the nron
severi group of @xmath6 and the endomorphism ring are connected by the relation @xmath144 , the even part of the associated clifford ring . from this , it is easy to compute that @xmath104 must have rank @xmath90 and discriminant @xmath141 . by the shioda
inose structure , the picard group of the associated k3 surface must have rank @xmath122 and discriminant @xmath141 .
therefore , the part of the branch locus corresponding to the curve @xmath145 ( and @xmath146 if @xmath34 is even ) corresponds to a sublocus of k3 surfaces having picard group of discriminant @xmath124 ( respectively @xmath123 ) .
section @xmath60 of @xcite explains how to correctly identify the branch locus and the arithmetic twist ; we will not restate the method or proofs here .
the discussion for the curves @xmath147 ( and @xmath148 if @xmath34 is even ) was omitted from @xcite ; we address it here .
let @xmath149 be an elliptic k3 surface with a reducible fiber of type @xmath99 and another of type @xmath100 .
furthermore , assume that @xmath149 has a section of height @xmath123 , for some natural number @xmath34 congruent to @xmath58 or @xmath59 modulo @xmath60
. then @xmath150 can not be a lattice of rank @xmath122 and discriminant @xmath0 or @xmath151 .
first , assume that @xmath149 has no other reducible fibers .
then , in particular , there can not be a torsion section : such a section would have height @xmath58 , which is impossible to obtain with only two reducible fibers of type @xmath54 and @xmath55 .
therefore , if @xmath150 had discriminant @xmath0 , the mordell weil lattice would be positive definite of rank @xmath0 and discriminant @xmath59 .
the hermite constant in dimension @xmath0 is @xmath152 , so this would imply that there is a section of height at most @xmath153 , which is impossible with the fiber configuration ( the best we can do is @xmath154 ) . similarly , if @xmath150 had discriminant @xmath151 , we would need a section of height @xmath155 , which is still forbidden .
note that we did not use the assumption of a section of height @xmath123 .
for the case when @xmath149 could have additional reducible fibers , we have to do some additional analysis , but a similar proof applies .
going back to our situation , let @xmath156 be the family of k3 surfaces lattice polarized by @xmath95 , over the base @xmath98 ( the coarse moduli space ) . in the third step of the method described above ,
we find a different elliptic fibration on the k3 surface @xmath157 ( uniformly over @xmath158 on the base ) , with @xmath99 and @xmath100 fibers .
however , according to the above proposition , for @xmath158 on one of the curves @xmath2 lying over @xmath147 or @xmath148 , we can not have such an elliptic fibration on @xmath159 .
therefore , the weierstrass equation for the resulting @xmath160 fibration must be undefined ( or not minimal , or one of these fibers must degenerate further ) for @xmath158 on such a curve @xmath2 .
this is reflected in the denominators of the expressions for the weierstrass coefficients and of the igusa
clebsch invariants , as well as the ( numerator of the ) invariant @xmath161 .
therefore , by adding their irreducible factors to that of the list for discriminant @xmath124 ( and @xmath123 when @xmath34 is even ) , we can then continue with step 4 of @xcite .
the remaining assertions and proofs of section @xmath60 of @xcite go through without any change .
therefore , we may now proceed with the computation of the hilbert modular surfaces @xmath45 for small values of @xmath4 . here
, we describe these surfaces for @xmath4 up to @xmath162 .
the cases @xmath48 have been treated in the previous literature on the subject , although by using completely different methods .
the cases @xmath163 show that one can quickly go beyond the current state of the art using our new techniques . for each @xmath4
treated in this paper , we also describe a tautological family of genus @xmath0 curves over the moduli space , whose jacobians are @xmath18-isogenous to a pair of elliptic curves , and we give the @xmath57-invariants of these elliptic curves .
is rational , we give the @xmath57-invariants in terms of the parameters on the moduli space . for the non - rational moduli spaces , we give @xmath164 and @xmath165 in terms of the parameters on the ( rational ) humbert surface @xmath31 , and verify that @xmath166 generates the function field of its double cover @xmath45 . ] to accomplish this , we use the eigenform location algorithm of @xcite to produce a tautological family of _ normalized _ genus @xmath0 curves ( i.e. for which the eigenforms for real multiplication are given by @xmath167 and @xmath168 ) , along with a method which involves integrating an eigenform . in principle
, we can even give the two maps of degree @xmath4 from the genus @xmath0 curve to the elliptic curves ; we indicate how this is done in the case @xmath23 .
there is no serious obstruction to computing these modular surfaces for @xmath169 , though of course the computations will get more challenging for large @xmath4 .
we list the geometric type of the surfaces treated here in the following table . for comparison with the literature
, we note that hermann @xcite described the geometric classification of modular surfaces @xmath170 of discriminant @xmath33 , which is the quotient of @xmath132 by an appropriate subgroup of @xmath171 . a different proof was given by kani and schanz @xcite , who also corrected a typo in @xcite , and described the connection with the moduli space of pairs of elliptic curves @xmath172 with an isomorphism on the @xmath4-torsion @xmath173 \cong e_2[n]$ ] .
the index @xmath174 is the factor which multiplies the determinant of the weil pairing under this isomorphism .
see also @xcite for a nice overview and connections to modular forms . for our particular situation , @xmath175 as the isomorphism is an anti - isometry ( see @xcite ) , and we can read off the corresponding results from the above papers ( note that the isomorphism type only depends on the square class of @xmath176 in @xmath177 ) . [ cols="^,^",options="header " , ]
we would like a family of k3 surfaces lattice polarized by @xmath178 .
it is easy to see that @xmath179 .
we now reverse - engineer tate s algorithm to find elliptic k3 surfaces with reducible fibers of types @xmath54 , @xmath55 and @xmath180 at @xmath181 , @xmath58 and @xmath59 respectively .
a general elliptic k3 surface with section has the form @xmath182 with @xmath183 being polynomials in @xmath184 of degree @xmath185 respectively . in order to have an @xmath54 fiber at @xmath181 , we may assume ( after shifting @xmath186 suitably ) that @xmath187 has degree at most @xmath0 .
similarly , the @xmath55 fiber at @xmath188 allows us to assume that @xmath189 , for some constant @xmath7 .
these normalizations involve shifting @xmath186 by a linear combination of @xmath190 and @xmath191 . finally ,
since there is an @xmath180 fiber at @xmath192 , we may shift @xmath186 by a suitable multiple of @xmath193 , to make @xmath194 and @xmath195 divisible by @xmath196 and @xmath197 respectively .
the final result is that the weierstrass equation has the form @xmath198 since we must quotient by the weierstrass scaling of @xmath186 and @xmath199 , we see that the resulting moduli space is a weighted projective space @xmath200 in the coordinates @xmath201 .
to compute the map to @xmath30 , we put the surface in the standard form described in @xcite . here
, it merely involves shifting @xmath186 so that the coefficient of @xmath203 is zero , and then rescaling so that the coefficient of @xmath204 is @xmath205 .
we obtain the weierstrass equation @xmath206 we may now read out the igusa
clebsch invariants , which are functions of @xmath207 and @xmath208 .
we obtain @xmath209 by considering the possible ways the rank could jump and give k3 surfaces whose nron
severi lattices have the appropriate discriminant , we obtain the following list of possible factors for the branch locus : @xmath210 an arithmetic verification as in @xcite pins down the correct double cover . a birational model for the surface @xmath202 ( equivalently , for @xmath211 ) is given by @xmath212 it is a rational surface .
the humbert surface is birational to the @xmath213-plane . in these coordinates , the igusa
clebsch invariants of a point on the moduli space are given by the formulas in above .
this is a conic bundle over @xmath214 .
setting @xmath215 makes this expression a square , so in fact this conic bundle has a section , and is a rational surface over the base field @xmath216 . to parametrize it ,
we complete the square by letting @xmath217 . solving for @xmath218
, we obtain @xmath219 the classical formulae over an algebraically closed field are described in @xcite , for instance .
we start with a genus @xmath0 curve @xmath220 which has two obvious maps to elliptic curves .
the group @xmath221 acts on the family of such weierstrass equations as follows : its two generators take @xmath222 to @xmath223 or to @xmath224 , where @xmath225 is a primitive sixth root of unity , @xmath226 and @xmath227 .
the invariants of this action on the polynomial ring generated by the parameters @xmath228 and @xmath229 are @xmath230 and @xmath231 . comparing igusa
clebsch invariants , we obtain the relation between our coordinates @xmath232 above and @xmath233 .
@xmath234 with inverse @xmath235 the branch locus , up to squares , equals @xmath236 . recall that interchanging @xmath228 and @xmath229 changes the equation of the genus @xmath0 curve by @xmath237 , and so switches the two elliptic subfields .
so it agrees with the classical double cover .
another explicit way to see this is through the @xmath57-invariants , which we compute next . over the moduli space @xmath211 ( or @xmath202 )
, it is possible to write down a tautological family of genus @xmath0 curves with @xmath238 reducible jacobian . in terms of @xmath239 above , the sextic defining the family is @xmath240 the @xmath57-invariants corresponding to the two elliptic subfields are @xmath241 which again shows that the involution switching the @xmath57-invariants is the one corresponding to the double cover .
since the hilbert modular surface is rational with parameters @xmath242 , we may further simplify the tautological family above . in terms of the parameters
@xmath243 on the hilbert modular surface , a tautological family of genus @xmath0 curve with @xmath238-reducible jacobian is given by @xmath244 we now describe some special curves on the hilbert modular surface . 1 .
the genus @xmath58 curve @xmath245 is part of the branch locus ; it is also part of the product locus , i.e. the boundary @xmath246 which corresponds to products of elliptic curves rather than jacobians of genus @xmath0 curves .
2 . for the curves @xmath247 , the invariant @xmath248 .
they correspond to degenerations of the genus @xmath0 curve to a rational curve .
3 . the curve @xmath249 on the humbert surface pulls back to a genus @xmath58 curve on @xmath211 , isomorphic to @xmath250 .
it is a modular curve , corresponding to jacobians with endomorphisms by a ( split ) quaternion algebra .
the minimal degree of the isogeny between the two elliptic curves we have associated above is @xmath90 .
this curve is the same as that defined in ( * ? ? ?
* formula ( 11 ) ) .
the curve @xmath251 also lifts to a @xmath250 on the hilbert modular surface , corresponding to the modular curve for which @xmath252 .
the curve @xmath253 corresponds to one of the @xmath57-invariants being @xmath58 ( i.e. having cm by @xmath254 $ ] ) .
the curve @xmath255 has genus @xmath58 and is part of the branch locus .
it is parametrized by @xmath256 and corresponds to the @xmath57-invariants being equal .
the curve @xmath257 lifts to a genus @xmath58 curve on the hilbert modular surface ; its parametrization on the humbert surface is given by @xmath258 , + @xmath259 .
it corresponds to the elliptic curves being @xmath0-isogenous .
there are also several simple non - modular curves , for example the curves @xmath260 , @xmath261 , and @xmath262 .
these lift to genus @xmath58 curves ( in fact , isomorphic to @xmath250 ) on @xmath211 .
since the hilbert modular surface is rational , in fact we have a @xmath0-parameter family of genus @xmath0 curves with @xmath238-split jacobian given by the equation above , and we can specialize it to produce infinitely many examples of @xmath59-parameter families of such genus @xmath0-curves .
to check the statement that elliptic curves corresponding to points on a specific curve on @xmath202 are ( say ) @xmath0-isogenous , one can simply substitute in the @xmath57-invariants into the appropriate classical modular polynomial . another way to verify that such a curve on the hilbert modular surface is modular is to check that the picard number of the elliptic k3 surface jumps ( for instance , due to an extra reducible fiber , or an extra section ) . from now on ,
we shall make such statements without further justification ; the interested reader may carry out the ( easy ) check .
next , we describe degree @xmath90 elliptic subfields , which were treated by shaska in @xcite .
a simple family of elliptic k3 surfaces which will allow us to recover this moduli space is that with @xmath54 and @xmath264 fibers .
as before , we put the @xmath54 fiber at @xmath265 , letting us write the weierstrass equation as @xmath266 with @xmath267 polynomials of degrees @xmath268 respectively . assuming the @xmath264 fiber is at @xmath188 , and shifting @xmath186 by a suitable quadratic polynomial , we can rewrite the weierstrass equation as @xmath269 this surface generically has an @xmath270 fiber at @xmath188 . to ensure an @xmath264 fiber , we need three more orders of vanishing for the discriminant at @xmath188 . since we need the components of the @xmath271 fiber to be defined over the ground field ,
the coefficient @xmath272 is a nonzero square ; by scaling @xmath186 we may assume it is @xmath59 .
similarly , by scaling @xmath184 , we can arrange @xmath273 .
the three orders of vanishing succesively yield @xmath274 , @xmath275 , and @xmath276 .
relabeling @xmath277 and @xmath278 , we obtain the final weierstrass equation @xmath279 over the rational moduli space with parameters @xmath232 . to compute the map to @xmath30 , we go to an elliptic fibration with @xmath54 and @xmath55 fibers via a @xmath0-neighbor step ( see @xcite ) . in terms of dynkin diagrams of rational curves on the k3 surface ,
the picture is as follows : ( -2,0)(7,0)(7.5,0.866)(7.5,0.866)(10.5,0.866)(10.5,-0.866)(7.5 , -0.866)(7,0 ) ; ( 0,0)(0,1 ) ; ( 9.5 , 0.866)(7.5,0.866)(7,0)(7.5,-0.866)(9.5,-0.866 ) ; ( 6,0)(7,0 ) ; ( 6,0 ) circle ( 0.2 ) ; ( -2,0 ) circle ( 0.1 ) ; ( -1,0 ) circle ( 0.1 ) ; ( 0,0 ) circle ( 0.1 ) ; ( 1,0 ) circle ( 0.1 ) ; ( 2,0 ) circle ( 0.1 ) ; ( 3,0 ) circle ( 0.1 ) ; ( 4,0 ) circle ( 0.1 ) ; ( 5,0 ) circle ( 0.1 ) ; ( 6,0 ) circle ( 0.1 ) ; ( 0,1 ) circle ( 0.1 ) ; ( 7,0 ) circle ( 0.1 ) ; ( 7.5,0.866 ) circle ( 0.1 ) ; ( 7.5,-0.866 ) circle ( 0.1 ) ; ( 8.5,0.866 ) circle ( 0.1 ) ; ( 8.5,-0.866 ) circle ( 0.1 ) ; ( 9.5,0.866 ) circle ( 0.1 ) ; ( 9.5,-0.866 ) circle ( 0.1 ) ; ( 10.5,0.866 ) circle ( 0.1 ) ; ( 10.5,-0.866 ) circle ( 0.1 ) ; ( -2,0 ) circle ( 0.1 ) ; ( -1,0 ) circle ( 0.1 ) ; ( 0,0 ) circle ( 0.1 ) ; ( 1,0 ) circle ( 0.1 ) ; ( 2,0 ) circle ( 0.1 ) ; ( 3,0 ) circle ( 0.1 ) ; ( 4,0 ) circle ( 0.1 ) ; ( 5,0 ) circle ( 0.1 ) ; ( 6,0 ) circle ( 0.1 ) ; ( 0,1 ) circle ( 0.1 ) ; ( 7,0 ) circle ( 0.1 ) ; ( 7.5,0.866 ) circle ( 0.1 ) ; ( 7.5,-0.866 ) circle ( 0.1 ) ; ( 8.5,0.866 ) circle ( 0.1 ) ; ( 8.5,-0.866 ) circle ( 0.1 ) ; ( 9.5,0.866 ) circle ( 0.1 ) ; ( 9.5,-0.866 ) circle ( 0.1 ) ; ( 10.5,0.866 ) circle ( 0.1 ) ; ( 10.5,-0.866 ) circle ( 0.1 ) ; the dark vertices form a sub - diagram cutting out an @xmath55 fiber .
the corresponding elliptic parameter ( unique up to fractional linear transformations ) is @xmath281 . substituting @xmath282 in to the weierstrass equation , and dividing the right hand side by @xmath283 , we obtain a quartic in @xmath184 , giving rise to a genus @xmath59 curve over @xmath284 .
in fact , the genus @xmath59 fibration has a section ( it is evident from the diagram above ) , so we may convert to the jacobian using classical formulas ( see @xcite , for instance ) . after some scaling and normalization
, we arrive at a weierstrass equation in standard form , with the following equation @xmath285 from this we may read out the igusa clebsch invariants .
@xmath286 a similar analysis as for discriminant @xmath60 computes the double cover of the humbert surface giving the hilbert modular surface . a birational model for the surface @xmath280 ( equivalently , for @xmath287 )
is given by @xmath288 it is a rational surface .
the humbert surface is birational to the @xmath213-plane . in these coordinates , the igusa
clebsch invariants of a point on the moduli space are given by the formulas in above .
the equation above describes a conic bundle over @xmath289 .
setting @xmath245 makes the right side a square , so the conic bundle has a section . to parametrize the surface , we set @xmath290 and solve for @xmath291 , obtaining @xmath292 a tautological curve over the hilbert modular surface ( which is rational in the parameters @xmath293 above ) is given by @xmath294 this presentation of the curve has the feature that the eigen - differentials for the action of real multiplication by @xmath295 ( alternatively , the pullbacks of the canonical differentials from the two elliptic curves ) are @xmath167 and @xmath296 .
we say that the weierstrass equation is normalized .- coordinate , and we make an arbitrary choice , designed to make the equation look `` nice '' . ]
the weierstrass equations given in this paper and in the auxiliary files for tautological families of genus @xmath0 curves ( for each of the discriminants ) will always be normalized . to check this property of being normalized amounts to a calculation using the `` eigenform location '' algorithm of @xcite . in this paper , we omit the details . next , we compute the @xmath57-invariants associated of the two associated elliptic curves
. to calculate the @xmath57-invariants in this and other sections , we used the following computational technique . for many specializations @xmath297 , we compute the genus @xmath0 curve from its igusa clebsch invariants .
we then decompose its jacobian and find out its simple factors , using analytic techniques ( using the function ` analyticjacobian ` in ` magma ` ; see @xcite ) .
more precisely , let @xmath6 be the analytic jacobian .
if the coordinates @xmath232 are chosen generically , its endomorphism ring will be @xmath295 , the quadratic ring of discriminant @xmath263 .
we solve the equation @xmath298 in @xmath299 .
then we can recover the period matrices of @xmath300 and @xmath301 , and from there compute the two @xmath57-invariants . in general , these are quadratic over the base field , but their sum and product are rational .
we use rational reconstruction to recover these numbers @xmath302 and @xmath303 from floating - point approximations .
finally , we apply the above procedure to enough specializations to enable us to recover @xmath164 and @xmath304 as rational functions , by interpolation / linear algebra ( using some upper bounds on the degree as in @xcite ) .
we obtain the following expressions : @xmath305 the discriminant of @xmath306 is , up to squares , the branch locus of the double cover defining the hilbert modular surface . solving in terms of @xmath307 and @xmath218 , we have ( up to interchange ) @xmath308 the above process may seem non - rigorous , but in fact can be made completely rigorous , for instance , by demonstrating the real multiplication by @xmath295 through an explicit correspondence ( as described in @xcite or in more detail in @xcite ) . for reasons of space , we do not describe this method explicitly here .
however , we will describe another method , suggested to us by noam elkies , which makes use of the fact that the differentials on the genus @xmath0 curve are normalized ( i.e. are eigenforms for real multiplication ) .
we start with the ( arbitrarily chosen ) point @xmath309 on @xmath280 . rescaling the @xmath186-coordinate
so the weierstrass equation has integer coefficients , the genus @xmath0 curve is given by @xmath310 for convenience , let us instead start from the curve @xmath2 @xmath311 ( obtained by the change of variables @xmath312 ) which has the rational point @xmath313 .
the form @xmath296 is a normalized differential on @xmath2 ( since @xmath314 under the above transformation ) .
assume that the associated elliptic curve @xmath3 is given by @xmath315 where @xmath316 , @xmath317 and @xmath318 are ( for the moment ) undetermined parameters .
also , let us assume without loss of generality that the morphism @xmath9 which we want to determine takes @xmath319 to @xmath320 .
note that the above model for @xmath3 can easily be converted to standard weierstrass form .
our normalization assumption says that @xmath321 maps the canonical differential @xmath322 on @xmath3 to @xmath323 on @xmath2 .
expanding in the formal ring about @xmath313 on these curves , we compute as follows : @xmath324 therefore @xmath325 on the other hand , we have @xmath326 so @xmath327 now , since @xmath328 , we have the equation @xmath329 , which we can write using the above expressions as @xmath330 where @xmath331
. we can invert the formal series on the right to write @xmath332 now , we know that @xmath331 must be a rational function of degree @xmath90 in @xmath186 ( for a degree @xmath4 map @xmath333 , it will have degree @xmath4 ) .
since its formal expansion starts @xmath334 , it must equal @xmath335 for some constants @xmath336 which are also undetermined .
we expand the above rational expression in a power series about @xmath337 and match with the previous formal expression for @xmath338 .
the first few coefficients give us linear equations which can be solved for @xmath339 and @xmath7 in that order , and then we can solve the next few polynomial equations for @xmath340 .
renormalizing the equations of @xmath2 and @xmath3 , we find that the elliptic curve @xmath3 is given by the weierstrass equation @xmath341 and the morphism @xmath333 by @xmath342 ( note that the @xmath57-invariant of @xmath3 is @xmath343 , which matches the expression given for @xmath344 at @xmath309 . )
the above calculation produced the morphism @xmath345 to one of the associated elliptic curves , for a single point on the moduli space .
to produce it over the whole base , one may work with the tautological genus @xmath0 curve , and with coefficients in the function field of @xmath280 .
alternatively , we may sample many such points and interpolate to find the coefficients of the rational map defining the morphism , and also the weierstrass coefficients of the elliptic curve .
once the map is produced , it is trivial to verify it ( by substituting in to the equation of the elliptic curve ) . to find the other map @xmath346
, we may use the other normalized differential on the genus @xmath0 curve and apply the above process to it .
more simply , we may rewrite the equation of @xmath46 and the map @xmath345 in terms of the coordinates @xmath232 on the humbert surface and the square root @xmath347 defining the double cover , and conjugate @xmath347 to @xmath348 everywhere .
we omit the details of these calculations for the higher discriminants treated in this paper , as the expressions become much more complicated with increasing discriminant .
we may , in fact , use the knowledge of the @xmath57-invariants obtained by the first ( sampling and interpolation ) method to reduce our work in this process , since they give us an extra equation satisfied by the weierstrass coefficients @xmath340 .
finally , the above process hinges on finding normalized weierstrass forms for the genus @xmath0 curves in the first place . for a single point @xmath349 on the moduli space
( with @xmath350 say ) , i.e. for a single genus @xmath0 curve , we do this by applying the eigenform location algorithm of @xcite , which indicates the fractional linear transformation needed to transform the curve to normalized form . at that point , the weierstrass coefficients are floating point numbers , but we may apply a version of rational reconstruction to recognize them as elements of @xmath351 . finally , with enough sampling , we may interpolate to write a normalized tautological curve over the moduli space . comparing our absolute invariants @xmath352 to those of shaska , we obtain the relation between the coordinates here and those in @xcite .
it is given by @xmath353 with inverse @xmath354 using this change of coordinates , we see that the double cover is the same as the @xmath355 double cover in @xcite and @xmath356 should be in the denominator rather than the numerator . ] .
let us check the compatibility of the @xmath57-invariants we found above with the values in @xcite . in the setup of that paper ,
the genus @xmath0 curve is given by @xmath357 where @xmath358 are related to the parameters @xmath233 on the hilbert modular surface by @xmath359 note that the parameters @xmath360 on the humbert surface are @xmath361 define should be the quantity @xmath142 as we have defined it , rather than its square . ] @xmath362 , @xmath363 , and consider the rational functions @xmath364 and @xmath365 .
then @xmath2 maps to the genus @xmath59 curve @xmath366 computing the @xmath57-invariant of this elliptic curve , and using the transformation formulas @xmath367 we see that it agrees with @xmath368 .
we now list several curves of interest on the hilbert modular surface . 1 . the rational curves @xmath369 correspond to @xmath370 .
the curve @xmath371 on the humbert surface lifts to a rational curve , whose points correspond to jacobians with endomorphisms by a split quaternion algebra .
the associated elliptic curves are related by a @xmath372-isogeny .
the curves @xmath373 are non - modular , but correspond to the `` degenerate '' case considered by shaska in ( * ? ? ?
* equation ( 12 ) ) .
4 . the modular genus @xmath58 curve @xmath374 is part of the branch locus .
it is parametrized by @xmath375 and corresponds to the two elliptic curves being isomorphic .
the curves @xmath376 , @xmath377 lift to non - modular rational curves on @xmath287 .
we start with a family of elliptic k3 surfaces with bad fibers of types @xmath55 and @xmath379 , and a section of height @xmath380 .
a weierstrass equation for such a k3 surface has the form @xmath381 with @xmath358 linear in @xmath184 and @xmath382 constant . since @xmath383 to have a fiber of type no worse than @xmath379 at @xmath58 , we may write @xmath384 we have a degree of freedom ( scaling @xmath184 ) , so we scale to have @xmath385 . to have the leaves of the @xmath379 fiber defined over the ground field , we set @xmath386 and @xmath387 .
next , we impose a section of height @xmath0 : its @xmath186-coordinate must have the form @xmath388 for some @xmath389 . substituting in to the weierstrass equation and completing the square
, we get a system of equations , which can be solved for @xmath390 and @xmath7 .
we finally use the remaining degree of freedom ( weierstrass scaling of @xmath186 and @xmath199 ) to set @xmath391 , since the variable @xmath392 has weight @xmath59 .
thus , we get a rational moduli space in the parameters @xmath272 and @xmath393 .
the weierstrass equation suggests the convenient change of variables @xmath394 .
the final weierstrass equation is @xmath395 note that the section has @xmath186-coordinate @xmath396 .
we go to an @xmath398 fibration by a @xmath0-neighbor step , as illustrated below .
( -2,0)(12,0 ) ; ( 1,0)(1,1 ) ; ( 7,0)(7,1 ) ; ( 11,0)(11,1 ) ; ( 5,0)(11,0)(11,1 ) ; ( 7,0)(7,1 ) ; ( 6,2 ) to ( 4,0 ) ; ( 6,2 ) to ( 11,1 ) ; ( 5,0 ) circle ( 0.2 ) ; ( -2,0 ) circle ( 0.1 ) ; ( -1,0 ) circle ( 0.1 ) ; ( 0,0 ) circle ( 0.1 ) ; ( 1,0 ) circle ( 0.1 ) ; ( 2,0 ) circle ( 0.1 ) ; ( 3,0 ) circle ( 0.1 ) ; ( 4,0 ) circle ( 0.1 ) ; ( 5,0 ) circle ( 0.1 ) ; ( 6,0 ) circle ( 0.1 ) ; ( 1,1 ) circle ( 0.1 ) ; ( 7,0 ) circle ( 0.1 ) ; ( 8,0 ) circle ( 0.1 ) ; ( 9,0 ) circle ( 0.1 ) ; ( 10,0 ) circle ( 0.1 ) ; ( 11,0 ) circle ( 0.1 ) ; ( 12,0 ) circle ( 0.1 ) ; ( 7,1 ) circle ( 0.1 ) ; ( 11,1 ) circle ( 0.1 ) ; ( 6,2 ) circle ( 0.1 ) ; ( -2,0 ) circle ( 0.1 ) ; ( -1,0 ) circle ( 0.1 ) ; ( 0,0 ) circle ( 0.1 ) ; ( 1,0 ) circle ( 0.1 ) ; ( 2,0 ) circle ( 0.1 ) ; ( 3,0 ) circle ( 0.1 ) ; ( 4,0 ) circle ( 0.1 ) ; ( 5,0 ) circle ( 0.1 ) ; ( 6,0 ) circle ( 0.1 ) ; ( 1,1 ) circle ( 0.1 ) ; ( 7,0 ) circle ( 0.1 ) ; ( 8,0 ) circle ( 0.1 ) ; ( 9,0 ) circle ( 0.1 ) ; ( 10,0 ) circle ( 0.1 ) ; ( 11,0 ) circle ( 0.1 ) ; ( 12,0 ) circle ( 0.1 ) ; ( 7,1 ) circle ( 0.1 ) ; ( 11,1 ) circle ( 0.1 ) ; ( 6,2 ) circle ( 0.1 ) ; an elliptic parameter is @xmath399 .
as in the previous section , we transform to the new elliptic fibration , which has the desired @xmath54 and @xmath55 fibers , and proceed to read out the igusa clebsch invariants .
they are given as follows : @xmath400 we then compute the double cover defining the hilbert modular surface . a birational model for the surface @xmath397 ( equivalently , for @xmath401 )
is given by @xmath402 it is a rational surface .
the humbert surface is birational to the @xmath213-plane . in these coordinates , the igusa
clebsch invariants of a point on the moduli space are given by the formulas in above .
the right hand side is quartic in @xmath291 , and we easily calculate that the surface is a rational elliptic surface with section . to find a rational parametrization , we make some elementary transformations to reduce the degree of the equation in @xmath291 and @xmath218 .
the net effect of these is to set @xmath403 the double cover is transformed to @xmath404 which is a conic bundle over the @xmath405-line with an obvious section @xmath406 .
parametrizing with a new parameter @xmath407 , we obtain finally @xmath408 where we have omitted @xmath347 for simplicity .
once again , it is possible to write down the tautological family of genus @xmath0 curves over the hilbert modular surface .
however , the expression is quite long , and so we will omit it in the main body of the paper
. however , it can be obtained from the auxiliary computer files .
similarly , for the higher discriminants treated in this paper , we will omit the expression for the tautological family of genus @xmath0 curves ( these are also in the computer files ) .
as before , it is possible to compute the equations of the two elliptic curves attached to the @xmath409-decomposable jacobian , and the morphisms from the tautological genus @xmath0 curve to the elliptic curves .
we omit this calculation , and simply display the elementary symmetric functions in the @xmath57-invariants of the two elliptic curves .
@xmath410 over the hilbert modular surface , which is rational in parameters @xmath411 , we can solve the quadratic equation to get the two @xmath57-invariants .
@xmath412 we compare our formulas with those of bolza @xcite , since those of bruin and doerksen are equivalent ( * ? ? ?
* appendix a ) .
bolza gives a family of genus @xmath0 curves with @xmath409-reducible jacobians over a weighted projective space with @xmath413 of weights @xmath414 respectively . by comparing the igusa
clebsch invariants , we obtain the following relation between those coordinates and ours : @xmath415 with inverse @xmath416 1 .
the curves @xmath417 and @xmath418 belong to the product locus .
the latter is also part of the branch locus for the double cover @xmath419 .
the curve @xmath420 on @xmath421 has genus @xmath58 , and is parametrized by @xmath422 .
its lift is also a rational curve , parametrized by letting @xmath423 .
it is a modular curve , corresponding to the elliptic curves being @xmath424-isogenous .
3 . the curve @xmath215 is part of the branch locus .
4 . the curve @xmath425 on the humbert surface has genus @xmath58 , and is parametrized by @xmath426 .
its lift is also rational , parametrized by letting @xmath427 .
this modular curve corresponds to the elliptic curves being @xmath90-isogenous .
5 . the curve @xmath428 is part of the branch locus .
it has genus @xmath58 , and is parametrized by @xmath429 6 .
there are also several simple non - modular curves : for instance , the specializations @xmath430 and @xmath431 gives genus @xmath58 curves with rational points .
next , we go on to discriminant @xmath432 , which was studied in @xcite .
we start with a family of elliptic k3 surfaces with bad fibers of type @xmath433 and @xmath434 , and a section of height @xmath435 .
a weierstrass equation for such a k3 surface has the form @xmath436 this has a @xmath433 fiber at @xmath181 and an @xmath180 fiber at @xmath437 .
we have shifted @xmath186 so that the generator of the mordell weil group has @xmath186-coordinate @xmath58 .
we need five more orders of vanishing for the discriminant at @xmath188 .
since the components of the @xmath434 fiber are defined over the ground field , @xmath272 is a square ; we may normalize it to be @xmath59 .
this takes care of the scaling in @xmath186 .
we then proceed to successively set the coefficients of @xmath193 through @xmath438 in the expression for the discriminant to be zero , which gives us the conditions @xmath439 to simplify the succeeding expressions , we define a new variable @xmath7 by @xmath440 .
the condition for the final order of vanishing gives us a quadratic equation in @xmath441 , whose discriminant is a square times @xmath442 .
therefore , setting @xmath443 makes the expression a square , and we solve for @xmath441 , obtaining @xmath444 we also scale @xmath445 to quotient by the scaling in @xmath184 , and let @xmath446 to simplify the expressions ; this change of variable is suggested by the final formulas below .
we obtain the final weierstrass equation @xmath447 as noted above , the section of height @xmath448 has @xmath186-coordinate @xmath58 .
we then proceed to an elliptic fibration with @xmath54 and @xmath450 fibers , via a @xmath0-neighbor step .
the elliptic parameter is simply @xmath186 .
( -2,0)(7,0 ) ; ( -1,0)(-1,1 ) ; ( 4,0)(4,1 ) ; ( 7,0)(7.5,0.866)(9.5,0.866)(9.5,-0.866)(7.5,-0.866)(7,0 ) ; ( 3,2 ) to [ bend right ] ( -1,1 ) ; ( 3,2 ) to [ bend left ] ( 7.5,0.866 ) ; ( -1,0)(6,0 ) ; ( 4,0)(4,1 ) ; ( 6,0 ) circle ( 0.2 ) ; ( -2,0 ) circle ( 0.1 ) ; ( -1,0 ) circle ( 0.1 ) ; ( 0,0 ) circle ( 0.1 ) ; ( 1,0 ) circle ( 0.1 ) ; ( 2,0 ) circle ( 0.1 ) ; ( 3,0 ) circle ( 0.1 ) ; ( 4,0 ) circle ( 0.1 ) ; ( 5,0 ) circle ( 0.1 ) ; ( 6,0 ) circle ( 0.1 ) ; ( 7,0 ) circle ( 0.1 ) ; ( 7.5,0.866 ) circle ( 0.1 ) ; ( 8.5,0.866 ) circle ( 0.1 ) ; ( 9.5,0.866 ) circle ( 0.1 ) ; ( 7.5,-0.866 ) circle ( 0.1 ) ; ( 8.5,-0.866 ) circle ( 0.1 ) ; ( 9.5,-0.866 ) circle ( 0.1 ) ; ( -1,1 ) circle ( 0.1 ) ; ( 4,1 ) circle ( 0.1 ) ; ( 3,2 ) circle ( 0.1 ) ; ( -2,0 ) circle ( 0.1 ) ; ( -1,0 ) circle ( 0.1 ) ; ( 0,0 ) circle ( 0.1 ) ; ( 1,0 ) circle ( 0.1 ) ; ( 2,0 ) circle ( 0.1 ) ; ( 3,0 ) circle ( 0.1 ) ; ( 4,0 ) circle ( 0.1 ) ; ( 5,0 ) circle ( 0.1 ) ; ( 6,0 ) circle ( 0.1 ) ; ( 7,0 ) circle ( 0.1 ) ; ( 7.5,0.866 ) circle ( 0.1 ) ; ( 8.5,0.866 ) circle ( 0.1 ) ; ( 9.5,0.866 ) circle ( 0.1 ) ; ( 7.5,-0.866 ) circle ( 0.1 ) ; ( 8.5,-0.866 ) circle ( 0.1 ) ; ( 9.5,-0.866 ) circle ( 0.1 ) ; ( -1,1 ) circle ( 0.1 ) ; ( 4,1 ) circle ( 0.1 ) ; ( 3,2 ) circle ( 0.1 ) ; after some easy weierstrass transformations and a fractional linear transformation of the elliptic parameter , the weierstrass equation for the new fibration is @xmath451 there is a section @xmath452 of height @xmath453 .
finally , we take a @xmath0-neighbor step to an elliptic fibration with @xmath54 and @xmath55 fibers . ( -2,0)(7,0)(7.5,0.866)(7.5,0.866)(9.5,0.866)(10,0)(9.5,-0.866)(7.5,-0.866)(7,0 ) ; ( 0,0)(0,1 ) ; ( 9.5 , 0.866)(7.5,0.866)(7,0)(7.5,-0.866)(9.5,-0.866 ) ; ( 6,0)(7,0 ) ; ( 6,0 ) circle ( 0.2 ) ; ( 6,1.5 ) to [ bend right ] ( 5,0 ) ; ( 6,1.5 ) to [ bend left ] ( 7.5,0.866 ) ; ( -2,0 ) circle ( 0.1 ) ; ( -1,0 ) circle ( 0.1 ) ; ( 0,0 ) circle ( 0.1 ) ; ( 1,0 ) circle ( 0.1 ) ; ( 2,0 ) circle ( 0.1 ) ; ( 3,0 ) circle ( 0.1 ) ; ( 4,0 ) circle ( 0.1 ) ; ( 5,0 ) circle ( 0.1 ) ; ( 6,0 ) circle ( 0.1 ) ; ( 0,1 ) circle ( 0.1 ) ; ( 7,0 ) circle ( 0.1 ) ; ( 7.5,0.866 ) circle ( 0.1 ) ; ( 7.5,-0.866 ) circle ( 0.1 ) ; ( 8.5,0.866 ) circle ( 0.1 ) ; ( 8.5,-0.866 ) circle ( 0.1 ) ; ( 9.5,0.866 ) circle ( 0.1 ) ; ( 9.5,-0.866 ) circle ( 0.1 ) ; ( 10,0 ) circle ( 0.1 ) ; ( 6,1.5 ) circle ( 0.1 ) ; ( -2,0 ) circle ( 0.1 ) ; ( -1,0 ) circle ( 0.1 ) ; ( 0,0 ) circle ( 0.1 ) ; ( 1,0 ) circle ( 0.1 ) ; ( 2,0 ) circle ( 0.1 ) ; ( 3,0 ) circle ( 0.1 ) ; ( 4,0 ) circle ( 0.1 ) ; ( 5,0 ) circle ( 0.1 ) ; ( 6,0 ) circle ( 0.1 ) ; ( 0,1 ) circle ( 0.1 ) ; ( 7,0 ) circle ( 0.1 ) ; ( 7.5,0.866 ) circle ( 0.1 ) ; ( 7.5,-0.866 ) circle ( 0.1 ) ; ( 8.5,0.866 ) circle ( 0.1 ) ; ( 8.5,-0.866 ) circle ( 0.1 ) ; ( 9.5,0.866 ) circle ( 0.1 ) ; ( 9.5,-0.866 ) circle ( 0.1 ) ; ( 10,0 ) circle ( 0.1 ) ; ( 6,1.5 ) circle ( 0.1 ) ; the elliptic parameter is @xmath454 . if @xmath316 is the class of the @xmath55 fiber , then @xmath455 , while @xmath456 , where @xmath2 is the remaining component of the @xmath450 fiber not included in @xmath316 .
since these numbers are coprime , the genus @xmath59 fibration defined by @xmath316 has a section .
we therefore convert to the jacobian , and read out the igusa
clebsch invariants , which are as follows .
@xmath457 computing the double cover , we obtain the following description of the moduli space . a birational model for the surface @xmath449 ( equivalently , for @xmath458 ) is given by @xmath459 it is a rational surface .
the humbert surface is birational to the @xmath213-plane . in these coordinates , the igusa
clebsch invariants of a point on the moduli space are given by the formulas in above .
once again , this is a conic bundle with a section . setting @xmath460 and solving for @xmath291
, we obtain @xmath461 let us compare our results with those of @xcite . there it is asserted that the moduli space @xmath462 of genus @xmath0 curves with a degree @xmath372 elliptic subfield ( which is birational to the humbert surface @xmath463 ) is birational to the hypersurface cut out by the following equation in @xmath464 : @xmath465 in fact , this is a rational surface : the discriminant with respect to @xmath135 is @xmath466 and the invertible change of variables @xmath467 converts it to @xmath468 , up to squares . setting this quantity equal to @xmath469 and solving for @xmath470 , we obtain the full parametrization @xmath471 unfortunately , there is a mistake in equation ( 10 ) should be @xmath472 rather than @xmath473 . ] in the paper @xcite : the moduli space described there , claimed to be birational to @xmath462 in theorem @xmath90 of their paper , is actually birational to the hilbert modular surface ( i.e. to @xmath449 ) , and the humbert surface is a quotient of it by an involution which we give explicitly in the auxiliary files ( we check directly that the absolute igusa clebsch invariants are not changed by applying this involution ) . by comparing the igusa
clebsch invariants , we obtain the relation between these coordinates and ours : @xmath474 with inverse @xmath475 the @xmath57-invariants of the associated elliptic curves are given by @xmath476 let us check that the @xmath57-invariant agrees with that implicit in @xcite .
it follows from the description there that @xmath2 maps to the elliptic curve @xmath477 where @xmath478 where @xmath147 and @xmath479 are the polynomials @xmath480 @xmath481 is a root of @xmath482 and @xmath233 are related to the parameters @xmath358 by @xmath483 after some computation , the value of the @xmath57-invariant is found to agree with @xmath344 .
we again list several interesting curves on the surface . 1
. the curves @xmath484 , @xmath485 are rational curves corresponding to @xmath370 , as is the lift of @xmath486 .
the curve @xmath487 on the humbert surface , parametrized by @xmath488 , lifts to an elliptic curve on @xmath458 .
it parametrizes the locus for which the two associated elliptic curves are related by a @xmath489-isogeny .
abstractly , this elliptic curve is isomorphic to @xmath490 of conductor @xmath489 , which has rank @xmath58 .
the curve @xmath491 lifts to an elliptic curve on @xmath458 , isomorphic to the elliptic curve @xmath492 of conductor @xmath162 and rank @xmath58 .
it parametrizes the locus of reducible jacobians for which the two associated elliptic curves have an @xmath162-isogeny .
4 . the curve @xmath493 of genus @xmath58 is part of the branch locus , and is parametrized by @xmath494 it corresponds to the two elliptic curves being isomorphic .
the curve parametrized by @xmath495 corresponds to the degenerate case i of @xcite .
the curve parametrized by @xmath496 corresponds to the degenerate case ii of @xcite .
the curve parametrized by @xmath497 corresponds to the degenerate case iii of @xcite .
next , we describe the moduli space for degree @xmath499 covers . the corresponding humbert surface is birational to the moduli space of elliptic k3 surfaces with @xmath54 , @xmath270 and @xmath500 fibers , with a section of height @xmath501 .
we briefly describe how to parametrize this moduli space and the family of k3 surfaces over it .
start with an elliptic surface @xmath502 where @xmath267 are polynomials of degrees @xmath503 respectively .
this surface already has an @xmath54 fiber at @xmath265 , an @xmath504 fiber at @xmath58 and an @xmath180 fiber at @xmath192 .
since the components of the eventual @xmath270 fiber at @xmath188 are rational , the constant term of @xmath187 is a square , and we may scale @xmath186 and @xmath199 so that @xmath505 .
similarly , @xmath506 is rational .
so we may write @xmath507 imposing two more orders of vanishing at @xmath188 , we may solve linear equations for @xmath508 and @xmath509 .
we need one more order of vanishing of the discriminant at @xmath192 .
this gives a quadratic equation in @xmath390 whose discriminant is @xmath510 times a square .
we set @xmath511 and solve for @xmath390 . finally , we need a section with the requisite height .
it must have @xmath186-coordinate @xmath512 times a linear factor , which we may normalize as @xmath513 . substituting in to the weierstrass cubic polynomial and completing the square
, we can compute the @xmath199-coordinate .
this leads to three equations involving the remaining variables .
we eliminate @xmath514 and @xmath515 to get a single equation connecting the remaining variables @xmath201 .
this equation is cubic in @xmath516 , but subtituting @xmath517 and @xmath518 makes it quadratic . setting the discriminant equal to a square , after some easy algebra we end up with a parametrization by two variables @xmath291 and @xmath218 .
the universal weierstrass equation is @xmath519 here we have introduced appropriate scalings of @xmath520 to eliminate denominators .
the @xmath186-coordinate of the section of height @xmath0 is given by @xmath521 the @xmath90-neighbor step corresponding to the figure below takes us to an @xmath398 fibration .
( 0,0)(9,0)(9.5,0.866)(10.5,0.866)(11,0)(10.5,-0.866)(9.5,-0.866)(9,0 ) ; ( 8,0)(8,1)(8.5,1.866)(7.5,1.866)(8,1 ) ; ( 2,0)(2,1 ) ; ( 7,3)(7.5,1.866 ) ; ( 7,3 ) to [ bend right ] ( 7,0 ) ; ( 7,3 ) to [ bend left ] ( 10.5,0.866 ) ; ( 7.5,1.866)(8,1)(8,0)(9,0)(9.5,-0.866)(10.5,-0.866)(11,0 ) ; ( 9,0)(9.5,0.866 ) ; ( 0,0 ) circle ( 0.1 ) ; ( 1,0 ) circle ( 0.1 ) ; ( 2,0 ) circle ( 0.1 ) ; ( 2,1 ) circle ( 0.1 ) ; ( 3,0 ) circle ( 0.1 ) ; ( 4,0 ) circle ( 0.1 ) ; ( 5,0 ) circle ( 0.1 ) ; ( 6,0 ) circle ( 0.1 ) ; ( 7,0 ) circle ( 0.1 ) ; ( 8,0 ) circle ( 0.1 ) ; ( 8,1 ) circle ( 0.1 ) ; ( 8.5,1.866 ) circle ( 0.1 ) ; ( 7.5,1.866 ) circle ( 0.1 ) ; ( 9,0 ) circle ( 0.1 ) ; ( 9.5,0.866 ) circle ( 0.1 ) ; ( 9.5,-0.866 ) circle ( 0.1 ) ; ( 10.5,0.866 ) circle ( 0.1 ) ; ( 10.5,-0.866 ) circle ( 0.1 ) ; ( 11,0 ) circle ( 0.1 ) ; ( 7,3 ) circle ( 0.1 ) ; ( 0,0 ) circle ( 0.1 ) ; ( 1,0 ) circle ( 0.1 ) ; ( 2,0 ) circle ( 0.1 ) ; ( 2,1 ) circle ( 0.1 ) ; ( 3,0 ) circle ( 0.1 ) ; ( 4,0 ) circle ( 0.1 ) ; ( 5,0 ) circle ( 0.1 ) ; ( 6,0 ) circle ( 0.1 ) ; ( 7,0 ) circle ( 0.1 ) ; ( 8,0 ) circle ( 0.1 ) ; ( 8,1 ) circle ( 0.1 ) ; ( 8.5,1.866 ) circle ( 0.1 ) ; ( 7.5,1.866 ) circle ( 0.1 ) ; ( 9,0 ) circle ( 0.1 ) ; ( 9.5,0.866 ) circle ( 0.1 ) ; ( 9.5,-0.866 ) circle ( 0.1 ) ; ( 10.5,0.866 ) circle ( 0.1 ) ; ( 10.5,-0.866 ) circle ( 0.1 ) ; ( 11,0 ) circle ( 0.1 ) ; ( 7,3 ) circle ( 0.1 ) ; we may now read out the igusa clebsch invariants .
they are fairly lengthy to write down , so we will not display them here .
instead they may be accessed in the auxiliary computer files . a birational model for the surface @xmath49 ( equivalently , for @xmath522 )
is given by @xmath523 it is a singular elliptic k3 surface ( i.e. of maximal picard number @xmath524 ) . the humbert surface is birational to the @xmath213-plane .
the equation above shows that @xmath49 has a genus @xmath59 fibration over @xmath214 , with an obvious section .
therefore it is an elliptic surface .
taking the corresponding jacobian fibration and after some simple transformations ( including setting @xmath525 ) , we obtain the following weierstrass equation , showing that it is a k3 surface . @xmath526
it has reducible fibers of type @xmath527 at @xmath188 and at the roots of @xmath528 , @xmath529 at @xmath530 and at @xmath181 , and @xmath531 at the roots of @xmath532 .
the trivial lattice therefore has rank @xmath533 , leaving room for mordell weil rank up to @xmath0 .
we find the sections @xmath534 the intersection matrix of these sections is @xmath535 it has determinant @xmath536 .
hence the span of these two sections and the trivial lattice has rank @xmath524 and discriminant @xmath537 .
we check that it is @xmath0- and @xmath90- saturated in the nron severi lattice , and therefore it must be all of @xmath538 .
we compute the symmetric polynomials in the two @xmath57-invariants , in terms of the parameters on the humbert surface , and check that the discriminant of the polynomial @xmath539 is , up to squares , the same as the branch locus for the hilbert modular surface .
we obtain the following expressions .
we next present some curves of low genus on the modular surface . 1 . the curves @xmath245 and @xmath541 lift to pairs of rational curves on @xmath522 .
they correspond to @xmath248 .
the curve @xmath542 lies in the branch locus .
it is also a component of the product locus ( i.e. the jacobian degenerates to a product of elliptic curves ) .
the curve @xmath543 also belongs to the product locus .
it lifts to a genus @xmath58 curve without any real points .
the curve @xmath215 lifts to a genus @xmath58 curve whose points have @xmath544 .
it too does not have any real points .
the curve @xmath376 lifts to an elliptic curve isomorphic to @xmath545 , of conductor @xmath162 and rank @xmath58 .
its points correspond to reducible jacobians such that the elliptic curves are related by an @xmath162-isogeny
the curve @xmath546 lifts to an elliptic curve isomorphic to @xmath547 , which has conductor @xmath524 and rank @xmath58 .
along it , the elliptic curves are related by a @xmath372-isogeny .
the curve @xmath548 lifts to an elliptic curve isomorphic to @xmath549 , which has conductor @xmath550 and rank @xmath58 .
it corresponds to elliptic curves linked by a @xmath550-isogeny . 8 .
the curve @xmath551 is part of the branch locus .
it has genus @xmath58 and can be parametrized as @xmath552 it corresponds to an isogeny of degree @xmath59 , i.e. @xmath544 .
the curve @xmath553 on the humbert surface has genus @xmath58 , and can be parametrized by @xmath554 it lifts to an elliptic curve isomorphic to @xmath555 , which has conductor @xmath556 and rank @xmath58 .
the points on this curve correspond to the two elliptic curves being related by a @xmath90-isogeny . 10 .
the curve @xmath557 on the humbert surface is non - modular .
it lifts to an elliptic curve isomorphic to @xmath558 , which has conductor @xmath559 and rank @xmath59 . 11 .
we can also produce non - modular curves of genus @xmath58 through the sections @xmath452 and @xmath560 and their linear combinations .
for instance , the section @xmath452 gives the curve @xmath561 while @xmath562 gives the curve @xmath563 the specializations of the tautological family along these genus @xmath58 curves are one - parameter families of genus @xmath0 curves with real multiplication by @xmath564 .
we start with the family of elliptic k3 surfaces having @xmath54 , @xmath566 , @xmath500 and @xmath180 fibers at @xmath567 respectively , and a section of height @xmath568 .
we start with a weierstrass equation @xmath569 with @xmath267 of degrees @xmath503 respectively .
we have shifted @xmath186 so that the section has vanishing @xmath186-coordinate at @xmath570 .
this fibration already has an @xmath54 fiber at @xmath265 and @xmath180 fibers at @xmath571 .
let us normalize @xmath572 where we have used that @xmath573 and @xmath506 must be squares in order for the components to be defined over the base field , and also used the weierstrass scaling to fix @xmath505 .
imposing two more orders of vanishing at @xmath188 gives linear equations for @xmath508 and @xmath509 , which are easily solved .
a third order of vanishing gives a quadratic equation for @xmath390 , whose discriminant equals @xmath574 , up to squares .
we set it equal to @xmath575 and solve for @xmath441 and then @xmath390 .
next , we impose an extra order of vanishing at @xmath192 , which gives a quadratic equation for @xmath405 , with discriminant up to squares equal to @xmath576 .
we therefore set @xmath577 and solve for @xmath405 .
next , scale @xmath578 and @xmath579 .
the section of height @xmath580 must have @xmath186-coordinate @xmath581 . substituting this in to the weierstrass equation , we obtain a square times an expression quartic in @xmath184 . completing the square
, we obtain two equations . solving these leads to a rational moduli space with parameters @xmath291 and @xmath218 .
the universal weierstrass equation is as follows , where we have scaled @xmath582 to avoid denominators : @xmath583 the @xmath500 fiber is now at @xmath584 .
the @xmath186-coordinate of a section @xmath585 of height @xmath580 is given by @xmath586 for brevity we omit the @xmath199-coordinate , though it may be found in the auxiliary files .
we now need to perform `` elliptic hopping '' to an @xmath398 fibration .
we start with a @xmath0-neighbor step to a fibration with @xmath54 and @xmath450 fibers .
( -2,0)(7,0)(7.69,0.95)(8.81,0.59)(8.81,-0.59)(7.69,-0.95)(7,0 ) ; ( 6,1)(6,-1)(6.5,-1.866)(5.5,-1.866)(6,-1 ) ; ( 0,0)(0,1 ) ; ( 6,0.97)(7,0.97 ) ; ( 6,1.03)(7,1.03 ) ; ( 6,0)(7,0)(7.69,0.95)(8.81,0.59)(8.81,-0.59 ) ; ( 6.5,-1.866)(6,-1)(6,0 ) ; ( 6,0)(7,0 ) ; ( 6,0 ) circle ( 0.2 ) ; ( 10.5,0 ) to [ bend right=45 ] ( 7,1 ) ; ( 10.5,0 ) to [ bend right=100 ] ( 5,0 ) ; ( 10.5,0 ) to [ bend left=45 ] ( 8.81,-0.59 ) ; ( 10.5,0 ) to [ bend left=60 ] ( 6.5,-1.866 ) ; ( 7 + 0.69,0.95 ) circle ( 0.1 ) ; ( 7 + 0.69,-0.95 ) circle ( 0.1 ) ; ( 7 + 1.81,0.59 ) circle ( 0.1 ) ; ( 7 + 1.81,-0.59 ) circle ( 0.1 ) ; ( 6,1 ) circle ( 0.1 ) ; ( 7,1 ) circle ( 0.1 ) ; ( 6,-1 ) circle ( 0.1 ) ; ( 5.5,-1.866 ) circle ( 0.1 ) ; ( 6.5,-1.866 ) circle ( 0.1 ) ; ( -2,0 ) circle ( 0.1 ) ; ( -1,0 ) circle ( 0.1 ) ; ( 0,0 ) circle ( 0.1 ) ; ( 1,0 ) circle ( 0.1 ) ; ( 2,0 ) circle ( 0.1 ) ; ( 3,0 ) circle ( 0.1 ) ; ( 4,0 ) circle ( 0.1 ) ; ( 5,0 ) circle ( 0.1 ) ; ( 6,0 ) circle ( 0.1 ) ; ( 0,1 ) circle ( 0.1 ) ; ( 7,0 ) circle ( 0.1 ) ; ( 10.5,0 ) circle ( 0.1 ) ; ( -2,0 ) circle ( 0.1 ) ; ( -1,0 ) circle ( 0.1 ) ; ( 0,0 ) circle ( 0.1 ) ; ( 1,0 ) circle ( 0.1 ) ; ( 2,0 ) circle ( 0.1 ) ; ( 3,0 ) circle ( 0.1 ) ; ( 4,0 ) circle ( 0.1 ) ; ( 5,0 ) circle ( 0.1 ) ; ( 6,0 ) circle ( 0.1 ) ; ( 0,1 ) circle ( 0.1 ) ; ( 7,0 ) circle ( 0.1 ) ; ( 7 + 0.69,0.95 ) circle ( 0.1 ) ; ( 7 + 0.69,-0.95 ) circle ( 0.1 ) ; ( 7 + 1.81,0.59 ) circle ( 0.1 ) ; ( 7 + 1.81,-0.59 ) circle ( 0.1 ) ; ( 6,1 ) circle ( 0.1 ) ; ( 7,1 ) circle ( 0.1 ) ; ( 6,-1 ) circle ( 0.1 ) ; ( 5.5,-1.866 ) circle ( 0.1 ) ; ( 6.5,-1.866 ) circle ( 0.1 ) ; ( 10.5,0 ) circle ( 0.1 ) ; the elliptic parameter is given by @xmath587 where @xmath588 converting to the jacobian , it is then an easy matter to take a second @xmath0-neighbor step to an @xmath398 fibration , as in the second step for discriminant @xmath432 .
we omit the details , and also the formulas for the igusa
clebsch invariants , which may be found in the auxiliary files .
we then compute the double cover defining the hilbert modular surface . a birational model for the surface @xmath50 ( equivalently , for @xmath589 )
is given by @xmath590 it is a singular elliptic k3 surface .
the humbert surface is birational to the @xmath213-plane .
this surface has a genus @xmath59 fibration over @xmath289 , with a section , for instance @xmath591 . converting to the jacobian fibration
, we get the weierstrass equation @xmath592 of an elliptic k3 surface .
it has an @xmath593 fiber at @xmath215 , an @xmath594 fiber at @xmath595 , a @xmath596 fiber at @xmath597 , an @xmath527 fiber at @xmath598 , and @xmath599 fibers at @xmath600 .
so the contribution to the picard number from the trivial lattice is @xmath96 .
in addition we find the sections @xmath601 which are linearly independent .
therefore , the picard number is @xmath524 , i.e. the hilbert modular surface is a singular k3 surface .
the intersection matrix of these sections is @xmath602 it has discriminant @xmath603 .
hence the span of these two sections and the trivial lattice has rank @xmath524 and discriminant @xmath604 .
we check that it is @xmath0-saturated in the nron severi lattice , and therefore it must be all of @xmath605 . once again , we check that the symmetric polynomials of the two @xmath57-invariants are rational functions of @xmath291 and @xmath218 ( not displayed here for brevity ) , and that the difference @xmath166 generates the function field of the hilbert modular surface over that of the humbert surface .
the equation of a tautological family of genus @xmath0 curves may also be found in the computer files . 1 . the rational curves @xmath606 correspond to @xmath370 .
2 . the curve @xmath607 lifts to a rational curve on @xmath589 .
it is part of the product locus .
the curve @xmath608 also lifts to a rational curve .
it corresponds to the elliptic curves being related by a @xmath609-isogeny .
the curve @xmath610 lifts to an elliptic curve on the hilbert modular surface , isomorphic to @xmath611 , of conductor @xmath550 and rank @xmath58 .
it corresponds to the elliptic curves being related by a @xmath550-isogeny .
5 . the curve @xmath612 lifts to an elliptic curve on @xmath613 , isomorphic to @xmath614 of conductor @xmath122 and rank @xmath58 .
it corresponds to the elliptic curves being related by a @xmath122-isogeny .
the curve @xmath615 lifts to an elliptic curve isomorphic to @xmath616 , of conductor @xmath617 and rank @xmath58 .
it corresponds to the elliptic curves being @xmath617-isogenous .
the rational curve @xmath618 on the humbert surface can be parametrized by @xmath619 .
it lifts to an elliptic curve on @xmath589 , isomorphic to @xmath555 , which has conductor @xmath556 and rank @xmath58 .
it corresponds to the elliptic curves being @xmath556-isogenous .
the rational curve @xmath620 lifts to an elliptic curve , isomorphic to @xmath621 of conductor @xmath524 and rank @xmath58 .
it corresponds to the elliptic curves being @xmath524-isogenous .
the curve @xmath622 has genus @xmath58 and is part of the branch locus .
it can be parametrized as @xmath623 and corresponds to the two elliptic curves being isomorphic .
again , we can obtain several genus @xmath58 non - modular curves from the sections .
for instance , the section @xmath452 gives rise to @xmath624 and the section @xmath560 to @xmath625 the tautological curve along this latter curve is short enough to note here ; it is given by @xmath626
we start with a family of elliptic k3 surfaces with reducible fibers of types @xmath55 , @xmath504 and @xmath628 , and a section of height @xmath629 . the weierstrass equation for this family
is given by @xmath630 the @xmath186-coordinate of a section of height @xmath0 is given by @xmath631 for brevity , we omit the description of the process of parametrization , which is similar to that for the smaller discriminants described in this paper .
a brief outline is given in the auxiliary files , for the interested reader .
to go to an elliptic fibration with @xmath54 and @xmath55 fibers , we proceed as follows .
first , we take a @xmath90-neighbor step with fiber class as shown below , to go to an elliptic fibration with @xmath379 and @xmath55 fibers .
( 0,0)(11,0 ) ; ( 3,0)(3,1 ) ; ( 7,0)(7,1)(7.707,1.707)(7,2.414)(6.293,1.707)(7,1 ) ; ( 9,0)(9,1 ) ; ( 10,0)(10,1 ) ; ( 6,3)(6.293,1.707 ) ; ( 6,3 ) to ( 6,0 ) ; ( 6,3 ) to ( 10,1 ) ; ( 7,1)(7,0)(10,0 ) ; ( 9,1)(9,0 ) ; ( 7,1)(6.293,1.707)(7,2.414 ) ; ( 6,3)(6.293,1.707 ) ; ( 0,0 ) circle ( 0.1 ) ; ( 1,0 ) circle ( 0.1 ) ; ( 2,0 ) circle ( 0.1 ) ; ( 3,0 ) circle ( 0.1 ) ; ( 3,1 ) circle ( 0.1 ) ; ( 4,0 ) circle ( 0.1 ) ; ( 5,0 ) circle ( 0.1 ) ; ( 6,0 ) circle ( 0.1 ) ; ( 6,3 ) circle ( 0.1 ) ; ( 7,0 ) circle ( 0.1 ) ; ( 7,1 ) circle ( 0.1 ) ; ( 7.707,1.707 ) circle ( 0.1 ) ; ( 6.293,1.707 ) circle ( 0.1 ) ; ( 7,2.414 ) circle ( 0.1 ) ; ( 8,0 ) circle ( 0.1 ) ; ( 9,0 ) circle ( 0.1 ) ; ( 10,0 ) circle ( 0.1 ) ; ( 9,1 ) circle ( 0.1 ) ; ( 10,1 ) circle ( 0.1 ) ; ( 11,0 ) circle ( 0.1 ) ; ( 0,0 ) circle ( 0.1 ) ; ( 1,0 ) circle ( 0.1 ) ; ( 2,0 ) circle ( 0.1 ) ; ( 3,0 ) circle ( 0.1 ) ; ( 3,1 ) circle ( 0.1 ) ; ( 4,0 ) circle ( 0.1 ) ; ( 5,0 ) circle ( 0.1 ) ; ( 6,0 ) circle ( 0.1 ) ; ( 6,3 ) circle ( 0.1 ) ; ( 7,0 ) circle ( 0.1 ) ; ( 7,1 ) circle ( 0.1 ) ; ( 7.707,1.707 ) circle ( 0.1 ) ; ( 6.293,1.707 ) circle ( 0.1 ) ; ( 7,2.414 ) circle ( 0.1 ) ; ( 8,0 ) circle ( 0.1 ) ; ( 9,0 ) circle ( 0.1 ) ; ( 10,0 ) circle ( 0.1 ) ; ( 9,1 ) circle ( 0.1 ) ; ( 10,1 ) circle ( 0.1 ) ; ( 11,0 ) circle ( 0.1 ) ; we then take a @xmath0-neighbor step to an @xmath398 fibration , as in the calculation for discriminant @xmath378 .
we omit the details . the equation for the double cover defining the hilbert modular surface can now be computed . a birational model for the surface @xmath632 ( equivalently , for @xmath633 )
is given by @xmath634 its minimal model is an honestly elliptic surface @xmath149 with @xmath635 and picard number @xmath636 .
the humbert surface is birational to the @xmath213-plane .
the equation above yields an elliptic surface : making the change of variables @xmath637 and absorbing @xmath638 in @xmath639 makes the right side quartic in @xmath218 , giving an elliptic fibration over @xmath640 , with an obvious section . converting to the jacobian form , after some simple transformations , including the substitution @xmath641
, we obtain the weierstrass equation @xmath642 which displays an obvious extra symmetry @xmath643 .
this is an honestly elliptic surface with @xmath644 .
it has reducible fibers of type @xmath645 at @xmath265 and @xmath58 , @xmath531 at @xmath59 , @xmath529 at @xmath205 , @xmath531 at the roots of @xmath646 and also at the roots of @xmath647 , and @xmath527 at the roots of @xmath648 .
the trivial sublattice has rank @xmath649 , leaving room for mordell weil rank up to @xmath0 .
in addition to the @xmath0-torsion section @xmath650 , we find the following sections : @xmath651 these sections are orthogonal of heights @xmath652 and @xmath653 respectively .
therefore , this elliptic surface has maximal picard number @xmath636 .
the discriminant of the lattice generated by @xmath654 and the trivial lattice is @xmath655 .
we checked that it is @xmath0- and @xmath90-saturated . therefore
, it must be the full picard group , and the mordell weil lattice is generated by these sections .
the igusa
clebsch invariants , the equation of a tautological family of genus @xmath0 curves , and the elementary symmetric polynomials in the @xmath57-invariants of the associated elliptic curves can be found in the auxiliary files . 1 .
the curve @xmath245 is part of the product locus .
it lifts to a genus @xmath58 curve on the hilbert modular surface , with no real points . 2 .
the curve @xmath656 has genus @xmath58 and is part of the locus @xmath370 , as is the curve @xmath657 , which lifts to a rational curve on @xmath632 .
the locus @xmath370 also contains the curves @xmath658 ( which lifts to an elliptic curve ) , @xmath659 ( lifts to a union of two rational curves ) , and @xmath660 , which has genus @xmath58 and is part of the branch locus .
the last curve can be parametrized as @xmath661 .
3 . the curve @xmath662 has genus @xmath58 and can be parametrized as @xmath663 .
it lifts to an elliptic curve of conductor @xmath489 , and is modular ( corresponding to the two elliptic curves being @xmath424-isogenous ) .
4 . the rational curve @xmath664 is part of the branch locus , as is the curve @xmath665 the latter can be parametrized as @xmath666 5 .
there are several non - modular specializations : for instance , @xmath667 or @xmath668 all yield elliptic curves with rational points . of course
, we may use the elliptic fibration to produce many elliptic curves , simply by specializing the parameter on the base .
the sections of the elliptic fibrations gives us several ( usually ) non - modular rational curves .
for instance , @xmath452 gives the curve parametrized by @xmath669 while @xmath560 gives the curve parametrized by @xmath670
we start with a family of elliptic k3 surfaces with reducible fibers of types @xmath672 , @xmath500 and @xmath450 , and a section of height @xmath673 .
a weierstrass equation is given by @xmath674 and the @xmath186-coordinate of the extra section is @xmath675 we first go to an elliptic fibration with @xmath54 and @xmath450 fibers , via the @xmath90-neighbor step corresponding to the figure below .
( 0,0)(6,0)(6.5,0.866)(8.5,0.866)(9,0)(8.5,-0.866)(6.5,-0.866)(6,0 ) ; ( 5,0)(5,1)(5.5,1.866)(4.5,1.866)(5,1 ) ; ( 2,0)(2,2 ) ; ( 5.5,3)(5.5,1.866 ) ; ( 5.5,3 ) to ( 2,2 ) ; ( 5.5,3 ) to ( 6.5,0.866 ) ; ( 2,0)(2,1 ) ; ( 0,0)(5,0)(5,1)(5.5,1.866 ) ; ( 0,0 ) circle ( 0.1 ) ; ( 1,0 ) circle ( 0.1 ) ; ( 2,0 ) circle ( 0.1 ) ; ( 2,1 ) circle ( 0.1 ) ; ( 2,2 ) circle ( 0.1 ) ; ( 3,0 ) circle ( 0.1 ) ; ( 4,0 ) circle ( 0.1 ) ; ( 5,0 ) circle ( 0.1 ) ; ( 5,1 ) circle ( 0.1 ) ; ( 4.5,1.866 ) circle ( 0.1 ) ; ( 5.5,1.866 ) circle ( 0.1 ) ; ( 5.5,3.0 ) circle ( 0.1 ) ; ( 6,0 ) circle ( 0.1 ) ; ( 6.5,0.866 ) circle ( 0.1 ) ; ( 6.5,-0.866 ) circle ( 0.1 ) ; ( 7.5,0.866 ) circle ( 0.1 ) ; ( 7.5,-0.866 ) circle ( 0.1 ) ; ( 8.5,0.866 ) circle ( 0.1 ) ; ( 8.5,-0.866 ) circle ( 0.1 ) ; ( 9,0 ) circle ( 0.1 ) ; ( 0,0 ) circle ( 0.1 ) ; ( 1,0 ) circle ( 0.1 ) ; ( 2,0 ) circle ( 0.1 ) ; ( 2,1 ) circle ( 0.1 ) ; ( 2,2 ) circle ( 0.1 ) ; ( 3,0 ) circle ( 0.1 ) ; ( 4,0 ) circle ( 0.1 ) ; ( 5,0 ) circle ( 0.1 ) ; ( 5,1 ) circle ( 0.1 ) ; ( 4.5,1.866 ) circle ( 0.1 ) ; ( 5.5,1.866 ) circle ( 0.1 ) ; ( 5.5,3.0 ) circle ( 0.1 ) ; ( 6,0 ) circle ( 0.1 ) ; ( 6.5,0.866 ) circle ( 0.1 ) ; ( 6.5,-0.866 ) circle ( 0.1 ) ; ( 7.5,0.866 ) circle ( 0.1 ) ; ( 7.5,-0.866 ) circle ( 0.1 ) ; ( 8.5,0.866 ) circle ( 0.1 ) ; ( 8.5,-0.866 ) circle ( 0.1 ) ; ( 9,0 ) circle ( 0.1 ) ; finally , a @xmath0-neighbor step ( as in the second step for discriminant @xmath432 ) takes us to an @xmath398 fiber .
we can then read out the igusa
clebsch invariants , which are in the auxiliary files , along with a tautological family of genus @xmath0 curves and the sum and product of the @xmath57-invariants of the elliptic curves .
we then compute the double cover defining the hilbert modular surface . a birational model for the surface @xmath676 ( equivalently , for @xmath677 )
is given by @xmath678 its minimal model is an honestly elliptic surface @xmath149 with @xmath635 and picard number @xmath679 .
the humbert surface is birational to the @xmath213-plane .
the equation above evidently defines an elliptic surface over @xmath289 with section . converting to the jacobian form , we obtain ( after weierstrass transformations and the change of variable @xmath680 ) @xmath681 it is honestly elliptic with @xmath644 , and reducible fibers of type @xmath645 at @xmath58 , @xmath682 at @xmath265 , @xmath527 at the roots of @xmath683 , @xmath531 at the roots of @xmath684 and @xmath527 at the roots of @xmath685 .
these give rise to a trivial lattice of rank @xmath556 , leaving room for mordell weil rank up to @xmath90 .
we find the sections @xmath686 with height matrix @xmath687 therefore the picard number @xmath688 is at least @xmath679 . on the other hand , by counting points modulo @xmath424 and @xmath609 and using the method of van luijk @xcite ( comparing the square classes of the discriminants of the picard groups of these reductions ) we obtain that the picard number can not be @xmath636 .
therefore @xmath689 exactly .
we checked that the lattice generated by @xmath452 and @xmath560 , which has discriminant @xmath690 , is @xmath0- and @xmath90-saturated .
therefore it is the full mordell weil lattice . 1 . the curves @xmath245 , @xmath691 and @xmath692 of genus @xmath58 on the humbert surface are all part of the product locus .
the first lifts to a rational curve on @xmath677 and the other two lift to elliptic curves .
the curves @xmath597 , @xmath693 and @xmath694 are part of the locus @xmath370 . the first two lift to unions of two rational curves , while the last clearly has no real points . 3 .
the curve @xmath695 lifts to an elliptic curve isomorphic to @xmath696 .
it corresponds to the two elliptic curves being @xmath162-isogenous .
4 . the branch locus has genus @xmath59 , and corresponds to @xmath544 .
the section @xmath452 gives rise to the following non - modular curve of genus @xmath58 : @xmath697 the section @xmath560 gives the non - modular curve @xmath698 the tautological curve over it is given by @xmath699
we start with a family of elliptic k3 surfaces with @xmath701 , @xmath434 and @xmath504 fibers , and a section of height @xmath702 .
a weierstrass equation is given by @xmath703 and the extra section having @xmath186-coordinate @xmath704 we first go to an elliptic fibration with @xmath55 , @xmath450 and @xmath180 fibers by a @xmath0-neighbor step .
( 0,0)(6,0)(6.5,0.866)(8.5,0.866)(8.5,-0.866)(6.5,-0.866)(6,0 ) ; ( 1,0)(1,1 ) ; ( 3,0)(3,1 ) ; ( 5,0)(5,1)(5.707,1.707)(5,2.414)(4.293,1.707)(5,1 ) ; ( 6,3)(5.707,1.707 ) ; ( 6,3 ) to ( 1,1 ) ; ( 6,3 ) to ( 6.5,0.866 ) ; ( 0,0)(5,0)(5,1 ) ; ( 3,0)(3,1 ) ; ( 0,0 ) circle ( 0.1 ) ; ( 1,0 ) circle ( 0.1 ) ; ( 1,1 ) circle ( 0.1 ) ; ( 2,0 ) circle ( 0.1 ) ; ( 3,0 ) circle ( 0.1 ) ; ( 3,1 ) circle ( 0.1 ) ; ( 4,0 ) circle ( 0.1 ) ; ( 5,0 ) circle ( 0.1 ) ; ( 5,1 ) circle ( 0.1 ) ; ( 5.707,1.707 ) circle ( 0.1 ) ; ( 4.293,1.707 ) circle ( 0.1 ) ; ( 5,2.414 ) circle ( 0.1 ) ; ( 6,0 ) circle ( 0.1 ) ; ( 6,3 ) circle ( 0.1 ) ; ( 6.5,0.866 ) circle ( 0.1 ) ; ( 6.5,-0.866 ) circle ( 0.1 ) ; ( 7.5,0.866 ) circle ( 0.1 ) ; ( 7.5,-0.866 ) circle ( 0.1 ) ; ( 8.5,0.866 ) circle ( 0.1 ) ; ( 8.5,-0.866 ) circle ( 0.1 ) ; ( 0,0 ) circle ( 0.1 ) ; ( 1,0 ) circle ( 0.1 ) ; ( 1,1 ) circle ( 0.1 ) ; ( 2,0 ) circle ( 0.1 ) ; ( 3,0 ) circle ( 0.1 ) ; ( 3,1 ) circle ( 0.1 ) ; ( 4,0 ) circle ( 0.1 ) ; ( 5,0 ) circle ( 0.1 ) ; ( 5,1 ) circle ( 0.1 ) ; ( 5.707,1.707 ) circle ( 0.1 ) ; ( 4.293,1.707 ) circle ( 0.1 ) ; ( 5,2.414 ) circle ( 0.1 ) ; ( 6,0 ) circle ( 0.1 ) ; ( 6,3 ) circle ( 0.1 ) ; ( 6.5,0.866 ) circle ( 0.1 ) ; ( 6.5,-0.866 ) circle ( 0.1 ) ; ( 7.5,0.866 ) circle ( 0.1 ) ; ( 7.5,-0.866 ) circle ( 0.1 ) ; ( 8.5,0.866 ) circle ( 0.1 ) ; ( 8.5,-0.866 ) circle ( 0.1 ) ; next , we go to a @xmath705 fibration , via the fiber class shown below .
( 0,0)(8,0)(8.5,0.866)(10.5,0.866)(11,0)(10.5,-0.866)(8.5,-0.866)(8,0 ) ; ( 3,0)(3,1 ) ; ( 7,0)(7,1 ) ; ( 6.95,1)(6.95,2 ) ; ( 7.05,1)(7.05,2 ) ; ( 7.5,3 ) to ( 7,1 ) ; ( 7.5,3 ) to ( 8.5,0.866 ) ; ( 7.5,3 ) to ( 6,0 ) ; ( 6,0)(8,0 ) ; ( 7,0)(7,1 ) ; ( 8.5,0.866)(8,0)(8.5,-0.866 ) ; ( 0,0 ) circle ( 0.1 ) ; ( 1,0 ) circle ( 0.1 ) ; ( 2,0 ) circle ( 0.1 ) ; ( 3,0 ) circle ( 0.1 ) ; ( 3,1 ) circle ( 0.1 ) ; ( 4,0 ) circle ( 0.1 ) ; ( 5,0 ) circle ( 0.1 ) ; ( 6,0 ) circle ( 0.1 ) ; ( 7,0 ) circle ( 0.1 ) ; ( 7,1 ) circle ( 0.1 ) ; ( 7,2 ) circle ( 0.1 ) ; ( 8,0 ) circle ( 0.1 ) ; ( 7.5,3 ) circle ( 0.1 ) ; ( 8.5,0.866 ) circle ( 0.1 ) ; ( 8.5,-0.866 ) circle ( 0.1 ) ; ( 9.5,0.866 ) circle ( 0.1 ) ; ( 9.5,-0.866 ) circle ( 0.1 ) ; ( 10.5,0.866 ) circle ( 0.1 ) ; ( 10.5,-0.866 ) circle ( 0.1 ) ; ( 11,0 ) circle ( 0.1 ) ; ( 0,0 ) circle ( 0.1 ) ; ( 1,0 ) circle ( 0.1 ) ; ( 2,0 ) circle ( 0.1 ) ; ( 3,0 ) circle ( 0.1 ) ; ( 3,1 ) circle ( 0.1 ) ; ( 4,0 ) circle ( 0.1 ) ; ( 5,0 ) circle ( 0.1 ) ; ( 6,0 ) circle ( 0.1 ) ; ( 7,0 ) circle ( 0.1 ) ; ( 7,1 ) circle ( 0.1 ) ; ( 7,2 ) circle ( 0.1 ) ; ( 8,0 ) circle ( 0.1 ) ; ( 7.5,3 ) circle ( 0.1 ) ; ( 8.5,0.866 ) circle ( 0.1 ) ; ( 8.5,-0.866 ) circle ( 0.1 ) ; ( 9.5,0.866 ) circle ( 0.1 ) ; ( 9.5,-0.866 ) circle ( 0.1 ) ; ( 10.5,0.866 ) circle ( 0.1 ) ; ( 10.5,-0.866 ) circle ( 0.1 ) ; ( 11,0 ) circle ( 0.1 ) ; the new elliptic fibration has mordell weil group of rank @xmath0 , generated by two sections @xmath452 and @xmath560 with height matrix @xmath706 with the entries arising as @xmath707 a @xmath0-neighbor step takes us to a fibration with @xmath708 fibers , a @xmath0-torsion section @xmath709 ( and a section of height @xmath710 ) .
the section @xmath560 intersects the fiber class in @xmath90 , whereas the remaining component of the @xmath628 fiber intersects it in @xmath0 .
since these are coprime , the new fibration has a section .
note also that the section @xmath711 of height @xmath712 does nt intersect the fiber class , giving rise to an extra @xmath180 fiber .
( 0,0)(9,0 ) ; ( 1,0)(1,1 ) ; ( 3,0)(3,1 ) ; ( 5,0)(5,1)(5.707,1.707)(5,2.414)(4.293,1.707)(5,1 ) ; ( 7,0)(7,1 ) ; ( 8,0)(8,1 ) ; ( 5,3.5)(4.293,1.707 ) ; ( 5,3.5 ) to ( 6,0 ) ; ( 5,3.5 ) to ( 4,0 ) ; ( 5,1)(5,0)(9,0 ) ; ( 8,0)(8,1 ) ; ( 5.707,1.707)(5,1)(4.293,1.707 ) ; ( 0,0 ) circle ( 0.1 ) ; ( 1,0 ) circle ( 0.1 ) ; ( 1,1 ) circle ( 0.1 ) ; ( 2,0 ) circle ( 0.1 ) ; ( 3,0 ) circle ( 0.1 ) ; ( 3,1 ) circle ( 0.1 ) ; ( 4,0 ) circle ( 0.1 ) ; ( 5,3.5 ) circle ( 0.1 ) ; ( 5,0 ) circle ( 0.1 ) ; ( 5,1 ) circle ( 0.1 ) ; ( 5.707,1.707 ) circle ( 0.1 ) ; ( 4.293,1.707 ) circle ( 0.1 ) ; ( 5,2.414 ) circle ( 0.1 ) ; ( 6,0 ) circle ( 0.1 ) ; ( 7,0 ) circle ( 0.1 ) ; ( 8,0 ) circle ( 0.1 ) ; ( 7,1 ) circle ( 0.1 ) ; ( 8,1 ) circle ( 0.1 ) ; ( 9,0 ) circle ( 0.1 ) ; ( 0,0 ) circle ( 0.1 ) ; ( 1,0 ) circle ( 0.1 ) ; ( 1,1 ) circle ( 0.1 ) ; ( 2,0 ) circle ( 0.1 ) ; ( 3,0 ) circle ( 0.1 ) ; ( 3,1 ) circle ( 0.1 ) ; ( 4,0 ) circle ( 0.1 ) ; ( 5,3.5 ) circle ( 0.1 ) ; ( 5,0 ) circle ( 0.1 ) ; ( 5,1 ) circle ( 0.1 ) ; ( 5.707,1.707 ) circle ( 0.1 ) ; ( 4.293,1.707 ) circle ( 0.1 ) ; ( 5,2.414 ) circle ( 0.1 ) ; ( 6,0 ) circle ( 0.1 ) ; ( 7,0 ) circle ( 0.1 ) ; ( 8,0 ) circle ( 0.1 ) ; ( 7,1 ) circle ( 0.1 ) ; ( 8,1 ) circle ( 0.1 ) ; ( 9,0 ) circle ( 0.1 ) ; ( 5,3.5 ) node@xmath560 ; finally , a @xmath0-neighbor step leads us to the desired @xmath398 fibration .
the fiber @xmath316 with an extended @xmath54 dynkin diagram is shown below , and the @xmath0-torsion section @xmath709 combines with some of the @xmath701 components and the non - identity component of the @xmath180 fiber to give an @xmath55 diagram disjoint from @xmath316 . also , the section @xmath452 of height @xmath710 , or its negative , has odd intersection number with @xmath316 , while the remaining component of the @xmath379 fiber has intersection @xmath0 with @xmath316 .
so the new fibration has a section .
we can read out the igusa
clebsch invariants , which are in the auxiliary files .
( 0,0)(12,0 ) ; ( 1,0)(1,1 ) ; ( 5,0)(5,1 ) ; ( 7,0)(7,1 ) ; ( 9,0)(9,1 ) ; ( 11,0)(11,1 ) ; ( 7,2)(6,2 ) ; ( 6.95,1)(6.95,2 ) ; ( 7.05,1)(7.05,2 ) ; ( 6,2 ) to ( 1,1 ) ; ( 6,2 ) to ( 11,1 ) ; ( 0,0)(7,0 ) ; ( 5,0)(5,1 ) ; ( 0,0 ) circle ( 0.1 ) ; ( 1,0 ) circle ( 0.1 ) ; ( 1,1 ) circle ( 0.1 ) ; ( 2,0 ) circle ( 0.1 ) ; ( 3,0 ) circle ( 0.1 ) ; ( 4,0 ) circle ( 0.1 ) ; ( 5,0 ) circle ( 0.1 ) ; ( 5,1 ) circle ( 0.1 ) ; ( 6,0 ) circle ( 0.1 ) ; ( 7,0 ) circle ( 0.1 ) ; ( 7,1 ) circle ( 0.1 ) ; ( 7,2 ) circle ( 0.1 ) ; ( 6,2 ) circle ( 0.1 ) ; ( 8,0 ) circle ( 0.1 ) ; ( 9,0 ) circle ( 0.1 ) ; ( 9,1 ) circle ( 0.1 ) ; ( 10,0 ) circle ( 0.1 ) ; ( 11,0 ) circle ( 0.1 ) ; ( 11,1 ) circle ( 0.1 ) ; ( 12,0 ) circle ( 0.1 ) ; ( 0,0 ) circle ( 0.1 ) ; ( 1,0 ) circle ( 0.1 ) ; ( 1,1 ) circle ( 0.1 ) ; ( 2,0 ) circle ( 0.1 ) ; ( 3,0 ) circle ( 0.1 ) ; ( 4,0 ) circle ( 0.1 ) ; ( 5,0 ) circle ( 0.1 ) ; ( 5,1 ) circle ( 0.1 ) ; ( 6,0 ) circle ( 0.1 ) ; ( 7,0 ) circle ( 0.1 ) ; ( 7,1 ) circle ( 0.1 ) ; ( 7,2 ) circle ( 0.1 ) ; ( 6,2 ) circle ( 0.1 ) ; ( 8,0 ) circle ( 0.1 ) ; ( 9,0 ) circle ( 0.1 ) ; ( 9,1 ) circle ( 0.1 ) ; ( 10,0 ) circle ( 0.1 ) ; ( 11,0 ) circle ( 0.1 ) ; ( 11,1 ) circle ( 0.1 ) ; ( 12,0 ) circle ( 0.1 ) ; ( 6,2 ) node@xmath709 ; we compute the double cover defining the hilbert modular surface . a birational model for the surface @xmath52 ( equivalently , for @xmath713 )
is given by @xmath714 its minimal model is an honestly elliptic surface @xmath149 with @xmath635 and picard number @xmath649 or @xmath679 .
the humbert surface is birational to the @xmath213-plane .
making the substitution @xmath637 and absorbing @xmath638 in to @xmath639 makes the right side quartic in @xmath218 , showing that @xmath715 is an elliptic surface with section over @xmath640 . converting to the jacobian form , after some weierstrass transformations and substitution @xmath716
, we obtain @xmath717 it is an honestly elliptic surface with @xmath644 , and reducible fibers of type @xmath718 at @xmath265 , @xmath719 at @xmath58 , @xmath596 at @xmath720 , @xmath531 at the roots of @xmath721 and also at the roots of @xmath722 , and @xmath527 at the roots of @xmath723 .
the trivial lattice has rank @xmath556 , leaving room for mordell weil rank up to @xmath90 .
we find the section @xmath724 of height @xmath725 .
therefore , the picard number @xmath688 is at least @xmath649 . on the other hand ,
counting points modulo @xmath424 and @xmath550 shows that @xmath726 .
therefore the picard number is either @xmath649 or @xmath679 ; we have not been able to determine it exactly .
the equation of a tautological family of genus @xmath0 curves and the elementary symmetric polynomials in the @xmath57-invariants of the two elliptic curves are omitted here , but can be found in the computer files . 1 . the rational curves @xmath727 ( parametrization @xmath728 , @xmath729 ) and @xmath730 ( parametrization @xmath731 , @xmath732 ) are part of the product locus , and lift to elliptic curves on @xmath52 . the rational curve @xmath733 ( parametrized by @xmath734 , @xmath735 )
is also part of the product locus , as well as the branch locus .
2 . the curves @xmath736 , @xmath737 , @xmath215 , @xmath738 and @xmath739 are part of the locus @xmath370 . the first two lift to rational curves , while the others lift to unions of two rational curves .
3 . the curve @xmath245 lifts to an elliptic curve isomorphic to @xmath696 .
it corresponds to the two elliptic curves being @xmath162-isogenous .
4 . the curve @xmath740 lifts to an elliptic curve isomorphic to @xmath741 .
it corresponds to the elliptic curves being @xmath122-isogenous .
the curve @xmath597 also lifts to an elliptic curve ( with conductor 36 and cm ) , and is modular : it corresponds to the elliptic curves being @xmath263-isogenous .
the curve @xmath742 lifts to a rational curve ; the two elliptic curves are again @xmath263-isogenous along that locus .
the branch locus is an elliptic curve , and corresponds to @xmath252 .
the section above gives rise to the non - modular rational curve parametrized by @xmath743
finally , we describe moduli space of elliptic subfields of degree @xmath162 via the hilbert modular surface of discriminant @xmath744 . to compute the surface , we work with the moduli space of elliptic k3 surfaces with @xmath434 , @xmath270 and @xmath504 fibers , and two sections with intersection matrix @xmath745 with the entries arising as @xmath746 we will be very brief in our description of all the steps , since the formulas are all fairly complicated and explained in greater length in the auxiliary files .
the moduli space of these k3 surfaces is again rational , and with chosen parameters @xmath291 and @xmath218 , the universal weierstrass equation is as follows .
@xmath747 where @xmath748 @xmath749 the @xmath186-coordinates of the two sections are given by @xmath750 first , we take a @xmath0-neighbor step to an elliptic fibration with @xmath672 , @xmath566 , @xmath504 and @xmath180 fibers , and two sections with intersection matrix @xmath751 with the entries being realized as @xmath752 in the figure below we only display the section of height @xmath753 , in order to not clutter the picture .
( 0,0)(1,0 ) ; ( 1,0)(6,0 ) ; ( 2,0)(2,2 ) ; ( 6,0)(5,1)(5 + 0.707,1.707)(5,2.414)(5 - 0.707,1.707)(5,1 ) ; ( 6,0)(7,1 ) ; ( 6,0)(8,0)(8 + 0.69,0.95)(8 + 1.81,0.59)(8 + 1.81,-0.59)(8 + 0.69,-0.95)(8,0 ) ; ( 7.03,1)(7.03,2 ) ; ( 6.97,1)(6.97,2 ) ; next , we go by a @xmath0-neighbor step to an @xmath754 fibration with two sections having intersection matrix @xmath755 with the entries being realized as @xmath756 we only show the section of height @xmath757 below .
( -1,0)(0,0 ) ; ( 0,0)(6,0 ) ; ( 2,0)(2,1 ) ; ( 6,0)(5,1)(5 + 0.5,1.866)(5 - 0.5,1.866)(5,1 ) ; ( 6,0)(6.5,1 ) ; ( 6,0)(7,0)(7 + 0.69,0.95)(7 + 1.81,0.59)(7 + 1.81,-0.59)(7 + 0.69,-0.95)(7,0 ) ; ( 6.53,1)(6.53,2 ) ; ( 6.47,1)(6.47,2 ) ; we next take a @xmath0-neighbor step to an @xmath758 fibration , with a section of height @xmath759 . finally , we take a @xmath0-neighbor step to go to an @xmath398 fibration , as in the second step for discriminant @xmath432 .
the formulas for the igusa
clebsch invariants can now be read off ; they may be found in the auxiliary files .
the expressions for the sum and product of the @xmath57-invariants of the two associated elliptic curves may be found in the auxiliary computer files , which also have the equation of a tautological family of genus @xmath0 curves . 1 .
the rational curve @xmath215 and the genus @xmath59 curve @xmath762 are contained in the product locus .
each lifts to a union of two curves on the hilbert modular surface ( of genera @xmath58 and @xmath59 respectively ) .
the curves @xmath763 all correspond to @xmath370 .
the first five are of genus @xmath58 and the remaining two have genus @xmath59 . the first two lift to unions of two genus @xmath58 curves with no rational points , the third to a union of two @xmath250 s , the fourth to an irreducible @xmath250 , the fifth to an elliptic curve , and the last two to curves of genus @xmath0 and @xmath90 respectively .
3 . the rational curve @xmath764 lifts to an elliptic curve isomorphic to @xmath765 , and corresponds to the elliptic curves being @xmath766-isogenous .
4 . the rational curve @xmath668 lifts to an elliptic curve isomorphic to @xmath767 , and corresponds to the elliptic curves being @xmath550-isogenous .
5 . the rational curve @xmath768 lifts to an elliptic curve isomorphic to @xmath769 , and corresponds to the the elliptic curves being @xmath770-isogenous . swww s. y. an , s. y. kim , d. c. marshall , s. h. marshall , w. g. mccallum , and a. perlis , _ jacobians of genus one curves _ , j. number theory * 90 * ( 2001 ) , no . 2 , 304315 .
n. bruin and k. doerksen , _ the arithmetic of genus two curves with @xmath409-split jacobians _ , canad . j. math . * 63 * ( 2011 ) , 9921021 .
n. bruin , e. v. flynn , d. testa , _ descent via @xmath771-isogeny on jacobians of genus @xmath0 curves _
, acta arith .
* 165 * ( 2014 ) , no . 3 , 201223 .
d. carlton , _
moduli for pairs of elliptic curves with isomorphic n - torsion _ , manuscripta math .
* 105 * ( 2001 ) , no . 2 , 201234 . j. w. s. cassels , e. v. flynn , _ prolegomena to a middlebrow arithmetic of curves of genus @xmath0 _ , london mathematical society lecture note series , 230 .
cambridge university press , cambridge , 1996 . i. dolgachev , _ endomorphisms of complex abelian varieties _ , available at http://www.math.lsa.umich.edu/~idolga/milanlect.pdf .
n. d. elkies and a. kumar , _
@xmath772 surfaces and equations for hilbert modular surfaces _ , algebra number theory , * 8 * ( 2014 ) , no . 10 , 22972411 . g. frey , _ on elliptic curves with isomorphic torsion structures and corresponding curves of genus @xmath0 _ , elliptic curves , modular forms & fermat s last theorem ( hong kong , 1993 ) , 7998 , ser .
number theory , i , int .
press , cambridge , ma , 1995 .
g. frey and e. kani , _ curves of genus @xmath0 covering elliptic curves and an arithmetical application _ , arithmetic algebraic geometry ( texel , 1989 ) , 153176 , progr .
89 , birkhuser boston , boston , ma , 1991
. b. h. gross and d. b. zagier , _ heegner points and derivatives of l - series _ , invent .
* 84 * ( 1986 ) , no . 2 , 225320 .
a. j. harder , _ moduli spaces of k3 surfaces with large picard number _ , master s thesis ( 2011 ) , queen s university .
w. hausmann , _ the fixed points of the symmetric hilbert modular group of a real quadratic field with arbitrary discriminant _ , math . ann .
* 260 * ( 1982 ) , no . 1 , 3150 . c. f. hermann , _ modulflchen quadratischer diskriminante _ , manuscripta math .
* 72 * ( 1991 ) , no . 1 , 95110 .
j. igusa , _ arithmetic variety of moduli for genus two _ , ann . of math .
( 2 ) * 72 * 1960 , 612649 .
a. krazer , _ lehrbuch der thetafunctionen _ ,
chelsea , new york , 1970 .
r. m. kuhn , _ curves of genus @xmath0 with split jacobian _ , trans .
* 307 * ( 1988 ) , no . 1 , 4149 .
a. kumar , _
@xmath772 surfaces associated with curves of genus two _ , int .
res . not .
imrn * 2008 * , no . 6 , art .
i d rnm165 , 26 pp .
a. kumar , _ elliptic fibrations on a generic jacobian kummer surface _
, j. algebraic geom .
* 23 * ( 2014 ) , 599667 .
a. kumar and r. mukamel , _ algebraic models and arithmetic geometry of teichmller curves in genus two _ , int .
res . not .
imrn ( to appear ) , .
a. kumar and r. mukamel , _ real multiplication through explicit correspondences _ , lms j. comput .
* 19 * ( special issue a ) ( 2016 ) , 2942 .
e. kani and w. schanz , _ modular diagonal quotient surfaces _ , math .
z. * 227 * ( 1998 ) , no . 2 , 337366 . b. mazur , _ modular curves and the eisenstein ideal _ , publications mathmatiques de linstitut des hautes tudes scientifiques * 47 * , no . 1 ( 1977 ) ,
k. magaard , t. shaska , and h. vlklein , _
genus @xmath0 curves that admit a degree 5 map to an elliptic curve _ , forum math .
* 21 * ( 2009 ) , no . 3 , 547566 .
c. t. mcmullen , _ foliations of hilbert modular surfaces _ , amer
. j. math .
* 129 * ( 2007 ) , 183215 .
n. murabayashi , _ the moduli space of curves of genus two covering elliptic curves _ , manuscripta math .
* 84 * ( 1994 ) , no . 2 , 125133 . t. shaska , _
genus @xmath0 fields with degree @xmath90 elliptic subfields _ , forum math . * 16 * ( 2004 ) no . 2 , 263280 .
t. shaska and h. vlklein , _ elliptic subfields and automorphisms of genus @xmath0 function fields _ , algebra , arithmetic and geometry with applications .
papers from shreeram s. abhyankar s 70th birthday conference , pg .
687707 , springer , 2004 .
t. shaska , g. s. wijesiri , s. wolf , l. woodland , _
degree @xmath60 coverings of elliptic curves by genus @xmath0 curves _ , albanian j. math .
* 2 * ( 2008 ) , no . 4 ,
. j. tate , _ algorithm for determining the type of a singular fiber in an elliptic pencil _ , modular functions of one variable , iv ( proc
. internat .
summer school , univ .
antwerp , antwerp , 1972 ) , pp .
lecture notes in math .
476 , springer , berlin , 1975 . g. van der geer , _ hilbert modular surfaces _ , ergebnisse der mathematik und ihrer grenzgebiete ( 3 ) , * 16*. springer - verlag , berlin , 1988 .
r. van luijk , _
k3 surfaces with picard number one and infinitely many rational points _ , algebra number theory 1 ( 2007 ) , no . 1 , 115 .
van wamelen , _ proving that a genus @xmath0 curve has complex multiplication _ , math . comp .
* 68 * ( 1999 ) no . 228 , 166367 .
van wamelen , _ computing with the analytic jacobian of a genus @xmath0 curve _ , in `` discovering mathematics with magma '' , pp .
117135 , springer , 2006 .
a. wiles , _ modular elliptic curves and fermat s last theorem _
, ann . of math .
( 2 ) * 141 * ( 1995 ) , no . 3 , 443551 . | we compute explicit rational models for some hilbert modular surfaces corresponding to square discriminants , by connecting them to moduli spaces of elliptic k3 surfaces .
since they parametrize decomposable principally polarized abelian surfaces , they are also moduli spaces for genus-@xmath0 curves covering elliptic curves via a map of fixed degree .
we thereby extend classical work of jacobi , hermite , bolza etc . , and more recent work of kuhn , frey , kani , shaska , vlklein , magaard and others , producing explicit families of reducible jacobians .
in particular , we produce a birational model for the moduli space of pairs @xmath1 of a genus @xmath0 curve @xmath2 and elliptic curve @xmath3 with a map of degree @xmath4 from @xmath2 to @xmath3 , as well as a tautological family over the base , for @xmath5 .
we also analyze the resulting models from the point of view of arithmetic geometry , and produce several interesting curves on them . |
in recent studies , the idea of establishing the existence of so - called critical elements " ( or the earlier minimal blow - up solutions " ) has led to significant progress in the theory of critical " dispersive and hyperbolic equations such as the energy - critical nonlinear schr " odinger equation @xcite , mass - critical nonlinear schr " odinger equation @xcite , @xmath0-critical nonlinear schr " odinger equation @xcite , energy - critical nonlinear wave equation @xcite , energy - critical and mass - critical hartree equations @xcite and energy - critical wave maps @xcite .
+ + in this paper we exhibit the generality of the method of critical elements " by applying it to a parabolic system , namely the standard . ] navier - stokes equations ( nse ) : @xmath5 typically , @xmath6 is interpreted as the velocity vector field of a fluid filling a region in space , and @xmath7 is the associated scalar pressure function . + + the critical " spaces are those which are invariant under the natural scaling of the equation . for nse , if @xmath8 is a solution , then so is @xmath9 for any @xmath10 .
the critical spaces are of the type @xmath11 , where @xmath12 . for the navier - stokes equations , one can take , for example , @xmath13 or the smaller space @xmath14 .
in fact , one has the chain of critical spaces " given by the continuous embeddings @xmath15 these are the spaces in which the initial data of solutions in the critical settings live , and we will also refer to them as critical spaces " that is , spaces of functions on @xmath16 whose norms satisfy @xmath17 . + + in the recent important paper @xcite , l. escauriaza , g. seregin and v. ver ' ak showed that any leray - hopf " weak solution which remains bounded in @xmath18 can not develop a singularity in finite time .
their proof used a blow - up procedure and reduction to a backwards uniqueness question for the heat equation , and was then completed using carleman - type inequalities and the theory of unique continuation . here ,
we approach the same problem using the method of critical elements " .
although we do also use the main tools appearing in @xcite to complete our proof , we hope that it is more intuitively clear in our exposition why those particular tools are needed .
+ + the precise statement of the main result we address in this paper is the following : [ thm : regularity ] let @xmath19 satisfy @xmath20 and let @xmath21 be the associated `` mild solution '' to nse satisfying @xmath22 .
suppose that there is some @xmath23 such that @xmath24 for all @xmath25 such that @xmath6 is defined
. then @xmath6 is defined ( and smooth ) for all positive times .
( theorem [ thm : regularity ] of course follows from the result in @xcite . )
we believe the methods given below will work as well if we replace @xmath0 by @xmath26 in theorem [ thm : regularity ] and hence can be used to give an alternative proof of the result in @xcite in the case of mild solutions . for technical reasons ( described below ) we start with the above result and plan to return to the more general case in a future publication .
+ + the mild solutions " to nse which we consider in our approach have the form @xmath27 for some divergence - free initial datum @xmath28 , with the linear solution operator @xmath29 and nonlinearity @xmath30 given by @xmath31 here @xmath32 is the convolution of @xmath28 with the heat kernel , and the equation ( [ eem ] ) , ( [ ns1 ] ) comes formally from applying the helmholtz projection operator @xmath33 to ( [ eq : nsea ] ) which fixes @xmath34 and eliminates the term @xmath35 , and then solving the resulting non - homogeneous heat equation by duhamel s formula . under sufficient regularity assumptions , mild solutions
are in fact classical solutions , and the existence of such a solution on some time interval is typically established via the contraction - mapping principle in an appropriate function space .
+ + we follow the method of c. kenig and f. merle . in a series of recent works @xcite
, they use the method of critical elements " to approach the question of global existence and scattering " ( approaching a linear solution at large times ) for nonlinear hyperbolic and dispersive equations in critical settings .
for example , for the 3d non - linear schr " odinger equation @xmath36 ( nls ) , they considered ( [ eem ] ) with @xmath37 the free schr " odinger operator , and both @xmath38 ( the focusing case ) , and @xmath39 ( the defocusing case ) , the exponents needed for the @xmath40-critical " and @xmath41-critical " settings , respectively .
note that in the case ( [ nls3 ] ) of cubic nonlinearity , the equation is invariant under the same scaling as the navier - stokes equations .
+ + in all cases , the general strategy was essentially the same , which we ll describe now in the @xmath0 setting : + + for any @xmath42 , one uses fixed - point arguments to assign a maximal time @xmath43 such that a solution @xmath6 to ( [ eem ] ) which remains in @xmath0 for positive time exists and is unique in some scaling - invariant space @xmath44 for any fixed @xmath45 , where @xmath44 denotes a space of functions defined on the space - time region @xmath46 .
define @xmath47 the type of result proved in @xcite ( variants of which were proved in @xcite and which we will prove here as well ) is that @xmath48 implies that @xmath49 and @xmath50 as @xmath51 for some @xmath52 . in other words ,
@xmath6 exists globally and scatters .
typically this is known to be true for @xmath53 for sufficiently small @xmath54 . in the case of nse
, the scattering condition is replaced by decay to zero in norm in other words , we set @xmath55 . for globally defined @xmath0-valued solutions ,
such decay was proved in @xcite ( see also @xcite ) .
+ + it is worth pointing out that the @xmath0 decay , which is proved quite easily in @xcite by decomposing the initial data into a large part in @xmath56 and a small part in @xmath0 , solving the corresponding equations and employing the standard energy arguments , is actually used heavily in our proofs , and significantly reduces the difficulty from the nls case treated in @xcite .
+ + the general method of proof , which can be referred to as `` concentration - compactness '' @xmath1 `` rigidity '' , is comprised of the following three main steps for a proof by contradiction : + 1 .
+ + assuming a finite maximal threshold @xmath57 for which @xmath58 implies global existence and scattering but such a statement fails for any @xmath59 , there exists a solution @xmath60 with @xmath61 , for which global existence or scattering fails .
+ + such a critical element @xmath60 produces a compact family " that is , up to norm - invariant rescalings and translations in space ( and possibly in time ) , the set @xmath62 is pre - compact in a critical space .
3 . + + the existence of the compact family produces a contradiction to known results .
+ steps 1 - 2 are accomplished by considering a minimizing sequence of solutions @xmath63 with initial data @xmath64 such that @xmath65 , @xmath66 for which global existence or scattering fails ( typically quantified by @xmath67 ) .
+ + then the main tool for realizing this program is a profile decomposition " associated to that sequence , which explores the lack of compactness in the embedding @xmath68 .
for example , in @xcite the following decomposition ( based on @xcite ) was used for treatment of the nls case : there exists a sequence @xmath69 , with associated linear solutions @xmath70 , and sequences of scales @xmath71 and shifts @xmath72 and @xmath73 , such that ( after a subsequence in @xmath74 ) @xmath75 where the profiles " in the sum are orthogonal in a certain sense due to the choices of @xmath76 , @xmath77 and @xmath78 , and @xmath79 is a small error for large @xmath80 and @xmath74 . ( from this , a similar decomposition is then established for the time evolution of @xmath81 . ) using this setup , it is shown that in fact the solution to nls with initial datum @xmath82 is a critical element for some @xmath83 , which establishes step 1 .
compactness ( step 2 ) is established using the same tools .
( we will discuss step 3 momentarily . )
+ + our ultimate goal here is to give such a proof in the context of the navier - stokes equations , where we would consider ( [ eem ] ) , ( [ ns1 ] ) for @xmath84 , with @xmath85;{l^3(\mathbb{r}^3 ) } ) \cap l^5({\mathbb{r}^3}\times ( 0,t))$ ] and @xmath86 defined accordingly , and @xmath87 since a profile decomposition has already been established by i. gallagher in @xcite for solutions to the navier - stokes equations evolving from a bounded set in @xmath0 ( based on the decomposition for the initial data in @xcite ) , we restrict ourselves in this paper to the case of data in @xmath0 .
we expect that the result in @xcite can be extended to the @xmath88 setting , which would then allow for an extension of our approach to that case .
we plan to return to this in a future publication profile decomposition has been established by the second author in @xcite , and the program in @xmath89 has been completed in a collaboration of the second author with i. gallagher and f. planchon , to appear soon . ] .
+ + we take @xmath90;{{\dot h^\frac{1}{2}}}({\mathbb{r}^3 } ) ) \cap l^2((0,t);{{\dot h^\frac{3}{2}}}({\mathbb{r}^3}))\ ] ] and as before let @xmath91 the profile decomposition of @xcite for solutions corresponding to a bounded sequence @xmath92 of divergence - free fields takes the form ( after a subsequence in @xmath74 ) @xmath93 where @xmath94 and @xmath95 are the solutions to the navier - stokes equations with initial data @xmath96 and @xmath97 , respectively , and @xmath79 and @xmath98 are again small errors for large @xmath80 and @xmath74 .
( we remark that the absence in the above decomposition of the time shifts @xmath99 which appeared in ( [ prof ] ) greatly simplifies matters on a technical level ; in @xcite , this was a significant consideration . )
the above program is then completed in theorems [ thm : a ] , [ thm : b ] and [ thm : c ] as follows : + + by known local - existence and small - data results , there exists a small @xmath100 such that @xmath101 implies that the solution exists for all time and tends to zero in the @xmath102 norm .
we thereby assume a finite critical value @xmath103 ( as in step 1 ) such that any solution @xmath6 with @xmath58 must exist globally and decay to zero in @xmath102 , and @xmath104 is the maximum such value .
+ + the failure of the global existence and decay property , which occurs for some solution @xmath6 with @xmath105 for any @xmath59 , is expressed by @xmath106 , where we define for @xmath107 @xmath108 (
so @xmath109 whenever @xmath110 and therefore also , by standard embeddings , @xmath111 so such solutions are smooth by the ladyzhenskaya - prodi - serrin condition " , see e.g. @xcite ) .
+ + in theorems [ thm : a ] and [ thm : b ] , we establish the existence of a solution ( critical element " ) @xmath60 with initial datum @xmath112 and @xmath113 such that @xmath114 and , further , that for any such solution , there exist @xmath115 , @xmath116 as @xmath117 and @xmath118 such that , for @xmath119 and @xmath120 , @xmath121 is pre - compact in @xmath122 .
+ + the pre - compactness in @xmath26 of such a @xmath123 is shown to be inconsistent with @xmath124 in theorem [ thm : c ] , by using the backwards uniqueness results for parabolic equations established in @xcite as well as the theory of unique continuation for parabolic equations to show that in fact @xmath125 .
( in general , step 3 requires something specific to the particular case being studied for example the morawetz - type estimate for nls used in @xcite as opposed to the methods used to establish steps 1 and 2 which are fairly general in nature . ) + + one interesting scenario in which such a @xmath123 would be compact in @xmath26 is if , in ( [ eq : kay ] ) , one could take @xmath126 for some @xmath127 ( with @xmath128 ) , @xmath129 and @xmath130 , and one imposes that @xmath131 ( i.e. , @xmath132 ) for some given non - zero @xmath133 .
this is the case of a self - similar " solution which was first ruled out in the important paper @xcite by j. neas , m. ruika and v. ver ' ak in the ( spatially global ) @xmath26 setting ( see @xcite for more general results ) .
theorem [ thm : c ] can therefore be thought of as a generalization of the result in @xcite .
( such solutions are of course ruled out as well by the more recent paper @xcite , but the proof here is much simpler for that purpose . ) + + * acknowledgements : * we would like to thank professors isabelle gallagher and vladim ' ir ver ' ak for many helpful comments and discussions . the second author would like to extend a personal and special thank you as well to vladim ' ir ver ' ak for his continuing help , guidance and support .
we ll say that @xmath6 is a mild " solution of nse on @xmath134 $ ] for some @xmath135 and @xmath107 if , for some divergence - free initial datum @xmath28 , @xmath6 solves ( in some function space ) for @xmath136 $ ] the integral equation @xmath137 we have used the following notation : for a tensor @xmath138 we define the vector @xmath139 by @xmath140 , and for vectors @xmath6 and @xmath141 , we define their tensor product @xmath142 by @xmath143 .
+ + we ll consider solutions in spatial dimension three ( @xmath144 ) , so @xmath145 , @xmath146 , @xmath147 . in that case , the projection operator @xmath33 onto divergence - free fields is defined on a vector field @xmath30 by @xmath148 , @xmath149 , and the riesz transform @xmath150 is defined on a scalar @xmath151 via fourier transforms by @xmath152 .
( one can also formally write this as @xmath153 . )
the heat kernel @xmath154 is defined by @xmath155^\vee ( x)= ( ( 4\pi t)^{-3/2 } \exp\{-|\cdot|^2/{4 t } \ } * g)(x)$ ] , and extended to act component - wise on vector fields .
+ + formally , ( [ eq : mild ] ) comes from applying @xmath156 to the classical navier - stokes equations which one can write as @xmath157 ( since @xmath158 due to the condition @xmath159 ) and solving the resulting heat equation ( since @xmath160 ) by duhamel s formula .
+ + in what follows , we ll set @xmath161 and @xmath162 for any @xmath163 $ ] , @xmath164 for any @xmath165 where @xmath166 denotes the space of tempered distributions , and for @xmath167 and any banach space @xmath168 , @xmath169 .
for any collection of banach spaces @xmath170 and @xmath171 , we ll always set @xmath172 .
similarly , for vector - valued @xmath173 , we define @xmath174 . for easy reference ,
we collect and state here the definitions of the main spaces to which we refer throughout the paper : @xmath175 @xmath176
in this section , we consolidate various known local existence results for the navier - stokes equations . we also unify the various theories through known persistency " results .
although the results presented in this section are well - known to experts , it seems to us that simple , self - contained proofs are often difficult to locate , so we present them here for the convenience of the reader .
our main results are given in section 3 , and the expert reader may prefer to skip directly to that section now .
+ + the goal in what follows is to establish the existence of local " solutions to ( [ eq : mild ] ) in some ( space - time ) banach space @xmath177 of functions defined on @xmath178 for some possibly small @xmath107 , with divergence - free initial datum @xmath28 in a banach space @xmath168 . in what follows
, we will let @xmath168 equal @xmath88 or @xmath0 .
( see , e.g. , @xcite and @xcite respectively for local well - posedness in @xmath179 and @xmath180 , and the recent ill - posedness result @xcite for @xmath181 . )
we ll re - write ( [ eq : mild ] ) as @xmath182 and , under the assumption that @xmath183 , try to solve the equation for some @xmath184 ( where @xmath11 will be chosen so that @xmath185 implies @xmath186 ) .
this will be accomplished by the following abstract lemma , using the contraction mapping principle : [ lemma : abs ] let @xmath11 be a banach space with norm @xmath187 , and let @xmath188 be a continuous bilinear form such that there exists @xmath189 so that @xmath190 for all @xmath191 and @xmath192 in @xmath11 . then for any fixed @xmath183 such that @xmath193 , equation ( [ eq : abs ] ) has a unique , see e.g. @xcite , formula ( 122 ) .
] solution @xmath194 satisfying @xmath195 , with @xmath196 * proof : * let @xmath197 using ( [ eq : f ] ) and the triangle inequality , one can verify directly that @xmath198 maps @xmath199 into itself .
moreover , @xmath198 is a contraction on @xmath200 as follows : suppose @xmath201 .
then @xmath202 @xmath203 and clearly @xmath204 by ( [ eq : g ] ) .
hence the contraction mapping principle guarantees the existence of a unique fixed - point @xmath205 of the mapping @xmath198 satisfying @xmath206 , which proves the lemma . @xmath207 can be found in @xcite .
for other versions , see for example the classical paper @xcite and a more modern exposition in @xcite.]suppose @xmath19 , and let @xmath208 , with norm @xmath209 note that @xmath210 , since h " older s inequality gives @xmath211 by the well - known ( see , e.g. @xcite , p. 120 ) inequality @xmath212 , we have @xmath213 and hence @xmath214 denoting @xmath215 we can write @xmath216 where @xmath217 . now by the maximal regularity theorem for @xmath218 ( see , e.g. , @xcite , theorem 7.3 ) , we have @xmath219 and so since @xmath220 , ( [ eq : i ] ) gives @xmath221 let s recall the following lemma ( see @xcite , lemma 14.1 ) : [ lemma : lr ] let @xmath222 $ ] and @xmath149 .
if @xmath223 , then @xmath224 .
( @xmath225 indicates bounded continuous functions . ) for @xmath226 , ( [ eq : i ] ) shows that @xmath227 , so lemma [ lemma : lr ] gives @xmath228 and hence @xmath229 .
moreover , the proof of lemma [ lemma : lr ] also gives the estimate @xmath230 which , together with ( [ eq : i ] ) gives @xmath231 using ( [ eq : h ] ) , ( [ eq : j ] ) and ( [ eq : k ] ) , we now conclude that @xmath232 for some @xmath233 .
( we remark that @xmath234 is independent of @xmath235 . ) by the standard @xmath56 energy estimates for the heat equation , @xmath19 implies that @xmath236 , where @xmath237 .
therefore we can take @xmath235 small enough that @xmath238 ( or , if @xmath239 is small enough one may take @xmath240 , giving a small data " global existence result ) , and hence by lemma [ lemma : abs ] with @xmath241 , there exists a unique small mild solution @xmath242 of nse on @xmath243 . note that ( [ eq : j ] ) and ( [ eq : k ] ) show that @xmath244 as well , so that more specifically @xmath245 . moreover , @xmath246 by lemma [ lemma : lr ] and the standard theory for the heat equation ( see , e.g. , @xcite ) .
the above result can be refined to give a solution which not only remains in @xmath0 , but belongs to @xmath248 as well for @xmath25 .
this will be shown by considering the spaces @xmath249 @xmath250 we will show that there exists some @xmath251 such that @xmath252 and hence there exists a unique small solution @xmath253 by lemma [ lemma : abs ] so long as @xmath254 satisfies @xmath255 note that young s inequality gives @xmath256 hence @xmath257 implies that indeed @xmath258 .
moreover , as before , the resulting solution in the case of ( [ eq : n ] ) will belong more specifically to @xmath259 .
+ + we claim now that @xmath260 which will show that ( [ eq : n ] ) will hold for any @xmath19 for @xmath235 sufficiently small . to prove ( [ eq : p ] ) , for any @xmath261 , take @xmath262 and @xmath263 large enough that @xmath264 , where @xmath265 for @xmath266 and @xmath267 otherwise .
then by young s inequality we have @xmath268 for small enough @xmath25 .
the bilinear estimate ( [ eq : m ] ) is a consequence of the continuous embedding @xmath269 , estimates ( [ eq : k ] ) and ( [ eq : l ] ) and the following claim : [ clm : b ] @xmath270 * proof : * we ll need the following facts : @xmath271 ( i ) is just young s inequality , and ( ii ) can be found for example in @xcite , prop .
11.1 on the oseen kernel " .
( see also @xcite , translated in @xcite ) . now write @xmath272 by ( ii ) and a change of variables we have @xmath273 @xmath274 and hence by ( i ) @xmath275 therefore , for @xmath276 , @xmath277@xmath278 which proves the claim .
( we remark that the constant in the claim does not depend on @xmath235 . )
take @xmath280 and let @xmath281 for some @xmath282 $ ] .
we will show local existence was presented in a course on mathematical fluid mechanics given by prof .
vladim ' ir ver ' ak at the university of minnesota in the spring of 2006 , and can be found as well in @xcite .
for other versions , see e.g. the classical paper @xcite or the more modern treatment in @xcite . ] in the space @xmath283 .
note first that @xmath284 where @xmath285 as follows : since @xmath286 , we can assume @xmath287 .
since @xmath288 for any @xmath289 , integration by parts yields @xmath290 taking @xmath291 and using the inequality @xmath292 which is due to the h " older , @xmath293 and @xmath294 . ] and sobolev inequalities , we have @xmath295 which is exactly @xmath296 recall now that @xmath297 where we can write ( see ( [ aa ] ) ) @xmath298 for some smooth @xmath299 . with a slight abuse of notation for simplicity , we now make the following claim : [ clm : c ] @xmath300 whenever the right - hand side is finite .
* proof : * @xmath301 @xmath302 where @xmath303 and we have used young s inequality for @xmath304 whenever @xmath305 in the @xmath191 variable with @xmath306 in the second line , and one - dimensional fractional integration , @xmath307 for @xmath308 and @xmath309 whenever @xmath310 ( see @xcite ) ] in the @xmath311 variable with @xmath312 in the last .
now since for @xmath276 we have @xmath313 for @xmath314 } f\otimes g$ ] where @xmath315}$ ] is the indicator function for the interval @xmath316 $ ] , for any fields @xmath317 claim [ clm : c ] gives @xmath318 for some @xmath319 independent of @xmath235 .
hence there exists a unique solution @xmath320 by lemma [ lemma : abs ] so long as @xmath321 which is true for small enough @xmath235 due to ( [ eq : five ] ) ( or , for @xmath322 sufficiently small , one may take @xmath240 ) .
moreover , estimate ( 7.26 ) of @xcite implies satisfies the differential nse for a suitable pressure @xmath7 ; however since the mild solution is smooth , we can reduce to this case , see the proof of theorem [ thm : c ] . ]
that @xmath323 the proof of which shows moreover that @xmath324 is continuous .
this , along with the fact that @xmath325 is weakly continuous in @xmath26 on @xmath326 ( see @xcite , ( 7.27 ) ) implies that @xmath327 .
the standard theory of the heat equation implies now that @xmath328 .
[ remark : bdd ] we can furthermore assume that @xmath329 ( for a possibly smaller @xmath235 ) in a manner similar to the proof of ( [ eq : m ] ) , using claim [ clm : b ] and estimate ( [ eq : r ] ) . for a construction of local solutions with higher regularity in similar spaces , see @xcite . at this point
, we also mention the following uniqueness theorem for solutions in the class @xmath330 established in @xcite ( see also @xcite for a simplified proof ) : [ unique ] let @xmath331 both satisfy ( [ eq : mild ] ) for a fixed @xmath280 . then @xmath332 on @xmath243 .
note that there is no size restriction on @xmath322 . for any @xmath280
, the local theory guarantees the existence of a mild solution @xmath333 on @xmath243 for some @xmath107 with @xmath334 , where we define @xmath335 by theorem [ unique ] , there can be at most one mild solution in @xmath336 with initial datum @xmath28 for a fixed @xmath107 , and hence there can be at most one contains the additional information that @xmath337 , one can derive uniqueness from the local theory without using theorem [ unique ] , but we proceed in this way for simplicity . ] such @xmath338 . define @xmath339 by @xmath340 again , by the @xmath88 uniqueness theorem
, @xmath341 implies that @xmath342 for all @xmath343 , hence there exists a unique mild solution @xmath344 on @xmath345 such that @xmath346 and @xmath347 for any @xmath348 .
we ll call @xmath349 the maximal time of existence in @xmath88 of the mild solution associated to @xmath28 .
+ + similarly , for @xmath19 , defining @xmath350 and @xmath351 by @xmath352 there exists a unique mild solution @xmath353 on @xmath354 such that @xmath355 and @xmath356 for any @xmath357 . we ll call @xmath358 the maximal time of existence in @xmath0 of the solution associated to @xmath28 .
+ + the standard embeddings and interpolation inequalities give the estimate @xmath359 and hence @xmath360 for any @xmath107 .
this implies that @xmath361 for any @xmath257 . we will now proceed to show that in fact @xmath362 for @xmath19 .
+ + the following lemma will show that a solution which is continuous in @xmath0 for some time can not leave @xmath0 while remaining continuous in @xmath88 ( or , equivalently , while remaining in @xmath363 ) , and hence that the strict inequality @xmath364 is impossible for @xmath365 .
furthermore , theorem [ unique ] implies that @xmath366 on @xmath367 where @xmath368 . in
what follows , for @xmath280 ( in particular , for @xmath19 ) we ll denote @xmath344 ( @xmath369 ) @xmath370 .
[ lemma : a ] suppose @xmath371;{l^3})$ ] is a mild solution for nse on @xmath316 $ ]
. then * @xmath372 ; * if moreover @xmath373 , then @xmath371;{{\dot h^\frac{1}{2 } } } ) \cap l^2((0,t);{{\dot h^\frac{3}{2}}})$ ] . in order to prove the lemma , we will need the following claim : [ clm : a ] suppose @xmath371;{l^3})$ ] , and fix some @xmath374 .
then * @xmath375 $ ] and * @xmath376 $ ] . if , moreover , @xmath371;{{\dot h^\frac{1}{2}}})$ ] , then * @xmath377 $ ] .
+ * proof of claim [ clm : a ] : * the claim is a simple consequence of the linearity of the operator @xmath218 , estimates ( [ aa](i ) ) and ( [ eq : five ] ) , and the estimate @xmath378 which follows from the @xmath56 energy inequality for the heat equation and the standard embeddings .
writing any one of these as @xmath379 where @xmath371;x)$ ] , we have @xmath380 which proves the claim . @xmath207 + + * proof of lemma [ lemma : a ] : * note first that we may actually assume @xmath6 is a solution in @xmath381 on @xmath382 for some @xmath383 .
this is guaranteed by the local theory in @xmath363 for initial data ( e.g. , @xmath384 ) in @xmath88 . by the same local existence theory ,
we are guaranteed a solution @xmath385 with initial datum @xmath386 for some small @xmath387 . by the uniqueness guaranteed by theorem [ unique ] , we know that in fact @xmath388 on @xmath389 .
+ + set @xmath390 .
if @xmath391 then we have proved ( a ) , so suppose @xmath392 .
since @xmath393 , ( [ eq : five ] ) shows that @xmath394 , hence one may take some @xmath395 small enough so that @xmath396 where @xmath397 is such that @xmath398 guarantees an @xmath399-solution on @xmath400 with initial datum @xmath141 by lemma [ lemma : abs ] and ( [ eq : q ] ) .
now , by claim [ clm : a ] ( ii ) , there exists @xmath401 such that @xmath402 therefore by the local existence theory , we are guaranteed a solution @xmath403 , where @xmath404 .
again by theorem [ unique ] , @xmath405 on @xmath406 . but now clearly @xmath407 , which is a contradiction to the definition of @xmath408 , and ( a ) is proved .
+ + let s now turn to ( b ) .
first suppose that we can show that @xmath19 implies @xmath371;{{\dot h^\frac{1}{2}}})$ ] .
then by replacing @xmath409 above by @xmath410 and replacing claim [ clm : a ] ( ii ) by claim [ clm : a ] ( iii ) , the method used to prove ( a ) can be adjusted to show that @xmath411 .
this fact , along with standard estimates ( see the proof of the local theory in @xmath0 ) , yields the conclusion in ( b ) .
it therefore remains only to show that @xmath412;{l^3 } ) , u_0 \in { { \dot h^\frac{1}{2}}}\ \
\longrightarrow \ \
u\in \mathcal{c}([0,t];{{\dot h^\frac{1}{2}}})\ .\ ] ] in order to do this , we turn to the local solutions @xmath413 of koch - tataru @xcite for ( possibly large , but somewhat regular ) initial data in the space @xmath414 , and a persistency " theorem given in @xcite which states that whenever the initial datum belongs to various spaces of regular data , including @xmath0 , the solution with that datum remains there for positive times .- theory , without resorting to the stronger tools in @xcite and @xcite which we have used here only in the interest of expediency . ] to complete the proof , we rely as before on the uniqueness provided by theorem [ unique ] .
+ + we recall briefly the space @xmath415 in which the koch - tataru solutions were constructed : @xmath416 @xmath417 where for @xmath418 we denote @xmath419 .
recall also the definition ( see , e.g. , @xcite ) that @xmath420 if and only if , but we consider some fixed @xmath107 . ] @xmath421 by ( [ eq : o ] ) , @xmath422 for any @xmath423 .
we know from ( [ eq : p ] ) that @xmath424 , and it can moreover be shown that @xmath425 , where @xmath426 , for any @xmath427 .
note also that @xmath428 , and hence by a change of variables we have @xmath429 which therefore tends to zero as well as @xmath430 by the preceding statement .
+ + the local result in @xcite is that for any @xmath107 there exists some small @xmath431 such that @xmath432 implies the existence of a unique small solution @xmath433 on @xmath243 satisfying @xmath434 . by the statements above , for any @xmath427 one can take @xmath435 small enough so that @xmath436 , @xmath437 . letting @xmath438 we obtain a solution @xmath141 on @xmath439 with initial datum @xmath440 . due to the natural scaling of the equation ,
@xmath441 is a solution on @xmath442 satisfying @xmath443 .
+ + we return now to the proof of ( [ eq : c ] ) .
suppose @xmath365 .
by the local existence theory , there exists a solution @xmath444 on @xmath389 with @xmath355 for some small @xmath445 . by uniqueness ( theorem [ unique ] ) , @xmath446 on @xmath389 hence ,
in particular , @xmath447 .
+ + define @xmath448 , and suppose @xmath449 .
+ + since @xmath450 , there exists some @xmath451 small enough that @xmath452^{{\lambda } _ 1}\|_\infty,\ \|[u({{\overline{t } } _ 1 } ) ] ^{{\lambda } _ 1}\|_{bmo^{-1 } } < { \epsilon } _ 1/2\ , \ ] ] guaranteeing a solution @xmath453 on @xmath454 satisfying @xmath455 , where @xmath456 . due to the embedding @xmath457
, we have @xmath458 , and in particular there exists some @xmath459 such that for @xmath460 small enough we have @xmath461^{\lambda } - [ u({{\overline{t } } _ 1 } ) ] ^{\lambda } \|_{bmo^{-1 } } & \lesssim & \|[u({\tilde{t_1 } } ) ] ^{\lambda } - [ u({{\overline{t } } _ 1 } ) ] ^{\lambda } \|_{3 } \\\\ & = & \|u({\tilde{t_1 } } ) - u({{\overline{t } } _ 1 } ) \|_{3 } \
< \ { \epsilon } _ 1/2 \ , \end{array}\ ] ] so that @xmath462^{\lambda } \|_{bmo^{-1 } } < { \epsilon } _ 1 $ ] .
also , for @xmath463 small enough , we have by ( i ) of claim [ clm : a ] @xmath464^{{\lambda } _ 1}\|_\infty = \sup_{t\in ( 0,t_1)}\sqrt{t}\|e^{t{\delta } } u({\tilde{t_1 } } ) \|_\infty \leq \qquad \qquad \qquad \qquad \qquad \qquad\ ] ] @xmath465^{{\lambda } _ 1}\|_\infty + { \epsilon } _
1/2 < { \epsilon } _ 1 .\ ] ] hence we are also guaranteed a solution @xmath453 on @xmath466 satisfying @xmath467 , where @xmath456 .
taking @xmath468 , we now have a koch - tataru solution @xmath453 on @xmath469 satisfying @xmath467 , where @xmath470 .
+ + since @xmath459 , we have @xmath471 by our definition of @xmath472 .
now , the persistency principle of @xcite states that a koch - tataru solution obtained by the fixed - point argument on an interval with initial datum in @xmath473 for @xmath474 will remain in @xmath473 on that interval and is moreover continuous in that norm . ] . since @xmath475
, we see that @xmath476 .
theorem [ unique ] then implies that @xmath477 on @xmath478 , and hence @xmath479 .
this , however , contradicts our assumption on @xmath472 , hence necessarily @xmath480 , and ( [ eq : c ] ) is proved , completing the proof of lemma [ lemma : a ] .
@xmath207 + + finally , we state here for convenience some known a priori results regarding globally - defined mild solutions to nse evolving from possibly large data : [ decay ] let @xmath481 be a global - in - time solution to ( [ eq : mild ] ) for some divergence - free @xmath482 , where @xmath168 is either @xmath483 or @xmath122 . then @xmath484
this is proved for @xmath485 in @xcite , and for @xmath486 in @xcite ( see also @xcite ) .
the following result is classical ( see , e.g. , @xcite for a proof ) : [ energysmall ] let @xmath487 be any divergence - free field , and let @xmath21 .
there exists a small universal constant @xmath488 such that @xmath489 implies that @xmath49 and for any @xmath490 the following energy inequality holds : @xmath491 together with theorem [ decay ] , we now have the following : [ energy ] let @xmath492 , @xmath493 and @xmath21 , and suppose @xmath494 .
then for some large @xmath495 we have , for all @xmath496 , @xmath497 finally , we remark that the following fact follows easily from the local theory and corollary [ energy ] ( recall @xmath498 ): [ thm : blowup ] for @xmath282 $ ] , take @xmath499 to be either @xmath500 or @xmath501 .
then for any @xmath502 , @xmath503
the following blow - up " criterion for the navier - stokes equations was proved in @xcite : if @xmath6 is a _ weak _ leray - hopf " solution of the navier - stokes equations ( essentially , a distributional solution to ( [ eq : nsea ] ) satisfying a natural energy estimate ) and if @xmath504 is the maximal interval on which @xmath6 is smooth , then if @xmath505 , we must have @xmath506 alternatively , one can say that @xmath507 implies that @xmath508 .
+ + we would like to suggest a different approach to this statement , using the method of critical elements " introduced in @xcite , in the setting of _ mild _ solutions .
we consider here a simpler special case of the above , namely the statement that for any @xmath365 , @xmath509 where @xmath510 ( see ( [ b ] ) ) is the associated mild solution .
theorem [ decay ] then implies moreover that @xmath511 , so that ( [ eq : s ] ) may be thought of as a statement of global existence and scattering " ( time - asymptotic convergence to a solution of the linear equation ) , such as was considered in @xcite for non - linear schr " odinger equations .
we start with a setup following @xcite as follows : + + suppose ( [ eq : s ] ) is false ; then there exists some maximal finite @xmath512 ( see ( [ eq : an ] ) below ) such that @xmath513 while , for any @xmath59 , there exists some initial datum @xmath514 and solution @xmath515 with @xmath516 and @xmath517 .
note that the local theory implies @xmath57 is well - defined since @xmath49 ( and moreover @xmath518 for @xmath240 as well by corollary [ energy ] ) for @xmath519 small enough .
+ + specifically , we define the critical value @xmath104 by @xmath520 under the assumption that @xmath521 , we ll prove the following two theorems , whose analogs were used to prove statements similar to ( [ eq : s ] ) in @xcite : [ thm : b ] suppose @xmath521 , and let @xmath524 with associated mild solution @xmath21 satisfy @xmath525 and @xmath526 .
for any @xmath527 such that @xmath528 , there exist sequences @xmath529 with @xmath530 , @xmath531 and @xmath532 with @xmath533 such that a subsequence of @xmath534 converges in @xmath88 .
this is a slightly weaker version of the compactness in @xmath88 of the closure of a family @xmath123 , where @xmath535 for some @xmath536 , @xmath537 , as was proved in @xcite .
note , however , that for any sequence @xmath538 , setting @xmath539 and @xmath540 , a convergent subsequence exists by the @xmath88-continuity of @xmath6 .
the only remaining scenario is therefore treated by the weaker statement in theorem [ thm : b ] .
+ + the idea is that @xmath521 implies the existence of a critical element which produces a type of compactness ( see ( [ eq : kay ] ) ) in the family @xmath123 . if one can rule out such compactness , then one has proved ( [ eq : s ] ) .
( the nonexistence of a particular example of a fully compact family k was established in @xcite and @xcite regarding the so - called `` self - similar '' solutions . )
the program is therefore completed with the following ridigity " theorem : theorem [ thm : c ] has the immediate corollary that no such element @xmath28 can exist , since @xmath542 whereas by assumption @xmath543 .
this implies , due to theorems [ thm : a ] and [ thm : b ] , that @xmath544 , which proves the regularity criterion ( [ eq : s ] ) ( theorem [ thm : regularity ] ) as desired . in order to prove theorem [ thm : a ] ( and also theorem [ thm : b ] )
, we will need the following profile decomposition " which was proved in @xcite .
recall the notation @xmath545 for the mild solution in @xmath546 associated to an initial datum @xmath365 .
[ thm : prof ] let @xmath64 be a bounded sequence of divergence - free vectors in @xmath0 .
then there exist @xmath547 and @xmath548 which are orthogonal " in the sense that , for @xmath549 , @xmath550 , @xmath551 @xmath552 and a sequence of divergence - free is a consequence of ( [ eq : aa ] ) and the orthogonality of the scales - cores .
indeed , @xmath553 where @xmath554 is divergence - free . ]
vector fields @xmath555 with @xmath556 for an at most finite number of @xmath557 such that the following is true : after possibly taking a subsequence in @xmath74 , @xmath558 and @xmath559 for any @xmath560 , for @xmath144 and @xmath561 where @xmath562 ( or @xmath563 if @xmath564 for all @xmath565 ) for any fixed numbers @xmath566 , @xmath567 , @xmath568 and @xmath569 for some @xmath570 and @xmath571 , and the following orthogonality properties hold : @xmath572 @xmath573 and @xmath574 recall that @xmath575 is defined for @xmath107 by ( [ eq : defet ] ) , and note that @xmath576 is simply some time such that for @xmath577 , ( [ eq : ab ] ) avoids the finite blow - up times of any of the @xmath578 s due to the natural scaling of the equation : if @xmath141 is a solution on @xmath326 then @xmath579 is a solution on @xmath580 for any @xmath581 . + +
we ll also need the following fact regarding the profiles " @xmath95 which appear in theorem [ thm : prof ] : this is in fact implicit in the statement of theorem [ thm : prof ] .
the proof can be found in @xcite , for example if one follows the proof of theorem 2 part ( iii ) in section 3.2.2 of that paper .
one sees that , under the assumptions of claim 3.5 above , the remainder @xmath98 belongs to @xmath586 for large @xmath80 and @xmath74 and hence @xmath94 belongs to the same space due to ( [ eq : ab ] ) and standard heat estimates .
theorem 2.12 then implies that @xmath587 for such @xmath74 .
@xmath207 + + finally , we will need the following backwards uniqueness " type of lemma ( a stronger version of which is proved in the last section ) : lemma [ lemma : zero ] will follow from the following backwards uniqueness theorem for systems of parabolic equations , which was proved in @xcite ( for the more general situation of a half - space see also @xcite , but we do not require such generality ) : [ thm : backwards ] fix any @xmath594 and @xmath595 .
let @xmath596 , and suppose a vector - valued function @xmath141 and its distributional derivatives satisfy @xmath597 for any bounded subset @xmath598 , @xmath599 for all @xmath600 , @xmath601 on @xmath602 and @xmath603 for all @xmath604 .
then @xmath605 in @xmath602 .
* proof of lemma [ lemma : zero ] : * we will apply theorem [ thm : backwards ] with @xmath606 , the associated vorticity to the velocity field @xmath6 .
suppose for the moment that @xmath607 satisfies the assumptions of theorem [ thm : backwards ] ( in the whole space , in fact ) .
then @xmath590 implies @xmath608 , and hence @xmath609 in @xmath610 , for any @xmath611 . since mild solutions to nse satisfy @xmath612 in the distributional sense
, we see that @xmath613 , @xmath614 for each @xmath615 which implies that @xmath616 in @xmath617 . taking @xmath618 and using continuity at @xmath619 ,
the theorem follows .
( in the forward direction , one uses the uniqueness of theorem [ unique ] . ) + + the vorticity satisfies the equation @xmath620 in the sense of distributions .
the assumptions of theorem [ thm : backwards ] follow therefore from the fact that @xmath621 for any @xmath622 , which follows from remark [ remark : bdd ] and arguments similar to those in the proof of lemma [ lemma : a ] .
the bounds on the derivatives of @xmath6 ( and hence of @xmath607 ) are then standard , see for example @xcite .
@xmath207 in this section we ll prove theorem [ thm : a ] , which establishes the existence of a critical element .
+ + * proof of theorem [ thm : a ] : * define @xmath104 by ( [ eq : an ] ) and assume that @xmath521 .
( note that @xmath104 is well - defined by global existence for small data . ) by definition of @xmath104 , we can pick @xmath66 as @xmath623 and @xmath624 , @xmath625 such that @xmath626 holds for all @xmath74 .
one can not prove , as one would naturally hope , that a subsequence of these @xmath96 converges to some @xmath112 satisfying the assertion of the theorem , but we will see that something similar is true .
we may assume that @xmath627 and hence we may apply the profile decomposition ( theorem [ thm : prof ] ) to the sequence @xmath628 . for the remainder of this section , @xmath629 etc .
will refer to this particular sequence .
we will complete the proof of the theorem by showing that @xmath82 satisfies the assertion of the theorem for some @xmath630 , and the profile @xmath631 will be our critical element .
+ + define @xmath632 , and note that ( [ eq : af ] ) and ( [ eq : ac ] ) imply that @xmath633 so that @xmath634 and therefore by the small data result , @xmath635 for large enough @xmath565 . corollary [ energy ] then implies that @xmath636 for large @xmath565 as well . + + note that theorem [ thm : blowup ] and claim [ lemma : aa ] together imply that there exists some @xmath637 such that @xmath638 , and hence we may re - order the profiles in the decomposition such that for some @xmath639 , @xmath640 and @xmath641 note also that theorem [ thm : blowup ] and corollary [ energy ] imply that @xmath642 to prove theorem [ thm : a ] , it suffices therefore to show that @xmath643 for some @xmath644 .
this will be accomplished using the following claim , which extends the orthogonality property ( [ eq : ac ] ) : [ clm : d ] set @xmath645 and @xmath646 for @xmath647 .
fix some @xmath648 , and let @xmath649 be any sequence such that @xmath650 $ ] for all @xmath74 .
then there exist subsequences @xmath651 and @xmath652 depending on @xmath653 and indexed by @xmath654 such that for @xmath655 and @xmath656 , @xmath657 where @xmath658 for each @xmath653 and we have set @xmath659 . assuming claim [ clm : d ] momentarily , we prove ( [ eq : ai ] ) ( and hence theorem [ thm : a ] ) as follows : + + let @xmath660 , so that @xmath661 . for any fixed @xmath653 ,
there exists some @xmath662 such that @xmath663 for infinitely many @xmath74 .
also , there exists some @xmath644 such that @xmath664 for infinitely many @xmath653 , say for a subsequence @xmath665 as @xmath666 . for each fixed @xmath667 , take @xmath668 to be a subsequence of @xmath654 s indexed by @xmath669 such that @xmath670 ( where @xmath651 , @xmath652 are as in claim [ clm : d ] ) , and @xmath671 as @xmath669 .
+ + since @xmath672 , we have @xmath673 by ( [ eq : ak ] ) .
let @xmath674 so that @xmath675 as @xmath676 .
note that @xmath677 , else we would have @xmath678 by the definition of @xmath104 which would contradict ( [ eq : al ] ) .
theorem [ thm : a ] will be proved if we show @xmath679 ( whereby we can set @xmath680 ) , so it remains only to show that @xmath681 , which we will now do .
+ + by continuity in @xmath0 , we may take @xmath682 $ ] such that @xmath683 .
set @xmath684 , and note that for @xmath685 and @xmath686 we have @xmath687 by ( [ eq : aj ] ) .
we may therefore apply claim [ clm : d ] with @xmath688 to conclude that for fixed @xmath667 , there exists a subsequence of @xmath689 s such that for @xmath686 , @xmath690 and @xmath685 we have @xmath691 where @xmath692 as @xmath669 .
recall from ( [ eq : ah ] ) that we defined @xmath693 so that @xmath694 therefore , by ( [ eq : am ] ) , we have @xmath695 keeping @xmath667 fixed and letting @xmath669 , this gives @xmath696 . finally ,
letting @xmath666 , we see @xmath697 as desired proving theorem [ thm : a ] .
@xmath207 [ remark : a ] these arguments can moreover be used to show that in fact @xmath698 .
that is , there is one and only one profile @xmath699 which blows up in finite time .
this is due to the fact that we have now shown that @xmath700 as @xmath666 , hence ( [ eq : am ] ) can be used to show that @xmath701 can be made arbitrarily small for any @xmath702 , which implies that @xmath635 by the global existence for small data result and the natural scaling of the equations .
the easiest to show is ( [ eq : ddf ] ) : since @xmath713 is defined for all @xmath637 by assumption , the properties of the profile decomposition imply , see ( [ eq : dda ] ) . ]
( see @xcite ) that @xmath714 for all @xmath74 , and so @xmath715 by ( [ eq : ah ] ) . using ( [ eq : ac ] ) and heat estimates , we see @xmath716 for large enough @xmath74 . for each @xmath654 ,
use ( [ eq : ae ] ) to take @xmath80 large enough , and a corresponding sufficiently large @xmath74 ( both in an increasing fashion ) such that @xmath717}\|r^j_n(t)\|_{{\dot h^\frac{1}{2}}}\leq 1\ , \ ] ] whereupon ( [ eq : ab ] ) gives @xmath718 .
+ + to prove ( [ eq : ddd ] ) , we take the following diagonal - type subsequence : if @xmath719 is bounded , pass to a subsequence such that @xmath720 . take @xmath721 to be the first element of this subsequence ( or of the original subsequence if @xmath722 is unbounded ) .
( this is @xmath723 . )
assume now that the first @xmath724 values of @xmath74 are chosen .
if @xmath725 is bounded , pass to a subsequence of the @xmath726st subsequence such that @xmath727 .
take @xmath651 to be the first element of this sequence which comes after @xmath728 ( or of the original @xmath726st subsequence if @xmath729 is unbounded ) . in this way , @xmath730 as @xmath731 ( @xmath655 ) for any @xmath565 such that @xmath78 is bounded .
+ + note that for @xmath732 , @xmath733 . therefore if @xmath734 , necessarily @xmath735 and @xmath736 .
hence by theorem [ decay ] , @xmath737 as @xmath51 , and hence @xmath738 in the expression in ( [ eq : ddd ] ) , whenever @xmath78 is bounded , using the @xmath0-continuity of @xmath95 in a neighborhood of @xmath739 , we can replace @xmath740 we can also assume that @xmath741 by approximating in @xmath742 .
if both @xmath78 and @xmath743 are unbounded , then @xmath744@xmath745 by ( [ eq : ddg ] ) .
if @xmath99 is bounded and @xmath746 is unbounded , we estimate the above expression by @xmath747 as @xmath748 since we have replaced @xmath78 by the constant @xmath739 in the first term , and the second term tends to zero .
if both @xmath99 and @xmath746 are bounded , replace @xmath99 by @xmath739 and @xmath746 by @xmath749 , and note by a change of variables that @xmath750 @xmath751 which , since we ve assumed the functions in the integrand are compactly supported , is easily seen to tend to zero as @xmath748 if @xmath752 or if @xmath753 and @xmath754 .
the case @xmath755 is handled similarly , and ( [ eq : ddd ] ) is proved .
+ + ( [ eq : dde ] ) is proved using ( [ eq : ddd ] ) as follows : note that @xmath756 suppose the sum on the right has @xmath757 terms in it . for a fixed @xmath80 , ( [ eq : ddd ] )
allows us to take @xmath74 large enough ( that is , we take a further subsequence of @xmath651 ) so that all the terms are smaller than @xmath758 , where @xmath759 . as @xmath80 increases with @xmath654 ,
we let @xmath74 increase sufficiently rapidly with @xmath654 as well so that this is true for all @xmath654 , and ( [ eq : dde ] ) is proved . + + returning now to the proof of claim [ clm : d ] , we want to show that for any @xmath261 , there exists some @xmath760 such that , for all @xmath761 , @xmath762 where @xmath763 we write @xmath764 for a sufficiently large @xmath765 to be chosen , expand the first term in @xmath766 and cancel the equal terms .
we ll then split @xmath766 into @xmath767 , where @xmath768 contains all terms with @xmath769 for @xmath770 ( see below ) .
note that for @xmath771 , @xmath736 and hence , for @xmath765 sufficiently large , theorem [ energysmall ] and ( [ eq : ag ] ) give the energy estimate @xmath772 .
then , using ( [ eq : ac ] ) , ( [ eq : ag ] ) and ( [ eq : dde ] ) , for any @xmath773 we have @xmath774 for @xmath775 sufficiently large .
now let @xmath776@xmath777 we estimate @xmath778@xmath779 noting that we have @xmath780 by ( [ eq : ddf ] ) , heat estimates and ( [ eq : ac ] ) , and by our choice of subsequences , respectively . using these and ( [ eq : ddj ] ) , for any @xmath261 we can make @xmath781 for large enough @xmath765 , which we now fix .
we are now left to estimate @xmath782 @xmath783@xmath784@xmath785 the first term is handled by ( [ eq : ddd ] ) . since we may pass to a subsequence such that @xmath786 as @xmath787 by ( [ eq : ae ] ) , using ( [ eq : ddf ] ) and the fact that @xmath788 we see that the terms with @xmath789 are small for large @xmath654 .
the only remaining issue is @xmath790 consider @xmath791 for @xmath792 .
if @xmath793 , then , since @xmath794 , we can use again the fact that @xmath795 as @xmath748 to see that the term is small for large @xmath654 ( recall that @xmath706 ) . otherwise , it suffices to consider ( by replacing @xmath99 by @xmath739 as before ) @xmath796@xmath797(y)\ dy\ .\ ] ] define @xmath798(y)$ ] .
we claim that @xmath799 in @xmath0 and hence this term is small for large @xmath654 as well ( for each @xmath800 $ ] ) , and claim [ clm : d ] is proved .
indeed , note that @xmath801 hence @xmath802 has a weak accumulation point in @xmath0 .
note also that @xmath803 as @xmath731 ( after passing to a subsequence ) by ( [ eq : ad ] ) , so in particular @xmath799 in @xmath26 .
any weak limit in @xmath0 for any subsequence is therefore also zero due to the embedding @xmath804 , hence zero is the only weak accumulation point of @xmath802 in @xmath0 and thus @xmath799 in @xmath0 .
this concludes the proof of claim [ clm : d ] ( and theorem [ thm : a ] ) .
@xmath207 in this section , we ll prove theorem [ thm : b ] which establishes the compactness of any critical element arising from theorem [ thm : a ] under the assumption that @xmath521 .
+ + * proof of theorem [ thm : b ] : * we will need the following claim : [ clm : e ] define @xmath104 by ( [ eq : an ] ) , and suppose that @xmath521 .
suppose @xmath805 and @xmath21 are such that @xmath525 and @xmath526 . for any sequence @xmath806 such that @xmath807 , let @xmath578 , @xmath95 be the profiles associated with the sequence @xmath808 .
then , after re - ordering , @xmath809 , and @xmath810 for all @xmath74 . moreover , after passing to a subsequence in @xmath74 , for @xmath811 either @xmath812 or @xmath813 * proof of claim [ clm : e ] : * first note that the following considerations hold for _ any _ sequence @xmath527 : + + letting @xmath814 , @xmath815 and noting @xmath124 implies that @xmath816 for all @xmath817 by theorem [ thm : blowup ] , we can think of @xmath81 as a minimizing sequence for @xmath104 , since @xmath818 for all @xmath74 .
proceeding as before and applying the profile decomposition , we see that , for a subsequence , @xmath819 and , by uniqueness , @xmath820 for @xmath821 $ ] , where @xmath822 for @xmath705 with the orthogonality properties ( [ eq : ac ] ) - ( [ eq : ae ] ) .
this is justified since we saw before ( see ( [ eq : al ] ) , ( [ eq : ai ] ) , remark [ remark : a ] , etc . )
that we may re - order the elements so that @xmath823 , @xmath564 for all @xmath811 and @xmath824 @xmath825 and it is noted moreover in @xcite that equation ( [ eq : dda ] ) holds for all @xmath74 .
( in fact , ( [ eq : dda ] ) is a simple consequence of the properties of the profile decomposition . ) + + suppose now that @xmath528
. then ( [ eq : dda ] ) gives @xmath826 we will now show that moreover , for @xmath811 , either @xmath812 or , after possibly passing to a subsequence , the limit ( [ eq : as ] ) holds as follows : take @xmath827 $ ] such that @xmath828 then , letting @xmath829 , we can apply claim [ clm : d ] ( where actually @xmath830 ) to see that there exist subsequences @xmath655 , @xmath656 indexed by @xmath654 for fixed @xmath653 such that @xmath831 where @xmath832 as @xmath731 for fixed @xmath653 . therefore , as in remark [ remark : a ] , one can take a subsequence @xmath833 such that , for @xmath811 , @xmath834 if along this subsequence @xmath835 we had that @xmath836 were bounded , one could therefore extract a convergent subsequence with limit @xmath837 , and therefore conclude that @xmath838 by the @xmath0-continuity and hence that @xmath812 by lemma [ lemma : zero ] . since @xmath839 ,
if @xmath840 then the limit ( [ eq : as ] ) must therefore hold for some subsequence , and claim [ clm : e ] is proved .
@xmath207 + + returning now to the proof of theorem [ thm : b ] , fix a sequence @xmath806 such that @xmath841 .
we define in the definition of @xmath842 is chosen only for simplicity ; one may in fact replace @xmath843 by @xmath234 for any fixed @xmath844 . ] a corresponding sequence @xmath529 by @xmath845 for each @xmath74 .
note that , for @xmath811 , ( [ eq : as ] ) gives ( after passing to a subsequence in @xmath74 ) @xmath846 note also that ( [ eq : dda ] ) implies that @xmath847 .
indeed , we have @xmath848 notice that , since @xmath849 , we may write @xmath850 and hence , by ( [ eq : ab ] ) we have ( for a further subsequence in @xmath74 ) @xmath851 fix any @xmath261 . since @xmath852 for all @xmath74 , the orthogonality properties ( [ eq : ad ] ) and ( [ eq : ae ] ) and young s inequality allow us to fix some @xmath80 so large that @xmath853 and @xmath854 for large enough @xmath74 .
for such a @xmath80 , we may take @xmath74 large enough that @xmath855 due to ( [ eq : ar ] ) and the fact that @xmath582 for @xmath811 which implies that @xmath856 by theorem [ decay ]
. we have therefore shown ( due to ( [ eq : at ] ) ) that for any @xmath261 , there exists @xmath857 such that @xmath858 for all @xmath859 .
a simple change of variables @xmath860 shows that @xmath861,s_n\right ) - u_1 \left ( \tfrac{1}{2}t^*(v_{0,1 } ) \right ) \right\|_3 < { \epsilon } \ ] ] for @xmath859 , and hence setting @xmath862 ( note that @xmath533 by ( [ eq : aaaaa ] ) ) and @xmath863 , we see that one can , of course , find a similar sequence which converges to @xmath864 . ]
@xmath865 in @xmath88 as @xmath748 and theorem [ thm : b ] is proved .
@xmath207 in this section , we ll prove theorem [ thm : c ] which denies the possibility of the compactness of a critical element as established in theorem [ thm : b ] .
this , due to theorems [ thm : a ] and [ thm : b ] , gives an alternative proof of the regularity criterion ( [ eq : s ] ) in the case of mild solutions ( which follows from the more general result in @xcite ) , and concludes the critical element proof of theorem [ thm : regularity ] .
+ + we will use three key known results .
the first two concern systems of parabolic equations : the backwards uniqueness theorem , theorem [ thm : backwards ] , proved in @xcite ( see also @xcite ) , and the following unique continuation " theorem , a proof of which can be found in @xcite but which was already a well - known fact from the unique continuation theory of differential inequalities ( see also @xcite ) : [ thm : uniquecont ] let @xmath866 for some @xmath867 , and suppose a vector - valued function @xmath141 and its distributional derivatives satisfy @xmath868 and there exist @xmath869 ( @xmath705 ) such that @xmath870 and @xmath871 for all @xmath872
. then @xmath873 for all @xmath874 .
the third result we will need is a local regularity criterion ( with estimates ) originating in @xcite and generalized in @xcite for the so - called suitable weak solutions " of the navier - stokes equations , which are defined as follows : [ def : suitable ] let @xmath875 be an open subset of @xmath876 , @xmath877 .
we call the pair @xmath878 a _ suitable weak solution _ of the navier - stokes equations in @xmath879 if @xmath880 , @xmath881 , @xmath6 and @xmath7 satisfy nse in distributions and the following local energy inequality holds : @xmath882 @xmath883 for almost all @xmath884 and for all non - negative cut - off functions @xmath885 which vanish in a neighborhood of the parabolic boundary @xmath886)$ ] .
[ lemma : smallness ] there exist positive absolute constants @xmath887 and @xmath888 for @xmath705 with the following property : if a suitable weak solution @xmath878 of nse on @xmath889 , where @xmath890 for @xmath418 , satisfies the condition @xmath891 then @xmath892 is h " older continuous on @xmath893 for any @xmath705 , and for each @xmath653 we have the estimate @xmath894 we will now prove theorem [ thm : c ] by using the backwards uniqueness , unique continuation and smallness criterion results ( theorem [ thm : backwards ] , theorem [ thm : uniquecont ] and lemma [ lemma : smallness ] ) to establish the following stronger version of lemma [ lemma : zero ] : [ lemma : final ] suppose @xmath127 and set @xmath895 .
suppose there exist some @xmath896 such that @xmath897 and @xmath898 and a sequence of numbers @xmath529 such that @xmath899 such that the following two properties hold : @xmath900 then @xmath901 .
( the case @xmath45 implies lemma [ lemma : zero ] due to the continuity properties of mild solutions . )
note the immediate corollary that there exists no @xmath127 such that @xmath124 and ( 1 ) and ( 2 ) hold with @xmath902 , since the conclusion of lemma [ lemma : final ] implies that @xmath903 by theorem [ unique ] .
this will exactly rule out the critical element produced in theorem [ thm : a ] and prove the regularity criterion ( [ eq : s ] ) .
let us postpone the proof of lemma [ lemma : final ] for the moment .
+ + * proof of theorem [ thm : c ] : * as in theorem [ thm : b ] , define @xmath104 by ( [ eq : an ] ) and suppose @xmath904 , and let @xmath524 with associated mild solution @xmath895 satisfy @xmath905 and @xmath526 .
+ + fix any @xmath906 such that @xmath841 , and let @xmath907 , @xmath533 and @xmath531 be the associated sequences guaranteed by theorem [ thm : b ] . using the conclusion of the same theorem , define @xmath908 for any @xmath144 and pass to a subsequence in @xmath74 so that @xmath909 in @xmath26 for some @xmath910 .
we now make the following important claim : * proof of claim [ clm : zero ] : * fix @xmath912 .
we calculate : @xmath914@xmath915 for any small @xmath261 , define @xmath916 .
we then estimate @xmath917 @xmath918 @xmath919 for sufficiently large @xmath74 .
the first term of the last line is arbitrarily small for small @xmath920 , and for fixed @xmath920 the second term is arbitrarily small for sufficiently large @xmath74 due to the fact that @xmath533 .
this proves the claim .
@xmath207 + + assuming lemma [ lemma : final ] holds and taking @xmath921 , claim [ clm : zero ] clearly completes the proof of theorem [ thm : c ] .
therefore it remains only to establish the following : + + * proof of lemma [ lemma : final ] : * since @xmath922 is the application we seek , we will consider only that case ( the case @xmath45 is even easier ) .
assume therefore that @xmath280 is such that @xmath124 , @xmath923 where @xmath21 , and there exists @xmath529 with @xmath907 as @xmath748 such that @xmath924 for any fixed @xmath912 .
+ + recall that @xmath925 for a vector field @xmath198 , and for a tensor @xmath926 we can write @xmath927 , where @xmath152 and we sum over repeated indices . therefore , letting @xmath928 , @xmath929 and defining @xmath930 , we have @xmath931 note that @xmath7 is well - defined as a distribution since @xmath932 boundedly for each @xmath653 by the calderon - zygmund theory , so that @xmath933 hence for @xmath934 , and denote the corresponding norm by @xmath935 . ]
moreover , @xmath7 satisfies the equation @xmath937 in distributions and therefore @xmath938 is smooth in @xmath191 for any @xmath939 since @xmath588 is smooth for such t. moreover , setting @xmath940 for any fixed @xmath941 and any @xmath418 , we have the following estimates from the classical elliptic theory ( see , e.g. , @xcite ) @xmath942 for any @xmath943 , we now have @xmath944(s ) \ ds\ .\ ] ] it is clear that @xmath945 , due to the fact that @xmath946 . therefore @xmath6 satisfies
( see @xcite , theorem 11.2 ) @xmath947 in distributions .
we now claim that the pair @xmath878 locally forms a suitable weak solution , in the sense of definition [ def : suitable ] .
+ + recall that , by construction , @xmath948 for any @xmath949 , so that the ladyzhenskaya - prodi - serrin " regularity condition ( see , e.g. , @xcite ) implies that @xmath6 is smooth on @xmath950 ( or , alternatively , the construction gives @xmath951 for any @xmath261 and hence is smooth , see , e.g. , @xcite , prop .
15.1 ) . therefore ( [ eq : aae ] ) and ( [ eq : aad ] ) together imply that @xmath938 is smooth in @xmath191 for @xmath952 and @xmath953 for any @xmath653 .
+ + we clearly have @xmath954 for any bounded @xmath955 .
since @xmath6 and @xmath7 are sufficiently smooth on @xmath950 we may multiply the equation ( [ eq : aaa ] ) by @xmath956 for @xmath957 as in definition [ def : suitable ] and integrate by parts to derive the local energy inequality in @xmath958 for a suitably larger region such that @xmath959 , which implies that @xmath960 .
( indeed , the integrations by parts happen only at the level of integration in @xmath191 , and then @xmath961 allows integration in @xmath311 . ) noting also that we have @xmath962 we see that @xmath963 .
the pair @xmath878 now clearly satisfies all the requirements to form a suitable weak solution to nse in @xmath964 for any @xmath955 .
+ + equation ( [ eq : aab ] ) implies moreover that for any @xmath261 one can find @xmath912 large enough that @xmath965 for all @xmath966 such that @xmath967 . therefore , due to lemma [ lemma : smallness ] ( and an appropriate scaling argument )
, we see that @xmath6 is smooth on @xmath968 $ ] for some sufficiently large @xmath969 , and by shifting spatial regions , we have the global bounds @xmath970 since @xmath6 is continuous up to @xmath971 outside @xmath972 , claim [ clm : zero ] now implies that @xmath973 for all @xmath191 such that @xmath974 .
we therefore also have @xmath975 on @xmath976 .
taking the curl of equation ( [ eq : aaa ] ) , we see that the vorticity @xmath607 satisfies the inequality @xmath977 in @xmath978 for some @xmath979 due to ( [ eq : aac ] ) .
since @xmath607 is bounded and smooth in @xmath978 , we can apply the backwards uniqueness theorem , theorem [ thm : backwards ] , to conclude that @xmath609 in @xmath978 .
+ + now , in @xmath980 $ ] we have @xmath6 , and hence @xmath607 , is smooth , hence @xmath981 for all @xmath982 .
we can therefore apply the theorem of unique continuation , theorem [ thm : uniquecont ] , to @xmath607 , since @xmath609 in @xmath983 , to conclude ( by appropriate shifting of local regions ) that @xmath609 in @xmath984 .
therefore , due to the divergence - free condition , @xmath6 is harmonic in @xmath984 which , along with the fact that @xmath946 , implies @xmath592 in @xmath984 .
it is then easy to apply theorem [ thm : backwards ] again ( see the proof of lemma [ lemma : zero ] ) to see that @xmath985 which proves lemma [ lemma : final ] and concludes the proof of theorem [ thm : c ] and theorem [ thm : regularity ] .
@xmath207 jean - yves chemin .
le systme de navier - stokes incompressible soixante dix ans aprs jean leray . in _ actes des journes mathmatiques la mmoire de jean leray _ ,
volume 9 of _ smin .
_ , pages 99123 .
france , paris , 2004 .
giulia furioli , pierre gilles lemari - rieusset , ezzedine zahrouni , and ali zhioua .
un thorme de persistance de la rgularit en norme despaces de besov pour les solutions de koch et tataru des quations de navier - stokes dans @xmath988 .
, 330(5):339342 , 2000 .
isabelle gallagher , drago iftimie , and fabrice planchon .
asymptotics and stability for global solutions to the navier - stokes equations . in _ journes `` quations aux drives partielles '' ( forges - les - eaux , 2002 ) _ ,
pages exp .
nantes , nantes , 2002 .
terence tao , monica visan , and xiaoyi zhang .
global well - posedness and scattering for the defocusing mass - critical nonlinear schrdinger equation for radial data in high dimensions . , 140(1):165202 , 2007 . | in this paper we present an alternative viewpoint on recent studies of regularity of solutions to the navier - stokes equations in critical spaces .
in particular , we prove that mild solutions which remain bounded in the space @xmath0 do not become singular in finite time , a result which was proved in a more general setting by l. escauriaza , g. seregin and v. ver ' ak using a different approach .
we use the method of concentration - compactness " @xmath1 rigidity theorem " which was recently developed by c. kenig and f. merle to treat critical dispersive equations . to the authors knowledge , this is the first instance in which this method has been applied to a parabolic equation .
+ + + * rsum : * dans cet expos , nous prsentons un point de vue diffrent sur les tudes rcentes concernant la rgularit des solutions des quations de navier - stokes dans les espaces critiques . en particulier , nous dmontrons que les solutions faibles qui restent bornes dans lespace @xmath2 ne deviennent pas singulires en temps fini .
ce rsultat a t dmontr dans un cas plus gnral par l. escauriaza , g. seregin et v. ver ' ak en utilisant une approche diffrente .
nous utilisons la mthode de @xmath3 concentration - compacit @xmath4 @xmath1 @xmath3 thorme de rigidit @xmath4 qui a t rcemment dveloppe par c. kenig et f. merle pour traiter les quations dispersives critiques .
la connaissance des auteurs , cest la premire fois que cette mthode est applique une quation parabolique .
[ section ] [ thm]lemma [ thm]claim [ thm]remark [ thm]definition [ thm]corollary |
let @xmath22 $ ] denote the set of @xmath3-hlder continuous functions @xmath23\to { \mathbb{r}}$ ] . in 2009 ,
kahane and katznelson @xcite proved the following result and asked whether it is sharp . for every @xmath0 there exists a function @xmath24 such that if @xmath9 $ ] and @xmath25 is of bounded variation , then the hausdorff dimension satisfies @xmath26 .
is the above result the best possible ?
we answer this question negatively and determine the optimal bound .
let @xmath27}\{\dim_{\mathcal{h}}a : f|_{a } \textrm { is of bounded variation}\},\ ] ] so that the above theorem states @xmath28 , see figure [ f : v ] . [
t : va ] for all @xmath0 we have @xmath29 ; ; ; ; ( 6.7,0 ) ; \to [ 0,1]$ ] with hlder exponent .
its graph consists of 6 affine copies of itself.,scaledwidth=90.0% ] kahane and katznelson also asked about dimensions of sets @xmath30 such that the restriction to @xmath30 is hlder continuous .
( see the next section for related results . )
we present two constructions , one deterministic and one stochastic , of functions that are not hlder on any set of high enough dimension .
first we consider self - affine functions .
these are constructed in definition [ d : sa ] below , see figure [ f:23 ] for illustration .
[ t : sa0 ] there is a dense set @xmath31 with the following property .
for each @xmath32 there is a self - affine function @xmath33 such that for all @xmath9 $ ] 1 .
if @xmath34 is @xmath14-hlder continuous for some @xmath15 , then @xmath35 ; 2 . if @xmath34 is of bounded variation , then @xmath36 . for a stochastically self - affine process , fractional brownian motion ( see definition [ d : fbm ] ) , we prove the following .
[ t : fbm0 ] let @xmath0 and let @xmath8\}$ ] be a fractional brownian motion of hurst index @xmath3 .
then , almost surely , for all @xmath9 $ ] 1 .
if @xmath37 is @xmath14-hlder continuous for some @xmath15 , then @xmath35 ; 2 . if @xmath37 is of bounded variation , then @xmath38 .
[ c : fbm ] let @xmath0 and let @xmath8\}$ ] be a fractional brownian motion of hurst index @xmath3 .
then @xmath39 let @xmath40 be the zero set of @xmath16 and let @xmath41 : b(t)=\max_{s\in [ 0,t ] } b(s)\}$ ] be the set of record times of @xmath16 .
it is classical that , almost surely , @xmath42 , see ( * ? ? ?
* chapter 18 ) . for the record ,
let us state the following , more subtle fact .
[ p : r ] almost surely , @xmath43 .
we could not find a reference for this in the literature , and include a proof in section [ s : fbm ] .
clearly @xmath40 and @xmath44 witness that theorem [ t : fbm0 ] and corollary [ c : fbm ] are best possible .
simon @xcite proved that a standard linear brownian motion is not monotone on any set of positive lebesgue measure .
theorem [ t : fbm0 ] for @xmath45 with hausdorff dimension in place of upper minkowski dimension is due to balka and peres @xcite .
the methods used there do not extend to minkowski dimension or to general exponents @xmath3 .
related results in the discrete setting , concerning non - decreasing subsequences of random walks , can be found in @xcite .
now we consider higher dimensional brownian motion .
let @xmath46 and @xmath23\to { \mathbb{r}}^d$ ] .
we say that @xmath47 is _ non - decreasing _ on a set @xmath9 $ ] if all the coordinate functions of @xmath48 are non - decreasing .
[ t:2dbm ] let @xmath8\}$ ] be a standard two - dimensional brownian motion .
then , almost surely , there exists a compact set @xmath18 $ ] such that @xmath16 is non - decreasing on @xmath49 and @xmath19 .
corollary [ c : fbm ] ( or ( * ? ? ?
* theorem 1.2 ) ) implies that , almost surely , the @xmath50-dimensional brownian motion @xmath16 can not be non - decreasing on any set of hausdorff dimension larger than @xmath51 . the following problem remains open in all dimensions @xmath46 .
let @xmath46 and let @xmath52 be a standard @xmath50-dimensional brownian motion .
what is the supremum of the numbers @xmath53 such that , almost surely , @xmath16 is non - decreasing on some set of hausdorff dimension @xmath53 ? finally , we prove restriction theorems for a generic @xmath3-hlder continuous function ( in the sense of baire category ) , see the following section for the details .
let @xmath54 $ ] denote the set of continuous functions @xmath23\to { \mathbb{r}}$ ] endowed with the maximum norm .
elekes ( * ? ? ?
* theorems 1.4 , 1.5 ) proved the following restriction theorem .
assume that @xmath55 .
for a generic continuous function @xmath56 $ ] ( in the sense of baire category ) for all @xmath9 $ ] 1 .
if @xmath48 is @xmath14-hlder continuous , then @xmath57 ; 2 . if @xmath48 is of bounded variation , then @xmath58 .
kahane and katznelson ( * ? ? ?
* theorems 2.1 , 3.1 ) , and independently mth ( * ? ? ?
* theorems 1.4 , 1.5 ) proved that the above result is sharp .
[ t : kkm ] let @xmath55 .
for every @xmath56 $ ] there are compact sets @xmath59 $ ] such that 1 .
@xmath60 and @xmath48 is @xmath14-hlder continuous ; 2 .
@xmath61 and @xmath62 is of bounded variation . in other words
there is always a set @xmath30 with the given properties and dimension , and for generic functions there is no @xmath30 of larger dimension .
let us recall that the _ @xmath14-variation _ of a function @xmath63 is defined as @xmath64 in the theorems above , bounded variation can be generalized to finite @xmath14-variation for all @xmath65 by similar methods .
for the following result see mth ( * ? ? ?
* theorem 5.2 ) .
[ t : m ] let @xmath65 and @xmath56 $ ] .
then there is a compact set @xmath9 $ ] such that @xmath66 and @xmath48 has finite @xmath14-variation .
our initial interest came from questions of kahane and katznelson @xcite on restrictions of hlder continuous functions .
first we need the following definition .
let @xmath67 be the set of @xmath3-hlder continuous functions @xmath63 .
for all @xmath0 and @xmath65 define @xmath68 } \sup_{a\subset [ 0,1 ] } \left\{\dim_{\mathcal{h}}a : f|_{a}\in c^{\beta}(a)\right\ } , \\ v(\alpha,\beta)&=\inf_{f\in c^{\alpha}[0,1 ] } \sup_{a\subset [ 0,1 ] } \left\{\dim_{\mathcal{h}}a : v^{\beta}(f|_{a})<\infty\right\}. \end{aligned}\ ] ] replacing hausdorff dimension by upper minkowski dimension yields @xmath69 } \sup_{a\subset [ 0,1 ] } \left\{\overline{\dim}_{\mathcal{m}}a : f|_{a}\in c^{\beta}(a)\right\ } , \\ \overline{v}(\alpha,\beta)&=\inf_{f\in c^{\alpha}[0,1 ] } \sup_{a\subset [ 0,1 ] } \left\{\overline{\dim}_{\mathcal{m}}a : v^{\beta}(f|_{a})<\infty\right\}.\end{aligned}\ ] ] as the hausdorff dimension is smaller than or equal to the upper minkowski dimension , @xmath70 and @xmath71 .
if @xmath72 then @xmath73 for all @xmath74 $ ] , so @xmath75 . for the following theorem see ( * ? ?
* theorems 5.1 , 5.2 ) . for all @xmath76 we have @xmath77 [ q : kk ]
are the above bounds optimal ?
we answer this question negatively and find the sharp bounds , which generalizes theorem [ t : va ] .
[ t : answer ] for all @xmath0 we have @xmath78 in [ s : slte ] we prove restriction theorems for functions which satisfy certain scaled local time estimates .
this allows us to prove the following more general version of theorems [ t : sa0 ] and [ t : fbm0 ] , see [ s : kk , s : fbm ] , respectively .
[ t : sa ] there is a dense set @xmath31 with the following property . for each @xmath79
there is a self - affine function @xmath80 $ ] such that for all @xmath9 $ ] 1 .
[ eq : sa1 ] if @xmath81 for some @xmath15 , then @xmath35 ; 2 .
[ eq : sa2 ] if @xmath82 for some @xmath65 , then @xmath83 .
[ t : fbm ] let @xmath0 and let @xmath8\}$ ] be a fractional brownian motion of hurst index @xmath3 .
then , almost surely , for all @xmath9 $ ] 1 . if @xmath84 for some @xmath15 , then @xmath35 ; 2 .
[ eq : fbm2 ] if @xmath85 for some @xmath65 , then @xmath86 .
the zero set of @xmath16 and the following result ( see ( * ? ? ?
* theorem 4.3 ) ) with lemma [ l : ub ] witness that theorem [ t : fbm ] is sharp for all @xmath87 .
[ t : a ] let @xmath0 and @xmath88 be fixed . then there is a compact set @xmath9 $ ] ( which depends only on @xmath3 and @xmath14 ) such that @xmath89 and if @xmath23\to { \mathbb{r}}$ ] is a function and @xmath90 such that for all @xmath91 $ ] we have latexmath:[\[\label{eq : mod } @xmath48 has finite @xmath14-variation . in section [ s : kk ]
we prove theorem [ t : answer ] by using [ t : sa ] to obtain the sharp upper bounds for @xmath93 and @xmath94 .
theorem [ t : fbm ] may be used there instead of theorem [ t : sa ] .
finally , theorems [ t : kkm ] , [ t : m ] and [ t : a ] provide the optimal lower bounds for @xmath95 and @xmath96 .
in section [ s:2d ] we consider higher dimensional brownian motion and prove theorem [ t:2dbm ] . in order to do so
, we establish a general limit theorem for random sequences with i.i.d
. increments , which is of independent interest . finally , in section [ s : generic ] we study generic @xmath3-hlder continuous functions in the sense of baire category . for @xmath0 let @xmath97 $ ] be the set of functions @xmath23\to { \mathbb{r}}$ ] such that for all @xmath91 $ ] we have @xmath98 let us endow @xmath97 $ ] with the maximum metric , then it is a complete metric space and hence we can use baire category arguments . we show that a generic @xmath99 $ ] witnesses @xmath100 and @xmath101 for all @xmath14 simultaneously .
[ t : gen ] let @xmath0 . for a generic @xmath99
$ ] for all @xmath9 $ ] 1 .
[ eq : gen1 ] if @xmath102 for some @xmath103 , then @xmath57 ; 2 .
[ eq : gen2 ] if @xmath104 for some @xmath65 , then @xmath105 .
let @xmath9 $ ] be non - empty and @xmath106 .
a function @xmath107 is called ( uniformly ) _
@xmath3-hlder continuous _ if there exists a constant @xmath108 such that @xmath109 for all @xmath110 . for
the definitions of @xmath54 $ ] , @xmath111 $ ] , and @xmath112 $ ] see section [ s : rel ] .
the diameter of @xmath30 is denoted by @xmath113 . for all @xmath114 the _ @xmath115-hausdorff
content _ of @xmath30 is @xmath116 the _ hausdorff dimension _ of @xmath30 is defined as @xmath117 let @xmath118 denote the cardinality of the set @xmath119 .
let @xmath120 be an integer .
for all @xmath121 and @xmath9 $ ] define @xmath122 the _ upper minkowski dimension _ of @xmath30 is defined as @xmath123 it is easy to show that this definition is independent of the choice of @xmath124 and we have @xmath125 for all @xmath9 $ ] . for more on these concepts see @xcite .
[ d : sa ] a compact set @xmath126 is called _ self - affine _ if for some @xmath120 there are injective and contractive affine maps @xmath127 such that @xmath128 a continuous function @xmath56 $ ] is _ self - affine _ if @xmath129 is a self - affine set .
[ d : fbm ] let @xmath0 .
the process @xmath130 is called a _ fractional brownian motion of hurst index @xmath3 _ if * @xmath16 is a gaussian process with stationary increments ; * @xmath131 and @xmath132 has standard normal distribution for every @xmath133 ; * almost surely , the function @xmath134 is continuous .
the covariance function of @xmath16 is @xmath135 .
it is well known that almost surely @xmath16 is @xmath53-hlder continuous for all @xmath136 , see [ l : ub ] below . for more information see ( * ? ? ?
* chapter 8) and ( * ? ? ?
* chapter 18 ) .
let @xmath137 be a _
complete _ metric space .
a set is _ somewhere dense _ if it is dense in a non - empty open set , otherwise it is _ nowhere dense_. we say that @xmath138 is _ meager _ if it is a countable union of nowhere dense sets , and a set is called _ co - meager _ if its complement is meager . by baire
s category theorem a set is co - meager iff it contains a dense @xmath139 set .
we say that the _ generic _ element @xmath140 has property @xmath141 if @xmath142 is co - meager .
let @xmath143,d_{{\mathcal{h}}})$ ] be the set of non - empty compact subsets of @xmath144 $ ] endowed with the _ hausdorff metric _ , that is , for each @xmath145 $ ] we have @xmath146 where @xmath147 .
then @xmath143,d_{h})$ ] is a compact metric space , see @xcite for more on this concept .
let @xmath148 stand for the support of the measure @xmath149 .
for @xmath150 let @xmath151 and @xmath152 denote the lower and upper integer part of @xmath153 , respectively .
in this section we prove restriction theorems for functions satisfying a scaled local time estimate .
first we need some notation .
let @xmath154 $ ] and an integer @xmath120 be fixed .
let @xmath121 and @xmath155 .
a _ time interval of order @xmath156 _ is of the form @xmath157 let @xmath158 .
a _ value interval of order @xmath156 _ is of the form @xmath159 for all @xmath160 define @xmath161 where @xmath162 is the set of time intervals of order @xmath156 .
clearly , @xmath163 .
[ d : an ] for a function @xmath23\to { \mathbb{r}}$ ] the _ scaled local time _
@xmath164 is the number of order @xmath156 intervals in @xmath165 in which @xmath47 takes at least one value in @xmath166 : @xmath167 it is easy to see that if @xmath47 is @xmath3-hlder continuous then for every @xmath168 , for some @xmath169 we have @xmath170 , since the function can not visit too many different value intervals in any given time interval .
finally , for each @xmath171 define @xmath172 thus the set @xmath173 includes @xmath3-hlder functions with scaled local times which are not much larger than the minimal values possible given their continuity .
we shall see below that the self - affine functions we define , as well as fractional brownian motion belong ( almost surely ) to this class .
the main goal of this section is to prove theorems [ t : holder ] and [ t : variation ] . for the notation @xmath173 and @xmath174 see definition [ d : an ] and , respectively .
[ t : holder ] let @xmath120 be an integer and let @xmath154 $ ] .
let @xmath175\to { \mathbb{r}}$ ] be a function such that @xmath176 for all large enough @xmath156 .
assume that @xmath15 and @xmath9 $ ] such that @xmath48 is @xmath14-hlder continuous
. then @xmath35 .
moreover , @xmath177 assume that @xmath9 $ ] and @xmath178 are fixed such that @xmath179 is @xmath14-hlder continuous .
choose @xmath180 such that @xmath176 for all @xmath181 .
clearly it is enough to prove . by decomposing @xmath30 into finitely many pieces of small enough diameters
, we may assume that @xmath48 is @xmath53-hlder continuous with hlder constant @xmath182 , that is , for all @xmath110 we have @xmath183 for all @xmath121 let @xmath184 let @xmath185 .
assume that @xmath186 such that @xmath187 and @xmath188 .
now we will prove that @xmath189 let us fix an arbitrary @xmath190 for some @xmath191 .
as @xmath192 , in order to show it is enough to prove that @xmath193 , therefore @xmath194 for some @xmath158 . as @xmath195
, we have @xmath196 for @xmath197 , which yields . hence follows .
fix an integer @xmath198 and let @xmath156 be an arbitrary integer such that @xmath199 . for
all @xmath200 let @xmath201 .
let @xmath202 be the minimal number such that @xmath203 .
note that @xmath204 for every @xmath205 , thus such a @xmath202 exists .
then the recursion and @xmath206 yield that @xmath207 therefore @xmath208 . applying repeatedly and using that @xmath209 and @xmath210 we obtain that @xmath211 hence follows , and the proof is complete .
the notation @xmath173 and @xmath174 are given in definition [ d : an ] and , respectively .
[ t : variation ] let @xmath120 be an integer , @xmath154 $ ] and @xmath65 .
let @xmath23\to { \mathbb{r}}$ ] be a function such that @xmath176 for all large enough @xmath156 .
assume that @xmath212 $ ] is such that @xmath48 has finite @xmath14-variation
. then @xmath213 .
moreover , @xmath214 if the theorem holds for @xmath215 , then it holds for every @xmath216 .
thus we may assume that @xmath217 , so @xmath218 .
suppose that @xmath219 $ ] such that @xmath179 has finite @xmath14-variation .
choose @xmath180 such that @xmath176 for all @xmath181 .
clearly it is enough to prove . by decomposing @xmath30 into finitely many pieces of small enough diameters
, we may assume that the @xmath14-variation of @xmath179 is at most @xmath182 , that is , @xmath220 let @xmath186 such that @xmath221 and @xmath188 .
assume that @xmath222 and @xmath223 contains @xmath224 sub - intervals in @xmath225 .
first we prove that @xmath226 let @xmath227 be the sub - intervals of @xmath223 in @xmath225 such that for every @xmath228 there is an _ even _
@xmath229 such that @xmath230 .
we may assume that @xmath231 , otherwise we switch to odd integers @xmath232 and repeat the same proof .
for all @xmath233 choose an @xmath234 such that @xmath235 .
we may assume that @xmath236 whenever @xmath237 .
let @xmath238 , and if @xmath239 is defined then let @xmath240 if the minimum exists . as @xmath195 , for all @xmath241 we have @xmath242 so if @xmath243 then @xmath244 is defined and @xmath245 .
thus the length of the defined sequence @xmath246 satisfies @xmath247 . by construction @xmath248 for all @xmath249 , which implies .
index the elements @xmath250 , and assume that each @xmath251 contains @xmath252 intervals of @xmath225 , so @xmath253 . inequality yields that @xmath254 therefore @xmath255 now assume that @xmath256 are fixed such that and @xmath257 , we prove that @xmath258 for all @xmath259 let @xmath260 applying inequality for @xmath261 and @xmath262 , and using that @xmath263 and @xmath264 imply that for every @xmath265 we have @xmath266 as @xmath267 , the above inequality implies that @xmath268 then @xmath269 , @xmath270 and imply .
finally , let @xmath156 be an arbitrary integer with @xmath271 .
let @xmath272 and @xmath273 , then @xmath274 , so @xmath275 . applying for @xmath276 easily yields that @xmath277 as @xmath278 , inequality follows .
the proof is complete .
the main goal of this section is to prove [ t : sa , t : answer ] .
first we define a family of self - affine functions @xmath279 , which will be used in section [ s : generic ] as well .
let @xmath280 be fixed integers such that @xmath281 is odd .
for every @xmath282 and @xmath283 define the one - to - one affine map @xmath284 as @xmath285 see figure [ f : maps ] . as the @xmath286 are contractions , hutchinson s contraction mapping theorem ( * ? ? ?
* page 713 ( 1 ) ) implies that there is a unique , non - empty compact set @xmath126 such that @xmath287 ( 0,0 ) rectangle ( 6,6 ) ; ( 0,0 ) rectangle ( 0.4,2 ) ; ( 0.4,2 ) rectangle ( 0.8,4 ) ; ( 0.8,4 ) rectangle ( 1.2,6 ) ; ( 0,0 ) ( 0.35,1.75 ) ; ( 0.4,2 ) ( 0.75,3.75 ) ; ( 0.8,4 ) ( 1.15,5.75 ) ; ( 1.2,4 ) rectangle ( 1.6,6 ) ; ( 1.6,2 ) rectangle ( 2,4 ) ; ( 2,0 ) rectangle ( 2.4,2 ) ; ( 1.2,6 ) ( 1.55,4.25 ) ; ( 1.6,4 ) ( 1.95,2.25 ) ; ( 2,2 ) ( 2.35,0.25 ) ; ( 2.4,0 ) rectangle ( 2.8,2 ) ; ( 2.8,2 ) rectangle ( 3.2,4 ) ; ( 3.2,4 ) rectangle ( 3.6,6 ) ; ( 2.4,0 )
( 2.75,1.75 ) ; ( 2.8,2 ) ( 3.15,3.75 ) ; ( 3.2,4 )
( 3.55,5.75 ) ; ( 3.6,4 ) rectangle ( 4,6 ) ; ( 4,2 ) rectangle ( 4.4,4 ) ; ( 4.4,0 ) rectangle ( 4.8,2 ) ; ( 3.6,6 ) ( 3.95,4.25 ) ; ( 4,4 ) ( 4.35,2.25 ) ; ( 4.4,2 )
( 4.75,0.25 ) ; ( 4.8,0 ) rectangle ( 5.2,2 ) ; ( 5.2,2 ) rectangle ( 5.6,4 ) ; ( 5.6,4 ) rectangle ( 6,6 ) ; ( 4.8,0 ) ( 5.15,1.75 ) ; ( 5.2,2 ) ( 5.55,3.75 ) ; ( 5.6,4 ) ( 5.95,5.75 ) ; ( 0,0 ) rectangle ( 6,6 ) ; ( 0,0 ) ( 0.08,2 ) ; ( 0.08,2 ) ( 0.16,0 ) ; ( 0.16,0 ) ( 0.24,2 ) ; ( 0.24,2 ) ( 0.32,0 ) ; ( 0.32,0 ) ( 0.4,2 ) ; ( 0.4,2 ) ( 0.48,4 ) ; ( 0.48,4 ) ( 0.56,2 ) ; ( 0.56,2 ) ( 0.64,4 ) ; ( 0.64,4 ) ( 0.72,2 ) ; ( 0.72,2 ) ( 0.8,4 ) ; ( 0.8,4 ) ( 0.88,6 ) ; ( 0.88,6 ) ( 0.96,4 ) ; ( 0.96,4 ) ( 1.04,6 ) ; ( 1.04,6 ) ( 1.12,4 ) ; ( 1.12,4 ) ( 1.2,6 ) ; ( 2.4,0 ) ( 2.48,2 ) ; ( 2.48,2 ) ( 2.56,0 ) ; ( 2.56,0 ) ( 2.64,2 ) ; ( 2.64,2 ) ( 2.72,0 ) ; ( 2.72,0 ) ( 2.8,2 ) ; ( 2.8,2 ) ( 2.88,4 ) ; ( 2.88,4 ) ( 2.96,2 ) ; ( 2.96,2 ) ( 3.04,4 ) ; ( 3.04,4 ) ( 3.12,2 ) ; ( 3.12,2 ) ( 3.2,4 ) ; ( 3.2,4 ) ( 3.28,6 ) ; ( 3.28,6 ) ( 3.36,4 ) ; ( 3.36,4 ) ( 3.44,6 ) ; ( 3.44,6 ) ( 3.52,4 ) ; ( 3.52,4 ) ( 3.6,6 ) ; ( 4.8,0 ) ( 4.88,2 ) ; ( 4.88,2 ) ( 4.96,0 ) ; ( 4.96,0 ) ( 5.04,2 ) ; ( 5.04,2 ) ( 5.12,0 ) ; ( 5.12,0 ) ( 5.2,2 ) ; ( 5.2,2 ) ( 5.28,4 ) ; ( 5.28,4 ) ( 5.36,2 ) ; ( 5.36,2 ) ( 5.44,4 ) ; ( 5.44,4 ) ( 5.52,2 ) ; ( 5.52,2 ) ( 5.6,4 ) ; ( 5.6,4 ) ( 5.68,6 ) ; ( 5.68,6 ) ( 5.76,4 ) ; ( 5.76,4 ) ( 5.84,6 ) ; ( 5.84,6 ) ( 5.92,4 ) ; ( 5.92,4 ) ( 6,6 ) ; ( 1.2,6 ) ( 1.28,4 ) ; ( 1.28,4 ) ( 1.36,6 ) ; ( 1.36,6 ) ( 1.44,4 ) ; ( 1.44,4 ) ( 1.52,6 ) ; ( 1.52,6 ) ( 1.6,4 ) ; ( 1.6,4 ) ( 1.68,2 ) ; ( 1.68,2 ) ( 1.76,4 ) ; ( 1.76,4 ) (
1.84,2 ) ; ( 1.84,2 )
( 1.92,4 ) ; ( 1.92,4 )
( 2,2 ) ; ( 2,2 ) ( 2.08,0 ) ; ( 2.08,0 )
( 2.16,2 ) ; ( 2.16,2 )
( 2.24,0 ) ; ( 2.24,0 )
( 2.32,2 ) ; ( 2.32,2 )
( 2.4,0 ) ; ( 3.6,6 ) ( 3.68,4 ) ; ( 3.68,4 )
( 3.76,6 ) ; ( 3.76,6 )
( 3.84,4 ) ; ( 3.84,4 )
( 3.92,6 ) ; ( 3.92,6 )
( 4,4 ) ; ( 4,4 )
( 4.08,2 ) ; ( 4.08,2 )
( 4.16,4 ) ; ( 4.16,4 )
( 4.24,2 ) ; ( 4.24,2 ) ( 4.32,4 ) ; ( 4.32,4 )
( 4.4,2 ) ; ( 4.4,2 )
( 4.48,0 ) ; ( 4.48,0 )
( 4.56,2 ) ; ( 4.56,2 )
( 4.64,0 ) ; ( 4.64,0 )
( 4.72,2 ) ; ( 4.72,2 )
( 4.8,0 ) ; it is easy to see that @xmath288 is a graph of a function @xmath289\to [ 0,1]$ ] , and @xmath279 can be approximated as follows .
let @xmath290\}$ ] be the diagonal of @xmath144 ^ 2 $ ] and define @xmath291 as @xmath292 if @xmath293 denotes the @xmath294 iterate @xmath295 , then @xmath296 is the graph of a piecewise linear function @xmath297 which converges uniformly to @xmath279 as @xmath298 .
clearly @xmath279 is a self - affine function with @xmath299 and @xmath300 , and the definition yields that @xmath279 is hlder continuous with exponent @xmath301 .
figure [ f : new ] shows the piecewise linear function @xmath302 , which approximates @xmath303 .
define the set @xmath304 then @xmath305 is a countable dense subset of @xmath306 , since every rational @xmath307 is in @xmath305 by @xmath308 and @xmath309 . for all @xmath32 fix @xmath276 such that @xmath310 , and define @xmath311 $ ] .
fix @xmath312 such that @xmath313 .
we use the scaled local times with @xmath314 . by [ t : holder , t : variation ] it is enough to show that @xmath315 for all @xmath316 .
let us fix @xmath316 .
clearly @xmath317 , and the construction of @xmath318 yields that for every @xmath319 , @xmath320 , and @xmath321 we have @xmath322 similarly , @xmath323 the above and @xmath324 yield that for all @xmath325 we have @xmath326 thus @xmath315 , and the proof is complete .
note that the bound we get for @xmath327 does not use the @xmath328 factor .
it is possible to go through the proof [ t : holder ] with such a stronger assumption , which would slightly improve the bounds on @xmath174 with @xmath329 in place of @xmath330 error term .
let @xmath0 . by [ t : kkm , t : m , t : a ]
it is enough to prove that @xmath331 for the first inequality let @xmath332 be arbitrary , then @xmath333 $ ] .
suppose that @xmath334 is @xmath14-hlder continuous on some set @xmath9 $ ] .
theorem [ t : sa ] yields that @xmath335 , thus @xmath336 .
since @xmath305 is dense in @xmath337 , we obtain that @xmath338 .
for the second inequality define @xmath339 and let @xmath340 be arbitrary . then @xmath333 $ ] .
assume that @xmath334 has finite @xmath14-variation on some @xmath30 .
theorem [ t : sa ] and @xmath341 imply that @xmath342 , so @xmath343 .
as @xmath305 is dense in @xmath344 , we have @xmath345 .
let @xmath0 be fixed and let @xmath346\to { \mathbb{r}}$ ] be a fractional brownian motion of hurst index @xmath3 .
we think of @xmath16 as a random function from @xmath144 $ ] to @xmath347 . the main goal of this section is to prove theorem [ t : fbm ] . by theorems [ t : holder ] and
[ t : variation ] it is enough to prove the following proposition . for the notation @xmath348 see definition [ d : an ] .
[ p : key ] let @xmath0 and let @xmath346\to { \mathbb{r}}$ ] be a fractional brownian motion of hurst index @xmath3 .
then , almost surely , @xmath349 for all @xmath156 large enough .
first we define a discrete ( truncated ) scaled local time . for all @xmath121 let @xmath350 for all @xmath160 let @xmath351 clearly , @xmath352 . for a function @xmath23\to { \mathbb{r}}$ ]
the _ discrete scaled local time _
@xmath353 is the number of points in @xmath354 which are mapped to @xmath166 by @xmath47 : @xmath355 for every @xmath171 define @xmath356 first we need to prove some lemmas . to avoid technical difficulties we assume that the domain of @xmath16 is extended to @xmath357 when necessary .
let @xmath358 be the probability space on which our fractional brownian motion is defined , and let @xmath359 be the natural filtration . if @xmath360 $ ] is a stopping time then define the @xmath361-algebra @xmath362 for all stopping times @xmath363 and integers @xmath160 and @xmath169 let @xmath364 [ l : sln ] there is a finite constant @xmath365 depending only on @xmath3 such that for every bounded stopping time @xmath363 and integers @xmath160 and @xmath169 we have , almost surely , @xmath366 pitt ( * ? ? ?
* lemma 7.1 ) showed that the property of _ strong local nondeterminism _ holds for fractional brownian motion , that is , there is a constant @xmath367 such that for all @xmath368 , almost surely , @xmath369 let us fix @xmath133 .
as @xmath16 is gaussian , almost surely the conditional distribution @xmath370 is normal , and implies that its density function is bounded by @xmath371 .
therefore , almost surely , @xmath372 where @xmath373 . applying for
finitely many @xmath374 implies that @xmath375 where @xmath376 .
the proof is complete .
[ l : snmpq ] there is a finite constant @xmath377 depending only on @xmath3 such that for all @xmath378 and @xmath379 we have @xmath380 let @xmath381 be the constant in lemma [ l : sln ] , clearly we may assume that @xmath382 .
we will show that @xmath383 satisfies the lemma .
we define stopping times @xmath384 .
let @xmath385 .
if @xmath386 is defined for some @xmath387 then let @xmath388 be the first time such that the contribution of @xmath16 to @xmath389 on @xmath390\cap { \mathcal{l}}_{n , m , p}$ ] is at least @xmath391 if such a time exists , otherwise let @xmath392 .
then @xmath382 and the definition of stopping times yield that @xmath393 note that we may assume that @xmath394 and hence the above conditional probabilities are defined , otherwise we are done immediately .
therefore it is enough to prove that @xmath395 for all @xmath396 .
the definition of @xmath397 , lemma [ l : sln ] , and the conditional markov s inequality imply that , almost surely , @xmath398 therefore the tower property of conditional expectation yields that @xmath399 which completes the proof .
the following lemma is a discrete version of proposition [ p : key ] .
[ l : dkey ] almost surely , @xmath400 for all large enough @xmath156 .
we give an upper bound for @xmath401 by applying lemma [ l : snmpq ] with @xmath402 to each of the relevant @xmath403 .
since @xmath160 , @xmath404 and @xmath405 , there are at most @xmath406 possibilities to choose @xmath403 .
therefore lemma [ l : snmpq ] implies that @xmath407 therefore @xmath408 , so the borel
cantelli lemma implies that @xmath409 the following lemma is well known , see the more general ( * ? ? ?
* corollary 7.2.3 ) .
[ l : ub ] almost surely , we have @xmath410 now we are ready to prove proposition [ p : key ] . by lemmas [ l : dkey ] and [ l : ub ] we can choose a random @xmath180 such that , almost surely , for every @xmath181 we have the following properties : 1 . [ i1 ] @xmath411 } |b(t)|<n-1
$ ] ; 2 .
[ i2 ] @xmath400 ; 3 .
[ i3 ] @xmath412 for all @xmath413 ; 4 .
[ i4 ] @xmath414 .
fix a path of @xmath16 for which the above properties hold .
let us fix an arbitrary @xmath181 , it is enough to prove that @xmath415 .
let @xmath160 , @xmath404 and @xmath158 be given , we need to show that @xmath416 property [ i1 ] yields that if @xmath417 with @xmath418 then @xmath419
. therefore [ i2 ] implies that for all @xmath417 we have @xmath420 let @xmath421 be a time interval of order @xmath156 such that @xmath422 and @xmath423 , then [ i3 ] yields that @xmath424 finally , , and [ i4 ] imply that @xmath425 hence holds , and the proof of proposition [ p : key ] is complete .
we include here a proof of proposition [ p : r ] , which we could not find in the literature .
recall that @xmath8\}$ ] is a fractional brownian motion of hurst index @xmath3 .
the lower bound is quite elementary , while the upper bound relies on a first moment computation and on a result of molchan .
he studied the distribution of the maximal value of fractional brownian motion on @xmath426 $ ] and of the time @xmath427 when it is achieved .
specifically , molchan ( * ? ? ?
* theorem 2 ) proved the following .
@xmath428 as @xmath429 . with probability
one @xmath16 is @xmath53-hlder continuous for every @xmath136 and maps @xmath44 to the non - degenerate interval @xmath430\}]$ ] , thus @xmath431 .
therefore it is enough to prove that @xmath432 almost surely .
using that @xmath427 and @xmath433 have the same distribution and @xmath434\}$ ] is also a fractional brownian motion , for all @xmath435 we obtain @xmath436 \neq\emptyset)&={\mathbb{p}}(\tau_{\max}(t ) \geq t-{\varepsilon})={\mathbb{p}}(\tau_{\max}(1)\geq 1-{\varepsilon}/t ) \\ & = { \mathbb{p}}(\tau_{\max}(1)\leq { \varepsilon}/t)=({\varepsilon}/t)^{1-\alpha+o(1)},\end{aligned}\ ] ] with the @xmath437 term tending to @xmath438 as @xmath439 .
let @xmath440 be the number of intervals @xmath441 $ ] which intersect @xmath44 .
let @xmath442 be arbitrary and fix @xmath443 such that @xmath444 \neq\emptyset ) \leq ( { \varepsilon}/t)^{1-\alpha-\delta}$ ] whenever @xmath445 . using this above for @xmath446 and the trivial bound @xmath182 for @xmath447 , for every large enough @xmath281 we obtain @xmath448 where @xmath449 is a finite constant depending only on @xmath450 . by markov s inequality @xmath451 for any @xmath281 large enough . applying this for @xmath452 yields that @xmath453 thus the borel
cantelli lemma implies that , almost surely , @xmath454 for all large enough @xmath156 .
therefore @xmath455 . as @xmath442 was arbitrary , @xmath456 almost surely .
the aim of this section is to prove theorem [ t:2dbm ] .
the idea is to find ( in a greedy manner ) large sets along which a simple random walk in @xmath457 is monotone .
since the scaled simple random walk converges to brownian motion , this gives sets along which @xmath16 is monotone . to control the dimension of the limit sets we estimate the energy of the discrete sets and apply a version of frostman s lemma to bound the dimension .
given a simple random walk @xmath458 , define the _ greedy increasing subset _ by @xmath459 our first task is to prove tightness for the number and structure of record times in @xmath460 . since our argument may apply in similar situations , we state some of our arguments in the more general context of sums of i.i.d.variables with power law tails . before focusing on the case of random walks , we prove theorem [ t : acc ] , a limit theorem , which will allow us to transfer estimates from the random walk setting to brownian motion . as theorem [ t : acc ]
below is a quite general result about random sequences with i.i.d .
increments , we hope that it will find further applications . fix @xmath0 .
let @xmath461 be some random variable , and assume that there are @xmath462 such that for all @xmath171 we have @xmath463 let @xmath464 be an i.i.d .
sequence with the law of @xmath363 .
define @xmath465 the number of steps before reaching @xmath156 is denoted by @xmath466 .
the following lemma is fairly standard .
[ l : m_n_tight ] there are constants @xmath467 such that for all @xmath468 we have 1 .
[ cl1 ] @xmath469 .
[ cl2 ] @xmath470 .
[ cl3 ] @xmath471 , and more generally , for all integers @xmath472 we have @xmath473 claim [ cl1 ] is given by ( * ? ? ?
* lemma 4.2 ) . inequality and @xmath474 imply that @xmath475 so [ cl2 ] holds .
the first bound of [ cl3 ] follow easily from [ cl2 ] .
the general bound holds since the first @xmath476 with @xmath477 ( if there is such @xmath476 ) is a stopping time .
applying [ cl2 ] to the sequence starting at time @xmath476 completes the proof .
let @xmath149 be a non - atomic _ mass distribution _ on a metric space @xmath478 , ( that is , a borel measure on @xmath479 with @xmath480 ) . for @xmath481
, define the _
@xmath53-energy of @xmath149 _ by @xmath482 for the following theorem see ( * ? ? ?
* theorem 4.27 ) .
[ t : energy ] let @xmath149 be a non - atomic mass distribution on a metric space @xmath479 with @xmath483
. then @xmath484 .
consider the set @xmath485 , endowed with lebesgue measure @xmath486 .
we next estimate the @xmath53-energy of @xmath486 .
[ l : energy_small ] let @xmath487 .
there is a finite constant @xmath488 , such that for all @xmath156 we have @xmath489 the argument is to consider the contribution to the energy from pairs @xmath490 with distance at various scales , and the largest scale will dominate the rest
. up to a factor of @xmath491 we may restrict the integral to @xmath492 .
we split the integral on @xmath493 into several parts .
note that @xmath494 is a disjoint union of unit intervals .
let @xmath495 be the contribution to @xmath496 from pairs @xmath497 and @xmath498 where @xmath499 and @xmath500 .
the number of such pairs @xmath501 is at most @xmath502 , so lemma [ l : m_n_tight ] [ cl3 ] yields that @xmath503 for @xmath504 let @xmath505 be the contribution to @xmath496 from pairs @xmath497 and @xmath498 where @xmath499 and @xmath506 .
let @xmath507 denote the number of such pairs @xmath501 .
for such @xmath501 the contribution from @xmath497 and @xmath498 to @xmath505 is at most @xmath508 where we used @xmath509 . thus @xmath510 .
lemma [ l : m_n_tight ] [ cl3 ] yields that for every @xmath511 , conditioned on @xmath512 , the expected number of @xmath513 in @xmath514 $ ] is at most @xmath515 .
lemma [ l : m_n_tight ] [ cl3 ] also implies that the expected number of @xmath516 in @xmath517 equals @xmath518 .
therefore @xmath519 and we obtain that @xmath520 this partition gives the identity @xmath521 . as @xmath522
whenever @xmath523 , we have @xmath524 . with the bounds above
, the largest @xmath202 dominates the sum and we arrive at the inequality @xmath525 .
we will wish to work with rescaled sets .
for all @xmath171 let @xmath526 define the measure @xmath527 , that is , @xmath528 times the lebesgue measure restricted to @xmath529 .
[ l : mu_tight ] let @xmath487 and @xmath530 .
then there exist @xmath531 depending on @xmath532 so that with probability at least @xmath533 we have @xmath534 ) \leq n \quad \text { and } \quad { \mathcal{e}}_\gamma(c_n,\mu_n ) \leq n_{\gamma}.\ ] ] let @xmath535 $ ] and let @xmath180 be arbitrary
. for all @xmath171 the following three inequalities hold .
markov s inequality and @xmath536 yield that @xmath537 lemma [ l : m_n_tight ] [ cl3 ] and [ cl1 ] yield that @xmath538 lemma [ l : energy_small ] yields that @xmath539 . indeed , @xmath540 and since @xmath541 on @xmath529 ,
changing to @xmath542 gives a further factor of @xmath543 .
now markov s inequality implies that @xmath544 the above three inequalities with large enough @xmath545 complete the proof .
theorem [ t : acc ] is concerned with a sequence of sequences of i.i.d .
variables @xmath546 satisfying .
that is , for each fixed @xmath156 the variables @xmath546 are i.i.d . ,
but there could be arbitrary dependencies between variables with different numbers @xmath156 .
the superscript @xmath156 is also the parameter used for scaling sums of the @xmath156th sequence .
thus we denote @xmath547 for all @xmath171 define @xmath548 [ t : acc ] with the notations above , almost surely , @xmath549 has an accumulation point @xmath49 in the hausdorff metric such that @xmath550 .
let @xmath542 be @xmath528 times the lebesgue measure on @xmath529 .
fix @xmath530 .
for some finite constants @xmath551 for all @xmath156 define the event @xmath552 ) \leq
n_0 \text { and } { \mathcal{e}}_{\gamma}(c_n,\mu_n ) \leq n_{\gamma } \textrm { for all } 0<\gamma<\alpha\}.\ ] ] since the map @xmath553 is non - decreasing , applying lemma [ l : mu_tight ] for a sequence of parameters @xmath554 with @xmath555 implies that there are constants @xmath556 such that @xmath557 for all @xmath156 .
let @xmath558 , then @xmath559 .
since @xmath530 was arbitrary , it is enough to prove that the theorem is satisfied whenever @xmath560 holds .
assume that @xmath560 holds , then there is a random subsequence @xmath561 such that the events @xmath562 hold for all @xmath200 .
since @xmath563)\leq n_0 $ ] for all @xmath511 , by passing to a subsequence we may assume that @xmath564 weakly , where @xmath149 is a measure on @xmath144 $ ] satisfying @xmath565)\leq n_0 $ ] .
similarly , we may assume that @xmath566 in the hausdorff metric for some compact set @xmath18 $ ] .
as @xmath567 $ ] is close in the hausdorff metric to @xmath568 , we get @xmath569 . for all @xmath487 we obtain @xmath570 for the first inequality see e.g. ( * ? ? ?
* lemma 2.2 ) .
theorem [ t : energy ] now implies that @xmath571 , and the proof is complete .
we now apply theorem [ t : acc ] to random walks on @xmath457 and thus prove theorem [ t:2dbm ] . for each @xmath156 , let @xmath572 be a simple random walk on @xmath457 , and define the rescaled random walks by @xmath573 .
it is well known that it is possible to construct the walks @xmath572 and two - dimensional brownian motion @xmath8\}$ ] on the same probability space so that @xmath574 converges uniformly to @xmath16 on the interval @xmath144 $ ] , see e.g. ( * ? ? ? * theorem 3.5.1 ) or @xcite .
we henceforth assume such a coupling . recall that for each walk we construct the _ greedy increasing subset _ by @xmath575 for every @xmath156 , this sequence has i.i.d .
increments with the law of @xmath576 we use the notation @xmath577 if @xmath578 as @xmath579 .
we need the following known estimate .
[ t : greedy ] let @xmath458 be a two - dimensional simple random walk .
let @xmath363 be the hitting time of the positive quadrant : @xmath580 .
then there is a @xmath90 so that @xmath581 for the above theorem see the general result of denisov and wachtel ( * ? ? ?
* theorem 1 ) , or a bit weaker one due to varopoulos ( * ? ? ?
* ( 0.3.3 ) and ( 0.4.1 ) ) . on exit times of planar
brownian motion from cones see evans ( * ? ? ?
* corollary 5(i ) ) or the somewhat weaker ( * ? ? ? * lemma 10.40 ) . in the brownian case
the exponent @xmath21 can be calculated by mapping the complement of @xmath582 onto a half plane by the conformal map @xmath583 and using the conformal invariance of planar brownian motion .
the continuous case can be transformed to the discrete one by coupling .
for the history of similar estimates and for further references see denisov and wachtel @xcite .
recall that @xmath572 are two - dimensional simple random walks so that the rescaled walks @xmath574 converge uniformly to a brownian motion @xmath16 .
let @xmath363 be the hitting time of the positive quadrant by @xmath584 , that is , @xmath585 .
for every @xmath586 , the greedy increasing subsequence of @xmath572 has i.i.d .
increments , distributed as @xmath363 . by theorem [ t : greedy ] we have @xmath587 with some @xmath90 .
thus we can apply theorem [ t : acc ] with @xmath588 .
this yields that , almost surely , there is an accumulation point @xmath49 of @xmath549 in the hausdorff metric such that @xmath19 . as @xmath589 uniformly , @xmath16 is non - decreasing on @xmath49 .
this completes the proof . for a higher dimensional simple random walk
@xmath590 define the hitting time @xmath591 then @xmath592 for some @xmath593 , see ( * ? ? ?
* theorem 1 ) .
our argument proves an analogue of theorem [ t:2dbm ] in higher dimensions with this @xmath3 instead of @xmath21 .
the goal of this section is to prove theorem [ t : gen ] .
first we need some preparation .
the following lemma is probably well known .
however , we could not find an explicit reference for its second claim , so we outline the proof .
[ l : ext ] let @xmath594 and @xmath595 .
assume that @xmath596 and @xmath63 is a function such that for all @xmath110 we have @xmath597 then @xmath47 extends to @xmath598 satisfying the above inequality for all @xmath599 .
if @xmath30 is closed then @xmath119 can be chosen to be linear on the components of @xmath600 .
as @xmath47 admits a unique continuous extension to the closure of @xmath30 which clearly satisfies , we may assume that @xmath30 is closed .
let @xmath223 be any component of @xmath600 , it is enough to prove that @xmath47 extends to @xmath601 such that holds . if @xmath602 or @xmath603 for some @xmath604 then @xmath605 works .
now let @xmath606 for some @xmath607 and let @xmath119 be the linear extension of @xmath47 to @xmath223 .
the concavity of the function @xmath608 implies that @xmath609 for all @xmath610 , the straightforward calculation is left to the reader .
let @xmath617 such that the oscillation of @xmath47 on @xmath618 $ ] is at most @xmath619 for all @xmath620 .
let @xmath621 be the piecewise linear function passing through the points @xmath622 , then clearly @xmath623 .
applying lemma [ l : ext ] @xmath476-times we obtain that @xmath624 $ ] .
we can choose @xmath625 such that @xmath626 satisfies @xmath627 .
hence for all @xmath91 $ ] we have @xmath628 let @xmath629 , then it is easy to see that every horizontal line segment of the graph of @xmath630 ( if there are any ) can be replaced by two line segments of nonzero slopes such that the resulting function @xmath613 satisfies @xmath631 and for all @xmath91 $ ] we have @xmath616 clearly @xmath615 , and the proof is complete .
now we are ready to prove the first part of theorem [ t : gen ] .
the concept of the proof is similar to that of ( * ? ? ?
* theorem 1.4 ) , but the technical details are much more difficult and in order to create appropriate hlder continuous functions some new ideas are needed as well .
let @xmath632 be arbitrarily fixed , and define @xmath633 : \dim_{\mathcal{h}}\{f = g\}\leq 1-\beta \textrm { for all } g\in c_1^{\beta}[0,1]\},\ ] ] where we use the notation @xmath634 : f(x)=g(x)\}$ ] .
first we show that it is enough to prove that @xmath635 is co - meager in @xmath97 $ ] .
indeed , since co - meager sets are closed under countable intersection , this implies that for a countable dense set @xmath636 the set @xmath637 is co - meager in @xmath97 $ ] .
now assume that @xmath638 and @xmath9 $ ] such that @xmath48 is @xmath14-hlder for some @xmath15 , we need to prove that @xmath57 .
choose a sequence @xmath639 such that @xmath640 and fix an @xmath171 .
as @xmath48 is @xmath14-hlder , there is an @xmath612 such that for all @xmath641 with @xmath642 the function @xmath643 is @xmath644-hlder with hlder constant @xmath182 .
let @xmath645 such that @xmath646 for all @xmath647 .
then @xmath648 are @xmath644-hlder with hlder constant @xmath182 , so by lemma [ l : ext ] there are functions @xmath649 $ ] such that @xmath650 for all @xmath511 .
therefore @xmath651 implies that @xmath652 for all @xmath511 , thus the countable stability of hausdorff dimension yields @xmath653 .
this holds for all @xmath171 , so @xmath57 .
now let @xmath632 be fixed , and for all @xmath654 define @xmath655 : { \mathcal{h}}_{\infty}^{1-\beta+1/n}(\{f = g\})\leq 1/m \textrm { for all } g\in c_1^{\beta}[0,1 ] \right\}.\ ] ] clearly @xmath656 , thus it is enough to show that each @xmath657 contains a dense open subset of @xmath97 $ ] .
assume that @xmath658 , @xmath659 , @xmath625 , and a piecewise linear function @xmath660 $ ] with nonzero slopes are given such that for all @xmath91 $ ] we have @xmath661 by lemma [ l : dense ] it is enough to find a function @xmath99 $ ] and @xmath662 such that @xmath663 where @xmath664 denotes the closed ball in @xmath97 $ ] centered at @xmath47 with radius @xmath224 .
now we define @xmath47 .
we can fix integers @xmath665 such that @xmath666 is odd and @xmath667 let @xmath668 and let @xmath669 $ ] be the self - affine function defined in section [ s : kk ] .
we will approximate @xmath670 by re - scaled copies of @xmath671 . as @xmath671 is @xmath53-hlder continuous
, there exists @xmath672 such that for all @xmath91 $ ] we have @xmath673 assume that @xmath674 such that @xmath670 is linear on each interval @xmath675 $ ] with nonzero slopes .
let @xmath676 for all @xmath677 , then @xmath678 for all @xmath679 .
let us define @xmath680 and @xmath681 such that @xmath682 for all @xmath620 and @xmath683 let @xmath684 now we are ready to define @xmath47 .
if for some @xmath620 , @xmath685 , and @xmath686 we have @xmath687 note that the linearity of @xmath670 implies that if @xmath153 satisfies , then @xmath688 now we prove that @xmath689 \quad \textrm{and } \quad f\in b(f_0,r_0/2).\ ] ] as the range of @xmath670 is @xmath144 $ ] , the definition of @xmath47 and @xmath690 imply that for all @xmath691 $ ] we have @xmath692 thus it is enough to prove for that @xmath99 $ ] .
assume that @xmath91 $ ] and @xmath492 , we need to prove that @xmath693 .
we consider three cases . _
first case _
: suppose that @xmath694 . then , @xmath660 $ ] and the definition of @xmath690 imply that @xmath695 _ second case _
: assume that @xmath490 are _ adjacent points _
, that is , @xmath696 $ ] for some @xmath697 and @xmath698 .
then the definitions of @xmath47 and @xmath699 , inequality , the definition of @xmath690 , and @xmath700 yield that @xmath701 _ third case _ :
suppose that there exists @xmath702 such that @xmath703 and @xmath704 are adjacent points .
the triangle inequality , the above inequality , the definition of @xmath690 , and @xmath700 imply that @xmath705 by the definition of @xmath706 for all @xmath490 at least one of the above three cases holds , which concludes the proof of .
finally , we prove that holds for some @xmath662 .
let us define @xmath442 as @xmath707 since @xmath708 , we have @xmath709 .
thus by @xmath710 we can fix an @xmath711 such that for all @xmath200 and @xmath712 we have @xmath713 since @xmath714 , we can define @xmath715 let us fix an integer @xmath716 such that @xmath717 let us define @xmath662 as @xmath718 then clearly @xmath719 .
let us fix @xmath720 , it is enough to show that @xmath721 .
fix @xmath620 and @xmath685 arbitrarily , and let @xmath722 $ ] .
let @xmath723 $ ] , by the subadditivity of @xmath724 it is enough to show that @xmath725 assume that @xmath121 and @xmath726 is a closed interval .
we divide @xmath223 into @xmath727 non - overlapping closed intervals of equal length , the resulting intervals are called the _ elementary intervals of @xmath223 of level @xmath156_. assume that @xmath728 and let @xmath729 be an elementary interval of @xmath730 of level @xmath731 .
now we show that @xmath732 intersects at most @xmath733 many first level elementary intervals of @xmath729 .
let us decompose @xmath729 into @xmath666 non - overlapping intervals of equal length , let @xmath734 be one of them .
let @xmath735 be two nonconsecutive first level elementary intervals of @xmath729 , it is enough to show that @xmath732 can not intersect both @xmath736 and @xmath737 .
assume to the contrary that there are @xmath738 and @xmath739 such that @xmath740 and @xmath741 .
if there are @xmath742 first level elementary intervals of @xmath729 between @xmath736 and @xmath737 , then @xmath743 therefore @xmath723 $ ] yields that @xmath744 on the other hand side , the definition of @xmath613 , @xmath47 and @xmath224 imply that @xmath745 the above inequalities and yield that @xmath746 , which is a contradiction
. therefore @xmath747 intersects at most @xmath748 many elementary intervals of @xmath730 of level @xmath749 .
since the length of these intervals is less than @xmath750 , the definition of @xmath361 and inequality yield that @xmath751 hence holds , and the proof is complete . in order to prove the second part of theorem [ t : gen ] we need a bridge between the notions of @xmath3-hlder continuity and @xmath14-variation .
the following lemma is ( * ? ? ?
* lemma 4.1 ) , see also ( * ? ? ?
* lemma 4.1 ) .
clearly it is enough to prove the theorem for a countable dense set of parameters @xmath14 .
since co - meager sets are closed under countable intersection , it is enough to show the statement for an arbitrary fixed @xmath65 . for all @xmath99 $ ] let @xmath757
$ ] be given such that @xmath758 .
fix an arbitrary @xmath759 and let @xmath760 .
it is enough to prove that @xmath761 for a generic @xmath99 $ ] . applying lemma [ l : bv ] we obtain that for all @xmath99 $ ] there are sets @xmath762 such that as @xmath765 and @xmath763 are @xmath53-hlder continuous , theorem [ t : gen ] and the definition of @xmath53
imply that for a generic @xmath99 $ ] for all @xmath171 we have @xmath766 inequalities , , and the countable stability of hausdorff dimension yield that @xmath761 for a generic @xmath99 $ ] .
the proof is complete . | for @xmath0 let @xmath1 denote the supremum of the numbers @xmath2 such that every @xmath3-hlder continuous function is of bounded variation on a set of hausdorff dimension @xmath2 .
kahane and katznelson ( 2009 ) proved the estimate @xmath4 and asked whether the upper bound is sharp .
we show that in fact @xmath5 .
let @xmath6 and @xmath7 denote the hausdorff and upper minkowski dimension , respectively .
the upper bound on @xmath1 is a consequence of the following theorem .
let @xmath8\}$ ] be a fractional brownian motion of hurst index @xmath3 .
then , almost surely , there exists no set @xmath9 $ ] such that @xmath10 and @xmath11 is of bounded variation .
furthermore , almost surely , there exists no set @xmath9 $ ] such that @xmath12 and @xmath13 is @xmath14-hlder continuous for some @xmath15 .
the zero set and the set of record times of @xmath16 witness that the above theorems give the optimal dimensions .
we also prove similar restriction theorems for deterministic self - affine functions and generic @xmath3-hlder continuous functions .
finally , let @xmath17\}$ ] be a two - dimensional brownian motion .
we prove that , almost surely , there is a compact set @xmath18 $ ] such that @xmath19 and @xmath20 is non - decreasing in each coordinate .
it remains open whether @xmath21 is best possible . |
this paper is about the use of the quantum action@xcite in physics .
let us start out by discussing some possible areas of application .
* high energy physics * + _ renormalisation_. the quantum action may turn out to provide a new definition of renormalisation . in conventional terminology , renormalisation means to extract physical observable parameters like mass , coupling etc . obtained from an interacting quantum field theory at the continuum limit .
those physical parameters differ from the so - called bare parameters and this difference represents the effects of the interaction . very similar to this , the quantum action has parameters ( mass , potential parameters ) different from the classical action and this difference represents the effects of quantum mechanics : q.m .
fluctuations occuring in the path integral are represented by a single path , however , for a particle with different properties of mass and interaction . in order to explore the use of the quantum action as means of renormalisation
, one could construct the quantum action for a many - body system ( like a chain of coupled anharmonic oscillators ) and explore its quantum mechanical continuum limit . in quantum field theory ,
one carries out renormalisation by computing n - point vertex functions , which represent vacuum - to - vacuum transition amplitudes for transition time @xmath0 . in the same time
limit , the existence of the quantum action has been proven rigorously in q.m .
, although so far only in the case of a single particle system .
+ _ cosmology and inflationary scenario of early universe_. inflation involves potentials with several minima and instanton solutions .
the instanton starts out as a quantum instanton and eventually turns into a classical instanton .
this has effects on the subsequent formation of galaxies@xcite .
using the effective potential or the quantum potential ( potential of the quantum action ) in general creates a potential different from the classical one .
in particular , it may have minima being absent in the classical potential .
consequently , this may create instantons being quite different from the classical instanton . in ref.@xcite instantons from the quantum action
were found to be `` softer '' than the corresponding classical instanton .
such quantum effects of the instanton may influence the outcome of the galaxy formation at the end of inflation .
+ _ hot and dense nuclear matter_. instantons are believed to play an important role in hot nuclear matter in the quark - gluon plasma phase@xcite .
instantons are important also for the mechanism of chiral symmetry breaking and for thoofts solution of the u(1 ) problem .
+ _ neutrino oscillations_. the process of oscillations of neutrino flavors@xcite may have to do with a process of tunneling in a potential with degenerate minima . *
condensed matter physics * + _ quantum dots , semi - conductor and quantum chaos_. advancing the speed of microprocessors may have technological obstacles but also physical limits . when reducing the size of a chip one
soon may enter the regime where quantum laws rule .
quantum chaos may become a very important issue , because it can hamper the flow of electric currents .
the quantitative determination of quantum chaos effects will possibly be of great importance for the development of future microprocessors . by use of some kind of effective action or the quantum action , one can study quantum corrals formed by atoms and quantum dots in semi - conductors .
in particular , this allows to study the temperature dependence of electron dynamics in atomic corrals , as well as for electrons moving in simple conductor - semiconductor - isolator geometries and to search for the possible presence of quantum chaos .
+ _ josephson junctions and superconducting quantum interference_. superconducting quantum interference devices ( squid ) have been used to demonstrate experimentally the phenomenon of quantum superposition in macroscopic states@xcite .
this involves josephson junctions .
the squid potential has a double - well structure .
the effective potential and the quantum potential should be useful tools to analyze quantum superposition in terms of such potential involving quantum effects .
+ _ quantum computers based on superconductors_. the symmetry of the order parameter in some triplet superconductors corrersponds to doubly degenerate chiral states .
gulian et al.@xcite predict that this degeneracy can be lifted via macroscopic quantum tunneling .
instanton - like quantum behavior may become important .
triplet superconducters may be used as basic elements of quantum computers . again the effective potential and the quantum potential should help to study quantum instantons and tunneling in such materials . *
atomic physics * + _ analogy of classical chaos in quantum physics_. attempts have been made to use the effective action in order to characterize chaotic behavior in quantum systems@xcite . in a a similar way
, the quantum action has been used also@xcite .
this allows to construct of a phase space portrait and poincar sections for a quantum system in analogy to classical physics . from this one
can obtain lyapunov exponents and kam surfaces for the quantum system .
a potentially quite interesting system to explore is the paul trap .
+ _ ultracold atoms in a billard formed by lasers_. trace formulas ( gutzwiller@xcite and generalisations ) have been used successfully@xcite to establish a relation between level densities and periodic semi - classical orbits .
it would be interesting to compare predictions of trace formulas in the semi - classical regime with the predictions obtained from the effective action or the quantum action .
+ _ dynamical tunneling_. steck et al.@xcite and hensinger et al.@xcite have demonstrated experimentally the phenomenon of dynamical tunneling ( where the classical transition is forbidden due to some conserved quantity different from energy ) .
it has been realized by arrays of cold atoms .
it has been observed that the presence of quantum chaos enhances the dynamical tunneling transition .
it would be instructive to reexamine dynamical tunneling using the phase space portrait constructed from a time - dependent effective action or quantum action .
* chemistry * + _ binding of macromolecules_. in the process of chemical binding of macromolecules , often a double well potential plays a role .
the effective action as well as the quantum action should be useful to find pathways in the formation of such macromolecules .
this paper is focussed on the quantum action@xcite .
we discuss some of its analytical properties .
we present results of numerical computations testing the goodness of the quantum action in fitting transition amplitudes . as a physical application we address the topic of quantum chaos , occuring in nuclear physics and atomic physics , from the viewpoint of the quantum action .
the quantum action parametrizes quantum transition amplitudes @xmath1 in terms of a new action @xmath2 defined by @xmath3/\hbar ] \nonumber \\ & & \tilde{s}[x ] = \int_{0}^{t } dt \frac{1}{2 } \tilde{m } \dot{x}^{2 } - \tilde{v}[x ] ~ , \end{aligned}\ ] ] where @xmath2 is a local action - the so called quantum action and @xmath4 is the classical trajectory of such action going from boundary point @xmath5 to @xmath6 .
for some potentials there may be multiple trajectories connecting those boundary points , then there is a sum over all trajectories .
@xmath7 is a normalisation factor .
one should note that the existence of such quantum action in general has not been proven yet . as an operational definition of the quantum action
, one may consider it as that action which gives the optimal fit of the transition amplitudes @xmath1 for all combinations of boundary points , for a fixed transition time . often such optimal fit using only a single trajectory is already very good , as can be seen in examples below . in the limit of large ( imaginary ) transition time
the quantum action has been proven rigorously to exist , giving an exact parametrisation of transition amplitudes@xcite .
the quantum action being local allows to construct a portrait of phase space of a quantum system by applying the tools of nonlinear dynamics to this action@xcite . in particular
, it allows to construct poincar sections and lyapunov exponents , in analogy to classical nonlinear dynamics ( note that the definition of lyapunov exponents requires the limit @xmath0 ) . , @xmath8 and @xmath9 ( full line ) , quantum potential ( dashed line ) and ground state wave function ( enhanced by factor 5 , dotted line ) . ]
consider the schrdinger equation in imaginary time ( @xmath10 ) with a given time - independent hamiltonian @xmath11 .
then the transition amplitude evolves according to @xmath12 < \psi_{n}|x_{in } > \nonumber \\ & = & \sim_{t \to \infty } ~ < x_{fi}|\psi_{gr } > \exp[-e_{gr}t/\hbar ] < \psi_{gr}|x_{in } > ~ .\end{aligned}\ ] ] in the large time limit ( feynman - kac limit ) the transition amplitude is projected onto the ground state contribution
. this may give the impression that for large time most of the physical information is lost .
however , this is not true .
the stationary schrdinger equation allows to reconstruct the potential @xmath13 mass from the ground state @xmath14 the mass can be reconstructed from the groundstate plus any one ( e.g. lowest ) of the excited states , @xmath15 in other words , all information on the hamiltonian is stored in the wave functions and energies of any two eigen states .
this is to some extent analogous to biology , where the genetic code is stored in all cells . similar to eq.([eq : psiclasspot ] ) there is a relation between the ground state and the quantum potential , proven in ref.@xcite , @xmath16 where @xmath17 denotes the minimum value of the quantum potential . in @xmath18 dimension
, this allows to express the wave function in terms of the quantum potential and quantum mass @xmath19 ~ .\ ] ] this expression holds in the case when the classical potential is parity symmetric , has a single minimum at @xmath20 and the ground state wave function a single maximum at @xmath20 .
the formula can be generalized to multi - valley potentials .
now we will establish a differential equation relating the classical potential to the quantum potential . defining @xmath21 the combination of eqs.([eq : psiclasspot],[eq : psiquantpot],[eq : defu ] ) yields
@xmath22 this is a riccati - type differential equation to determine the function @xmath23 when @xmath24 is given .
in @xmath18 it can be expressed more explicitely by @xmath25 ( bottom ) , @xmath26 ( center ) and @xmath27 ( top ) as function of transition time @xmath28 .
classical parameters @xmath29 , @xmath30 and @xmath31 . ]
in @xmath18 ideas from supersymmetry have been used to improve the computation of tunneling amplitudes@xcite . for a given potential
, there is a supersymmetric partner potential , which yields the same spectrum as the original potential , except for the ground state .
consider a hamiltonian of the form ( taking @xmath32 ) , @xmath33 there is a supersymmetric partner potential @xmath34 , given by @xmath35 defining @xmath36 yields the relation @xmath37 and obeys the riccati differential equation @xmath38 eqs.([eq : w2],[eq : riccati1d ] ) are identical to eq.([eq : riccati ] ) in @xmath18 . as a lesson
, one finds that the formalism of the supersymmetric partner potential has a close correspondence in @xmath18 to the quantum action .
the concept of the supersymmetric partner potential and the quantum action , however , differ in two respects : first the supersymmetric partner potential has no physical interpretation in contrast to the quantum action , which gives a representation of q.m .
transition amplitudes .
second , the concept of the supersymmetric partner potential is restricted to @xmath18 dimension@xcite , in contrast to the quantum action which is valid also in @xmath39 dimensions .
it would be tempting to apply the quantum action to the tunneling problem in @xmath39 dimensions .
some work is in progress . .
we consider the following classical potential in 1-d @xmath40 the corresponding classical action is given by @xmath41 .
the potential is parity symmetric .
because it has an infinite barrier ( for @xmath42 ) at the origin , the system at @xmath43 is separated from the system at @xmath44 .
we consider only the motion in the domain @xmath44 .
we have chosen to consider the inverse square potential , because of the distinct feature that the corresponding quantum mechanical transition amplitudes are known analytically@xcite .
the classical potential compared to the quantum potential as well as the ground state wave function is shown in fig.[[fig : potwave ] ] .
the behavior of the parameters of the quantum action as a function of transition time @xmath28 are shown in fig.[[fig : actpar ] ] .
the goodness of the fit of transition amplitudes by the quantum action , using only one trajectory is shown in fig.[[fig : e44_s ] ] .
it shows the global relative error of the quantum action versus transition time .
one observes a maximal error in the order of @xmath45 for @xmath46 .
when @xmath28 approaches zero , the error goes to zero .
it becomes quite small also for @xmath0 , consistent with the analytic behavior of the quantum action .
moreover , the figures shows that the error decreases , if the quantum potential allows for the presence of terms like @xmath47 and @xmath48 , which are absent in the classical potential .
fig.[[fig : estand_mv-2 ] ] shows the behavior of the parameter combination @xmath49 versus transition time , computed by doing an optimisation of the quantum action parameters to globally fit the transition amplitudes .
fig.[[fig : eqdiff_mv-2 ] ] shows the same quantity , computed in an alternative way from the flow equation , discussed in ref.@xcite .
one observes that both methods yield asymptotically stable results which agree among each other and their asymptotic values agree also with the analytical value @xmath50 , predicted by eq.([eq : transformlaw ] ) . obtained from global fit method vs. transition time @xmath28 .
] obtained from flow equation vs. transition time @xmath28 .
dependence on location of interval @xmath51 $ ] of final boundary points . ]
quantum chaos has been experimentally observed in irregular energy spectra of nuclei , and also in atoms exposed to strong electromagnetic fields@xcite .
classically chaotic billiard systems have a quantum counterpart being atoms moving in a 2-d billiard , where bounday reflections are caused by controlled interaction with laser beams@xcite .
irregular patterns have been predicted by computer simulations and were found experimentally in wave functions of quantum billards@xcite , where scars are reminders of classical motion@xcite .
recently , dynamical tunneling in atoms between regular islands has been observed .
the transition is enhanced by chaos@xcite .
spectra of fully chaotic quantum systems can statistically be described by random matrices@xcite , which corresponds to a level density distribution of wigner - type , while integrable ( nonchaotic ) quantum system yield a poissonian distribution .
here we ask : what happens between these two extremes ? for example , an hydrogen atom exposed to a weak exterior magnetic field shows a level distribution , which is neither poissonian nor wignerian .
can we compare classical chaos with quantum chaos ? is the quantum system more or less chaotic than the corresponding classical system ? moreover , an understanding of how classically mixed phase space is reflected in quantum systems is an open problem .
semiclassical methods of quantisation ( ekb , gutzwiller s trace formula ) are not amenable to mixed dynamical systems@xcite . here
we want to address those questions using the concept of the quantum action@xcite .
we have considered the following @xmath52 hamilton system , which is classically chaotic , @xmath53 in the quantum action we made an ansatz of the following form for the quantum potential , @xmath54 the parameters of the quantum action have been determined by a global fit to transition amplitudes .
the terms @xmath55 , @xmath56 , @xmath57 , @xmath58 , @xmath59 , were found to be quite small or compatible with zero .
a comparison of poincare sections of the classical action and the quantum action shows that there are slightly more and larger regular islands in the quantum case than in the classical case .
the ratio of chaotic phase space is displayed in fig.[[fig : ratio ] ] .
it shows that overall the quantum system is less chaotic .
+ h.k . and k.j.m.m .
are grateful for support by nserc canada . | we consider a mixed chaotic hamiltonian system and compare classical with quantum chaos . as alternative to the methods of energy level spacing statistics and trace formulas
, we construct a quantum action and a quantum analogue phase space to analyse quantum chaos . |
since early 1994 , when the borocarbide family ( , where r = rare earth ) was discovered@xcite a series of interesting results have been reported on these materials .
the most remarkable feature is the coexistence of superconductivity and antiferromagnetism for r = tm , er , ho and dy . the high superconducting transition temperatures @xmath4 and the broad variation of the ratio @xmath5 ( where @xmath6 is the neel temperature , ranging from @xmath7 k to @xmath8 k )
make this family particularly appropriate to explore that coexistence@xcite .
interesting properties are also found in the non - magnetic borocarbides , r = y , lu .
these materials exhibit a non - exponential temperature dependence@xcite and a non - linear field dependence@xcite of the electronic specific heat in the superconducting phase ; an anomalous upper critical field @xmath9 with an upward curvature near @xmath4@xcite and a four - fold oscillation when the field is rotated within the basal plane@xcite ; and a square flux line lattice ( fll ) at high fields@xcite . although these properties were initially taken as evidence of a non conventional pairing mechanism@xcite , both the unconventional @xmath9 and the square fll
can also be satisfactorily explained by assuming a non - local electrodynamics .
traditionally , nonlocality was expected to be relevant only in materials with a ginzburg - landau parameter @xmath10 and sufficiently clean to have an electronic mean free path @xmath11 .
however , the large vortex cores make theoretical analysis very difficult in that case .
the borocarbides , in contrast , have intermediate @xmath12 values ( @xmath13 ) that make core effects much smaller .
this allows the non - local effects to be introduced perturbatively either in the london model , as done by kogan , gurevich et al.@xcite , or in the ginzburg - landau theory , as shown by de wilde et al.@xcite and park and huse@xcite . in both cases ,
the result is a coupling of the supercurrents to the underlying crystal symmetry .
the first success of the kogan - gurevich nonlocal model was the explanation@xcite of the deviations of the reversible magnetization @xmath14 of bi:2212 in the mixed state from the logarithmic dependence on magnetic field , @xmath15 , predicted by the london model@xcite .
later on , song et al.@xcite found similar deviations in yni@xmath0b@xmath0c when @xmath16-axis , that could also be quantitatively accounted for by the model .
recently we have extended that study to all directions of @xmath17 and showed that this generalization of the london theory provides a satisfactory complete description of the anisotropic @xmath18 with a self - consistent set of parameters@xcite . in the local london model anisotropy
is introduced via a second rank mass tensor @xmath19 . in tetragonal materials such as borocarbides @xmath20 ,
thus the properties should be isotropic in the basal plane
. however , non - local corrections introduce@xcite a fourfold anisotropy as a function of the magnetic field orientation within the @xmath21 plane , reflecting the square symmetry .
this @xmath22 periodicity was indeed observed in @xmath14 in the mixed state of both yni@xmath0b@xmath0c@xcite and luni@xmath0b@xmath0c@xcite .
kogan s model also predicted@xcite that two structural transitions in the fll should occur in borocarbides for @xmath16 , a first order reorientation transition between two rhombic lattices at a field @xmath23 and a second order transition from rhombic to square at @xmath24
. small angle neutron scattering ( sans ) studies@xcite confirmed those predictions in yni@xmath0b@xmath0c .
a jump in the apical angle @xmath25 of the rhombic lattice , discontinuous within the resolution , occurs at @xmath26 to @xmath27 , and the lattice becomes square ( @xmath28 ) at @xmath29 to @xmath30 . according to a recent analysis by knigavko et al.@xcite ,
@xmath23 really consists of two second order transitions taking place in a very narrow field range .
although the role of nonlocality on the @xmath31 properties of the fll is by now convincingly established , less is known about its effects on the @xmath32 vortex response . since nonmagnetic borocarbides exhibit a very low critical current @xmath33 for @xmath16 , pinning correlation volumes were expected to be large , as indeed observed@xcite .
thus , the elastic properties of the fll must play a key role in the pinning . as the shear modulus @xmath34 depends on @xmath25@xcite , and @xmath25
undergoes a discontinuous jump at @xmath23 , it is bound to happen that @xmath34 and therefore the pinning properties change abruptly at this field@xcite . in other words , vortex pinning , which involves distortions from equilibrium vortex configurations ,
should be affected by the symmetry changes in the vortex lattice .
recently we showed@xcite that , in yni@xmath0b@xmath0c and for @xmath35 , the reorientation transition at @xmath23 induces a kink in the pinning force density @xmath36 .
we also observed anisotropies in @xmath1 both between the c - axis and the basal plane ( out - of - plane anisotropy ) and within the plane ( in - plane anisotropy ) .
we found that @xmath1 for @xmath2 is one order of magnitude larger than for @xmath35 and has a quite different field dependence .
we argued that this surprising behavior is unlikely to arise either from pinning by magnetic impurities or from non - local effects , although we recognized that the evidence supporting those claims was not conclusive .
here we report further studies of @xmath1 in non - magnetic borocarbides .
we show that the kink in @xmath36 is also visible in luni@xmath0b@xmath0c and coincides with the field @xmath23 for this compound , thus confirming that it is a signature of this nonlocality - induced transition .
we find that @xmath37 slightly decreases as @xmath38 increases , in contrast to @xmath39 .
we also study the effect of co - doping in lu(ni@xmath40co@xmath41)@xmath0b@xmath0c .
we observe that @xmath23 decreases as the nonlocal effects are progressively suppressed by increasing @xmath42 , in agreement with the @xmath38 dependence .
all the luni@xmath43co@xmath41b@xmath0c samples exhibit an enormous out - of - plane anisotropy .
this unambiguously demonstrates that this anisotropy is due neither to the magnetic impurities ( as those crystals have a density of impurities much smaller than the yni@xmath0b@xmath0c crystal ) , nor to nonlocality ( as it does not dissapear with increasing @xmath42 ) .
we also rule out the precence of surface barriers for @xmath44 by performing minor hysterisis loops .
we report magnetization measurements , performed with a quantum design squid magnetometer , on yni@xmath0b@xmath0c ( y-0 ) , and lu(ni@xmath40co@xmath41)@xmath0b@xmath0c with @xmath45 ( lu-0 ) , @xmath46 ( lu-1.5 ) and @xmath47 ( lu-3 ) single crystals .
the dimensions , @xmath4 and estimated @xmath48 for each sample are summarized in table i. the y-0 crystal is the same one that was previously investigated in refs.@xcite .
the normal state magnetization follows a curie law that indicates the presence of a very dilute distribution of localized magnetic moments .
it corresponds to a rare - earth impurity content of @xmath49 at .
@xmath50 relative to y , probably due to contaminants in the y starting material@xcite .
the lu(ni@xmath40co@xmath41)@xmath0b@xmath0c crystals , grown as described elsewhere@xcite , show a much weaker curie tail at low temperatures ( @xmath51 ) which might arise , for example , from a @xmath52 magnetic impurities of gd in the lu site@xcite .
isothermal magnetization loops in the superconducting mixed state were measured , and the critical current density @xmath33 was then calculated using the bean s critical state model@xcite .
figure [ fpvsh](a ) shows the pinning force density @xmath53 for the sample as a function of the applied field for @xmath54-axis at several @xmath38 .
we observe that at low fields @xmath36 decreases strongly with increasing @xmath3 , but above a field @xmath55 the field dependence becomes much weaker . in a recent work@xcite we showed that this `` kink '' in @xmath36 at @xmath56 is a signature of the reorientation phase transition in the fll .
we based this claim in several facts .
first , the position of the kink for @xmath54 coincides with the value of @xmath23 .
second , @xmath56 is rather insensitive to the field orientation , in agreement with the behavior of @xmath23 observed in sans experiments@xcite .
finally , @xmath57 is an increasing function of @xmath38 , as predicted by knigavko et al.@xcite . [
cols="^,>,>,>,^",options="header " , ] we now expand the previous study and analyze the behavior of the field @xmath56 when the nonlocality is reduced . in the kogan - gurevich description@xcite the strength of the nonlocal perturbations is parametrized by a new characteristic distance , the nonlocality radius @xmath58 : the weaker the nonlocality effects , the smaller @xmath58 .
this means that @xmath58 decreases with increasing @xmath38 or decreasing @xmath48 .
thus , if this model is correct , an increase in @xmath38 or a decrease in @xmath48 should produce qualitatively similar effects on the nonlocality - induced transitions @xmath23 and @xmath59 .
-axis for ( a ) y-0 sample at several temperatures and ( b ) lu-0 , lu-1.5 and lu-3 samples at @xmath60.,width=340 ] a typical way to reduce @xmath48 is to introduce impurities .
it has been shown that the increase of @xmath42 in lu(ni@xmath40co@xmath41)@xmath0b@xmath0c crystals decreases @xmath48 without significantly increasing @xmath33@xcite .
recent results confirmed that the analogy between @xmath38 and @xmath48 is indeed valid in this material in the case of @xmath59 .
first , gammel et al.@xcite showed that @xmath59 increases as @xmath48 is reduced by increasing @xmath42 .
later on , eskildsen et al.@xcite showed that @xmath39 also rises as @xmath38 increases . if @xmath56 is indeed a signature of @xmath23 , we should also find a correlation between its @xmath38 and @xmath48 dependencies . in figure [ phasediagram ]
we plot @xmath61 , together with @xmath62 .
we observe that @xmath56 is almost constant at low @xmath38 and , unlike the rhombic to square second order transition @xmath39 , it slightly decreases with @xmath38 at higher temperatures .
.,width=340 ] to determine the @xmath48 dependence of @xmath56 we performed measurements on the lu-0 , lu-1.5 and lu-3 samples at @xmath63 for @xmath54 ( estimates of @xmath48 are given in table i ) .
the results are shown in figure [ fpvsh](b ) , where we plotted @xmath36 .
we observe that the kink in @xmath1 ( indicated by the arrows ) is still visible in the field range where the transition @xmath23 should appear , and that it shifts to lower fields with increasing @xmath42 , in agreement with the @xmath38 dependence .
in other words , the analogy between @xmath42 and @xmath38 is satisfied for @xmath56 , thus confirming that the kink arises from nonlocality . another
interesting fact is that the lu-0 sample has a larger @xmath1 than the y-0 sample at the same @xmath38 , even though it has a lower density of magnetic impurities .
this indicates that the magnetic moments in the y-0 are not the relevant pins for the flux lines when @xmath35 .
we now turn to the pinning properties for @xmath2 .
figure [ outofplane](a ) shows @xmath1 at @xmath64 for the y-0 sample for @xmath65 $ ] , as a function of the reduced field @xmath66 . the @xmath67 $ ] data , already shown in fig .
[ fpvsh](a ) , is included for comparison . it is evident that the behavior for @xmath65 $ ] is very different from that observed for @xmath67 $ ] .
first we note that a broad maximum develops at intermediate fields @xmath68 .
second , @xmath1 is much larger than for @xmath16 over most of the field range .
the maximum of this out - of - plane anisotropy , @xmath69/f_p[001]$ ] , is @xmath70 as shown in figure [ outofplane](b ) . at higher @xmath38 the overall behavior and
the @xmath71 values are similar .
finally , for @xmath72 $ ] ( not shown ) , we found that @xmath1 is slightly but systematically smaller ( @xmath73 ) than for @xmath74 $ ] . and
( c ) lu-1.5 at @xmath63 for @xmath35 and @xmath44 .
( b ) and ( d ) show the out of plane anisotropy @xmath71 corresponding to ( a ) and ( c ) respectively.,width=453 ] hereafter we will focus on the origin of the difference in @xmath1 between the @xmath75-axis and the @xmath21-plane . as we pointed out in ref.@xcite , the large out - of - plane
@xmath1 anisotropy sharply contrasts with the very small ( @xmath76 ) mass anisotropy@xcite .
therefore , explanations based on the anisotropic scaling frequently used in high @xmath4 superconductors@xcite are ruled out .
as we mentioned above , the y-0 crystal contains a small amount of magnetic impurities@xcite , which align preferentially along the ab - plane and thus are a potential source of the @xmath1 anisotropy .
however , we had previously argued@xcite that this was unlikely .
indeed , since the alignment of these localized moments increases with @xmath3 , pinning should become more directional as the field increases and thus @xmath71 should increase monotonically with @xmath3 . this is not in agreement with the data in fig.[outofplane](b ) , where @xmath71 first grows with @xmath77 , maximizes at @xmath78 and decreases again .
in particular , @xmath1 is almost isotropic at the peak effect ( see dotted line which corresponds to @xmath79 ) .
thus , although we were not able to totally rule out magnetic pinning on crystal y-0 , the data suggested that this was not the case .
conclusive evidence that the localized magnetic moments are not responsible for the large @xmath71 comes from the persistence of this effect in the samples lu-0 , lu-1.5 and lu-3 , which have a magnetic impurity content two order of magnitude lower than the y-0 crystal . in figure
[ outofplane](c ) we show @xmath80 for the lu-1.5 for @xmath16 and @xmath2 , at @xmath63 , and in figure .
[ outofplane](d ) we plotted the corresponding @xmath81 .
we observe that the out - of - plane anisotropy is even larger than in the y-0 sample .
in contrast to the behavior observed in the y-0 , @xmath71 in the lu-1.5 increases monotonically with @xmath77 up to near the peak effect region , where it suddenly drops approaching to the isotropic limit .
measurements on the lu-3 crystal at several @xmath38 show a similar behavior and exhibit a @xmath71 larger than in the y-0 .
we had also previously argued@xcite that @xmath71 in the y-0 sample seems too large to be ascribed to nonlocality , which should appear as a perturbatively small effect .
this conclusion is unambiguously confirmed by the presence of even larger @xmath71 values in the doped samples lu-1.5 and lu-3 , where nonlocal effects are strongly suppressed .
the existence of significant surface barriers for @xmath2 , have also been ruled out by performing minor hysteresis loops with @xmath82 at several @xmath38 and @xmath3 .
examples for the y-0 and lu-1.5 samples are shown in figure [ minorloops ] .
if hysteresis were due to surface barriers no flux changes would occur in the bulk while @xmath3 is changing from one branch of the main loop to the other one , hence the data of the minor loop connecting the lower and upper branches would be meissner - like straight lines@xcite .
in contrast , in the case of bulk pinning , the lines connecting both branches are curved ( parabolic in the simplest bean model for an infinite slab ) just as we observe in the insets of figure [ minorloops](a ) and ( b ) . at @xmath63 for ( a ) y-0 sample and ( b ) lu-1.5 sample . the insets show a blow up of the minor hysterisis loops corresponding to the full loop showed in the main panel .
the dotted lines show the behavior expected for surface barriers.,width=340 ] moreover , the @xmath1 calculated from these minor loops assuming only bulk pinning are in good agreement with those obtained from the main loops ( fig .
[ outofplane ] ) . indeed , according to the isotropic bean model
the critical state profile can be completely inverted if we change the external field in @xmath83 , where @xmath84 , and @xmath85 and @xmath86 are the shortest and largest sample dimensions perpendicular to the field direction , respectively . from the inset of figure [ minorloops](a ) we find that the width of the hysterisis loop is @xmath87 .
for this sample @xmath88 , thus it is expected that the connecting line between the two branches of the loop , reaches the upper branch , when the field is decreased about @xmath89 .
inspection of figure [ minorloops](a ) shows that @xmath90 , in reasonably good agreement given the simplicity of the model .
the same analysis on the data of the inset of figure [ minorloops](b ) , leds to a similar conclusion .
non - magnetic borocarbides provide a very attractive ground to study the limit of very low pinning forces .
the high quality of the crystals have the double benefit of producing large mean free paths , resulting in measurable non - local effects , and a very dilute distribution of vortex pinning centers .
the combination of both factors allows the observation of the influence of nonlocality on vortex pinning .
this is a rather unexplored field , and further studies , both experimental and theoretical , are needed . on the other hand ,
the origin of the large out - of - plane anisotropy in @xmath1 is unclear and deserves further investigation .
we have conclusively ruled out explanations based on the mass anisotropy , pinning by magnetic impurities , non - local effects and surface barriers . a simple explanation for it could be the presence of some still unidentified anisotropic pinning centers , such as planar defects .
work partially supported by anpcyt , argentina , pict 97 no.01120 , and conicet pip 4207 .
a.v.s . would like to thank to conicet for financial support .
research sponsored by the u.s .
department of energy under contract de - ac05 - 00or22725 with the oak ridge national laboratory , managed by ut - battelle , llc .
cava , h. takagi , h. w. zandbergen , j. j. krajewski , w. f. peck , jr .
, t. siegrist , b. batlogg , r. b. van dover , r. j. felder , k. mizuhashi , j. o. lee , h. eisaki , and s. uchida , _ nature ( london ) _ * 367 * , 252 ( 1994 ) .
cava , h. takagi , b. batlogg , h. w. zandbergen , j. j. krajewski , w. f. peck , jr . , r. b. van dover , r. j. felder , t. siegrist , k. mizuhashi , j. o. lee , h. eisaki , s. a. carter , and s. uchida , _ nature ( london ) _ * 367 * , 146 ( 1994 ) . m. r. eskildsen ; k. harada ; p. l. gammel ; a. b. abrahamsen ; n. h. andersen ; g. ernst ; a. p. ramirez ; d. j. bishop , k. mortensen , d. g. naugle , k. d. d. rathnayaka , and p. c. canfield , _ nature _ * 393 * , 242 ( 1998 )
. m.r .
eskildsen , p.l .
gammel , b.p .
barber , u. yaron , a.p .
ramirez , d.a .
huse , d.j . bishop , c. bolle , c.m .
lieber , s. oxx , s. sridhar , n.h .
andersen , k. mortensen , and p. c. canfield , _ phys .
lett . _ * 78 * , 1968 ( 1997 ) . | we present an overview of the temperature , field and angular dependence of the irreversible magnetization of non - magnetic borocarbides ( y;lu)ni@xmath0b@xmath0c . we show that nonlocal electrodynamics influences pinning via the unusual behavior of the shear modulus in non - hexagonal lattices .
on top of that , we observe that the pinning force density @xmath1 exhibits a rich anisotropic behavior that sharply contrasts with its small mass anisotropy . when @xmath2 , @xmath1 is much larger and
has a quite different @xmath3 dependence , indicating that other pinning mechanisms are present . |
nova - like ( nls ) stars are non - eruptive cataclysmic variables ( cvs , see * ? ? ?
* ) , classified into several subtypes depending on evidence of a strong magnetic field on the white dwarf ( wd ) , and on spectral and photometric characteristics ( see * ? ? ?
we will focus on the vy scl - type nova - likes or ` anti - dwarf ' novae characterized by the presence of occasional dips on the light curve , so - called low states , defined by @xcite as a fading of the optical light by more than 1.5 mag in less than 150 days .
the drop in luminosity can reach 7 mag and may last from weeks to years .
the large optical and uv luminosity seems to imply that the vy scl - type nova - like , in their longer lasting optically high state , are undergoing mass transfer onto the wd at the high rate @xmath0 m@xmath1 yr@xmath2 , sustaining an accretion disk in a stable hot state in which dn outbursts are suppressed ( see e.g. the disk thermal instability model of * ? ? ?
the low states have been attributed to a sudden drop of @xmath3 from the secondary , or even to a total cessation of mass transfer ( see e.g. * ? ? ?
* ; * ? ? ?
the reason of this dramatic decrease of @xmath3 is still unclear .
the most probable cause may be spots on the surface of the secondary , covering the l1 point and causing the mass - transfer cut off @xcite .
@xcite have suggested instead non - equilibrium effects in the irradiated atmosphere of the donor . if the transition from the high to low state occurs because of a decrease in @xmath3 , the accretion disks in these systems should move from the equilibrium region to the one of dwarf novae ( dn ) instabilities , so we should observe ( dn ) outbursts with recurrence times of 12 to 20 days during the low state , caused by thermal viscous instabilities in the accretion disk @xcite .
however , outbursts in the low state of these objects are extremely rare ( see * ? ? ? * for mv lyr ) .
dn outbursts must be suppressed during the low state despite the low @xmath4 ; this can be explained by a wd effective temperature high enough ( 30,00050,000 k ) to irradiate the inner accretion disk and maintain it in the stable ` hot ' state @xcite , while the incoming mass accretion stream stops or decreases @xcite .
the wd in the known dn never reaches this temperature range , but high wd effective temperatures have indeed been inferred via spectroscopic observations in the _ uv _ and _ fuv _ ranges for the vy scl objects ( see table 1 ) .
@xcite suggested instead that the dn outbursts are prevented by the periodic disruption of the disc by a magnetic field of the wd , and in this model the vy scl would be intermediate polars ( ip ) , in which the wd magnetic field is of the order of 10@xmath5 gauss .
@xcite proposed a link between the vy scl - type stars and super soft x - ray sources ( sss ) based on a _ rosat _ observations of v751 cyg .
they found an anti - correlation in the optical and x - ray intensity , and despite the very poor spectral resolution of the _ rosat _ hri , the spectrum appeared to be soft in the low state .
the authors suggested that quasi - stable thermonuclear burning occurs on the surface of the wd in the low state , preventing dn outbursts . in this framework , vy scl stars are key objects in the evolution of interacting wd binaries , in which hydrogen burning occurs periodically without outbursts .
accretion goes on at a very high rate without ever triggering a thermonuclear runaway , because of the recurrent interruptions of the high @xmath3 regime .
thus there is a possibility that vy scl stars reach the chandrasekhar mass and the conditions for type ia supernovae outbursts . @xcite and @xcite suggested also thermonuclear burning in the low states of v sge and bz cam .
however , these two objects had not been actually observed as sss ; in more recent years an x - ray observation of the vy scl system v504 cyg in the low state failed to reveal a luminous sss @xcite . using archival x - ray observations , in this paper we compare high and low state x - ray data , and some new _ uv _ data , for four vy scl - type stars .
we seek clues to the complex evolution of these systems , and explanations for the changes that take place during the transition from the high to low state .
in table 1 we report parameters from the published results of observations in the optical , near ( _ nuv _ ) and far ( _ fuv _ ) ultraviolet wavelength ranges .
we see that these objects have an orbital period just above the period gap , in a narrow range between 3.2 and 3.7 hours . for the three systems mv lyr ,
tt ari and v794 aql the wd effective temperature t@xmath6 was estimated in previous low states ( not shown in fig .
1 ) from _ uv _ and _ fuv _ observations , in the range between 39,000 k and 47,000 k. these systems could not have been sss at the time of those observations , because t@xmath6 places the flux peak in the _ fuv _ range . on the other hand , we can not rule out ignition of thermonuclear burning , neither the possibility that the wd may become hotter with time in subsequent low states .
the low state t@xmath6 of mv lyr tabulated in tab .
1 was measured while the _ fuv _ flux , corresponding to most of the bolometric one , was about 3.6 @xmath7 erg s@xmath2 ( see table 1 ) . in fig .
5 of @xcite we see that at the tabulated distance these values may be consistent with thermonuclear burning with @xmath8 m@xmath9 yr@xmath2 and a wd mass less than 1 m@xmath9 .
@xcite suggested that mv lyr becomes hotter in the high state , hypothesizing a lower limit t@xmath10 k , consistent with a measured _ fuv _ flux of 2.5 @xmath11 erg s@xmath2 in the high state .
if the wd really becomes hotter while emitting x - rays at this level , the possibility of thermonuclear burning would be even more likely in the high state . for a wd mass of 0.7 m@xmath9 and the value of @xmath4 in table 1 , assuming that the extreme _ uv _ ( _ euv _ ) luminosity is close to the total ( bolometric ) luminosity , nuclear burning according to @xcite occurs with t@xmath12 k ( see their fig .
5 ) , which implies a peak luminosity in the _ euv _ range and no detectable sss .
an accurate determination of t@xmath6 in the high state is important : even an upper limit inferred from the absence of an sss in the x - rays is useful to constrain the evolutionary models . = 0.1 cm l|cccc & bz cam & mv lyr & tt ari & v794 aql + dist(pc ) & @xmath13}$ ] & @xmath14}$ ] & @xmath15}$ ] & @xmath16}$ ] + p@xmath17 & @xmath18}$ ] & @xmath19}$ ] & @xmath20}$ ] & @xmath21}$ ] + @xmath22 & 1240@xmath23}$ ] & 1013@xmath24}$ ] & 1722@xmath25}$ ] & @xmath26}$ ] + m@xmath27(m@xmath1 ) & & @xmath28}$ ] & 0.571.2@xmath29}$ ] & @xmath30}$ ] + @xmath31(m@xmath32 yr@xmath2 ) & & 23@xmath33@xmath34 @xmath35,[6]}$ ] & @xmath36@xmath37@xmath25}$ ] & 10@xmath3810@xmath39@xmath40@xmath41}$ ] + @xmath42(m@xmath32 yr@xmath2 ) & & 3@xmath43@xmath37@xmath44}$ ] & 10@xmath45@xmath46@xmath47@xmath48}$ ] & + t@xmath49 & & @xmath50}$ ] & & + t@xmath51 & & @xmath52}$ ] & @xmath53}$ ] & @xmath54}$ ] + t@xmath55 & & @xmath56}$ ] & & + _ fuv _
@xmath58 s@xmath2 ) & & 1.4@xmath59@xmath35}$]@xmath60 & @xmath61@xmath62}$]@xmath63 & + _ fuv _ flux@xmath64 ( erg @xmath58 s@xmath2 ) & & 9.4@xmath65@xmath44}$]@xmath63 & & + + + [ tab : par ]
we examined the archival x - ray data of vy scl - type stars obtained with _ swift _ , _ rosat _ , _ suzaku _ and _ chandra _ and chose the objects that were observed both in the high and low states : bz cam , mv lyr , tt ari and v794 aql .
the data are summarized in table 2 .
all the data were never published except for a _ rosat _ observations of tt ari , which we examined again and which were also analysed by @xcite and @xcite . in order to assess
when the low and high optical states occurred , we relied on the data of the variable star network ( vsnet ) collaboration @xcite , the american association of variable star observers ( aavso ) and asas databases .
the optical light curves are presented in figure [ fig : lc ] .
the epoch of the x - ray observation is marked with an arrow in each plot .
we did not find optical data for v794 aql around the epoch of the x - ray observations taken on 15 march 2011 .
however , from the photometric observations of the object before and after this date presented in @xcite it is reasonable to assume that v794 aql was in the intermediate state during the x - ray observation .
we used the heasoft version 6 - 13 to extract the _ swift _ and _ suzaku _ spectra , and the xspec version 12.8.0 for spectral modelling .
we also measured the _
uv _ magnitudes of the objects in both the _ swift_/uvot observations and in additional _ galex _ archival exposures . the _ chandra _
acis - s+hetg grating spectra were extracted with ciao version 4.3 .
four different partial exposures were added for both the heg and meg gratings spectra with the ftool addascaspec , written originally for _ asca _ , but also useful for all x - ray telescopes and detectors . for better resolution and in order to use the full exposure , we also combined the data from two _ suzaku _ front illuminated detectors ( xis 0 and xis 3 ) taken in 3x3 and 5x5 modes .
the timing analysis of the _ suzaku _ data was performed with the xronos sub - package of ftools after the barycentric correction .
tt arietis is one of the brightest cvs , usually between _ v _ magnitude 10 and 11 .
sometimes it abruptly falls into an ` intermediate state ' at _ v_@xmath6614 or even into a ` low state ' reaching _ v_@xmath6618 . according to @xcite
this binary system consists of a 0.571.2 m@xmath1 white dwarf and a 0.180.38 m@xmath1 secondary component of m 3.5@xmath67 spectral type @xcite .
the only low state before the one discussed in this paper was observed in the years 19801985 @xcite .
the first panel of figure [ fig : lc ] shows the long term light - curve of tt ari between 1990 and 2013 .
the optical brightness started declining dramatically at the beginning of 2009 and the low state lasted for about 9 months , with a drop in optical luminosity of about 7 mag .
however , in the low state the optical luminosity is not constant , with variations between _ v_=15 and _ v_=18 .
the high state x - ray spectrum of tt ari was at first obtained by _
exosat _ in aug 21/22 1985 @xcite .
authors found that the x - ray flux in the range 0.24.0 kev was about 1.9@xmath68 erg @xmath58 s@xmath2 .
they also proposed that there are two or more hot emitting x - ray regions and two or more cold absorbing or scattering regions in tt ari . on january 20 to 21 , 1994 tt ari was also observed with _ asca _ with an effective exposure time @xmath69 s. detailed analysis of these data was performed by @xcite .
one of the most interesting findings of the previous x - ray observations is the rapid variability of the x - ray flux , a quasi - periodic oscillations ( qpo ) with a semi - period of 1526 minutes ( @xcite , @xcite ) .
we will show that qpo with periods in this range are observed in all the high state observations we examined . in 2005
the _ chandra _ hetg spectra of tt ari were obtained by c. mauche ( first shown in a presentation by mauche , 2010 ) . below
we will discuss this set of observations in details .
= 0.1 cm clcccc name & state & date & instrument & exposure(s ) & count rate ( cts s@xmath2 ) + bz cam & high & 21/12/2012 & _ swift_-xrt & 15001 & @xmath70 + & low & 15/05/2011 & _ swift_-xrt & 2710 & @xmath71 + mv lyr & high & 08/06/2012 & _ swift_-xrt & 7569 & @xmath72 + & low & 29/07/2011 & _ swift_-xrt & 3282 & @xmath73 + tt ari & high & 01/08/1991 & _ _ rosat__-pspcb & 24464 & @xmath74 + & high & 06/09/2005 & _ chandra_-heg & 95362@xmath60 & @xmath75 + & high & 06/09/2005 & _ chandra_-meg & 95362@xmath60 & @xmath76 + & high & 06/07/2009 & _
suzaku_-xis fi@xmath63 & 28617 & @xmath77 + & high & 06/07/2009 & _
suzaku_-xis bi@xmath63 & 28617 & @xmath78 + & intermediate & 16/10/2009 & _ swift_-xrt & 4421 & @xmath79 + & low & 22/11/2009 & _ swift_-xrt & 12030 & @xmath80 + v794 aql & intermediate & 15/03/2011 & _ swift_-xrt & 6148 & @xmath81 + & low & 12/07/2011 & _ swift_-xrt & 4629 & @xmath82 + + + [ tab : obs ] a set of four _ chandra _ hetg exposures were obtained within 5 weeks in 2005 ( for details see table 2 ) . in the optical , the source was undergoing a ` shallow decline ' from the average optical magnitude in the high state , from _
[email protected] to _ [email protected] .
[ fig : ttarichandra ] shows the coadded meg and coadded heg spectra of four subsequent eposures .
there was no significant flux or spectral variability between the different exposures , and we shall describe the coadded meg and heg spectra .
a rich emission line spectrum was measured , with strong emission lines of mg , si , ne and s. the strongest lines of the _ chandra _ meg spectrum are listed in table 3 . for the h - like lines we evaluated the flux with a gaussian fit to the line ; we also estimated the flux in the he - like triplet lines ( ne ix , mg xi , si xiii ) , but we could only do it with larger uncertainty because the lines are blended and we needed to fit three gaussians ( note that the intercombination line is not resolved ) .
moreover , the triplets of he - like lines are observed in a region of the spectrum which is rich in other lines , like those due to transitions of iron . despite these difficulties
, we performed the fit with three gaussians for the triplets of si xiii and mg xi .
we added a fourth line of fe xviii at 13.509 for ne ix .
we thus evaluated the r ratio @xmath84 ( intensity of the forbidden to the intercombination line ) and the g ratio @xmath85 ( where r is the intensity of the resonance line ) .
we estimated an uncertainty of about 20% on both these ratios .
we obtained r=0.63 and g=0.78 for ne ix , r=0.36 and g=0.66 for mg ix , r=0.33 and g=0.66 for si xiii .
we consulted @xcite , who explored the dependence of these ratios on electron density and plasma temperature .
the authors assumed a photoionezed plasma , with or without additional collisional ionization .
we see from their figure 8 that the r ratios we obtained corresponds to high density ; we obtain a lower limit on the electron density n@xmath86 @xmath87 , @xmath88 @xmath87 and @xmath89 @xmath87 for ne , mg and si , respectively . however , it is known that the r ratios appears smaller , as if the density was higher than its actual value , when there is also photoexcitation by a strong _ uv_/_euv _ source , exciting the @xmath90 level electrons into the @xmath91 level @xcite .
we do expect additional photoexcitation if the lines are produced very close to the hot and luminous wd of tt ari , thus , we do not know that in this case the r ratio is a completely reliable indicator .
the g ratio , on the other hand , is reliable and indicates plasma temperature @xmath92 k , @xmath93 k and @xmath94 for ne , mg and si , respectively .
the next step was to fit the observed spectra with a physical model .
a fit with a plasma in collisional ionization equilibrium is not acceptable because of too high @xmath95 , but adding a second temperature we obtained a more reasonable fit , with @xmath95=1.2 .
we adopted two bvapec models in xspec , which describe the emission spectrum of collisionally ionized diffuse gas , calculated using the atomdb code v2.0.1 with variable abundances at different temperature ( see table 4 and fig .
[ fig : ttarichandra ] ) and with velocity broadening . by letting the abundances of single elements vary
, we found the best fit with the following values for the abundances : [ ne / h]=@xmath96 , [ mg / h]=@xmath97 , [ si / h]=@xmath98 , [ s / h]=@xmath99 , [ fe / h]=@xmath100 , [ o / h]=@xmath101 . the emission measure of the cooler component is 3.1@xmath102 @xmath103 and the emission measure of the hotter component is 3.6 @xmath104 @xmath103 .
we note that , if these two regions are related to accretion , for an electron density of order of 10@xmath105 @xmath87 ( the minimum electron density derived from the g ratio for si ) , the linear dimension of the emission region is of order of 3.1 @xmath106 cm and 7.1 @xmath106 cm , respectively .
this is , of course , an purely phenomelogical model ; two large and distinct regions with different plasma temperature are difficult to explain in a physically realistic way .
bvapec model performs the gaussian fitting of the lines and gives the @xmath107 for two systems of lines associated with the two components ( see table 4 ) .
the full width at half maximum that corresponds to these values of @xmath107 is about 1100 and 1500 km s@xmath2 .
we also wanted to try and better understand the physical scenario by adopting a more physically realistic model . @xcite
have shown that accretion in all non magnetic cvs , and often even in magnetic ones , is best described by a stationary cooling flow model .
we thus used the cooling flow vmcflow model in xspec ( a cooling flow model after * ? ? ?
we see however that this fit yields a larger @xmath95 value than the previous simplified model , and this is mainly because there is more flux in the he - like lines than predicted by the model
. this may be due to additional photoionization , for instance in a wind from the system , implying that not all the x - ray flux is produced in the accretion flow .
the main problem , however , is that the cooling flow model includes the mass accretion rate @xmath4 as a parameter , but predicts a very low value for @xmath4 , only @xmath108 m@xmath9 yr@xmath2 , while the uv and optical observations indicate @xmath109 m@xmath9 yr@xmath2 for the high state @xmath4 ( see table 1 ) .
we thus conclude that either the observed x - ray flux does not originate in the accretion flow that produces the luminous accretion disk , or that accretion energy is mostly re - radiated in the _ euv _ and not in the x - ray range .
@xmath110 ( @xmath111 ) & meg flux@xmath112}$ ] + s xv & 2.4606@xmath113 & 5.0387 & 2.8 + & 2.448@xmath114 & 5.064 & 3.3 + & 2.4260@xmath115 & 5.1106 & 3.4 + si xiv & 2.0061@xmath116 & 6.1803@xmath117 & 6.12 + si xiii & 1.8650@xmath113 & 6.6479 & 3.4 + & 1.854@xmath114 & 6.687 & 1.7 + & 1.8396@xmath115 & 6.6739 & 0.57 + mg xii & 1.4733@xmath118 & 8.4154@xmath119 & 3.07 + mg xi & 1.3522@xmath113 & 9.1687 & 2.0 + & 1.3434@xmath114 & 9.2291 & 1.6 + & 1.3312@xmath115 & 9.3136 & 0.59 + ne x & 1.0211@xmath120 & 12.142@xmath121 & 5.25 + ne ix & 0.9220@xmath113 & 13.44 & 3.3 + & 0.9149@xmath114 & 13.55 & 1.8 + & 0.9051@xmath115 & 13.69 & 0.59 + fe xviii & 0.8735 & 14.19 & 0.74 + fe xvii & 0.8256@xmath122 & 15.017@xmath123 & 2.98 + fe xvii & 0.7388@xmath124 & 16.781@xmath125 & 4.8 + + [ tab : lines ] the light curves of these _ chandra _ exposures still show a qpo , although the modulation has a @xmath6621 min period .
we discuss in detail below the similar light curve we extracted from an additional archival observation done with _ suzaku _ , which has higher s / n .
_ suzaku _ observations of tt ari were obtained by saitou in 2009 just before the low state ( see top panel of fig .
[ fig : lc ] ) .
the average x - ray flux during this observation was higher by almost a factor of 2 than during the _ chandra _ observation . in order to exclude the effect of a slightly different energy ranges of the detectors we compared the x - ray flux in the range 0.510.0 kev , common for both instruments .
the difference between the x - ray flux measured with _
chandra _ and _ suzaku _ may be correlated with the optical one .
hetg observations were taken at the time when tt ari was optically less luminous ( @xmath126 1 mag ) .
the broad band spectrum of tt ari observed with _
is presented in fig .
[ fig : ttarisuzaku ] .
emission lines of ne , mg , si are clearly seen , like in the _ chandra _ spectrum , together with s , fe xxv , and fe xxvi lines .
we fit this spectrum either with a two - component thermal plasma model with temperatures 0.80 and 7.1 kev ( for details see table 3 ) and the abundances that were derived from the _ chandra _ spectra fit .
the residuals of the fit indicate an extra line feature at 6.4 kev in the _ suzaku _ spectrum that is most probably the iron fluorescence line . in the inset in fig .
[ fig : ttarisuzaku ] we show the fit of the 5.58.0 kev region with three gaussians at 6.41 , 6.68 and 6.96 kev .
the equivalent width of the k @xmath127 iron line at 6.41 kev is @xmath128 ev .
this line implies that the x - ray emission region is close to a ` cold ' source , which may be the wd surface and/or an optically thick accretion disk , and it is consistent with the wd being less hot than about 100,000 k. the cooling flow model can be used also for the _ suzaku _ spectrum because we do measure spectral lines to constrain the model .
the fit is not optimal , and we run into the same problem of low @xmath4 .
the _ suzaku _ light curve is shown in fig .
[ fig : ttarixrlc ] and is extremely similar to the light curve previously observed with _
the data were integrated in bins of 100 s. the qpo have an amplitude of 50 % . in fig .
[ fig : fourier ] we present the fourier power spectrum of _ suzaku _ light curve of tt ari .
the highest peak corresponds to 0.839 mhz or 19.8 min oscillations .
according to @xcite at the time of _ suzaku _ observations tt ari also showed qpo in the optical range , with several peaks , at 2.5 , 1.1 and 0.38 mhz .
in @xcite the frequency of qpo in x - ray was explained as the beat frequency between the kepler frequency at the inner edge of the accretion disk and and the star s rotation rate .
nevertheless , from the qpo semi - periods measured by us in the _ chandra _ and _ suzaku _ observations and from those derived by @xcite and @xcite , no correlation emerges between the observed frequency of qpo and the x - ray flux of tt ari ( see the bottom panel of fig . [
fig : fourier ] ) . in 1991
tt ari was observed with _
rosat_. the _ rosat _ x - ray spectrum is shown in fig .
[ fig : rosat ] .
@xcite analysed these data and found that the best fitting model is an absorbed black - body .
we reanalysed the data and found that a black - body can represent only the soft part of the spectrum ( the _ rosat _ energy range is 0.22.4 kev ) and for more sophisticated fit we need two - component thermal plasma model .
parameters of the best fitting model are presented in table 4 .
emission lines at 6.41 , 6.68 and 6.96 kev and their fit with three gaussians . ] in the intermediate and low state tt ari was observed with _
the observations in the intermediate state were presented by @xcite .
the unabsorbed flux was about the same as during the _ suzaku _ observation , 1.5 @xmath68 ergs @xmath58 s@xmath2 , the spectrum could only be fitted with a multi - temperature plasma , and a new quasiperiod of 0.4 days was also measured in optical . the low state spectrum , presented in the top panel of fig .
[ fig : highlowmod ] , is best fitted with two components of absorbed thermal plasma in collisional ionization equilibrium with a fixed metallicity apec model at 0.7 kev and 3.9 kev respectively ( we only set a fixed metallicity with solar abundances because of the poorer data quality of this dataset ) .
the low state x - ray flux appeared to be about ten times smaller than in the high and intermediate state , and definitely no luminous supersoft x - ray phase was detected . in table 4
we show also present the comparison with only one component at 3.4 kev .
an upper limit to the temperature of a blackbody - like component is approximately 150,000 k. = 0.1 cm satellite & _ rosat _ & & & + models & 2 apec & 2 bvapec & vmcflow & 2 vapec+gauss@xmath60 & vmcflow+gauss@xmath60 & 2 apec & apec + n(h)@xmath129(@xmath130 ) & @xmath131&@xmath132 & @xmath133 & @xmath134 & @xmath135 & @xmath136 & @xmath137 + n(h)@xmath138(@xmath130 ) & & @xmath139 & & & & & + t@xmath129 ( kev ) & @xmath140 & @xmath141 & & @xmath142 & & @xmath143 & @xmath144 + t@xmath138 ( kev ) & @xmath145 & @xmath146 & & @xmath147 & & @xmath148 & + sigma@xmath129 ( @xmath149 ) & & @xmath150 & & & & & + sigma@xmath138 ( @xmath149 ) & & @xmath151 & & & & & + em@xmath129 ( @xmath152)@xmath63 & & @xmath153 & & & & & + em@xmath138 ( @xmath152)@xmath63 & & @xmath154 & & & & & + t@xmath155 ( kev ) & & & @xmath156 & & @xmath157 & & + t@xmath158 ( kev ) & & & @xmath159 & & @xmath160 & & + @xmath161 ( @xmath162 ) & & & @xmath163 & & @xmath164 & & + flux@xmath165 & @xmath166 & @xmath167 & @xmath168 & @xmath169 & @xmath170 & @xmath171 & @xmath172 + flux@xmath173 & @xmath174 & @xmath175 & @xmath176 & @xmath177 & @xmath178 & @xmath179 & @xmath180 + @xmath95 & 1.0 & 1.2 & 1.6 & 1.0 & 1.2 & 1.0 & 1.2 + + + + [ tab : ttarimod ] in the first panel of figure [ fig : lc ] , in addition to the optical light curve of tt ari , the red points show the _ galex _ near _ uv _ ( _ nuv _ ) observations . in table 6 we give exposure times and the mean ab magnitude in the _ u_/_uv _ filters during the low and high states .
the amplitude of the low to high state transition in _ nuv _ was much lower than in optical : 3 versus 7 magnitudes .
like in the optical range , tt ari shows flaring activity in the _ nuv _ , with amplitudes up to 1 mag . however , the _ uv _ and optical flares occur at different times , and do not appear to correlated , neither anti - correlated
. it may be due to the different origine of the _ uv _ and optical radiation .
bz cam shows brightness variations around an average value _ v _
= 1213 , with rare occasional transitions to low states with _ v _
= 1414.5 .
besides the low state studied here , two additional low states were detected in 1928 and at the beginning of 2000 ( @xcite and @xcite , respectively ) .
bz cam is surrounded by a bright emission nebula with a bow - shock structure , first detected by @xcite and also studied by @xcite , @xcite , @xcite .
@xcite proposed that the bow shock structure is due to the interactions of the wind of bz cam with the interstellar medium .
the wind in bz cam was also studied by @xcite .
@xcite suggested that this nebula is photoionized by a bright central object that must be a super soft x - ray source , while the bow shock structure is due to the high proper motion of bz cam , moving while it emits the wind . the second plot of fig .
[ fig : highlowmod ] shows the x - ray spectra of bz cam observed with the _ swift _ xrt . the luminosity is higher in the low state , however , in the very soft spectral region , at energy @xmath830.5 kev ,
the x - ray flux is almost twice higher in the high state , which is exactly the opposite of the scenario predicted by @xcite .
interestingly , the spectral fits in both states indicate that we may be observing an unresolved , strong ne x lyman @xmath127 line at 1.02 kev . in the high state spectrum
the fe xxv line at 6.7 kev is clearly detected .
bz cam was also previously observed with _
rosat _ in the high state .
@xcite and @xcite fitted the spectrum either with one component blackbody or with a highly absorbed bremsstrahlung ( or mewe ) model .
@xcite favored the blackbody .
however , with the larger energy range of _ swift _ we see that a fit including also the high energy part of the spectrum , is only possible with at least two components , and that the blackbody is not adequate .
the black dots in the second panel of fig .
[ fig : highlowmod ] show the low state spectrum of bz cam , and a fit with a two - component vapec model . with only a broad band spectrum and no detected emission lines
, we could not adequately fit the cooling flow model .
the same is true for other low resolution x - ray spectra described below . the high state of bz cam ( red dots in the second panel of fig .
[ fig : highlowmod ] ) is best fitted with a two components vapec model .
the fitting parameters are listed in table 5 .
we find the best fit with non solar abundances , [ fe / h]=(@xmath181 ) and [ ne / h]=(@xmath182 ) . in the low state
, the increased flux seems to be due to a higher maximum temperature , 64 kev instead of about 10 kev .
the column density n(h ) appears to increase in the low state .
again , a lower limit to the blackbody temperature is of about 150,000 k. l|ccc|ccc & & + & & low state & & low state + satellite & & _ swift _ & & _ swift _ + models & vapec & 2 vapec & 2 apec & vapec+plow & 2 apec & 2 apec + n(h)(@xmath130 ) & @xmath183 & @xmath184 & @xmath185 & @xmath186 & @xmath187 & @xmath188 + photon index & & & & @xmath189 & & + t@xmath129 ( kev ) & @xmath190 & @xmath191 & @xmath192 & @xmath193 & @xmath194 & @xmath195 + t@xmath138 ( kev ) & & @xmath196 & @xmath197 & & @xmath198 & @xmath199 + flux@xmath200@xmath60 & @xmath201 & @xmath202 & @xmath203 & @xmath204 & @xmath205 & @xmath206 + flux@xmath207@xmath60 & @xmath208 & @xmath209 & @xmath210 & @xmath211 & @xmath212 & @xmath213 + @xmath95 & 1.6 & 1.1 & 1.0 & 1.0 & 1.2 & + + [ tab : bzmvmod ] bz cam was observed in _ uv _ with _ swift _ during the low state with a 1172.43 s exposure and with a 121 s exposure during the high state .
the uv magnitudes in the ab system in the high and low states , given in table 6 , indicate a smaller variation than observed in the other objects .
this is explained by figure [ fig : bzcamim ] in which we show the uv image of the nebula obtained with _ swift_/uvot observations .
even with the poor spatial resolution of uvot , we detect an extended object ; obviously the ionized nebula also emits copious uv flux . comparing the image of bz cam in figure [ fig : bzcamim ] and the optical image in figure 4 of @xcite
, we see a trace of the bow shock oriented in the south - slightly south - west direction , but there seems to be additional _ uv _ emission of the nebula in the north - east direction , unlike in the in o and h @xmath127 images taken in september of 2000 with the wiyn telescope @xcite .
= 0.05 cm lcccccc object & state & date & instrument & exp.(s ) & mag.@xmath214 & filter@xmath60 + bz cam & high & 21/12/2012 & _ swift _ uvot & 1200.74 & @xmath215 & uvw2
+ & low & 15/05/2011 & _ swift _ uvot & 2737.32 & @xmath216 & u + mv lyr & low & 29/07/2011 & _ swift _ uvot & 2092.81 & @xmath217 & u + tt ari & high & 01/11/2005 & _ galex _ & 1468.3 & @xmath218 & fuv + & high & 13/11/200303/11/2007 & _ galex _ & 801686 & @xmath219 & nuv + & low & 15/11/200902/12/2009 & _ galex _ & 14661702 & @xmath220 & nuv + & low & 22/11/2009 & _ swift _ uvot & 6311.28 & @xmath221 & uvw2 + v794 aql & intermediate & 15/03/2011 & _ swift _ uvot & 6186.67 & @xmath222 & uvw1
+ & low & 12/07/2011 & _ swift _ uvot & 4732.45 & @xmath223 & uvm2 + + + [ tab : uv ] mv lyr spends most of the time in the high state , having brightness in the range _ v _
= 1213 , and during the occasional short low states it is in the range _ v _
. historical light curves of mv lyr can be found in @xcite , @xcite , @xcite and @xcite . with their _ fuse _ ( far ultraviolet spectroscopic explorer ) observations @xcite estimated that during the low state @xmath224 m@xmath1 yr@xmath2 , a four orders of magnitude decrease from the value of @xmath4 estimated by @xcite in the high state .
the third plot of figure [ fig : highlowmod ] shows the comparison between the high and low state x - ray spectra of mv lyr obtained with _
swift _ together with the spectral fits . unlike in bz cam
, the high state x - ray flux of mv lyr is higher by an order of magnitude than in the low state .
the spectrum is also harder , with an additional component prominent above 1.7 kev .
we fitted the high state spectrum of mv lyr with a two components thermal plasma model , but a good fit is also obtained with a thermal plasma and a power law model .
we fitted the low s / n , low state spectrum with a two components plasma model with fixed solar abundances , without any attempt to explore the role of the abundances ( see table 5 ) .
a _ rosat _ observation of mv lyr in november 1992 in the high state was studied by @xcite .
the authors fitted the spectrum with a black body , however , like in the case of bz cam , this model fails to fit the high energy part of the spectrum that we measured with _
@xcite observed mv lyr at the end of the 9 week optical low state in 1996 and obtained only an upper limit for the x - ray luminosity of 10@xmath225 erg s@xmath2 assuming a distance of 320 pc ( smaller than the most current estimate of 505@xmath22650 pc we give in table 1 ) . assuming a distance of 320 pc , the flux measured during low state observation done with _
swift _ ( see tab .
5 ) four months after the beginning of the decline to the low state , one month after minimum , would be 10@xmath227 erg s@xmath2 , more than a factor of 10 higher than this upper limit .
thus , it seems that the x - ray flux of mv lyr in the low state is not quite constant . in the high optical state v794 aql
varies between 14th and 15th magnitude , and in the low states it can plunge to 1820 mag .
( in the _ b _ filter , see * ? ? ?
* ) . @xcite fitted spectra of the hubble space telescope space telescope imaging spectrograph ( _ hst - stis _ ) and of _ fuse_. they derived the following binary system parameters : m@xmath228 m@xmath1 , high state @xmath229 = @xmath230 m@xmath1 yr@xmath2 , inclination @xmath231 , and distance to the system d = 690 pc .
the spectra and their fits of v794 aql in the intermediate state ( _ [email protected] ) and in the low state are presented in the bottom panel of fig .
[ fig : highlowmod ] .
the x - ray flux is three times higher in the intermediate than in the low state .
we fitted the intermediate state spectrum of v794 aql with two vapec components ( see table 7 ) .
in both components we need high abundance of mg ( @xmath232 ) .
again , there is no luminous supersoft x - ray source in the low state , with an upper limit of about 150,000 k. satellite & _ swift _ & _ swift _ + models & 2 vapec & apec + n(h)(@xmath130 ) & @xmath233 & @xmath234 + t@xmath129 ( kev ) & @xmath235 & @xmath236 + t@xmath138 ( kev ) & @xmath237 & + flux@xmath200@xmath60 & @xmath238 & @xmath239 + flux@xmath207@xmath60 & @xmath240 & @xmath241 + @xmath95 & 1.0 & 0.7 + + [ tab : v794mod ]
an important motivation for this research has been the claim by @xcite and @xcite that some of the wd in vy scl - type stars must be burning hydrogen quietly in the low state , without ever triggering thermonuclear flashes because of the short duration of the burning .
we found that the predicted supersoft x - ray source does not appear in the low states , so thermonuclear burning at high temperature is ruled out .
however , as we see in table 1 at least 3 of the 4 objects we investigated have low mass m@xmath242 . according to @xcite thermonuclear burning in wds
whose mass is lower than 1 m@xmath9 occurs quietly with atmospheric temperature below 150,000 k , outside of the sss window , except for very high values of the mass accretion rate ( see fig.5 of these authors ) . for wolf et al .
we see that wds with mass up to 0.7 m@xmath9 burn hydrogen in the stable regime ( without nova eruptions ) with @xmath3 of a few times 10@xmath243 m@xmath9 year@xmath2 have t@xmath244200,000 k. we note that also cal 83 , hosting a wd burning h at a much higher atmospheric temperature , has low , intermediate and high states in the optical and x - rays , although with a smaller amplitude in the optical than the vy scl binaries .
these variations were associated with the changes in the amount of irradiation of the accretion disc ( see e.g. @xcite , @xcite ) .
all the x - ray spectra of the vy scl systems we examined appear complex , and the _ chandra _ and _ suzaku _ spectra of tt ari clearly indicate more than one emission region or mechanism .
the best - fitting model for all the 0.310.0 kev broad band spectra is a , probably still simplistic , two - component absorbed thermal plasma model .
here we discuss the possible origins of the observed x - ray emission .
we found that in three of the systems the x - ray luminosity decreases during the optical and uv low states , although the x - ray flux variation is the smallest .
the x - ray flux seemed to be anti - correlatd with the optical and uv only in bz cam .
thus , if in the other three systems the main source of _ uv_/_fuv _ and optical luminosity is the accretion disk , it seems unlikely that the x - ray flux is due to the innermost portion of the disk .
in fact , our fits with the cooling flow model , which generally yields good results assuming that all the x - ray flux is emitted in an accretion flow , for all four objects return unreasonably low values of @xmath3 , which can not be reconciled with measurements at other wavelengths . this is not completely unexpected , since the accretion disks of systems accreting at high @xmath4 , close to 10@xmath245 m@xmath9 yr@xmath2 , seem to re - radiate mostly or completely in the _ euv _ range @xcite , because the boundary layer is optically thick . for tt ari , the semi - regular variability ( qpo ) with periods of 1726 minutes in the high state
is best explained with the flickering of an accretion disk , however we also found that there is no correlation between the x - ray flux and the frequency of the qpo , which would be expected for accretion disk flickering ( @xcite , @xcite ) .
if the origin of the x - ray emission is not in the boundary layer of the accretion disk , it may originate in a wind , either from the wd or from the accretion disk , depleting matter from the system .
such a wind may play an important role in the evolution , preventing the wd from reaching the chandrasekhar mass .
the fit of the tt ari emission lines observed with _
chandra _ indicates a fwhm in the range 11001500 km s@xmath2 .
however , the lines do not display any measurable blue or red shift to prove a wind scenario .
there is significant broadening , but it may be due to collisional ionization in the accretion flow , or to matter in almost - keplerian rotation .
the wd effective temperature and _ fuv _ flux reported in table 1 are consistent with a line driven wind , although if nuclear burning takes place , the radius of wd at some stage may increase , and we can not rule out that at some ( still not observed ) brief stage the wd reaches a luminosity where also electron scattering opacity starts playing a role ( a nova - like radiation pressure wind ) . in either case ,
the most likely origin for the x - ray flux in the observations we examined is circumstellar material , shocked when it collides with a new outflow , possibly at a large distance from the wd .
there may be circumstellar material left from the agb phase of primary or old remnant of a previous nova , or a previous ` thicker ' wind caused by enhanced luminosity due to nuclear burning , that has slowed down .
a strong stellar wind is very likely to play a role in the extended bz cam nebula , which was initially classified as a planetary nebula .
instead we would argue that for tt ari this explanation can not account for the largest portion of the x - ray flux , because this system shows a 6.4 kev reflection line , which indicates that a large fraction of x - rays ( at least x - rays above 7 kev ) must originate close to the white dwarf or to the disk .
there is a secure observation of x - rays far away from the accretion disk in ux uma ( see @xcite ) , an eclipsing nova - like with a hard , absorbed , eclipsed x - ray component and a soft , unabsorbed , uneclipsed x - ray component .
the soft x - rays in ux uma may indeed originate in a wind from the system .
a fast wind is also known to occur in cal 87 ( @xcite , @xcite ) , another system that may be closely related to the vy scl - type stars .
the x - rays and optical flux variations anticorrelate only in bz cam , so it is possible that in this system the wind increases in the low state , causing additional absorption and obscuring the accretion disk .
disk winds are observed in many types of compact objects , while the mechanism that causes them is not completely clear . at optical and uv wavelengths
a mass outflow from the disk has been inferred in some cv via the observation p cygni profiles , most notably the civ @xmath2461549 line ( @xcite , @xcite , @xcite ) .
p cygni line profiles or / and absorption features have also been detected in x - rays in low mass x - ray binaries ( @xcite , @xcite ) and are assumed to originate in a high - velocity outflow from a flared and x - ray - heated accretion disk .
disk winds also cause additional circumstellar , sometimes time - dependent , absorption components in the soft x - rays in non - magnetic cv s ( @xcite , @xcite ) .
a tempting hypothesis is that , while one component of the x - ray flux is due to a mass outflow from the system , another component originates in a different , and coexistent mode of accretion other than the disk , i.e. a stream to the polar caps . in short ,
the vy scl would be intermediate polars ( ip s ) .
this scenario explains the lack of a clear correlation of _ uv_/optical versus x - ray flux variations . as in the model proposed by @xcite
the stream to the polar caps still continue , at decreased rate , when the accretion disk is periodically disrupted in the optically low state . in an ip ,
the disk would emit in optical and _ uv _ , but it would be truncated instead of having an x - ray emitting boundary layer , no matter what the value of @xmath4 is .
mauche ( 2010 ) compared spectra of magnetic and non - magnetic cv s and made a point that division into the two classes is not clear - cut on the basis of the x - ray spectrum alone , because of the large variety of observed x - ray spectra of magnetic cv s .
there is no ` typical ' spectrum among polars and ip s .
there is evidence for and against the magnetic scenario for tt ari , but the x - ray spectrum alone does not prove or disprove it .
an x - ray flux modulation due to the wd rotation period , which is very typical and is considerd the smoking gun to classify ip s , has not been detected in these systems so far . for
three of them the reason may be low inclination , but not so for v794 aql .
however , if the major component of the x - ray flux in the high state is not the accretion stream to the poles , but it is associated instead with a wind , isolating the accretion component for the timing analysis is a serious hurdle in detecting a periodicity due to the wd rotation .
the vy scl binaries are critical to understand the evolution of wd interacting binaries .
they pose several riddles for the theories and understanding them well is a key to a consistent evolutionary picture .
are these systems almost always quietly accreting at a high rate , with short intervals of low @xmath4 that prevent the occurence of a thermonuclar flash and mass loss in nova outbursts ?
is thermonuclear burning of hydrogen on - going at all phases , and how do we find evidence since we do not observe their wd at the high effective temperature necessary to emit supersoft x - rays ?
we analyzed a number of x - rays and uv observations of four vy scl systems comparing phenomena occurring during the optically ` high ' and ` low ' state .
we did not detect supersoft x - ray emission in both states , however , we can not exclude h burning at a lower temperature , outside of the sss window , as can be expected from the low masses of the white dwarfs in these systems .
the data collected and examined in this paper suggest that the x - ray emission has more than one component in all the four systems .
we concluded that one component most likely originates in the circumstellar material , shocked by the wind , possibly at a large distance from the wd while the second component can be x - ray emission from the polar caps .
however , we are not able to prove neither clearly disprove an ip scenario for these systems .
it can be argued that the x - ray observations at this stage have posed more new puzzles .
we suggest that monitoring these systems over the years in optical , _ uv _ and x - rays as frequently and simultaneously as possible is a key to understand how accretion occurs and how it interplays with the thermal state of the secondary . more intensive monitoring , that may be done with
_ swift _ , would be very rewarding , allowing to understand whether an evolutionary path at high mass transfer rate without mass loss in nova outbursts can be sustained for a long time , and whether it leads to ` quieter ' outflows preventing the wd growth in mass , or to evolution towards a type ia supernova .
polina zemko acknowledges the grant of the national scholarship programme saia and a pre - doctoral grant of the cariparo foundation at the university of padova .
dr . shugarov acknowledges the vega grant no.2/0002/13 .
k. v. , suleimanov v. f. , nikolaeva e. a. , borisov n. v. , 2010 , in werner k. , rauch t. , eds , american institute of physics conference series vol .
1273 of american institute of physics conference series , modeling of the spectral energy distribution of the cataclysmic variable tt ari and evaluation of the system parameters .
pp 342345 j. , iyudin a. , jimenez - garate m. , burwitz v. , schwarz r. , distefano r. , schulz n. , 2004 , in tovmassian g. , sion e. , eds , revista mexicana de astronomia y astrofisica conference series vol . 20 of revista mexicana de astronomia y astrofisica , vol . 27 , resonant scattering and recombination in cal 87 .
pp 1820 r. f. , szymkowiak a. e. , 1988 , in fabian a. c. , ed . , nato asic proc .
229 : cooling flows in clusters and galaxies einstein observatory solid state detector observations of cooling flows in clusters of galaxies . pp 5362 m. , ebisawa k. , heise j. , hartmann j. , 2004 , in tovmassian g. , sion e. , eds , revista mexicana de astronomia y astrofisica conference series vol .
20 of revista mexicana de astronomia y astrofisica , vol .
27 , a new view of the supersoft x - ray source cal 87 observed with xmm - newton . | four vy scl - type nova - like systems were observed in x - rays during both the low and the high optical states .
we examined _ chandra _ , _ rosat _ , _ swift _ and _ suzaku _ archival observations of bz cam , mv lyr , tt ari , and v794 aql .
the x - ray flux of bz cam is higher during the low state , but there is no supersoft x - ray source ( sss ) as hypothesized in previous articles .
no sss was detected in the low state of the any of the other systems , with the x - ray flux decreasing by a factor between 2 and 50 .
the best fit to the _ swift _ x - ray spectra is obtained with a multi - component model of plasma in collisional ionization equilibrium .
the high state high resolution spectra of tt ari taken with _
chandra _
acis - s and the hetg gratings show a rich emission line spectrum , with prominent lines of in mg , si , ne , and s. the complexity of this spectrum seems to have origin in more than one region , or more than one single physical mechanism . while several emission lines are consistent with a cooling flow in an accretion stream , there is at least an additional component .
we discuss the origin of this component , which is probably arising in a wind from the system .
we also examine the possibility that the vy scl systems may be intermediate polars , and that while the boundary layer of the accretion disk emits only in the extreme ultraviolet , part of the x - ray flux may be due to magnetically driven accretion .
[ firstpage ] cataclysmic variables nova - likes : stars . |
synchronization of interacting elements is an emergent phenomenon in complex systems @xcite .
rhythms and synchronization of neuronal activities are essential for odor discrimination @xcite , visual feature integration @xcite , and brain computation @xcite , but perfect synchronization is not always desirable , but can sometimes be disastrous in mental disorders such as epilepsy @xcite . however ,
considering the enormous number of connections between cells and organs in the body , the apparent independence or desynchronization between rhythms in different organs may be a bigger puzzle rather than their synchronization @xcite .
indeed electrical engineers design desynchronization to implement time division multiple access ( tdma ) that prevents message collisions and provides asynchronous sleep cycles for nodes on wireless sensor networks @xcite .
complex systems sometimes show partial synchronization .
two types of partial synchronization have been recognized .
the _ chimera state _ shows spatial separation of synchronized and desynchronized domains @xcite , whereas the _ periodic synchronization _ shows temporal alternation between synchronized and desynchronized states @xcite .
unihemispheric sleep is an example of the chimera state where one half of the brain sleeps while the other half remains awake ; some animals adopt this strategy when predation risk is high @xcite .
context - dependent control of synchronization is therefore necessary to ensure that complex systems function appropriately .
in particular , to disrupt synchronization and achieve desynchronization , various methods have been proposed ; these include simply decreasing interaction strengths to stimulating with a short pulse @xcite , giving linear / nonlinear delayed feedback @xcite , and introducing inhibitory interactions @xcite .
recent studies have investigated the synchronization - desynchronization transition in complex networks @xcite .
these studies have revealed that in locally - coupled networks , this transition can be controlled by adapting the topology of networks @xcite . however , those controls require parameter optimization , because the transition between synchronized and desynchronized states is sharp . on top of the fine - tuning problem ,
real systems generally have high complexities : ( i ) complex networks have heterogeneous couplings with excitatory and inhibitory links @xcite ; ( ii ) the strength of the couplings depends on the activities of nodes @xcite ; and ( iii ) the networks are hierarchically organized @xcite .
therefore , precise control of synchronization looks implausible in complex systems . we ask how such complex systems realize precise and robust control of synchronization between interacting elements . in a hierarchical system ,
generation of coherent behavior becomes increasingly difficult as the complexity of subsystems increases .
this simple observation suggests a simple and effective way to control synchronization of a complex system by changing the _ complexity _ of its subsystems .
here we formulate a minimal model to demonstrate this idea by using coupled stuart - landau oscillators .
synchronization phenomena can be generally described by considering oscillators and their coupling . in particular , we adopt stuart - landau oscillators in which amplitude and phase are variable @xcite : @xmath0 first , we consider a _ subsystem _ that consists of three coupled oscillators with @xmath1 as a minimal set for generating diverse complexity .
the dynamics of the complex variable @xmath2 can be understood by decomposing amplitude and phase parts : @xmath3 in the absence of coupling ( @xmath4 ) , the amplitude converges to a stable focus ( @xmath5 ) for @xmath6 , but the focus loses stability at the hopf bifurcation point ( @xmath7 ) , and a stable limit - cycle emerges with amplitude @xmath8 and frequency @xmath9 for @xmath10 .
note that @xmath11 is another solution for @xmath10 .
here we used only positive - definite @xmath12 ; when @xmath12 became negative in simulations , we made it positive with the following transformation , @xmath13 and @xmath14 . once the coupling is applied ( @xmath15 ) , the subsystem produces ample dynamics depending on the adjacency matrix @xmath16 with entries @xmath17 that determine the coupling sign from oscillator @xmath18 to oscillator @xmath19 . the phase part of eq .
( [ phase ] ) is a generalized kuramoto model @xcite , in which oscillators have positive and negative couplings , and their coupling strengths depend on their amplitudes .
the amplitude dependence has the physical meaning that when oscillator @xmath18 affects oscillator @xmath19 , the coupling strength is proportional to the affecter amplitude @xmath20 , but inversely proportional to the receiver amplitude @xmath12 ; i.e. , the pair of strong affecter and weak receiver exhibits a maximal coupling . and @xmath21 , and ( b ) @xmath22 .
black points : attractors ; arrows : vector flows @xmath23 on the plane ( @xmath24 ) .
( b ) the basins of the three attractors , @xmath25 , @xmath26 , and @xmath27 , are painted with red ( top left ) , blue ( bottom right ) , and green ( middle ) , respectively . ] when coupling is weak , the amplitudes can be approximated as @xmath28 , and the three phase equations of eq .
( [ phase ] ) can be reduced to two equations of the phase differences , @xmath29 and @xmath30 . assuming identical intrinsic frequencies ( @xmath31 ) for simplicity , we obtain @xmath32 where @xmath33 .
their steady states ( @xmath34 and @xmath35 ) are governed by the parameters @xmath36 and the adjacency matrix @xmath16 . here
self - couplings can be safely ignored ( @xmath37 ) , because their contribution is absent in eq .
( [ phase ] ) . considering that the off - diagonal elements @xmath17 for @xmath38 can take either 1 or -1 for positive or negative couplings , the adjacency matrix @xmath16 has a total of @xmath39 possibilities . leaving index degeneracy aside , 16 cases remain ( fig .
[ fig : motif ] ) .
most of them drive @xmath40 to single attractors at steady states , regardless of @xmath36 . however , two anti - symmetric matrices , @xmath41 which correspond to networks 9 and 10 in fig . [ fig : motif ] , are exceptional in that they produce multiple attractors for similar @xmath36 ( @xmath42 ) .
however , the two anti - symmetric matrices also generate single attractors for largely dissimilar @xmath36 .
the effective coupling @xmath33 can be different from the topological coupling @xmath17 for largely dissimilar @xmath36 .
therefore , we focus on networks 9 and 10 , which can alter their complexity ( number of attractors ) by controlling @xmath36 . in particular , network 9 has three populations of distinguishable oscillators : the first oscillator attracts the other two ; the second repels the other two ; and the third attracts one and repels one .
in contrast , network 10 has three populations of indistinguishable oscillators .
network 9 produces single attractors for dissimilar @xmath36 ( fig .
[ fig : vectorflow]a ) , but three attractors for similar @xmath36 ( fig .
[ fig : vectorflow]b ) . for @xmath43 ,
the @xmath40 phase plane has three attractors ( @xmath44 ) , ( @xmath45 ) , and ( @xmath46 ) .
we label them @xmath25 , @xmath26 , and @xmath27 , because @xmath25 and @xmath26 have small basins of attractors , whereas @xmath27 has a big basin ( fig .
[ fig : vectorflow]b ) . when initial conditions [ @xmath47 , @xmath48 are given in the small basins , the subsystem
is quickly attracted to @xmath25 or @xmath26 , but when initial conditions are given in the big basin , the subsystem approaches @xmath27 slowly ; near the center @xmath27 , the change in the radius of the cycle is not readily apparent . when the amplitude dynamics in eq .
( [ amplitude ] ) is off ( @xmath49 ) , and the phase model in eq .
( [ phase ] ) is only considered with frozen amplitudes , the subsystem is not attracted to the centers , but revolves endlessly around the centers at given initial radii @xcite .
@xmath25 , @xmath26 , and @xmath27 are special positions in which the coupling terms , @xmath50 in eq .
( [ amplitude ] ) , vanish and amplitudes @xmath51 become fixed .
network 10 has the same phase plane as network 9 , with simple translations @xmath52 and @xmath53 .
, which then affect every unit in reverse . , width=340 ] to demonstrate that the synchronization of a hierarchical system can depend on the complexity ( number of attractors ) of subsystems , we construct a hierarchical system composed of multiple units .
each unit corresponds to one subsystem of the three coupled oscillators .
the amplitude and phase of the oscillator @xmath19 in the @xmath54th unit are represented as @xmath55 .
we consider an _
inter - unit _ coupling in addition to the _ intra - unit _ coupling ( fig .
[ fig : cell_diagram]a ) . in particular , we start by simply copying the intra - unit coupling into the inter - unit coupling : @xmath56 where @xmath57 is introduced to represent weaker inter - unit couplings compared with intra - unit couplings ( @xmath58 ) . by using an arithmetic average of @xmath59 for @xmath60 units , @xmath61 we can rearrange eq .
( [ slmodel2a ] ) as @xmath62 where @xmath63 and @xmath64 .
henceforth , we use the rescaled parameters as @xmath65 and @xmath66 .
this equation can be interpreted as the mean fields @xmath67 affect every unit ( fig .
[ fig : cell_diagram]b ) . to probe the inter - unit synchronization ,
we define order parameters for three populations of oscillators : @xmath68 with @xmath69 .
because three populations usually have the same degree of synchronization , hereafter , we represent them simply as @xmath70 , unless otherwise specified . for the anti - symmetric intra - unit couplings ,
the arithmetic mean field leads to desynchronize units ( @xmath71 ) , because the inter - unit coupling in eq .
( [ slmodel2 ] ) effectively generates repulsive interactions between units . for an example of two - unit systems ( @xmath72 )
, @xmath73 indirectly interacts with @xmath74 through @xmath75 or @xmath76 .
these three - step interactions always yield a negative loop as a net for the anti - symmetric matrices @xmath16 .
the effective repulsion between units leads them to stay as far away as possible .
this state has been referred as the _ splay state _
@xcite . therefore ,
under the arithmetic mean field , the anti - symmetric intra - unit coupling can provide an effective scheme for the desynchronization of hierarchical systems . for the inter - unit coupling , we also consider a geometric average ( log - average ) of @xmath59 , @xmath77^{1/n}.\ ] ] the geometric average is frequently suitable in biological systems @xcite . unlike the arithmetic average in eq .
( [ arithmetic_average ] ) , the geometric average decouples amplitude and phase averages : @xmath78^{1/n } , \\
\bar{\theta}_\sigma & \equiv & \frac{1}{n } \sum_{n=1}^n \theta_{n\sigma}.\end{aligned}\ ] ] the average phase @xmath79 is given as an arithmetic average of bare phases , independent on amplitudes .
after adopting the geometric mean field @xmath80 of eq .
( [ geometric_average ] ) , we decompose eq .
( [ slmodel2 ] ) into amplitude and phase parts : @xmath81 , \\
\label{phase2 } \dot{\theta}_{n\sigma } & = & \omega_{n\sigma } + k \sum_{\sigma ' \neq \sigma } a_{\sigma \sigma ' } \bigg [ \frac{r_{n\sigma'}}{r_{n\sigma } } \sin(\theta_{n\sigma ' } - \theta_{n\sigma } ) + \epsilon \frac{\bar{r}_{\sigma'}}{r_{n\sigma } } \sin(\bar{\theta}_{\sigma ' } - \theta_{n\sigma } ) \bigg].\end{aligned}\ ] ] in this study , we adopt the geometric mean - field coupling , because it produces broader spectrum of synchronization between units .
this completes our formulation of hierarchical oscillators . and @xmath82 ) and identical intrinsic frequencies ( @xmath83 ) for @xmath84 , the hierarchical system generates seven dynamic states : ( i ) @xmath85 state , in which two units are attracted into the same attractor @xmath25 or @xmath26 ; ( ii ) @xmath86 state , in which one unit rotates around @xmath25 or @xmath26 , and the other rotates around @xmath27 ; ( iii ) @xmath87 state , in which one unit rotates around @xmath27 , and the other unit jumps alternately between @xmath25 and @xmath26 .
( iv ) @xmath88 state , in which two units slowly approach @xmath27 ; ( v ) @xmath89 ; ( vi ) @xmath90 ; ( vii ) @xmath91 states , in which two units slowly approach horizontally- , vertically- , and diagonally - shifted attractors from @xmath27 .
different colors represent different pairs of two units.,width=453 ] x & @xmath92 & @xmath93 & @xmath94 & @xmath95 + @xmath85 & @xmath96 & @xmath97 & @xmath98 & @xmath99 + @xmath100 & @xmath101 & @xmath97 & @xmath102 & - + @xmath86 & @xmath103 & @xmath97 & @xmath104 & @xmath97 + @xmath87 & @xmath102 & - & @xmath105 & @xmath106 + @xmath88 & @xmath107 & @xmath97 & @xmath108 & @xmath109 + @xmath89 & @xmath102 & - & @xmath110 & @xmath111 + @xmath90 & @xmath102 & - & @xmath112 & @xmath111 + @xmath91 & @xmath102 & - & @xmath113 & @xmath111 +
we investigate the synchronization of hierarchical oscillators governed by eqs .
( [ amplitude2 ] ) and ( [ phase2 ] ) .
we first consider a two - unit system ( @xmath72 ) , because it contains essential ingredients in the hierarchical system ; then , we extend the model into large systems . to examine the dynamics of two - unit systems ,
we again focus on the phase differences ( @xmath114 and @xmath115 ) between three oscillators within each unit , and the phase difference @xmath116 between two units .
the phase variables ( @xmath117 ) of two units are transformed to ( @xmath118 ) where @xmath119 , @xmath120 , @xmath121 , @xmath122 , and @xmath123 . here
we assume identical intrinsic frequencies ( @xmath124 ) for simplicity .
first , when @xmath36 are widely dissimilar , individual units have single and identical attractors ( section [ section2 ] ) .
thus the phase differences ( @xmath125 , @xmath126 ) and ( @xmath127 , @xmath128 ) for two units are attracted into the same attractor in the absence of inter - unit coupling .
the attraction is not perturbed even with weak coupling ( @xmath82 ) between units .
therefore , the inter - unit coupling contributes to fully synchronize the two units ( @xmath129 ) .
second , when @xmath36 are similar ( e.g. , @xmath84 ) , individual units of network 9 have three attractors ( fig . [
fig : vectorflow]b ) , and two units can be attracted to different attractors ( @xmath25 , @xmath26 , @xmath27 ) .
this complexity impedes the synchronization between the two units . in the absence of inter - unit coupling ,
the two units have four states if one considers the symmetry of @xmath25 and @xmath26 : ( i ) @xmath85 state , in which two units stay in the same small basin ; ( ii ) @xmath100 state , in which two units stay in different small basins .
( iii ) @xmath86 state , in which one unit stays in the big basin , while the other unit stays in one of the two small basins ; and ( iv ) @xmath88 state , where two units stay in the big basin . under weak inter - unit coupling ( @xmath82 )
, we numerically solved eqs .
( [ amplitude2 ] ) and ( [ phase2 ] ) , and observed ample dynamics of the two - unit system ( fig .
[ fig : phase_two ] ) : * @xmath85 state .
two units arrive at the same attractor , either @xmath25 or @xmath26 , and are fully synchronized ( @xmath129 ) .
* @xmath100 state .
this state is unstable and excluded for @xmath82 .
therefore , two units never sit on @xmath25 and @xmath26 .
* @xmath86 state .
two units rotate around @xmath27 and @xmath130 instead of being attracted into fixed points .
* @xmath87 state .
one unit shows a precessional cycle around @xmath27 , and the other unit jumps alternately between @xmath25 and @xmath26 .
when two units sit on different basins ( @xmath86 and @xmath87 states ) , they show partial synchronizations ( @xmath131 ) .
regarding the @xmath88 state , we found new stationary solutions in the two - unit system that satisfy @xmath132 with no contribution of coupling terms in eq .
( [ amplitude2 ] ) : @xmath133 + \epsilon k \big [ \cos \big ( \bar{x } + \frac{z}{2 } \big ) + \cos \big ( \bar{y } + \frac{z}{2 } \big ) \big ] = 0 , \\ k \big [ \cos x_1 - \cos ( x_1 - y_1 ) \big ] + \epsilon k \big [ \cos \big ( x_1 - \frac{z}{2 } \big ) - \cos \big ( x_1 - \bar{y } - \frac{z}{2 } \big ) \big ] = 0 ,
\\ k \big [ \cos y_1 + \cos ( x_1 - y_1 ) \big ] + \epsilon k \big [ \cos \big ( y_1 - \frac{z}{2 } \big ) + \cos \big ( \bar{x } - y_1 + \frac{z}{2 } \big ) \big ] = 0 ,
\\ k \big [ \cos x_2 + \cos y_2 \big ] + \epsilon k \big [ \cos \big ( \bar{x } - \frac{z}{2 } \big ) + \cos \big ( \bar{y } - \frac{z}{2 } \big ) \big ] = 0 , \\
k \big [ \cos x_2 - \cos ( x_2 - y_2 ) \big ] + \epsilon k \big [ \cos \big ( x_2 + \frac{z}{2 } \big ) - \cos \big ( x_2 - \bar{y } + \frac{z}{2 } \big ) \big ] = 0 , \\ k \big [ \cos y_2 + \cos ( x_2 - y_2 ) \big ] + \epsilon k \big [ \cos \big ( y_2 + \frac{z}{2 } \big ) + \cos \big ( \bar{x } - y_2 -\frac{z}{2 } \big ) \big ] = 0\end{aligned}\ ] ] with @xmath134 and @xmath135 . under these constraints , eq .
( [ phase2 ] ) yields five phase difference equations : @xmath136 \nonumber \\ & & + \epsilon k \big [ \sin \big(\bar{x } + \frac{z}{2 } \big ) + \sin \big(\bar{y } + \frac{z}{2 } \big ) - \sin \big(x_1-\frac{z}{2 } \big ) + \sin \big(x_1 - \bar{y } - \frac{z}{2 } \big ) \big ] , \\ { \dot{y}_1 }
& = & k \big [ \sin x_1 + \sin(x_1 - y_1 ) \big ] \nonumber \\ & & + \epsilon k \big[\sin \big(\bar{x } + \frac{z}{2 } \big ) + \sin \big(\bar{y } + \frac{z}{2 } \big ) - \sin \big(y_1-\frac{z}{2 } \big ) + \sin \big ( \bar{x } - y_1 + \frac{z}{2}\big ) \big ] , \\ { \dot{x}_2 } & = & k\big [ \sin y_2 + \sin(x_2 - y_2 ) \big ] \nonumber \\ & & + \epsilon k \big [ \sin \big ( \bar{x } - \frac{z}{2 } \big ) + \sin \big ( \bar{y } - \frac{z}{2 } \big ) - \sin \big(x_2+\frac{z}{2 } \big ) + \sin \big(x_2 - \bar{y } + \frac{z}{2 } \big ) \big ] , \\ { \dot{y}_2 } & = & k \big [ \sin x_2 + \sin(x_2 - y_2 ) \big ] \nonumber \\ & & + \epsilon k \big [ \sin \big ( \bar{x } - \frac{z}{2 } \big ) + \sin \big ( \bar{y } - \frac{z}{2 } \big ) - \sin \big(y_2+\frac{z}{2 } \big ) + \sin \big(\bar{x } - y_2 - \frac{z}{2 } \big ) \big ] , \\ \label{eq : z } { \dot{z } } & = & k \big [ \sin x_1 + \sin y_1 - \sin x_2 - \sin y_2 \big ] \nonumber \\ & & + \epsilon k
\big [ \sin \big(\bar{x}+\frac{z}{2}\big ) -\sin \big(\bar{x } -\frac{z}{2}\big ) + \sin \big(\bar{y}+\frac{z}{2 } \big)- \sin \big(\bar{y}-\frac{z}{2 } \big ) \big].\end{aligned}\ ] ] we examine the stationary conditions , @xmath137 .
it is interesting that @xmath138 is automatically satisfied , once @xmath139 with constraints , @xmath140 and @xmath141 .
the condition gives four @xmath88 states : * @xmath88 state ( @xmath142 and @xmath143 ) .
two units arrive at @xmath27 ( @xmath144 and @xmath145 ) with arbitrary phase difference @xmath116 between two units .
any @xmath116 values can satisfy @xmath146 in eq .
( [ eq : z ] ) ; i.e. , two units are coupled , but still behave independently .
* @xmath89 state ( @xmath147 and @xmath143 ) .
two units ( @xmath125 , @xmath126 ) and ( @xmath127 , @xmath128 ) are horizontally located from their center ( @xmath102 , @xmath102 ) ; @xmath145 , and @xmath148 and @xmath116 satisfy @xmath149 * @xmath90 state ( @xmath142 and @xmath150 ) .
two units ( @xmath125 , @xmath126 ) and ( @xmath127 , @xmath128 ) are vertically located from their center ( @xmath151 , @xmath151 ) ; @xmath144 , and @xmath152 and @xmath116 satisfy @xmath153 * @xmath91 state ( @xmath147 and @xmath154 ) .
two units ( @xmath125 , @xmath126 ) and ( @xmath127 , @xmath128 ) are diagonally located from their center ( @xmath102 , @xmath151 ) ; @xmath155 , and @xmath156 and @xmath116 satisfy @xmath157 the complex eqs .
( [ complexeq1])-([complexeq3 ] ) give exact values of the phase differences ( @xmath158 ) and ( @xmath159 ) between three oscillators within each unit , and the phase difference @xmath116 between two units . however , if @xmath57 is too small , they do not have solutions .
therefore , @xmath89 , @xmath90 , and @xmath91 states can only emerge under sufficient inter - unit coupling . for @xmath82 ,
the three states show partial synchronization ( @xmath160 ) .
the realization of these seven states depends on initial conditions , [ @xmath161 . however , the initial - condition dependence is chaotic in the two - unit system .
although we can not predict final states given initial conditions , the realization probability of each state is not uncertain ( table [ tab : table1 ] ) .
the probability of @xmath85 state , which generates full synchronization ( @xmath129 ) , is relatively small because the probability is small that two units will enter the same small basin .
therefore , as the number of units considered increases , the likelihood that each unit sits on a different attractor also increases , so the units in a hierarchical system may become partially synchronized .
we performed the same analysis for network 10 , and confirmed that the two - unit system of network 10 also generates seven states like network 9 .
their realization probabilities are also the same except for the switch between @xmath88 and @xmath89 states : @xmath162(@xmath163network10)=@xmath162(@xmath164network9 ) @xmath165 and @xmath162(@xmath164network10)=@xmath162(@xmath163network9 ) @xmath166 . .
given network 9 ( red ) and network 10 ( blue ) , single attractors are generated for dissimilar @xmath36 ( @xmath167 and @xmath21 ) , while three attractors are generated for similar @xmath36 ( @xmath22 ) .
the dissimilar @xmath36 is changed to the similar @xmath36 at @xmath168 , and back to the dissimilar @xmath36 at @xmath169 . for the simulation , coupling parameters
were set to @xmath170 and @xmath171 , and uniformly - distributed intrinsic frequencies were used as @xmath172 $ ] .
, width=453 ] and @xmath173 , the degree of synchronization between 1000 units is recorded for a long period after equilibration . its time average @xmath174
is represented by colors ( blue : desynchronization ; red : synchronization ) .
( a ) network 9 and ( b ) network 10 .
axes have base-10-logarithmic scales to allow presentation of a large range of @xmath36 .
for the simulation , coupling parameters were set to @xmath170 and @xmath171 , and uniformly - distributed intrinsic frequencies were used as @xmath172 $ ] .
, width=604 ] we now consider a hierarchical system that is composed of a large number of subsystems ( @xmath175 ) . when the system is sufficiently large ( @xmath176 ) , it produces homogeneous synchronization patterns , independent of initial conditions , unlike the two - unit system .
however , like the two - unit system , we could control the synchronization of the large system simply by changing the complexity ( number of attractors ) of subsystems by adjusting @xmath36 ( fig .
[ fig : control ] ) . for largely dissimilar @xmath36 ,
individual units have single attractors , and thus become easily synchronized ( @xmath129 ) .
in contrast , similarity of @xmath36 increases , it becomes more probable that they will enter basins of different attractors .
this circumstance leads to partial synchronization or desynchronization of units ( @xmath177 ) .
network 9 was slightly more effective for synchronization , but slightly less effective for desynchronization than network 10 .
this result is similar to the observation that in the two - unit system , network 10 had a higher probability than network 9 of generating the bb state , in which individual units behave independently despite the inter - unit interaction .
to check the robustness of controllability , we examined how parameters @xmath36 affect the degree of synchronization .
we obtained similar degrees of synchronization over large ranges of @xmath36 ( fig . [
fig : parameter ] ) .
this demonstrates robust control of synchronization based on changes of the number of attractors , not based on changes of the attractor positions .
we introduced a simple and effective way to control synchronization between elements in complex systems , in which hierarchically - organized heterogeneous elements have asymmetric and activity - dependent couplings .
we formulated this idea by using stuart - landau oscillators .
each subsystem consisted of three oscillators that interacted to each other positively or negatively .
anti - symmetric couplings between three oscillators were special to generate single and multiple dynamic attractors depending on the amplitudes of the oscillators . here
the number of attractors could specify the complexity of subsystems .
when the subsystems were connected through a mean field of their average activities , the degree of synchronization between subsystems was controllable by changing the complexity of subsystems .
in particular , we considered two kinds of mean fields using arithmetic average and geometric average , and found that the geometric mean field that counts large outliers less important was effective for controlling synchronization . the controllable synchronization can be applied to understand the synchronization behavior of complex biological networks .
interestingly , anti - symmetric couplings of three populations ( network 9 ) are realized in pancreatic islets @xcite , and the synchronization between pancreatic islets is an important requirement in glucose metabolism @xcite .
our finding may help to understand how islet - cell networks are synchronized .
this research was supported by basic science research funded by ministry of science , ict & future planning no .
2013r1a1a1006655 and by the max planck society , the korea ministry of education , science and technology , gyeongsangbuk - do and pohang city . | the controllability of synchronization is an intriguing question in complex systems , in which hiearchically - organized heterogeneous elements have asymmetric and activity - dependent couplings . in this study
, we introduce a simple and effective way to control synchronization in such a complex system by changing the complexity of subsystems .
we consider three stuart - landau oscillators as a minimal subsystem for generating various complexity , and hiearchically connect the subsystems through a mean field of their activities .
depending on the coupling signs between three oscillators , subsystems can generate ample dynamics , in which the number of attractors specify their complexity .
the degree of synchronization between subsystems is then controllable by changing the complexity of subsystems .
this controllable synchronization can be applied to understand the synchronization behavior of complex biological networks .
_ keywords _ : synchronization , stuart - landau model , controllability , complexity |
the unusual properties of high temperature cuprate superconductors continue to challenge our understanding of the physics of electrons in metals .
the related ideas of `` quantum criticality '' @xcite and density wave ( or `` stripe '' ) order and fluctuations @xcite have been much debated , but in the _ hole - doped _ cuprates the discussion has been inconclusive . further , while it has been proposed @xcite that many of the anomalous properties of the materials could be explained by a quantum critical point at or near `` optimal doping '' , there is no unambiguous evidence of long ranged order setting in near optimal doping in the hole - doped cuprates .
the _ electron - doped _ cuprates offer a new perspective on these issues . in these materials commensurate @xmath5 spin density wave order has been detected by muon spin rotation @xcite and neutron scattering measurements .
@xcite the order exists over a wide range of dopings and vanishes at a critical doping @xmath6 near the `` optimal '' doping @xmath7 .
thus , in this material an unambiguous quantum critical point exists , separating an apparently disordered phase from a phase with a well - defined ( @xmath5 commensurate magnetic ) order .
it is therefore of interest to examine the effect of the order and criticality on material properties . in this paper
we consider the low temperature hall resistance , which is sensitive to the rearrangment of the fermi surface caused by the onset of the spin density wave order .
we apply to the electron - doped cuprate case a mean - field analysis presented previously in the literature , @xcite and compare the results to experiments .
@xcite the rest of this paper is organized as follows .
[ sec : model ] defines the model to be studied .
the formal solution to boltzmann equation with relaxation time approximation for 2-dimensional systems is derived in sec .
[ sec : formalism ] .
[ sec : weak field limit ] gives the numerical results for @xmath8 in weak magnetic field and compares them to experimental data . in sec .
[ sec : critical behavior ] we analyze our formal solution in the critical region , @xmath9 .
the sensitivity of our numerical results to variation of parameters is studied in sec .
[ sec : sensitivity ] .
the possibility of explaining the @xmath10 experimental results by calculating the real part of the self energy due to spin fluctuations is investigated in sec .
[ sec : normal state region ] .
we conclude our calculations and discuss our results in sec .
[ sec : conclusion and discussion ] .
we consider electrons moving in a square lattice of unit lattice constant with dispersion @xmath11 the `` canonical '' values of the band parameters are @xmath12ev , @xmath13 and @xmath14 .
as will be shown below , our main results are not sensitive to the precise parameter choices .
we are interested in electron dopings , corresponding to two dimensional carrier densities per unit cell @xmath15 . ] for @xmath16 , band parameters @xmath17 , @xmath13 , @xmath14 and backscattering ( `` gap '' ) values @xmath18 ( dashed line ) , @xmath19 ( light solid line ) and @xmath20 ( heavy solid line ) .
the dotted line is given by @xmath21.__,width=240 ] we assume that for @xmath4 less than a critical value @xmath6 a commensurate spin density wave ( sdw ) order occurs , so the electrons are subject to a coherent backscattering of amplitude @xmath22 and wavevector @xmath23 which doubles the unit cell , changes the dispersion to @xmath24 and reconstructs the fermi surface . if @xmath18 then ( at the relevant dopings @xmath4 ) the fermi surface consists of one large hole surface centered at the @xmath25 point @xmath26 ( dashed line in fig [ fig : fermi surface ] ) . if @xmath27 then the fermi surface is reconstructed .
the details depend on the band filling and the magnitude of @xmath22 . if the chemical potential @xmath28 is such that @xmath29 then both bands cross the fermi level and the fermi surface involves two symmetry - inequivalent hole pockets centered at @xmath30 and one electron pocket centered at @xmath31 .
one of the hole pockets and two portions of the electron pocket are shown as the light solid lines in fig [ fig : fermi surface ] . however , if @xmath32 , the lower band is completely filled and only the electron pocket remains ( heavy solid line in fig [ fig : fermi surface ] ) .
for the parameters used in fig [ fig : fermi surface ] , @xmath33ev . as @xmath4
is decreased below @xmath6 we assume that the gap magnitude increases . however , the band filling and hence the chemical potential decreases .
therefore , depending on the relative growth ( with decreasing @xmath4 ) of @xmath28 and @xmath22 either the hole pockets vanish at a critical @xmath34 , or it persists to @xmath35 .
the comparison to data given below suggest that the former circumstance obtains in the actual materials , and that indeed the gap opens so rapidly that @xmath36 is of order @xmath37 , not much less than @xmath38 .
we require the @xmath0 transport coefficients . in the @xmath39 limit the only scattering is due to impurities , and we assume a boltzmann equation description is adequate , so that the physics is described by a momentum space distribution function @xmath40 with @xmath41 the fermi - dirac distribution .
we also adopt the relaxation time approximation , linearize in the applied electric field @xmath42 and take the @xmath39 limit so that for each band @xmath43 , the boltzmann equation becomes ( @xmath44 ) : @xmath45 we are interested in two dimensional electrons .
we take @xmath46 and note that @xmath47 corresponds to a derivative along the fermi contour . defining @xmath48 to be the coordinate along the fermi arc , noting that the structure of the equation implies @xmath49 and temporarily dropping , for ease of writing , the band superscript b allow us to rewrite eq [ [ be ] ] as @xmath50h(s)=e(s ) \label{be2}\ ] ] with ( @xmath51 is the mean free path , and @xmath52 is the superconducting flux quantum ) : @xmath53 eq [ [ be2 ] ] may be formally solved @xcite in terms of the green function @xmath54 satisfying @xmath55k(s , s')=\delta(s - s ' ) \label{kdef}\ ] ] with boundary condition @xmath56 as @xmath57 the currents longitudinal and transverse to @xmath42 may then be constructed by averaging over the velocity as usual . assuming a square lattice , restoring the factors of velocity and the band label ( if necessary ) , and symmetrizing or antisymmetrizing for @xmath58 , we find @xmath59 \label{sigxx}\\ \sigma_{xy}&=\sigma_{q}\oint \frac{dsds'}{p_b^2 } k(s , s ' ) \sin[\phi(s)-\phi(s ' ) ] \label{sigxy}\end{aligned}\ ] ] where @xmath60 is conductance quantum , @xmath61 is proportional to the inverse of the magnetic length @xmath62 , and @xmath63 is the polar angle of fermi velocity on the corresponding fermi contours .
the formulae are valid as long as magnetic breakdown can be neglected . a criterion for this
is given below .
, shown in fig [ fig : fermi surface ] .
p , sp are respectively the fermi lines for the normal state and the normal state fermi line shifted by @xmath64 .
the curvature has been neglected .
@xmath65 denotes the electron pocket of sdw state and @xmath66 denotes the hole pocket .
arrows along fermi lines show the circulation direction determined by @xmath67 with @xmath68 .
they also indicate the direction of increase in arclength coordinate , @xmath48 .
@xmath69 , @xmath70 , @xmath71 and @xmath72 are fermi velocities along corresponding fermi lines .
the dashed line is the diagonal of the first quadrant of brillouin zone connecting @xmath31 and @xmath73 .
the dot - dashed line is perpendicular to the dashed line : if the normal state energy dispersion is linearized , and the momentum dependence of velocity is neglected , the problem is symmetric under reflection through the dot - dashed line.__,width=240 ] the problem has several momentum scales .
obtaining a non - vanishing hall coefficient requires breaking of particle - hole symmetry , either by a variation of scattering rate , a fermi surface curvature , or a variation of velocity .
all of these phenomena are characterized by momentum scales of the order of the size of the brillouin zone ; we refer to these scales generically as @xmath74 .
we take these to be large , so we may neglect landau quantization .
a second important scale @xmath75 is set by the sdw backscattering amplitude , @xmath22 , and vanishes as @xmath76 .
a third scale is the range , @xmath77 , of the kernel , @xmath78 ( eq [ [ adef ] ] ) .
we distinguish weak field ( @xmath79 ) and strong field ( @xmath80 ) regimes .
a fourth significant scale is the inverse of the mean free path , @xmath81 .
the results we present are valid for @xmath82 .
in addition , we expect disorder to modify the sdw behavior significantly for @xmath83 ; so our results do not apply for @xmath84 . in this case
we may neglect the periodicity and write @xmath85\theta(s - s ' ) \label{ksoln}\ ] ] finally , we note that all of the formulae are derived on the assumption that a state evolves along a given fermi contour and does not jump to another one , _
i.e. _ , that magnetic breakdown may be neglected . by applying the considerations of ref . to the model studied here
, we find the condition for absence of magnetic breakdown is , in order of magnitude , that the inverse of the magnetic length @xmath86 be less than the gap momentum scale @xmath87 .
our results are only valid if @xmath88 . the condition that breakdown occurs at a field higher than the weak - strong field crossover is @xmath89 , which as noted above is a necessary condition for the validity of the approach .
in this section we analyse in detail the case in which the field is weak enough that @xmath77 is small relative to all scales in the problem . in this case the @xmath90 integrals in eqs [ [ sigxx],[sigxy ] ] are dominated by @xmath91 and may be performed , leaving @xmath92 for the longitudinal and hall conductance per plane .
eq [ [ sigxxlow ] ] is exactly the drude result ( average of mean free path over the fermi arc ) , while eq [ [ sigxylow ] ] reproduces ong s result @xcite @xmath93 further analysis requires information about the scattering rates .
the two commonly made assumptions are : * mean free path is constant over the fermi surface * scattering rate is constant over fermi surface the constant scattering rate assumption may be derived theoretically by assuming weak ( born approximation ) point - like scatterers .
constant mean free path is more appropriate for very strong ( unitary ) scattering .
ong @xcite shows that the constant mean free path approximation implies that the hall conductance for each fermi surface pocket is a constant , depending only on the value of the mean free path and the sign ( electron- or hole - like ) of the fermi pocket .
therefore , in the constant mean free path approximation , @xmath94 does not change as @xmath4 is decreased through @xmath6 .
the reconstruction of the fermi surface means that one goes from one hole pocket ( @xmath95 ) to two hole pockets ( @xmath96 ) and one electron pocket ( @xmath97 ) so that the total @xmath98 .
however , at @xmath34 the hole pockets vanish , and in the constant mean free path approximation the hall coefficient would jump discontinuously from a value close to the @xmath99 implied by the initial , large , hole - like surface , to the value @xmath100 implied by a single electron - like pocket .
this behavior is in evident disagreement with data , @xcite which indicates instead a smooth change beginning at @xmath101 , which is approximately the doping at which antiferromagnetism sets in . @xcite , [ ongformula ] ] for parameters @xmath12ev , @xmath13 , @xmath102 and @xmath103=0.6\sqrt{1-x/0.16}$ ] .
long dashed line : hall coefficient calculated for parameters @xmath12ev , @xmath13 , @xmath104ev and @xmath103=0.7\sqrt{1-x/0.16}$ ] .
they are calculated with constant @xmath105 approximation .
dotted line : curve @xmath106 for @xmath107 and @xmath108 for @xmath109 .
stars : experimental data .
@xcite solid squares : experimental data .
@xcite the open box shows the approximate size of the region for which the weak field - strong field crossover may be observable with present accessible fields.__,width=240 ] we therefore turn to the constant @xmath105 approximation . in this case , @xmath94 for each pocket is weighted by the square of the fermi velocity for that pocket and more variation occurs .
the solid line and long dashed line in fig [ fig : hall ] show typical results obtained by evaluating eqs [ [ sigxxlow],[ongformula ] ] in the constant scattering rate approximation and assuming a backscattering amplitude growing proportional to @xmath110 .
one sees that for @xmath10 the hall resistance is hole - like and depends only very weakly on @xmath4 . at @xmath111
a nonanalyticity occurs , and @xmath8 begins to decrease . for the parameters chosen , @xmath112 . at this doping
the hole pockets vanish , a further change in curvature occurs , and @xmath8 approximates very closely the @xmath113 behavior expected from the single remaining electron pocket ( shown as the dotted line ) .
fig [ fig : hall ] also shows , as stars and heavy squares , the experimental results .
@xcite these data are in qualitative agreement with calculation , but important quantitative differences occur . at the lowest doping ( nominal carrier concentration @xmath114 for stars @xcite and @xmath115 for heavy squares @xcite ) the measured @xmath8 is substantially below both the calculated value and the @xmath113 curve .
it seems to us essentially impossible to obtain a hall resistance satisfying @xmath116 ; we therefore suggest that the nominal doping level , @xmath4 , differs from the actual doping .
the presently available data do not allow us to locate the point @xmath36 with precision , but strongly suggest that @xmath117 .
in this section , we study in more detail the behavior of hall conductance @xmath94 in the vicinity of @xmath111 .
when the system is tuned through the critical point ( which we believe to be @xmath118 in the electron - doped cuprates ) , the coherent backscattering introduces reconstruction of fermi surface , changes its local curvature , and thus affects the hall conductance .
the weak field limit was previously studied by bazaliy _
et al _ @xcite in the context of the 3d sdw compound @xmath119 .
they obtained a result equivalent to our eq [ [ weakfieldsigmaxx ] ] and a qualitative result equivalent to the weak field limit of our eq [ [ scalingfunction ] ] . and
the strong field limit was studied by fenton and schofield , @xcite who found a result equivalent to our eqs [ [ strongfieldresult],[scalingrelaxation ] ] .
the two dimensional nature of the present problem allows us to carry the analysis further , obtaining prefactors for @xmath94 in the weak field limit and a treatment of the weak to strong field crossover .
the detailed analysis of this equation is somewhat invovled , except for the strong magnetic field limit to be discussed below , partly because it is not easy to obtain an analytic relationship between the arclength coordinates @xmath48 and momentum coordinate-*p * for sdw fermi line .
we shall argue that this @xmath126 has a scaling form : @xmath156 such that @xmath157 and @xmath158 , with the prefactor @xmath159 of order @xmath149 , which will be computed in the special case of constant relaxation time in the following , and the exact form of @xmath160 depending on the manner in which the symmetry between @xmath65 and @xmath66 branches is broken .
it has been noted by bazaliy _
et al _ @xcite that in the vicinity of the crossing points , @xmath161 can be parametrized by @xmath162 and @xmath163 : first expanding the dispersions @xmath164 where , @xmath165 , @xmath166 , @xmath167 , @xmath168 , and the superscript @xmath169 stands for the corresponding value evaluated at the crossing point , and then inverting them @xmath170 with @xmath171 } { ( { \bf v}^{p\star}\times{\bf v}^{sp\star})^2}$ ] and @xmath172 } { ( { \bf v}^{p\star}\times{\bf v}^{sp\star})^2}$ ] . from eq [ [ inverse ] ] : @xmath173 where the second equality of eqs [ [ inverse],[dp ] ] only holds on the sdw fermi surface .
thus , @xmath174 and the sdw fermi velocities are : @xmath175 now return to eq [ [ criticaldeltasig ] ] .
@xmath176 , where @xmath177 is the relaxation time at crossing point , @xmath178 is the sdw fermi velocity when mirror symmetry is present , in which case @xmath179 and @xmath180 in eqs [ [ convertds],[sdwvelocity ] ] are constant vectors . substituting this and eq [ [ convertds ] ] into eq [ [ criticaldeltasig ] ] ,
scaling energy variables in unit of @xmath22 , and defining @xmath181 , where @xmath182 and @xmath183 is the angle made by the normal state fermi velocities , @xmath179 and @xmath180 , shown in fig [ fig : localview ] , we can obtain the above scaling form eq [ [ scalingfunction ] ] . in the following
, we shall discuss eq [ [ scalingfunction ] ] in two limits : weak magnetic field , @xmath184 , and strong magnetic field , @xmath185 . in weak magnetic field
, we can start from eq [ [ ongformula ] ] .
both @xmath179 and @xmath132 can vary in momentum space .
to lowest order , these effects are additive .
if we keep the * p*-dependence of relaxation time @xmath132 , assuming its dependence has form : @xmath186 \label{relaxationfunction}\ ] ] and keep velocity of normal state , @xmath187 and @xmath180 , to be constant , eq [ [ ongformula ] ] can be evaluated for each fermi pocket , giving @xmath188 \label{weakfieldtau}\ ] ] in the case in which the normal state velocities @xmath187 and @xmath180 are * p*-dependent and relaxation time @xmath132 is constant , differentiating eq [ [ expansion ] ] , substituting eq [ [ inverse ] ] into it , and defining vectors : @xmath189 we then have @xmath190 and @xmath191 , ( @xmath192 ) .
eq [ [ ongformula ] ] then gives : @xmath193 \label{weakfieldcurvature}\ ] ] these weak magnetic field results , eqs [ [ weakfieldtau],[weakfieldcurvature ] ] , confirm our scaling form for @xmath139 , eq [ [ scalingfunction ] ] , in the limit @xmath194 . also the prefactor @xmath159 can be calculated from these two equations .
it would be of interest to study eq [ [ weakfieldcurvature ] ] in more detail , since all the parameters in @xmath195 are set by normal state energy function eq [ [ dispersion ] ] .
taking @xmath196 from numerical calculation , @xmath197 , shown as solid line in the upper left panel of fig [ fig : linearbehavior ] , which is compared with the results of numerical evaluation of eq [ [ ongformula ] ] ( solid circles in this panel ) . in weak magnetic field , the longitudinal conductivity @xmath198 has been treated in great detail for the case of the 3d sdw transition in @xmath119 .
@xcite following essentially the same procedure , we obtain a similar expression for @xmath199 : @xmath200 taking @xmath201 from numerical calculation , @xmath202 , shown as solid line in upper right panel of fig [ fig : linearbehavior ] , compared with the results taken from numerical evaluation of eq [ [ sigxxlow ] ] ( solid squares ) .
the numerical results of @xmath203 are shown as solid diamonds in the lower panels of fig [ fig : linearbehavior ] .
the reconnection of fermi surface suppresses both longitudinal and transverse conductivity . in the vicinity of criticality ,
since @xmath205 , we expect a square root behavior in @xmath206 .
there actually is , as shown in the lower right plot of fig [ fig : linearbehavior ] , however , this regime is so small that it may be difficult to observe in practice . as a function of @xmath22 near critical doping ( solid circle ) and function @xmath207 ( solid line ) .
upper right : numerical results for @xmath208 as function of @xmath22 ( solid square ) and function @xmath209 ( solid line ) .
lower left : numerical results for @xmath8 as a function of @xmath22 ( solid diamond ) and function @xmath210 ( solid line ) .
lower right : numerical results for @xmath8 as a function of @xmath4 near critical doping .
the parameters are : @xmath12ev , @xmath13 , @xmath211ev , and @xmath103=0.6\sqrt{1-x/0.16}$].__,width=288 ] in strong magnetic field regime @xmath212 , the integration eq [ [ criticaldeltasig ] ] is dominated by ( @xmath213 ) , with @xmath214 , @xmath215 , and @xmath216 , then @xmath217 \label{strongfieldresult}\ ] ] with @xmath218 and @xmath219 the mean free path in normal state . this result , @xmath220 and independent of @xmath22 , confirms our scaling function for @xmath212 . to justify this result ,
let s consider a simple case in which the normal state velocities @xmath135 are constant , while relaxation time is assumed to be of form : eq [ [ relaxationfunction ] ] . under this condition ,
the only contribution to @xmath126 in eq [ [ criticaldeltasig ] ] comes from the second line containing @xmath153 . changing to @xmath221 variable
, it gives @xmath222 @xmath223 is a function of only one variable , numerical evaluation of which gives the solid line in fig [ fig : vary_relaxationtime ] , along which shown as dashed line is the linear function @xmath224 , representing the calculated asymptotic behavior as @xmath225 .
the prefactor 2 is calculated by relating @xmath226 in eq [ [ strongfieldresult ] ] and @xmath227 in eq [ [ relaxationfunction ] ] in the limit @xmath76 : @xmath228 .
this solid line clearly shows the expected behavior of @xmath223 : @xmath229 for @xmath230 , and @xmath231 , with a crossover in between .
the crossover regime can be estimated by setting @xmath232 .
the relaxation time can be inferred from the _ ab _ plane resistivity measurement : @xcite at b=10 t , @xmath233 , which gives @xmath234 sec .
so , for @xmath235ev , corresponding to @xmath236 if we take @xmath237ev , the crossover is @xmath238 t , where @xmath239ev@xmath240 . for these parameters ,
the breakdown field is slightly larger . for @xmath139 under the linearized dispersion , constant velocity approximation .
the unit of @xmath241 is @xmath242.__,width=240 ]
in this section we examine the sensitivity of our results to different choices of parameters .
let us first consider the @xmath243 region .
fig [ fig : compareorderedhall ] shows numerical results of @xmath8 for parameter choices : @xmath12ev , @xmath13 , @xmath211ev and @xmath244ev ( dashed line ) , and @xmath13 , @xmath14 and @xmath245ev ( solid line ) , which can be compared with the solid and dashed lines in fig [ fig : hall ] .
all of these curves share the same feature : the existence of @xmath36 . in fig
[ fig : hall ] , @xmath246 , while in fig [ fig : compareorderedhall ] , @xmath247 ( dashed line ) and @xmath248 ( solid line ) . for @xmath249 , both electron pocket ( centered at @xmath31 ) and hole pockets ( centered at @xmath30 )
are present , and as doping decreases , both of them contribute to the decrease of @xmath250 , while for @xmath251 , the hole pockets vanish ( the @xmath252 band is filled ) , and thus only electron pocket contributes to @xmath250 , which leads to the slowing down of decrease in @xmath250 as decreasing doping . for @xmath251 , the behavior of @xmath250 is quite universal , approximately @xmath253 ( the hall coefficient for free electrons , with @xmath4 electron density ) , as shown in fig [ fig : hall ] . different choices of parameters change the actual value of @xmath36 .
the numerical amplitude of @xmath254 is approximately @xmath255 , by continuity .
there is still another possibility that if the opening of sdw gap is not fast enough , the hole pockets survive for @xmath4 down to 0 . in this case
, there is an upturn when @xmath4 is smaller than some value , with a much smaller relative change in the magnitude of @xmath8 , as shown in the inset of fig [ fig : compareorderedhall ] .
we do not believe this possibility is relevant to the electron doped cuprates .
ev , @xmath13 , @xmath211ev and @xmath244ev .
solid line : @xmath14 , @xmath13 , @xmath245ev .
inset : @xmath8 for @xmath256ev , with the same band parameters as the solid line in the main panel.__,width=240 ] in the regime @xmath10 , we can change band parameters @xmath257 and @xmath258 in eq [ [ dispersion ] ] , and calculate @xmath8 for these choices of parameters . fig [ fig : comparenormalhall ] shows the results for different choices of band parameters : @xmath13 and @xmath211ev ( heavy solid line ) ; @xmath259 and @xmath14 ( dashed line ) ; @xmath260 and @xmath14 ( dotted line ) .
the fermi surfaces corresponding to these choices of parameters at @xmath7 are shown in fig [ fig : comparenormalfs ] .
it can be seen that ( i ) different choices of band parameters do not lead to drastic difference in @xmath8 ; ( ii ) for each choice of band parameters , @xmath250 is quite insensitive to change in doping , in contrast to the experimental data , shown as solid circles ; ( iii ) our numerical @xmath8 are close to @xmath261 ( the hall coefficient for circular fermi surface , with @xmath262 hole density ) , shown as light solid line in fig [ fig : comparenormalhall ] .
both our numerical results and @xmath261 are not close to the experimental data of ref . .
the difference is largest at @xmath263 , the point farthest from criticality , where one would expect the difference to be smallest . and @xmath211ev ; dashed line : @xmath259 and @xmath14 ; dotted line : @xmath260 and @xmath14 ; light solid line : curve @xmath261 ; solid circles : experimental data .
@xcite__,width=240 ] .
solid line : @xmath13 and @xmath211ev ; dashed line : @xmath259 and @xmath14 ; dotted line : @xmath260 and @xmath14.__,width=240 ]
fig [ fig : comparenormalhall ] shows that the measured @xmath8 at @xmath10 shows a pronounced doping dependence which is incompatible with the mean field calculation at fixed band parameters .
fig [ fig : comparenormalhall ] also shows that a variation of the correct order of magnitude can be obtained if a doping dependence of the band parameters is assumed . in this section
we investigate whether this doping dependence can be viewed as a precursor of the spin density wave instability by calculating the effect of spin fluctuations on the fermi surface .
this is given by the real part of the self energy .
the leading approximation to this is : @xmath264 where @xmath265 and @xmath266 are electron and spin fluctuation matsubara green s functions , respectively . in our calculation
, we shall take : @xmath267 and @xmath268 , with @xmath269 , where @xmath270 is the energy scale characteristic of spin fluctuations , @xmath271 is coherent length , @xmath272 characterizes the distance to criticality , and @xmath273 .
this self energy for @xmath274 can be integrated immediately , at @xmath0 , if both @xmath275 ( @xmath276 is an energy cutoff ) and @xmath272 are small .
the result may be represented as @xmath277 where , @xmath278 is a dimensionless constant
. we can take @xmath279 as our new electronic energy dispersion relation , and integrate eqs [ [ sigxxlow],[ongformula ] ] numerically .
the hall coefficient @xmath8 as function of @xmath272 , for @xmath280 ( solid line ) and @xmath281 ( dashed line ) , is shown in fig [ fig : fsrenormalhall ] , where the parameters are taken as : @xmath12ev , @xmath282 , @xmath211ev , and @xmath283ev .
because the relationship between @xmath272 and doping has not yet established , in our numerical calculation , we keep the chemical potential fixed at the unrenormalized value for @xmath7 . fig [ fig : renormalizedfs ] shows fermi surfaces for the case @xmath280 , @xmath284 and the case @xmath281 , @xmath285 , as well as the original unrenormalized one . both of these fermi surfaces give @xmath8 about 25% less than the unrenormalized value .
when electronic energy dispersion is @xmath279 , for coupling constant @xmath280 ( solid line ) and @xmath281 ( dashed line ) .
other parameters are : @xmath12ev , @xmath282 , @xmath211ev , the energy cutoff @xmath283ev.__,width=240 ] ] : @xmath286 ( solid line ) ; @xmath280 and @xmath284 ( dashed line ) ; @xmath281 and @xmath285 ( dotted line ) .
the numerical @xmath8 corresponding to the dashed line is about 20% less than that corresponding to the solid line , and @xmath8 corresponding to dotted line is about 30% less than that of the solid line.__,height=240 ]
we have used mean field theory to calculate the changes in the hall resistivity of a two dimensional system which undergoes a spin density wave instability .
the model studied was chosen to be appropriate to the electron - doped cuprates .
the important physics is that as @xmath4 is decreased below a critical value @xmath6 , the fermi surface is rearranged ( fig [ fig : fermi surface ] ) , leading to changes in @xmath8 .
these changes were discussed , in the context of a theory of @xmath119 , by bazaliy _
et al _ , @xcite and the related discussions , in the context of the d - density wave model of cuprates , were given by ref . .
the new results presented here include consideration of a wide range of dopings and gap values ( fig [ fig : hall ] ) , a systematic treatment of all relevant effects ( sec .
[ sec : formalism ] ) , a discussion of the weak field - strong field crossover ( sec .
[ sec : weak field limit ] ) , a treatment of precursor effects ( sec . [ sec : normal state region ] ) , and detailed application of the model to the electron - doped cuprates ( fig [ fig : hall ] and conclusion ) .
we note , unfortunately , that the specifics of the critical behavior and weak to strong field crossover will be visible ( at presently accessible fields @xmath287 t ) only at samples doped very close ( @xmath288 ) of the critical doping , and at the present level of sample purity , these effects will be complicated by magnetic breakdown .
the data of ref .
suggest @xmath289 while the data of ref .
indicate @xmath290 and is most probably @xmath291 .
thus the gap apparently opens quite fast with doping .
however , important discrepancies remain between data and calculation . at @xmath251
our calculation shows @xmath292 ( as expected for a fermi surface with one electron pocket containing @xmath4 electrons ) .
the data of ref .
are qualitatively consistent with this trend , but are offset , suggesting that the actual doping is slightly lower than the nominal one .
the @xmath114 data point of ref .
is qualitatively inconsistent with theory . also , the magnitude reported @xcite of @xmath8 at @xmath10 is inconsistent with theory .
one expects that as @xmath4 is increased beyond @xmath6 , @xmath8 should revert to the band value , @xmath293 .
while the sign and order of magnitude are correct , the numerical value is too large ; there is no reasonable choice of band parameters which reproduces the magnitude to better than a factor of two .
our work suggests several important directions for future research . on the experimental side
, we suggest that our finding that @xmath292 for @xmath251 can be used to determine the actual doping of lightly doped samples ( @xmath294 ) .
we also note that experiments which pinned down the behavior in the @xmath295 regime and more precisely located @xmath36 would be very helpful .
finally , clarifying the behavior at @xmath10 is potentially of considerable interest , as it gives ( in principle ) evidence of the modification of the fermi suface by finite range sdw correlations . on the theoretical side ,
important directions include combining the present considerations with those of ref .
to obtain a theory of the nernst effect , and improving the theoretical treatment of the sdw precursor effects .
work on the ac hall effect is in progress .
_ acknowledgements _ we thank r. greene , y. dagan and m. r. norman for helpful discussions and a. j. schofield for drawing our attention to the strong - field crossover and the possibility of magnetic breakdown .
this work was supported by nsf - dmr-0431350 .
100 s. sachdev , rev .
mod . phys . * 75 * , 913 ( 2003 ) .
p. monthoux , a. v. balatsky , and d. pines , phys . rev
* 67 * , 3448 ( 1991 ) .
s. a. kivelson , i. p. bindloss , e. fradkin , v. oganesyan , j. m. tranquada , a. kapitulnik , and c. howald , rev .
mod . phys . * 75 * , 1201 ( 2003 ) . c. m. varma , phys .
lett . * 83 * , 3538 ( 1999 ) .
g. m. luke et .
al . , phys . rev . *
b 42 * , 7981 ( 1990 ) .
m. matsuda , y. endoh , k. yamada , h. kojima , i. tanaka , r. j. birgeneau , m. a. kastner , and g. shirane , phys .
b 45 * , 12548 ( 1992 ) .
p. k. mang , o. p. vajk , a. arvanitaki , j. w. lynn , and m. greven , phys .
* 93 * , 027002 ( 2004 ) .
m. r. norman , q. si , ya .
b. bazaliy , and r. ramazashvili , phys .
lett . * 90 * , 116601 ( 2003 ) .
s. chakravarty , c. nayak , s. tewari , and x. yang , phys .
. lett . * 89 * , 277003 ( 2002 ) .
v. oganesyan and i. ussishkin , phys . rev .
* b 70 * , 054503 ( 2004 ) .
y. dagan , m. m. qazilbash , c. p. hill , v. n. kulkarni , and r. l. greene , phys .
lett . * 92 * , 167001 ( 2004 ) .
y. onose , y. taguchi , k. ishizaka , and y. tokura , phys .
lett . * 87 * , 217001/1 - 4 ( 2001 ) .
r. g. chambers , proc .
( london ) * a238 * , 344 ( 1957 ) .
e. i. blount , phys . rev . * 126 * , 1636 ( 1962 ) .
n. p. ong , phys .
b 43 * , 193 ( 1991 ) .
b. bazaliy , r. ramazashvili , q. si , and m. r. norman , phys .
b 69 * , 144423 ( 2004 ) .
j. fenton and a. j. schofield , cond - mat/0507245 . | a mean field calculation of the @xmath0 limit of the hall conductance of electron - doped cuprates such as @xmath1 is presented .
the data are found to be qualitatively consistent with the reconstruction of the fermi surface expected upon density wave ordering .
the magnitude of the density wave gap is found to be large .
the hall resistance exhibits a nonanalyticity at the quantum critical point for density wave ordering , but the amplitude of the anomaly is found to be unobservably small . the quantum critical contribution to @xmath2
is determined .
quantitative discrepancies between calculation and data remain , suggesting that the experimental doping is not identical to the @xmath3 concentration @xmath4 . |
oxide systems possessing k@xmath0nif@xmath3 and related structures have been extensively studied owing to their interesting structural , magnetic , and electrical properties .
recently , it was discovered that sr@xmath0ruo@xmath3 , a noncuprate member of @xmath7o@xmath3 family , exhibits @xcite superconductivity at 0.93 k. this report has naturally provoked many researchers to study the similarity and the dissimilarity between the superconductivity in this compound and that in the well - known la@xmath0cuo@xmath3 with the same crystal structure , but a much higher @xmath8 [ ref . 2 ] .
the technique of substituting cu by impurity ions in la@xmath0cuo@xmath3 has been successfully used as a powerful probe to reveal the unconventional properties of this cuprate .
@xcite similarly , various attempts have been made with sr@xmath0ruo@xmath3 by replacing ru by ir [ ref .
4 ] , fe [ ref . 5 ] and
6 ] ions , in order to investigate the effect of cationic substitutions .
nonmagnetic impurities are known to induce local moments and suppress @xmath8 severely @xcite and therefore , replacement of ru@xmath9 ( 4@xmath10 ) ions by nonmagnetic ti@xmath9 ( 3@xmath11 ) ions is an interesting topic of study .
the synthesis of these doped ruthenates is relatively easy , since the ti@xmath9 ion has the same oxidation state , coordination number as well as very similar ionic radius when compared with ru@xmath9 .
in fact , the compound sr@xmath0tio@xmath3 is iso - structural with sr@xmath0ruo@xmath3 with very similar lattice parameters , making the substitution facile .
it has been reported @xcite that the ti@xmath9 impurity indeed suppresses the superconductivity and induces localized moments even for very low ti concentration , @xmath4 @xmath120.03 in sr@xmath0ru@xmath1ti@xmath2o@xmath3 . during the past few years
, a few reports have appeared @xcite investigating ti doped sr@xmath0ruo@xmath3 system ; all these studies focus on the limit of very low ti doping .
interestingly , sr@xmath0ru@xmath1ti@xmath2o@xmath3 is also expected to show a metal to insulator transition as a function of ti doping , because the two end members possess widely different electrical properties with sr@xmath0ruo@xmath3 being a paramagnetic metal and sr@xmath0tio@xmath3 being a band insulator . however , there are only few reports on the full series ( 0 @xmath13 1 , in sr@xmath0ru@xmath1ti@xmath2o@xmath3 ) of compounds @xcite investigating the transport properties , while no detailed electronic structure study has been reported to date .
these transport studies reveal that there is a strong anisotropy between the interlayer resistivity ( @xmath14 ) and in - plane resistivity ( @xmath15 ) , observed in single crystal sr@xmath0ruo@xmath3 .
@xcite @xmath14 of sr@xmath0ruo@xmath3 shows pure metallic conductivity below 130 k and nonmetallic @xmath16 / @xmath17 @xmath12 0 at higher temperatures .
however , the metal - nonmetal crossover temperature decreases with increasing @xmath4 ( ti concentration ) and for @xmath4=0.2 , @xmath14 never enters a metallic regime .
the in - plane resistivity remains metallic throughout the temperature range for sr@xmath0ruo@xmath3 , while with increasing @xmath4 , an insulating upturn is observed in the resistivity data . for @xmath4=0.2 sample
, such semiconducting - like behavior is found below 40 k [ ref .
moreover , oswald _
et al . _
@xcite reported a steep increase in the room temperature resistivity of polycrystalline sr@xmath0ru@xmath1ti@xmath2o@xmath3 compounds at around @xmath4 = 0.3 .
therefore , all these observations indicate the occurrence of a metal to insulator transition ( mit ) as a function of ti doping in this ruthenate series , the critical composition being close to @xmath4=0.2 . in order to understand the electronic structure and thereby the nature of the mit in this series ,
we have carried out a detailed high - energy spectroscopic study on sr@xmath0ru@xmath1ti@xmath2o@xmath3 for @xmath4=0.0 , 0.1 , 0.2 , 0.5 and 0.6 , employing x - ray photoelectron ( xp ) , ultraviolet photoelectron ( up ) , and bremsstrahlung isochromat ( bi ) spectroscopies ( s ) .
these experimental studies , supplemented by band structure calculations for sr@xmath0ruo@xmath3 and sr@xmath0tio@xmath3 , using an _ ab initio _ self - consistent linearized - muffin - tin orbital ( lmto ) method within the atomic sphere approximation ( asa ) , establish the importance of coulomb correlation effects in understanding metal insulator transition as a function of ti concentration , which can only be analyzed in terms of mott transition .
all electron spectroscopic studies reported here were carried out on polycrystalline samples which were synthesized by the conventional ceramic route .
stoichiometric quantities of high purity srco@xmath18 and the respective transitional metal oxides were thoroughly ground and the mixture was heated at 600 @xmath19c for 24 h to avoid ru evaporation , then at 930 @xmath19c for 24 h , then at 1050 @xmath19c for 24 h , and finally at 1200 @xmath19c for 36 h with intermittent grindings .
the product was then pressed into pellets and finally sintered at 1200 @xmath19c for 36 hours and cooled to room temperature in air .
the phase purity of each composition was checked by powder x - ray diffraction ( xrd ) technique .
a transport study of these compounds has already been published .
@xcite electron spectroscopic studies were carried out in a commercial spectrometer from vsw scientific instruments ltd .
, united kingdom , equipped with a monochromatized al @xmath20 photon source for xps , a he discharge lamp for ups , and an electron gun for bis .
the samples were mounted on copper stubs using ultrahigh vacuum ( uhv ) compatible resins , while the electrical contact of the sample with the instrument ground was maintained using uhv compatible ag paste .
the pressure of the experimental chamber was better than 5 @xmath21 10@xmath22 mbar during the experiment and the sample surface was cleaned _ in situ _ by mechanical scraping using an alumina file until the various core levels ( o 1@xmath23 and sr 3@xmath24 ) showed reproducible spectral features and also the intensity of adventitious c 1@xmath23 peak was reduced to a minimum .
all experiments were performed at room temperature .
the binding energy of all spectral features were calibrated to the instrument fermi level which was determined by recording the fermi edge region from a clean silver sample .
self - consistent lmto - asa calculations were performed for stoichiometric sr@xmath0ruo@xmath3 and sr@xmath0tio@xmath3 in the experimentally determined tetragonal structure with seven atoms per unit cell and with 128 @xmath25 points in the irreducible part of the brillouin zone .
the lattice parameters used for sr@xmath0ruo@xmath3 are @xmath26=@xmath27=3.8603 and @xmath28=12.729 ( ref .
13 ) and for sr@xmath0tio@xmath3 are @xmath26=@xmath27 = 3.88 and @xmath28=12.6 ( ref .
in fig . 1(a ) , we show the spectral region for ru 3@xmath29 and ti 2@xmath29 core levels for these compounds . for sr@xmath0ruo@xmath3 ( _ i.e _ , @xmath4=0 sample ) , we observe a doublet feature with peaks at 463.8 ev for ru 3@xmath30 and 486.4 ev for ru 3@xmath31 core level spectra with a spin - orbit splitting of 22.6 ev .
weak satellite features are also seen at around 476.5 and 499 ev , corresponding to main ru 3@xmath30 and 3@xmath31 features , respectively .
a new distinct peak appears next to the ru 3@xmath30 peak at 458 ev , gaining intensity monotonically with increasing ti doping in sr@xmath0ru@xmath1ti@xmath2o@xmath3 ; this is identified as the ti 2@xmath30 signal .
the spin - orbit splitting in 2@xmath29 level of tio@xmath0 [ ref.15 ] is around 5.7 ev , therefore , ti 2@xmath31 signal appears at around 463.7 ev binding energy , overlapping the ru 3@xmath30 signal . in order to investigate the ti 2@xmath29 core level spectral features without the overlapping ru 3@xmath29 core level signal , we used the ru 3@xmath29 signal from the sr@xmath0ruo@xmath3 ( @xmath4=0 ) sample as the reference spectrum for ru 3@xmath29 spectral shape for all the compounds .
the basic idea is to remove the ru 3@xmath29 contribution from the spectral region in fig .
1(a ) by subtracting the reference ru 3@xmath29 spectrum , obtained from the parent undoped sr@xmath0ruo@xmath3 , after a proper multiplicative normalization ; thus extracted ti 2@xmath29 spectra are shown for all @xmath4@xmath320 in fig .
resulting ti 2@xmath30 and 2@xmath31 peaks appear at 458 and 463.7 ev binding energies , with the expected spin - orbit splitting of 5.7 ev for a ti@xmath9 species .
moreover it is well known that ti@xmath9 oxides , as in srtio@xmath18 [ ref .
16 ] or tio@xmath0 [ refs .
15 , 16 ] , show weak satellite features , at about 13 ev away from the main peak in the ti 2@xmath29 spectral region , characteristic of the ti@xmath9 state . in the present case
, we can see similar weak satellite features at about 13 ev above the main peaks in the extracted ti 2@xmath29 spectra , [ marked by arrows in fig .
this confirms that ti is in the formally ti@xmath9 3@xmath33 state in this family of compounds , establishing this as a homovalent substituted ( ti@xmath9 in place of ru@xmath9 ) family of compounds .
it is important to note here that such homovalent substitutions leave the electronic configuration of ru sites essentially unchanged , in contrast to more usual hetrovalent substitutions ( e.g. , sr@xmath34 in place of la@xmath35 ) leading to a doping of charge carriers into the system . in fig .
2 , we show the calculated density of states ( dos ) of the two end members , namely , sr@xmath0ruo@xmath3 and sr@xmath0tio@xmath3 . in the upper panel [ fig .
2(a ) ] , we show the total dos along with the ru @xmath24 , o @xmath29 , sr @xmath24 , and sr @xmath23 partial dos for sr@xmath0ruo@xmath3
. the total dos has a finite value at @xmath36 , consistent with the metallic property of sr@xmath0ruo@xmath3 .
the occupied part of the dos is dominated by o @xmath29 and ru @xmath24 states .
significant amounts of o @xmath29 states extend down to about -7.8 ev with three distinct groups of features , namely , those between -3.5 and -7.8 ev , between -1.7 and -3 ev and very close to the fermi energy .
the ru @xmath24 states are found to spread over the complete valence band range down to -7.8 ev .
the feature near -6 ev is contributed by the bonding o @xmath29 states with ru @xmath24 states .
the most intense feature around -2.5 ev in the total dos corresponds to primarily o 2@xmath29 states nonbonding with respect to ru - o interactions and therefore , with a negligible contribution from ru states .
ru ions in sr@xmath0ruo@xmath3 remains in the low - spin configuration ( 4@xmath10 : @xmath37 ) because of a strong crystal field effect @xcite splitting the @xmath38 and the @xmath39 levels significantly .
therefore , it has been shown that the states at @xmath36 arises from the antibonding ru @xmath40(@xmath41 ) and o @xmath42 states having four electron occupancy in these triply degenerate bands .
@xcite the unoccupied dos between 0 and 3.5 ev is dominated by ru @xmath24 @xmath38-derived states , while the states around 7 ev is contributed mainly by sr @xmath24 related states .
all the features above 10 ev in the unoccupied dos are mainly contributed by sr @xmath23 and @xmath29 states along with o @xmath23 states . the calculated band structure for sr@xmath0tio@xmath3 is shown in fig .
it exhibits a clear gap between the occupied and the unoccupied states , in agreement with the observed insulating property of the compound , in contrast to the metallic state of sr@xmath0ruo@xmath3 . the fermi level in fig .
2(b ) has been aligned at the bottom of the conduction band .
the calculation indicates a band gap of 1.3 ev .
expectedly , the occupied part of the band is completely dominated by the o @xmath29 states with small contributions from sr @xmath24 and ti @xmath24 states , consistent with the idea of a formal ti@xmath9 ( @xmath11 ) valence state in the compound .
the ti @xmath24 level has some occupancy arising from the hybridization of ti @xmath24 states with the o @xmath29 states .
the feature centered at about -5 ev in the total dos is found to have significant contributions from such hybridized ti @xmath24 states along with the dominant o @xmath29 contributions .
the features around -3 ev are primarily from an admixture of o @xmath29 and sr @xmath24 states . the first part ( 0 to 2.7 ev ) of the unoccupied density of states
is completely dominated by ti @xmath24 states of @xmath39 symmetry , while the ti @xmath38 states appear between 2.5 and 7.5 ev , strongly admixed with a contribution from sr @xmath24 states . the features above 12 ev has contributions from sr @xmath23 , o @xmath23 , and ti @xmath29 states .
the experimental xps valence band spectra from all the samples are shown in fig .
the overall spectral features of the valence band are similar for all the compounds , though there are some systematic variations in the relative intensities as well as in spectral details of different features .
the xp spectra exhibit two broad features with peaks at about 1.5 and 5.7 ev binding energies , respectively .
the broad feature centered near 5.7 ev binding energy is easily seen to be due to primarily o @xmath29 bonding states for all compositions ; this is evident from the fact that both sr@xmath0ruo@xmath3 and sr@xmath0tio@xmath3 have o @xmath29 dos over this energy range ( see fig .
the spectral intensity of the 1.5 ev feature decreases monotonically with increasing ti doping .
an intensity decrease of certain parts over the occupied states is indeed to be expected in view of a progressive substitution of ru@xmath9 4@xmath10 state by ti@xmath9 3@xmath11 state with an increasing @xmath4 in this series . a comparison with the band structure results of sr@xmath0ruo@xmath3 [ see fig .
2(a ) ] reveals that the spectral feature with the peak at 1.5 ev is mainly contributed by o @xmath29 nonbonding states over the higher energy between 1.5 and 3.5 ev binding energy and the antibonding ru @xmath40-o @xmath42 states at a lower energy near the fermi level . in the case of sr@xmath0tio@xmath3 ,
on the other hand , a gap opens up near @xmath36 over the energy window where ru 4@xmath24 states are found in sr@xmath0ruo@xmath3 , due to the 3@xmath11 configuration of the ti@xmath9 in contrast to 4@xmath10 configuration of ru@xmath9 .
the o @xmath29 nonbonding contributions in sr@xmath0tio@xmath3 appear around 1.52 ev binding energy as also in the case of sr@xmath0ruo@xmath3 .
therefore , with increasing ti concentration , the ru - o antibonding states close to fermi energy are progressively removed , resulting in a decrease in the overall spectral intensity in the 02 ev binding energy region , as observed in the experimental spectra .
while such considerations provide an understanding of the overall decrease in the spectral intensity within the first 2 ev of the fermi energy with increasing @xmath4 , these fail to explain a more interesting , though subtle , aspect of the spectral change .
first we note that the spectral intensity at @xmath36 is substantial for sr@xmath0ruo@xmath3 , as is evident in fig .
3 , consistent with its metallic ground state .
it is also evident from the figure that there is negligible intensity at @xmath36 for @xmath4=0.6 , also consistent with its insulating ground state , thereby capturing the change in the electronic structure associated with the experimentally observed metal - insulator transition .
however , it is clear that the previous discussion indicating a decrease in the spectral intensity will invariably suggest a finite intensity at the @xmath36 , proportional to the ru content .
in other words , the naive description of an overall decrease of the ru 4@xmath24 associated spectral intensity and the associated changes in the electronic structure can not explain the absence of any spectral intensity at @xmath36 for any finite @xmath4 less than 1 and therefore , such an approach can not explain the metal - insulator transition at a fractional @xmath4 value in this series .
we note that the spectral behavior in the near @xmath36 region suggests the formation of a gap even at a finite ru concentration ; such a gap formation is well known in the context of insulating transition metal compounds , driven by coulomb interactions . however , a poor instrumental resolution in xps precludes any careful investigation of the gap formation as a function of @xmath4 in these spectra .
this issue will be discussed in greater detail at a later stage , making use of the higher resolution in up spectra .
however , we note here that the correlation effect in this ru 4@xmath24 band system has been shown @xcite to be unexpectedly strong and can not be ignored .
changes in the electronic structure near the fermi edge region are most directly responsible for the metal - insulator transition in any system and therefore , this spectral region is the most important one to understand the nature of the mit in this series of compounds .
we have carefully studied the narrow spectral region close to the fermi energy for all the samples , using he i ups , in view of its much better energy resolution .
the set of he i narrow scan spectra , emphasizing the changes in the electronic structure close to the fermi energy , is shown in fig .
all the spectra are normalized at 2.8 ev binding energy , because this intensity is almost completely derived from o 2@xmath29 nonbonding states which remain invariant with ti doping .
we observe nearly the same intensity for all the spectra at about 2 ev binding energy ; these spectral features are primarily contributed by nonbonding o @xmath29 states ( see fig .
2 ) providing an explanation for the near constancy of this spectral feature across the series .
the regular decrease in spectral intensity with increasing ti concentration in the 0 to 2 ev energy region is evident from the figure ; this is essentially due to the substitution of the ru@xmath9 @xmath10 with ti@xmath9 @xmath11 configuration with increasing @xmath4 , decreasing the total electron count . but more interestingly , we observe a very substantial gap ( @xmath430.10.2 ev ) opening at @xmath36 for @xmath440.2 compounds , thereby clearly signalling the metal - insulator transition , consistent with reports @xcite based on transport measurements .
we have also probed the evolution of the unoccupied states as a function of ti doping using bi spectroscopy .
the experimental bi spectra , shown in fig . 5 ,
exhibit three broad features with peaks at 1.8 , 10 , and 17 ev .
the spectra clearly exhibit finite dos at @xmath36 for @xmath4=0 and 0.1 compounds ; interestingly , the spectral intensity at @xmath36 vanishes for @xmath440.2 , thereby defining an energy gap and an insulating state , consistent with the valence band results , discussed above .
the feature extending from @xmath36 up to about 5 ev decreases in intensity systematically with increasing @xmath4 .
this can be understood by considering the cross section of different states .
it is known @xcite that the cross section of ru 4@xmath24 states is almost 100 times that of ti 3@xmath24 states at an excitation energy of 1486.6 ev , employed in the present study .
therefore , spectra in the 03 ev range are dominated by ru 4@xmath24 states , with only minor contributions from ti 3@xmath24 states ; this immediately explains the experimental observation of decreasing intensity of the leading spectral feature with increasing ti doping in place of ru .
in contrast to the clear and systematic change in the spectral feature close to @xmath36 , the higher energy features remain almost identical with changing ti concentration .
this is not surprising , as the band structure results indeed show that the features above 6 ev are dominated by o and sr states , which are essentially the same for both sr@xmath0ruo@xmath3 and sr@xmath0tio@xmath3 .
while all these spectroscopic results , establishing a metal - insulator transition , are consistent with transport properties , @xcite these spectroscopic results provide some critical inputs in understanding the possible mechanism responsible for this mit , clearly making several other alternative mechanisms untenable .
though the mechanism of the mit in this system has hardly been discussed in the literature so far , our results , for example , clearly suggest that the mit in this system can not be understood in terms of percolation or any other inhomogeneous model , with the assumption that the doping of ti in place of ru in sr@xmath0ruo@xmath3 does not result in homogeneously doped samples , but only some inhomogeneous physical mixtures of the end compositions , namely , sr@xmath0ruo@xmath3 and sr@xmath0tio@xmath3 .
according to such a description , increasing ti concentration should gradually increase the concentration of sr@xmath0tio@xmath3 at the cost of metallic sr@xmath0ruo@xmath3 and consequently , a gradual reduction in the intensity of the spectral feature at @xmath36 with increasing ti doping , is expected .
more importantly , the spectra for even the insulating range of compounds with @xmath440.2 are expected to exhibit finite spectral weight at @xmath36 , contributed by the large metallic fraction present in the inhomogeneous system representing a mixture of metallic and insulating phases .
evidently , this is not observed experimentally where a finite and substantial gap is found to appear in the spectral weight near @xmath36 with only 20% ti doping .
specifically , the changes in the lattice parameters between the end members [ for sr@xmath0ruo@xmath3 @xmath26=@xmath27=3.8603 and @xmath28=12.729 ( ref . 13 ) and for sr@xmath0tio@xmath3 @xmath26=@xmath27=3.88 and @xmath28=12.6 ( ref .
14 ) ] are not large enough to make any perceptible change in the various hopping strengths that contribute to the bandwidth .
therefore , the only significant change in the bandwidth will be due to the decrease in the average coordination number by 20% at the same level of doping .
since the bandwidth is linearly proportional to the coordination number within a tight binding model , we may expect a decrease of about 20% in the ru 4@xmath24 bandwidth from this homovalent substitution .
another possible mechanism for the mit is the disorder driven anderson transition arising from random replacement of ru ions by ti in the sr@xmath0ruo@xmath3 lattice .
it is indeed true that the extent of disorder is expected to increase with increasing ti concentration at least until the point of 50% doping .
such disorder driven metal - insulator transition is likely to be signalled by a finite spectral weight at @xmath36 , since the anderson - type transitions are characterized by finite , but localized , density of states at @xmath36 . in presence of long range coulomb interactions such
a system may even open up a soft gap , namely , coulomb gap .
@xcite electron spectroscopic investigations have indeed revealed @xcite a host of interesting spectroscopic features associated with disorder effects in substituted transition metal compounds . however , the experimental observation of a substantial hard gap in the present case is clearly not compatible with ideas concerning a disorder - driven mit .
the electronic configuration of ru@xmath9 ( 4@xmath10 ) ion in sr@xmath0ruo@xmath3 is low - spin .
while a strong crystal field would split the @xmath24 levels into @xmath39 and @xmath38 states , the jahn - teller distortion in sr@xmath0ruo@xmath3 with an elongation of the ruo@xmath45 octahedra along the @xmath46-axis further splits the triply degenerate @xmath39 level , leading to a stabilization of the doubly degenerate @xmath47 , @xmath48 bands with respect to the @xmath49 band .
if this effect is very strong , the system could go into an insulating state with a fully filled doubly degenerate @xmath47 , @xmath48 band .
however the extent of distortion is not very severe in sr@xmath0ruo@xmath3 and the spread of the @xmath49 band is wide enough to overlap with the @xmath50 band .
@xcite as a consequence , the system remains metallic in spite of the modest jahn - teller distortion .
doping with ti would certainly narrow the ru bandwidth ; however , this effect is not expected to be large enough to remove the overlap of the @xmath49 band with the @xmath48 and @xmath47 bands altogether for as little as 20% ti doping .
it has been already reported @xcite that the electron correlation effect is not negligible in sr@xmath0ruo@xmath3 in spite of the broad ru 4@xmath24 band .
therefore , the physical properties of these compounds are controlled by the relative strengths of on - site coulomb interaction energy @xmath51 and the bandwidth @xmath52 . in the limit of @xmath51 being significantly less than @xmath52 , the system is metallic , while a progressive reduction in @xmath52 can drive a metal to insulator transition in a system with integral @xmath24 occupancy .
we believe that this system is another example of such a bandwidth controlled mit , in contrast to any band - filling controlled physics in spite of its obvious composition - driven change .
our core level study has shown that ti retains its stable 4 + oxidation state throughout the series ; as a consequence , the ru ions throughout the series has an integral filling with 4@xmath10 electronic configuration .
random replacement of ru ions by ti ions has two distinct effects on the electron motion .
effects of disorder in such a case will definitely be present ; however , as discussed above , it can not explain the mit .
the second , and more important , consequence of the substitution in the present context is to reduce the ru - ru coordination ; this evidently leads to decrease in the ru 4@xmath24 bandwidth .
such a progressive reduction in @xmath52 with increasing ti concentration may lead to the value of @xmath53 to cross the critical value for the metal - insulator transition leading to the insulating state with a finite mott - hubbard gap at the fermi energy .
this description of the mit appears to be the only one compatible with all the experimental observations presented here . in conclusion
, we have investigated the electronic structure of sr@xmath0ru@xmath1ti@xmath2o@xmath3 with @xmath4=0.0 , 0.1 , 0.2 , 0.5 , and 0.6 using x - ray and ultraviolet photoemission spectra and high - energy bremsstrahlung isochromat spectra in conjunction with _
_ _ initio _ band structure results .
core level spectra establish that substituted ti remains in the ti@xmath9 state throughout the series , ensuring a ru@xmath9 4@xmath10 with integral ru 4@xmath24 occupancy per ru site .
spectra from the occupied and unoccupied parts show the formation of a gap ( @xmath43200 mev ) at the fermi energy for samples with ti content , @xmath4 @xmath54 0.2 , establishing a metal - insulator transition . these available results can only be understood in terms of a bandwidth controlled metal - insulator transition as a function of the composition , arising from a decreasing ru - ru coordination with increasing ti content .
the authors acknowledge the department of science and technology , and the board of research in nuclear sciences , government of india , for financial support .
present address : materials and structures laboratory , tokyo institute of technology , 4259 nagatsuta , midori - ku , yokohama 226 - 8503 , japan . also at jawaharlal
nehru centre for advanced scientific research , bangalore , and center for condensed matter theory , indian institute of science , bangalore , india .
electronic address : [email protected] y. maeno , h. hashimoto , k. yoshida , s. nishizaki , t. fujita , j. g. bednorz , and f. lichtenberg , nature * 372 * , 532 ( 1994 ) .
j. g. bednorz and k. a. muller , z. phys .
b : condens . matter * 64 * , 189 ( 1986 ) .
y. maeno , t. tomita , m. kyogoku .
s. awaji , y. aoki , k. hoshino , a. minami , and t. fujita , nature * 328 * , 512 ( 1987 ) .
r. j. cava , b. batlogg , k. kiyono , h. takagi , j. j. krajewski , w. f. peck jr .
, l. w. rupp jr . , and c. h. chen , phys .
b * 49 * , 11890 ( 1994 ) .
r. greatrex , m. m. greenwood , and m. lala , mater .
bull . * 15 * , 113 ( 1980 ) .
h. r. oswald , s. felder , e. casagrande , and a. reller , solid state ionics * 6365 * , 565 ( 1993 ) .
g. xiao , m. z. cieplak , j. q. xiao and c. l. chien , phys .
b * 42 * , r8752 ( 1990 ) ; k. ishida , y. kitaoka , k. yamazoe , k. asayama , and y. yamada , phys . rev .
lett * 76*. 531 ( 1996 ) ; m .- h .
julien , t. feher , m. horvatic , c. berthier , o. n. bakharev , p. segransan , g. collin , and j .- f .
marucco , _ ibid _ * 84 * , 3422 ( 2000 ) .
m. minakata and y. maeno , phys .
b * 63 * , 180504 ( r ) ( 2001 ) .
k. pucher , j. hemberger , f. mayr , v. fritsch , a. loidl , e .- w .
scheidt , s. klimm , r. horny , s. horn , s. g. ebbinghaus , a. reller , and r. j. cava , phys .
b * 65 * , 104523 ( 2002 ) .
m. braden , o. friedt , y. sidis , p. bourges , m. minakata , and y. maeno , phys .
lett . * 88 * , 197002 ( 2002 ) .
n. kikugawa and y. maeno , phys .
lett . * 89 * , 117001 ( 2002 ) .
k. chandrasekaran , r. vijayaraghavan , and u. v. varadaraju , mater .
phys . * 56 * , 63 ( 1998 ) .
t. oguchi , phys .
b * 51 * , 1385 ( 1995 ) . w. wong - ng , h. mcmurdie , b. paretzkin , a. kuchinski , and a. dragoo , powder diff . * 3 * , 179 ( 1988 ) . c. n.r .
rao , d. d. sarma , s. vasudevan , and m. s. hegde , proc .
royal soc .
( london ) , ser .
a * 367 * , 239 ( 1979 ) .
a. e. bocquet , t. mizokawa , k. morikawa , a. fujimori , s. r. barman , k. maiti , d. d. sarma , y. tokura , and m. onoda , phys .
b * 53 * , 1161 ( 1996 ) .
s. nakatsuji and y. maeno , phys .
b * 62 * , 6458 ( 2000 ) .
t. yokoya , a. chainani , t. takahashi , h. katayama - yoshida , m. kasai , y. tokura , n. shanthi , and d. d. sarma , phys .
b * 53 * , 8151 ( 1996 ) . j. j. yeh and i. lindau , at .
. data tables * 32 * , 1 ( 1985 ) .
m. pollak , discuss .
faraday soc .
* 50 * , 7 ( 1970 ) ; v. ambegaokar , b. i. halperin , and j. s. langer , phys .
b * 4 * , 2612 ( 1971 ) ; a. f. efros and b. i. shklovskii , j. phys .
c : solid state phys .
* 8 * , l49 ( 1975 ) ; a. l. efros , j. phys .
c : solid state phys .
* 9 * , 2021 ( 1976 ) ; j. h. davies , p. a. lee , and t. m. rice , phys . rev .
b * 29 * , 4260 ( 1984 ) ; b. i. shklovskii , and a. l. efros , sov .
.- semicond .
* 14 * , 487 ( 1980 ) .
d. d. sarma , s. r.barman , h. kajueter , and g. kotliar , europhys .
lett . * 36 * , 307 ( 1996 ) ; h. kajueter , g. kotliar , d. d. sarma , and s. r. barman , int . j. mod
b * 11 * , 3849 ( 1997 ) ; k. maiti , priya mahadevan , and d. d. sarma , phys . rev . lett . *
80 * , 2885 ( 1998 ) ; d. d. sarma , a. chainani , s.r .
krishnakumar , e. vescovo , c. carbone , w. eberhardt , o. rader , ch .
jung , ch .
hellwig , w. gudat , h. srikanth , and a.k .
raychaudhuri , _ ibid _ * 80 * , 4004 ( 1998 ) ; k. maiti , a. kumar , d.d .
sarma , e.weschke , and g. kaindl , phys . rev .
b * 70 * , 195112 ( 2004 ) .
fig . 1 : ( a ) ru 3@xmath29 spectral region with coincident ti 2@xmath29 contribution from the sr@xmath0ru@xmath1ti@xmath2o@xmath3 series .
( b ) ti 2@xmath29 spectra , extracted from the region shown in ( a ) , from the same . | we investigate the nature and origin of the metal - insulator transition in sr@xmath0ru@xmath1ti@xmath2o@xmath3 as a function of increasing ti content ( @xmath4 ) . employing detailed core , valence , and conduction band studies with x - ray and ultraviolet photoelectron spectroscopies along with bremsstrahlung isochromat spectroscopy
, it is shown that a hard gap opens up for ti content @xmath50.2 , while compositions with @xmath60.2 exhibit finite intensity at the fermi energy .
this establishes that the metal - insulator transition in this homovalent substituted series of compounds is driven by coulomb interaction leading to the formation of a mott gap , in contrast to transitions driven by disorder effects or band filling .
pacs number(s ) : 71.30.+h , 79.60.-i , 71.20.-b , 71.28.+d |
radio polarization is generally considered as an important parameter for measuring the local magnetic field , as well as for determining the intrinsic mode of electromagnetic emissions from the local plasmas .
the spatial resolvable data of stokes i and v components are provided by a series of solar radio instruments in different frequencies , such as nobeyama radioheliograph ( norh ) , nancy radio heliograph ( nrh ) , siberian solar radio telescope ( ssrt ) , owens valley solar arrays ( ovsa ) , and etc .
especially , high - quality images of stokes i and v components at 17 ghz are continuously recorded from 1992 up to date in norh , and intensively studied by solar astronomers in all over the world .
but only limited authors paid particular attention to the radio polarization observed by norh . in five events ( kundu _ et al .
_ 2001 ) , the norh images show direct evidence that the radio sources are compact bipolar loops : source sizes are less than 5 " , and three of the five events studied show closely spaced oppositely polarized components in the circular polarization maps .
all five events are located directly over magnetic neutral lines in the photosphere , which is consistent with a single - loop scenario in which magnetic energy release and acceleration of nonthermal electrons are confined to a compact localized region .
the polarization is analyzed in four microwave bursts observed by norh with one looptop ( lt ) and two footpoint ( fp ) sources ( su and huang 2004 ) .
the three sources in each given burst are always polarized in the same sense .
this may be interpreted in terms of extraordinary mode ( x - mode ) emission , by taking into account the polarity of the underlying magnetic field and propagation effects , which may lead to inversion of the sense of polarization in the limbward fp and lt sources of the flaring loop .
huang and lin ( 2006 ) found a quasi - periodic reversal between left - circular polarization ( lcp ) and right - circular polarization ( rcp ) in the lt source , five minutes before the 2002 april 21 flare observed by norh , and the polarization gradually turned to lcp . during this period , the polarization of the corresponding fp source maintained the rcp sense . it may be taken as evidence that magnetic energy is released or energetic particles are produced at the magnetic reconnection site in a quasi - periodic fashion .
moreover , in the 2001 march 10 flare observed simultaneously by norh and ssrt , the emission at 17 ghz with lcp is generated in positive magnetic field polarity of the michelson doppler imager ( mdi ) photospheric magnetogram .
it is considered as an evidence of the ordinary mode ( o - mode ) emission and explained by strong anisotropy of the pitch angle distribution of the nonthermal electrons that responsible for the radio emission ( altyntzev _ et al .
_ 2008 ) . in a recent statistics of huang _
( 2010 ) , two footpoint emissions are compared in a total of 24 norh events with loop - like structures .
the polarization degrees in the two fps are well correlated , while , they mostly have the same sense ( rcp or lcp ) .
there is a comparable proportion of the normal and abnormal events , which are defined as whether a stronger emission corresponds to a stronger or weaker polarization in the two fp sources ( kundu _ et al .
_ 1995 ) . on the other hand , the observed polarization sense ( rcp or lcp )
may be reversed due to the linear mode coupling mechanism ( cohen 1960 ; melrose and robinson 1994 ; zheleznyakov _ et al .
_ 1996 ) , on the basis of plasma dispersion relation under approximations of locally homogeneous corona and geometrical optics , which is widely used to analyze the polarization of solar radio emission ( e.g. , peterova and akhmedov 1974 ) .
when the emission propagating along the line - of - sight is nearly perpendicular to the local magnetic field ( qt - region ) , the observed polarization sense is reversed in those frequencies below the transition frequency ( weakly coupled ) , defined by cohen ( 1960 ) .
otherwise , the sense is unchanged in those frequencies above the transition frequency ( strongly coupled ) . for two footpoint emissions with opposite magnetic polarities
, they should show the same polarization sense in the weakly coupled case , but opposite polarization senses in the strongly coupled case ( melrose and robinson 1994 ) .
however , it is not easy to judge if the observed emission does actually propagate through the qt - region , and the observations were not always compatible with that predicted by the theory ( mclean and sheridan 1972 ; white _ et al .
_ 1992 ) .
it is predicted by ramaty ( 1969 ) for the nonthermal gyrosynchrotron ( gs ) emission that the extraordinary mode ( x - mode ) is dominated in an optically - thin source , and the o - mode is only dominated in an optically - thick source .
one case to observe the optically - thin o - mode emission in microwave band is caused by the thermal gyroresonance absorption at the third harmonic ( preka - papadema and alissandrakis 1988 ) , the suppression of x - mode emission , originating in the lower layers of the loop , will produce excess o - mode radiation at the diskward foot of a flaring loop .
this requires that the magnetic field is strong enough to bring the third harmonic layer inside the flaring loop .
thus , alissandrakis _ et al . _
( 1993 ) have found evidence for the o - mode emission in two classes of events . in one class
the o - mode comes from the regions overlying the strong magnetic field , which can be interpreted in terms of the thermal gyroresonance absorption at the third harmonic . in the other class
the entire burst emits in the o - mode , which may be attributed to high gs optical depth .
the second case of optically - thin o - mode emission in microwave band was reported by ledenev _
( 2002 ) in the 2000 july 14 event , due to the scattering of electromagnetic radiation during propagation through a plasma layer with developed langmuir turbulence .
the ordinary component is slightly lowered , while the extraordinary component undergoes the most effective scattering .
the third case of optically - thin o - mode emission in microwave band was suggested by white et al .
( 1992 ) and gopalswamy _ et al . _
( 1994 ) , and the linear mode coupling may take place in the plasma current sheet with a transverse component of magnetic field .
it was also reported by vourlidas _
( 1997 ) that the o - mode emission in 4.7 ghz observed by vla in a solar active region , in contradiction with the gyroresonance models .
the excess of the o - mode emission is attributed to the magnetic field configuration and the temperature inhomogeneities across the spot , and it may originate from the hotter penumbral loops .
therefore , it is convenient to study the linear mode coupling and the intrinsic mode of electromagnetic emissions from the observed polarization sense and photospheric magnetic polarity .
however , it is difficult to measure the local magnetic field of radio sources with lower plasma density in solar corona than that in the photosphere .
for example , the radio emission at 17 ghz is generated at least ten thousand kilometers above the photosphere , and the photospheric field lines may turn to different directions in solar corona . section 2 shows one method to calculate the propagational angle of microwave emission between the line - of - sight and the local magnetic field , thus , to determine the linear mode coupling process and the intrinsic mode in solar microwave bursts .
section 3 analyzes one example observed by norh .
relevant discussions are given in section 4 .
there are two equations derived in huang ( 2006 ) for the coronal magnetic field strength ( @xmath1 ) , the angle ( @xmath2 ) between the the coronal magnetic field and the line - of - sight ( i.e. , eqs.(8 ) and ( 9 ) in huang 2006 ) , based on the nonthermal gs formulae in dulk and marsh ( 1982 ) , with an error of about 30% in respect to the full expressions of nonthermal gs emission , for a limited range of electron spectral index , harmonic number , propagation angle etc .
the detailed derivations and discussions are given in huang ( 2006 ) , and followed an earlier paper of zhou and karlick ( 1994 ) .
@xmath3 @xmath4 @xmath5 @xmath6 @xmath7 @xmath8 here , @xmath9 , the electron gyrofrequency @xmath10 .
all coefficients in eq.(3)-(6 ) depend on five observable values , i.e. , radio frequency @xmath11 , the turnover frequency @xmath12 , the brightness temperature @xmath13 at a given frequency @xmath11 , the polarization degree @xmath14 , and the electron spectral index @xmath15 , which is calculated by @xmath16 , here , @xmath17 is the photon spectral index in the optically - thin part ( dulk and marsh 1982 ) .
thus , it is easy to cancel out the terms with @xmath1 in eqs.(1 ) and ( 2 ) , and to obtain a new equation of @xmath9 as following .
@xmath18 @xmath19 the theoretical error of the approximations in dulk and marsh ( 1982 ) is better than @xmath20 in respect to the full expressions of nonthermal gs emission ( takakura and scalise 1970 ) , under the conditions of @xmath21 , @xmath22 , the propagation angle @xmath23 , the low cutoff energy @xmath24 , and the emission is dominated by the extraordinary mode , but the accuracy worsens at @xmath25 , especially , at extremes of @xmath2 and @xmath26 ( dulk 1985 ) .
now , we pay particular attention to the propagation angle @xmath2 in eq.(7 ) .
when @xmath27 , it means a positive magnetic polarity in respect to ordinary and extraordinary mode emission , together with lcp and rcp sense along the line - of - sight , respectively .
in contrast , when @xmath28 , it means that a negative magnetic polarity appears in the radio source , the emission with lcp sense along the line - of - sight will turn to the x - mode , and the emission with rcp sense along the line - of - sight will turn to the o - mode .
however , the polarization is always positive , and the angle @xmath2 is always smaller than 90 degrees in the approximated formulae of dulk and marsh ( 1982 ) . actually , the absolute degree of polarization must be the same at either side of @xmath29 , but the sense of polarization is opposite in the two cases , and the polarization sense changes its sign around @xmath29 according to the general plasma dispersion relation .
hence , we can not determine the exact direction of the longitudinal magnetic field with such a 180-degree ambiguity . with typical values of @xmath13 , @xmath14 , @xmath11 , @xmath12 , and @xmath17 in microwave bursts , when @xmath30 or @xmath31 , we find the left side of eq.(7 ) varies monotonously with @xmath32 from positive values to negative values , which means that a unique solution possibly exists for eq.(7 ) .
figure 1 shows the solution of eq.(7 ) with various values of @xmath13 , @xmath14 , @xmath12 , and @xmath17 , here , the radio frequency @xmath11 is fixed as 17 ghz for norh .
thus , we can actually minimize the value of the left side of eq.(7 ) to a small value ( e.g. , @xmath33 ) to estimate the solution of @xmath2 .
( solid ) , @xmath34 ( dashed ) , and @xmath35 ( dot - dashed ) .
right top panel : the solution of eq.(7 ) with different turnover frequencies of 4.0 ( solid ) , 5.0 ( dashed ) , and 6.0 ghz ( dot - dashed ) .
right bottom panel : the solution of eq.(7 ) with different photon spectral indices of -2.5 ( solid ) , -3.0 ( dashed ) , and -3.5 ( dot - dashed ) the other parameters are always fixed as the middle value in their varied ranges for all the panels.,width=8 ] in principle , we may determine the intrinsic mode and linear mode coupling in solar microwave bursts by three steps .
\1 ) at first , the mode coupling may happen only when the angle approximately equals to 90 degrees , i.e. , when the emission propagates through the quasi - transverse region ( cohen 1960 ) .
\2 ) secondly , the inversion of polarization sense due to the weakly mode coupling takes place only when the transition frequency defined by cohen ( 1960 ) is larger than the frequency of microwave emission , and an observable criterion of the weakly mode coupling in flaring loops is indicated by the same polarization sense in the two footpoints of a flaring loop ( melrose and and robinson 1994 ) .
\3 ) finally , the intrinsic mode of microwave emission is determined by the observed polarization and the calculated direction of local magnetic field according to the general plasma dispersion relation , together with the mode coupling process
. however , a 180-degree ambiguity still exists in the direction of longitudinal magnetic field , to produce an uncertainty of the intrinsic mode .
an m7.6 flare in active region noaa 9628 ( s21e01 ) start at 04:25:09 ut , peak at 04:34:00 ut , and end at 05:11:47 ut in 2001 september 25 , is selected from the flare list in the web of norh . the full sun observations of nobeyema radio polarimeters ( norp ) at 1 - 35 ghz in the same event ( figure 2 )
are used to discuss the radiation mechanism of 17 and 34 ghz observed by norh .
it includes light curve , polarization , spectrum and turnover frequency during the event .
there are different polarization senses in different frequencies , such as the lcp sense in 1 ghz , and rcp sense at 17 ghz , which is consistent with the polarization at 17 ghz observed by norh as shown in the left bottom panel of figure 4 .
the spectra at three selected times show the typical features of nonthermal gs emission , but we still ca nt exclude the plasma emission that possibly existed at 1 - 2 ghz ( i.e. , the cd type with second peak in decimeter band by nita _
the turnover frequency increased fast in the rising phase , then , it varied in 10 - 15 ghz . due to larger errors of peak frequency that fitted from only six frequencies of norp , we suppose that 35 ghz always belongs to optically thin , but 17 ghz is somehow close to the peak frequency in this event , which may cause an error in the calculated spectral indices of norh .
the left top panel of figure 3 shows a bright soft x - ray loop observed by yohkoh / sxt in this event , overlaid by the contours of the norh/17ghz brightness temperature ( solid ) and polarization degree ( dashed ) .
the soft x - ray emission in looptop ( lt ) is much stronger than almost symmetric emissions in two footpoints ( fp1 and fp2 ) , while , the microwave emissions in the two fps are very asymmetric , with a large difference of almost one order of magnitude during the burst , as shown in the left bottom panel of figure 3 . therefore , we calculate the propagational angle @xmath2 from eq.(7 ) in this event with the brightness temperature ( @xmath13 ) at 17 and 34 ghz , and the polarization degree ( @xmath14 ) at 17 ghz observed by norh , and the spectral index ( @xmath17 ) calculated with the particular software of norh at 17 and 34 ghz with different spatial resolutions , together with the peak frequency ( @xmath12 ) estimated by norp .
it is most interesting that the propagation angle @xmath2 in fp1 is much larger than that in fp2 , and overlaid on the photospheric magnetogram of mdi in this event ( see the right top panel of figure 3 ) . to confirm this result ,
we draw the time variation of the propagation angle @xmath2 in the two fps in the right bottom panel of figure 3 .
it is evident that @xmath2 reaches @xmath0 around the peak time in fp1 ( the variation of @xmath28 is marked by dotted line ) , but @xmath2 in fp2 mostly varies around @xmath36 .
the calculated propagation angles in the two fps just mean that there are opposite magnetic polarities in the two fps during the impulsive phase of the event , i.e. , negative in fp1 and positive in fp2 .
it is predicted that the polarization degree in the optically thin limit is inversely proportional to the propagation angle from figure 3d and eq.(16 ) of dulk and marsh ( 1982 ) , which seems to conflict with the results in figure 4 that shows stronger polarization and larger propagation angle in fp1 , while , weaker polarization and smaller propagation angle in fp2 .
the apparent discrepancy between the theoretical prediction and the results in figure 4 may be caused by the asymmetric magnetic fields in the two fps .
the averaged photosphere magnetic field in fp1 is about 5 - 10 times larger than that in fp2 , which may have stronger effect on the polarization degree than that of the propagation angle .
the weak polarization commonly corresponds to the small magnetic field , and the calculations in figure 3d of dulk and marsh ( 1982 ) should be obtained with the same magnetic field strength .
moreover , the reason why the apparent photospheric magnetic polarities in the two fps are opposite to the calculated coronal magnetic polarities may be not true due to the 180-degree ambiguity as mentioned above .
in the general plasma dispersion equation , the extraordinary mode ( x - mode ) emission along the line - of - sight corresponds to the observed right - circular polarization ( rcp ) emitted from a positive magnetic polarity , and the ordinary mode emission ( o - mode ) along the line - of - sight corresponds to the observed rcp emitted from a negative magnetic polarity .
note that the left - circular polarization ( lcp ) is detected as frequently as the rcp in solar radio observations ( see the recent statistics of huang and nakajima 2009 ) . in this case
, the x - mode emission along the line - of - sight may correspond to the observed lcp emitted from a negative magnetic polarity , and the o - mode emission along the line - of - sight corresponds to the observed lcp emitted from a positive magnetic polarity ( e.g. , altyntsev _ et al .
_ 2008 ) .
the key problem is that we may make mistake by simply using the photospheric magnetic polarities to determine the intrinsic mode of the electromagnetic waves detected by solar radio telescopes .
if we use the observed photospheric magnetic polarities to determine the intrinsic modes in the 2001 september 25 flare , the result seems to be opposite to that using the calculated longitudinal magnetic polarities in this event .
but , we still have a 180-degree ambiguity about the direction of longitudinal magnetic field in present case .
regarding the linear mode coupling , the selected event is very closed to the central meridian passage ( cmp ) , thus it is difficult to distinguish the diskward and limbward sources in the two fps , and to judge the emission from which fp source propagate through the qt - region .
we ca nt be simply neglect the mode coupling in the two fp sources nearby the cmp , which may be caused by a small deviation of the finite size of the fp sources from the cmp .
moreover , the observed polarization at 17 ghz in this event has same rcp sense in both two fps together with opposite magnetic polarities , which implies that the transition frequency should be larger than 17 ghz , i.e. , in the weakly coupled case ( melrose and robinson 1994 ) . the calculated angle between the local magnetic field and the line - of - sight in fp1 varied around @xmath0 at four times ( @xmath37 , @xmath38 , @xmath39 and @xmath40 ) of the impulsive phase ( figure 3 ) , which gives a direct evidence of the fp1 emission propagating through a qt region where the linear mode coupling took place . while , the calculated angle in another fp was always much smaller than @xmath0 , which implies that the fp2 emission does not propagate through the qt region .
on the other hand , the observed photospheric magnetic field in fp1 is about one order of magnitude larger than that in fp2 ( figure 3 ) , it is ready to compare the the transition frequency defined by cohen ( 1960 ) and the observed frequency ( 17 ghz ) of norh in both two fp sources , and it is confirmed that the mode coupling in fp1 actually belongs to the weakly coupled case .
moreover , the calculated angle between the line - of - sight and the local magnetic field becomes large with increasing of radio emissions , which just means an enhancement of the transverse magnetic field during the burst , and it may support the existence of the qt - region .
now , we try to understand the intrinsic mode of the two fp emissions in this event . from the norp data
as mentioned above , the 35 ghz emission is located in the optically - thin region , which belongs to the x - mode according to ramaty s theory ( 1969 ) , while , the 17 ghz emission is close to the peak frequency and it may belong to x - mode or o - mode .
if the angle @xmath27 or the longitudinal magnetic field is positive , when the 17 ghz emission in fp1 propagates through the qt - region , where the intrinsic mode with the rcp sense should be reversed from the x - mode to the o - mode in the two time intervals of the impulsive phase .
the reversion from the x - mode to the o - mode took place at @xmath37 and @xmath39 , and the reversion from the o - mode to the x - mode took place at @xmath38 and @xmath40 . while , the 17 ghz emission in fp2 does not propagate through the qt region , and the intrinsic mode in fp2 always belongs to the x - mode with the rcp sense . on the other hand , due to the 180-degree ambiguity
, the angle @xmath2 may equal to @xmath41 minus the calculated value .
thus , the intrinsic mode in fp1 is just reversed from the o - mode to the x - mode at @xmath37 and @xmath39 , and the reversion from the x - mode to the o - mode took place at @xmath38 and @xmath40 . while , the 17 ghz emission in fp2 always belongs to the o - mode .
\1 ) according to the theoretical prediction of melrose and robinson ( 1994 ) , for two fp emissions with opposite magnetic polarities ( such as the event in this paper ) , they should show the same polarization sense in the weakly coupled case , but opposite polarization senses in the strongly coupled case .
the problem is that the flare is at the central meridian , and there should be rough symmetry in the ray paths to earth .
hence , the first new result in this paper is just to give an evidence of the weakly mode coupling near by the central meridian , which is consistent with the theoretical prediction of melrose and robinson ( 1994 ) .
\2 ) secondly , we have derived a new equation of @xmath42 based on the earlier paper of huang ( 2006 ) , and shown the existence of the solution of this equation under some typical parameters of microwave bursts , which should be helpful for studying the model coupling and coronal magnetic field quantitatively . though the uncertainty of the sign of @xmath42 does exist , but it does not affect the prediction of the mode coupling around @xmath29 . regarding the problem of intrinsic modes , we may compare the extrapolation of photosphere magnetic field with the calculated coronal magnetic field , and determine the exact direction of coronal magnetic field , which we plan to do in a forthcoming paper . in this paper ,
if we believe that the magnetic polarities in photosphere and corona are commonly coincident , the intrinsic modes can be determined therein .
the remained problem that the authors could not solve so far is that the uncertainty of peak frequencies in different locations .
it is expected to study this problem when the new instruments ( such as fasr and chinese radio heliograph with multiple frequencies ) are completed in near future .
the study is supported by two nfsc projects respectively with no.11073058 and 10833007 , as well as the `` 973 '' program with no.2011cb811402 .
the authors would like to thank the norh teams for their dedicated data preparing and preprocessing . | an explicit equation of the propagational angle of microwave emission between the line - of - sight and the local magnetic field is newly derived based on the approximated formulae of nonthermal gyrosynchrotron emission ( dulk and marsh 1982 ) .
the existence of the solution of propagational angle is clearly shown under a series of typical parameters in solar microwave observations .
it could be used to determine the intrinsic mode and linear mode coupling in solar microwave bursts by three steps .
1 ) the mode coupling may happen only when the angle approximately equals to 90 degrees , i.e. , when the emission propagates through the quasi - transverse region ( cohen 1960 ) .
2 ) the inversion of polarization sense due to the weakly mode coupling takes place only when the transition frequency defined by cohen ( 1960 ) is larger than the frequency of microwave emission , and an observable criterion of the weakly mode coupling in flaring loops was indicated by the same polarization sense in the two footpoints of a flaring loop ( melrose and and robinson 1994 ) .
3 ) finally , the intrinsic mode of microwave emission is determined by the observed polarization and the calculated direction of local magnetic field according to the general plasma dispersion relation , together with the mode coupling process .
however , a 180-degree ambiguity still exists in the direction of longitudinal magnetic field , to produce an uncertainty of the intrinsic mode .
one example is selected to check the feasibility of the method in the 2001 september 25 event with a loop - like structure nearby the central meridian passage observed by nobeyama radio heliograph and polarimeters . the calculated angle in one footpoint ( fp ) varied around @xmath0 in two time intervals of the maximum phase , which gives a direct evidence of the emission propagating through a quasi - transverse region where the linear mode coupling took place , while , the angle in another fp was always smaller than @xmath0 where the mode coupling did not happen .
moreover , the right - circular sense at 17 ghz was always observed in both two fps during the event , which supports that the transition frequency should be larger than 17 ghz in the first fp together with strong magnetic field of over 2000 gauses in photosphere , where the weakly coupled case should happen .
moreover , there are two possibilities of the intrinsic mode in the two fps due to the 180-degree ambiguity .
1 ) the emission of extraordinary ( x ) mode from the first fp turns to the ordinary ( o ) mode in the two time intervals of the maximum phase , while , the x - mode is always emitted from the second fp .
2 ) the inversion from o - mode to x - mode takes place in the first fp , while the o - mode keeps in the second fp .
if the magnetic polarities in photosphere and corona are coincident in this event , the intrinsic mode belongs to the second case . |
by now , there have been a number of experiments which point to a flavor asymmetry in the sea of the proton @xcite .
although the exclusion principle @xcite and charge asymmetry @xcite may contribute to the observed asymmetries , they are not sufficiently large to explain the data @xcite .
the two most noteworthy experiments are the deviation of the gottfried sum rule from 1/3 , i.e. , @xmath1 @xcite and the recent drell - yan measurements of @xmath2 by e866 @xcite and na51 @xcite .
the cern measurement gives @xmath3 at @xmath4 and the fermilab one obtains the major part of the @xmath5-distribution of @xmath6 , finding @xmath7 at @xmath8 .
the most reasonable explanation of these large deviations from unity or 1/3 is that there is a flavor asymmetry in the sea quark distributions ; this flavor asymmetry is readily understood if a meson cloud surrounds the quarks , in that a proton can change into a neutron by emitting a @xmath9 , as first pointed out by thomas @xcite and later by henley and miller @xcite .
the sullivan process allows one to calculate the @xmath5-distribution of the asymmetry , as has been done by speth and collaborators @xcite and others ( see @xcite for references ) ; reasonable agreement with experiment is obtained .
+ more recently , alberg et al .
@xcite have pointed out that a test of the meson cloud model is feasible by carrying out drell - yan experiments with @xmath10 beams on protons and deuterium . in this model , the flavor asymmetry in the @xmath0 , for instance , is expected to be even larger than in the proton ( @xmath11 ) , so that @xmath2 should be much larger than 1 .
this is in contrast to the prediction of su(3 ) , under which @xmath12 by @xmath13 , from which it follows that @xmath14 in the @xmath0 .
+ in this paper we use the convolution method with the sullivan process to calculate the valence and sea quark distributions of the @xmath0 in the meson cloud model .
in the convolution method , the physical @xmath0 wave function is composed of the following fock states @xmath15 where @xmath16 is the probability amplitude to find a physical @xmath0 in a state consisting of a virtual baryon @xmath17 and a virtual meson @xmath18 with longitudinal momentum fractions @xmath19 and @xmath20 , and transverse momenta @xmath21 and @xmath22 , respectively .
the wave function renormalization factor @xmath23 is a measure of the probability of finding the `` bare '' @xmath0 , that is the @xmath0 without a meson cloud in the physical @xmath0 . + we assume that the @xmath24 will have components @xmath25 , + @xmath26 , @xmath27)$ ] and @xmath28 @xcite .
we neglect higher mass components .
+ in the infinite momentum frame ( imf , @xmath29 with @xmath30 the @xmath0 momentum ) the contribution of a certain fock state , @xmath31 , to the @xmath0 quark distribution can be written in terms of its quark components as @xmath32 where the splitting functions @xmath33 and @xmath34 are related to the probability amplitude @xmath35 in the imf via @xmath36 @xmath37 in terms of these splitting functions the wave function renormalization constant z is given by @xmath38^{-1}\equiv [ 1+\sum_{bm}\int_0 ^ 1f_{bm/ \sigma^+}(y)dy \;]^{-1 } , \ ] ] and the quark distribution functions @xmath39 of a @xmath0 within the fock state expansion are given as @xmath40 where @xmath41 is the quark distribution of the bare @xmath0 .
+ the next step is to calculate the splitting functions @xmath42 and @xmath43 .
we do this using time ordered perturbation theory ( topt ) in the imf , following the steps of reference @xcite . in topt in the imf
one can write the probability amplitudes @xmath44 explicitly as @xmath45 where @xmath46 is the invariant mass squared of the intermediate @xmath31 fock state @xmath47 and @xmath48 denotes the vertex function in the imf - limit .
vertices involving point - like particles automatically fulfill the symmetry relation @xmath49 but since hadrons have an extended structure one has to introduce phenomenological vertex form factors which parameterize the unknown microscopic effects .
therefore the vertex function @xmath50 is replaced by @xmath51@xmath50 and from equation ( 9 ) we get the restriction @xmath52 this is satisfied by the exponential form @xcite @xmath53 where @xmath54 is a cut - off parameter which we use in our further calculations .
+ equations ( 3 ) and ( 7 ) allow us to write the splitting functions as @xmath55 ^ 2}.\ ] ] one gets the spin averaged vertex functions in the imf from the interaction lagrangian density @xmath56 where @xmath57 denotes a baryon @xmath58 , @xmath59 , @xmath0 , @xmath60 and @xmath61 a pseudo scalar field @xmath62 , @xmath63 , @xmath64 @xcite . for the vertex @xmath65 the vertex function @xmath66
is given by @xmath67 @xmath68 and when spin averaged we obtain @xmath69 hence , in our case the splitting function is given by @xmath70 ^ 2}.\ ] ] using the following coupling constants g @xcite + @xmath71 = @xmath72 + @xmath73 @xmath74 + @xmath75 @xmath74 + @xmath76 @xmath77 + and the cutoff parameter @xmath78 gev for most of our work , @xcite we just need the quark distributions in the bare particles as an additional input . + for some of the following , we use holtmann s parameterization of the quark distribution function in the bare nucleon @xcite
in which he assumes a symmetric sea , @xmath79 @xmath80
@xmath81 @xmath82 @xmath83 in addition to this form for @xmath79 , we also take a form which is tied to the recent determination of the gluon distribution @xcite , @xmath84 since the gluon splits into the `` perturbative '' or bare sea quarks , this distribution should be close to that of those quarks . the advantage of this choice will become clear in the next section . for the quark distribution in the pion
@xcite we take @xmath85 @xmath86 where 20% of the pion s momentum is assumed to be carried by the symmetrical sea , which is presumed to be due to gluon splitting .
+ to begin with , we determined the quark distributions of the bare hyperons by using su(3 ) symmetry for the valence quarks ; that is we neglected the mass difference between the @xmath87 and @xmath88 and @xmath89 quarks .
we do not show the results because we do not believe that this is a realistic choice .
look at the @xmath90 .
experiments @xcite show that @xmath91 using this we can take the parametrization for the quark distribution in the pion to get the @xmath92 distribution in the @xmath93 and through charge independence the @xmath94 distribution in the @xmath90 . to get the s distribution in the @xmath90 we assume that the gluon and the light sea quark distributions in the kaon and the pion are the same and
hence carry the same momentum fraction in both particles .
we also assume the following form of the s quark distribution @xmath95 where the parameter @xmath96 is determined by matching up the right momentum fraction .
the valence quark distributions in the @xmath90 are then given by @xmath97 @xmath98 compared to the pion , where @xmath99 and @xmath100 look the same , the s distribution in the @xmath90 now peaks at higher @xmath5 than the @xmath94 distribution , which reflects the fact that the s quark is heavier than the u and the d. + our starting point for the bare @xmath0 is holtmann s parametrization for the bare nucleon , equations 18 - 21 .
again we assume only a change in the @xmath101-part of the parametrizations and that the light sea quarks and gluons in the bare @xmath0 and the bare nucleon look the same . to account for the higher mass of the s quark
we then make the ansatz @xmath102 where we take model dependent constituent quark masses to get @xmath103 the quark distributions in the bare @xmath0 are then @xmath104 @xmath105 @xmath106 we can also determine the distributions in the bare @xmath59 and @xmath107 via charge independence and su(3 ) : @xmath59 : = @xmath108 = @xmath107 : = @xmath109 + @xmath110 @xmath111 + @xmath112 @xmath113 + @xmath114 @xmath115 .
+ since we assume no change in the symmetric sea of the bare particles , the only variance in the antiquark distributions of the physical @xmath0 comes from the different parametrization of the @xmath90 which only affects @xmath116 .
there is also a change in @xmath99 due to the more realistic @xmath87 quark distributions for the @xmath59 and the @xmath107 .
the splitting @xmath117 is nearly negligible . hence taking into account
the larger s quark mass does not noticeably change our results for @xmath2 and the difference @xmath118 from those using su(3 ) parameters .
however , the valence quark distributions in the physical @xmath0 do change .
we first test our model for the measured sea quark distributions in the proton .
we show both @xmath119 and @xmath2 in our model with no @xmath120 and no mesons more massive than the pion in figs . 1 and 2 .
the comparison with the e866 data shows that @xmath119 agrees reasonably with experiment , but that the ratio @xmath2 does not turn over towards 1 at higher value of @xmath5 with the parameterization of the bare sea used by holtmann@xcite .
this problem is also found for other meson cloud and chiral models , as has been recently noted by peng et al.@xcite .
however , we find that the ratio does turn over ( although not sufficiently fast ) for @xmath121 given by eq . 22 .
it is true that the inclusion of the @xmath120 would help , but we believe that the splitting of the gluon into @xmath122 pairs is the dominant cause of the return of the ratio @xmath123 towards unity at @xmath124 .
we believe that a parameterization for @xmath79 can be found that is consistent with the gluonic data and with the ratio @xmath2 . for a further discussion , see also @xcite .
+ the four splitting functions for the @xmath0 are shown in fig .
3 ; it is clear that the contribution from the @xmath125 state is very small . in fig .
4 we show the calculated valence quark momentum distributions . the s quark distribution peaks at a slightly higher value of @xmath5 than that of the @xmath88 quark momentum distribution due to its larger mass . the momentum distribution of the sea quarks is shown in figs .
we show the distributions for both the holtmann @xcite and the gluonic @xcite bare @xmath0 distributions .
the @xmath126 and @xmath127 distributions are slightly different , but the difference is so small that we do not attempt to show it .
the difference @xmath128 shown in fig .
8 is , of course , independent of the bare @xmath0 s quark distribution , since it is due to gluon splitting .
it is interesting to see in fig .
9 how much @xmath129 distribution depends on that of the bare sea . in both cases considered the ratio is larger than @xmath130 in the proton . for the holtmann @xcite bare quark distribution @xmath131 increases with increasing @xmath5 and does not turn over . for the gluonic distribution , on the other hand , @xmath132 does not rise half as high and approaches 1 as @xmath133 .
we show @xmath134 in fig .
10 and compare @xmath131 to @xmath135 in fig . 11 . in the latter figure
we note that @xmath136 falls below 1 for the gluonic - like distribution at high @xmath5 , but the distribution functions are very small here .
+ we also tested the influence of the cut - off parameter @xmath54 on our results .
since a higher cut - off primarily makes the amplitudes of the splitting functions larger but does not change their shape very much , the sea quarks will carry more of the @xmath0 momentum while the momentum fraction carried by the valence s quarks stays the same and that of the valence u quarks gets smaller .
the ratio @xmath137 gets smaller and the difference @xmath138 gets larger with growing @xmath54 [ at all x ] .
in conclusion , we have used the meson cloud model to calculate the valence and sea quark distributions in the @xmath0 . the same model can of course
be used to calculate the valence and sea quark distributions in the @xmath139 and @xmath140 .
now that quark distributions in the proton have been well - determined , it is possible to use @xmath0 and @xmath140 beams and inclusive drell - yan reactions on protons , alone , to obtain similar information for the hyperons .
nuclear binding corrections for deuterium targets are thereby avoided .
for instance if @xmath141 is large and @xmath142 is small , the measurement @xmath143 is sensitive to @xmath119 in the @xmath0 .
we wish to thank j.c .
peng for fruitful discussions and for providing helpful information .
this work has been supported in part by the u. s. department of energy , contract no .
de - fg03 - 97er41014 , and by the national science foundation , grant no .
phy-97 - 22023 . | we use the meson cloud model and the sullivan mechanism to estimate the sea flavor asymmetry in the @xmath0 baryon and calculate the distribution functions of both sea and valence quarks .
we find large deviations from su(3 ) . |
it was almost ten years ago that the atmospheric neutrino anomaly@xcite could be explained by the oscillations of muon neutrinos into tau neutrinos @xcite .
this interpretation has been confirmed by the observation of the zenith angle dependence of the neutrino fluxes by the superkamiokande@xcite .
while the zenith angle dependence of the electron - like events is consistent with the theoretical predictions@xcite@xcite , that of the muon - like events disagree with the theory , especially for the up - going events . the neutrino state with flavor basis and that of mass basis is related to each other by the mixing matrix as follows , @xmath2 where @xmath3 the super - kamiokande shows that @xmath4 the equation ( [ eq : mu - tau_osc ] ) means that @xmath5 which is consistent with the choos results@xcite .
once the the matter enhanced neutrino oscillations was proposed @xcite , it has soon been applied to the atmospheric neutrinos crossing the earth@xcite .
recently much attention@xcite has been devoted to this subject .
however , it seems that no systmatic study has not been made yet , especially for the three flavor model of neutrinos .
the purpose of this paper is to reveal the conditions that amplify the effect of @xmath1 through matter oscillations so that we can determine this small mixing angle .
i assume that the masses of neutrinos are hierarchical , namely , @xmath6 @xmath7 , furthermore i assume that the solar neutrino oscillations are attributed to @xmath8 .
the atmospheric neutrino oscillations are derived by @xmath9 @xmath10 , the value of which is @xmath11 .
the solar neutrino problem may be solved by the matter enhanced , or just so oscillations . in either case
, @xmath12 is much smaller than @xmath13 , and irrelevant to our concerns . in the following , we assign @xmath14 , and @xmath15 .
the parameters contained in the model are @xmath16 , @xmath1 , and @xmath17 .
in calculating the matter effect of the earth , i use the density profile given by the `` preliminary reference earth model''(prem)@xcite .
the prominent feature of our model is that the survival probability of electron neutrinos @xmath18 is completely determined by the three parameters , _
i.e. _ , @xmath1 , @xmath0 where @xmath19 is the neutrino energy , and the zenith angle @xmath20 .
it is natural to define the resonance condition as @xmath21 the resonance states form curves in the the three parameter space @xmath22 @xmath23 corresponding to discretization of the resonance wavelength .
i solve eq.([eq : resonance ] ) numerically , and show in fig.1 and fig.2 the values of the three parameters at the resonance . fig.1 and
fig.2 correspond to the longest and the second longest wavelength , respectively .
the other probabilities @xmath24 , @xmath25 , @xmath26 , and @xmath27 @xmath28 also depend on the @xmath29 mixing angle @xmath16 .
i have investigated the resonance conditions of neutrinos passing through the earth in a wide range of parameters , although only a small mixing angle @xmath30 has a practical meaning as has been suggested by the choos experiments .
the results show that it is necessary to observe the high energy electron neutrinos crossing the depth of the earth to investgate the small mixing angle @xmath1 .
for example , the resonance occurs at @xmath31 , @xmath32 , and @xmath33 .
the energy of electron neutrino is @xmath34 if @xmath35 .
this indicates the need to assemble the data of the zenith angle distribution of the electron neutrinos whose energy is of order @xmath34 in order to determine the lepton mixing angle @xmath1 .
see for example , j. pantaleone , phys .
81 ( 1998 ) 5060 ; m.v .
chizhov and s.t .
petcov , phys .
83 ( 1999 ) 1096 ; e.kh .
akhmedov , a. dighe , p. lipari , and a.yu .
smirnov , nucl .
phys . b3542 ( 1986 ) 1454 ; o. yasuda , acta phys .
b30 ( 1999 ) 3089 ; t. ohlsson and h. snellman , phys .
b474 ( 2000 ) 153 ; i. mocioiu and r. shrock , hep - ph/0002149 ( 2000 ) . | the effect of matter enhanced neutrino oscillations on atmosheric neutrinos is investigated systematically in the framework of one mass dominant model of three neutrinos .
the resonance conditions of neutrino crossing the earth are determined by the three parameters , namely , the zenith angle , @xmath0 , and the mixing angle @xmath1 of the electron neutrinos with tau neutrinos .
the values of the triplet under the resonance is found numerically . |
it has long been known that cooling by molecular hydrogen , @xmath6 , plays a major role in regulating star formation in primordial protogalaxies @xcite .
the importance of @xmath6 stems from the fact that in most circumstances it is the dominant coolant in primordial gas at @xmath7 . as a result ,
the chemistry of @xmath6 in primordial gas has attracted considerable study @xcite and it is now generally accepted that only a small chemical network , of maybe 2030 reactions , is required to model @xmath6 chemistry over a very wide range of conditions in these systems . the introduction of metals into the gas , as will occur following enrichment of their surroundings by the first generation of supernovae , complicates matters enormously .
many other atomic and molecular coolants become available and the associated chemistry is highly complex : for instance , a reasonably complex model of purely gas - phase chemistry can easily stretch to @xmath8 reactants and almost 4000 reactions ( e.g. * ? ? ?
this would not matter were it not for two important points .
first , many astrophysicists believe that metal enrichment above a certain level the so - called ` critical metallicity ' leads to a significant change in the stellar initial mass function ( imf ) , from an imf dominated by massive stars to one that looks far more like the familiar salpeter imf of local star formation @xcite .
second , testing this idea numerically using three - dimensional hydrodynamical simulations requires us to model the chemistry of the gas , but the highly detailed chemical models mentioned above are impractical to use in high - resolution numerical simulations , owing to their high computational cost .
this is a consequence of the fact that chemical rate equations are frequently stiff and so for reasons of stability must be solved implicitly , with a computational cost that scales as the cube of the number of chemical species involved .
it is therefore important to look for ways to simplify the chemistry without unduly compromising the accuracy of the resulting model .
major simplifications can be made if we make the reasonable assumption that the main coolants in low metallicity , high redshift gas will be similar to those in the local interstellar medium which should be true provided that the abundance ratios of the various metals in low metallicity gas are not _ too _ unusual and also if we restrict the range of physical conditions in which we are interested . in this paper , we present a chemical model designed to model the chemistry of the major coolants in cool ( @xmath9 ) low - metallicity gas at low gas densities . in the absence of ultraviolet radiation , and in gas with @xmath2 , our model is valid for atomic hydrogen number densities @xmath3 , where @xmath4 is the size in myr of the characteristic physical timescale of interest in the problem , corresponding to an overdensity of @xmath10 with respect to the cosmological mean background density . at lower metallicities , or if a moderately strong uv field is present , our model remains accurate up to significantly higher densities .
we do not treat the cooling or chemistry of hot gas ( @xmath11 ) as this has already been treated in detail elsewhere ( see e.g. * ? ? ?
the structure of this paper is as follows . in [ chem ]
we present our simplified chemical network and discuss the principles determining our choice of reactants ( [ chem_choice ] ) and reactions ( [ react_select ] ) , as well as our treatment of grain surface chemistry ( [ surface_chem ] ) and photochemistry ( [ photochem ] ) . in
[ therm ] , we discuss our treatment of the main thermal processes included in our model , with a particular emphasis on atomic fine structure cooling .
we conclude in [ appl ] with a brief discussion of possible applications of our model .
our choice of which chemical species to include in our simplified network was guided by two main considerations . in order to properly model the thermal evolution of the gas
, we must be able to accurately model the evolution of the chemical abundances of all of the major coolants . at the same time , in order to keep our chemical treatment computationally efficient , we do not want to include more species than are strictly necessary . in some cases , the decision on whether or not to include a species was obvious .
for instance , neutral atomic hydrogen , @xmath12 , is a major constituent of the gas and is also a major coolant at @xmath13 and above .
molecular hydrogen , @xmath6 , also must be included as it has long been recognised to be the dominant coolant in metal - free gas at @xmath14 @xcite , and it is now clear that it also remains significant in metal - poor gas @xcite . a good case can also be made for the inclusion of @xmath15 , which dominates the cooling in zero - metallicity gas at @xmath16 @xcite , and which remains important at least up to metallicities @xmath17 @xcite . as far as metals go , however , the choice is less obvious .
metal atoms and ions produce little in the way of resonance line cooling at @xmath7 , but many can act as sources of fine structure line emission or emission from metastable transitions at these temperatures . therefore , in order to determine which of the various species are important coolants in the region of parameter space that our model is designed to cover , we directly compared the cooling rates produced by each species , under the assumption that the relative abundances of the various elements were the same as in the local interstellar medium .
the species we investigated in this comparison were the same as those included in the shock models of @xcite : @xmath18 , @xmath19 , @xmath20 , @xmath21 , @xmath22 , @xmath23 , @xmath24 , @xmath25 , @xmath26 , @xmath27 , @xmath28 , @xmath29 , @xmath30 , @xmath31 , @xmath32 , @xmath33 and @xmath34 .
this list includes all of the neutral or singly ionized species that have both non - negligible abundances and emission lines that are accessible at @xmath7 .
doubly ionized species are unlikely to be abundant in gas of this temperature , and so are unlikely to contribute significantly to the total cooling rate . atomic data for @xmath18 , @xmath19 , @xmath29 , @xmath35 and @xmath36 was taken from the sources listed in tables [ fs_data][fs_coll_rates ] .
for the remaining species , we used data taken from @xcite . for the relative abundances of c ,
n , ne , ni , o , s and si , we used values taken from @xcite , while for cl we used a value taken from @xcite .
we investigated three different scenarios for the ionization state of the gas , which we here refer to as the ` no ionization ' , ` moderate ionization ' and ` high ionization ' cases . in the no ionization case
, we assumed that the gas was completely neutral , and so set the electron abundance to zero , along with the abundances of all of the ionized species . in the moderate ionization case
, we assumed that the fractional ionization of hydrogen was such that @xmath37 , that species with ionization potentials lower than that of hydrogen c , cl , fe , ni , s and si were fully ionized , and that all of the other species remained fully neutral .
finally , in the high ionization case , we assumed that @xmath38 , that species with ionization potentials lower than that of hydrogen were again fully ionized , and that the fractional ionization of the other species was the same as that of hydrogen , i.e. 50% . in each case , we computed the cooling rate due to fine structure and/or metastable transitions from each of the listed species for a large number of temperatures in the interval @xmath39 and number densities in the interval @xmath40 . in these calculations , we assumed that the total metallicity of the gas was @xmath41 , and that the gas was optically thin in all of the relevant transitions .
we also calculated the contributions to the total cooling rate made by lyman-@xmath42 emission and compton cooling ; the rate of the latter was calculated for an assumed redshift @xmath43 .
we used the results of our comparison to select the set of major coolants that it was necessary to include in our thermal model by identifying all of the coolants that contributed more than 25% of the total cooling rate for any of the combinations of temperature , density and ionization that we examined . the resulting set consisted of @xmath18 , @xmath19 , @xmath29 , @xmath36 , compton cooling and lyman-@xmath42 cooling .
therefore , the only metals that we include in our chemical and thermal model are carbon , oxygen and silicon .
the importance of carbon and oxygen is unsurprising they are well known to play a major role in the cooling of the local atomic ism @xcite , and previous authors have also predicted that they will play a key role at high redshift ( see e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
silicon has attracted less attention ( although see @xcite ) , but @xmath36 proves to be the dominant coolant in highly ionized gas with @xmath44 at @xmath45 .
although gas with the very high fractional ionization assumed here will recombine quickly at temperatures @xmath7 , it nevertheless seems prudent to include silicon in our model .
moreover , since it is necessary to include neutral silicon in our chemical model if we are to calculate the @xmath36 abundance accurately , it requires little extra effort to include the effects of fine - structure cooling from .
our choice of a 25% cutoff in this analysis is somewhat arbitrary .
if we were to decrease the size of this cutoff , we would find that the number of species that must be included would increase , as both @xmath23 and s are important coolants at the 1020% level in portions of our parameter space .
however , as the cooling rates of the dominant coolants often have uncertainties that are comparable to or larger than the size of the contributions from these minor coolants , the accuracy we would gain by including them is less than might be expected , and does not ( in our opinion ) justify the additional complexity and computational expense that would be required in order to treat them . finally , it is clear that our conclusions here are sensitive to our choice of elemental abundance ratios .
for instance , an increase in the iron abundance ( relative to the other metals ) of 12 orders of magnitude would render fe and @xmath23 important coolants in portions of our parameter space , necessitating their inclusion in our model .
such an increase would be expected if the enrichment of the gas were dominated by pair - instability supernovae with masses close to the top end of the 140@xmath46 allowed mass range @xcite .
however , studies of the abundance ratios found in extremely metal - poor stars in the galactic halo , which at present give us the best picture available of the elemental composition of very metal - poor gas , do not find evidence for significant enrichment by pair - instability supernovae ( see e.g. the discussion in * ? ? ? * ) and so at the present time there is no compelling reason to include iron in our chemical model .
our chemical model therefore consists of the eight coolants discussed above
@xmath12 , @xmath6 , @xmath15 , @xmath18 , @xmath19 , @xmath29 , @xmath35 and @xmath36 together with ten additional species that play key roles in determining one or more of the abundances of the coolants : @xmath47 , @xmath48 , @xmath49 , @xmath50 , @xmath51 , @xmath52 , @xmath53 , @xmath54 , @xmath55 and @xmath56 .
our rationale for including @xmath51 and @xmath52 in our model is that in the presence of a significant flux of hard uv photons , x - rays or cosmic rays , ionized helium can act as an important source of free electrons , and moreover can transfer charge to neutral carbon or silicon ( but not oxygen ) far more effectively than @xmath48 can
. we do not include minor primordial coolants such as @xmath57 or @xmath58 .
these are never important at low densities and rarely important at high densities ( see @xcite , @xcite ) .
more significantly , we do not include molecular coolants such as co or @xmath59 , assuming instead that the bulk of the carbon and oxygen in the gas remains in atomic or ionized form .
this assumption dramatically simplifies the chemical modelling of the gas , but at the same time restricts the range of physical conditions over which the resulting model is useful . to assess the conditions for which this approximation is justified , we need to know two things .
first , what fraction of the available carbon and oxygen must be locked up in co or @xmath59 in order for cooling from these molecules to dominate over fine structure cooling ?
second , under what conditions are these fractions achievable within a dynamically interesting timescale ? to answer the first of these questions , we have performed calculations using the treatment of carbon and oxygen fine structure cooling discussed in
[ fscool ] below , together with a table - based treatment of co and @xmath59 taken from @xcite and @xcite .
we have computed @xmath60 , defined as @xmath61 where @xmath62 is the cooling rate per unit volume due to species @xmath63 , for a wide range of temperatures and densities .
we assume that @xmath64 and that all of the cooling occurs in the optically thin regime .
we adopt a nominal redshift @xmath43 and consider only temperatures @xmath65 , under the assumption that heating from the cmb will prevent the gas from cooling appreciably below this temperature . in figure
[ fmol ] , we show how @xmath60 varies as a function of temperature and density .
figure [ fmol ] demonstrates that over most of the parameter space that we have examined , @xmath60 is of order unity .
we find that @xmath66 only for temperatures very close to the cmb temperature ( which is likely an numerical artifact , a result of the fact that our treatment of fine structure cooling includes the effects of radiative pumping by the cmb , while our treatment of co and @xmath59 cooling does not ) , and at high densities , where the level populations of the fine structure coolants start to reach their local thermodynamic equilibrium values , causing the fine structure cooling rate per atom to saturate . for co or @xmath59
cooling to be effective , we therefore require that about as many carbon and oxygen atoms be incorporated into molecules as remain in atomic form . with these values of @xmath60 in hand , we can now turn to the question of whether it is possible to produce enough co and @xmath59 in the gas within an interesting timescale .
since we require a significant amount of oxygen to be in the form of coolant molecules , the timescale of interest , @xmath67 , is given approximately by @xmath68 where @xmath69 is the net rate of formation of coolants per unit volume .
therefore , to estimate @xmath67 we must first estimate @xmath69 .
although a complete discussion of the formation and destruction mechanisms of co and @xmath59 in low metallicity gas is beyond the scope of this paper , we will briefly summarize the most important points . in hot gas ( @xmath70 ) , most co and water molecules form via reaction pathways initiated by the hydroxyl radical
, oh ( see e.g. * ? ? ?
* ; * ? ? ?
this is formed by @xmath71 but most is then destroyed by @xmath72 however , a small fraction instead reacts to form other molecular species , such as water or co , e.g.@xmath73 the resulting @xmath59 and co molecules can be destroyed by collisions with atomic hydrogen : @xmath74 if , as will generally be the case in low - metallicity gas , @xmath75 ( where @xmath76 is the fractional abundance of @xmath6 relative to the total number of hydrogen nuclei ) , then the destruction of water by reaction [ water1 ] is far more effective than its formation by reaction [ ohd2 ] , and water will never account for more than a small fraction of the available oxygen . on the other hand ,
the co formed by reaction [ ohd3 ] can potentially account for almost all of the available oxygen or carbon whichever is present in the smaller amount as the destruction of co by reaction [ cod1 ] is ineffective at @xmath77 .
therefore , the net rate of formation of coolant molecules ( primarily @xmath78 ) is given approximately by @xmath79 where @xmath80 is the rate coefficient of reaction @xmath63 , @xmath81 and @xmath82 are the fractional abundances of atomic hydrogen and atomic carbon , and @xmath83 and @xmath84 are the number densities of @xmath29 and @xmath6 respectively .
the timescale to convert significant quantities of oxygen to co is then given approximately by @xmath85 for @xmath86 , which is a reasonable value for low metallicity gas if @xmath6 formation on dust is unimportant , this gives a timescale of approximately @xmath87 where @xmath88 is the number density of hydrogen nuclei , and where we have adopted the values for the various rate coefficients that are given in @xcite .
in cold gas , all of these reactions ( except for reaction [ ohd3 ] ) are ineffective , and other processes dominate the formation of water and co. a good summary of the relevant chemistry is given in @xcite . when @xmath89 , the most important mechanism involves the formation of the @xmath90 ion via the rapid ion - neutral reaction @xmath91
if @xmath92 , then most of the resulting @xmath90 ions simply dissociatively recombine : @xmath93 otherwise , they can then react further with @xmath6 to give @xmath94 and @xmath95 , with dissociative recombination of the latter producing @xmath96 and water .
co formation follows through reaction [ ohd3 ] .
all of these reactions occur rapidly , and so the net rate of formation of coolant molecules is given by the rate of formation of @xmath90 multiplied by the fraction of @xmath90 that is not destroyed by dissociative recombination , ( 1 - @xmath97 : @xmath98 for @xmath99 , we have @xmath100 , and hence @xmath101 . in that case , the time required to convert most of the oxygen to @xmath90 and thence to other molecules is @xmath102 where we have used the fact that @xmath103 owing to the rapid transfer of charge between oxygen and hydrogen ( see [ metal_chem ] below ) . if we assume that @xmath104 and that @xmath86 , then this gives us a value for @xmath67 of @xmath105 decreasing the fractional ionization of the gas will increase @xmath67 , but increasing it will have no significant effect .
finally , if the fractional ionization of the gas is too low for reaction [ ohpf ] to operate effectively , then the formation of coolant molecules again occurs primarily via hydroxyl , which in this case is formed mainly by direct radiative association @xcite @xmath106 in this case , @xmath67 is simply @xmath107 comparing the three values of @xmath67 derived above , we see that in gas with @xmath2 , @xmath108 , regardless of the gas temperature , provided that @xmath109 and that the fractional ionization satisfies the constraint given above .
if we compare this with the characteristic physical timescale of the problem of interest , @xmath4 , then it is simple to show that for densities @xmath110 co and @xmath59 will not form in quantities large enough to dominate the cooling , and hence it is valid to ignore these molecules and all of their associated chemistry .
note also that if we were to include the effects of photodissociation of @xmath96 , @xmath59 and co in the above analysis , then this would push the required density to an even larger value . in a gravitationally collapsing protogalaxy ,
a reasonable value for @xmath4 is the gravitational free - fall timescale @xmath111 , and in this scenario , our neglect of the molecular chemistry is valid as long as @xmath112 . on the other hand ,
if we want to simulate the thermal evolution of the interstellar medium in a metal - poor dwarf galaxy , a more reasonable timescale may be the sound - crossing time of the disk , which is of order @xmath113 for a @xmath114 disk and a sound speed of @xmath115 . in this case ,
our model is valid only for @xmath116 . as a consistency check on our conclusions here
, we examined the results of @xcite , who model the thermal and chemical evolution of freely - falling gas at a range of metallicities far below solar , using a detailed treatment of the gas chemistry .
they find that for @xmath117 , significant conversion of carbon and oxygen to molecular form does not occur until @xmath118 , in line with the value derived here . at lower metallicities ,
an even higher gas density is required . as another check ,
we have computed the evolution of the @xmath78 and @xmath59 abundances in the gas at the center of several of the simulated protogalactic halos discussed in @xcite ( 2007 ; hereafter paper ii ) , using values for the density , temperature , @xmath6 abundance and @xmath48 abundance taken from our simulations , and modelling the chemistry with the full umist99 chemical network @xcite .
we find that in most of these runs , our neglect of the molecular coolants is justified , as their abundances never become large enough for them to significantly affect the cooling .
our approximation begins to break down in the runs with @xmath2 , where about 10%20% of the total carbon and oxygen are incorporated into co and @xmath59 , which is just enough to affect the cooling at high temperatures and/or high densities .
however , only in our runs with @xmath119 does it break down completely .
therefore , the use of our highly simplified chemical model would appear to be justified in low - density gas with @xmath120 .
we are thus left with a set of eighteen chemical species that must be modelled : @xmath47 , @xmath48 , @xmath12 , @xmath49 , @xmath50 , @xmath6 , @xmath54 , @xmath53 , @xmath15 , @xmath51 , @xmath52 , @xmath18 , @xmath19 , @xmath29 , @xmath55 , @xmath35 , @xmath36 and @xmath56 .
the combined evolution of the abundances of these species is described by a chemical network consisting of 74 reactions : 47 collisional gas - phase reactions ( summarized in table [ tab : chem_gas_coll ] ) , 12 photochemical gas - phase reactions ( summarized in table [ tab : chem_gas_photo ] ) , 7 grain surface reactions ( summarized in table [ tab : chem_grain ] ) , and 8 reactions involving cosmic rays ( summarized in table [ tab : cosmic ] ) .
the abundances of these species are also constrained by seven conservation laws : @xmath121 where @xmath122 is the fractional abundance of chemical species @xmath63 relative to the total abundance of hydrogen nuclei , and where @xmath123 , @xmath124 , @xmath125 , @xmath126 , and @xmath127 are the total abundances of deuterium , helium , carbon , oxygen and silicon , respectively .
furthermore , we assume in our modelling that @xmath49 and @xmath50 are in chemical equilibrium , allowing us to write their abundances as : @xmath128 and @xmath129 where @xmath88 is the number density of hydrogen nuclei .
this assumption is generally justified in simulations of the cooling and gravitational collapse of gas in protogalactic halos , as the timescales on which @xmath49 and @xmath50 reach chemical equilibrium are much shorter than the cooling or free - fall timescale .
( for a more detailed discussion of this point , see @xcite ) .
the constraints represented by equations [ chg_cons][eq_h2p ] allow us to reduce the total number of chemical rate equations that must be solved to only nine . in practice
, we generally choose to solve for the abundances of the ionized species ( @xmath48 , @xmath54 , @xmath52 , @xmath19 , @xmath55 , @xmath36 , @xmath56 ) , @xmath6 and @xmath15 , but alternative choices are possible and would not significantly alter the results obtained .
the number of chemical reactions that could be included in our chemical network is very large , despite the limited number of chemical species involved .
fortunately , many of these reactions have little or no impact on the evolution of the abundances of our main coolants and so the number of reactions that need to be included in our chemical network remains reasonably small .
we can divide the reactions that must be included into two subsets .
the first subset consists of the reactions required to model the chemistry of hydrogen , helium and deuterium , including the formation and destruction of @xmath6 and @xmath15 ( reactions 129 , 4855 , 6063 , 6770 ) .
the second subset consists of the reactions required to model the carbon , oxygen and silicon chemistry ( reactions 3047 , 5659 , 6466 , 7174 ) .
the amount of @xmath48 present in the gas is controlled by seven main reactions : collisional ionization of @xmath12 by electrons ( reaction 11 ) , charge transfer with helium ( reactions 26 & 27 ) , photoionization ( reaction 48 ) , cosmic ray ionization ( reaction 67 ) , gas - phase recombination ( reaction 13 ) and recombination on the surface of dust grains ( reaction 61 ) . similar reactions ( nos . 12 , 28 , 29 , 49 , 68 , 14 & 62 ) partially determine the @xmath54 abundance , but in this case , charge transfer to and from hydrogen ( reactions 18 & 19 ) is also of great importance , owing to the very large abundance of hydrogen relative to deuterium . finally , the @xmath52 abundance is controlled primarily by collisional ionization ( reaction 24 ) ,
photoionization ( reaction 50 ) , cosmic ray ionization ( reaction 69 ) , gas - phase and grain - surface recombination ( reactions 25 & 63 ) and charge transfer with hydrogen ( reactions 26 & 27 ) .
the remaining 24 reactions in this subset control the formation and destruction of @xmath6 and @xmath15 .
@xmath6 forms in the gas phase via the intermediate ions @xmath49 and @xmath50 ( reactions 2 & 4 ) , as well as on the surface of dust grains ( reaction 60 ) .
it is destroyed by collisions with @xmath48 , @xmath47 , @xmath12 and @xmath6 ( reactions 710 ) , and can also be photodissociated or photoionized by uv radiation ( reactions 53 & 54 ) , or ionized by cosmic rays ( reaction 70 ) .
collisions with @xmath51 @xcite @xmath130 and @xmath52 @xcite @xmath131 can also destroy @xmath6 , but in general these processes are not as effective as collisions with hydrogen , and so they can be omitted from our simplified chemical model without significantly affecting its accuracy .
the @xmath49 and @xmath50 ions required for gas - phase @xmath6 formation are formed primarily by the radiative association of atomic hydrogen with free electrons or protons respectively ( reactions 1 & 3 ) , and while reactions 2 & 4 generally dominate the removal of @xmath49 and @xmath50 from the gas , in hot or highly ionized gas a number of other processes become competitive ( reactions 5 , 6 & 15 - 17 ) photodissocation of @xmath49 and @xmath50 ( reactions 51 & 52 ) can also become important if the incident radiation field is strong . finally , although @xmath15 can form from intermediate ions such as @xmath132 or @xmath133 in a manner analogous to @xmath6 ( see e.g. * ? ? ?
* ) , most actually forms from @xmath6 via reaction 20 : @xmath134 the resulting @xmath15 can be destroyed by the inverse of this reaction ( reaction 21 ) , or by photodissociation ( reaction 55 ) . in hot gas , the @xmath15 abundance is also influenced by reactions between @xmath6 and @xmath53 and @xmath15 and @xmath12 ( nos .
22 & 23 respectively ) .
note that with the exception of the grain surface reactions ( which are discussed in [ surface_chem ] below ) , the reactions required to model the hydrogen , helium and deuterium chemistry accurately in metal - enriched gas are just the same as those required to model primordial gas .
the values for many of these rate coefficients are known to within a small amount of uncertainty at the temperatures and densities of interest , and so our choice of the particular values used here should be uncontroversial .
however , a few of our assumptions demand further comment .
first , we note that the rates listed for reactions involving @xmath6 or @xmath15 as a reactant generally assume that these molecules are not vibrationally excited .
this assumption is reasonable at the low densities treated here , but breaks down at densities @xmath135 .
second , we note that for several reactions involving deuterium where no readily available rate coefficient exists in the astrophysical literature , we have assumed that the rate is the same as for the analogous reaction involving @xmath12 or @xmath48 .
the rates of processes involving cosmic rays ( see table [ tab : cosmic ] ) depend on the energy spectrum and energy density of cosmic rays .
these are poorly known at the relevant energies even in the local ism , and far less is known concerning high - redshift cosmic rays . for this reason
, we do not give absolute values for these rates , but instead parameterize them in terms of the cosmic ray ionization rate for atomic hydrogen , @xmath136 , which can then be considered an adjustable parameter of the model .
finally , we note that two of the reactions that regulate the @xmath49 abundance have large uncertainties in their rate coefficients .
the reactions in question are the associative detachment of @xmath49 with @xmath12 ( reaction 2 ) and the mutual neutralization of @xmath49 with @xmath48 ( reaction 5 ) .
as discussed in @xcite , the uncertainties in the rates of both of these reactions may be as large as an order of magnitude . in gas with a high fractional ionization , this uncertainty can lead to a significant uncertainty in the @xmath6 formation rate and in the final @xmath6 fractional abundance , particularly in the presence of a strong ultraviolet background radiation field . in the chemical model presented here
, we have followed @xcite and have adopted a rate coefficient for reaction 2 taken from @xcite and a rate coefficient for reaction 5 taken from @xcite .
however , we caution the reader that this should not be regarded as an endorsement of the accuracy of these particular values . as discussed in [ chem_choice ] above , we do not include the molecular chemistry of these elements in this simplified model .
our treatment of the carbon , oxygen and silicon chemistry is therefore purely a treatment of the charge balance of these species .
we begin with oxygen , in many respects the simplest of the three to treat .
the ionization potential of neutral oxygen is only 0.02 ev larger than that of neutral hydrogen , and so charge transfer between @xmath55 and @xmath12 or @xmath48 and @xmath29 ( reactions 36 & 37 ) occurs rapidly .
since the hydrogen abundance is orders of magnitude larger than the oxygen abundance , this means that the ratio of ionized to neutral oxygen is controlled by the ratio of ionized to neutral hydrogen , i.e. that @xmath137 in most circumstances , reactions 36 & 37 are the only reactions required in order to accurately model the oxygen chemistry . nevertheless ,
for completeness we also include several other processes : radiative recombination ( reaction 32 ) , grain surface recombination ( reaction 65 ) , collisional ionization ( reaction 35 ) , charge transfer with @xmath52 ( reaction 38 ) , photoionization ( reaction 57 ) and cosmic ray ionization ( reaction 72 ) .
in the case of carbon , the situation is rather different .
first , charge transfer from @xmath48 to @xmath18 ( reaction 39 ) is much less effective than charge transfer from @xmath48 to @xmath29 , and so reaction 39 plays a far less important role in the carbon chemistry than reaction 37 does in the oxygen chemistry .
second , carbon has an ionization potential of only 11.26 ev .
this means that charge transfer from @xmath19 to @xmath12 is significantly endothermic , rendering it unimportant at @xmath7 ( although we include it here for completeness ) .
it also means that neutral carbon can be photoionized by ultraviolet photons with wavelengths @xmath138 , which can penetrate easily into low metallicity protogalactic gas .
consequently , photoionization of @xmath18 ( reaction 56 ) plays an important role in the carbon chemistry , whereas photoionization of @xmath55 is unimportant outside of @xmath139 regions .
in addition to charge transfer and photoionization , ionized carbon can also be produced by collisional ionization ( reaction 33 ) , although this is important only for temperatures @xmath140 , by charge transfer with @xmath52 ( reaction 41 ) and by cosmic ray ionization ( reaction 71 ) .
@xmath19 is removed from the gas primarily by recombination in the gas phase ( reaction 30 ) and on grain surfaces ( reaction 64 ) , although if the free electron abundance is small , charge transfer from @xmath19 to @xmath35 ( reaction 44 ) can also become important . in the case of silicon , a similar set of processes operate : collisional ionization ( reaction 34 ) , cosmic ray ionization ( reaction 73 ) , gas - phase recombination ( reaction 31 ) , grain surface recombination ( reaction 66 ) , charge transfer from @xmath48 , @xmath52 or @xmath19 to @xmath35 ( reactions 42 , 43 & 44 ) , and photoionization by an ultraviolet radiation field ( reaction 58 ) , assuming one is present . in the case of silicon , however , we also include the doubly ionized ion , @xmath56 , in our chemical model , since this can be produced by charge transfer from @xmath48 to @xmath36 ( reaction 45 ) with an endothermicity of only a few ev , rendering it potentially important at temperatures @xmath141 .
it can also be produced by photoionization ( reaction 59 ) , although in this case the 16.35 ev energy requirement renders this process important only within regions , or by cosmic ray ionization ( reaction 74 ) .
the @xmath56 produced by these processes can be destroyed by charge transfer with neutral hydrogen ( reaction 46 ) , or by recombination ( reaction 47 ) . regarding the accuracy of our adopted reaction rates , we note that while many have well determined rate coefficients , there are two notable exceptions .
the only available rate coefficients for reactions 43 and 44 , which involve the transfer of charge from @xmath52 to @xmath35 and from @xmath19 to @xmath35 respectively , assume that the reactions both proceed at the langevin rate . in practice ,
highly exothermic charge transfer reactions often proceed at a rate far below the simple langevin rate , and so the rates of these reactions may be overestimated in our model .
if dust grains are present in the metal - enriched gas , then our chemical model must account for the effects of reactions occurring on the surface of the grains , as these reactions are often far more effective than their gas - phase counterparts .
we therefore include a small number of grain surface reactions in our model , as summarized in table [ tab : chem_grain ] .
the only grain surface reaction between neutral species that is included in our model is the formation of molecular hydrogen ( reaction 60 ) .
this is a hugely important reaction in the local interstellar medium , since at typical interstellar densities it is the only process capable of producing the large quantities of molecular hydrogen that are directly observed ( see e.g. * ? ? ? * ) or inferred from other molecular tracers , such as co. theoretical modeling @xcite suggests that it remains important down to metallicities of order @xmath142 , within the range of applicability for the chemical model presented here . the rate that we adopt for this process is based on the widely used rate of @xcite .
this was derived for solar metallicity gas , assuming a distribution of grain sizes as given in @xcite , and to adapt it for use in low metallicity gas , we simply assume that the rate scales linearly with the metallicity @xmath143 .
the validity of this assumption is open to question , as many of the features of the grain population , such as the grain size distribution or the mix of compositions , may differ greatly between milky way dust and protogalactic dust .
an alternative , physically motivated approach would be to adopt a grain size distribution and mix of compositions based on the results of numerical modeling of dust formation in high redshift supernovae @xcite and then to compute the @xmath6 formation rate expected for this grain population .
this is the approach used by @xcite .
however , the uncertainties associated with this approach are considerable . to begin with ,
the predictions of the numerical models are highly sensitive to the degree of mixing assumed to occur within the supernova ejecta @xcite . in addition , dust destruction in the reverse shock is typically not taken into account in these models , and to the best of our knowledge , the amount of dust that survives the passage of the shock has yet to be fully quantified ( although research in this area is actively proceeding ; r. schneider , priv .
finally , the changes wrought on the grain population by subsequent processing in the interstellar medium or intergalactic medium ( see e.g. * ? ? ?
* ) are not fully understood . in light of these uncertainties
, we do not believe that this approach currently offers much of an advantage over our simple assumption of an @xmath6 formation rate that scales with metallicity .
we do not include any neutral - neutral surface reactions in our model other than @xmath6 formation .
for neutral - neutral surface reactions to be able to significantly affect the ability of the gas to cool , they must be able to alter the abundances of neutral carbon , oxygen or silicon by a substantial amount .
we can place an upper limit on the rate at which these reactions occur if we assume that the reaction probability @xmath144 , i.e. that every collision between a metal atom and a grain results in a reaction . in that case , the reaction rate per unit volume for an atomic species @xmath63 is given by @xmath145 where @xmath146 is the thermal velocity of atomic species @xmath63 , and @xmath147 is the total surface area of grains per unit volume of gas . for milky way dust of the type assumed by @xcite , @xmath148 , where @xmath88 is the number density of hydrogen nuclei . at lower metallicity
, our assumption that the grain size distribution does not change with metallicity implies that the value of @xmath147 in gas with a metallicity @xmath149 is simply @xmath150 .
we can therefore rewrite equation [ grain_rate ] as @xmath151 where @xmath152 is the mass of species @xmath63 in atomic mass units .
the corresponding conversion timescale @xmath153 is then @xmath154 if @xmath155 , then this expression becomes : @xmath156 if we equate this to a characteristic physical timescale @xmath4 , then we can show that @xmath157 as long as @xmath158 for gas with a temperature @xmath159 and metallicity @xmath2 , and with atomic carbon as the colliding species , this gives : @xmath160 at higher temperatures , the limiting @xmath88 will be slightly smaller , but at @xmath161 , collisional dissociation of most molecular species is highly effective , and so in this temperature regime , grain surface reactions are unlikely to be important . from this analysis
, we see that even if @xmath144 , neutral - neutral grain surface reactions are unimportant at gas densities @xmath162 . in gravitationally collapsing gas with @xmath163 ,
this corresponds to @xmath164 . note that a key point here is that for neutral - neutral grain surface reactions involving our atomic coolants ( c , o etc . ) to significantly affect the thermal behaviour of the gas , a large fraction of the total number of coolant atoms must react , whereas for grain surface reactions to affect the @xmath6 cooling rate by significantly altering the @xmath6 abundance , only a small fraction of the total number of hydrogen atoms must react .
this means that in the case of @xmath6 , the relevant conversion timescale is several orders of magnitude shorter , and hence our density limit would be significantly smaller if we were to omit grain surface @xmath6 formation from our model .
one way in which this simple analysis could break down is if @xmath165 were much larger for reactions involving the coolant atoms than for @xmath6 formation .
however , a recent analysis of @xmath6 formation on grain surfaces by @xcite that takes both physisorbed ( i.e. van der waals bonded ) and chemisorbed ( i.e.chemically bonded ) hydrogen into account demonstrates that in the conditions of interest in this paper , @xmath166 ( with the result that the computed @xmath6 formation rate is very similar to the widely - used rate of @xcite that is used in our model ) . a comparable analysis has not been performed for reactions involving c , o or si , but clearly they can not have @xmath167 , and so there is little scope for these reactions to occur significantly faster than @xmath6 formation . finally , we include in our surface chemistry model six important reactions involving ions : the recombination of @xmath48 , @xmath54 , @xmath52 , @xmath19 , @xmath55 and @xmath36 with electrons on the surface of grains . as grain surface recombination is a non - radiative process , it proceeds at a much faster rate than radiative recombination in the gas phase .
moreover , if a typical grain is negatively charged , then the effective cross - section for collisions will be much enhanced over the geometric cross - section due to coloumb focussing .
grain charging is largely determined by the parameter @xcite @xmath168 where @xmath169 is a measure of the radiation energy density between @xmath170 and @xmath171 relative to the @xcite field .
when @xmath172 is small , most grains are negatively charged , and so in these conditions grain surface recombination can be important even if neutral - neutral grain surface reactions are unimportant . to model grain surface recombination , we follow @xcite .
we adopt their rate coefficients for the recombination of @xmath48 , @xmath52 , @xmath19 and @xmath36 . for @xmath55 and @xmath54
, we use the facts that @xmath55 and @xmath54 have almost the same ionization potential as @xmath48 and that the ion arrival rate at the grain scales as @xmath173 ( where @xmath152 is the mass of the ion in atomic mass units ) to derive rates by appropriately scaling the @xmath48 rate .
the @xcite rates were all computed for milky way dust , and to adapt them for use in low metallicity gas , we again assume that they scale linearly with @xmath143 , with the same caveats as before .
in table [ tab : chem_gas_photo ] , we list the cross - sections for all but two of the photochemical reactions included in our model .
the two exceptions are @xmath6 photodissociation and @xmath15 photodissociation , which are caused by absorption in a large number of discrete spectral lines , and which are discussed separately in [ h2_photodiss ] below . given the cross - section , @xmath174 , the corresponding photochemical rate can be obtained from @xmath175 { \rm d}e\ ] ] where @xmath176 is the photon energy , @xmath177 is the energy threshold , @xmath178 and @xmath179 are the mean intensity ( in units of @xmath180 ) and optical depth for a photon of energy @xmath176 ( both of which are problem dependent ) , and where @xmath181 is a factor that accounts for the effects of secondary ionizations .
it is generally a reasonable approximation to set @xmath182 unless the ionizing component of the radiation field is dominated by x - rays @xcite . in the case that x - rays dominate ,
fits for @xmath181 for @xmath12 and @xmath51 photoionization as a function of the fractional ionization of the gas can be found in @xcite and @xcite .
the effects of secondary ionizations on the other processes listed here are generally negligible , owing to the large abundances of neutral @xmath12 and @xmath51 relative to all other species .
if the gas is optically thick at the @xmath51 photoionization threshold , then an additional process that must be taken into account is the photoionization of @xmath12 by the diffuse emission produced by @xmath52 recombination . in the limit of high optical depth ,
the on - the - spot approximation applies , and we can model this process as a local ionization rate with a value @xcite @xmath183 n_{\rm e } n_{{\rm{he}^{+ } } } \ : { \rm cm^{-3 } } \ : { \rm s^{-1}},\ ] ] where @xmath184 is given by @xmath185 where @xmath186 is the @xmath51 ionization threshold .
( note that if @xmath187 , as is the case in primordial gas with a low fractional ionization and low @xmath6 abundance , then @xmath188 ) .
if the on - the - spot approximation does not apply , then the radiative transfer of this diffuse emission must be modelled in some fashion .
however , a discussion of appropriate techniques for doing so lies well beyond the scope of this paper .
in some circumstances it may also be necessary to take account of the photodissociation of @xmath49 and @xmath50 by photons produced by ionized hydrogen and helium ( both recombination emission and bremsstrahlung ) .
this process is generally important only in neutral gas close to a significant volume of dense , ionized gas ( e.g. in the neutral gas immediately surrounding an expanding ionization front ) .
it is discussed in detail in @xcite and so we do not discuss it further here . finally , we note that for each of the photoionization or photodissociation rates listed in table [ tab : chem_gas_photo ] , there is a corresponding photoheating rate , given by @xmath189 where @xmath190 is the energy of the primary photoelectron and @xmath191 gives the fraction of this energy converted to heat , which can be calculated using the results of @xcite or @xcite . in practice , photoheating from the photoionization of @xmath12 and @xmath51 usually dominates over the other contributions by a wide margin .
although the binding energy of @xmath6 is only 4.48 ev , photons of this energy are not able to dissociate @xmath6 effectively , as the simplest dissociative transition excitation to the vibrational continuum of the ground state is strongly forbidden @xcite .
transitions to the repulsive b@xmath192 state , the least energetic of the excited electronic states of @xmath6 , are also forbidden , and so photodissociation takes place primarily through excitation to the lyman ( b@xmath193 ) or werner ( c@xmath194 ) electronic states followed by radiative decay to the vibrational continuum of the ground state . as a number of vibrational levels
are accessible in each excited state , photodissociation takes place through a number of discrete absorption lines , known as the lyman and werner band systems @xcite . in optically thin gas
, the photodissociation rate can be written as @xmath195 where @xmath196 is the fraction of @xmath6 molecules that have vibrational and rotational quantum numbers @xmath197 in the electronic ground state , and @xmath198 is the photodissociation rate due to transitions out of @xmath197 .
the latter can be written as @xmath199 where @xmath200 is the pumping rate from level @xmath197 in the electronic ground state to level @xmath201 in either the lyman or werner states , @xmath202 is the fraction of decays from @xmath203 which end in the vibrational continuum of the ground state ( rather than back in some bound state ) , and where we sum over all accessible levels . given appropriate molecular data , calculation of @xmath204 is straightforward for each bound level @xmath197 in the electronic ground state . to then calculate @xmath205 , one also needs to know the level populations @xmath196 . if we assume that the mean intensity @xmath206 is independent of energy , and that all of the @xmath6 is in the @xmath207 level ( i.e. the para - hydrogen ground state ) , then @xmath205 evaluates to @xmath208 where we have made use of molecular data taken from @xcite and @xcite .
this expression remains a good approximation in the more general case that @xmath209 is allowed to vary with frequency , provided that the variations are not too extreme and that we replace @xmath209 in equation [ h2_thin_flat ] with @xmath210 , where @xmath211 @xcite : @xmath212 relaxing the assumption that all of the @xmath6 has @xmath213 also makes little difference to @xmath205 @xcite .
vibrational excitation of the @xmath6 makes a much larger difference @xcite , but at low gas densities we would expect the populations of the vibrational levels of @xmath6 to be very small .
equation [ h2_thin ] therefore gives a reasonable estimate of the optically thin @xmath6 photodissociation rate within the regions of parameter space for which our chemical model is valid .
if enough @xmath6 is present in the gas , then the lyman - werner lines can become optically thick , leading to a reduction in the @xmath6 photodissociation rate , an effect known as @xmath6 self - shielding .
if the gas is at rest , then the effects of @xmath6 self - shielding can be treated quite accurately using the prescription of @xcite .
they parameterize the self - shielding with a shielding function @xmath214 , defined to be the ratio of the @xmath6 photodissociation rate in self - shielded gas to the rate in optically thin gas .
they demonstrate how to calculate @xmath214 as a function of the gas temperature and the @xmath6 column density and also construct the following useful fitting function : @xmath215 , \label{db_fsh}\ ] ] where @xmath216 , @xmath217 is the @xmath6 column density , @xmath218 and @xmath219 is the doppler broadening parameter .
although @xcite assume a semi - inifinite slab geometry in their models , their approach is easy to extend to more complicated geometries . unfortunately , their simple treatment breaks down in gas which is not at rest .
doppler shifts due to the motions of the gas cause @xmath6 in different regions to absorb at slightly different wavelengths , and if these doppler shifts are comparable to or larger than the thermal linewidth of the gas ( as will be the case in transonic or supersonic gas respectively ) , then the effect is to reduce the amount of self - shielding that occurs .
an accurate treatment of @xmath6 self - shielding in this regime probably requires one to solve the full frequency - dependent transport equation , which can not currently be done in a computationally efficient manner within a three - dimensional hydrodynamics code .
consequently , various different approximations have been used to study @xmath6 photodissociation in this regime .
the simplest approach is to ignore self - shielding entirely ( see e.g. * ? ? ?
* ; * ? ? ?
this is a good approximation if the velocities in the gas are large and the @xmath6 column densities are small , but otherwise will significantly overestimate the photodissociation rate . at the other extreme , one can ignore the effects of doppler shifts ( e.g. * ? ? ?
* ; * ? ? ?
this is a good approximation if the @xmath6 column density is sufficiently large ( @xmath220 ) that the lorentz wings of the lyman - werner line profiles dominate the line widths , as in this case the line widths will be much larger than any likely doppler shifts within the molecular gas . on the other hand , this approach will underestimate the true photodissociation rate when @xmath221 , as is the case in many interesting low - metallicity systems .
another approximation has recently been suggested by @xcite .
they use equation ( [ db_fsh ] ) to compute @xmath214 , but adopt a value for @xmath219 that includes both a thermal contribution and one arising due to the velocity dispersion of the gas . in practice , this means that they treat their @xmath6 as having an effective @xmath219 equivalent to that in a purely thermal gas with @xmath222 .
the accuracy of this approximation depends on the correlation length of the velocity field .
if this is small compared to the other length scales of interest , then treating the velocity dispersion in this fashion is reasonable and should give a fairly accurate result .
on the other hand , if the velocity field is dominated by large - scale bulk motions ( such as infall into a protogalaxy ) , then this approximation will be significantly less accurate .
finally , in @xcite and in paper ii , we use a local approximation in which only the @xmath6 within a single sph smoothing length is assumed to contribute to the shielding ( see also * ? ? ?
* ; * ? ? ?
* for a grid - based version of this approach ) .
this fairly crude approximation is intended to take account of the fact that @xmath6 close to a given point of interest is more likely to have only a small relative velocity than gas a large distance away .
it will generally underestimate the amount of self - shielding , but nevertheless represents an improvement over neglecting self - shielding entirely . aside from its inevitable inaccuracy , this approximation also suffers from the disadvantage of being resolution dependent , as increasing the number of sph particles in the simulation will generally decrease all of the sph smoothing lengths and hence will cause a systematic increase in @xmath214 . on the other hand , it has the significant advantages of being computationally efficient ( as only local data is required ) as well as being very easy to implement . to sum up ,
a number of different approximate methods exist for treating @xmath6 self - shielding in large numerical simulations , but none are entirely satisfying .
further work on this problem is definitely called for .
turning now to @xmath15 , we note that @xmath15 photodissociation in optically thin gas can be treated in much the same way as @xmath6 photodissociation .
the necessary molecular data for @xmath15 can be found in @xcite , and the resulting photodissociation rate for a radiation field with a flat spectrum can be written as @xmath223 which is only @xmath224 larger than the @xmath6 rate . above an hd column density @xmath225 , self - shielding of the hd lines significantly reduces the photodissociation rate . for a static gas
, this process can again be modeled using the approach of @xcite , although the same problems arise when one tries to extend this approach to a gas distribution which is not static .
however , in the case of @xmath15 , we face an additional complication : if the @xmath15 column density is sufficiently high for hd self - shielding to be significant , then the @xmath6 column density will be very much larger ( since even in significantly fractionated regions , one typically has an hd:@xmath6 ratio of no more than about @xmath226 ) .
consequently , the line widths of the @xmath6 lyman - werner lines are not negligible , and some degree of overlap between these lines and the hd absorption lines will occur . additionally , if the @xmath6:h ratio is small , as will often be the case in the systems of interest , then a significant hd column density implies a large neutral hydrogen column density , which means that absorption of radiation in the lyman series lines of atomic hydrogen must also be taken into account .
these effects are difficult to include accurately in a simple treatment of hd self - shielding and before attempting to do so it is reasonable to ask whether an accurate treatment of hd self - shielding is really required .
we argue that in many cases of interest it is not .
comparison of the rate at which @xmath15 is photodissociated in optically thin gas with the rate at which it is destroyed by reaction 21 demonstrates that the latter dominates whenever @xmath227 , where @xmath228 is the strength of the radiation field at @xmath211 in units of @xmath229 .
photodissociation therefore dominates only when the uv field is strong or the proton number density is small .
however , in either case , it is difficult to see how the large column densities of hd and @xmath6 required for effective shielding could be built up or maintained .
therefore , we suspect that for most applications , treating the hd in the optically thin limit is probably sufficient , as in the conditions where this approximation breaks down , photodissociation is unlikely to be important .
as we do not include molecular coolants such as co or @xmath59 in our chemical model of metal - enriched gas , for the reasons outlined in [ chem_choice ] , the main contribution that the metals make to the cooling of the gas is through fine structure line emission from neutral @xmath18 , @xmath29 and @xmath35 atoms and @xmath19 and @xmath36 ions . to model this emission , we assume that the populations of all the electronically excited levels of these atoms and ions are negligible , an approximation which should be highly accurate at the gas densities considered in this study .
this assumption allows us to model @xmath19 and @xmath36 as two - level systems and @xmath18 , @xmath29 and @xmath35 as three - level systems , allowing us to compute their effects in a straightforward fashion . for a two - level ion ,
if we denote the ground state as level 0 and the excited state as level 1 , then the power radiated per unit volume can be written as @xmath230 where @xmath231 is the number density of ions in level 1 , @xmath232 is the einstein coefficient for spontaneous emission for the transition from level 1 to level 0 , @xmath233 is the corresponding coefficient for stimulated emission , @xmath234 is the energy of the transition , and @xmath235 is the mean specific intensity at the frequency of the transition . if @xmath236 , then the ions will also absorb energy from the radiation field , at a rate @xmath237 where @xmath238 is the number density of ions in level 0 and @xmath239 is the einstein coefficient for absorption from level 0 to level 1 , which is related to @xmath233 by @xmath240 , where @xmath241 and @xmath242 are the statistical weights of levels 0 and 1 respectively .
the net loss of energy per unit time per unit volume is therefore @xmath243 \ } \right),\end{aligned}\ ] ] and it will be seen that if @xmath244 then the ions will absorb more energy than they emit , and so the gas will actually gain energy . to compute @xmath245 , we need to know several pieces of atomic data the values of @xmath232 , @xmath234 , @xmath241 and @xmath242 , which are summarized for @xmath19 and @xmath36 in table [ fs_data ] together with the values of @xmath246 , @xmath238 and @xmath231 . to compute @xmath238 and @xmath231
, we assume that the levels are in statistical equilibrium , in which case : @xmath247 where @xmath248 and @xmath249 are the total rates of collisional excitation and de - excitation respectively .
these are related by @xmath250 and so once one is known , the other can be computed easily . in table
[ fs_coll_rates ] , we list collisional de - excitation rates for collisions between @xmath19 or @xmath36 and various possible collision partners such as @xmath12 , @xmath6 or @xmath47 . given the number densities of these species , @xmath249 can be easily computed , since @xmath251 where @xmath252 is the collisional de - excitation rate for a collision with chemical species @xmath253 with number density @xmath254 . for @xmath19
, we include the effects of collisions with electrons , atomic hydrogen and molecular hydrogen ( in both ortho and para forms ) .
the collision rate with @xmath48 is negligible at the temperatures of interest due to the strong coloumb repulsion , and the abundances of the other chemical species included in our model are too small for them to be important collision partners . for @xmath36 , we include only the effects of collisions with @xmath12 and @xmath47 , as rates for collisions with @xmath6 are not available .
however , provided that the @xmath6 abundance is small compared to the atomic hydrogen abundance , this is unlikely to be a major source of error .
finally , to compute @xmath235 , we assume that the only significant radiation field present at the infra - red and sub - millimeter wavelengths of the fine structure transitions is the cosmic microwave background . in that case
, @xmath235 is simply given by the value of the planck function at the frequency of the transition for a radiation field with temperature @xmath255 .
for the three - level atoms ( @xmath18 , @xmath29 and @xmath35 ) , we use a very similar approach . in this case , the power radiated per unit volume is @xmath256 and the power absorbed per unit volume is @xmath257 where 0 , 1 and 2 denote the ground state and the two excited states respectively .
the level populations @xmath238 , @xmath231 and @xmath258 are found by solving @xmath259 with all symbols having their obvious meanings . to compute the total collisional excitation and de - excitation rates for carbon and oxygen
we include the effects of collisions with ortho and para-@xmath6 , atomic hydrogen , protons and electrons . for silicon , we include only the effects of collisions with @xmath12 and @xmath48 , as rates for collisions with @xmath6 or electrons do not appear to be available .
the rates used are summarized in table [ fs_coll_rates ] .
apart from fine structure emission , we also include in our thermal model several other processes that can lead to the cooling of the gas .
these are summarized in table [ cool_other ] , along with a reference to the source ( or sources ) from which the associated cooling rate has been taken . in most cases ,
the rate itself is also listed . in hot , ionized gas ,
cooling is dominated by electron impact excitation of atomic hydrogen ( lyman-@xmath42 cooling ) , atomic helium and @xmath52 .
excitation of atomic helium occurs from both the @xmath260s ground state and the @xmath261s metastable state . in common with previous authors , we assume that the population of the @xmath261s state is set by the balance between radiative recombination to triplet states and radiative decay to the ground state .
one consequence of this assumption is that the number density of @xmath51 atoms in the @xmath261s state , @xmath262 , is proportional to the product of the number densities of free electrons and of @xmath52 , i.e. @xmath263 , which means that the cooling rate from metastable helium scales as @xmath264 . to model cooling from @xmath12 , the @xmath51 metastable state , and @xmath52
, we use rates from @xcite , which themselves were based on earlier work by @xcite . to model cooling from the @xmath51 ground state
, we use our own fit to the data of @xcite .
a number of other processes are of importance in ionized gas .
we include cooling due to collisional ionization of atomic hydrogen and atomic helium , the gas phase recombination of @xmath48 and @xmath52 , ionic recombination on dust grains , compton scattering of cmb photons by free electrons ( compton cooling ) , and thermal bremsstrahlung .
the rates adopted for all of these processes are summarized in table [ cool_other ] . as the chemical and thermal model presented here
is not designed to be used for the study of very hot gas , we do not include processes involving @xmath265 ( although rates for these processes can be found in @xcite ) . in neutral gas ,
all of the aforementioned processes become ineffective . in cool , neutral gas ,
most of the cooling comes from @xmath6 or from the fine structure lines of carbon , oxygen and silicon , which have already been discussed above .
cooling from @xmath6 is treated in our model through use of the cooling function of @xcite , which we have extended to temperatures below @xmath266 by assuming that only the @xmath267 and @xmath268 transitions contribute significantly to the cooling rate . @xcite
tabulate the @xmath6 cooling rate as a function of temperature , density , @xmath12:@xmath6 ratio and ortho : para ratio .
for simplicity , in our implementation we do not track the evolution of the @xmath6 ortho : para ratio , instead keeping it fixed at 3:1 , but we note that variations in this ratio are unlikely to significantly affect the @xmath6 cooling rate at temperatures at which it contributes significantly to the total cooling rate ( see , for instance , figure 5 in @xcite ) .
a comparison of the @xcite @xmath6 cooling rate with various other rates that have been used in the literature is given in paper ii ( figure 1 ) .
at low temperatures ( @xmath269 ) , @xmath15 cooling becomes increasingly important and can dominate the total cooling rate if sufficient fractionation occurs @xcite . to model @xmath15 cooling , we use the recent cooling function of @xcite . they provide a complicated fit as a value of temperature and density that is valid over a wide range of both .
for further details , the interested reader should consult their paper .
we include two final processes that can become important in some circumstances .
cooling due to @xmath6 collisional dissociation is modelled under the assumption that each dissociation removes @xmath2704.48 ev of thermal energy from the gas .
it is an effective source of cooling only at temperatures above a few thousand kelvin . in practice ,
the @xmath6 abundance in zero metallicity or low metallicity gas is frequently too small for this process to be important .
we also include the effects of energy transfer from the gas to the dust grains ( if present ) , using a rate from @xcite . at solar metallicity ,
this becomes important at a density @xmath271 , but at lower metallicities , it does not dominate until significantly larger densities are reached ( see e.g. * ? ? ? * ) .
we include in our model the effects of several processes that can heat the gas .
most of these processes operate only if a radiation background is present .
the first of these is photoelectric emission from dust grains .
this operates as follows : photons that interact with dust grains can cause the ejection of energetic electrons from the grain if the photon energy exceeds the work function of the grain . as the energy carried by the electrons
is quickly thermalized , this leads to the heating of the gas .
this process is of great importance in the local ism and has been examined in detail by a number of authors ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . to accurately compute the effects of photoelectric emission , we need to know the grain size distribution and the composition of the grains
however , as discussed previously , large uncertainties exist concerning the properties of grains in low metallicity protogalactic gas .
we therefore make the same assumption here as we did in our treatment of grain surface chemistry , i.e.that the dust has the same properties as milky way dust , but has an abundance that is reduced by a factor @xmath272 .
this assumption allows us to use the following expression for the photoelectric heating rate , taken from @xcite : @xmath273 where @xmath169 is a measure of the radiation energy density between @xmath170 and @xmath171 relative to the @xcite field , and where @xmath274 is the photoelectric heating efficiency , given by @xmath275 where @xmath276 .
the parameter @xmath277 that controls the photoelectric heating efficiency is given by @xmath278 @xcite ; note that this differs by a factor of 2 from the parameter @xmath172 introduced in [ surface_chem ] .
for small @xmath277 , most grains are negatively charged and @xmath279 . on the other hand , for large @xmath277
, most grains are positively charged , and @xmath274 is small , as it is difficult for photons with energies @xmath280 to dislodge further electrons from the grains .
a second source of radiative heating is the photodissociation of @xmath6 .
our treatment of photodissociation heating follows @xcite : we assume that each photodissociation deposits @xmath281 of themal energy into the gas . as well as photodissociating some of the @xmath6 , an ultraviolet background
will also produce vibrationally excited @xmath6 via radiative pumping of the excited levels . in dense gas ,
this pumping leads to heating as most of the excited molecules undergo collisional de - excitation .
we include the effects of radiative pumping by adopting a pumping rate that is 8.5 times larger than the photodissociation rate @xcite , and assuming that each excitation transfers an average of @xmath282 to the gas @xcite , where @xmath283 is the critical density at which collisional de - excitation of vibrationally excited @xmath6 occurs at the same rate as radiative de - excitation .
our value for @xmath283 is a weighted harmonic mean of the value for @xmath6-@xmath12 collisions given by @xcite , reduced by a factor of ten as advised by @xcite , and the value for @xmath6-@xmath6 collisions given by @xcite .
heating due to the photoionization of @xmath12 or @xmath51 has already been discussed in [ photochem ] and we do not discuss it further here . as is also discussed in that section , we do not include heating due to the photodetachment of @xmath49 , photodissociation of @xmath50 or photoionization of @xmath18 , @xmath29 , @xmath35 or @xmath36 , as the contribution from these processes is not significant .
we also include the effects of heating due to @xmath6 formation .
the formation of an @xmath6 molecule via reaction 2 releases 3.53 ev of energy , while formation via reaction 4 releases 1.83 ev , and formation on a grain surface ( reaction 60 ) releases 4.48 ev .
we assume that essentially all of this energy goes into rotational and vibrational excitation of the resulting @xmath6 molecule , and hence is radiated away at low gas densities and is converted by collisional de - excitation into heat at high gas densities .
finally , we include heating due to the ionization of the gas by cosmic rays .
following @xcite , we assume that every ionization deposits 20 ev of heat in the gas , and so derive a heating rate that scales with the total cosmic ray ionization rate of the gas ( i.e. the sum of the ionization rates for the various individual chemical species , weighted by the fractional abundances of those species ) . since considerable uncertainty exists concerning the value of the cosmic ray ionization rate in the local interstellar medium ( see the discussion in * ? ? ?
* ) , let alone concerning the appropriate rate to use in high - redshift protogalaxies , a more detailed treatment does not appear to be warranted at this time .
the chemical network and thermal model described in the preceding sections have a number of potential applications .
one of the more obvious applications is the study of the cooling and gravitational collapse of gas in low metallicity protogalaxies .
as previously noted , the density range for which our model is valid corresponds to a wide range of cosmological overdensities .
it can therefore be used to study the thermal and chemical evolution of the majority of the gas within a given protogalaxy .
for example , suppose we model the gas distribution within a @xmath43 halo as a singular isothermal sphere with mean number density within the virial radius @xmath284 ( a reasonable zeroth - order approximation for a protogalactic halo with a mean overdensity @xmath285 ; see e.g. @xcite ) . in this case , only @xmath2704% of the gas within the halo has a density @xmath286 at which our model may break down .
therefore , although our model is of limited usefulness for studying gravitational fragmentation and star formation within this dense central region , it does allow one to study many other important problems , such as how the minimum protogalactic mass at which cooling becomes effective is affected by the presence of metals , or how the gas responds to the presence of an ultraviolet background .
we use a version of this model to address some of these questions in paper ii for the case of small protogalactic halos within initially ionized regions .
a second possible application is the study of the evolution of the ism in high redshift , metal - poor dwarf galaxies . in this case , much of the gas involved is often gravitational stable , and so the characteristic timescales are longer than the gravitational free - fall time .
this limits the applicability of the model to lower densities than in the case of gravitationally collapsing protogalaxies .
however , even if we adopt a relatively long characteristic timescale @xmath287 , corresponding to the sound crossing time of a @xmath114 disk at a sound speed of @xmath288 , the model remains valid for densities up to @xmath289 even in the least auspicious case ( @xmath2 gas , with no significant uv radiation field present ) .
our model is therefore well - suited for use in the study of the evolution of the warm neutral component of the ism in such galaxies and will in some cases also be useful in the study of the cold neutral component .
however , the reader is reminded that our model does not treat hot ( @xmath11 ) , highly - ionized gas , and can not be used to study the effects of stellar feedback ( regions , supernovae , etc . ) unless coupled with some existing model capable of treating this hot gas ( see e.g. @xcite ) . as well as studying the chemical evolution of gas _ within _ dwarf galaxies and protogalaxies , we can also use it to study the chemical evolution of gas _ between _ galaxies , i.e. the intergalactic medium ( igm ) .
one important reason to do so is the fact that several different obervational techniques have been suggested that may allow one to probe the thermal and chemical state of this gas .
@xcite has suggested that if large regions of the igm remain neutral after the turn - on of the first observable ionizing sources , then atomic oxygen in the igm may be detectable through its uv absorption , as it will produce an forest analogous to the @xmath290 lyman-@xmath42 forest .
@xcite also argue that a substantial fraction of the metals in high - redshift galactic winds are likely to be in low - ionization states and may be observable in absorption .
metals in the high redshift igm may also leave detectable imprints in the fluctuation spectrum of the cosmic microwave background @xcite .
finally , @xcite have recently suggested that fine structure emission from atomic oxygen may also produce a detectable spectral distortion of the cmb if the excited fine - structure levels can be populated by radiative pumping via the balmer-@xmath42 transition .
accurate modelling of many of these effects requires accurate modelling of the temperature and chemical make - up of the intergalactic gas , and hence a model such as that presented here .
we thank r. klessen and m .-
mac low for useful discussions , and the anonymous referee for valuable feedback on earlier drafts of this paper .
we also thank r .- s .
ciobanu for pointing out a serious typographical error in an earlier version of this paper . s.c.o.g . acknowledges support from nsf grant ast03 - 07793 during the early phases of this work .
acknowledges support from the emmy noether program of the deutsche forschungsgemeinschaft ( grant no .
kl1358/1 ) .
llllc & & & & + 1 & @xmath291 & @xmath292 & & 1 + & & @xmath293 $ ] & @xmath294 & + & & & & + & & @xmath295 & & + & & @xmath296 & & + & & @xmath297 $ ] & @xmath298 & + & & & & + & & & & + 2 & @xmath299 & @xmath300 & @xmath301 & 2 + & & & & + & & @xmath302 & @xmath303 & + & & & & + & & & & + 3 & @xmath304 & @xmath305 & & 3 + & & @xmath306 $ ] & & + & & & & + & & & & + 4 & @xmath307 & @xmath308 & & 4 + & & & & + & & & & + 5 & @xmath309 & @xmath310 & & 5 + & & @xmath311 & + & & & & + & & & & + 6 & @xmath312 & @xmath313 & @xmath314 & 6 + & & & & + & & @xmath315 & @xmath316 & + & & & & + & & & & + 7 & @xmath317 & @xmath318 & & 7 + & & @xmath319 & & + & & @xmath320 & & + & & @xmath321 & & + & & @xmath322 & & + & & @xmath323 & & + & & @xmath324 & & + & & @xmath325 $ ] & & + & & @xmath326 & & + & & & & + & & & & + 8 & @xmath327 & @xmath328 & & 8 + & & & & + & & & & + 9 & @xmath329 & @xmath330 $ ] & & 9 + & & & & + & & & & + 10 & @xmath331 & @xmath332 & & 10 + & & & & + & & & & + 11 & @xmath333 & @xmath334 & & 11 + & & @xmath335 & & + & & @xmath336 & & + & & @xmath337 & & + & & @xmath338 & & + & & @xmath339 & & + & & @xmath340 & & + & & @xmath341 & & + & & @xmath342 $ ] & & + & & & & + & & & & + 12 & @xmath343 & @xmath344 & & + & & & & + & & & & + 13 & @xmath345 & @xmath346 & case a & 12 + & & @xmath347^{-1.923 } $ ] & & + & & @xmath348 & case b & 12 + & & @xmath349^{-2.242 } $ ] & & + & & & & + & & & & + 14 & @xmath350 & @xmath351 & & + & & & & + & & & & + 15 & @xmath352 & @xmath353 & & 11 + & & @xmath354 & & + & & @xmath355 & & + & & @xmath356 & & + & & @xmath357 & & + & & @xmath358 & & + & & @xmath359 & & + & & @xmath360 & & + & & @xmath361 $ ] & & + & & & & + & & & & + 16 & @xmath362 & @xmath363 & @xmath364 & 11 + & & & & + & & @xmath365 & & + & & @xmath366 & & + & & @xmath367 & & + & & @xmath368 & & + & & @xmath369 & & + & & @xmath370 & & + & & @xmath371 & & + & & @xmath372 & & + & & @xmath373 & & + & & @xmath374 $ ] & @xmath375 & + & & & & + & & & & + 17 & @xmath376 & @xmath377 & @xmath378 & 13 + & & & & + & & @xmath379 & @xmath380 & + & & & & + & & & & + 18 & @xmath381 & @xmath382 & & 14 + & & @xmath383 & & + & & & & + & & & & + 19 & @xmath384 & @xmath385 & @xmath386 & 14 + & & @xmath387 & & + & & & & + & & @xmath388 & @xmath389 & + & & & & + & & & & + 20 & @xmath390 & @xmath391 \times 10^{-9}$ ] & & 15 + & & & & + & & & & + 21 & @xmath392 & @xmath393 & & 15 + & & & & + & & & & + 22 & @xmath394 & @xmath395 & @xmath396 & 16 + & & & & + & & @xmath397 & @xmath398 & + & & & & + & & & & + 23 & @xmath399 & @xmath400 & @xmath396 & 17 + & & & & + & & @xmath401 & @xmath398 & + & & & & + & & & & + 24 & @xmath402 & @xmath403 & & 11 + & & @xmath404 & & + & & @xmath405 & & + & & @xmath406 & & + & & @xmath407 & & + & & @xmath408 & & + & & @xmath409 & & + & & @xmath410 & & + & & @xmath411 $ ] & & + & & & & + & & & & + 25 & @xmath412 & @xmath413 & case a & 18 + & & @xmath414 $ ] & & + & & & & + & & @xmath415 & case b & 18 + & & @xmath416 $ ] & & + & & & & + & & @xmath417 & & + & & @xmath418 $ ] & & 19 + & & & & + & & & & + 26 & @xmath419 & @xmath420 & & 20 + & & & & + & & & & + 27 & @xmath421 & @xmath422 & @xmath423 & 21 + & & & & + & & @xmath424 & @xmath425 & + & & & & + & & & & + 28 & @xmath426 & @xmath427 & & + & & & & + & & & & + 29 & @xmath428 & @xmath429 & & + & & & & + & & & & + 30 & @xmath430 & @xmath431 & @xmath432 & 22 + & & & & + & & @xmath433 & @xmath434 & + & & & & + & & @xmath435 & @xmath436 & + & & & & + & & & & + 31 & @xmath437 & @xmath438 & @xmath439 & 23 + & & & & + & & @xmath440 & @xmath441 & + & & & & + & & @xmath442 & @xmath443 & + & & & & + & & & & + 32 & @xmath444 & @xmath445 & @xmath446 & 24 + & & & & + & & @xmath447 & & + & & @xmath448 $ ] & @xmath449 & + & & & & + & & & & + 33 & @xmath450 & @xmath451 & @xmath452 & 25 + & & & & + & & & & + 34 & @xmath453 & @xmath454 & @xmath455 & 25 + & & & & + & & & & + 35 & @xmath456 & @xmath457 & @xmath458 & 25 + & & & & + & & & & + 36 & @xmath459 & @xmath460 & & 26 + & & & & + & & & & + 37 & @xmath461 & @xmath462 & & 27 + & & @xmath463 \exp \left(-\frac{227}{t}\right)$ ] & & + & & & & + & & & & + 38 & @xmath464 & @xmath465 & & 28 + & & @xmath466 & & + & & & & + & & & & + 39 & @xmath467 & @xmath468 & & 27 + & & & & + & & & & + 40 & @xmath469 & @xmath470 & & 27 + & & & & + & & & & + 41 & @xmath471 & @xmath472 & @xmath396 & 29 + & & & & + & & @xmath473 & @xmath474 & + & & & & + & & @xmath475 & @xmath161 & + & & & & + & & & & + 42 & @xmath476 & @xmath477 & @xmath478 & 30 + & & & & + & & @xmath479 & @xmath443 & + & & & & + & & & & + 43 & @xmath480 & @xmath481 & & 31 + & & & & + & & & & + 44 & @xmath482 & @xmath483 & & 31 + & & & & + & & & & + 45 & @xmath484 & @xmath485 & & 30 + & & @xmath486 { \exp \left(-\frac{3.178}{t_{\rm e}}\right)}$ ] & & + & & & & + & & & & + 46 & @xmath487 & @xmath488 & & 30 + & & @xmath489 $ ] & & + & & & & + & & & & + 47 & @xmath490 & @xmath491 & & 32 + & & & & + & & @xmath492 & & 33 + & & @xmath493 & & + & & & & + llllll 48 & @xmath494 & @xmath495 & @xmath496 & 1 + & & @xmath497^{-1}$ ] & @xmath498 & + & & & & + 49 & @xmath499 & @xmath500 & @xmath496 & 1 + & & & & + 50 & @xmath501 & @xmath502 & @xmath503 & 2 + & & @xmath504 & & + & & @xmath505 & & + & & @xmath506 $ ] & & + & & & & + 51 & @xmath507 & @xmath508 & @xmath509 & 3 + & & & & + 52 & @xmath510 & @xmath511 & @xmath512 & 4 + & & @xmath513 $ ] & @xmath514 & + & & & & + & & @xmath515 & @xmath516 & + & & @xmath517 $ ] & & + & & & & + 53 & @xmath518 & see [ h2_photodiss ] & & 5 + & & & & + 54 & @xmath519 & @xmath520 & @xmath521 & 6 + & & & @xmath522 & + & & & & + & & @xmath523 & @xmath524 & + & & & & + & & @xmath525 & @xmath526 & + & & & & + 55 & @xmath527 & see [ h2_photodiss ] & & 7 + & & & & + 56 & @xmath528 & @xmath529 & @xmath530 & 8 + & & & @xmath531 & + & & & @xmath532 & + & & & @xmath533 & + & & & @xmath534 & + & & & @xmath535 & + & & & & + 57 & @xmath536 & @xmath537 & @xmath538 & 8 + & & & @xmath539 & + & & & @xmath540 & + & & & @xmath541 & + & & & @xmath542 & + & & & @xmath543 & + & & & & + & & & & + 58 & @xmath544 & @xmath545 & @xmath546 & 8 + & & & @xmath547 & + & & & @xmath548 & + & & & @xmath549 & + & & & @xmath550 & + & & & @xmath551 & + & & & & + 59 & @xmath552 & @xmath553 & @xmath554 & 8 + & & & @xmath555 & + & & & @xmath556 & + & & & @xmath557 & + & & & @xmath558 & + & & & @xmath559 & + lllc & & & + 60 & @xmath560 & @xmath561 & 1 + & & @xmath562^{-1 } \left[1.0 + 10^{4 } \exp \left(-\frac{600}{t_{\rm gr}}\right ) \right]^{-1}$ ] & + & & & + 61 & @xmath563 & @xmath564 & 2 + & & @xmath565^{-1}$ ] & + & & & + 62 & @xmath566 & @xmath567 & 3 + & & & + 63 & @xmath568 & @xmath569 & 2 + & & @xmath570^{-1}$ ] & + & & & + 64 & @xmath571 & @xmath572 & 2 + & & @xmath573^{-1}$ ] & + & & & + 65 & @xmath574 & @xmath575 & 3 + & & & + 66 & @xmath576 & @xmath577 & 2 + & & @xmath578^{-1}$ ] & + & & & + clcc 67 & @xmath579 & 1.0 & @xcite + 68 & @xmath580 & 1.0 & @xcite + 69 & @xmath581 & 1.09 & @xcite + 70 & @xmath582 & 2.0 & @xcite + 71 & @xmath583 & 3.83 & @xcite + 72 & @xmath584 & 5.67 & @xcite + 73 & @xmath585 & 6.5 & @xcite + 74 & @xmath586 & 2.5 & @xcite + cclllll @xmath18 & @xmath587 & 3 & 1 & 609.2 & 24 & @xmath588 + @xmath18 & @xmath589 & 5 & 1 & 229.9 & 63 & @xmath590 + @xmath18 & @xmath591 & 5 & 3 & 369.0 & 39 & @xmath592 + @xmath29 & @xmath587 & 3 & 5 & 63.1 & 230 & @xmath593 + @xmath29 & @xmath589 & 1 & 5 & 44.2 & 330 & @xmath594 + @xmath29 & @xmath591 & 1 & 3 & 145.6 & 98 & @xmath595 + @xmath35 & @xmath587 & 3 & 1 & 129.6 & 110 & @xmath596 + @xmath35 & @xmath589 & 5 & 1 & 44.8 & 320 & @xmath597 + @xmath35 & @xmath591 & 5 & 3 & 68.4 & 210 & @xmath598 + @xmath19 & @xmath587 & 4 & 2 & 157.7 & 92 & @xmath599 + @xmath36 & @xmath587 & 4 & 2 & 34.8 & 410 & @xmath600 + cclcc & & & & + @xmath18 & o-@xmath6 & @xmath601 & & + & & @xmath602 & & 1 + & & @xmath603 & & 1 + & & @xmath604 & & 1 + & & & & + @xmath18 & p-@xmath6 & @xmath605 & & + & & @xmath606 & & 1 + & & @xmath607 & & + & & @xmath608 & & 1 + & & @xmath609 & & + & & @xmath610 & & 1 + & & & & + @xmath18 & @xmath12 & @xmath611 & & 2 + & & @xmath612 & & 2 + & & @xmath613 & & 2 + & & & & + @xmath18 & @xmath48 & @xmath614 & @xmath615 & 3 + & & @xmath616 & @xmath617 & 3 + & & & & + & & @xmath618 & @xmath615 & 3 + & & @xmath619 & @xmath617 & 3 + & & & & + & & @xmath620 & @xmath615 & 3 + & & @xmath621 & @xmath617 & 3 + & & & & + @xmath18 & @xmath47 & @xmath622 & & + & & @xmath623 & & + & & @xmath624 $ ] & @xmath625 & 4 + & & & & + & & @xmath626 & & + & & @xmath627 & & + & & @xmath628 $ ] & @xmath629 & 4 + & & & & + & & @xmath630 & & + & & @xmath631 & & + & & @xmath632 $ ] & @xmath625 & 4 + & & & & + & & @xmath633 & & + & & @xmath634 & & + & & @xmath635 $ ] & @xmath629 & 4 + & & & & + & & @xmath636 & & + & & @xmath637 & & + & & @xmath638 $ ] & @xmath625 & 4 + & & & & + & & @xmath639 & & + & & @xmath640 & & + & & @xmath641 $ ] & @xmath629 & 4 + & & & & + & & & & + @xmath29 & o-@xmath6 & @xmath642 & & 5 + & & @xmath643 & & 5 + & & @xmath644 & & 5 + & & & & + @xmath29 & p-@xmath6 & @xmath645 & & 5 + & & @xmath646 & & 5 + & & @xmath647 & & 5 + & & & & + @xmath29 & @xmath12 & @xmath648 & & 5 + & & @xmath649 & & 5 + & & @xmath650 & & 5 + & & & & + @xmath29 & @xmath48 & @xmath651 & @xmath652 & 6 + & & @xmath653 & @xmath654 & + & & @xmath655 & @xmath656 & + & & & & + & & @xmath657 & @xmath658 & 6 + & & @xmath659 & @xmath660 & + & & @xmath661 & @xmath662 & + & & & & + & & @xmath663 & @xmath664 & 6 + & & @xmath665 & @xmath666 & + & & & & + @xmath29 & @xmath47 & @xmath667 & & 7 + & & @xmath668 & & 7 + & & @xmath669 & & 7 + & & & & + & & & & + @xmath35 & @xmath12 & @xmath670 & & 2 + & & @xmath671 & & 2 + & & @xmath672 & & 2 + & & & & + @xmath35 & @xmath48 & @xmath673 & & 2 + & & @xmath674 & & 2 + & & @xmath675 & & 2 + & & & & + & & & & + @xmath19 & o-@xmath6 & @xmath676 & @xmath677 & 8 , 9 + & & @xmath678 & @xmath679 & + & & & & + @xmath19 & p-@xmath6 & @xmath680 & @xmath677 & 8 , 9 + & & @xmath681 & @xmath679 & + & & & & + @xmath19 & @xmath12 & @xmath682 & @xmath683 & 2 , 10 + & & @xmath684 & @xmath685 & + & & & & + @xmath19 & @xmath47 & @xmath686 & @xmath683 & 11 + & & @xmath687 & @xmath685 & + & & & & + & & & & + @xmath36 & @xmath12 & @xmath688 & & 12 + & & & & + @xmath36 & @xmath47 & @xmath689 & & 13 + & & & & + llc & & + & & + h excitation & @xmath690 & 1 + & & + he excitation ( @xmath260s state ) & @xmath691 & 2 + & & + he excitation ( @xmath261s state ) & @xmath692 & 1 + & & + @xmath52 excitation & @xmath693 & 1 + & & + @xmath12 collisional ionization & @xmath694 & 3 + & & + @xmath51 collisional ionization & @xmath695 & 3 + & & + compton cooling & @xmath696 & 1 + & & + bremsstrahlung & @xmath697 & 4 + & & + & @xmath698 & + & @xmath699 & + & & + @xmath48 recombination ( radiative ) & @xmath700 & 5 + & & + @xmath52 recombination ( radiative ) & @xmath701 & 6 + & & + @xmath52 recombination ( dielectronic ) & @xmath702 & 7 + & & + grain surface recombination & @xmath703 & 8 + & & + @xmath6 rovibrational lines & see [ cool : other ] & 9 + & & + @xmath15 rovibrational lines & see [ cool : other ] & 10 + & & + @xmath6 collisional dissociation & @xmath704 & 11 + & & + gas - grain energy transfer & @xmath705 \left(\frac{{\rm z}}{{\rm z_{\odot}}}\right ) n^{2}$ ] & 12 + & & + & & + * heating : * & & + & & + photoelectric effect & @xmath706 & 13 + & & + & @xmath707 & + & & + @xmath6 photodissociation & @xmath708 & 14 + & & + uv pumping of @xmath6 & @xmath709 & 15 + & & + h photoionization & dependent on incident spectrum ; see [ photochem ] & 16 + & & + he photoionization & dependent on incident spectrum ; see [ photochem ] & 17 + & & + gas - phase @xmath6 formation & @xmath710 n_{{\rm{h } } } \left(\frac{n}{n + n_{\rm cr}}\right)$ ] & 18 + & & + @xmath6 formation on dust grains & @xmath711 & 19 + & & + cosmic - ray ionization & @xmath712 & 20 + | we present a simplified chemical and thermal model designed to allow computationally efficient study of the thermal evolution of metal - poor gas within large numerical simulations .
our main simplification is the neglect of the molecular chemistry of the heavy elements .
the only molecular chemistry retained within the model is the formation and destruction of molecular hydrogen . despite this major simplification ,
the model allows for accurate treatment of the thermal evolution of the gas within a large volume of parameter space .
it is valid for temperatures @xmath0 and metallicities @xmath1 . in gas with a metallicity @xmath2 , and in the absence of an incident ultraviolet radiation field , it is valid for hydrogen number densities @xmath3 , where @xmath4 is the size in myr of the characteristic physical timescale of interest in the problem .
if @xmath5 , or if a strong ultraviolet radiation field is present , then the model remains accurate up to significantly higher densities .
we also discuss some possible applications of this model . |
_ swift _ j164449.3 + 573451 ( hereafter , swift j1644 + 57 ) was first discovered by the _ swift _
burst alert telescope ( bat ) at 12:57:45 ut on 28 march 2011 @xcite .
some evidences suggest that swift j1644 + 57 is a tidal disruption of a star by a supermassive black hole ( smbh ) .
this phenomenon triggered the bat three times after the initial trigger during the first few days @xcite .
the late - time x - ray light curve was extended to a longer period by following the expected power - law decay for the tidal disruption of a star , i.e. , @xmath12 ( e.g. , * ? ? ?
finally , the source of x - ray , ir , and radio emissions were well matched up with the center of the host galaxy where a smbh resides @xcite .
there have been many studies performed to understand the nature of this event , such as the characteristics of the star that was disrupted .
such a question is closely connected to the smbh mass ( @xmath13 ) .
the disruption of a solar - type star is possible for all @xmath14 @xcite , but compact stars like a white dwarf can be disrupted only if @xmath15 @xcite .
if so , then this kind of event provides the interesting possibility of discovering intermediate - mass black holes .
unfortunately , there has been controversy concerning the mass of the smbh .
@xcite provided a rough estimate of the smbh mass of @xmath16 using a black hole mass luminosity relation and the lower limit of @xmath17 based on the x - ray variability .
similarly , @xcite estimated @xmath13 to be @xmath18
@xmath19 , derived from @xmath10-band luminosity , but at that time , @xmath10-band luminosity contained a significant amount of the transient light .
@xcite utilized a relation between the black hole mass , the radio luminosity , and the x - ray luminosity , and found @xmath13 @xmath20 . using a quasi - periodic oscillation resonance hypothesis ,
@xcite provided an @xmath13 estimate of @xmath21 .
@xcite concluded that a white dwarf was tidally disrupted and the mass of smbh is less than @xmath22 in light of the short timescales of the x - ray light curve . in summary ,
the @xmath13 estimates have centered around the two discrepant values of @xmath19 and @xmath22 or less . a better understanding of the host galaxy properties is needed to clear up the situation .
cccccc + 2011 - 04 - 04 & 55655.147614 & 6.6 & f606w & 1260 & [email protected] + 2011 - 08 - 04 & 55777.276876 & 129 & f606w & 4160 & [email protected] + 2011 - 12 - 02 & 55897.684390 & 249 & f606w & 1113 & [email protected] + 2013 - 04 - 12 & 56394.429204 & 746 & f606w & 2600 & [email protected] + + + 2011 - 04 - 04 & 55655.132654 & 6.6 & f160w & 997 & [email protected] + 2011 - 08 - 04 & 55777.257148 & 129 & f160w & 1412 & [email protected] + 2011 - 12 - 02 & 55897.702220 & 249 & f160w & 1209 & [email protected] + 2013 - 04 - 12 & 56394.295795 & 746 & f160w & 2812 & [email protected] + [ hsttab ] cccccc + 2011 - 04 - 28 & 55679.975316 & 31.4 & 3.6@xmath24 m & 1250 & [email protected] + 2011 - 10 - 31 & 55865.023052 & 216 & 3.6@xmath24 m & 1253 & [email protected] + 2012 - 02 - 24 & 55981.541644 & 333 & 3.6@xmath24 m & 1252 & [email protected] + + + 2011 - 04 - 28 & 55679.975316 & 31.4 & 4.5@xmath24 m & 1250 & [email protected] + 2011 - 10 - 31 & 55865.023052 & 216 & 4.5@xmath24 m & 1253 & [email protected] + 2012 - 02 - 24 & 55981.541644 & 333 & 4.5@xmath24 m & 1252 & [email protected] + [ spitzertab ] in order to more accurately estimate the smbh mass and better constrain the properties of the host galaxy , we analyze the morphology and the surface brightness profile of the host galxaxy based on high - resolution _ hubble space telescope _
( _ hst _ ) images and estimate the multi - band fluxes of the host galaxy using our long - term monitoring data lasting more than 2.4 years .
we fit the multi - band spectral energy distribution ( sed ) of the host galaxy luminosity with stellar population synthesis models , and then obtain the properties of the galaxy .
finally , we provide our best estimate of @xmath13 based on the host galaxy properties .
this is the second of a series of two papers . in the first paper ( m. i m et al .
2015 in preparation , hereafter , im15 ) , we present the dataset of the long - term monitoring campaign and an analysis of the late - time light curve . throughout the paper , we selected _ _
s@xmath26mpc@xmath26 , @xmath27 , and @xmath28 as cosmological parameters and adopt the ab magnitude system .
we observed _ swift _ j1644 + 57 using wide field camera ( wfcam ) on united kingdom infrared telescope ( ukirt ) for nearly 2.4 years following the burst as a part of our gamma - ray burst ( grb ) and transient observation program @xcite .
we observed intensively in the @xmath10 band among @xmath29 , and @xmath10 bands of wfcam .
the number of epochs of @xmath10 band data used for thie analysis is @xmath30 and the last data were observed at @xmath31 days after the initial bat trigger .
the numbers of epochs for the @xmath32-band data are @xmath33 and the last data were observed at @xmath34 , @xmath35 , and @xmath31 days , respectively , where @xmath36 is the number of days since the initial bat trigger .
we have only one epoch of data for the ukirt @xmath37 band which was observed at @xmath38 days .
we also observed _ swift _ j1644 + 57 using camera for quasars in early universe ( cquean ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) on the 2.1 m otto - struve telescope of the mcdonald observatory in @xmath39 , and @xmath40 bands .
the numbers of epochs for the @xmath41 and @xmath40 band data are @xmath42 , and @xmath43 and the last data were observed at @xmath44 , and @xmath45 days respectively . the ukirt and cquean observation logs and photometry results
are described in im15 .
in addition , we also used data from @xcite and @xcite for the earlier optical and near - infrared ( nir ) data .
morphology analysis requires high - resolution images because this object is so compact that it is virtually a point source in the ukirt and cquean images . for the high - resolution images
, we obtained _ hst _
wfc3 multi - drizzled , stacked images available in the mast database .
we used f606w- , f160w - band data and the number of epochs in each two band is four .
these data were observed at @xmath46 and @xmath47 days .
the _ hst _
wfc3 data are summarized in table [ hsttab ] . and a point source component .
the upper panel shows the results for the f606w - band images , while the lower panel shows the results for the f160w - band images .
magnitudes of the model components are also shown .
[ hstfig ] ] to supplement the nir observation data , we used the _ spitzer _
irac 3.6 , 4.5@xmath24 m post basic calibrated data ( pbcd ) from the nasa / ipac infrared science archive .
these were observed at @xmath48 , and @xmath49 days . a log of the _ spitzer _ irac 3.6 , 4.5@xmath24 m data are shown in table [ spitzertab ] .
the flux measurements were performed by sextractor software @xcite except for the _ hst _ images for which galfit @xcite models were used for the flux measurements . for the x - ray data
, we used _
swift_/xrt data taken from the _ swift _ archive and _ xmm - newton _ data from the _ xmm - newton _ science archive .
we analyzed the surface brightness profile of the host galaxy , in order to determine the bulge fraction and its nature .
we used the _ hst _ images and galfit software to fit two - dimensional models to the light distribution of the host galaxy . to construct the point spread function ( psf ) , we selected @xmath50 isolated stars with signal - to - noise ratios @xmath51 in the vicinity of _ swift _ j1644 + 57 , and co - added them .
we used error images that are created by galfit for the fitting . for galfit to create the error image properly , we modified the unit of adu and the image header values such that @xmath52}$ ] as recommended in galfit website .
a crucial factor affecting the fitting results is background subtraction . for the background determination
, we set 6 annuli with the radii logarithmically increasing between 2.5 and 9 times the radius of an ellipse for which pixel values are @xmath53 of the background noise .
we centered the annuli on the center of the host galaxy , set the minimum width of the annuli to be @xmath54 arcsec ( @xmath55 pixel ) for f606w images and @xmath56 arcsec ( @xmath57 pixel ) for f160w images , and augmented the widths in step with the logarithmically growing radii .
we then derived the mean pixel values of each annulus .
finally , we adopted their mean value as the background value .
our surface brightness fit was carried out using a deep , stacked image of the data taken with f606w at @xmath58 129 , 249 , and 749 days .
it has been known that the transient component is negligible in the optical bands bluer than @xmath59 even at the early time @xcite .
the use of the stacked , late - time image in the f606w band makes the transient component more negligible . on the other hand , nir - bands , including f160w ( similar to @xmath60 band of wfcam ) , are known to contain a significant transient component which may affect the host galaxy analysis .
furthermore , the spatial resolution of the f606w images is better by a factor of three than that of f160w , which greatly helps the surface brightness fitting .
the other _ hst _ data were also analyzed to understand the importance of the transient component , and the results for the transient component are presented later in this section .
for the galaxy models , we used the srsic @xcite , de vaucouleurs @xcite , and exponential disk profiles or a combination of thereof .
the srsic profile is described as @xmath61,\ ] ] where @xmath62 is the surface brightness at the effective radius @xmath63 , and @xmath64 is the srsic index .
@xmath65 is a variable parameter denpendent on @xmath64 , where @xmath66 and @xmath67 correspond to the de vaucouleurs and exponential profiles , respectively .
although @xmath66 is commonly quoted for the ellipticals and classical bulges , the srsic index of ellipticals and classical bulges can assume a value in the range @xmath68 , whereas pseudobulges have @xmath69 @xcite .
all of the model parameters such as ellipticity and center positions of the different components , were set free in the fitting procedure .
figures [ sbfig ] and [ sb2fig ] show images of the host galaxy , the two - dimensional models , and the residuals ( i.e. , the model subtracted images ) , for four different models : ( 1 ) a single srsic ; ( 2 ) a srsic bulge + exponential disk ; ( 3 ) an exponential disk ; and ( 4 ) a double exponential profile models . the figures also show one - dimensional profiles ( along the major axis ) of the host galaxy and those of each model component , which are converted through the iraf ellipse task .
in addition to the profiles , the differences between the data and the model profiles are shown .
the results of each fit are summarized in table [ galfittab ] . both the single srsic model with @xmath0 and the srsic bulge with @xmath70 @xmath71 exponential disk model fit the data well ( @xmath72 @xmath73 @xmath74 ) .
when the disk component is added , the bulge to total host galaxy flux ratio ( b / t ) is @xmath1 . on the other hand
, the single exponential disk model provides a poor fit to the data as shown in figure 2 and with @xmath75 .
the double exponential profile model fits the data nearly as well as the single srsic model and the srsic bulge+disk model in terms of @xmath76 .
however , the analysis of the one - dimensional surface brightness profile shows that the model does not follow the outer part of the profile well , demonstrating a relatively steeper decline than that of the single srsic model and the srsic bulge @xmath71 exponential disk model .
this model gives b / t @xmath77 , suggesting a significant bulge component .
therefore , we conclude that the host galaxy of _ swift _ j1644 + 57 is bulge - dominant .
we also conclude that the bulge is likely to have a srsic index higher than @xmath78 regardless of the existence of the disk .
this value corresponds to the range of the classical bulges @xcite .
we can not completely exclude the case where the bulge is pseudobulge with @xmath79 , but even in this case , the object has a significant bulge .
ccccccccccc b & @xmath80 & @xmath81 & @xmath82 & & & & & & @xmath80 & 1.322 + b@xmath71d & @xmath83 & @xmath84 & @xmath85 & & @xmath86 & @xmath87 & & @xmath1 & @xmath88 & 1.223 + + b@xmath89 & @xmath90 & 4 ( fixed ) & @xmath91 & & & & & & @xmath90 & 1.275 + d & & & & & @xmath92 & @xmath93 & & & @xmath92 & 6.547 + b@xmath89@xmath71d & @xmath83 & 4 ( fixed ) & @xmath94 & & @xmath95 & @xmath96 & & @xmath97 & @xmath98 & 1.228 + s@xmath99@xmath71d & @xmath100 & 1 ( fixed ) & @xmath101 & & @xmath102 & @xmath103 & & @xmath104 & @xmath105 & 1.351 [ galfittab ] 500 days since the bat trigger .
[ xfig ] ] - and @xmath10-band light curves in linear scale .
some of the data points at very early times are cut in order to highlight the late - time light curves .
the magnitudes of the latest @xmath106 bands converge to single values , suggesting that the transient component has disappeared at @xmath107 days .
[ linfig ] ] ] additionally we fit the observed surface brightness profile with a single de vaucouleurs bulge model and a de vaucouleurs bulge @xmath71 exponential disk model .
the results of these fits are nearly identical to that of the single srsic and the srsic bulge+disk models .
to estimate the transient component flux , we fit all the f606w and f160w images with a model containing both the point source ( transient ) and the host galaxy components . here
, we adopt a single srsic profile with a fixed srsic index ( @xmath108 ) for the host galaxy component , and a psf profile for the transient component .
the compactness of the host galaxy and the bright transient component in the f160w images create a serious degeneracy , particularly between the effective radius and the srsic index , when fitting multi - component models . to alleviate the degeneracy , we fixed the srsic index to be @xmath108 , similar to that of the f606w band .
the flux fractions of the models as a function of time are shown in figure [ hstfig ] .
in the case of f606w , the flux fraction from the transient component is very small or nonexistent , while the transient component is very bright in the earliest f160w band , even brighter than the entire host galaxy .
the result reflects a very red color to the transient well , and justifies the exclusion of the point source component in the late - time f606w images during the host galaxy analysis .
the point source contribution declines rapidly as time goes on in f160w , but it contributes to the total flux of the object until around @xmath109 days . on the other hand , the fluxes of the host galaxy component are almost constant in both bands over the entire period .
the magnitude of host galaxy in the f160w band is @xmath110 mag , and as we shall see in the next section , this is the same as for the last data point of the ukirt @xmath60-band light curve , suggesting that the flux of the last data point in the nir light curve represents the host galaxy flux . +
in this section , we show long - term observation results and estimate multi - band fluxes of the host galaxy of _ swift _ j1644 + 57 .
figure [ ukirtfig ] shows the @xmath111 , and @xmath10 light curves .
the gray data points in the background show _
swift_/xrt @xmath112 @xmath113kev data .
the @xmath114 light curves resemble each other .
the nir light curves rapidly decline until @xmath115 days , turn up again with a second peak at @xmath116 days , and decline again steadily .
the behaviors of these nir light curves are very similar to that of the x - ray light curve except that the x - ray light curve appears shifted ahead of the nir light curves at a time of @xmath11715 days .
the similar shapes of these light curves indicate that the origins of the x - ray emission and nir emission are related to each other . on the other hand , the time gap between these two emissons denotes that the x - ray source and nir source are separated from each other as much as the time gap .
@xcite suggested that x - ray source is in the close vicinity of the black hole due to the fact that the x - ray emission shows very rapid , high variability , while the ir and the radio sources are located a large distance from the black hole on account of the relatively smooth and small variability .
they argued that the jet generated by black hole collides with the surrounding medium where the electrons are accelerated by the jet .
these high - speed electrons emit the ir to radio photons through synchrotron radiation .
the jet seems to be nearly turned off at @xmath118 days in light of the fact that there is an abrupt decrease in flux of a factor of @xmath119 or more , which can be seen in all the _
swift_/xrt , _ chandra _ , and _ xmm - newton _
data @xcite .
this late - stage turn - off of x - ray flux is also shown in figure [ xfig ] .
if the jet was turned off , then the transient components of the nir fluxes must be quenched following the x - ray flux , and it is expected that the fluxes of the pure host galaxy of _ swift _ j1644 + 57 were revealed at that time . as we can see in figure [ linfig ] , which shows the @xmath60- , @xmath10- bands light curves in linear scale , the fluxes of the @xmath106 bands converge to single values at late - time .
furthermore , the latest @xmath60-band magnitude is nearly the same as that of the host galaxy of the f160w - band images , shown in the results of the model fitting in
[ sec : mor ] .
this evidence indicates that it is reasonable to regard the nir ( @xmath111 , and @xmath10 band ) fluxes of the last data , taken at @xmath120 or @xmath121 days , as those of the pure host galaxy .
the left panel of figure [ lcfig ] shows the light curves of the cquean @xmath59- and @xmath122-band and the ukirt @xmath37-band data . in the case of the @xmath122 band ,
the fluxes from the object slightly decrease with time .
we also regard the flux of the last data , that is the ukirt @xmath37-band data , as the @xmath37-band flux of the host galaxy since it is observed far beyond expected quenching time of jet . on the other hand ,
@xmath59-band fluxes are virtually constant , demonstrating that the transient components are basically non - existent in the @xmath59 or bluer bands ( figure [ hstfig ] ; * ? ? ?
* ; * ? ? ?
we take the flux of the last data of the @xmath59 band as that of the host galaxy .
the number of cquean @xmath123- and @xmath124-band data points are scarce compared to the nir data and the last data were observed at early time ( @xmath125 days , respectively ) .
however , there is little or no change between the very early - time magnitudes from @xcite and our @xmath123- and @xmath124-band magnitudes in the same way as the fluxes of the @xmath59 and f606w bands .
therefore , we consider the @xmath123- and @xmath124-band fluxes of the last epoch data as those from the host galaxy .
furthermore , we also consider the magnitude of the single srsic model of the stacked _ hst _
wfc3 f606w image as that of the host galaxy since the point source contribution to whole flux is negligible .
ll @xmath126 & [email protected] @xcite + @xmath123 & [email protected] + f606w & [email protected] + @xmath124 & [email protected] + @xmath59 & [email protected] + @xmath37 & [email protected] + @xmath40 & [email protected] + @xmath127 & [email protected] + @xmath60 & [email protected] + @xmath10 & [email protected] + 3.6@xmath24 m & [email protected] ( including transient ) + 4.5@xmath24 m & [email protected] ( including transient ) [ hosttab ] rl @xmath128 ( @xmath129-folding time scale of stellar population ) & @xmath130 @xmath131[@xmath128/yr ] @xmath132 with a step size of 0.1 + @xmath133 ( age of stellar population ) & @xmath134 @xmath131[@xmath133/yr ] @xmath135 with a step size of 0.1 + imf ( initial mass function ) & @xcite + z ( metallicity ) & 0.004 , 0.008 , 0.020 , 0.050 + extinction law & @xcite + @xmath136 ( @xmath137-band attenuation for stellar population in magnitude ) & @xmath138 @xmath136 @xmath139 with a step size of 0.1 [ sedtab ] rr stellar mass [ @xmath140 & @xmath141 + + sfr [ @xmath142/yr ] & @xmath143 + + specific sfr [ @xmath144 & @xmath145 + + @xmath133 [ gyr ] & @xmath5 + + @xmath128 [ gyr ] & @xmath146 + + @xmath136 & @xmath147 + + z & @xmath148 + + @xmath149 & 1.64 [ sed2tab ] the right panel of figure [ lcfig ] shows the light curves of the _ spitzer _ irac 3.6@xmath24 m and 4.5@xmath24 m bands .
the magnitude changes in these bands are the more significant than those for the other optical / nir bands .
the fluxes from the transient component seem to be non - negligible even in the last epoch data ( @xmath150 days ) , because the last irac epochs were still in the rapidly decreasing phase .
therefore , we consider the fluxes of the last epoch irac data to be the upper limit fluxes of the host galaxy
. the multi - band magnitudes of the host galaxy of _ swift _ j1644 + 57 are shown in table [ hosttab ] .
we took the galactic extinction of photometric data into account based on @xcite .
we added the @xmath126-band photometric data from @xcite to expand the data points to the short wavelength band . the temporal change of the observed sed of _ swift _ j1644 + 57 are summarized in figure [ changefig ] .
the fluxes in redder bands show substantial changes with time . meanwhile , bluer band fluxes are constant .
this red feature of the transient has been suggested to be due to dust extinction . from previous studies
, it is known that the hydrogen column density of the source of x - ray is large ( @xmath151@xmath152 ) , meaning the line of sight to the smbh has a very large extinction value ( @xmath153 @xmath113 ) @xcite .
+ . two _
data are the upper limit fluxes .
the @xmath149 value for the fit is 1.64 .
[ sedfig ] ]
we performed sed model fitting of the multi - band fluxes of host galaxy of _ swift _
j1644 + 57 in order to determine the properties of the host galaxy such as the stellar mass ( @xmath154 ) which is an important parameter for the @xmath13 estimation .
we utilized the code fitting and assessment of synthetic templates ( fast;mariska / fast.html ] * ? ? ? * ) , which is a public sed fitting tool for the investigation of the galaxy properties using the photometric data ranging from uv to ir .
the code is based on the idl and fits of uv to ir stellar population templates to photometric data or galaxy spectra .
fast runs with the method of @xmath149 fitting and using stellar population grids to derive the best - fit model and its parameters .
we used the 2003 version of the bruzual & charlot ( bc03 ) model @xcite for the stellar population model .
there are three initial mass functions ( imfs ) available in the fast @xcite .
we chose the salpeter imf . to define the star formation history ( sfh ) , we assumed an exponentially decreasing star formation rate ( sfr ) .
the stellar population was modeled with the @xmath129-folding time scales , @xmath130 @xmath131[@xmath128/yr ] @xmath132 with a step size of 0.1 and ages of @xmath134 @xmath131[@xmath133/yr ] @xmath135 with a step size of 0.1 .
we used several metallicity values such as z = 0.004 , 0.008 , 0.02 , and 0.05 .
the model seds were attenuated by dust , for which we used attenuation curves based on @xcite .
we adopted @xmath138 @xmath136 @xmath139 with a step size of 0.1 .
all the input parameters for the sed fitting are summarized in table [ sedtab ] .
figure [ sedfig ] shows the sed fitting result .
the two _ spitzer _ data were treated as the upper limit fluxes of the host galaxy .
the estimated stellar mass of the host galaxy is @xmath3 .
the @xmath129-folding time scale is @xmath155 gyr and the age of the stellar population is @xmath156 gyr .
the sfr of galaxy is @xmath157/yr , and the specific sfr is @xmath158 .
@xcite derived sfr of @xmath112
@xmath159/yr from the h@xmath160 and [ o ii ] emission line luminosities .
the value of @xmath161/yr from h@xmath160 is consistent with our 1@xmath162 upper limit .
the sfr from [ o ii ] line ( @xmath159/yr ) is about twice larger but the [ o ii ] line based sfrs are known to be dependent on physical condition such as the reddening and metallicity ( e.g. , * ? ? ?
* ) , and less reliable than h@xmath160 based sfrs .
another possible cause of the discrepancy is the different timescales that are probed by different sfr indicators ( emission line indicators probing recent star formation ) . @xcite
estimated @xmath163mag , i.e. , no extinction using the intensity ratio of h@xmath160 and h@xmath164 lines .
this is consistent with our sed fitting result @xmath165 .
the @xmath149 value for the fit is 1.64 . the host galaxy of _ swift _
j1644 + 57 is a low mass , low sfr galaxy with a low extinction .
also it seems to have experienced a rapid decline of sfr not very long ago .
this fits in well with a recent suggestion by @xcite that host galaxies of tidal disruption events are e+a galaxies with @xmath166 gyr stellar population and low or no sfrs .
the output parameters are given in table [ sed2tab ] .
the errors correspond to 1@xmath162 confidence intervals derived from 100 times monte carlo simulations in which the input photometric data are changed according to their errors .
we also tried the chabrier imf instead of the salpeter imf for the fit .
the change of the imf influenced to the stellar mass , decreasing the stellar mass by @xmath167 dex .
we also fitted the sed model with the @xcite stellar population instead of the bc03 model .
we set the input parameter ranges identical to the case of the bc03 stellar populations .
the results were nearly identical to the bc03 result .
our analysis of the host galaxy shows that the host galaxy is a bulge - dominated and nearly extinction free ( @xmath168 mag ) . on the other hand ,
the spectral properties of the nuclear transient suggests a high extinction ( @xmath169 mag ) .
these two facts may appear contradictory , but we note that a significant amount of dust can be found easily in nuclear region of bulge - dominated galaxies when their nuclei are acitve . for example , hosts of luminous agns are mostly early - type , bulge - dominated galaxies ( e.g. , * ? ? ? * ) , and such agns are known to contain a significant amount of dust in nuclear region as a form of hot or warm dusty torus ( e.g. , * ? ? ?
our results on the properties of the host galaxy of _ swift _ j1644 + 57 can be summarized as follows .
it is a bulge - dominant galaxy ( b / t=@xmath1 ) .
the mass of the host galaxy is somewhat low at @xmath170 , even though the galaxy is bulge - dominated .
now we estimate the mass of the smbh that played the main role in the transient phenomenon .
it is now generally accepted that the smbhs ( @xmath171 @xmath172 ) reside in the bulges of all massive galaxies .
tight scaling relations have been derived between smbh mass and several physical properties of the bulges ( velocity dispersion , mass , luminosity , etc . ) in many previous studies @xcite .
some argue that ellipticals and classical bulges follow identical relations , while the pseudobulges follow a somewhat different relation with large scatter @xcite .
we conclude from the best - fit galaxy models that the host galaxy of _ swift _ j1644 + 57 has a classical bulge , and have also found a minor possibility of the pseudobulge with b / t=@xmath173 . for now , we consider only the best model , that is , the case of the host galaxy having a classical bulge and being bulge - dominant . and
the arrow indicates the upper bound value .
[ bhfig ] ] in order to esimate the central smbh mass in the host galaxy of _ swift _ j1644 + 57 , we used the scaling relation between @xmath13 and the stellar mass of the bulge ( @xmath8 ) .
we expect that a large part of the stellar mass derived in [ sec : sed ] belongs to the bulge component . @xcite present the @xmath7
@xmath8 relation , where @xmath8 is directly obtained from the bulge luminosity ( @xmath9 ) of _ spitzer _
3.6@xmath24 m and the calibrated @xmath8 @xmath9 relation .
they excluded pseudobulges when constructing the relation .
the relation is @xmath174,\ ] ] where @xmath175 , @xmath176 , and the intrinsic scatter is @xmath177 .
the estimated mass of the smbh is @xmath178 based on the stellar mass of the host galaxy and the above relation . if we consider the b / t=0.83 and assume that the mass - to - light ratio is constant in the bulge and disk , the stellar mass is decreased by @xmath179 dex .
it leads to a decrease in @xmath7 by @xmath179 dex .
the tight scaling relations between @xmath7 and host galaxy properties suggest a close link between the smbh growth and the galaxy evolution . there may be a cosmic evolution of the scaling relations , for which there have been various studies @xcite .
the evolution of the scaling relations is still controversial in terms of the selection effects in the high redshift regime .
nevertheless , we can consider a case where the growth of @xmath7 happened ahead of the assembly of the stellar mass as suggested by some of these studies .
@xcite suggest the redshift evolution out to @xmath180 , in the form of @xmath181 , based on 11 x - ray - selected broadline agns . considering this evolution effect and the redshift of _ swift _
j1644 + 57 @xmath182 , the mass of the smbh could be larger by @xmath183 dex .
we also estimated @xmath7 through the @xmath7
@xmath10 band luminosity of bulge ( @xmath184 ) relation in @xcite , assuming that most of the nir fluxes come from the bulge .
their relation is much improved compared with previous studies in view of serveral things .
they excluded galaxies with black hole monsters which have over - massive smbhs despite having relatively small bulges and ellipticals .
they also omitted galaxies with black hole masses that are measured based on the kinematics of ionized gas without taking line widths into account , since this method may yield underestimated masses .
galaxies in the process of merging generally have low - mass black holes for their luminosities .
thus they excluded these galaxies from the relation .
they also did not include pseudobulges .
the relation of @xcite is @xmath185 where the @xmath160 is @xmath186 , the @xmath164 is @xmath187 , and intrinsic scatters is 0.30 .
@xmath188 is the @xmath10-band absolute magnitude of the bulge based on the photometric system of 2mass . using the best - fit sed derived earlier , we applied k - correction and evolution correction .
we find @xmath189 vega mag including the evolutionary correction of 1.38 mag , which is derived by the difference in the @xmath10-band magnitude between 0.63 gyr old population as in table [ sed2tab ] and 4.5 gyr old population , which is the age of this host galaxy in the local universe .
then , the mass of the smbh is estimated to be @xmath190 . if we take b / t=0.83 into account and assume that this value is also applicable to the nir bands , then @xmath7 decreases by @xmath179 dex .
if we use an @xmath191 value that has not been corrected for evolution , then @xmath7 becomes @xmath192 dex larger .
the tidal disruption of normal stars by a black hole is not possible for @xmath193 , since the tidal radius where the disruption can occur is located inside the schwarzschild radius @xcite .
our results on the mass of the smbh satisfy the condition for a tidal disruption event .
although we favor the model in which the host galaxy of _ swift _ j1644 + 57 is a classical bulge , our analysis shows that it could be a galaxy with a pseudobulge with b / t=0.36 ( table [ galfittab ] ) .
if so , it is rather difficult to obtain an @xmath13 value , since the scaling relation is not well established for pseudobulges , especially in the low mass range of @xmath194 for the host galaxy .
several works have shown that the @xmath195 host galaxy scaling relations are weak or zero with a large scatter for pseudobulges . over the @xmath196 range of @xmath197 to @xmath198 where such a relation has been studied
, @xmath13 can have any value between @xmath199 to @xmath200 ( figure 21 of * ? ? ?
* ) . to reach down to @xmath201 as implied from the pseudobulge fit of our data ,
currently one can barely do so by relying on results from low mass agns @xcite .
in such a case , an @xmath13 value between @xmath202 to @xmath203 is possible ( figure 32 of * ? ? ? * ) .
overall , if the host galaxy harbors a pseudobulge , then we can only loosely constrain @xmath13 to have a value between @xmath202 to @xmath19 considering our current poor knowledge of the @xmath13 value in pseudobulges .
it is also known that a small fraction of pseudobulges have a srsic index of @xmath204 .
the best example is pox 52 , for which @xmath205 @xmath206 , @xmath207 , and @xmath208 @xcite . therefore , even if we accept the srsic index of @xmath108 as the best - fit result , we need to keep this kind of caveat in mind .
figure [ bhfig ] shows our overall results on @xmath7 and the results from the previous studies we mentioned in [ sec : intro ] .
it shows that our favorite results are compatible with the previous rough estimates from @xcite and @xcite , who also used scaling relations .
however , our results are improved compared to the previous results , by revealing that the host galaxy has a significant bulge component through a two - dimensional bulge + disk decomposition of the surface brightness profile , and removing the transient component in nir light using a long - term light curve .
the @xmath13 limit could be much looser ( the dashed line ) if the host galaxy harbors a pseudobulge .
a critical test of the pseudobulge model would be to obtain a deep , high - resolution image to see how the surface brightness profile behaves at the outer region of the host galaxy . +
we investigated the host galaxy properties of tidal disruption event , _ swift _
j1644 + 57 through morphology analysis , light curve analysis , and sed fitting .
we also estimated @xmath7 which played the main role of this phenomenon , through scaling relations .
we decomposed the surface brightness profile of the host galaxy based on high - resolution _ hst _
wfc3 images .
we found that the host galaxy of _ swift _
j1644 + 57 is a bulge - dominated galaxy which is well described by a single srsic model with the srsic index , @xmath0 .
if we add a disk component , the bulge to total host galaxy flux ratio ( b / t ) is @xmath1 , still indicating a bulge - dominant galaxy .
we conclude that the host galaxy of _ swift _
j1644 + 57 has a classical bulge from the best - fit galaxy models , although we can not completely exclude the possibility of this galaxy containing a pseudobulge with b / t=@xmath173 .
the nir light curves enabled us to isolate the fluxes from the host galaxy after @xmath209 days following the dissipation of the x - ray flux . on the other hand
, we found that there are no significant changes in the light curves of the short wavelength bands , supporting the red feature of the transient possibly being caused by severe dust extinction .
we fit seds to the multi - band fluxes of the host galaxy which are derived in the light curve analysis .
the estimated stellar mass of the host galaxy is @xmath3 .
the @xmath129-folding time scale @xmath128 is @xmath146 gyr and the age of stellar population is @xmath5 gyr .
the sfr of galaxy is @xmath210/yr . in terms of the surface brightness profile and the stellar mass ,
this galaxy resembles m32 , a small companion galaxy of m31 .
we estimated the central @xmath7 through scaling relations .
the mass of the smbh is estimated to be @xmath178 from @xmath7 @xmath8 and @xmath7 @xmath9 relations for the @xmath10 band .
however , the limit on @xmath7 can be much looser if the host galaxy has a pseudobulge .
future high - resolution , deep imaging should be able to unambiguosly distinguish the two possibilities . +
this work was supported by the national research foundation of korea ( nrf ) grant , no .
2008 - 0060544 , funded by the korea government ( msip ) .
we thank the observers who obtained the cquean and ukirt data that were used in our analysis .
this paper includes the data taken at the mcdonald observatory of the university of texas at austin . at the time of the ukirt observation
, ukirt was operated by the joint astronomy centre on behalf of the science and technology facilities council of the u.k .
this work is based in part on observations made with the spitzer space telescope , which is operated by the jet propulsion laboratory , california institute of technology under a contract with nasa .
we acknowledge the use of public data from the swift data archive .
cp , ts , and ng acknowledge support from the nasa research grant , nnx10af39 g .
mi gratefully acknowledges the hospitality and the support from the korea institute of advanced study where part of this was carried out . | we study the host galaxy properties of the tidal disruption object , swift j164449.3 + 573451 using long - term optical to near - infrared ( nir ) data .
first , we decompose the galaxy surface brightness distribution and analyze the morphology of the host galaxy using high resolution _ hst _
wfc3 images .
we conclude that the host galaxy is a bulge - dominant galaxy that is well described by a single srsic model with srsic index @xmath0 . adding a disk component , the bulge to total host galaxy flux ratio ( b / t ) is @xmath1 , which still indicates a bulge - dominant galaxy .
second , we estimate multi - band fluxes of the host galaxy through long - term light curves .
our long - term nir light curves reveal the pure host galaxy fluxes @xmath2 days after the burst .
we fit spectral energy distribution ( sed ) models to the multi - band fluxes from the optical to nir of the host galaxy and determine its properties . the stellar mass , the star formation rate , and the age of stellar population are @xmath3 , @xmath4/yr , and @xmath5 gyr .
finally , we estimate the mass of the central super massive black hole which is responsible for the tidal disruption event . the black hole mass is estimated to be @xmath6 from @xmath7 @xmath8 and @xmath7 @xmath9 relations for the @xmath10 band , although a smaller value of @xmath11 can not be excluded convincingly if the host galaxy harbors a pseudobulge . |
relativistic mean - field ( rmf ) models are widely used to construct a realistic equation of state ( eos ) , which could satisfy various experimental constraints @xcite .
the most challenging task is to reconcile the constraint on the pressure in the isospin - symmetrical matter ( ism ) , the so - called `` flow constraint '' @xcite , which favors a soft eos , and the existence of the neutron star ( ns ) , with the mass @xmath1 @xcite ( @xmath2 is the sun mass ) , which favors a stiff eos . in standard rmf models
hyperons may appear in ns cores .
this results in a decrease of the maximum ns mass below the observed limit , cf .
@xcite and references therein .
the appearance of @xmath0 isobars in nss would lead to a further decrease of the maximum ns mass @xcite .
these problems are dubbed in the literature as the hyperon and @xmath0 puzzles . in
@xcite we proposed an rmf model with scaled hadron masses and coupling constants ( labeled there as mkvor ) , which solves the hyperon puzzle and fulfills successfully the maximum mass constraint , the flow constraint and various other constraints , but the @xmath0s were not included . here
we show that our model passes the constraints also with the inclusion of @xmath0s .
a more detailed consideration can be found in @xcite .
the lagrangian of the model is formulated in @xcite .
the model is a generalization of the non - linear walecka model with effective coupling constants @xmath3 and hadron masses @xmath4 dependent on the @xmath5 field @xcite , here @xmath6 lists the included mesonic fields , @xmath7 indicates the baryon species ( @xmath8 ; @xmath9 is the @xmath10 hyperon ; @xmath0 labels the @xmath0 isobar ) and @xmath11 labels the hadrons ; @xmath12 and @xmath13 are the corresponding scaling functions . in the rmf approximation
the contribution of @xmath0s to the energy density has the same form as for spin-@xmath14 fermions but with the spin degeneracy factor @xmath15 . the energy density of our model , @xmath16 = \sum_b e_{\rm kin}\big(p_{{\rm f},b } , m_b^ * ( f ) , s_b \big ) + \sum_{l = e,\mu } e_{\rm kin}(p_{{\rm f},l},m_l , s_l ) + \frac{m_n^4 f^2}{2 c_\sigma^2 } \eta_\sigma(f ) \nonumber \\ & + \frac{1}{2 m_n^2 } \big[\frac{c_\om^2 n^2 } { \eta_\om(f ) } + \frac{c_\rho^2 n_i^2}{\eta_\rho(f ) } + \frac{c_\phi^2 n_s^2}{\eta_\phi(f ) } \big ] , \ , n=\sum_b x_{\om b } n_b,\ , n_i=\sum_b x_{\rho b } t_{3 b } n_b,\ ,
n_\phi=\sum_{h } x_{\phi h } n_h , \label{edensity}\end{aligned}\ ] ] is expressed in terms of the scalar field @xmath17 , @xmath18 and particle densities @xmath19 , where @xmath20 , @xmath21 is the fermion spin and @xmath22 stands for the isospin projection of baryon @xmath23 .
the fermion kinetic energy density is defined as @xmath24 .
meson coupling constants , masses and scaling functions enter the energy density only in combinations @xmath25 therefore , studying infinite matter we do not have to specify the coupling constants , meson masses and their scaling factors separately . the baryon mass scaling function
is @xmath26 .
the particle densities as functions of the total baryon density @xmath27 in beta - equilibrium matter ( bem ) follow from the conditions : @xmath28 , where @xmath29 is the baryon charge and @xmath30 are the fermion chemical potentials ; @xmath31 ; and the charge neutrality condition @xmath32 .
these equations are solved self - consistently with the equation of motion for the scalar field @xmath33 .
finally , the pressure is given by @xmath34 .
the parameters of the model in the nucleon sector are fitted to reproduce the nuclear matter saturation properties , which are defined as coefficients of the taylor expansion of the energy per particle in ism in terms of @xmath35 and @xmath36 , @xmath37 and @xmath38 .
we consider an extension of the mkvor model introduced in @xcite , labeled as mkvor * , for details see @xcite .
the input parameters are the same as in the mkvor model : the nuclear saturation density @xmath39 , the binding energy @xmath40 , the incompressibility @xmath41 , the symmetry energy @xmath42 and the nucleon effective mass @xmath43 . scaling functions of the mkvor * model as functions of the scalar field
are shown in figure [ eta_f ] .
compared to mkvor model in the @xmath44 function an additional sharp decrease at @xmath45 is introduced .
a sharp decrease of at least one of @xmath46 scaling functions for @xmath47 leads to a stiffening of the eos , cf .
it limits the scalar field growth with the increasing density and prevents the nucleon effective mass from vanishing , which would occur in the original mkvor model in the ism if @xmath0 were included . dashes in figure [ eta_f ]
mark maximum values of the scalar field @xmath48 , reachable in the nss . in the bem
all results for mkvor and mkvor * models coincide for densities reachable in nss .
the behavior of the scaling functions for @xmath49 is irrelevant in bem .
the hyperon and @xmath0 couplings with vector mesons are related to the nucleon ones via the su(6 ) symmetry relations @xcite : @xmath50 coupling constants with the @xmath5 meson follow from the baryon potentials in ism at the saturation density @xmath51 , where we set @xmath52 , @xmath53 , and @xmath54 .
the @xmath0 potential is poorly constrained by the data .
we explore the range @xmath55 , keeping in mind that the realistic value of @xmath56 is close to the nucleon one @xmath57 .
the scaling function for @xmath58 meson is chosen as @xmath59 , which corresponds to the @xmath60 set of models in @xcite .
below we label the mkvor * model with @xmath0s and without hyperons as mkvor*@xmath0 , and the model with @xmath0s and full baryon octet with @xmath60 scaling as mkvor*h@xmath61 .
neutron star masses and radii are obtained by integrating the tolman
oppenheimer volkoff equation . for the bem for densities @xmath62
our eos is supplemented by the eos of the ns crust .
the details can be found in @xcite .
concentrations in ism as functions of the density .
lines are labeled with the potentials @xmath56 , in @xmath63 .
right panel : pressure as a function of the density for our model without and with @xmath0s for @xmath64 , @xmath65 and @xmath66 . for @xmath65 and @xmath67 bold lines
show the maxwell constructions and thin lines show the original pressure curves .
hatched area shows the flow constraint from @xcite . ]
on the left panel in figure [ dsym ] we show the @xmath0 concentrations , determined in the ism by the chemical equilibrium condition @xmath68 . we see that with a decrease of @xmath56 the critical density for the @xmath0 appearance decreases .
for @xmath69 the @xmath0 baryons appear in a third - order phase transition .
interestingly , for @xmath70 there exists a density range with multiple solutions for @xmath71 at a given density . among these solutions the one with greater @xmath71 is energetically favorable .
this means that @xmath0s appear with a jump from @xmath72 to a finite value of @xmath71 in a first - order phase transition .
the pressure in the ism as a function of the density is shown on the right panel in figure [ dsym ] .
for @xmath73 in the equilibrium the system follows the maxwell - construction line . for @xmath70 the region with multiple solutions
manifests itself also in the pressure .
note that for @xmath74 the equilibrium pressure curve lies fully within the flow constraint .
it means that if in the future the constraint sharpens , it could be used to constrain the @xmath0 potential . as functions of the density in bem in the mkvor*@xmath0 ( left panel ) and mkvor*h@xmath61 ( right panel ) models for @xmath75 and @xmath76 . ]
the particle fractions as functions of the density in the bem are presented in figure [ nd_bem ] for the mkvor*@xmath0 model ( left panel ) and for the mkvor*h@xmath61 model ( right panel ) , for @xmath75 and @xmath77 . with an increase of the density
@xmath78s appear in both cases at density @xmath79 for @xmath75 and @xmath80 for @xmath81 .
the @xmath0 fraction increases sharply , and for the mkvor*@xmath0 model reaches @xmath82 and @xmath83 for @xmath84 and @xmath67 , respectively . as in @xcite , the inclusion of hyperons reduces the amount of @xmath0s at a given density . ,
mkvor*@xmath0 ( @xmath81 ) and mkvor*h@xmath61 ( @xmath85 and @xmath67 ) models .
right panel : the gravitational ns mass as a function of the baryon mass for mkvor*h@xmath61 model for @xmath75 ( @xmath0s do not yet appear for stars with masses in the shown range ) and @xmath81 .
shaded rectangle shows the constraint from @xcite , two empty rectangles show the change of the constraint at the assumption of progenitor mass loss by @xmath86 and @xmath87 . ] despite the @xmath0 fraction in the ns medium is large , the effect of @xmath0s on the ns masses and radii appears to be small . on the left panel in figure [ nsprops ]
we show the ns mass as a function of the central density for our models , and on the middle panel the mass - radius plot is shown .
the maximum decrease of a ns mass at a given central density in mkvor*h@xmath61 model is less than @xmath88 for @xmath75 and does not exceed @xmath89 for @xmath81 .
the decrease of the maximum ns mass is even smaller , being less than @xmath90 for @xmath81 .
the change of the ns radius for a given ns mass is not more than @xmath91 .
thus , within the mkvor*h@xmath61 model both @xmath0 and hyperon puzzles are resolved .
an interesting observation is that the inclusion of @xmath0s allows to improve the constraint extracted from analysis of gravitational and baryon masses of the pulsar j0737 - 3039(b ) @xcite . on the right panel in figure [ nsprops ]
we show the gravitational ns mass as a function of the baryon ns mass for @xmath75 and @xmath76 together with the constraint from @xcite . for @xmath75
the curve is the same as in the case without @xmath0s , since @xmath0s do not appear at the densities corresponding to the central density of the ns with @xmath92 . for @xmath81 the constraint
is passed better .
we studied the effect of inclusion of @xmath0 isobars on the eos of the ns matter with hyperons within the mkvor - based rmf models of @xcite with scaled hadron coupling constants and masses .
we varied the @xmath0 potential at the saturation density , @xmath56 , in the range @xmath93 to estimate the maximum effect of the @xmath0 appearance on ns properties . for the ism @xmath0 baryons can appear in third- or first - order phase transitions , depending on the value of @xmath56 . in the ns matter
the @xmath0 fraction is not small for @xmath93 , which we studied . in spite of that ,
a decrease of the maximum ns mass does not exceed @xmath94 , so the maximum mass constraint remains satisfied . in the presence of @xmath0s the constraint on the relation between the gravitational and baryon masses of j0737 - 3039(b ) @xcite proves
to be better fulfilled .
the reported study was funded by the russian foundation for basic research ( rfbr ) according to the research project no 16 - 02 - 00023-a .
the work was also supported by the slovak grant no .
vega-1/0469/15 , by `` newcompstar '' , cost action mp1304 and by the ministry of education and science of the russian federation ( basic part ) .
99 klhn t _ et al .
_ 2006 _ phys .
_ c * 74 * 035802 . | knowledge of the equation of state of the baryon matter plays a decisive role in the description of neutron stars . with an increase of the baryon density
the filling of fermi seas of hyperons and @xmath0 isobars becomes possible .
their inclusion into standard relativistic mean - field models results in a strong softening of the equation of state and a lowering of the maximum neutron star mass below the measured values .
we extend a relativistic mean - field model with scaled hadron masses and coupling constants developed in our previous works and take into account now not only hyperons but also the @xmath0 isobars .
we analyze available empirical information to put constraints on coupling constants of @xmath0s to mesonic mean fields .
we show that the resulting equation of state satisfies majority of presently known experimental constraints . |
domains of different magnetic orientations spontaneously nucleate in all magnetic materials in order to minimize the magnetostatic energy ( dipolar interaction ) of the system .
these domains are separated by domain or bloch walls where the magnetization is changing over , rapidly , from one orientation to an other . in this paper
we shall develop a theoretical account of these transition regions in terms of mobile electrons and their spins , that is to say from first principles .
the understanding and control of the domain structure , magnetic morphology , is central to the effective use of magnetic materials in many technologies @xcite .
moreover , the nature , shape , and motion of bloch walls have always attracted fundamental scientific interest @xcite .
nevertheless , all the theoretical discussions so far , of the entire subject , have been strictly phenomenological . in the present paper
we wish to lay the foundation of a first - principles approach to the problem . to start with a tractable problem , henceforth , we shall focus on that of a single bloch wall .
note that whilst its width , @xmath1 @xmath2 , much larger than the lattice spacing @xmath3 , it is smallest of all the other length scales such as the sizes of the domains , @xmath4 @xmath5 , magnetic texture , @xmath6 @xmath7 , etc . in the problem .
thus , by deploying our first - principles methodology , designed to treat variations in properties on the length - scale @xmath3 , to bloch wall problems , we are making the natural choice of crossing the smallest scale - gap first .
the established theory of bloch walls is a topic in micro - magnetics @xcite and it is wholly phenomenological .
it consists of the equations of landau and lifshitz @xcite as generalized by brown @xcite . for clarity and easy future reference ,
we summarize here the part of this theoretical framework relevant to our present concern . in short ,
the magnetization density is described by the vector field @xmath8 , which evolves in time according to the landau - lifshitz equation , @xmath9}{\delta \vec{m } } \label{eq : ll}\ ] ] where @xmath10 is the bohr magneton , @xmath11 is the planck constant and @xmath12 $ ] is a generalized free - energy functional of @xmath8 . for our present purposes it will be sufficient to take the @xmath12 $ ] to be given by the usual ginzburg - landau expansion , @xmath13 \approx \int d\vec r \left\
{ { \cal a } \sum_{i=1}^{3 } |\nabla m_i|^2 + \sum_{i , j=1}^{3 } { \cal b}_{ij}m_im_j + ... \right\ } \label{eq : lg}\ ] ] where @xmath14 and @xmath15 refer to cartesian axes @xmath16 and @xmath17 , and the coefficients @xmath18 , @xmath19 are material specific parameters which depend on such thermodynamic variables as temperature , @xmath20 , and pressure , @xmath21 . evidently , the equilibrium magnetization , @xmath22 , which minimizes @xmath12 $ ] is a stationary solution of eq .
( [ eq : ll ] ) .
the simplest bloch wall like solution of these equations follow from parameterizing the cartesian components @xmath23 in terms of spherical coordinates @xmath24 and @xmath25 ( see fig . 1 . ) @xmath26 where @xmath27 is a constant equal to the saturation magnetization , and stopping the expansion in eq .
( [ eq : lg ] ) at the @xmath28 order . for a cubic system ( @xmath29 )
this means @xmath30 where the new coefficients are defined as @xmath31 and @xmath32 , respectively .
taking @xmath33 and @xmath25 a function of the @xmath17 coordinate only , it can be readily shown @xcite that , subject to the boundary conditions @xmath34 and @xmath35 , the function @xmath36 minimizes @xmath37 in eq .
( [ eq : fcubic ] ) . in
what follows we shall study this simple example of a 90@xmath38 bloch wall from the point of view of mobile electron whose spin gives rise to such magnetic configuration .
given the success and the general satisfaction with the established phenomenological theory outlined above one may , at this stage , protest that such microscopic considerations are not needed .
to answer this objection we shall now pause briefly to argue the case to the contrary . in the interest of economy we present our arguments as a list of brief statements : * firstly , the phenomenological theory assumes that @xmath39 rotates as @xmath17 goes from @xmath40 to @xmath41 without a change in magnitude .
a first - principles theory should tell us if and when this is true .
* because , as yet , experiments can not tell us otherwise @xcite , the phenomenological theory accepts the form of @xmath42 in eq .
( [ eq : soliton ] ) as immutable .
by contrast , the microscopic theory could discover situations ( materials ) where this is not
the same thing can be said about deviations of @xmath24 from @xmath43 as a function of @xmath17 or , indeed , the @xmath44 dependence of both @xmath24 and @xmath25 .
* of course , first - principles calculations not only yield @xmath45 but also the changes in the electronic structure due to the deviation of @xmath45 from its saturation value , namely the presence of a bloch wall .
in other words , they describe the electronic structure which supports , consistent with , a bloch wall .
this novel information becomes available for the first time in the calculations we shall report . * the change in the electronic states due to the presence of a bloch wall is , or
can be viewed as a description of the electron
bloch wall interaction .
clearly , a study of this interaction will make an important contribution to understanding the widely observed magneto - transport phenomena associated with electrons scattering off bloch walls @xcite .
* inevitably , a magnetic inhomogeneity will distort the underlying lattice and such distortions give rise to bloch wall
bloch wall interactions and pinning of bloch walls by lattice defects like impurities vacancies or grain boundaries @xcite .
although we do not deal with the appropriate generalization of our first - principles approach , we wish to stress that it can be readily adopted to address these issues and this possibility is an important part of the motivation for perusing it . finally , we comment on our choice of bcc iron as the host to the bloch wall in our calculations .
as is well known the saturation magnetic moment per atom @xmath46 at @xmath47 and the effective moment @xmath48 deduced from the curie constants , measured at @xmath49 , for metallic ferromagnets are usually not the same .
in fact , the deviation of @xmath46/@xmath48 from 1 can be taken as a good measure of how independent the local magnetic moments are from their relative orientation . since
bcc iron is a famously good moment system , namely @xmath50 , we expect the first - principles calculations to map optimally onto the phenomenological theory summarized by eqs .
( [ eq : fcubic ] ) and ( [ eq : soliton ] ) .
this is indeed what we find and we take this fact as a validation of our conceptual framework as well as our numerical procedures for implementing it .
the prize to pay for having done this easy case is that the importance of all the novel features listed above ( _ a .- e .
_ ) are minimized .
thus , if we disregard these interesting but small effects , our calculations reduce to a new , if complicated , way of computing the material dependent parameters , e.g. the spin - wave stiffness constant @xmath51 and the anisotropy parameter @xmath52 . as these
are also available from other type of first - principles calculations @xcite , the comparison of our results with those of others in the field is a useful exercise in its own right . in the next section
we shall describe the first - principles theoretical framework , based on a relativistic , spin - polarized density functional theory , for our calculations.this is followed by a section describing the computational procedures .
we shall present and discuss our results in sec .
4 . , whereas in sec . 5 .
we shall evaluate the progress we have been able to make and the prospects of the microscopic approach we have advocated .
a fully relativistic density functional theory ( dft ) @xcite , which includes the dipolar interaction between electrons , would , presumably , yield an inhomogeneous ground state with the domain structure determined by the size and shape of the sample . clearly , the corresponding calculations are out of the question and we shall follow the logic of the phenomenological theory .
namely , we neglect the dipolar interaction and study a single bloch wall engendered by the constraint that the magnetization @xmath45 is oriented along two different easy axes at @xmath53 . this problem is readily encompassed by the spin - polarized relativistic density functional theory in the local density approximation ( lda ) as usually applied in solid - state physics @xcite .
there are only two features of the way we shall proceed which deserve further general comment .
firstly , we note that we shall not be looking for a ground state , as usual , but a lowest energy state consistent with a constraint which prescribes a symmetry different from the ground state . fortunately
, as it is well known , dft covers this eventuality @xcite .
secondly , we shall not attempt to address the vexing conceptual difficulties that arise in connection with imposing constraints while solving self - consistently the kohn - sham - dirac equation of the theory @xcite , but in this preliminary exploration of the subject implement our strategy in the simplest possible way as outlined below .
there is no reason to doubt that the established phenomenological theory @xcite in the introduction gets the essential physics of the bloch wall formation right .
from the point of view of formulating a first - principles version of this theory it may be summarized as follows : the changing orientation of @xmath39 , namely the variation of @xmath42 in eq .
( [ eq : soliton ] ) , across the bloch wall implies an exchange energy cost which is lower the slower the variation , whereas the anisotropy energy favors a rapid change from one easy direction , @xmath54 , to the other , @xmath55 , and the width of the transition region of @xmath42 in eq .
( [ eq : soliton ] ) , @xmath56 , is determined by the balance of these two tendencies .
indeed , we can bypass the solution of the euler - lagrange equation , which in the case of minimizing the free - energy @xmath37 in eq .
( [ eq : fcubic ] ) is the famous sine - gordon equation whose solution is @xmath42 in eq .
( [ eq : soliton ] ) , and take @xmath42 to be a simple function , @xmath57 , which goes from 0 to @xmath43 in a distance of @xmath58 and minimize the corresponding free - energy @xmath59 with respect to the width @xmath58 . as can be readily shown , by substituting @xmath57 into eq .
( [ eq : fcubic ] ) , for @xmath33 @xmath60 where @xmath61 clearly , the first term in eq .
( [ eq : fsimple ] ) , proportional to 1/@xmath58 , is due to the exchange interaction measured by @xmath51 and the second term , which is proportional to @xmath58 and the constant @xmath52 , represents the contribution of the magneto - crystalline anisotropy energy for cubic systems .
furthermore , @xmath59 reaches its minimum at @xmath62 given that @xmath63 and @xmath64 are simple dimensionless numbers which parameterize the profile @xmath57 , this result agrees well with the exact soliton solution for @xmath65 .
evidently , the above discussion suggests a rather straightforward strategy for first - principles calculations : * assume a profile @xmath57 . *
carry out a density functional calculation for the magnetization @xmath39 constrained to follow the profile @xmath57 and calculate the total energy or grand potential @xmath66 . *
minimize @xmath66 with respect to @xmath58 . for the profile we shall always take the particularly simple choice @xmath67 which is compared to the soliton solution in fig . 2 . for this function @xmath68 and @xmath69 and hence if @xmath66 is given by eq .
( [ eq : fsimple ] ) @xmath70 which implies @xmath71 .
note that this bloch wall thickness is @xmath72 times larger than that deduced from eq .
( [ eq : soliton ] ) .
thus the bloch wall thickness is to some extent an ill - defined concept in the present prescribed profile approach to the problem .
as will be clear presently , our calculated @xmath66 can be fitted by the above functional form very accurately and hence it determines the coefficients @xmath51 and @xmath52 from first principles . for orientation we note that via the landau - lifshitz equation , eq .
( [ eq : ll ] ) , the free - energy functional given in eq .
( [ eq : fcubic ] ) implies a spin - wave dispersion relation . in the long wave - length limit
this yields @xmath73 thus , the spin - wave stiffness constant @xmath74 and the usual cubic ( fourth - order ) anisotropy constant @xmath75 as expected , in the presence of magneto - crystalline anisotropy the spin - wave spectrum is gaped . without any account of dipolar interactions , in non - relativistic quantum mechanics
the orientation of the magnetization is independent from the crystal axis .
as is clear from the above discussion , under this circumstance ( @xmath76 ) there is no bloch wall of finite width .
thus , our calculation of @xmath66 must include spin - orbit coupling which is the major source of magneto - crystalline anisotropy in transition metals @xcite .
rather than treating the spin - orbit interaction in perturbation theory , we use a fully relativistic spin - polarized density functional theory @xcite .
the natural variable in the microscopic theory is the magnetization averaged over a unit cell @xmath77 where @xmath78 is the volume of the @xmath14th unit cell .
we may then define the local orientation @xmath79 by @xmath80 where @xmath81 denotes the length of the vector @xmath82 .
it is this effective local orientation that we wish to constrain to follow the discrete version of the profile in eq .
( [ eq : linbw ] ) .
namely , we assume that @xmath83 is the same within an atomic plane with position @xmath84 and take @xmath85 where for a 90@xmath38 bloch wall , which is @xmath86 layer thick , @xmath87 thus , we have to solve the kohn - sham - dirac equation of dft for the circumstance where the orientation of the magnetization , @xmath88 , is changing from layer to layer within a slab of @xmath86 atomic layers and is uniformally @xmath89 and @xmath90/2 to the left and right , respectively , of the slab as shown in fig . 2 .
evidently , the relativistic spin - polarized screened - kkr method @xcite , which treats just such geometries , is ideal for tackling this problem .
fortunately , the code which implements this method scales linearly with @xmath86 and , as we shall demonstrate presently , can handle 400 - 500 layers with readily available computer power .
we performed calculations proposed in the previous section by using the spin - polarized relativistic screened korringa - kohn - rostoker ( spr - skkr ) method @xcite .
although the method is by now well - established ( see also refs .
@xcite ) , for completeness and since we also made developments specific to the bloch wall problem , we briefly outline our computational strategy . for an ensemble of individual scatterers
the multiple scattering theory ( mst ) yields the one - electron green function for an arbitrary complex energy @xmath17 and real space coordinates @xmath91 , @xmath92 @xmath93 @xmath94 where @xmath95 and @xmath96 label two specific sites of position vectors @xmath97 and @xmath98 , while @xmath99 and @xmath100 denote pairs of angular momentum quantum - numbers @xmath101 and @xmath102 , respectively .
the functions @xmath103 and @xmath104 , regular and irregular at @xmath105 , respectively are properly normalized solutions of the kohn - sham - dirac equation related to a single , finite - ranged potential well for which we now assume a spherically symmetric effective potential , @xmath106 , and effective field , @xmath107 , pointing along a local ( positive ) @xmath17 coordinate axis . in short , @xmath108 where @xmath109 with the usual pauli matrices @xmath110 and the two - dimensional unity matrix @xmath64 .
the numerical solution of the above equation with a corresponding expression for the so - called single - site _
t_-matrix was originally given in refs .
a particular feature resulting from the approach they used is that , although the _ t_-matrix has necessarily off - diagonal elements , no coupling between different @xmath111 values is present .
we further simplify the problem by adopting the atomic sphere approximation ( asa ) in which the volume of the sphere is taken equal to the volume of the corresponding wigner - seitz cell .
clearly , asa deals with overlapping potentials which is , in strict sense , prohibited within the mst , however , mostly for inhomogeneous systems , this approach is widely used , since on one hand it gives a better description of the interstitial region than the muffin - tin approach , on the other hand it is conceptually much simpler than a full potential description .
as implied above , our approach allows the orientation of the magnetization to vary from site to site as required in the case of a bloch wall where we keep the orientation to be constant within the atomic planes but let it rotate from layer to layer around the global @xmath17 direction .
therefore , in each layer there will be a rotation @xmath112 , corresponding to @xmath33 and @xmath113 ( see eq . ( [ eq : phip ] ) ) , which relates the @xmath114-matrix in the local to that in the global coordinate system as follows @xmath115 where @xmath116 denotes a matrix containing , block - wise , the irreducible representations of @xmath112 .
as the solution of the above single - site problem is relatively easy and universal in applications of mst , the evaluation of the scattering path operator ( spo ) , @xmath117 in eq .
( [ eq : greenf ] ) , is , essentially , the main difficulty in such calculations . the geometrical arrangement of the scatterers involved in the system is put , in mst , into the so - called structure constants , @xmath118 ( @xmath119 ) , which for the relativistic case can be obtained by a transformation in terms of the clebsh - gordon coefficients from its non - relativistic counterpart ( see e.g. ref .
@xcite ) . by defining the corresponding matrices having both site and angular momentum indices @xcite @xmath120 the spo
is given by the following matrix inversion @xmath121 a particular problem arises from the fact that @xmath122 is long - ranged , therefore , the inversion in eq .
( [ eq : rsspo ] ) can not be directly performed . for a system with three - dimensional periodicity
the problem can be exactly handled by making use of the lattice fourier - transformation , which splits eq .
( [ eq : rsspo ] ) into the corresponding @xmath123-projections which has to be solved in angular momentum space only .
the bloch wall problem we deal with exhibits , however , two - dimensional ( 2d ) periodicity in the @xmath124 plane , while in the @xmath17 direction the translational symmetry is broken due to the variation of the orientation of the magnetization . thus employing two - dimensional lattice fourier - transforms
@xmath125 where @xmath21 and @xmath126 denote atomic layers , generated by @xmath127 and @xmath128 , respectively , @xmath129 are 2d lattice vectors and @xmath130 is a vector in the first 2d brillouin - zone ( bz ) , and the new matrix notation in terms of layer indices @xmath131 one can write @xmath132 evaluating the matrix inversion in eq .
( [ eq:2dspo ] ) is still demanding since the structure constants involved are long - ranged as far as the interlayer distances are concerned . in order to render this problem tractable the concept of screening has been introduced in the middle of nineties @xcite . without going into details , a canonical transformation of the _ t_-matrices and the structure constants , which leaves the green function invariant , in terms of repulsive scattering potentials makes it possible to reduce the spatial range of the effective structure constants . as
this transformation is independent of the real scatterers in the system , it is very useful to perform self - consistent calculations , since it has to be performed for the structure constants only once at the beginning of calculation . in terms of the screened quantities
( [ eq:2dspo ] ) has exactly the same form .
however , since @xmath133 is now well - localized , it can be truncated for @xmath134 at a given @xmath95 ( @xmath135 for fcc and bcc principal facets ) , which in turn implies a block - tridiagonal form for the matrix @xmath136 .
this blocks are related to the so - called principal layers containing @xmath95 subsequent atomic layers . by splitting our system into left and right perfect semi - infinite subsystems ,
in each of them the scatterers are all identical ( bulk ) , and into a central region , where the potentials as well as the orientation of the magnetization can vary , the projection of the spo , eq .
( [ eq:2dspo ] ) , onto the central region can be calculated _ exactly _ , i.e. taking into account all the scattering events to the left and right semi - infinite regions @xcite .
a remarkable feature of the method is that , if only the layer diagonal blocks of the spo need to be calculated , it scales linearly with the size of the central region , namely @xmath86 @xcite .
thus it opens the way for investigating systems with inhomogeneities extending much beyond the atomic scales such as bloch walls .
let us now turn to the task of performing the brillouin - zone integration @xmath137 where @xmath138 and @xmath139 , while @xmath140 denotes the volume of the 2d bz . in ref .
@xcite we described a method for reducing the demand of the above bz integration using the symmetry operations of the underlying lattice .
this is obviously useless for the present case of the 90@xmath38 bloch wall , since the direction of the magnetization rotates from layer to layer , say , from the @xmath141 axis to the @xmath142 axis , therefore the bloch wall itself is not invariant under any of the symmetry operations of the @xmath143 group characteristic to the bcc(001 ) bz . however , it is still useful to note that the magnetization direction described by the angle @xmath42 in a 90@xmath38 bloch wall perpendicular to the ( 001 ) direction of a bcc lattice and satisfying the boundary conditions @xmath144 and @xmath145 has the symmetry property @xmath146 taking into account also the symmetry of the underlying lattice this implies that the bloch wall is invariant under a @xmath147 rotation around the axis ( 110 ) which is in three dimensional space represented by the matrix @xmath148 whereby the axis of the rotation should cross the @xmath17 axis at @xmath149 , i.e. for which @xmath150 .
quite clearly , @xmath151 . also evidently , the 2d square bz is invariant under @xmath152 . to make use of this symmetry ,
let @xmath153 be the unitary matrix which represents @xmath152 in the @xmath101-space and @xmath154 denote the layer onto which a particular layer @xmath21 is mapped by @xmath152 .
then for the corresponding single - site @xmath114-matrices we can write @xmath155 by introducing @xmath156 or alternatively , by taking the choice of @xmath157 , @xmath158 it follows that @xmath159 a relationship between the 2d fourier - transformed structure constants with arguments @xmath130 and @xmath160 can also be established using the transformation of the real - space structure constants @xmath161 therefore , by using eq .
( [ eq:2dft ] ) we can proceed as follows @xmath162 where we defined @xmath163 and @xmath164 , both being 2d lattice vectors . thus , similar to eq .
( [ eq : umat1 ] ) , introducing @xmath165 eq .
( [ eq:2dgtraf1 ] ) can be written compactly @xmath166 clearly , the matrix @xmath167 can also be used instead of @xmath168 in eq .
( [ eq : ttraf2 ] ) , which immediately implies the following transformation property for the spo @xmath169 as the above relationship makes possible to calculate the spo only in half of the bz when performing the integration in eq .
( [ eq : bzint ] ) , the computational time and memory storage request of the computer code can be reduced by a factor of two .
alternatively , in the actual calculations we have made use of the above symmetry of the bloch wall by halving the computational demand of the inversion in eq .
( [ eq:2dspo ] ) , while keeping all the @xmath130 points in the bz .
although , this latter procedure is almost equivalent to what has been described above , as it facilitates the tridiagonal shape of the corresponding matrix @xcite , its use is rather limited to localized schemes whereas the former one can be regarded to be quite general for calculations in bloch wall problems . for our present purposes the one - electron green function , eq .
( [ eq : greenf ] ) , can be used to calculate several quantities of interest such as the electronic density of states ( dos ) @xmath170 where the energy @xmath171 is real , @xmath172 is a small imaginary part serving as parameter of lorentzian broadening , and @xmath173 denotes trace of a matrix in the four - dimensional dirac - space , the charge density @xmath174 where @xmath175 is a semi - circle contour in the upper complex semi - plane starting at the bottom of the valence band , @xmath176 , and ending at the fermi energy , @xmath177 , and the spin - density @xmath178 furthermore , the charge , the spin - moment and the band - energy can be obtained straightforwardly as @xmath179 @xmath180 and @xmath181 respectively . from eq .
( [ eq : greenf ] ) it is quite obvious that the above quantities can readily be resolved into components with respect to cells ( layers ) as we shall show them when we present are results in the forthcoming sections . by solving the poisson - equation for the electrostatic potential and employing spin - density functional theory in the local approximation @xcite for the exchange - correlation potential and exchange field
, self - consistent calculations can be performed .
first we carried out self - consistent calculations for bcc bulk iron , where we have used a theoretical lattice constant of @xmath3=5.204 a.u .
derived by recent careful full - potential calculations @xcite . in that calculation we kept the orientation of magnetization along the ( 100 ) easy axis and used 91 @xmath182-points in the irreducible wedge of the 2d bz ( ibz ) for the bz - integrations , eq .
( [ eq : bzint ] ) . for the energy integrals , eqs .
( [ eq : charge ] ) and ( [ eq : spin - moment ] ) we used 16 points along the semi - circle sampled according a gaussian - quadrature , while the corresponding summations in angular momentum space were subject to a cut - off of @xmath183=2 .
we converged the fermi energy to get the corresponding charge ( z=27 ) to an accuracy of 10@xmath184 electrons .
note , that in our present approach , because of the equilibrium with the left ( and right ) semi - infinite regions , this fermi level has to be also used in the calculations of the ( finite - size ) bloch wall .
as mentioned earlier , we have made use of different approximations to calculate the free - energy .
the first , computationally less demanding , one is based on the frozen - potential approximation frequently used also in magneto - crystalline anisotropy calculations ( see ref .
@xcite and refs .
therein ) .
briefly , we used the self - consistent bulk potential and effective field magnitude in each layer of the bloch wall and we set the orientation of the exchange field by rotating successively the bulk @xmath114-matrix from layer to layer , eq .
( [ eq : trot ] ) , according to the prescribed function @xmath185 . in this case
only one iteration , to calculate the single - particle ( band ) energy and the charges , were carried out . as the charge neutrality ( number of particles )
is not preserved for the fixed volume of the central region , the grand - canonical potential has to be considered . for any thickness @xmath86 of the bloch wall , taking always the difference with respect to the ferromagnetic state , this is approximated by @xmath186 the other approach is a fully self - consistent one .
unfortunately , the computer power available to us was sufficient for only a few such calculations . therefore , only for the case of @xmath187 shall we present and discuss the changes in the electronic structure and magnetic moments due to the bloch wall .
we have calculated the bloch wall formation energy @xmath188 for various values of @xmath86 using the frozen potential approximation .
note , that the summation in eq .
( [ eq : omegan ] ) has to be , in principle , taken over all the layers in the system .
our calculations show that layers more than about ten layers away from the wall do not contribute significantly to the sum although @xmath189 and @xmath190 by themselves differ considerably from zero even far away from the wall .
as well - known , in the phenomenological theory @xcite , the exchange energy and the magneto - crystalline energy contribute equally to the bloch wall energy .
experimentally , for bcc fe the magneto - crystalline anisotropy constant , @xmath191 , is found to be 0.3 @xmath192ryd per atom@xcite , while careful ( non - orbital - polarized ) lda calculations predict 0.1 - 0.2 @xmath192ryd ( see ref .
@xcite and refs .
therein and note that for a cubic system @xmath193 $ ] . )
this , on one hand , implies , that the bloch wall energy normalized to one layer is expected to be of the same order of magnitude as the anisotropy energy and , on the other , we have to calculate the bloch wall energy to the same accuracy as is necessary in magneto - crystalline anisotropy calculations .
the main difficulty arises from the fact that , in particular , close to the real energy axis one has to sample a high number of @xmath130-points when performing the bz - integration in eq .
( [ eq : bzint ] ) . to reduce this problem we smoothened the energy integrals in eqs .
( [ eq : charge ] ) and ( [ eq : bande ] ) by the fermi - function at a finite temperature @xmath20 , taking into account the poles of it below the contour @xmath175 .
both from an analysis of the integrand and by checking it numerically , it turns out that the contour can be deformed to infinity in the upper complex semi - plane .
moreover , only a finite number of matsubara poles , @xmath194 @xmath195 , has to be considered . assuming a quadratic @xmath20 dependence of @xmath196 due to the sommerfeld expansion , in order to perform extrapolation to @xmath20=0 it was necessary to take two different @xmath20 values only . in our calculations
we choose 300k and 150k for these two temperatures by using 32 and 40 matsubara poles with 1275 and 2926 @xmath130-points in the 2d ibz for @xmath197 ( 3 and 1.5 mryd ) , respectively .
these values were shown to be sufficiently high to yield converged bulk anisotropy energies which we calculated to check the reliability of our numerical evaluations with respect to other methods and also to compare to the value that can be deduced from the bloch wall energies ( see eq .
( [ eq : k4 ] ) . for t=300k and 150k we got @xmath191=0.142 and 0.154 @xmath192ryd , respectively , which were extrapolated to @xmath191=0.158 @xmath192ryd at t=0 . obviously , this is a very good agreement with the results of other first - principles calculations .
let s now turn to the results for the bloch wall energies , associated with an area of size @xmath198 , where @xmath3 is the lattice constant of our bcc lattice , as a function of thickness @xmath58 , measured in units of @xmath3 ( @xmath199 ) , shown in fig .
[ frozen ] for t=150k and t=300k as diamonds and crosses , respectively .
the first thing to note is that the calculated points are very well fitted by the expression eq .
( [ eq : linomega ] ) as drawn by solid and dashed lines , respectively .
the resulting values for @xmath200 are listed in table i together what fitted for t=0 .
the fitted curves have their minima at 424 and 401 lattice parameters for t=300k and t=150k , respectively .
note that there is a point on the @xmath201 curve which is beyond the minimum and hence we can be said to have crossed the scale gap .
the minima are very shallow due to the smallness of the anisotropy energy , thus , this is likely to be a feature of calculations for all cubic systems . from the extrapolated values of @xmath202 and @xmath203 to @xmath47 , @xmath191=0.16 @xmath192ryd and @xmath204=262 mev@xmath205 can be derived via eqs .
( [ eq : swstiffness ] ) and ( [ eq : k4 ] ) ) . by using the expression of @xmath206 from the phenomenological theory we also obtained a thickness of 394 lattice parameters for the bloch wall
several experimental values can be found for @xmath204 , e.g. @xmath207 mev@xmath205 by stringfellow @xcite , @xmath208 mev@xmath205 by collins et al .
@xcite or @xmath209 mev@xmath205 by mook and nicklow @xcite .
so our value is only slightly below the experimental findings .
the same is true for the comparison with other theoretical results found from bulk calculations @xcite .
one source of discrepancies between our and other calculations is certainly the use of different lattice constants .
it is , however , worth noting that by using the same computer code and a spin - flip technique , in the scalar - relativistic limit , a value of @xmath204=300 mev@xmath205 was calculated @xcite , in better agreement with those calculated by others .
this suggest that another source of the deviation of our present value from those of others arises from the relativistic approach we used , that is , the spin - orbit coupling gives not only rise to the magneto - crystalline anisotropy , but to some extent influences the spin - spin interaction parametrized by @xmath204 .
the value of the anisotropy constant is of greater concern .
as it is generally the case in lda calculations @xcite , the experimental value , 0.3 @xmath192ryd , is roughly a factor of two bigger than the theoretical results .
however , the important point here is that we find the same value for @xmath191 as given above when we determine @xmath191 from bulk calculations demonstrating the internal consistency of our calculations . as implied by eq .
( [ eq : omegan ] ) our calculation allows to resolve the bloch wall formation energy into contributions related to layer .
this than can be compared to the free - energy density of the ginzburg - landau theory @xcite . taking the mean value for each layer and in terms of the parameters introduced above , for a linear 90@xmath38 bloch wall this free - energy density reads @xmath210 in fig .
[ layer800 ] the solid line and crosses in the inset display the layer - resolved contributions to the bloch wall energy as calculated by our first - principles method for a 800 layer thick wall and @xmath20=150k , while the dashed curve and the diamonds in the inset are evaluated by using eq .
( [ eq : endens ] ) with the the corresponding fitted parameters , @xmath202 and @xmath203 ( see table i. ) , that is no further fitting have been used .
apparently , the coincidence of the two curves is nearly perfect even in this atomic scale resolution of the bloch wall energy . as is clear from eq .
( [ eq : endens ] ) , in the phenomenological theory the exchange contribution to the layer resolved energy is constant , which equals @xmath211 , inside the wall and zero outside . a characteristic deviation from that behavior is found for the first - principles values near the edge of the wall . the exchange contribution to the first layer outside the bloch wall , i.e. , to layers numbered by 0 and 801 in fig .
[ layer800 ] , is exactly half of the above constant value ( see inset of fig . [ layer800 ] ) .
this is due to the non - local nature of the exchange couplings not accounted properly within the phenomenological theory .
we note that bloch wall formation can be discussed also by using an effective heisenberg hamiltonian with a cubic anisotropy term , in which the above feature is readily recovered .
small oscillations can also be seen near the edge of the bloch wall .
these are most likely related to friedel - type oscillations , which in homogeneous systems arise due to any imperfections , as relaxations in the electronic structure . in order to access reliably the electronic structure in the presence of a bloch wall selfconsistent calculations
have to be performed .
these calculations are not ground state calculations because the magnetic moment in every layer is forced to point along a prescribed direction .
the proper way to perform such calculations is to introduce a constraining field which forces the moment to point along the chosen direction @xcite . for our first attempts we ignored the constraining field and , instead ,
took the projection of the magnetic moment onto the prescribed direction after every iteration .
we have performed this calculation for a 60-layer bloch wall with a magnetization profile according to eq .
( [ eq : linbw ] ) . we allowed 21 layers on the two sides outside the bloch wall to relax .
that is , given the symmetry of the bloch wall , 51 _ different _ potentials were involved in this calculation .
the charging due to the bloch wall per atom for the different layers works out to be smaller than @xmath212 and , therefore , can be considered to be zero within the accuracy of our calculations .
this means we observed no charge redistribution ( transfer ) across the bloch wall . despite of this fact , as inferred from fig .
[ dos ] , the densities of states show characteristic changes for different layers in the bloch wall .
interestingly , the biggest changes are found at energies with also big peaks in the bulk dos , which presumably indicates lifting of some degeneracies , related to the cubic symmetry , due to the presence of bloch wall .
contrary to the charges , the magnetic moments display a small but clear deviations from their bulk value . in fig .
[ mag ] we show the deviation of the moment from the bulk value .
obviously , the moments decrease gradually when approaching the center of the wall .
surprisingly , however , the moments outside the wall still quite differ from their bulk value , approaching it relatively far from the wall only .
similarly , at the center of the wall one would expect a moment equal to that calculated for a bulk with ( 110 ) hard axis .
this value is shown by dashed line in fig .
[ mag ] .
the component of the moment parallel to the wall but perpendicular to the exchange field is shown in fig . [ magy ] .
( one should note that the frame of reference is a local one which turns round with the exchange field . )
not surprisingly , the strong peaks appear at those layers which are just outside the bloch wall , as these are the layers with the most asymmetric neighborhood . clearly , these peaks are an artifact due to the prescribed magnetization profile which display pronounced kinks at the two borders of the bloch wall ( see fig .
2 ) . the component perpendicular to the wall is smaller than @xmath213 and therefore negligible .
these calculations were performed with 91 @xmath130-points in the two dimensional irreducible brillouin zone . to achieve convergence roughly 100 iterations
have been necessary ( where one iteration took roughly 14 min on 16 nodes on a t3e ) .
clearly , the number of selfconsistent iterations necessary to obtain reasonable convergence goes up with the number of layers and we have not been able to achieve convergence for thicker bloch walls within the limited cpu time available to us .
due to the relatively small number of @xmath130-points used in this calculations we do not claim convergence for our results .
but we expect the results to be qualitatively correct .
in short , we have presented a first - principles , that is to say parameter - free and yet materials specific , description of bloch walls in ferromagnets . as an example we have prescribed the orientation of the magnetization density , @xmath45 , to evolve from the easy axis ( 100 ) to ( 010 ) in bcc iron and calculated its energy , @xmath214 , as a function of the width , @xmath206 , of the transition region . to do this we used fully relativistic spin - polarized density functional theory ( dft ) and solved the kohn - sham - dirac equation by the spr - skkr method@xcite . due to the relativistic description of the electrons ,
their spin and orbital degrees of freedom were treated on equal footing and , hence , the calculation gave a full account of the magneto - crystalline anisotropy .
consequently , the bloch wall energy had a minimum .
the equilibrium width as well as the full curve , @xmath215 displayed in fig .
3 . was found to be in good agreement with available experimental data .
the novel feature of this kind of electronic theory of a bloch wall , as compared to its conventional phenomenological description @xcite , is that it provides an account of the distortions of the electronic structure due to the presence of the bloch wall as well as its shape , width and energy .
in particular , we have calculated the variations of the size of the magnetic moment , related to the local exchange splitting , and the local densities of states from layer to layer across the transition region .
these results are displayed in figs .
5 . and 6 .
evidently , these changes are small as expected on the grounds that bcc fe is a good moment system .
nevertheless , even in this case they contain the essential information to describe the scattering of electrons by bloch walls as one needs to do in a study of magneto - transport @xcite .
finally , we note that our calculations scale linearly with the number of layers , @xmath86 , within the bloch wall . due to this fact we were able to perform calculations for up to 800 layers .
that is to say , for the first time , we were able to describe a mesoscopic magnetic defect in fully first - principles terms .
this work resulted from a collaboration partially funded by the tmr network ( contract no .
emrx - ct96 - 0089 ) and the hungarian national science foundation ( contract no .
otka t030240 and t029813 ) .
one of us ( j.s . ) would like to thank to computational collaborative project 9 of uk for financial support .
most of the computations were performed on the 3te machine at the manchester computer center of the epsrc .
.parameters derived from a least square fit of the data in fig . 3 for t=300k and 150k to the function eq .
( ) . in the last row the corresponding values from a quadratic interpolation to t=0 are found . [ cols="^,>,^,^,^,^,^,^,^",options="header " , ] | we present a first principles account of bloch walls on the basis of relativistic spin - density functional theory in the local density approximation ( lda ) .
we argue that calculations based on this description will provide useful and novel informations about the physics of this important element of micro - magnetics . to illustrate the points we wish to make we implement the proposed calculations for fe and determine the bloch wall thickness @xmath0 .
moreover we study features of the electronic structure which arise due to the presence of the bloch wall .
= 10000 |
the gamma - ray binary system psr b1259 - 63/ls 2883 consists of a 48 ms pulsar orbiting around a be star .
the orbit is eccentric ( @xmath0 ) with an orbital period of @xmath1 yr @xcite .
the system is particularly interesting because it is the only gamma - ray binary system where the nature of the compact object is known .
unpulsed tev gamma - ray emission has previously been detected around periastron by the high energy stereoscopic system ( h.e.s.s . )
previous observations have also detected unpulsed radio emission near periastron ( e.g * ? ? ? * ) and x - ray emission across the whole orbit ( see e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the discovery of extended radio and x - ray structures have been recently reported @xcite .
these observations have shown that the unpulsed emission is variable at all wavelengths with respect to the orbital period .
the system was observed around the 2010 periastron passage ( @xmath2 14 december 2010 ) with _
fermi_/lat in the high energy ( he , @xmath3 gev ) regime and showed a flux modulation during the periastron passage , peaking @xmath4 days after periastron @xcite .
the non - thermal radiation from the system is powered by the spin - down luminosity of the pulsar and the unpulsed radiation originates from the stand - off shock front which forms between the pulsar and stellar wind .
extensive modelling of this process has been presented in @xcite , where the x - rays are produced via synchrotron radiation of ultra - relativistic electrons ( lorentz factor @xmath5 ) and gamma - rays through the inverse compton ( ic ) scattering of target photons from the be star .
the ratio between the x - ray and gamma - ray flux is dependent on the radiative and adiabatic cooling times of the post - shocked pulsar wind .
ic cooling was predicted to increase near periastron .
ic cooling of the pre - shocked pulsar wind was also suggested for psr b1259 - 63 by @xcite .
the optical companion in the system , ls 2883 , is a be star , and produces an infrared ( ir ) flux which is higher than predicted by kurucz stellar atmosphere models @xcite .
this ir excess is produced by free - free and free - bound emission in the circumstellar disc which surrounds the be star .
these be star circumstellar discs are decretion discs which form around be stars and are regions of enhanced stellar outflow .
optical interferometry observations of these extended circumstellar envelopes have shown that they are symmetrical around the rotation axis ( see e.g. * ? ? ?
the decretion discs are variable and grow and shrink over periods of hundreds to thousands of days . see e.g. @xcite for a review of classical be stars .
the ir excess from the circumstellar disc provides an additional photon source to the blackbody flux from the optical star .
@xcite showed that in the isotropic case this ir excess could increase the gamma - ray flux , particularly at gev energies , by a factor @xmath6 .
it is possible that near the disc crossing the ir excess may have a much greater effect . in this study
we discuss the effects of the ir excess near periastron , adopting an anisotropic inverse compton model ( e.g. * ? ? ?
* ; * ? ? ?
the soft photon excess provides target photons for ic scattering in the thomson limit where the scattering cross section is a maximum .
the additional effects of the different electron cooling processes in the plerion shock front ( e.g. * ? ? ?
* ) , as well as the effects of gamma - ray absorption @xcite , will not be discussed in this paper and the reader is directed to the given reference for relevant discussions . in order to include the ir excess
, we determine the target photon distribution using a modified version of the curve of growth method which was first proposed by @xcite ( referred to hereafter as the cog method ) .
this modified cog model takes into account the changing optical depth as seen by the pulsar during its orbit .
we constrain the cog model using literature data and vlt observations which were taken during january 2011 , approximately 22 days after the december 2010 periastron passage .
this paper is structured as follows : an outline of previous observations and modelling of psr b1259 - 63 is presented in section [ psrb ] ; section [ sec : observations ] discusses observations we obtained with the vlt , salt , saao 1.9 m and irsf telescopes after the 2010 periastron passage ; section [ be - stars ] discusses the modified cog method that we have derived ; the anisotropic ic scattering is discussed in section [ ic ] ; section [ sec : modelling ] outlines our modelling of the anisotropic inverse compton scattering around periastron ; sections [ sec : discussion ] and [ sec : conclusion ] discuss the results and implications thereof .
a few optical and ir photometric observations of ls 2883 are reported in literature and available in bright star catalogues .
a comparison of the catalogues observations show little change in the optical magnitude . @xcite
confined the spectral range of ls 2883 to a o9 - b2 type star with a colour excess of @xmath7 and assumed a spectral type b2e , with @xmath8 m@xmath9 and @xmath10 r@xmath11 .
analysis of spectroscopic observations also showed that h@xmath12 emission occurred at a radius of @xmath13 , assuming a keplerian circumstellar disc , which implies that the circumstellar disc has a radius @xmath14 @xcite .
more recently , @xcite undertook spectroscopic observations of ls 2883 and proposed new system parameters .
these authors estimate an o9.5 v star with a mass of 31 m@xmath11 ( for a non - rotating star ) , a radius of 9.2 r@xmath11 and an effective temperature of 33500 k. the authors also suggest a new binary inclination of @xmath15 and a lower colour excess of @xmath16 . the size of the be star s circumstellar disc can be constrained by the observed pulsar eclipse around periastron .
observations of the system during consecutive periastron passages show that the pulsed radio signal is eclipsed between roughly 20 days before periastron until 20 days after .
this is believed to be the result of the pulsar passing behind the be star s circumstellar disc .
the binary separation at @xmath17 days is @xmath18 ( assuming @xmath19 and @xmath20 ) which implies that the disc radius is @xmath21 .
for this reason we have adopted a disc radius of @xmath22 in this study .
it should be noted , however , that smooth particle hydrodynamic ( sph ) simulations by @xcite and @xcite suggest that the circumstellar disc might be `` blown - away '' by the pulsar wind if the base density of the disc is low .
however , for larger densities , the disc remains more stable and these authors suggest that the density near the base of the disc is @xmath23 g @xmath24 .
this study only considers a constant disc size .
psr b1259 - 63 was detected by _
fermi_/lat around the 2010 periastron with unexpected results .
while the source remained faint before and just after periastron , approximately @xmath25 days after periastron , the source became very active and much brighter @xcite .
this was unexpected as it was assumed that the system would show fairly symmetric emission around periastron .
radio and x - ray observations around the same period showed no unusual or flaring events and were consistent with previous periastron passages @xcite .
significantly , the peak luminosity detected by _
fermi_/lat was @xmath26 erg s@xmath27 , which corresponds to approximately 100 per cent of the spin - down power of the pulsar .
no pulsed gamma - ray emission was detected with _
fermi_/lat from psr b1259 - 63 during this period .
this is unusual as psr b1259 - 63 fits most of the criteria normally associated with gamma - ray pulsars @xcite .
the lack of a pulsed gamma - ray signal from a known radio pulsar poses interesting questions regarding the location of gamma - ray pulse production .
see e.g. @xcite for a discussion of the gamma - ray pulsars detected by _ fermi_/lat .
possible explanations for the _ fermi_/lat flare have been suggested in terms of ic scattering of the cold pulsar wind @xcite , or doppler shifted synchrotron emission from the tail of the pulsar bow shock @xcite .
we obtained mid - ir observations of psr b1259 - 63/ls 2883 with the vlt telescope during january 2011 .
a number of observations were also undertaken at the saao sutherland observatory : we obtained near - ir observations with the irsf telescope during december 2010 , and spectroscopic observations with the saao 1.9 m telescope and with salt in april 2011 and may 2012 respectively .
the observations are summarised below .
we obtained ( in service mode ) mid - ir observations of psr b1259 - 63/ls 2883 with the vlt imager and spectrometer of the mid infrared ( visir ; * ? ? ?
* ) with the vlt on 5 january 2011 .
this is approximately three weeks after the 14 december 2010 periastron passage .
the observations consisted of one pointing in multiple bands in the broad n - band region ( 8 13 @xmath28 m ) .
the observations were executed with a perpendicular chop / nod direction , a chopping amplitude of 10 and a random jitter of 3 .
the total integration time was of the order of 60 seconds per filter and achieved @xmath29 ratios of @xmath30 .
the observations were calibrated with the standard star hd92682 , which was observed as part of the standard eso procedure , allowing for an average calibration accuracy within @xmath31 per cent .
@xcite .
the calibration star data were reduced with the ` visir_img_phot ` pipeline and the science target observations were reduced using the standard ` visir_img_combine ` pipeline to correct for the image nodding and chopping .
both reduction pipelines produce a single fits @xcite file with four measurements of the targets contained within each file .
we calculated a flux conversion ratio for each filter by comparing the measured flux from the standard star to the defined standard flux .
the flux conversion ratio was used to calibrate the measured flux from the science target .
the final error in the vlt observation of ls 2883 was calculated by combining the estimated photometric error of each flux measurement and the standard deviation of the four measurements in each fits file , for both the science target and standard star .
not included in the error calculation is the @xmath32 error associated with the calibration process and any error in the standard star s catalogue flux value .
the calibrated ( but not de - reddened ) flux at each filter is summarized in table [ tab : vlt ] . .vlt
/ visir data for ls 2883 observed on 5 january 2011 .
flux measurements were calibrated using the standard star hd92682 , but are not corrected of interstellar reddening . [ cols="<,^,^ " , ] we considered three electron distributions : a post - shock power law distribution with @xmath33 between @xmath34 ; a pre - shock mono - energetic distribution centred at @xmath35 ; and a pre - shock mono - energetic distribution centred at @xmath36 .
the ic scattering results from these distributions are discussed below . the predicted aic scattering rate , ( @xmath37 ) , from the post - shocked , adiabatically cooled electron spectrum at periastron is shown in fig .
[ fig : scattering_bw ] .
the solid black line shown in the figure represents the scattering rate due to photons originating from the ( unobscured ) star .
the dashed lines show the contribution due to the circumstellar disc photons , for the different orientations considered .
also shown are the combined contributions ( star+disc ) for the different orientations . the scattering rate has been scaled ( normalized ) to the peak ( unobscured ) stellar contribution , to illustrate better the ratio between the star and disc contributions .
the disc contribution is highest at lower energies , since the scattering occurs in the thomson limit , while the stellar contribution peaks at a @xmath38 ev , due to the scattering occurring mainly in the klein - nishina limit . as is shown in fig .
[ fig : scattering_bw ] , the 45 and 90 disc contributions exceed the stellar contribution below @xmath39 and @xmath40 gev respectively .
this effectively means that the ic spectrum at gev energies is dominated by scattering from the circumstellar disc for these two orientations .
the modelled spectrum shows that the total scattering rate ( star+disc contribution ) is significantly greater than the stellar contribution alone .
the predicted integrated fluxes around periastron , using the parameters above , are shown in figs .
[ fig : aic_light_curve_fermi_new ] & [ fig : aic_light_curve_hess_new ] for the _ fermi_/lat and h.e.s.s .
energy ranges respectively .
the modelled light curves show the integrated flux around periastron , which has been scaled to the maximum ( unobscured ) stellar contribution . for the _ fermi_/lat light curve
the integration energy range is @xmath41 gev , while for the h.e.s.s .
light curves the energy range is @xmath42 tev .
the h.e.s.s .
range was chosen to correspond to the results presented in @xcite .
the lower limit corresponds to their fig . 5 , which presents the integrated flux above 380 gev , while the upper limit , 2.321 tev , is the
highest reported energy in their fig .
the stellar contribution in fig .
[ fig : scattering_bw ] is calculated assuming the full solid angle of the star is observable by the pulsar / electron cloud ( an unobstructed star , applicable to no circumstellar disc ) . however ,
if the disc is thick enough to obscure part of the star , the photon contribution from the star will decrease , reducing the stellar contribution to the ic scattering .
also shown in figs .
[ fig : aic_light_curve_fermi_new ] & [ fig : aic_light_curve_hess_new ] are the modelled gamma - ray light curves if the disc obscures the star , which changes the shape and decreases the flux contribution from the star around periastron .
the gaps in the modified stellar light curves correspond to periods when the pulsar is passing through the disc .
the light curves show that the disc contribution is a minimum around the disc crossing epoch ( @xmath43 days ) and a maximum around periastron .
gev ) , normalized to the maximum stellar contribution .
for each disc orientation , i.e. 90 ( top ) , 45 ( middle ) and 10 ( bottom ) , the disc contribution is compared with the unobscured and the obscured stellar contribution . ] , but for the h.e.s.s . energy range . ] the modelled normalized aic scattering spectrum for pre - shock electrons with energies @xmath35 and @xmath36 at periastron are shown in figs .
[ fig : scattering_four ] and [ fig : scattering_six ] respectively . only a 90 disc has been considered for the pre - shock electrons .
the two electron energies which are considered , were suggested by @xcite ( @xmath35 ) and @xcite ( @xmath36 ) . the @xmath35 spectrum ( fig .
[ fig : scattering_four ] ) shows no indication of a peak in the scattering rate , usually associated with scattering in the klein - nishina regime .
the highest scattering rate occurs at low energies for both the stellar and disc contributions , indicating that the scattering occurs in the thomson limit . the predicted scattering spectrum for the @xmath36 electrons ( fig .
[ fig : scattering_six ] ) shows that the stellar photons scatter in the klein - nishina regime ( implied by the peak in the scattering rate at @xmath44 ev ) , while the disc photons scatter in the thomson regime .
the scattering from the disc is higher than the stellar contribution for energies below @xmath45 ev and the inclusion of the disc contribution increases the total scattering rate by a factor of @xmath46 . .
the figure is normalized to the relevant maximum stellar contribution . ] .
the figure is normalized to the relevant maximum stellar contribution . ]
the predicted integrated flux at _ fermi_/lat energies around periastron are shown for @xmath35 and @xmath36 pre - shock electrons in fig . [ fig : fermi_multiplot_preshock ] , scaled to the maximum stellar contribution .
only the 90 disc has been considered , using the same orientation as for the post - shock electrons .
while the @xmath35 electrons do not significantly alter the light curve , the @xmath36 electrons produce a disc contribution which is significant across the period under consideration , resulting in a higher flux than the stellar contribution at periastron .
the disc minimum occurs around the disc crossing epoch ( @xmath47 days ) while the maximum occurs around periastron .
( top ) and @xmath36 ( bottom ) .
the figures are normalized to the relevant maximum stellar contribution . ]
the predicted increase in the flux at periastron due to the contribution of photons from the disc is shown in fig .
[ fig : ratio_peri_neg ] for different electron distributions ( pre- and post - shock ) .
the predicted increase in the flux is of the same order as the isotropic approximation @xcite and is found to be greater than a factor @xmath6 below gev energies . and @xmath36 ) .
] since the anisotropic calculation is highly directional , photons that arrive over the wide solid angle of the disc do not all scatter with the same energy . unlike the isotropic approximation where
an average scattering angle was considered and the photons were considered to be isotropically distributed , the anisotropic model is based on a photon distribution which is spread out over the solid angle of the disc .
this results in a slight decrease in the influence of the circumstellar disc in the anisotropic calculation .
the increase in the aic spectrum is of the same order as the isotropic approximation because the ir excess predicted in this paper is higher than considered in @xcite .
this is due to the change in the colour excess .
a colour excess of e(b - v)@xmath48 was previously proposed by @xcite , however @xcite have proposed a newer value of e(b - v)@xmath49 . since the reddening effects are lower , less correction is applied to the observed flux . since interstellar
reddening influences the optical region far more than the ir region of the spectrum , this results in a lower optical flux relative to the ir flux .
this implies that the ir excess is higher than previously considered .
the disc is therefore brighter because of a higher disc temperature and is brighter relative to the star .
this relative increase in the photon spectrum increases the influence of the disc photon contribution . the predicted aic , gamma - ray light curves ( figs . [
fig : aic_light_curve_fermi_new ] , [ fig : aic_light_curve_hess_new ] & [ fig : fermi_multiplot_preshock ] ) show that the disc contribution exhibits three properties : firstly , the contribution is a maximum around periastron ; secondly , there is a slight increase just before the disc crossing epoch and a minimum at the disc crossing ; and lastly , the rise after the disc crossing is much faster than the slow rise just before the disc crossing .
this modelled disc contribution light curve is a combination of the photon distribution of the circumstellar disc and the ic scattering angle , as discussed below . as shown by the profile of the normalized photon number density originating from the disc ( fig .
[ fig : disc_ir_profile ] ) , higher frequency photons occur mainly near the centre of the circumstellar disc while longer wavelengths remain constant further out in the disc .
however , the photon number density is much higher at optical frequencies , peaking near @xmath50 hz ( fig .
[ fig : cog_ls2883_fixed ] ) .
these higher frequency photons mainly originate near the centre of the disc , close to the surface of the star . at the position where the pulsar passes through the disc , @xmath51 from the centre of the disc ,
the photon density is much lower .
while the target photons from the disc originate from a wide solid angle near the disc crossing epoch , the bright central region only extends over a small solid angle at large polar angles .
near periastron the pulsar is closer to the star and the bright region of the disc , providing a larger solid angle of brighter photons for ic scattering .
this results in a much larger disc contribution towards the aic scattering at periastron and not at the disc crossing epoch . and @xmath52 hz , from the circumstellar disc at @xmath53 days before periastron plotted through the centre of the disc .
note the y - axis is scaled logarithmically . ]
the scattering from the disc increases rapidly after the disc crossing epoch due to more favourable scattering angles . due to the system orientation ,
before the disc crossing , all scatterings occur predominantly head - to - tail ( @xmath54 ) while after the disc crossing predominantly head - to - head ( @xmath55 ) collisions occur . on the other hand , the aic scattering of the stellar photons follows a smooth transition which results in a smooth light curve and not the sharp increase and decrease seen for the disc contribution .
the combined effects of the photon density and the ic scattering angle result in the predicted aic light curves .
additional effects , such as the heating of the circumstellar disc due to the pulsar crossing or changes in the disc size near periastron , have not been considered .
it may be possible that shock heating of the circumstellar disc could locally increase the disc intensity , providing additional photons for ic scattering , and significantly influence the gamma - ray light curve .
@xcite have suggested that the observed _
fermi_/lat flare is due to the shock heating of the disc and the ic scattering of pre - shocked pulsar wind with a lorentz factor of @xmath56 .
however , the model proposes a large opening angle for the circumstellar disc and requires the pulsar to leave the disc at @xmath57 days near the observed _
fermi_/lat flare .
this is in contradiction to the radio observations which detect the pulsed radio emission @xmath58 days after periastron @xcite . by considering only the ir excess from the circumstellar disc and not considering any of the other consequences of changes in the density ,
the size of the pulsar wind nebula with orbital motion or radiative cooling effects , the total gamma - ray flux increases at gev energies by a factor @xmath6 . while it was thought that the extended size of the disc may completely dominate at the disc cross - epoch , it is shown by the modified cog method that the ir intensity at the distance of the disc crossing is faint and the majority of the emission occurs close to the central star . fig .
[ fig : fermi_hess_model ] shows the calculated gamma - ray energy spectrum at periastron in comparison to the h.e.s.s .
and _ fermi_/lat observations .
the model curve has been been linearly scaled to the observations .
the predicted gamma - ray flux lies orders of magnitudes below the _
fermi_/lat detection limit and is therefore undetectable with current gamma - ray telescopes . at very high energies ( vhe ) , while the flux is high enough for detection , the predicted variability which would result from a change in the disc ir contribution is smaller than the error in the h.e.s.s .
measurements and therefore would not be detected at a significant level .
however , be stars are also known to be variable at optical frequencies . if the optical magnitude of the system changes significantly , this could result in a higher variability at tev energies .
sph simulations @xcite of psr b1259 - 63/ls 2883 around periastron have suggested that the disc could be disrupted by the passage of the pulsar during the periastron passage . as shown by these authors , the degree
to which the disc is disrupted depends on the density of the disc . for a larger density ( @xmath23 g @xmath24 )
the disc remains more stable .
this could result in changes in the observed optical and ir emission from the disc .
however , observations near periastron on 5 january 2011 with visir / vlt found a flux that was of the same order as earlier band a measurements with the _ midcourse space experiment _ ( msx ) ( table [ tab : ls2883_all ] ) . similarly , spectroscopic observations of ls 2883 with the saao 1.9 m telescope , taken four months after periastron show a very prominent h@xmath59 emission line and no indication of stellar absorption , indicating that the circumstellar disc completely dominates the stellar absorption at the same frequency .
these observations , of the ir excess and h@xmath60 emission line , which took place after periastron , show that the be star is still surrounded by a substantial circumstellar disc .
the observed equivalent width for the h@xmath60 line is ew @xmath61 , very similar to the measurement reported by @xcite ( ew @xmath62
) . this may be an indication that the disc remains prominent despite the pulsar passage or , alternatively , may indicate a very rapid recovery .
similarly the observations obtained with salt during may 2012 , far away from periastron , still show a prominent h@xmath60 line .
the flare detected with _
fermi_/lat after the last periastron passage was unexpected .
we have suggested that the ir excess associated with the circumstellar disc can have a large influence on the gamma - ray production , increasing the flux by a factor @xmath46 .
however this increase is too low to explain the _
fermi_/lat flare . as is shown by the cog model
, the majority of the ir emission is produced close to the stellar surface and not near the location of the pulsar - disc crossing .
this results in the majority of the disc contribution occurring close to periastron and not near the disc crossing phase .
it has been suggested by @xcite that the pre - shock scattering of @xmath35 electrons could cause the observed _
fermi_/lat flare if the photon density from this disc is large enough .
these authors have suggested that a luminosity of 40 per cent of the stellar luminosity is required , and that the additional photons must be originate over a large region .
however , the observations undertaken during and after the previous periastron passage do not seem to suggest such a large heating effect . the ir emission from the circumstellar disc remains fairly consistent with the _
2mass _ observations , while the h@xmath60 observations four months after periastron is consistent with earlier observations reported by @xcite .
these observations then suggest that the shock heating which should occur during the pulsar - disc crossing will be more localised around the pulsar .
@xcite discussed thermal x - ray emission in gamma - ray binaries as a result of the shock heating of the stellar wind .
these authors estimated the total luminosity of the thermal x - rays as some fraction of the total kinetic luminosity of the stellar wind .
we can make a similar approximation of the total amount of energy available for optical / ir emission from shock heating by estimating the kinetic luminosity of the stellar wind from ls 2883 .
the total kinetic luminosity can be approximated by @xmath63 , where @xmath64 is the mass loss rate , and @xmath65 is the speed of the wind .
similarly to @xcite we can estimate the kinetic luminosity for the polar wind to be , @xmath66 and for the equatorial wind to be @xmath67 the kinetic luminosity of the circumstellar disc can also be estimated by considering the disc to be keplerian , with a velocity @xmath68 .
if the density through the circumstellar disc is given by equation [ eqn : rho_distance ] , the mass flow rate of the disc through an area d@xmath69 is given by @xmath70 , where the density and velocity depend on the distance @xmath71 from the centre of the star .
integrating over the area of the disc perpendicular to the direction of the disc motion ( d@xmath72 ) gives the total kinetic luminosity as @xmath73 here @xmath74 has been calculated from equation [ eqn : rho_distance ] and the fitted cog parameters have been used . since
only a fraction of the available kinetic luminosity can be directed into shock heating , it seems likely that any increase in the ir / optical emission will have a luminosity @xmath75 erg s@xmath27 .
it has also been shown by @xcite that doppler boosting of the synchrotron emission could produce the required energy level to explain the _
fermi_/lat flare .
the resulting flare occurs because of the preferential direction of the shock outflow after the second disc crossing which results in bulk doppler boosting of material in the cone of the pulsar wind shock front .
this suggests that the _ fermi_/lat flare is the result of the combined effects of shock heating and doppler shifting .
the sph simulations undertaken by @xcite show that circumstellar disc can be deformed around the pulsar , creating a cavity - like structure around the pulsar .
such a shock heated cavity may then allow for additional photons , which will allow for head - to - head gamma - ray production after the second disc crossing . dedicated optical and ir campaigns around future disc crossing phases are required to investigate the degree to which the circumstellar disc is influenced by the pulsar and pulsar wind .
the binary star system psr b1259 - 63/ls 2883 has remained a fascinating source since the detection of the pulsar approximately two decades ago .
the interaction between the be star and the pulsar produces multi - wavelength emission from radio up to tev gamma - rays , covering 19 decades in energy .
the optical companion in the system is a be type star , known to possess a circumstellar disc which produces an ir flux .
this flux is observed as an ir excess which is higher than predicted by blackbody and stellar atmosphere models .
since the gamma - rays in the system are produced by the up - scattering of photons , the presence of an extended and additional photon source could potentially have a significant effect on the gamma - ray production in psr b1259 - 63/ls 2883 .
we modified the cog method , first proposed by @xcite , which allowed us to predict the flux originating from the circumstellar disc at any position near the disc .
this method has been used to predict the photon density as observed from the point - of - view of the pulsar .
the cog method was constrained using optical and ir data obtained from literature and vlt observations .
the photon density was calculated around periastron and we applied it to the problem of aic scattering .
the anisotropic scattering considered the change in the photon distribution during the orbit , the variation in the solid angle subtending the disc , and the direction of scattering .
the variation in the gamma - ray spectrum around periastron was calculated for different disc orientations .
in this study the leptons in the pulsar wind are assumed to be mono - energetic before the stand - off shock and a power - law energy distribution in the post - shock region .
both a pre and post shock distribution have been considered . for psr
b1259 - 63 , it can be shown that adiabatic cooling dominates near the stand - off shock and , for this reason , only adiabatic cooling was considered for the aic modelling .
the ic modelling showed that the inclusion of the ir excess associated with the circumstellar disc increased the gev gamma - rays by a factor @xmath76 .
this relative increase in the flux is partly dependent on the choice of the electron distribution . for a wide distribution
( e.g. @xmath77 ) , as was shown for the isotropic modelling ( fig . 3 in * ? ? ?
* ) , or an electron distribution centred around @xmath78 ( fig .
[ fig : fermi_multiplot_preshock ] ) , the ic scattering is dominated by the scattering of stellar photons in the thomson limit . however , this study has shown that for the electron distribution proposed by @xcite , which has been shown to provide a reasonable fit to the 2004 h.e.s.s .
observations , the ir excess from the circumstellar disc increases the gamma - ray flux by a factor @xmath79 at gev energies . the predicted light curves show ( figs .
[ fig : aic_light_curve_fermi_new ] , [ fig : aic_light_curve_hess_new ] & [ fig : fermi_multiplot_preshock ] ) that the circumstellar disc is most influential close to periastron and , generally , after the disc crossing .
this is a result of the disc being brightest near the stellar surface and the associated favourable scattering angles after the disc crossing .
while it was thought the predicted gamma - ray flux may be highest near the disc crossing epoch ( @xmath80 days ) , the low intensity of the disc at @xmath40 stellar radii from the star s surface does not provide a sufficient number of photons to increase significantly the gamma - ray flux in this region .
this study has only considered a steady state disc and has not considered what effects the pulsar passage may have introduced on the circumstellar disc .
shock heating of the disc or clumping of material may result in regions of higher intensity which could result in more localised flux increases .
the additional effects such as changes in the electron cooling have not been considered , as these have been addressed by other authors .
this study has focussed on the effect of the ir excess from the circumstellar disc in order to determine its influence on the production of gev energy gamma - rays .
the study shows that the gamma - ray flux could be increased by a factor greater than two at gev energies , which is a significant increase .
however , the predicted gamma - ray flux lies far below that detected by _
fermi_/lat at the previous periastron passage . a combination of a number of effects , including e.g. doppler boosting , shock heating and changes in the shock front are required to explain the flare that was observed by _
fermi_/lat approximately one month after periastron .
the authors are very grateful to the late o.c .
de jager for suggesting this study .
numerical modelling was undertaken using the university of the free state s high performance computing division and the authors are particularly grateful for the help of a. van eck .
this research was funded by the south african square kilometre array project .
some of the observations reported in this paper were obtained with the southern african large telescope ( salt ) .
this paper uses observations made at the south african astronomical observatory ( saao ) .
this publication makes use of data products from the two micron all sky survey , which is a joint project of the university of massachusetts and the infrared processing and analysis center / california institute of technology , funded by the national aeronautics and space administration and the national science foundation .
this research made use of data products from the midcourse space experiment .
processing of the data was funded by the ballistic missile defense organization with additional support from nasa office of space science .
this research has also made use of the nasa/ ipac infrared science archive , which is operated by the jet propulsion laboratory , california institute of technology , under contract with the national aeronautics and space administration .
the authors would like to thank the anonymous reviewer for their very helpful comments which have improved this study .
abdo a. a. , ackermann m. , ajello m. et .
, 2010 , apj suppliment , 187 , 460 abdo a. a. , ackermann m. , ajello m. et .
al , 2011 , apj , 736 , l11 aharonian f. , akhperjanian a. g. , aye k .- m .
et al . , 2005 ,
a&a , 442 , 1 aharonian f. , akhperjanian a. g. , anton g. et .
, 2009 , a&a , 507 , 389 ball l. , kirk j. g. , 2000 , aph , 12 , 335 carciofi a. c. 2011 , iaus 272 , 325 chernyakova m. , neronov a. , lutovinov a. , rodriguez j. , johnston s. , 2006 , mnras , 367 , 1201 chernyakova m. , neronov a. , aharonian f. , uchiyama y. , takahashi t. , 2009 , mnras , 397 , 2123 cominsky l. , roberts m. , johnston s. , 1994 , apj .
427 , 978 crawford s.m . , still m. , schellart p. , et al . , 2010 ,
spie proc .
7737 , 54 dubus g. , 2006 , a&a , 451 , 9 dubus g. , cerutti b. , henri g. , 2008 , a&a , 477 , 691 fargion d. , konoplich r. v. , salis a. , 1997 , z. phys .
c 74 , 571 johnston s. , manchester r. n. , lyne a. g. , bailes m. , kaspi v. m. , qiao g. , damico n. , 1992 , apj , 387 , l37 johnston s. , manchester r. n. , lyne a. g. , nicastro l. , spyromilio , j. , 1994 , mnras , 268 , 430 johnston s. , manchester r. n. , mcconnell d. , campbell - wilson d. , 1999 , mnras , 302 , 277 kaspi v. m. , tavani m. , nagase f. , hirayama m. , hoshino m. , aoki t. , kawai n. , arons , j. , 1995 , apj , 453 , 424 khangulyan d. , aharonian f. a. , bogovalov s. v. , ribo m. , 2012 , apj , 752 , l17 kirk j. g. , ball l. , skjaeraasen o. , 1999 , astropart .
phys . , 10 , 31 kong s. w. , yu y. w. , huang y. f. , cheng k. s. , 2011 , mnras , 416 , 1067 kong s. w. , cheng k. s. , huang y. f. , 2012 , apj , 753 , 127 kurucz r. l. , 1979 , apj suppliment , 40 , 1 lagage p. o. , pel j. w. , authier , m. et al . , 2004 , the messenger , 117 , 12 moldn j. , johnston s. , rib m. , paredes j. m. , deller a. t. , 2011 , apj , 732 , l10 moskalenko i. v. , strong a. w. , 2000 , apj , 528 , 357 negueruela i. , rib m. , herrero a. , lorenzo j. , khangulyan d. , aharonian f. a. , 2011 , apj , 732 , l11 okazaki a. t. , nagataki s. , naito t. , kawachi a. , hayasaki k. , owocki s. p. , takata j. , 2011 , pasj , 63 , 893 pavlov g. g. , chang c. , kargaltsev o. , 2011 , apj , 730 , 2 pomm s. , johansson l. , sibbens g. , denecke b. , 2003 , nimpa , 505 , 286 porter j. m. , rivinius t. , 2003 , pasp , 115 , 1153 press w. h. , teukolsky s. a. , vetterling w. t. , flannery b. p. , 2007
, numerical recipes : the art of scientific computing , cambridge univ . press , cambridge price s. d. , egan m. p. , carey s. j. , mizuno d. r. , kuchar t. a. , 2001 , aj , 121 , 2819 quirrenbach a. , buscher d. f. , mozurkewich d. , hummel c. a. , armstrong j. t. , 1994 , a&a , 283 , l13 quirrenbach a. , bjorkman k. s. , bjorkman j. e. et al . , 1997 , apj , 479 , 477 skrutskie m. f. , cutri r. m. stiening r. etal . , 2006 ,
aj , 131 , 1163 smette a. , vanzi l. , 2007 , very large telescope paranal science operations visir user manual , 81 takata j. , okazaki a. t. , nagataki s. et al . , 2012 , apj , 750 , 70 tam p. h. t. , huang r. h. h. , takata j. , hui c. y. , kong a. k. h. , cheng k. s. , 2011 , apj , 736 , l10 tavani m. , arons j. , 1997 , apj , 477 , 439 telting j. h. , waters l. b. f. m. , roche p. , boogert a. c. a. , clark j. s. , de martino d. , persi p. , 1998 ,
mnras , 296 , 785 tody d. , 1993 , in hanisch r. j. , brissenden r. j. v. , barenes j. , eds , asp conf .
52 , 173 tryka s. , 1997 opt .
co. 137 , 317 van soelen b. , meintjes p. j. , 2011 , mnras , 412 , 1721 waters l. b. f. m. , 1986 , a&a , 162 , 121 waters l. b. f. m. , lamer h. j. g. l. m. , 1984 , a&as , 57 , 327 wells d. c. , greisen e. w. , harten r. h. , 1981 , a&as , 44 , 363 westerlund b. e. , garnier r. , 1989 , a&as , 78 , 203 wood k. , bjorkman k. s. , bjorkman j. e. , 1997 , apj , 477 , 926 wright a. e. , barlow m. j. , 1975 , mnras , 170 , 41 zabalza v. , bosch - ramon v. , paredes j. m. , 2011 , apj , 743 , 7
the specific intensity which orientates along any line - of - sight in the disc can be calculated by considering an impact vector @xmath81 relative to the disc . in fig .
[ fig : cog_fig_paper ] , the origin of the coordinate system is at @xmath82 ( the centre of the pulsar ) , the unit vector @xmath83 points towards the centre of the disc ( and star ) and @xmath84 in the direction of the incoming photon / line - of - sight .
the value of @xmath85 is then defined as @xmath86 where @xmath87 is the radial distance in units of @xmath88 and @xmath89 is defined by @xmath90 any line - of - sight in a direction defined by @xmath91 , @xmath92 intercepts the disc mid - plane at a position ( figs . [
fig : cog_fig_paper ] & [ fig : disc - geometry ] ) @xmath93 where @xmath94 and @xmath95 are measured relevant to the disc mid - plane such that the origin is in the centre of the star / disc and the @xmath94 and @xmath95 are measured in the plane of the disc .
the height of the disc at any point in this coordinate system is @xmath96 where @xmath97 is the opening angle of the disc . for any line - of - sight ( @xmath84 )
, we can construct a distance @xmath98 which is perpendicular to the disc mid - plane and intercepts the line - of - sight .
if @xmath98 intercepts the disc mid - plane at any distance @xmath99 and @xmath100 from the point of the disc plane intercept ( @xmath101 ) , then the length along @xmath98 between the disc mid - plane and the line - of - sight intercepts is @xmath102 where @xmath103 . to find the point of the disc intercept @xmath99 and @xmath100 can be solved from @xmath104 , and therefore @xmath105 since in the plane of the disc , @xmath106 the equation above can be solved for @xmath99 and hence @xmath100 can be calculated .
the solution gives two values for both @xmath99 and @xmath100 , which lie below and above the disc plane .
special cases occur when @xmath99 or @xmath100 are equal to zero .
the correct solution for the intercept above and below the plane is determined by the direction of the line - of - sight and its orientation to the disc .
the values of the ( @xmath107 ) and ( @xmath108 ) coordinates are then given by @xmath109 and the z coordinates are calculated by @xmath110 using the solution to the disc intercepts , we then calculate the angles @xmath111 and @xmath112 in fig .
[ fig : cog_fig_paper ] .
the two disc intercepts lie at @xmath113 and @xmath114 while the centre of the electron cloud / pulsar lies at @xmath115 .
the angles @xmath111 and @xmath112 are the angles between the disc intercepts @xmath113 & @xmath114 and the centre of the electron cloud , and are calculated from @xmath116 | the gamma - ray binary system psr b1259 - 63/ls 2883 consists of a 48 ms pulsar orbiting a be star .
the system is particularly interesting because it is the only gamma - ray binary system where the nature of the compact object is known .
the non - thermal radiation from the system is powered by the spin - down luminosity of the pulsar and the unpulsed radiation originates from the stand - off shock front which forms between the pulsar and stellar wind .
the be star / optical companion in the system produces an excess infrared flux from the associated circumstellar disc .
this infrared excess provides an additional photon source for inverse compton scattering .
we discuss the effects of the ir excess near periastron , for anisotropic inverse compton scattering and associated gamma - ray production .
we determine the infrared excess from the circumstellar disc using a modified version of a curve of growth method , which takes into account the changing optical depth through the circumstellar disc during the orbit .
the model is constrained using archive data and additional mid - ir observations obtained with the vlt during january 2011 .
the inverse compton scattering rate was calculated for three orientations of the circumstellar disc .
the predicted gamma - ray light curves show that the disc contribution is a maximum around periastron and not around the disc crossing epoch .
this is a result of the disc being brightest near the stellar surface .
additional spectroscopic and near - infrared observations were obtained of the system and these are discussed in relation to the possibility of shock heating during the disc crossing epoch .
radiation mechanisms : non - thermal pulsars : individual : psr b1259 - 63 x - rays : binaries |
the blazhko effect is a modulation of the pulsations found in some variable stars of the rr lyrae type .
the origin of the effect has been the subject of debate ever since it was first reported in 1907 by russian astronomer sergey nikolaevich blazhko @xcite .
see the recent reviews by @xcite , @xcite .
various models have been proposed over the years to explain the effect including the beating - modes model " which was recently presented by the author @xcite .
other recent models can be found and discussed in section 4 of @xcite . in a recent paper , @xcite
hypothesize that the origin of the blazhko effect is a dynamical interaction between a multi - shock structure and an outflowing wind in a coronal structure .
the beating - modes model " proposes that the modulation is not a true modulation at all , but rather a beat - frequency between two vibrational modes with slightly different frequencies , one of which is highly non - sinusoidal .
that model achieves a good fit with kepler data for rr lyr and , more importantly , reproduces key features of the observed light curve including the pulsation waveform , the upper and lower blazhko envelope functions which are offset from each other in phase , and the motion , disappearance , and reappearance of the so called bump " that appears on the pulsation waveform . the basic idea for the model
is not new ( see , e.g. * ? ? ?
* ) , but previous efforts made no attempt to show that the model could achieve a good fit to actual data , and it has been given little , if any , serious consideration .
other evidence in its favor include an ability to explain the mysterious double - maxima waveform found in the star v445 lyr ( section 3.1 in * ? ? ?
* ) and also , as is shown in section [ modulation ] below , the ability to explain a double - blazhko modulation observed in v808 cyg .
this paper examines an important modification of the beating - modes model in which the two modes may interact through nonlinear coupling .
this will be referred to as the interactive - beating - modes model " . in section [ analytic ]
we analyze the interaction in the adiabatic limit .
we consider the case where one of the modes ( presumably the fundamental ) is much more robust than the other and is consequently affected much less by the interaction .
the lowest order nonlinear coupling is considered , and the dynamics are found to generate an opposing side - peak , producing a triplet spectrum with asymmetry dependent on the coupling strength . in the adiabatic limit , this is a pure triplet with no additional side - peaks .
this model can match any possible side - peak amplitude ratio . that ratio is known to vary widely in observational data .
rr lyr , for example , is known to have a very asymmetric triplet spectrum with ratio measured using period04 software to be 0.194 for second quarter _
data with the larger side - peak on the right of the fundamental , while the ratio for v808 cyg was measured to be 0.857 with the larger side - peak on the left .
although the idea that the opposing side - peak is generated by a nonlinear mixing process is implicit in the beating - modes model , it was explicitly discussed in a general way in @xcite . in section [ dissipative ] nonlinear dissipative effects
are considered , and it is found that additional side - peaks of odd order emerge in the spectrum .
consistent with this , data from v808 cyg is shown to favor third order over second order side - peaks of the fundamental .
( the presence of the weaker second order peaks can also be accounted for . )
modes with extremely small detuning from the fundamental become phase - locked to it and effectively disappear from the spectrum . assuming that the phase - locked state is stable ( and this is generally the case ) these modes _ will not _ produce a blazhko effect . on the other hand ,
when the modes are not phase - locked , the relative phase continues to increase by @xmath0 every blazhko period .
the cases that produce the most dramatic blazhko effect are those for which the detuning is just slightly too large to allow phase - locking to occur .
the interaction is found to pull the frequency of the nonradial mode toward the fundamental , substantially reducing blazhko frequency from the value of the detuning in the absence of coupling .
note that this model is quite different from one claiming that the blazhko effect comes from a dynamical oscillation between the fundamental and a mode that is phase - locked to it .
phase - locked " in that case does not mean fixed relative phase , but rather a relative phase that oscillates within a range of values and does not increase without bound .
@xcite found such an oscillatory case for a model involving the ninth overtone ( oscillatory cases 2a and 2b rather than the fixed point case 1c on their page 2 ) .
@xcite tried to find such an oscillatory phase - locked solution between the fundamental and a near - resonant nonradial mode and were unable to do so , suggesting that this is not a viable mechanism for the blazhko effect .
there appears to be good observational evidence of nonradial modes in rr lyrae stars .
@xcite , in a study of multiperiodic rr lyrae variables of the galactic bulge , find 25 stars of the rr0-@xmath11 type which are fundamental mode pulsators that have a near - resonant peak in the spectrum forming a doublet , with no opposing side - peak ( at least not visible above the noise ) .
since these peaks seem too close together to be generated by two different radial modes , this would seem to be good evidence that the smaller of the peaks was generated by a nonradial mode .
the only alternative is to imagine that the single side - peak is generated by modulating the fundamental in a manor that is special in the sense that it suppresses all side - peaks except the one observed .
@xcite state
it is not clear at present if so strongly asymmetric triplets ( or pure doublets ) can be reproduced by the resonant mode coupling theory " .
@xcite find a multiplet structure in variations of the line profile of rr lyr which they suggest may be indicative of the presence of a nonradial mode .
there is also observational evidence for nonradial modes that are not near - resonant with the fundamental or with the first radial overtone . in the spectrum of v445 lyr ,
@xcite suggest that the peak designated @xmath2 is likely nonradial . in the spectrum of v1127 aql
, @xcite suggest that a peak designated @xmath3 could be a nonradial p - mode and a peak designated @xmath4 , found below the fundamental , could be a g - mode .
another likely example is the doubled peak in v808 cyg described below in section [ modulation ] .
@xcite in their figure 1 , indicate that nonradial modes very close in frequency to the fundamental may have a growth rate advantage , possibly explaining the tendency for these modes to be found in very close proximity .
their results suggest the observed modes might be dipolar modes of radial order near 140 .
results of @xcite and @xcite suggest a much lower radial order ( 4 and 12 respectively ) .
all of these calculations are for a static star , without the presence of the high amplitude fundamental mode . because of its extreme nonlinear character
, the fundamental would seem likely to have very dramatic effects on the characteristics of the other modes such as their frequencies , growth rates , spatial waveform , etc .
, which may be very difficult to predict . in section [ modulation ] it is shown that the model provides an explanation for a peak observed in v808 cyg that is apparently modulated at twice the blazhko frequency .
the explanation is based on the effects of the fluctuating environment provided by the blazhko effect and specifically depends on the involvement of the nonradial mode .
the effect therefore would seem to be incompatible with blazhko models that assume the fundamental mode is being modulated by some process , which is true for almost all blazhko models .
there are terms in the amplitude equations that generate these environmental effects , which in some other papers are referred to as saturation " terms .
see e.g. equations 6 and 7 in @xcite in which the r " terms are the resonant terms and the s " terms are the saturation or environmental terms .
environmental terms can provide a limiting effect on the growth of a mode , as implied by the word saturation " , but they can also result in frequency shifts and modulation .
these terms will not result in phase - locking or direct transfer of energy between modes .
for relatively small excitations from equilibrium the stellar dynamics can be described ( in the adiabatic approximation ) in terms of a set of normal modes characterized by a set of eigenfunctions and corresponding set of normal mode displacements @xmath5 ( see e.g. * ? ? ?
* ) . in the treatment presented here ,
these displacements are real , not complex , and the modal interactions are described by a set of real differential equations .
amplitude equations in the astrophysical literature are most commonly found in the complex form ; the conversion of the results presented here to the complex form is straightforward and given in appendix [ complex ] .
both the real and complex forms are discussed in @xcite .
the equations for the adiabatic case are generated using a potential energy approach .
this approach does not seem to be discussed much in papers on stellar dynamics , but can be found in other literature relating to normal modes ( see e.g. * ? ? ?
* section 2 , paragraph 1 ) .
one advantage to this approach is to automatically generate equations that satisfy energy conservation .
an arbitrary displacement from equilibrium will result in an increase in the potential energy @xmath6 of the system which is the combination of the gravitational potential energy and the internal energy .
if this displacement is expanded in terms of the eigenfunctions , then the potential energy can be expressed in terms of the normal mode displacements .
since the potential is a function of the normal mode displacements , it can be expanded as a power series in them .
the lowest order terms in this power series will be quadratic with no cross terms between them , i.e. no terms such as @xmath7 ( the eigenfunctions are chosen such that this is the case ) . as a result , in the limit of low level excitations , the normal modes are completely independent harmonic oscillators . but for any finite level of excitation , there are higher order terms in the potential energy , cubic and above , and these may include cross terms between the normal mode displacements
. this will result both in nonlinearity within the equations of motion for the @xmath8 but will also result in coupling terms between them . the effective force " that appears in the equation of motion of a particular mode
is determined by the negative of the derivative of the potential energy with respect to the displacement variable corresponding to that mode .
each mode will have an effective mass " or inertia @xmath9 associated with it .
however , by independently rescaling the amplitudes of each of the modes , the @xmath9 can all be set to unity .
the kinetic energy of mode @xmath10 can then be expressed as @xmath11 and its equation of motion is given by : @xmath12 where @xmath13 is the nonlinear part of the effective force for mode @xmath10 , and @xmath14 is the potential energy , both as nonlinear functions of all the displacements @xmath5 .
there are terms in the power series expansion of the potential involving only one mode : @xmath15 where @xmath16 is a coefficient .
the lowest nonzero term in this expansion is the quadratic , and it determines the frequency : @xmath17 .
the higher order terms can result in a non - sinusoidal distortion to the oscillation and/or a shift in the frequency .
most terms in the potential involve two or more modes and can result in coupling between them . note that although this approach is most accurate when all nonlinear terms are small , we know that for rr lyrae and some other types of variable stars the pulsation is often highly non - sinusoidal .
it is not clear to what extent it is useful to try and model this using the @xmath15 terms or some other representation of a nonlinear function of @xmath8 .
@xcite studied a simple one zone " nonlinear oscillator model of this type which can generate the sawtooth - like velocity function that is ubiquitous in cepheid and rr lyrae stars . but this kind of approach may be lacking in certain respects due to its simplicity .
in particular , the spacial character of a mode at high amplitude may deviate substantially from its eigenfunction and may be more non - sinusoidal near the surface than internally .
@xcite show how nonlinear normal modes can be treated in a more precise way , while at the same time be parameterized as a single position / velocity pair even though the state for that nonlinear mode is no longer an eigenfunction and can be expanded as a linear combination of the eigenfunctions ; see their equation 12 .
we also know from the stellar spectra that one or more low amplitude modes can coexist with the main oscillation ( see , e.g. , * ? ? ?
* ) and so it is reasonable to assume that the system can still be approximately described in terms of nonlinear interactions between this set of modes .
the cross terms can lead to phase - locking if sufficiently strong , or can cause modulation as a result of the slipping relative phase between the modes and corresponding periodic change in the effects of the cross term on each of the modes .
these effects are the most significant when there is a near - resonance . in this case the corresponding term or terms in the potential will have some component at a very low ( or zero ) frequency . when a mode is non - sinusoidal , such as the fundamental in this case , it will have harmonic components and the presence of these additional frequencies can sometimes cause a term in the potential to become near - resonant when it was not previously . within the adiabatic approximation , energy lost by one mode must be gained by the modes with which it is interacting through the nonlinear coupling terms . for the case of interest , there are two modes of nearly the same frequency : the fundamental which is assumed to be very robust and highly non - sinusoidal , and a weaker mode that is nearly sinusoidal and is assumed to be a nonradial mode .
we will refer to the fundamental as mode 0 and the nonradial mode as mode 1 , represented respectively by @xmath18 and @xmath19 .
the corresponding angular frequencies are @xmath20 and @xmath21 , and it should be noted that these are the frequencies in the absence of nonlinear effects ; the peaks observed in the stellar spectra , especially that of mode 1 , may be shifted away from these values as will be shown below . for the sake of convenience , the dominant side - peak in the spectrum
may be referred to as @xmath22 even though technically it should be referred to as the observed @xmath22 " .
there is no peak in the spectrum at the actual ( unshifted ) @xmath22 .
the robust character of the fundamental should be reflected in a comparison of their average kinetic energies : @xmath23 .
the character of the nonradial mode as a function of polar and azimuthal angles is derived from a spherical harmonic @xmath24 [ see e.g. @xcite , equation 13.60 or @xcite , equations 1.1 , 1.2 and 1.3 ] .
because @xmath19 is nonradial , the potential energy @xmath14 can only contain cross - terms with powers of @xmath19 greater than one .
so terms of the form @xmath25 are not allowed for any @xmath26 .
this is because a radial mode to any power is still an entirely radial function .
so it has no angular dependence , which in terms of spherical harmonics is the constant @xmath27 .
this is of course orthogonal to any other @xmath24 including the one corresponding to @xmath19 .
the net potential energy for such terms , being determined by an integral over the sphere , will thus necessarily be zero .
on the other hand , putting any spherical harmonic to any even power will produce a function that is everywhere positive or zero .
expanding this resulting function in terms of spherical harmonics will always produce a @xmath27 component , and this can interact with the radial mode .
this is also true for odd powers of 3 or more , for nonradial modes which have @xmath28 and @xmath29 even some modes have symmetry properties that prevent odd powers from producing a @xmath27 component .
this includes all modes with @xmath30 and also the modes with @xmath28 for which the spherical harmonic is antisymmetric about the equator , i.e. those with @xmath29 odd . in the latter case
if we make a change of variables @xmath31 we reproduce the exact same mode with inverted amplitude .
since the radial mode has no @xmath32 dependence , this should mean that terms in the potential involving both @xmath18 and @xmath19 should be invariant under a sign change for @xmath19 , e.g. @xmath33 .
for @xmath34 odd this can only be true if @xmath35 , i.e. , odd powers of @xmath34 are disallowed .
similarly , if we consider a mode with @xmath30 we can make a change of variables @xmath36 we again have an inverted amplitude and the same argument holds . ] .
so the allowed cross terms have the form @xmath37 where @xmath38 with the additional condition that if @xmath34 is odd then the nonradial mode must have @xmath28 and @xmath29 even .
we will mainly be interested in terms ( if any ) which are near - resonant ; the other allowed terms will typically be neglected .
we will initially assume that @xmath28 ; the case of nonzero @xmath39 is discussed briefly in section [ nonzerom ] .
here we will consider only the cases that are quadratic in @xmath19 as these are lowest order and therefore likely to be much stronger than those being neglected .
two terms in the potential energy that might result in a significant interaction are @xmath40 and @xmath41 .
the @xmath42 term is only relevant if we allow the fundamental to be strongly non - sinusoidal , possibly resulting in @xmath18 having a strong second harmonic component .
this possibility is not generally considered in the astrophysical literature . however , as is discussed below , the presence or absence of this term has no qualitative effect on the results presented in this paper , as the two terms have similar effects on the dynamics .
the @xmath43 term does not depend on any harmonic content but is a higher order term .
the corresponding force terms that appear in the equation for @xmath44 are @xmath45 and @xmath46 .
note that the second of these , if converted to the complex form as described in appendix [ complex ] , will generate terms corresponding to both the second and fourth terms on the right hand side of equation 7 in @xcite as well as other terms that can be neglected as nonresonant .
the remaining nonlinear terms in their equation 7 ( terms 3 , 5 and 6 ) can be neglected in our application as they are higher order in @xmath19 .
( terms 5 and 6 are also disallowed by symmetry for @xmath29 odd as mentioned in their paper . )
term 3 in their equation 7 does play an important role in nonlinear damping as we discuss in section [ dissipative ] and appendix [ complex ] .
the fundamental , due to its higher energy content , will be effected only slightly by its coupling to the nonradial mode .
we will initially make the approximation that we can neglect the effect of @xmath19 on @xmath18 entirely and that therefore @xmath18 can be approximated as a fixed periodic function with frequency @xmath20 .
( the more general case with separate equation for @xmath47 and dissipation is discussed briefly in section [ x0x1 ] and is found to reproduce the results given here . )
if this function has a significant second harmonic component , it will produce a near - resonance in the @xmath42 term .
the the square of this function will generate a second harmonic component that will produce a near - resonance in the @xmath43 term .
these terms have the same @xmath19 dependence so they can be combined together and the time origin can be chosen so that the net second harmonic is a pure cosine with some nonnegative amplitude .
contributions to this combined term from terms with higher powers of @xmath18 could also be included if they have any significance .
the detuning parameter is defined by @xmath48 .
we assume that the detuning is small compared to the fundamental frequency : @xmath49 .
both the @xmath42 and @xmath43 terms can generate a component that is a constant times @xmath19 . in this small detuning limit ,
the effect of this term is a simple shift in the detuning to the new value @xmath50 note that @xmath51 will typically be near zero when @xmath18 is approximately sinusoidal .
for convenience we rescale time so that the fundamental frequency @xmath52 and we define the net coupling coefficient @xmath53 as the combined amplitude of second harmonic components divided by @xmath54 .
the resulting equation for the dynamics of @xmath19 is : @xmath55 x_1\ ] ] this equation has near - resonant terms that can lead to a slow evolution of the amplitude of @xmath19 . usually such equations are reduced by an averaging procedure to equations that show the slow dynamics of the envelope of the oscillation and also its relative phase .
for example , this is the case in @xcite where fast oscillating equation set ( 1 ) is converted to the slow dynamics in equation set ( 2 ) .
however , in the current case , it is easier and more useful to solve the fast equation .
we try a solution of the form : @xmath56 - a_{1b } \cos[(1-\omega_b)t],\ ] ] where @xmath57 is the blazhko frequency and @xmath58 and @xmath59 are the upper and lower side - peak amplitudes respectively . as is shown below
, the parameters of the problem will determine the ratio of the two amplitudes and also the blazhko frequency . while the fundamental frequency itself does not appear in the dynamics of @xmath19 , it is of course present in @xmath18 .
note that @xmath57 is not necessarily equal to the detuning @xmath60 .
this is because the nonlinear interaction can shift the average frequency of the nonradial mode , pulling both side - peaks in towards the fundamental . substituting the expression for @xmath19 into equation [ nonrad1 ] and using a trigonometric identity to expand the cosine product , one obtains the following equation relating the coefficients of the terms with frequency @xmath61 : @xmath62 and for @xmath63 one obtains : @xmath64
there are also terms of frequency @xmath65 and @xmath66 generated by the cosine product , however because these terms are far from resonance they can be neglected provided the detuning @xmath67 is small . off
resonance terms are always dropped from amplitude equations because they do not stay in phase long enough to have much impact on the slow averaged dynamics .
for example , in @xcite equation ( 1 ) for @xmath68 there is a resonant coupling term proportional to @xmath69 , but there are dozens of lower order non - resonant coupling terms that are simply ignored and omitted from the equations ( e.g. terms proportional to @xmath70 , @xmath71 , @xmath72 , etc . )
it was verified numerically that the peaks at @xmath65 and @xmath66 are small and go to zero in the limit of small @xmath67 . by also neglecting as small the quadratic terms @xmath73 and @xmath74
, one can solve for @xmath57 obtaining : @xmath75 we will adopt the convention that @xmath57 is always positive , even when the adjusted detuning @xmath67 is negative .
note that the solution for @xmath57 only exists if @xmath76 .
( recall that @xmath53 is nonnegative . )
one can also solve for @xmath77 , the ratio of the amplitudes of the side - peaks .
we will define @xmath77 differently depending on the sign of @xmath67 . for positive @xmath67
we define @xmath78 and for negative @xmath67 we define @xmath79 .
one finds with these definitions that @xmath77 is restricted to the range @xmath80 and is given by : @xmath81 note that equation [ nonrad1 ] is a second order linear homogeneous differential equation with variable coefficients .
such equations are expected to have two linearly independent solutions .
the second solution is obtained by replacing the cosines in equation [ x_1 ] with sines and changing the sign of the second term to plus .
this leads to the same expressions for @xmath77 and @xmath57 .
the most general solution is a linear combination of the sine and cosine forms .
two approaches are useful to examining these equations . in the first we assume that the @xmath67 has a fixed value and see what happens as the coupling strength @xmath53 is varied .
this is shown in figure [ fig_dfix ]
. shows @xmath77 and @xmath82 as a functions of @xmath83 , where @xmath77 is the side - peak amplitude ratio , @xmath53 is the coupling strength , @xmath67 is the adjusted detuning ( see equation [ deltashift ] ) and @xmath57 is the blazhko frequency .
for @xmath84 the modes become phase - locked and there are no side - peaks ( assuming the phase - locked state is stable ) .
@xmath85 is a perfectly symmetric triplet with both side - peaks of equal height .
as @xmath77 is reduced toward zero the triplet becomes increasingly asymmetric . ]
note that the adjusted detuning @xmath67 is typically shifted from the original detuning @xmath86 in proportion to coupling strength according to equation [ deltashift ] .
note that as @xmath53 is increased toward @xmath60 , the blazhko frequency decreases toward zero and the side - peak amplitude ratio approaches unity . for @xmath87 phase locking occurs between the fundamental and the nonradial mode and a net rate of energy flow may occur between them . at the other end of the range ,
when the coupling @xmath53 is decreased to zero , @xmath57 increases to become exactly equal to the detuning @xmath60 and the side - peak amplitude ratio @xmath77 goes to zero . in the weak coupling limit ,
i.e. for @xmath88 , equation [ r ] simplifies to : @xmath89 in the second approach we assume that the coupling strength is fixed and examine what happens if the detuning is changed .
this is shown in figure [ fig_cfix ]
. shows @xmath77 and @xmath90 as functions of @xmath91 , where @xmath77 is the side - peak amplitude ratio , @xmath53 is the coupling strength , @xmath67 is the adjusted detuning ( see equation [ deltashift ] ) and @xmath57 is the blazhko frequency . for @xmath92
there is phase - locking , which ( if stable ) will result in a spectrum containing the fundamental and its harmonics only , with no side - peaks . for @xmath60 slightly greater than @xmath53 , the blazhko frequency emerges from zero and is considerably smaller than the detuning .
the triplet that appears in the spectrum is nearly balanced in side - peak amplitudes .
as the detuning is increased , the side - peak ratio decreases toward zero making the triplet increasingly asymmetric and the blazhko frequency starts to approach the value of the detuning . ] since the detuning can be negative , the phase - locked region actually extends from @xmath93 to @xmath94 . in terms of the unadjusted detuning @xmath86
this range is offset as determined by equation [ deltashift ] , and it is worth noting that if the offset is greater than @xmath53 then the phase - locking range will not include @xmath95 . if there is a cluster of peaks in the vicinity of the fundamental , we would expect the interaction to cause the peaks with the smallest adjusted detuning to become phase - locked and disappear " from the spectrum , with the remaining peaks now generally closer together , each with its own side - peak on the opposite side of the fundamental .
we can also study equation [ nonrad1 ] numerically . in figure [ fig_adiawave ]
we show @xmath96 for a detuning @xmath97 and several values of the coupling strength @xmath53
. shows @xmath19 as a function of time for different values of the coupling strength .
all have detuning @xmath97 . curve ( a ) has @xmath98 , ( b ) has @xmath99 , ( c ) has @xmath100 and ( d ) has @xmath101 .
phase - locking occurs for @xmath102 .
the apparent modulation is caused by the interaction with the fundamental that is constantly changing due to the changing relative phase between the two modes .
note the increasing depth of the modulation and the decreasing modulation frequency as the coupling is increased.,height=576 ] the phase - locking transition point was found numerically to be about @xmath103 .
so it is about 4% higher than the expected value of 0.05 .
the discrepancy is due to the fact that @xmath67 has a finite but small size and thus the analytic results are only approximately correct .
note that the waveform shown is only for the nonradial mode @xmath19 and does not include the fundamental @xmath18
. these will be combined together in some way in the real star to produce the observed light curve .
the modulated appearance of @xmath19 is due to its interaction with the fundamental .
the corresponding spectra are shown in figure [ fig_adiaspec ]
. shows spectra for @xmath19 corresponding to the waveforms shown in figure [ fig_adiawave ] .
the four cases are superimposed for comparison .
each case consists of one left and one right side - peak of the fundamental labeled ( a ) through ( d ) in correspondence with figure [ fig_adiawave ] .
note that as the coupling strength is increased , the peaks move inwards and the amplitude ratio approaches 1.0 . note that the fundamental is not present in the spectra of @xmath19 . ]
only the side - peaks are present in the spectra of @xmath19 ; the fundamental is absent .
the modulation that is observed is at twice the blazhko frequency because this is the spacing between the two side - peaks .
now we will briefly discuss the case of nonzero @xmath39 .
due to stellar rotation , these modes take the form of traveling waves that circulate around the star parallel to the equator . by convention ,
negative values of @xmath39 correspond to prograde modes which travel in the same direction as the stellar rotation , while positive values of @xmath39 correspond to retrograde modes .
the rotation also causes a splitting of the frequencies as a function of @xmath39 which is reminiscent of the well known zeeman splitting in atomic spectra : @xmath104 where @xmath105 is a mode dependent constant .
if one approximates the stellar rotation rate @xmath106 as a constant ( ignoring any dependence of @xmath106 on @xmath107 and @xmath32 ) then it has been found ( see e.g. * ? ? ?
* section 19.2 and equation 19.45 ) that for moderate to large values of the radial index @xmath26 , p - modes have @xmath108 and g - modes have @xmath109\}\omega$ ] .
so in general , @xmath105 is expected to be in the range @xmath110 to @xmath106 , with the low end of the range corresponding to dipolar g - modes .
one of these traveling wave modes by itself can not couple resonantly to the fundamental .
however , a pair with opposite values of @xmath39 can couple to the fundamental and if phase - locked to the fundamental can extract energy from it .
it will generate a triplet in the spectrum and as such is a potential explanation for the blazhko effect @xcite .
due to the frequency splitting , the sum of the pair produces a standing wave pattern that slowly rotates at the frequency @xmath105 . if not phase - locked to the fundamental and if viewed in a frame of reference that co - rotates at frequency @xmath105 , this combined mode can interact with the fundamental in the same manner described in this section . in the nonrotating frame
however , the observed light curve will appear amplitude modulated at the frequency @xmath111 as a result of the changing portion of the stellar surface that is visible from the direction of observation .
if phase - locked to the fundamental this will create a triplet spectrum , otherwise this will create a more complex spectrum .
it is also worth noting that it is possible , at least in theory , for a pair with opposite values of @xmath39 to produce a triplet spectrum with strong amplitude asymmetry .
this was demonstrated in @xcite using amplitude equations .
since saturation effects limit the energy that the pair can acquire through the opacity mechanism , there are nonlinear environmental - type terms appearing in the amplitude equations that generate this limiting effect .
apparently the cross terms between the modes can result in one mode being strongly favored over the other even when the various coefficients for the two modes are only slightly different , i.e. the various coefficients in their equation set 3 .
however , energy can also be acquired by the pair through nonlinear resonant - type coupling to the fundamental mode ( assuming phase - locking ) , and this process favors the pair of modes having equal amplitudes . to the extent that this process dominates , the asymmetry between the pair will be diminished toward zero . retaining the approximation that the fundamental is very robust compared to the nonradial mode
, we can ask what is the effect of the weak coupling terms from @xmath19 on the waveform of @xmath18 . since we know from the observed the light curve and radial velocity waveforms for rr lyrae stars that @xmath18 is operating in a highly nonlinear regime
, we will discuss the effects qualitatively .
the two coupling terms coming from the potential that would appear in the equation for @xmath47 are @xmath112 and @xmath113 .
these terms are assumed to be very weak compared to their counterparts in the equation for @xmath44 because of our assumption of a robust fundamental which leads to @xmath114 being substantially larger than @xmath115 .
the square of @xmath19 appearing in both of these includes the frequency @xmath116 and , when the opposing side - peak is present in @xmath19 , also the frequency @xmath117 .
so the @xmath42 term may put small peaks at those two frequencies in the spectrum of @xmath18 .
the @xmath43 term contains an additional factor of @xmath18 and thus it may generate small peaks in the @xmath18 spectrum with frequencies @xmath118 and @xmath119 and , when the opposite side - peak is present in @xmath19 , also the frequencies @xmath120 and @xmath121 . since both @xmath18 and @xmath19 have an effect on the observed light curve , these peaks may be visible in the spectrum if they are large enough to appear above the noise .
one can also construct a dynamical model that includes both modes explicitly , i.e. one that has equations for both @xmath47 and @xmath44 .
this model , with damping effects also included , is presented and discussed in section [ x0x1 ] .
real stars have excitation and damping , and the equations of motion then contain forcing terms dependent on the velocity
. the linear term in the velocity can result in excitation or damping depending on the sign of the coefficient .
if there is excitation then some nonlinear damping may be required to prevent the mode from increasing without bound .
this could for example take the form of a cubic term in the velocity of the mode in question .
another possibility ( not applicable to the current problem ) is that the excess energy may be transferred to a resonant mode where it is dissipated . for the current problem
, analytic analysis becomes rather difficult or impossible with the additional terms in the equations , so we will explore the results numerically . the dynamics are started from an initial condition and then allowed to settle onto the attractor before the results are analyzed . to equation [ nonrad1 ]
we add a linear excitation term and a cubic damping term : @xmath122 x_1 + 2\kappa \dot{x}_1 + q \dot{x}_1 ^ 3\ ] ] where @xmath123 is the growth rate coefficient .
it must be positive to make the mode unstable .
@xmath124 is the cubic damping coefficient
. it must be negative to restrain the system from unlimited growth . as it is the only nonlinear term in the equation , it is possible to rescale the mode amplitude in a way that makes @xmath125 . as discussed in appendix [ complex ] and also verified numerically , there is only one other possible quadratic or cubic term that can act to restrain the growth of the linear excitation term .
this term would have the form @xmath126 and it will generate nearly identical dynamics when @xmath67 is small ( though with a different scale factor ) . in figure [ fig_disswave ]
we show time series for equation [ nonrad2 ] with two values of the growth rate @xmath123
. shows waveforms for two values of the growth rate @xmath123 .
note the change in the modulation envelope .
both cases have detuning @xmath97 , coupling @xmath100 and nonlinear damping coefficient @xmath127 . ( a ) @xmath128 ( b ) @xmath129 . compare these to the results for @xmath130 given in figure [ fig_adiawave](c ) . ] the detuning @xmath67 and coupling @xmath53 fixed at the same values as used previously in figures [ fig_adiawave](c ) and [ fig_adiaspec](c ) .
as can be seen , in the small @xmath123 limit the dynamics approach the adiabatic results . as @xmath123 is increased , we see the emergence of additional side - peaks in the spectrum and a distortion of the modulation envelope giving it somewhat of a saw - tooth character .
the spectra for the same values of @xmath123 are shown in figure [ fig_dissspec ]
. shows spectra corresponding to the waveforms shown in figure [ fig_disswave ] for two values of the growth rate @xmath123 .
note the emergence of the third and fifth order side - peaks as @xmath123 is increased.,height=576 ] the nonlinearity in the equations generates nonlinear mixing products between the first order side - peaks . as a result ,
the additional side - peaks so created are also spaced by twice the blazhko frequency , i.e. this will only generate side - peaks of odd order .
evidence for this can be found in observational data ; see section [ odd ] for discussion of results from v808 cyg .
most previous blazhko models involve some process that results in the actual modulation of the fundamental mode at the blazhko frequency and it is this modulation that produces the observed effect . there appears to be no obvious reason for this modulation to favor odd - order side - peaks .
we can explicitly include the fundamental by adding an equation for @xmath47 .
although this mode is strongly non - sinusoidal near the surface , one approach is to nevertheless use a simple model that produces a sine wave that may slowly change in response to the interaction .
this is what is usually done , ( see e.g. * ? ? ?
* equation set 1 ) .
one might try to justify this by claiming that most of the energy of the fundamental is far below the stellar surface , and that the oscillation is relatively sinusoidal in that region . for our current case
the resulting equations are : @xmath131 x_0 + 2\kappa_0 \dot{x}_0 + q_0 \dot{x}_0 ^ 3\ ] ] @xmath132 x_1 + 2\kappa_1 \dot{x}_1 + q_1 \dot{x}_1 ^ 3\ ] ] as before , time has been scaled so that the frequency of the fundamental mode ( mode 0 ) is exactly 1.0 and @xmath133 is the unadjusted detuning of the nonradial mode ( mode 1 ) . the @xmath42 term is not used because it is incompatible with our modeling scheme for the fundamental , i.e. lacking strong second harmonic content in the fundamental , this term is nonresonant and will have no impact . deriving the @xmath43 terms from a potential guarantees that energy leaving the fundamental via this term is transferred to the nonradial mode and vice versa . to be near the limit of a robust fundamental we simply need to choose parameters such that @xmath134 , where @xmath135 and @xmath136 are the amplitudes of the two modes which may be slowly varying because of the interaction between them .
provided @xmath137 is small , mode 0 should oscillate with an approximate sine wave of amplitude @xmath135 , and @xmath138 which appears in equation [ nonrad1b ] will have a second harmonic component of amplitude @xmath139 .
thus in the robust fundamental limit , the dynamics of @xmath19 should approximately match the dynamics of the earlier results provided we make the connection : @xmath140 so the range of values of @xmath67 that results in phase - locking is from @xmath141 to @xmath142 . with @xmath18 being approximately a sine wave
, we have @xmath143 , and therefore from equation [ deltashift ] we have @xmath144 .
so the range of values of @xmath86 that results in phase - locking is from @xmath145 to @xmath141 which does not include zero .
the lowest blazhko frequencies should occur just outside this range .
based on the motion being an approximate sine wave , one can solve explicitly for the amplitude by calculating the energy input each cycle by the growth rate and setting it equal to the energy loss through nonlinear damping .
this yields : @xmath146 this equation could be applied to @xmath136 as well as @xmath135 , but in the case of significant coupling , @xmath136 will be varying with time and so the equation may then provide a rough estimate of the average value of @xmath147 . as a test of the previous results the following parameter values were chosen : @xmath148 , @xmath149 , @xmath150 and @xmath151 . for convenience @xmath152 was set to @xmath153 , which by equation [ amplitude ] sets @xmath154 and by equation [ c_2vsc ] sets @xmath155 . numerically integrating the equations , with @xmath86 set to @xmath156
, we obtain a close match to the dynamics shown previously in figure [ fig_disswave](a ) . by equation [ deltashift ] this corresponds to an adjusted detuning @xmath97 so the parameter values are all matched to that previous case , thereby verifying that the approximations used in the simplified equations were valid .
the effect of the interaction on the fundamental is very slight as expected , the modulation envelope of @xmath18 having an amplitude of about 0.003 or about 0.15% of @xmath157 . for a slightly more negative @xmath86 , mode 1
becomes phase - locked to mode 0 , extracts energy from it and increases in amplitude .
so , for example , for @xmath158 , mode 1 is phase - locked and has a constant amplitude of @xmath159 , whereas the maximum amplitude achieved for @xmath160 is about 0.163 .
the fundamental is also loaded down slightly now having an amplitude of about 1.83 , down from its original value of 2.00 .
since this a stable locking of the two modes , there is no modulation and a spectrum would show only a single peak at the fundamental frequency
. thus it would be difficult to determine from observational data that mode 1 was excited .
going still more negative in the detuning , phase - locking is again lost but this time with the adjusted detuning being negative .
one again sees dynamics closely matching figure [ fig_disswave](a ) with @xmath161 .
more can be done with these equations , in particular , one can see how the behavior changes as the amplitude of the nonradial mode is allowed to be closer to that of the fundamental . from a preliminary examination it appears that qualitatively similar results may persist even in cases where @xmath136 is on the order of half the value of @xmath135 .
this may warrant more study at a future date .
two interesting pieces of evidence in support of the model were found in the spectra of v808 cyg : a preference for odd order side - peaks of the fundamental and a mode that appears to be modulated at twice the blazhko frequency . these are discussed below in sections [ odd ] and [ modulation ] respectively .
results given below were calculated from _
project kic 4484128 , q1 through q17 , long cadence corrected flux data using period04 software .
the fundamental frequency @xmath162 is about 1.825273 cycles per day .
the blazhko frequency @xmath57 is about 0.01085 .
average ( or zero point ) flux for this data is about 8835 .
flux of the fundamental is 2857 and the first order side - peaks , left and right , are 539 and 463 respectively .
this is about 1460 days or 4 years worth of data .
the time - step is 0.020435 days and the nyquist frequency is 24.4708 which falls between the 13th and 14th harmonics of the fundamental . the full width at half maximum ( fwhm ) for the spectral peaks is about 0.0008 .
v808 cyg has relatively little asymmetry which indicates according to the theory that there may be a fairly strong interaction between mode 1 and mode 0 ( the fundamental ) . as discussed in section [ dissipative ] , stars with a strong interaction between the modes may tend to favor the generation of odd order side - peaks .
this is in fact found to be the case for v808 cyg as shown in figure [ fig_f0sidepeaks ] , which strongly favors the third order side - peaks over the second order .
part ( a ) shows septuplet spectrum for v808 cyg .
the inner triplet spectral peaks including the fundamental itself have been removed leaving only the second and third order side - peaks on each side as indicated by the labels . note that the second order peaks are smaller than the third order peaks when they might be expected to be much larger .
this is consistent with the effects of nonlinear damping on @xmath19 which generates odd - order side - peaks ( see text and figure [ fig_dissspec ] ) .
the large central peaks ( unlabeled ) are windowing artifacts , i.e. they result from the modulation of the fundamental by the window " made up of the discontinuous chunks of data available for analysis .
part ( b ) shows the window spectrum .
it is offset in frequency by @xmath162 in order to show the effect of the window modulation on that peak .
the artifacts would be larger if the fundamental had not been removed .
the remnant is presumably due to slight variations in the frequency and amplitude of @xmath162 over the time of the data set.,height=480 ] the third order should be much smaller but are instead somewhat larger than the second order .
it should be noted that there is at least two other mechanisms that can generate these peaks including the second order ones .
the first is the nonlinear light - curve function ( or functional ) that transforms the dynamical variables of the excited modes into the observed light curve .
this function ( or functional ) is unknown but is observed in some cases to be approximated by a nonlinear function of the radial velocity ( see e.g. * ? ? ?
this can generate nonlinear mixing products between mode 0 ( the fundamental ) and mode 1 including side - peaks of any order .
the second mechanism comes from the effect of the interaction on the fundamental as discussed in section [ x0effects ] . in the adiabatic case
this will put small second order side - peaks in the spectrum . in the dissipative case
it can also put higher even - order peaks in the spectrum .
the more robust the fundamental is compared to the nonradial mode , the smaller these even - order side - peaks will be .
neither of these other mechanisms can explain the observed preference for the odd - order side - peaks .
the fourth and fifth order peaks ( not shown ) are only marginally detected , being less than a factor of two higher than neighboring noise bumps .
the amplitude of the fourth order are down by about a factor of 5 from the third order , while the fifth order are down about another 30% .
this again indicates a preference for the odd - order cases since one might otherwise expect another factor of 5 drop in amplitude between the fourth and fifth order .
the other mechanisms mentioned above may be playing a more significant role as the order increases .
higher order peaks are not visible .
the measured peak amplitudes are given in appendix [ freqtab ] .
an interesting peak is to be found in the spectrum of v808 cyg .
this peak appears to have modulation side - peaks as shown in figure [ fig_mod ] .
part ( a ) shows @xmath163 a modulated peak found in the spectrum for v808 cyg . note that the side - peaks are displaced from the main peak by multiples of twice the blazhko frequency .
this double blazhko modulation is shown to be consistent with the interactive - beating - modes model of the blazhko effect ( see text ) .
note that the peaks are split , possibly indicating two closely spaced modes or a mode split by stellar rotation and also indicating that at least one and probably both peaks correspond to nonradial modes .
part ( b ) shows the window spectrum .
its frequency is offset by @xmath163 in order to show that window modulation is the source of the noise skirt at the base of this peak in part ( a).,width=576 ] it is close in frequency to @xmath164 but is displaced from it by about 50 times the fwhm of the spectral peaks .
we therefore assume that the peak is generated by an excited mode which we will refer to as mode 2 with frequency @xmath163 and amplitude @xmath165
. we will continue to refer to the fundamental as mode 0 and the mode that is near - resonant with it as mode 1 .
note that @xmath163 is actually split into two peaks with frequencies of @xmath166 and @xmath167 .
one explanation for this split would be that the noisy turbulent environment of the star is causing the amplitude to fluctuate in a way that happens to look like the sum of these two frequencies .
a second explanation is that there really are two closely spaced modes that generate these peaks . in favor of
the second explanation is the fact that the difference between the peak frequencies is 0.002062 and from this one can determine that the two modes slipped in phase by about 3.0 cycles over the time span of the full data set ( approximately 4 years of data from the kepler project ) .
this seems like an excessive number of consistent cycles to be generated by a random process .
also , when the data is divided into two parts of approximately 2 years each , the same split peaks are seen in both spectra with approximately the same amplitudes
. if the split peak is the result of random noise , then it would be quite unlikely to get the same spectra for two non - overlapping time intervals .
thus there probably are two distinct modes involved here and since they are so close together it is reasonable to assume that at least one is a nonradial mode .
a third possibility , which is really a special case of the second possibility , is that the peak is split by stellar rotation , implying mode 2 is a nonradial mode with a nonzero index @xmath39 . since the magnitude of the splitting is related to the rotation rate as discussed in section [ nonzerom ]
, this would imply , for @xmath168 a stellar rotation period for v808 cyg of between 455 and 909 days depending on the characteristics of mode 2 .
there is a constraint on rotation periods based on studies that analyze atomic line - widths .
these studies specify a maximum value for @xmath169 , where @xmath170 is the rotational velocity at the equator and @xmath10 is the angle between the line of sight and the stellar rotation axis . @xcite and @xcite give a value of 10km / s while @xcite give a value of 6km / s .
@xcite , table 1.1 , states that the radii of rr lyrae stars are in the range of 4 to 6 solar radii .
if we choose @xmath170 = 10km / s and @xmath107 = 4 solar radii , we get a minimum acceptable value for the period of about 20 days .
this is well below the values indicated above for rotational splitting .
the side - peaks are also split into pairs of the same spacing .
a total of five pairs of side - peaks are visible , three on the left and two on the right .
the fact that it appears modulated is a strong indication that this peak ( or pair of peaks ) does correspond to an actual vibrational mode ( or pair of modes ) as opposed to the explanation that it is some kind of noisy period - doubling byproduct .
note that any modulation of the fundamental in that case would necessarily produce side - peaks centered on @xmath171 rather than centered on the offset frequency @xmath163 as is observed .
reduced amplitude copies of the modulation pattern can be found at locations that are offset by multiples of @xmath162 .
this shows that these frequencies are mixing with the fundamental and that therefore we can rule out the possibility of @xmath163 being from some other star in the background that appears in very close proximity to v808 cyg .
frequency tables for v808 cyg are given in appendix [ freqtab ] .
perhaps even more interesting is that the apparent modulation frequency , based on the side - peak spacing , is twice the blazhko frequency .
based on measurements from the entire data set , the blazhko frequency is about 0.01085 , so twice this is 0.02170 .
the side - peak spacing is determined to be about 0.02165 so this is a very accurate match .
we have an explanation for this doubled frequency in terms of the interactive - beating - modes model , i.e. the blazhko model presented in this paper .
as explained below , the modulation of mode 2 observed in the current problem may be due to the type of nonlinear interaction that is not phase - dependent .
the blazhko effect generates an oscillatory environment that can have a modulatory effect on the frequency of mode 2 .
this can be seen through the @xmath172 " terms that often appear in the complex amplitude equations , see , e.g. equation ( 1 ) in @xcite , equations ( 1a ) and ( 1b ) in @xcite and equations ( 70 ) and ( 71 ) in @xcite .
( see appendix [ complex ] of this paper for discussion of the complex equations ) . in the equation for @xmath173 in the cited papers
, this term has the form @xmath174 where @xmath172 is a complex coefficient and @xmath175 and @xmath176 are the complex amplitudes of modes a and b. the imaginary component of @xmath172 results in the frequency of mode a being shifted in proportion to the square of the amplitude of mode b , while the real component of @xmath172 results in a similar shift in the growth rate of mode a. note that this term does not depend on the relative phase of mode a and will not cause phase - locking between the modes .
this is in contrast , for example , to the c " terms appearing in equation ( 1 ) in @xcite which can result in phase - locking between modes a and b. so the presence of the @xmath172 term in the equation for @xmath173 can be thought of as an effect of the environment " generated by mode b on the dynamics of mode a. if @xmath177 is constant in time then the result is simply a constant shift in the oscillatory properties of mode a. but if it is slowly varying the result will be to modulate mode a. note that environmental terms are sometimes referred to as saturation terms , see e.g. equations 6 and 7 in @xcite . extending this analysis to the current problem we consider 3 cases . in the first case
we assume there is an interaction between modes 0 and 1 as discussed earlier in this paper which results in @xmath19 having a modulated amplitude as was shown in figures [ fig_adiawave ] and [ fig_disswave ] , with corresponding spectra shown in figures [ fig_adiaspec ] and [ fig_dissspec ] .
since the fundamental does not appear in the spectrum of @xmath19 , the modulation frequency is twice the blazhko frequency ( i.e. it is the spacing between the left and right side - peaks ) .
thus a term of the form @xmath178 appearing in the equation for @xmath179 will generate a shift in the frequency and/or growth rate that is varying at twice the blazhko frequency . a second case to consider is the combined effect of modes 0 and 1 on mode 2 .
the lowest order terms of interest which might appear the equation for @xmath179 are of the form @xmath180 , where @xmath181 and @xmath182 are complex coefficients .
the dominant frequencies of the two terms in the parentheses are @xmath57 and @xmath183 respectively so these terms are slowly varying environmental terms that will modulate mode 2 at the blazhko frequency . however , there are symmetry conditions that apply in a lot of cases that will cause @xmath181 and @xmath182 to be exactly zero .
this occurs when mode 1 is antisymmetric under certain transformations
. modes with @xmath28 and odd @xmath29 are antisymmetric under pole reversal , i.e. looking at the inverted star .
modes with @xmath30 are antisymmetric under rotation about the pole by @xmath184 . the environment for mode 2 is generated by the combined oscillations of modes 0 and 1 and as a result is slowly varying as the relative phase between modes 0 and 1 changes due to the blazhko effect .
due to the antisymmetry , if we move forward in time by one half blazhko cycle the relative phase in the transformed coordinates will exactly match that of the original time with the original coordinates .
but this choice of coordinates can not affect the dynamics and therefore the environmental effect must repeat every half blazhko cycle and have period @xmath185 .
thus @xmath181 and @xmath182 must be exactly zero for these cases .
the only cases which do not have such a symmetry are those with @xmath28 and even @xmath29 .
only these choices for mode 1 can result in modulation of mode 2 at once , rather than twice , the blazhko frequency .
note that as the degree @xmath29 increases , this distinction between odd and even diminishes in importance and so the @xmath186 coefficients for even @xmath29 should tend toward zero to match the odd results as @xmath29 is increased . for the dipolar modes ( @xmath187 ) and other modes with the antisymmetry condition we must use the next higher order term , @xmath188 , which will modulate at the frequency of @xmath185 .
a third case for modulation of mode 2 occurs if the blazhko effect results from the split - pair model of @xcite in which the blazhko effect is produced by a phase - locking between the fundamental and a mode pair , designated 1a and 1b , with opposite values of @xmath30 that is split into two slightly different frequencies by stellar rotation ( see section [ nonzerom ] ) .
these two modes form the two side - peaks of the fundamental . for v808 cyg ,
this would imply , for @xmath168 , a stellar rotation period between 46 and 92 days depending on the characteristics of this mode .
note that this is above the minimum of 20 implied by atomic line - width analysis ( see discussion earlier in this section ) .
this case can also produce the double blazhko modulation , with terms in the equation for @xmath179 of the form @xmath189 .
because of the symmetry properties mentioned above , the terms which could produce modulation at once rather than twice the blazhko frequency are eliminated .
other recent blazhko models would appear to be incompatible with the observed double blazhko modulation of mode 2 .
these generally assume that the blazhko effect involves some process modulating the fundamental mode at the blazhko frequency
. the dominant environmental term in the equation for @xmath179 would then be of the form @xmath190 .
but since fundamental amplitude @xmath191 is fluctuating at the blazhko frequency this will simply modulate mode 2 at that same frequency and not double it . in the purely adiabatic limit
, the terms discussed above can be derived from corresponding terms in the potential energy . if , for example , we consider a term in the potential energy of the form @xmath192 , where @xmath186 is a real coefficient , the equation for @xmath193 will have a corresponding term @xmath194 ( i.e. negative of the derivative with respect @xmath165 ) . following the discussion in appendix [ complex ] ,
the corresponding resonant terms in the equation for @xmath179 are @xmath195 .
thus in this case , @xmath181 and @xmath182 as defined previously are equal and purely imaginary . since the expression multiplying @xmath196 is purely imaginary this corresponds to a pure frequency modulation .
similar results apply to the @xmath172 , @xmath197 and @xmath198 coefficients ; they are all purely imaginary with @xmath199 and @xmath200 and all result in frequency modulation . however ,
if we are not close to the adiabatic limit , then all the coefficients should be regarded as independent complex numbers .
an interesting question is whether the symmetry properties of the dynamics are retained when the amplitude is large and therefore strongly nonlinear .
this may be indicated in the current case by the highly non - sinusoidal nature of the fundamental oscillation .
studies of simpler nonlinear systems have indicated that symmetrical quasiperiodic oscillations may be quite common ; see e.g. figure ( 2b ) in @xcite . as discussed in @xcite
another possibility is complementary asymmetric attractor pairs , though this would seem unlikely except when phase locking occurs ; see e.g. figure ( 3b ) in @xcite .
in this paper we have explored a new model of the blazhko effect called the interactive beating - modes model " .
the model is generated by including a nonlinear interaction between the two modes that make up a model described in a previous work @xcite . in that model the blazhko effect results from the combination of two active modes of nearly the same frequency , one highly non - sinusoidal and one approximately sinusoidal .
that model was shown to accurately model results from rr lyr , which is known to have a very asymmetric triplet spectrum , with the peak to the right of the fundamental being about 5 times the amplitude of the one on the left ( measured ratio @xmath201 for q2 _ kepler _ data ) . in the current work ,
the inclusion of the lowest order interaction terms between the equations governing the two modes is shown to allow the adjustment of the side - peak ratio through the strength of the coupling , including ratios that are fairly close to 1.0 as is the case for v808 cyg ( measured ratio @xmath202 with main side - peak on the left of the fundamental ) . in order to achieve this effect
, the coupling strength must be below a critical value beyond which phase - locking occurs between the modes .
the adiabatic case produces a pure triplet spectrum , but when nonlinear damping is also included , it is found that additional odd - order side - peaks are generated .
this agrees with results from v808 cyg which appears to favor the odd - order side - peaks over the even - order ones .
there appears to be no reason to expect a suppression of even - order side - peaks in other recent blazhko models such as the ninth overtone model .
the concept of environment " is introduced , in which strongly excited modes can alter the frequency and amplitude of weakly excited modes by effectively altering the environment in which they reside .
this is a nonlinear effect and can be characterized by environmental terms ( sometimes referred to as saturation terms ) that appear in the equations of motion . in the case of the blazhko effect ,
the environment so produced is slowly varying at either the blazhko frequency or at twice that frequency depending on the nature of the modes involved .
this varying environment leads to modulation of the weakly excited mode .
an example was found in the case of v808 cyg for a mode which is apparently modulated at twice the blazhko frequency .
this is shown to be consistent with the model , provided the two main modes are the fundamental and a nonradial mode with indices @xmath28 and odd @xmath29 .
other popular blazhko models , such as the ninth overtone model , would appear to be unable to account for this doubling of the modulation frequency .
the complex amplitude equations may be obtained from the real equations as follows .
start with the real set of equations governing the dynamics of a set of interacting modes : @xmath203 where @xmath204 is the @xmath205th mode amplitude , @xmath206 is the corresponding frequency and @xmath207 is a general nonlinear function of all mode amplitudes and velocities [ @xcite also considers possible dependence on accelerations , but we do not discuss this here ] .
we introduce a complex variable @xmath208 , the real part of which is the mode amplitude and the imaginary part is the negative of the mode velocity divided by the frequency : @xmath209 taking the time derivative : @xmath210 however , @xmath211 and @xmath212 and from equation [ realeqn ] @xmath213 where @xmath214 is just @xmath215 expressed in terms of the complex amplitudes and velocities , i.e. replacing @xmath216 with @xmath217 and @xmath218 with @xmath219 for all @xmath34 .
thus from equation [ zjdot ] , @xmath220 we should note that @xmath207 includes all additional terms appearing on the right hand side of equation [ realeqn ] including linear and nonlinear damping and excitation for the @xmath205th mode as well as cross terms involving other modes .
so for example , if @xmath207 contains the linear excitation / damping term @xmath221 then @xmath214 will contain the terms @xmath222 and @xmath223 .
note that @xmath224 is the exponential growth rate for mode @xmath205 which combines both excitation and damping .
if negative , then damping exceeds excitation and the mode will decay exponentially to zero amplitude . if positive then one or more nonlinear terms are needed to prevent the growth from continuing without bound . if @xmath224 is small compared to @xmath206 , then it may be appropriate to ignore the term @xmath223 as non - resonant since its dominant frequency is @xmath225 instead of @xmath206 .
in other words , if we can assume the @xmath208 is approximately proportional to @xmath226 then @xmath227 is approximately proportional to @xmath228 which is strongly nonresonant .
a possible nonlinear damping term in @xmath207 would be @xmath229 with @xmath53 negative .
this expands to four terms in @xmath214 , but the only resonant one is @xmath230 .
the term @xmath231 produces a similar resonant term : @xmath232 .
there are two other cubic terms : @xmath233 and @xmath234 , both of which produce resonant terms that are purely imaginary functions times @xmath208 and so they will affect the phase but not the amplitude of the mode and they can not provide damping .
none of the quadratic terms is resonant .
the data used from v808 cyg is _ kepler _ project kic 4484128 , q1 through q17 , long cadence corrected flux data .
the data was analyzed using period04 software which attempts to fit the data to the form : @xmath235 where @xmath236 is the zero point ( constant component of the flux ) , @xmath237 is the frequency , @xmath238 is the amplitude and @xmath239 is the phase for peak @xmath10 in the spectrum .
note that the phase goes from 0 to 1 rather than @xmath0 .
the calculations are done in _ kepler _ flux units .
@xmath236 for this data is 8835.47 . in table
[ tabf0 ] is shown the fundamental mode @xmath162 , its harmonics and associated side - peaks , while in table [ tabf2 ] is shown the mode @xmath163 and its side - peaks , which may be offset by multiples of the fundamental .
the two components of the split peak @xmath163 are labeled @xmath240 and @xmath241 .
frequencies in the two tables were calculated using @xmath242 , @xmath243 , @xmath244 and @xmath245 .
other frequencies in the tables are linear combinations of these .
they are assumed to be phase - locked and are therefore not allowed to independently vary from their calculated values .
period04 was used to simultaneously determine the optimal amplitude and phase for all of these preset frequencies .
reliability of these results was determined by examining the peaks in the spectrum and characterizing their visibility relative to the background noise .
each peak gets one of three visibility ratings : good , which is used for peaks that are at least twice the height of nearby noise bumps in the data ; low , which is used for peaks that are distinguishable from the noise bumps but are less than twice their level ; and zero , which is used for peaks that are either not present or are not distinguishable from a noise bump .
the low cases could have a fairly high inaccuracy in the amplitude and phase determination .
lllll @xmath246&1.771023&8.79&0.975&low + @xmath247&1.781873&12.91&0.985&low + @xmath248&1.792723&49.27&0.190&good + @xmath249&1.803573&38.13&0.266&good + @xmath250&1.814423&538.87&0.649&good + @xmath162&1.825273&2857.07&0.364&good + @xmath251&1.836123&463.26&0.811&good + @xmath252&1.846973&29.98&0.976&good + @xmath253&1.857823&56.07&0.281&good + @xmath254&1.868673&7.65&0.164&low + @xmath255&1.879523&5.66&0.632&low + @xmath256&3.596296&5.73&0.313&good + @xmath257&3.607146&3.75&0.791&zero + @xmath258&3.617996&49.40&0.392&good + @xmath259&3.628846&8.22&0.016&good + @xmath260&3.639696&449.92&0.815&good + @xmath261&3.650546&1507.69&0.599&good + @xmath262&3.661396&457.72&0.026&good + @xmath263&3.672246&68.62&0.308&good + @xmath264&3.683096&52.21&0.508&good + @xmath265&3.693946&7.82&0.700&low + @xmath266&3.704796&4.49&0.871&zero + @xmath267&5.421569&4.24&0.417&low + @xmath268&5.432419&5.82&0.806&zero + @xmath269&5.443269&50.34&0.645&good + @xmath270&5.454119&70.75&0.206&good + @xmath271&5.464969&421.69&0.048&good + @xmath272&5.475819&839.92&0.873&good + @xmath273&5.486669&391.00&0.299&good + @xmath274&5.497519&85.09&0.654&good + @xmath275&5.508369&49.78&0.761&good + @xmath276&5.519219&12.62&0.077&low + @xmath277&5.530069&2.74&0.213&zero + @xmath278&7.246842&3.16&0.782&low + @xmath279&7.257692&13.80&0.099&good + @xmath280&7.268542&44.11&0.872&good + @xmath281&7.279392&83.58&0.471&good + @xmath282&7.290242&348.07&0.349&good + @xmath283&7.301092&487.07&0.150&good + @xmath284&7.311942&312.60&0.583&good + @xmath285&7.322792&96.64&0.953&good + @xmath286&7.333642&37.60&0.090&good + @xmath287&7.344492&14.87&0.374&good + @xmath288&7.355342&3.15&0.654&zero + lllll @xmath289&0.80738&1.05&0.936&zero + @xmath290&0.809443&4.12&0.009&low + @xmath291&0.82908&5.34&0.573&good + @xmath292&0.831143&6.40&0.871&good + @xmath293&0.85078&4.11&0.363&low + @xmath294&0.852843&3.25&0.192&low + @xmath295&0.87248&22.64&0.725&good + @xmath296&0.874543&16.50&0.264&good + @xmath297&0.89418&6.62&0.372&low + @xmath298&0.896243&8.25&0.882&good + @xmath299&0.91588&2.16&0.103&zero + @xmath300&0.917943&5.56&0.586&low + @xmath301&0.93758&2.51&0.343&zero + @xmath302&0.939643&5.69&0.883&good + @xmath303&2.632653&3.37&0.856&low + @xmath304&2.634716&9.09&0.042&good + @xmath305&2.654353&11.63&0.600&good + @xmath306&2.656416&12.44&0.886&good + @xmath307&2.676053&10.03&0.389&good + @xmath308&2.678116&3.50&0.115&low + @xmath240&2.697753&64.51&0.741&good + @xmath241&2.699816&46.00&0.267&good + @xmath309&2.719453&15.46&0.396&good + @xmath310&2.721516&22.57&0.933&good + @xmath311&2.741153&13.22&0.172&good + @xmath312&2.743216&15.64&0.632&good + @xmath313&2.762853&1.04&0.658&zero + @xmath314&2.764916&4.36&0.122&low + @xmath315&4.457926&2.42&0.143&zero + @xmath316&4.459989&6.14&0.316&good + @xmath317&4.479626&7.66&0.865&good + @xmath318&4.481689&9.61&0.182&good + @xmath319&4.501326&10.61&0.676&good + @xmath320&4.503389&5.32&0.388&zero + @xmath321&4.523026&39.98&0.994&good + @xmath322&4.525089&26.84&0.535&good + @xmath323&4.544726&12.73&0.658&good + @xmath324&4.546789&17.48&0.187&good + @xmath325&4.566426&7.81&0.486&good + @xmath326&4.568489&10.06&0.924&good + @xmath327&4.588126&0.60&0.987&zero + @xmath328&4.590189&3.28&0.311&low + @xmath329&6.283199&2.33&0.433&low + @xmath330&6.285262&5.68&0.565&good + @xmath331&6.304899&5.64&0.059&good + @xmath332&6.306962&6.64&0.401&good + @xmath333&6.326599&6.85&0.896&good + @xmath334&6.328662&3.28&0.676&zero + @xmath335&6.348299&13.42&0.206&good + @xmath336&6.350362&5.32&0.779&good + @xmath337&6.369999&8.94&0.872&good + @xmath338&6.372062&12.79&0.426&good + @xmath339&6.391699&5.14&0.697&good + @xmath340&6.393762&8.28&0.132&good + @xmath341&6.413399&0.72&0.316&zero + @xmath342&6.415462&2.60&0.587&low + | this paper considers a nonlinear coupling between a radial and a nonradial mode of nearly the same frequency .
the results may be of general interest , but in particular have application to the beating - modes model " of the blazhko effect which was recently shown to accurately reproduce the light curve of rr lyr . for weak coupling , the two modes do not phase - lock and they retain separate frequencies , but the coupling nevertheless has important consequences . upon increasing the coupling strength from zero ,
an additional side - peak emerges in the spectrum forming an asymmetric triplet centered on the fundamental . as the coupling
is further increased , the amplitude of this side - peak increases and the three peaks are also pulled towards each other , decreasing the blazhko frequency . beyond a critical coupling strength ,
phase - locking occurs between the modes . with appropriate choice of coupling strength ,
this interactive beating - modes model " can match the side - peak amplitude ratio of any star .
the effects of nonlinear damping are also explored and found to generate additional side - peaks of odd order .
consistent with this , the odd side - peaks are found to be favored in v808 cyg .
it is also shown that the blazhko effect generates a fluctuating environment " that can have a modulatory effect on other excited modes of the star .
an example is found in v808 cyg where the modulation is at double the blazhko frequency .
an explanation is found for this mysterious doubling , providing additional evidence in favor of the model . |
this paper recounts the results of a project to simulate the structure of decagonal quasicrystal aluminum - nickel - cobalt @xmath2(alconi ) in the `` basic co '' ( cobalt rich ) phase purely from energy minimization principles . of the equilibrium decagonal quasicrystals , @xmath2(alconi )
has ( in some of its modifications ) the highest structural quality and has received the most study .
studies of the phase diagram indicate that , e.g. , decagonal @xmath2(al@xmath3co@xmath4ni@xmath5 ) is stable ( at higher temperatures only ) , whereas @xmath2(alco ) is metastable only @xcite .
recently , a computational approach was proposed for discovering the atomic structure of any decagonal quasicrystal , given no information except a set of pair potentials , the quasilattice constant , and the periodic lattice constant ; it was applied first to @xmath2(alconi ) in the `` basic ni '' phase ; @xcite . in the study described here ( and briefly reported elsewhere @xcite ) ,
the same approach is applied to `` basic co '' for the first time , and shown to work . as in the `` basic ni '' case ,
our final structure description is in terms of a supertiling with a large quasilattice constant , but here different clusters and tiles enter than in the `` basic ni '' case . since the sensitivity of the structure to the precise composition is one of the issues in this paper ( e.g. in sec .
[ sec : bni - compare ] ) , and since known structures of crystalline `` approximant '' phases greatly illuminated our understanding of the related quasicrystals in the past , we shall pause to review what is known in the al - co - ni phase diagram .
the decagonal portion of the al - co - ni phase diagram is fragmented into several modifications occupying small domains .
@xcite . of these ,
those showing the simplest diffraction patterns are the so - called `` basic nickel '' phase near the ni - rich composition al@xmath6co@xmath7ni@xmath8 and so - called `` basic cobalt '' near the co - rich composition al@xmath6co@xmath8ni@xmath7 .
several high - resolution x - ray structure determinations were carried out for the `` basic ni '' phase @xcite , but studies of the `` basic co '' phase have lagged .
an interesting aspect of the co - rich portion of the phase diagram is the fivefold ( rather than tenfold ) symmetric decagonal phase , @xcite .
in particular @xmath2(al@xmath3co@xmath8ni@xmath9 and @xmath2(al@xmath3co@xmath10ni@xmath11 @xcite , also @xmath2(al@xmath12co@xmath13ni@xmath14 ) @xcite .
this and other co - rich modifications show superstructure diffraction peaks , indicating modulations of the `` basic '' structure : @xmath2(al@xmath3co@xmath15ni@xmath7 ) , similar to the `` fivefold '' modification @xcite , and @xmath2(al@xmath16co@xmath8ni@xmath17 ) , which has a period of 61 in one direction and thus was called the `` one - dimensional quasicrystal '' @xcite . throughout the phase diagram ,
the _ quasilattice constant _
@xmath18 is close to @xmath19 .
there is a solved periodic crystal approximant of `` basic co '' , @xmath1(alconi ) structure @xcite .
there are further modifications near to the `` basic ni '' composition as well as near @xmath2(al@xmath6co@xmath20ni@xmath20 , and another ( partially solved ) approximant @xcite , with unit cell @xmath21 @xmath22 @xmath23 , and composition al@xmath16co@xmath24ni@xmath24 .
the `` basic co '' phase has a @xmath25 stacking period , like `` basic ni '' , but it shows much stronger diffuse scattering than `` basic ni '' , @xcite in such a way as to indicate a local doubling of the @xmath26 periodicity ( to @xmath27 ) ; the known large approximants also have @xmath0 . [
later in this paper ( sec .
[ sec : doubled ] ) , we shall address the stabilizing effects of this period doubling . ]
our general technique is the same as those used in the previous work on `` basic ni '' .
we begin with a small - scale rhombus tiling and discover general motifs and patterns .
these patterns usually have a geometry consistent with an inflated penrose tiling : we define a new model using that tiling , and the patterns observed at the small scale are promoted to be fundamental objects on the inflated tiling . by restricting configurations and increasing the size of fundamental objects , we can run simulations on larger and larger unit cells without excess degrees of freedom , speeding up the mmc process considerably .
the outline of the paper is as follows . after reviewing the technique and the information needed in its set - up ( sec .
[ sec : set - up ] ) , we present initial results in sec .
[ sec : fixed - site ] from monte carlo lattice - gas simulations using a discrete site - list , both with the initial edge-@xmath18 rhombi and also with @xmath28-edge bilayer rhombi ; in particular , the whole structure is built from two cluster modifs
the and the .
next , sec .
[ sec : ideal - deco ] codifies this by describing an ideal decoration , which requires specification of the orientations as well as the optimum placement of a subset of easily moved al atoms . in sec .
[ sec : relax ] , we pass on to molecular dynamics and relaxation studies that break free of the discrete - site lists ; these reveal troughs ( which we call `` channels '' ) in the potential energies felt by al atoms which lead to local disruption of the layering of the atoms , and a breaking of the two - layer peridiocity assumed in previous sections . here and in secs .
[ sec : declro ] and [ sec : puckerlro ] , we take up the correlations in the atoms displacements , and also how this determines the an ordering of the orientations of clusters which reduces the system s symmetry to pentagonal .
we conclude with an application to w(alconi ) , the best - known approximant of co - rich quasicrystal @xmath2(alconi ) , in sec .
[ sec : w ] , and a discussion ( sec .
[ sec : discussion ] ) of the key results and the limits on their validity .
in this section , we lay out the procedures of the simulation , as well as the assumptions and facts that all our results depend on .
our methods are a combination of metropolis monte carlo ( mmc ) , relaxation , and molecular dynamics ( md ) .
we first perform mmc on a set of fixed sites .
we create this set by make use of a _ tiling _ of penrose rhombi on each layer and by placing atomic sites on each of the rhombi using a _
decoration_. fig .
[ fig : tileflips ] shows penrose rhombi and two decorations that we use .
[ see appendix [ app : code ] for a more detailed description of the decoration and tiling . ]
penrose rhombus tilings ( even random ones ) have a natural inflation rule whereby the same space can be retiled with rhombi whose edges are a power of the golden ratio @xmath29 multiplied by the orginal edge length . in this paper , we will make use of rhombi with edges @xmath30 , @xmath313.96 , and @xmath3210.4 ; we shall also mention a similar tiling with edges @xmath33 that applies to the ni - rich @xmath2(alnico ) phase , a structure closely related to the one we are investigating .
a unit cell can be tiled in many different ways with the same number of penrose rhombi ; this is physically important since the different tilings correspond to many different configurations of atomic sites that are consistent with the same physical cell and the assumptions based on the lattice constants . to explore this degree of freedom
, we perform mmc on the rhombic tiles by using rearrangements of three penrose rhombi ( and the atoms on them ) that preserve their collective hexagon outline .
the two rhombus configurations for which this is possible are shown in fig .
[ fig : tileflips ] .
, can be flipped with respect to their asymmetric axes as a way to move sites around . , width=163 ] the mmc is performed on a temperature schedule specified by the beginning and ending inverse temperatures @xmath34 along with an inverse temperature step @xmath35 . at each temperature step ,
a set number of mmc operations per site are performed .
after we find a configuration with this `` fixed - site '' method , we can remove the site list restriction , and use relaxation and md to find a structure that is more energetically favorable by our potentials .
why is our method to start with tilings , decorations , and discrete sites , and iterate this ( as outlined in the introduction ) , rather than immediately perform md ?
the reason is that the energy surface of @xmath2(alconi ) in configuration space contains many local energy minima
. a pure md program would be almost certain to be trapped in a glassy configuration .
[ even with a small number of atoms and a simpler set of potentials , extremely long md coolings were necessary in order to produce recognizable ( but still quite defective ) quasicrystals by brute force .
@xcite ] the only experimental inputs into the procedure are lattice constants , composition , and and density ; the only theoretical input is the pair potentials . for the initial trials
, one must also make a discrete choice of which size of rhombus to use the quasilattice constant of a decagonal tiling is defined only modulo factors of @xmath36 and one must decide how many atomic layers are to be simulated . .
pair potential minima @xmath37 for al - co - ni .
the `` ( 0 ) '' well is the hard core radius , defined as the minimum radius actually found in a relaxation ( after molecular dynamics annealing ) of an example configuration ; @xmath38 is defined by @xmath39ev .
minima are listed only for @xmath400.025 ev . [ cols="<,>,^,^,>",options="header " , ]
in this appendix , we compare our results to three recent experiment - based structure models , all based on some sort of 20 diameter decagon ( the first two being similar to the burkov cluster of subsec .
[ sec : ideal - burkov ] ) .
we then report on a trial simulation of our own , using the methods of sec .
[ sec : fixed - site ] but with an enriched site list for the lattice gas , from which a 20 decagon packing emerges that is competitive in energy with the 13 model developed in this paper .
after most of our work was completed , a tentative structure solution appeared for the approximant pd4 of the co - rich phase @xcite with nominal composition al@xmath3co@xmath4ni@xmath5 .
the @xmath26 projection ( see their fig .
3 ) clearly shows an arrangement of decagonal clusters of diameter @xmath41 ( larger than the by exactly the factor @xmath36 ) .
this 20 cluster appears practically the same as the burkov cluster which as we explained in subsec .
[ sec : ideal - burkov ] appears in our structure model .
namely , this decagon has rings 1,2 , and 3 like our of sec .
[ sec : clusters ] ; their ring 3 al show deviations into ring 2.5 , similar to what occurs in our structures . in half of their clusters , ring 1
is missing a couple of al atoms ; these have more irregularity of their ring 2.5/3 al atoms .
the big difference is that , in pd4 , the 20 decagons do not overlap ; instead , they decorate a fat rhombus with edge 10.4 .
thus , the cluster - cluster linkages in this model are of length @xmath42 , for 20 decagons sharing an edge , or @xmath43 , for 20 decagons related by the short diagonal of a fat rhombus .
it may also be noted that the cluster orientation pattern in pd4 is neither `` ferromagnetic '' nor `` antiferromagnetic '' ; this would suggest that ( in terms of sec .
[ sec : declro ] the 12.2 linkage induces an `` antiferromagnetic '' interaction .
although the `` '' motifs no longer sit at vertices of this 19.7 edge network , that atom cluster is still in evidence .
the difference is that in the pd4 structure , every is encircled by ten such , whereas this number was smaller in our model of fig .
[ fig : ruleswideal ] . the pd4 atomic structure , like the rigid - site - list models of sec . [ sec : ideal - deco ] , can be decomposed into a dhbs ( decagon - hexagon - boat - star ) tiling with edge @xmath44 .
in fact , a decagonal model based on pd4 is a concrete example of the structure models intermediate between the `` basic ni '' decoration of ref . and the decoration of our sec .
[ sec : ideal - deco ] .
such a hybrid model would have 8 decagons , which are absent in the former decoration , but have a larger proportion of 2.45 hbs tiles than in the latter decoration .
( in particular , every edge of the 20 decagon is the centerline of a 2.45 hexagon , decorated typically by a ni - ni pair . )
we expected such an intermediate model to be favored at an intermediate composition such as al@xmath6co@xmath20ni@xmath20 , but the energy differences are quite small , so it may well be competitive at the compositions the present paper focuses on . another decagonal approximant was discovered with composition al@xmath16ni@xmath45co@xmath46 , close to the `` basic ni '' phase , and a structure model developed based on electron diffraction and z - contrast imaging @xcite .
this model consists of edge - sharing 20 decagons of the kind we have been describing .
they have al@xmath47tm@xmath48 cores , _ alternating _ in orientation ( the cluster network happens to be bipartite ) .
there are no ring 2.5 al ; about half the edges of the ( contained in the 20 one ) have two ring 3 al , the other half of the edges have only one ring 3 al . the edges of the 20 decagons , without exception , have two tm atoms . a good interpretation of the cluster network in ref .
is that the 20 ( edge 6.4 ) decagons occupy the large sites of a binary tiling with edge 16.8 , while pentagonal bipyramids ( ) occupy the small sites .
the vertices of the 20 decagons , as in the pd4 model , are occupied by motifs like the .
such also occur between pairs of adjacent .
there , they define additional tile vertices which divide the area between the 20 decagons into hexagons , boats , and stars with edge 6.5 . deloudi et al @xcite have presented a structure model ( formulated in a 5d - cut framework ) for the same co - rich composition we address here .
they formulate a 20 pentagonal cluster , which is built around the w - phased pentagon cluster ; these can overlap in various ways .
the al / tm assignments in the model are based entirely on those reported in sugiyama s @xmath1(alconi ) refinement @xcite .
examination of their fig .
1 reveals the relation between their cluster and our motif .
take each vertex of their 20 decagon that is _ not _ a vertex of the inscribed large pentagon , and draw an arc around it through the adjacent decagon vertices ( which _ are _ vertices of the inscribed pentagon ) ; this arc encloses 3/10 of a .
their model , being based on a stacking period @xmath49 , accomodates not only the pentagonal bipyramid ( ) , but also the puckering of the al atoms we called ring 2.5/ring 3 ( or better , channel atoms ) along the arcs just mentioned . in fig .
3(a ) of ref .
it can be seen that , although not recognized at all in their formulation , are naturally generated in the interstices between the large pentagons .
the overall pattern could be decribed as a packing of edge - sharing 32 super - decagons ( inflated by one factor of @xmath36 ) , with placed on the super - decagon centers and odd vertices , and on the super - decagon even vertices .
( it is not surprising to find descriptions on different length scales : their model , being essentially a decoration of the quasiperiodic penrose tiling , acquires its inflation symmetry . )
we question whether the details of this model are good energetically .
our study suggests that good models are built by combinations of the 8 decagon ( the central part of the ) , of 2.45 edge hbs tiles , plus the . to the extent that this model appears to have incomplete fragments of the 8 decagon , we suspect it will have an increased energy .
it is conceivable that only al atoms are wrong , and the tm atoms are correctly placed , which would still give good agreement with electron micrographs of all kinds .
the show up prominently in the experimental haadf - stem image of ref .
( their fig . 3(b ) ) as white pentagons , all oriented the same way .
these are not as frequent as they would be in the decoration of our fig .
[ fig : ruleswideal ] , suggesting that our model is not correct for that composition . fig .
[ fig:20a ] shows an exploratory fixed - site lattice - gas simulation on a 4 tiling in the 32@xmath5023 cell , similar to sec .
[ sec : mc-4a ] , but with a richer set of candidate sites decorating the tiles , and a much longer annealing time .
two 20 decagons are seen in this configuration , in place of the four 13 decagons that typically emerged in our main simulations .
the atomic structure is quite similar to those described above ; a minor difference is that every edge of the has exactly one ring 3 al , and their placement alternates perfectly .
( that alternation was necessarily disrupted in our model of sec .
[ sec : ideal - deco ] , wherever the s shared edges . )
also , the edges of the 20 decagon are not only ni - ni , but often al - co or ni - co in fig .
[ fig:20a ] .
it would take much more work to settle the optimum decoration of these sites , and the optimum atom content for a decoration based on such configurations . the 20 based configuration of fig .
[ fig:20a ] is lower in energy by 1.8 mev / atom as compared to the best result of the 13 sort produced by our 4 simulations , if ideal - site configurations are compared .
however , after the rmr protocol in an 8 cell ( see sec . [
sec : relax ] ) , the 20 type structure was slightly _ higher _ in energy , by 0.6 mev / atom , than the 13 type structure .
( the latter probably contains more al channels , and thus offers more opportunity to reduce energy by puckering . )
overall , it is our impression that despite the deficiencies ( known and still unknown ) in our potentials , the biggest problem for achieving a correct structure is the search pathway to find it . on the one hand , we think our method has been remarkably successful at predicting characteristic features such as the appearance of the in co - rich compositions .
on the other hand , we have recounted four ways in which our procedure misled us : ( i ) the fixed site list ( ii ) the 8 periodicity ( iii ) the possibility of a larger cluster than ( iv ) the cluster orientation order a separate note is that our calculations are good only to predict stability between similar decagonal approximant structures .
many deficiencies of our potentials the dependence on electron density ( and hence on composition , in principle ) ; inaccuracies in the nearest - neighbor tm - tm and tm - al potential wells ; cutoffs at some interaction radius ; omission of three- or four - body terms ; and poor handling of vacancies will tend to cancel , in such a comparison .
but any phase may , of course , be preempted by a coexistence of two dissimilar phases that happens to have a slightly lower energy ; our pair potentials are more likely to give a wrong answer in this situation . to construct a global phase diagram ,
it is necessary to follow up the kind of search described in this paper , by ab - initio total energy calculations , with an attempt to imagine _ all _ possible competing phases and include them in this database . since it would be prohibitive to try out a large set of candidate structures with the ab - initio codes , the present sort of study is a prerequisite to the phase - diagram studies .
s. ritsch , c. beeli , h. u. nissen , t. gdecke , m. scheffer , and r. lck , phil .
lett . , 74 , ( 1996 ) 99 - 106 s. ritsch , c. beeli , h. u. nissen , t. gdecke , m. scheffer , and r. lck , phil . mag . lett .
78 , 67 - 76 ( 1998 ) .
philos . mag . lett .
78 , 67 ( 1998 ) h. takakura , a. yamamoto , and a. p. tsai , acta crystallogr . a 57 : 576 ( 2001 ) ; a. cervellino , t. haibach , and w. steurer , acta crystallogr .
b 58 , 8 ( 2002 )
. b. grushko , d. holland - moritz , and k. bickmann , j. alloys comp .
236 , 243 ( 1996 ) decagonal quasicrystals in al - co and ternary x. z. li , r. c. yu , k. h. kuo , and k. hiraga , phil . mag
73 , 255 ( 1996 ) .
s. ritsch , c. beeli , and h .- u .
nissen , phil . mag . lett .
74 , 203 ( 1996 ) .
s. ritsch , h .- u .
nissen , and c. beeli , phys .
76 , 2507 ( 1996 ) .
s. ritsch , o. radulescu , c. beeli , d. h. warrington , r. lck , and k. hiraga , phil . mag .
80 , 107 ( 2000 ) .
k. sugiyama , s. nishimura , and k. hiraga , j. alloy comp .
342 , 65 ( 2002 ) .
w. steurer , personal communication .
f. frey , e. weidner , k. hradil , m. de boissieu , r. currat , k. shibata , a. p. tsai , and t. j. sato , phil . mag .
a. 80 , 2375 - 91 ( 2000 ) f. frey , mater .
eng . a 294 , 178 - 85 ( 2000 ) .
m. mihalkovi , i. al - lehyani , e. cockayne , c. l. henley , n. moghadam , j. a. moriarty , y. wang , and m. widom , phys .
b 65 , 104205 ( 2002 ) .
n. gu , c. l. henley , and m. mihalkovi , phil . mag .
86 , 593 ( 2006 ) .
m. mihalkovi , c. l. henley , and m. widom , j. non - cyst .
sol . 334 : 177 ( 2004 )
. j. roth and c. l. henley , philos . mag .
a , vol 75 ( no 3 ) , 861 ( 1997 ) . j. a. moriarty and m. widom , phys . rev .
b 56 , 7905 ( 1997 ) .
i. al - lehyani , m. widom , y. wang , n. moghadam , g.m .
stocks and j.a .
moriarty , phys .
b 64 , 075109 ( 2001 ) .
the gpt potentials , although systematically founded , appear to underrepresent the real interactions in @xmath2(alnico ) by a factor of up to 2 , as inferred from comparisons to experimental phonon spectra ( m. mihalkovic , unpublished ) . m. widom and e. cockayne , physica a 232 , 713 ( 1996 ) .
s. katrych , m. mihalkovi , v. gramlich , m. widom and w. steurer , phil . mag .
a 86 , 451 ( 2006 ) .
m. mihalkovi and m. widom , phil . mag .
a 86 , 557 ( 2006 ) . in the initial stage monte carlo simulations ,
a better procedure would have been to use ( at least ) two densities , bracketing the physical one , as elaborated in sec .
[ sec : disc - discrete ] . in some runs ,
samples were only cooled to @xmath51 , corresponding to @xmath523000 k ( which would be well above the melting point , but the constraint of a discrete site list makes a large difference ) but low energy structures obtained at these high temperatures were quite similar in configuration to those obtained from low temperatures , in that the decagonal rings were pretty well bound together : the structures that are favorable @xmath53 are , in general , the same as those favorable at @xmath54 . the initial stage simulations in ref .
for the `` basic ni '' phase used a smaller unit cell that was probably _ too _ small .
if used in the `` basic co '' case , it would have hindered us from discovering the right motifs . c. l. henley , phil .
a 86 , 1123 ( 2006 ) ; c. l. henley , m. de boissieu , and w. steurer ( eds . ) , phil . mag . a 86 , 1131 ( 2006 ) .
k. saitoh , m. tanaka , a. p. tsai , and c. j. rossouw , j. phys .
69 , 2379 ( 2000 ) .
b. grushko and m. dblinger , z. kristallogr .
219 , 447 ( 2004 ) .
more generally , a potential map may defined whenever a securely occupied subset of sites can be identified .
its efficacy depends on how nicely the problem of placing the remaining atoms factorizes into independent problems . in the present case ,
we might ( after filling the deep wells with al ) define a second potential map , which would be more realistic since the undetermined al atoms obviously do interact with the well - determined ones .
however , the images of such a potential function ( which is dominated by forbidden zones where the atoms have been filled in ) are harder to relate to the atomic structures , and we did not use them in the present work . observe how the ring-1 al well extends in the @xmath55 plane outward from the cluster center ( in the as well as ) . helping explain shifts in the position under relaxation ( subsec . [
sec : relax - bilayer ] ) , and predicting an anisotropic debye - waller factor in diffraction experiments owing to fluctuations in that well .
if we simply take the idealized model of fig .
[ fig : ruleswideal ] , and rotate every alternate core by the 10-fold screw axis without any other modification , the energy increases by by @xmath561 ev per cluster .
this confirms how the ring 3 and atoms adapt to various orientations so as to make them nearly degenerate .
note that @xmath57 is _ not _ the same as the variable @xmath58 that was defined in sec .
[ sec : channels ] and used in appendix [ app : channel - anal ] . in fact , the @xmath59 are well - defined only if certain layers ( which we called 0 and 2 ) are mirror layers globally , and that happens only when @xmath60 takes the same value everywhere . in terms of the collective coordinate @xmath61 defined in eq .
( [ eq : zm ] ) to parametrize the sliding of three al atoms in a channel , we could say @xmath62 , whereas ( assuming a first symmetry - breaking @xmath63 ) @xmath64 . in the simulations reported here , runs were performed as a series of repeated cooling cycles .
each succeeding cycle started from the last configuration of the previous cycle , at a very high temperature , so that we did nt need to re - initialize random seed .
eventual variant runs were performed with different run parameters , or started from different seed configuration .
p. oleynikov , i. demchenko , j. christensen , s. hovmller , t. yokosawa , m. dblinger , b. grushko , and x. d. zou , phil . mag .
86 , 457 ( 2006 ) .
their figure lacks @xmath72 information ( which might reveal puckering ) . | a systematic , decoration - based technique to discover the atomic structure of a decagonal quasicrystal , given pair potentials and experimentally measured lattice constants , is applied to the `` basic '' cobalt - rich decagonal al - co - ni quasicrystal .
first lattice - gas monte carlo simulations are performed , assuming the atomic sites are vertices of a rhombus tiling with edge 2.45 .
this phase is found to be dominated by 13 diameter decagon - shaped clusters , each with a pentagon of co atoms at the center .
these , and another smaller cluster , decorate vertices of a `` binary tiling '' with rhombus edge 10.4 . further simulations with a restricted site list show that al arrangements on the borders of the 13 decagon cluster form hexagon , boat , and star tiles with edge 2.45 ; they indicate specific sites for co versus ni atoms , and how the structure adapts to small composition changes . in the second half of the paper , relaxation ( augmented by molecular dynamics annealing )
is used to obtain realistic structures .
the dominant new feature is a set of linear `` channels '' attractive to al atoms and running transverse to the layers .
each is typically occupied by three atoms in four layers , implying puckering and a spontaneous period doubling to @xmath0 .
puckering favors pentagonal long range order of the cluster orientations .
our simulation captures most features of the related @xmath1-alconi crystal , except for its pentagonal bipyramid motif .
# 1 # 1 # 1 # 1 # 1 # 1 # 1 # 1 # 1j^puck_#1 |
with remarkable physical properties , graphene is currently in the scientific limelight not only for possessing tremendous technological potential but also for having opened several avenues of basic science exploration @xcite .
the honeycomb lattice structure of graphene and its two sublattices are reponsible for a variety of novel physics phenomena .
recent theoretical reports suggest that doping one sublattice with vacancies @xcite , adsorbates @xcite and substitutional impurities @xcite might lead to the appearance of band gaps , and in the substitutional case induce spin - polarized current @xcite .
this unusual feature may pave the way to engineer the transport properties of doped graphene leading to a new species of transistor @xcite . among the unresolved issues in the field
is the recent observation in cvd grown graphene on cu substrates that substitutional nitrogen dopants display a rather unusual asymmetry - heavily favouring sites on one of the two graphene sublattices @xcite .
density functional theory ( dft ) calculations by zabet - khosousi et al .
@xcite suggest individual nitrogen impurities attaching to the edge of a graphene sheet prefer to occupy one sublattice over the other , which would lead to domains of segregation agreeing with the graphene grain boundaries .
however , such a relationship between crystal grain and the sublattice segregation domains of the impurities has not been seen experimentally . with this as a motivation , we address the problem of why substitutional impurities in graphene display such an asymmetry in their configuration . besides offering a possible explanation for the observed sublattice asymmetry , our results suggest that this feature could be produced with other impurities under certain circumstances .
graphene sublattices , hereafter referred to as a and b , are triangular lattices that span the hexagonal structure when they are superimposed a suitable distance apart .
being perfectly equivalent to one another , the only way that impurities may prefer one of the sublattices is if this sublattice symmetry is broken .
symmetry breaking operations are known to induce some degree of segregation in the way impurities are spread across the structures .
graphene edges , for example , can break the sublattice symmetry leading to a modulation in the spatial distribution of impurities @xcite .
in fact , theoretical calculations performed by some of the authors have suggested that it is possible to have one energetically preferential sublattice when impurities are near an edge .
while this might indicate a possible explanation for the puzzling asymmetry seen in the case of nitrogen doping , such effects are strongly dependent on the edge geometry and only occur in the close vicinity of edges , which can not explain the experimentally observed doping asymmetry across the entire sample . here
we argue that , rather than edge effects , the substitutional impurities themselves are responsible for breaking the symmetry and for making one of the graphene sublattices more energetically favourable than the other .
similar symmetry breaking arguments have been put forward for the case of adsorbed impurities @xcite , but the effect is less robust than the substitutional case .
most importantly , in this manuscript we point to the existence of a critical impurity concentration beyond which the asymmetry disappears and one would expect to find dopants evenly distributed between sublattices .
to demonstrate that impurities are the key symmetry - breaking agents we must account for the total energy balance of a system with many impurities
. we will start with the energy contribution from a single impurity , then a pair of impurities , and finally a system with a finite concentration of randomly distributed impurities . to consider such a range of systems , a theoretical framework which can account for a scalable number of randomly placed impurities is essential .
one such method can be found within the tight - binding ( tb ) model which allows us to describe the electronic structure of graphene as well as that of the substitutional impurities .
the tb hamiltonian for a pristine graphene sheet is @xmath0 , where @xmath1 is the electronic hopping between nearest neighbour carbon atoms , the basis @xmath2 represents a single atomic orbital located at a site that is identified by the unit cell @xmath3 in the sublattice @xmath4 , and the sum runs over all nearest neighbours .
one key quantity in this study is the total energy change that results from the introduction of a perturbing potential @xmath5 due to the presence of substitutional impurities .
the electronic contribution to the total energy change ( @xmath6 ) of the system is fully described by the lloyd formula method @xcite and results from the changes in the electronic density by the introduction of @xmath5 : @xmath7 where @xmath8 is the fermi function , @xmath9 is the single - particle green function ( gf ) associated with the tb hamiltonian of a pristine graphene sheet and @xmath10 is the unit operator .
the integrand above depends on the form of the perturbation @xmath5 , which for the case of a single impurity is @xmath11 , corresponding to a substitutional impurity with onsite energy @xmath12 localized at the unit cell @xmath13 in sublattice @xmath14 .
alternative forms of @xmath15 that include extra matrix elements may be used but do not modify the results found here , due to the underlying symmetry contained in the perturbation potential . in the case of a single impurity , eq.([lloyd ] ) becomes @xmath16 where @xmath17 is the diagonal matrix element of the single - particle gf , and we have omitted the integration range and the fermi function for conciseness . since the diagonal element of the single - particle gf carries no position dependence it is not necessary to solve the integral of eq.([lloyd-1imp ] ) to show that the energy cost @xmath18 is position independent . when a second impurity is added to the system , at some site @xmath19 , on sublattice @xmath20
, the energy is not simply twice @xmath18 , but is given by @xmath21 , where @xmath22 contains multiple scattering effects between the two impurities and is written as @xmath23 since @xmath22 contains off - diagonal matrix elements of the gf , it will depend on the relative position of impurities 1 and 2 , and we note that with only two impurities embedded in an infinite system the fermi energy ( @xmath24 ) will not be shifted from that of pristine graphene .
the functional form of @xmath22 has been derived based on the oscillations in the local density of states ( ldos ) of graphene in the presence of a pair of substitutional impurities , and is seen to behave as a decaying sinuosoid @xmath25 , where @xmath26 , @xmath27 and @xmath28 are functions of the fermi energy and @xmath29 is the separation of the two impurities .
[ fig1 ] plots @xmath22 against @xmath29 using the parameterisation @xmath30 , showing a negative value for @xmath31 and a positive value for @xmath32 at @xmath33 for all separations , i.e. it is energetically more favourable to accommodate a second impurity on the same sublattice at @xmath33 , regardless of separation , while for @xmath34 the energetic favourability is separation dependent .
note the lack of oscillations in fig .
[ fig1]a despite its explicit dependence on @xmath35 . here
the oscillatory character of @xmath22 is suppressed by the commensurability of the possible separations and the oscillation period @xmath36 .
such a commensurability feature has been previously reported for nanotubes but is also prevalent in graphene .
the physical mechanism that drives this energetic preference for one of the sublattices can be understood as a result of the ldos oscillations .
these oscillations arise from the presence of nitrogen dopants , which cause significant changes in the ldos ( see references @xcite and fig .
[ fig : ldos ] ) . in general terms ,
ldos oscillations induce oscillations in other quantities ( often referred to as friedel oscillations ) that depend on the spatial distribution of impurities @xcite .
because the total energy is one such a quantity , @xmath22 tends to oscillate between positive and negative values but the commensurability effect masks these oscillations in such a way that impurities on the same ( different ) sublattices always yield negative ( positive ) values .
another way of understanding the energetic preference for one of the sublattices is to recall that graphene electrons are multiply scattered by the impurities .
these interference effects between scatterers are responsible for minimizing the total energy . in other words ,
impurities become aware of their mutual presence through an indirect interaction between them that is mediated by the conduction electrons of graphene .
this is somewhat similar to the rkky interaction that exists between localized magnetic moments embedded in a medium with conduction electrons .
in fact , the theoretical framework presented here is analogous to the rkky theory in graphene , the main difference being that there are no magnetic moments involved in this case .
for instance , it can be shown that when the integrand in eq.([c12 ] ) is expanded to lowest order in powers of @xmath12 the expression for @xmath22 acquires exactly the same functional form as the rkky coupling in graphene .
we now consider a dilute concentration of @xmath37 substitutional impurities randomly distributed in a graphene sheet .
the perturbation @xmath15 will now consist of many additive terms of the form used to derive eq.([lloyd-1imp ] ) .
following similar steps the total change in energy can be approximated as @xmath38 where the sum runs over all @xmath39 pairwise interactions in the system and terms beyond two - body interaction are neglected .
it should be noted that despite the small magnitude of @xmath22 , which is typically @xmath40 to @xmath41 at the average nearest neighbour separation , the magnitude of the sum over all pairwise interactions , @xmath42 , grows quadratically with the number of impurities in the system .
this leads to a sizable correction to the energy balance for a large number of impurities because of the commensurability effect discussed previously .
for the single and double impurity cases the fermi energy is not modified , however for low concentrations , @xmath43 , the fermi energy will shift as @xmath44 and the pairwise impurity interaction will be oscillatory , resembling fig .
[ fig1]b and fig .
[ fig : ldos ] . to demonstrate that asymmetric doping is indeed a feature of the multiple scattering effects between pairs of impurities
, we shall consider the energy balance in two extreme limits - fully symmetric and fully asymmetric . to do this
we need to consider the concentration of impurities on each sublattice , a and b , given by @xmath45 . rather than considering a sum over all sites , we instead define a probability density as the total pairwise interaction energy with a single impurity , on site 1 , from all other impurities
. this density , which depends on the sublattice occupied by the impurity at site 1 , can be calculated as @xmath46 for impurities on sublattice @xmath14 and @xmath47 respectively , where @xmath48 is the distance from a site on the @xmath49 sublattice to the nearest site on the @xmath20 sublattice . by symmetry
we can write @xmath50 and @xmath51 . the total pairwise interaction energy in the system , @xmath42 ,
can now be found by summing over all impurities @xmath52 where @xmath53 denotes the number of impurities on the @xmath14 sublattice ( and similarly for @xmath54 and @xmath47 ) in a system with radius @xmath55 , and the half prevents double - counting of interactions . to find the impurity segregation that minimizes the system energy we let @xmath56 and @xmath57 , where @xmath58 denotes the fraction of impurities occupying the @xmath14 sublattice .
we can now write the total pairwise interaction energy of the system per unit area as @xmath59 provided @xmath60 , this function has minima at @xmath61 and @xmath62 ( fig .
[ fig2 ] solid black line ) meaning the most energetically favourable configuration is the fully asymmetric one .
this suggests the impurity sublattice segregation in graphene is indeed a feature of the multiple scattering effects between pairs of impurities , captured in the quantities @xmath63 and @xmath64 . the dashed red line in fig .
[ fig2 ] shows the situation where @xmath65 . here
, the total pairwise interaction energy has a minimum at @xmath66 and the system does not favour any sublattice asymmetry .
when @xmath67 a critical dopant concentration , @xmath68 , is reached , beyond which one would not expect to see sublattice segregation . to illustrate the magnitude of @xmath69 we consider its value at some typical concentration , and also the difference between its maxima and minima and the size of this difference . as @xmath70 is an energy per unit area , it is useful to consider the difference between the energies of the sublattice segregated ( @xmath61 or @xmath62 ) and sublattice unsegregated ( @xmath66 ) states over the typical area occupied by a single impurity .
for example , a concentration @xmath71 and parameterisation @xmath72 yields a difference in energy of @xmath73 . a numerical calculation of @xmath63 and @xmath64 as a function of @xmath12 and @xmath43 gives us an insight into the types of impurities and concentrations that lead to sublattice segregation .
[ fig3 ] shows a plot of critical concentration , @xmath68 , versus onsite energy , @xmath12 .
the onsite energy for nitrogen takes different values depending on the method of calculation .
the two most commonly used are @xmath74 , calculated through a self - consistent tight binding model @xcite , and @xmath75 , through fitting of tight binding ldos profiles to dft results @xcite .
we see that for @xmath74 the sublattice segregation should occur up until concentrations of @xmath76 , as demonstrated in fig.[fig3 ] , well above the experimentally reported values of @xmath77 @xcite .
however , in the case of @xmath75 the critical concentration falls to @xmath78 , falling only slightly below the experimentally reported values . regarding the robustness of our calcuations , it is worth analysing how our results depend on the choice of the tb parameter @xmath12 used to represent the nitrogen impurity . according to fig.[fig3 ] , @xmath68 does depend on @xmath12 .
therefore , we may conclude that the exact value of the critical concentration above which the sublattice asymmetry will no longer exist does depend on the choice of impurity parametrization .
however , fig.[fig3 ] also points to the existence of a critical concentration for all values of @xmath12 , which indicates that the physical mechanism that we propose to explain the appearance of sublattice asymmetry in graphene also gives rise to a maximum concentration for which the effect is observable .
consequently , our results point to a simple way in which the physical mechanism behind the sublattice asymmetry can be experimentally tested . in other words , one may conclude that the electron mediated interaction between impurities is responsible for the observed sublattice asymmetry if such an effect disappears when the nitrogen concentration is increased .
we emphasize that the existence of a critical concentration is robust to the values of the tb parameters used to represent the electronic structure of graphene and of its substitutional dopants , however its value is indeed dependent on the parameterisation chosen .
furthermore , the finite size of the experimental domains and their clearly defined edges are not handled in our model but can be explained by non - uniformities in the substrate structure , for example step edges , leading to a disruption in the inter - impurity interactions .
our work here opens up the possibility of extending this method to other impurities , though this may prove more complicated .
nitrogen , which sits next to carbon in the periodic table , is an especially simple dopant to work with in graphene , bonding similarly to carbon , and having only a single extra electron .
it would be natural to extend this to boron , which sits the other side of carbon , except that it is known to interact strongly with cu substrates , destroying the effect @xcite .
it is clear from fig .
[ fig3 ] that the critical concentration for nitrogen depends on the parameterisation @xmath12 , which comes from the perturbing potential @xmath15 that models the impurity . by extending this potential to include more terms e.g. hopping and nearest neighbour perturbations
, we can model other impurities , where the exact values of the parameters can be calculated self - consistently @xcite .
in summary , we have shown that the ldos oscillations that arise as a result of the presence of substitutional impurities are responsible for breaking the sublattice symmetry displayed by pristine graphene .
when considered over a large range of distances they are averaged out and are often disregarded for cases of randomly dispersed impurities . however , in graphene such oscillations have either the same periodicity of the lattice or are very close to it , leading to a net result that affects the behaviour of several physical quantities in doped graphene .
one such quantity is the indirect interaction energy that exists between impurities embedded in a conduction - electron - filled medium . in graphene
this indirect interaction yields a finite contribution , which in turn drives the doping asymmetry .
our model predicts the existence of a critical concentration for nitrogen dopants beyond which the sublattice asymmetry should no longer be visible . while the exact value for such a concentration is not accurately predicted by our simple model
, it nevertheless suggests that concentrations above 1% are already above this critical limit .
most importantly , from a qualitative point of view , the mechanism we propose to explain the sublattice asymmetry of substitutional dopants in graphene leads to a simple consequence that can be easily tested experimentally .
should the asymmetry cease to exist as dopant concentrations are increased , we may conclude that the doping asymmetry arises due to fluctuation of the ldos that break the natural symmetry that otherwise exists between the two sublattices of graphene .
the authors acknowledge financial support from the programme for research in third level institutions ( prtli ) , science foundation ireland ( grant no .
sfi 11/ rfp.1/mtr/3083 ) , the center for nanostructured graphene ( cng ) which is sponsored by the danish national research foundation , project no .
dnrf58 , and capes ( grant no .
10144 - 12 - 9 ) , fapern and inct of space studies .
a. cresti , f. ortmann , t. louvet , d. van tuan and s. roche . broken symmetries , zero - energy modes , and quantum transport in disordered graphene : from supermetallic to insulating regimes , phys .
lett . * 110 * , 196601 ( 2013 ) , doi : 10.1103/physrevlett.110.196601 .
h. park , a. wadehra , j. w. wilkins , and a. h. castro neto .
spin - polarized electronic current induced by sublattice engineering of graphene sheets with boron / nitrogen , phys .
b * 87 * , 085441 ( 2013 ) , doi : 10.1103/physrevb.87.085441 .
p. rakyta , a. kormanyos and j. cserti .
sublattice asymmetry and spin - orbit interaction induced out - of - plane spin polarization of photoelectrons , phys .
b * 83 * , 155439 ( 2011 ) , doi : 10.1103/physrevb.83.155439 .
y. lin , k. a. jenkins , a. valdes - garcia , j. p. small , d. b. farmer and p. avouris .
operation of graphene transistors at gigahertz frequencies , nanoletters , * 9 * , 422 - 426 ( 2009 ) , doi : 10.1021/nl803316h .
s. v. morozov , k. s. novoselov , m. i. katsnelson , f. schedin , d. c. elias , j. a. jaszczak , and a. k. geim .
giant intrinsic carrier mobilities in graphene and its bilayer , phys .
lett . , * 100 * , 016602 ( 2008 ) , doi : 10.1103/physrevlett.100.016602 .
l. zhao , m. levendorf , s. goncher , t. schiros , l. palova , a. zabet - khosousi et al , local atomic and electronic structure of boron chemical doping in monolayer graphene , nanoletters * 13 * , 4659 - 4665 ( 2013 ) , doi : 10.1021/nl401781d .
r. lv , q. li , a. r. botello - mndez , t. hayashi , b. wang , a. berkdemir , q. hao_et al .
nitrogen - doped graphene : beyond single substitution and enhanced molecular sensing , scientific reports * 2 * , 586 ( 2012 ) , doi:10.1038/srep00586 .
a. zabet - khosousi , l. zhao , l. palova , m. s. hybertsen , d. r. reichman , a. n. pasupathy , and g. w. flynn .
segregation of sublattice domains in nitrogen - doped graphene , j. am .
soc . , ( 2014 ) , 136 ( 4 ) , pp 13911397 doi : 10.1021/ja408463 g .
lambin , h. amara , f. ducastelle and l. henrard .
long - range interactions between substitutional nitrogen dopants in graphene : electronic properties calculations , phys .
b * 86 * , 045448 ( 2012 ) , doi : 10.1103/physrevb.86.045448 . v. v. cheianov and v. i. falko .
friedel oscillations , impurity scattering , and temperature dependence of resistivity in graphene , phys .
lett . * 97 * , 226801 ( 2006 ) , doi : 10.1103/physrevlett.97.226801 . | motivated by the recently observed sublattice asymmetry of substitutional nitrogen impurities in cvd grown graphene , we show , in a mathematically transparent manner , that oscillations in the local density of states driven by the presence of substitutional impurities are responsible for breaking the sublattice symmetry .
while these oscillations are normally averaged out in the case of randomly dispersed impurities , in graphene they have either the same , or very nearly the same , periodicity as the lattice . as a result ,
the total interaction energy of randomly distributed impurities embedded in the conduction - electron - filled medium does not vanish and is lowered when their configuration is sublattice - asymmetric .
we also identify the presence of a critical concentration of nitrogen above which one should expect the sublattice asymmetry to disappear .
this feature is not particular to nitrogen dopants , but should be present in other impurities . |
the central unknown quantity required for reliable calculations of weak decay amplitudes are the hadronic matrix elements . in this respect ,
the semileptonic @xmath0 decay is an interesting process : pseudoscalar mesons are well - understood as goldstone bosons , only the vector part of the weak interaction is involved , and there are experimental data available which allows one to judge various theoretical ( model ) calculations .
these calculations generally fall into two different classes : effective theories using meson degrees of freedom , e.g. chiral perturbation theory @xcite , and models employing quark and gluon degrees of freedom , e.g. constituent quark models @xcite and the dyson schwinger approach @xcite .
in particular , the light - front approach has been a popular framework for a hamiltonian approach to analyze exclusive hadronic processes @xcite , and the pseudoscalar - to - pseudoscalar weak transition form factors have been studied within a light - front constituent quark model in ref .
one complication of the light - front formalism is that in the analysis of timelike exclusive processes one needs to take into account light - front nonvalence contributions , which are absent in a manifestly poincar invariant approach .
recently , an effective treatment has been presented @xcite to incorporate such contributions , and a systematic program has been laid out @xcite to take into account the nonvalence contributions
. however , the systematic program explicitly requires all higher fock - state components while there has been relatively little progress in computing the basic wave - functions of hadrons from first principles .
furthermore , the ( nonperturbative ) dressing of the quark - w vertex and the dynamical chiral symmetry breaking are yet underdeveloped aspects in this approach .
in particular the latter aspect needs to be seriously considered for processes involving pions and kaons , which are the pseudo - goldstone bosons associated with dynamical chiral symmetry breaking @xcite . in this work we use the set of dyson
schwinger equations [ dses ] to calculate the @xmath0 transition form factors .
the dses provide us with a manifestly covariant approach which is consistent with dynamical chiral symmetry breaking @xcite , electromagnetic current conservation @xcite , and quark and gluon confinement @xcite . for recent reviews on the dse and its application to hadron physics ,
see refs .
our calculation of the @xmath0 transition form factors is analogous to recent calculations of the pion and kaon electromagnetic form factors in impulse approximation @xcite .
here we demonstrate that in the same framework , with a consistent dressing of the quark - w vertex , we can also describe the semileptonic decay of both the neutral and charged kaons , without any readjustment of the model parameters . in sec .
ii we review the formulation that underlies a description of the meson form factors within a modeling of qcd through the dses , and discuss the details of the model . in sec .
iii we present our numerical results for the form factors and decay widths .
a discussion of our results is given in sec .
the matrix elements describing kaon semileptonic decays are @xmath1 for the charged kaon , and for the neutral kaon @xmath2 where @xmath3 and @xmath4 , with @xmath5 for on - shell pions and kaons for timelike vectors , @xmath6 , and @xmath7 . ] .
alternatively , we can decompose @xmath8 into its transverse and longitudinal components @xmath9 with @xmath10 the transverse projection operator @xmath11 in the isospin - symmetric limit , which we employ here , the form factors @xmath12 and @xmath13 are the same for the @xmath14 and the @xmath15 . in impulse approximation ,
these matrix elements are given by @xmath16 \ ; , \label{impulse}\end{aligned}\ ] ] where @xmath17 is the dressed quark propagator with flavor index @xmath18 , @xmath19 the pion ( kaon ) bethe
salpeter amplitude [ bsa ] , and @xmath20 the dressed @xmath21 vertex , each satisfying their own dse .
note that the coupling constant and the ckm matrix element @xmath22 are removed from the definition of the quark - w vertex .
the dse for the renormalized quark propagator in euclidean space is @xmath23 where @xmath24 is the dressed - gluon propagator , @xmath25 the dressed - quark - gluon vertex , and @xmath26 .
the most general solution of eq .
( [ gendse ] ) has the form and is renormalized at spacelike @xmath27 according to and with @xmath28 the current quark mass .
the notation @xmath29 represents a translationally invariant regularization of the integral with @xmath30 the regularization mass - scale ; at the end of all calculations this regularization scale can be removed .
mesons can be studied by solving the homogeneous bethe
salpeter equation [ bse ] for @xmath31 bound states , with @xmath32 and @xmath33 flavor indices , @xmath34 where @xmath35 and @xmath36 are the outgoing and incoming quark momenta respectively , and similarly for @xmath37 .
the kernel @xmath38 is the renormalized , amputated @xmath39 scattering kernel that is irreducible with respect to a pair of @xmath39 lines .
this equation has solutions at discrete values of @xmath40 , where @xmath41 is the meson mass .
together with the canonical normalization condition for @xmath39 bound states , it completely determines @xmath42 , the bound state bsa .
the different types of mesons , such as ( pseudo-)scalar , ( axial-)vector , and tensor mesons , are characterized by different dirac structures .
the most general decomposition for pseudoscalar bound states is @xcite @xmath43\,,\end{aligned}\ ] ] where the invariant amplitudes @xmath44 , @xmath45 , @xmath46 and @xmath47 are lorentz scalar functions of @xmath48 and @xmath49 .
note that these amplitudes explicitly depend on the momentum partitioning parameter @xmath50 .
however , so long as poincar invariance is respected , the resulting physical observables are independent of this parameter @xcite . in order to describe weak ( and electromagnetic ) form factors one also needs the dressed @xmath51 ( and @xmath52 ) vertices .
these vertices satisfy an inhomogeneous bse : e.g. the @xmath53 vertex satisfies @xmath54 to solve the bse , we use a ladder truncation , with an effective quark - antiquark interaction that reduces to the perturbative running coupling at large momenta @xcite @xmath55 where @xmath56 is the free gluon propagator in landau gauge .
the corresponding rainbow truncation of the quark dse , eq .
( [ gendse ] ) , is given by together with .
this combination of rainbow and ladder truncation preserves both the vector and axial - vector ward
takahashi identities .
this ensures that pions are ( almost ) massless goldstone bosons associated with dynamical chiral symmetry breaking @xcite ; in combination with impulse approximation for meson form factors , eq .
( [ impulse ] ) , it also ensures current conservation .
furthermore , this truncation was found to be particularly suitable for the flavor octet pseudoscalar and vector mesons since the next - order contributions in a quark - gluon skeleton graph expansion , have a significant amount of cancellation between repulsive and attractive corrections @xcite .
the model is completely specified once a form is chosen for the `` effective coupling '' @xmath57 .
we employ the ansatz @xcite @xmath58 } \;,\end{aligned}\ ] ] with and .
the ultraviolet behavior is chosen to be that of the qcd running coupling @xmath59 ; the ladder - rainbow truncation then generates the correct perturbative qcd structure of the dse - bse system of equations .
the first term implements the strong infrared enhancement in the region phenomenologically required @xcite to produce a realistic value for the chiral condensate .
we use , , , , and a renormalization scale which is well into the perturbative domain @xcite .
the remaining parameters , and along with the quark masses , are fitted to give a good description of the chiral condensate , @xmath60 and @xmath61 . within this model
, the quark propagator reduces to the perturbative propagator in the ultraviolet region .
however , in the infrared region both the wave function renormalization @xmath62 and the dynamical mass function @xmath63 deviate significantly from the perturbative behavior , due to chiral symmetry breaking .
it is interesting to note that the typical results obtained using the rainbow dse , a significant enhancement of @xmath64 below @xmath65 and also an enhancement of @xmath66 , have recently been confirmed by lattice simulations @xcite ; with the present ladder dse model , the functions @xmath67 and @xmath68 are in semiquantitative agreement with the forms obtained in lattice simulations @xcite .
the vector meson masses and electroweak decay constants one obtains in this model are in good agreement with experiments @xcite . without any readjustment of the parameters ,
one gets remarkable agreement with the most recent jlab data for @xmath69 @xcite , and also for other electromagnetic charge radii and form factors @xcite .
the strong decays of the vector mesons into a pair of pseudoscalar mesons are also well - described within this model @xcite . here
we apply the same approach to the kaon semileptonic decay .
to calculate the form factors @xmath70 , we start by solving the quark dse in rainbow approximation , and subsequently use its solution to solve the homogeneous bse , eq .
( [ hombse ] ) , for the pseudoscalar bound states , and the inhomogeneous bse , eq .
( [ verbse ] ) , for the quark - w vertex , using the ladder truncation , eq .
( [ ourbseansatz ] ) . with these bs amplitudes and quark propagators
, we then calculate the form factors in impulse approximation using eq .
( [ impulse ] ) . in the su(3 ) flavor limit , @xmath71 , the pion electromagnetic form factor , and @xmath72 .
we therefore expect @xmath73 to be of order one for small @xmath74 . on the other hand , @xmath75 , which is a measure of the constituent mass ratio @xmath76 @xcite ,
is expected to be significantly smaller .
this expectation is indeed confirmed by our calculations , see fig .
[ fig1 ] . in the physical region , @xmath77 ,
our results for @xmath78 are almost linear .
the experimental data are often analyzed in terms of @xmath73 and @xmath79 , using a linear parametrisation @xmath80 such a linear parametrisation for both @xmath73 and @xmath79 implies that @xmath75 is independent of @xmath74 , since @xmath79 is related to @xmath81 via @xmath82 to within a few percent , @xmath75 is indeed almost constant in the physical region , @xmath83 , as can be seen from fig .
our results are summarized in table [ table ] , where we have included a slope parameter @xmath84 for @xmath85 as well .
our results agree quite well with the available data for @xmath86 , see fig .
also the results for other observables compare reasonably well with the experimental data , given the error bars , and the differences between the form factor parameters extracted from the charged and the neutral kaon decay , see table [ table ] . the partial decay width for @xmath87 and @xmath88
is obtained by integrating the decay rate @xmath89 \;,\end{aligned}\ ] ] where @xmath90 and @xmath91 is the lepton mass .
again , we find good agreement with the data , see table [ table ] . for comparison , we have also included the results from some other model calculations @xcite and from 1-loop chiral perturbation theory @xcite .
our results compare quite well with chiral perturbation theory , not only for the slope parameters , but also for the curvature of the form factors .
a prediction from current algebra is the value of @xmath13 at the callan
treiman point , @xmath92 , namely @xmath93@xcite .
the reason we get @xmath94 instead of @xmath95 is related to our value of @xmath96 : we obtain @xmath97 in our model @xcite , compared to the experimental value @xmath98 . because our approach satisfies constraints coming from current conservation and the axial - vector ward
takahashi identity , it is not surprising that we indeed agree with these predictions from chiral perturbation theory . outside the physical region , @xmath78 and @xmath13 are clearly not linear , as shown in fig .
[ fig3 ] . in the region we considered
, our result for @xmath86 is very well approximated by a monopole , with monopole mass @xmath99 .
the reason for this monopole behavior is easy to understand : the vertex function @xmath20 has a resonance peak in its transverse vector components at the @xmath100 mass , because the homogeneous version of eq .
( [ verbse ] ) has a solution at @xmath101 , corresponding to the @xmath100 vector meson .
the longitudinal part of the quark - w vertex has a resonance peak at the location of the @xmath102 scalar bound state .
we tentatively identify this with the @xmath103 meson ; however , in contrast to vector states , the use of ladder approximation is expected to be inadequate for scalars @xcite . near these bound states
, the vertex behaves like @xmath104 where @xmath105 and @xmath106 are the bound state bsas , and @xmath107 and @xmath108 the residues of the vertex at these poles @xcite . in our model
we have actual poles in the quark - w vertex , rather than resonance peaks , because we have not included meson loop corrections in our kernel , and therefore we do not generate a width for the @xmath100 nor the @xmath109 . in the spacelike region and at small timelike @xmath110
we expect these corrections to be small .
however , they will modify our results significantly close to the resonance peak .
as can be seen from eq .
( [ impulse ] ) , these poles in @xmath20 lead a similar behavior of @xmath111 near these poles . from eq .
( [ fzerodef ] ) it is clear that the vector meson pole leads to a pole in @xmath86 , but not in @xmath13 , whereas the scalar meson induces a pole in @xmath13 but not in @xmath86 @xmath112 note that @xmath85 will in general be sensitive to both the scalar and the vector meson bound state .
the present model has a vector bound state at @xmath113 ( cf .
the experimental mass @xmath114 ) @xcite .
thus the monopole behavior of the form factor @xmath86 in fig .
[ fig3 ] simply reflects the existence of this pole .
a similar behavior was found in the pion and kaon electromagnetic form factors @xcite .
the physical @xmath86 can be expected to rise more slowly with @xmath74 than our present calculations due to meson loop corrections , and will develop a nonzero imaginary part above the threshold for intermediate @xmath115 states . the scalar form factor @xmath13 does also exhibit the expected pole behavior , and we are able to identify the scalar and the vector poles separately , see fig . [ fig3 ] .
the lowest @xmath102 scalar bound state in this model has a mass of @xmath116 , which is rather low compared to the @xmath117 with a mass @xmath118 ; on the other hand , it could be an indication that there exists a light scalar resonance @xmath119 , as has been speculated @xcite .
something similar happens in in the @xmath120 quark sector : in rainbow - ladder truncation , one typically finds a scalar @xmath120 bound state around @xmath121 to @xmath122 @xcite , just below the @xmath123 mass .
this bound state could either correspond to a broad @xmath124 , or to the @xmath79 and/or @xmath125 , in which case the calculated mass is about 30 to 40% too low .
it is known that the leading perturbative corrections to the ladder kernel cancel to a large extent in both the pseudoscalar and vector channel , but not in the scalar channel ; we therefore expect in the scalar channel significant repulsive corrections to the ladder kernel which could increase the mass of the scalar bound state @xcite .
we also expect meson loop corrections to be more important in the scalar channel than in the vector channel .
therefore , our calculation for @xmath13 should not be trusted quantitatively in the timelike region beyond the callan treiman point , @xmath126
. in the spacelike asymptotic region @xmath73 and @xmath79 seem to behave differently : a monopole fit does not work for @xmath79 . in order to fit @xmath79 , we either need to add a nonzero constant , or use at least two monopoles . simple power counting indicates that @xmath8 scales like @xmath127 ; inserting this behavior in eqs .
( [ fzero ] ) and ( [ fplus ] ) implies that @xmath86 falls like @xmath128 , but @xmath13 goes to a constant ( up to logarithmic corrections ) at large spacelike @xmath74 ( remember that @xmath129 scales like @xmath130 in the asymptotic region ) .
this behavior of @xmath73 agrees with the pqcd predictions for the electromagnetic pion and kaon form factor @xcite ; the behavior of @xmath79 can be understood if one realizes that the combination @xmath131 does fall like @xmath128 if @xmath13 goes to a constant .
using the rainbow - ladder truncation of the set of dses with a model for the effective quark - antiquark interaction that has been fitted to the chiral condensate and @xmath132 , we study the @xmath0 decay in impulse approximation .
our results , both for the form factors and for the decay width , are in good agreement with experimental data , without any readjustment of the parameters ; they also compare quite well with chiral perturbation theory .
note however that in chiral perturbation theory the pion charge radius is used as input , in order to fix a low - energy constant , which is important for these form factors as well , in particular for @xmath133 . in our calculation , the only model parameters are in the infrared behavior of the effective quark - antiquark interaction , which were fitted to the chiral condensate and the pion decay constant @xcite ; the pion charge radius follows from a calculation similar to the one presented here for the @xmath0 decay @xcite .
our approach is based on previous work by kalinovsky _
et al . _
@xcite , and the results are quite similar .
however , an important difference with this earlier study is that we dress the quark - w vertex by solving the inhomogeneous bse for this vertex .
the main advantage of doing so is that we thus automatically include effects coming from intermediate vector and scalar mesons .
another difference is that here we use actual solutions of a dse for the propagators and bs amplitudes in eq .
( [ impulse ] ) , rather than phenomenological parametrisations , which reduces the number of parameters in the calculation .
we have demonstrated explicitly that both @xmath86 and @xmath13 exhibit resonance peaks in the timelike region due to the existence of vector and scalar bound states respectively .
the effects of meson loops are not included in our present calculation , which is why we find a pole behavior rather than a resonance peak . in ref . @xcite it was already demonstrated that non - analytic effects from such loops could contribute significantly to the behavior of the form factors , not only beyond the threshold for @xmath134 production , but also in the physical region .
we hope to be able to incorporate these effects in future work .
our results are also similar to those obtained recently by ji and choi @xcite in a light - front calculation , at least for @xmath86 , which dominates this decay , even though details of the calculation are quite different . in light - front calculations of timelike processes
one has to include particle - number - nonconserving fock states to recover the lorentz covariance , which complicates the calculation @xcite .
since our approach is manifestly covariant , such contributions are automatically included in our impulse approximation .
another difference is that we use momentum - dependent quark self - energies , consistent with dynamical chiral symmetry breaking , whereas in ref .
@xcite constituent quarks with fixed masses are used .
recently , the effects of a running quark mass ( instead of a fixed constituent mass ) have been explored in a light - front calculation of the pion form factor @xcite .
it would be interesting to see the effect of such a running mass on the @xmath0 form factors , in particular on @xmath85 , which in general appears to be quite sensitive to details of the calculation .
we would like to thank p.c .
tandy , c.d .
roberts , k. maltman , s.r .
cotanch , and h. choi for stimulating discussions and useful suggestions .
this work was supported by the us doe under grants no .
de - fg02 - 96er40947 and de - fg02 - 97er41048 , and benefited from the resources of the national energy research scientific computing center .
h. leutwyler and m. roos , z. phys .
* c25 * , 91 ( 1984 ) . j. gasser and h. leutwyler , nucl
. phys . * b250 * , 517 ( 1985 ) . n. isgur , phys . rev .
d * 12 * , 3666 ( 1975 ) .
h. choi and c. ji , phys .
d * 59 * , 034001 ( 1999 ) [ hep - ph/9807500 ] .
h. choi and c. ji , phys .
* b460 * , 461 ( 1999 ) [ hep - ph/9903496 ] .
a. afanasev and w. w. buck , phys .
d * 55 * , 4380 ( 1997 ) [ hep - ph/9606296 ] .
y. kalinovsky , k. l. mitchell and c. d. roberts , phys .
b399 * , 22 ( 1997 ) [ nucl - th/9610047 ] . m. a. ivanov , y. l. kalinovsky and c. d. roberts , phys .
rev . * d60 * , 034018 ( 1999 ) [ nucl - th/9812063 ]
. s. j. brodsky , h. pauli and s. s. pinsky , phys .
rept . * 301 * , 299 ( 1998 ) [ hep - ph/9705477 ] .
c. ji and h. choi , hep - ph/0009281 .
s. j. brodsky and d. s. hwang , nucl .
* b543 * , 239 ( 1999 ) [ hep - ph/9806358 ] .
p. maris , c. d. roberts and p. c. tandy , phys .
* b420 * , 267 ( 1998 ) [ nucl - th/9707003 ] .
maris and c. d. roberts , phys .
rev . * c56 * , 3369 ( 1997 ) [ nucl - th/9708029 ] .
see e.g. k. higashijima , phys .
d * 29 * , 1228 ( 1984 ) ; v.a .
miransky , phys .
b * 165 * , 401 ( 1985 ) ; d. atkinson and p.w .
johnson , phys .
d * 37 * , 2290 and 2296 ( 1988 ) ; c.d .
roberts and b.h .
mckellar , phys .
d * 41 * , 672 ( 1990 ) .
p. maris and p. c. tandy , phys .
rev . * c61 * , 045202 ( 2000 ) [ nucl - th/9910033 ] ; ibid , phys .
* c62 * , 055204 ( 2000 ) [ nucl - th/0005015 ] .
burden , c.d .
roberts , and a.g .
williams , phys .
b * 285 * , 347 ( 1992 ) ; g. krein , c.d .
roberts , and a.g .
williams , int . j. mod .
phys . a * 7 * , 5607 ( 1992 ) ; p. maris , phys .
d * 52 * , 6087 ( 1995 ) . c. d. roberts and s. m. schmidt , prog . part .
* 45s1 * , 1 ( 2000 ) [ nucl - th/0005064 ] .
r. alkofer and l. von smekal , hep - ph/0007355 .
p. maris and p. c. tandy , phys .
* c60 * , 055214 ( 1999 ) [ nucl - th/9905056 ] .
a. bender , c. d. roberts and l. von smekal , phys .
lett . *
b380 * , 7 ( 1996 ) [ nucl - th/9602012 ] . f. t. hawes , p. maris and c. d. roberts , phys . lett .
* b440 * , 353 ( 1998 ) . j. i. skullerud and a. g. williams , hep - lat/0007028 .
p. maris , nucl - th/0009064 .
j. volmer _ et al .
_ [ the jefferson lab f(pi ) collaboration ] , nucl - ex/0010009 .
p. maris , nucl - th/0008048 .
d. jarecke , p. maris , and p.c .
tandy , in preparation .
a. apostolakis _ et al .
. lett .
* b473 * , 186 ( 2000 ) .
particle data group , c. caso _ et al .
_ , eur .
phys . j. * c3 * , 1 ( 1998 ) .
j. bijnens , g. colangelo , g. ecker and j. gasser , hep - ph/9411311 .
c.g . callan and s.b .
treiman , phys .
* 16 * , 153 ( 1966 ) .
a. bender , c. d. roberts and l. von smekal , phys .
b * 380 * , 7 ( 1996 ) [ nucl - th/9602012 ] ; c. d. roberts , in _ quark confinement and the hadron spectrum ii _ , edited by n. brambilla and g.m .
propseri ( world scientific , singapore , 1997 ) , pp .
224 - 230 [ nucl - th/9609039 ] .
d. black , a. h. fariborz and j. schechter , hep - ph/0008246 ; j. a. oller and e. oset , phys .
d * 60 * , 074023 ( 1999 ) [ hep - ph/9809337 ] .
p. jain and h. j. munczek , phys .
d * 48 * , 5403 ( 1993 ) [ hep - ph/9307221 ] . c. j. burden , l. qian , c. d. roberts , p. c.
tandy and m. j. thomson , phys .
c * 55 * , 2649 ( 1997 ) [ nucl - th/9605027 ] .
p. maris , c. d. roberts , s. m. schmidt and p. c. tandy , phys .
c * 63 * , 025202 ( 2001 ) [ nucl - th/0001064 ] .
g. p. lepage and s. j. brodsky , phys . lett .
b * 87 * , 359 ( 1979 ) .
l. s. kisslinger , h. choi and c. ji , hep - ph/0101053 .
.our results , compared to data @xcite where available and some other calculations .
the partial decay width is in @xmath135 ; because @xmath136 in the exact isospin limit , the experimental value for the @xmath14 partial decay width should be multiplied by a factor of two , in order to compare it with our calculation and with the @xmath15 partial decay width . [
cols="<,^,^,^,^,^,^ " , ] | the rainbow truncation of the quark dyson
schwinger equation is combined with the ladder bethe
salpeter equation for the meson bound state amplitudes and the dressed quark - w vertex in a manifestly covariant calculation of the @xmath0 transition form factors and decay width in impulse approximation . with model gluon parameters previously fixed by the chiral condensate , the pion mass and decay constant , and the kaon mass , our results for the @xmath0 form factors and the kaon semileptonic decay width are in good agreement with the experimental data . |
metamaterials consisting arrays of artificial particles usually work in the frequency regime in which the free - space wavelength is larger than the inter - particle distance by a factor around 10 .
their electromagnetic properties can therefore be understood by treating a metamaterial as a homogeneous medium . more often , a local effective medium of permittivity ( @xmath0 ) and permeability ( @xmath1 ) is assumed to be valid .
there are several approaches in obtaining the local effective medium of a metamaterial .
one approach is the s - parameter retrieval method . @xcite it obtains the material parameters by comparing the transmission and the reflection coefficients of a slab of metamaterial and the corresponding homogeneous slab .
another approach relies on a homogenization theory which averages the microscopic field by line and area integrals . @xcite both approaches can cater for complex geometry and microstructure of the metamaterial and they work very well at frequencies far away from the resonance gap where spatial dispersion becomes important so that higher order eigenmodes ( including the non - propagating ones ) can be excited inside the metamaterial .
@xcite in this work , we will take another approach to deal with the spatial dispersion at a frequency near the resonance gap . instead of picking or assuming a dominant eigenmode within the metamaterial
, we will find the non - local effective medium directly for an arbitrary bloch wave vector @xmath2 .
for instance , the permittivity @xmath3 now depends on both the angular frequency @xmath4 and the wave vector .
apart from working near the resonance gap , some metamaterials like the wire medium , @xcite the three - dimensional wire mesh @xcite and the structured metal surface @xcite also indicate that spatial dispersion ( the non - local effective medium ) can become important in order to have an accurate description of the metamaterial . in this paper
, we will establish a numerical scheme in getting the non - local effective medium and some effects of spatial dispersion on metamaterial will be examined .
section [ sect2 ] outlines the numerical scheme in obtaining the non - local effective medium . we will employ a dipolar model for simplicity .
it has the advantage that analytic formulas can be obtained in some cases .
moreover , all the material parameters , two tensors ( @xmath5 and @xmath6 ) together with two pseudo tensors ( @xmath7 and @xmath8 ) can be obtained easily in the whole frequency regime .
all the four constitutive tensors will be extracted as ref . and pointed out that a metamaterial should be described as a bianisotropic medium if the artificial atoms suffer from cross - coupling ( electric field generates magnetic dipole and/or magnetic field generates electric dipole ) . in section [ sect3 ] , the local effective medium of a double - lattice metamaterial will be first examined .
one sublattice consists of electric artificial atoms and another sublattice consists of magnetic artificial atoms .
the condition that the local effective medium becoming free of bianisotropy will be discussed . in the last section , the non - local effective medium of the double - lattice metamaterial will be examined .
the relationship between the non - local effective medium to the corresponding local one will be discussed . on the other hand ,
a lorentz - type model is suggested for fitting the frequency dispersion of the non - local effective medium .
some effects of the spatial dispersion , including the structure - induced bianisotropy and the cross - coupled resonance , will also be discussed .
here , we give an outline of the numerical scheme in obtaining the non - local effective medium using a dipolar model . the artificial particles ( split - rings , short wires , or electric atoms , @xcite etc . )
are placed in an infinite lattice of lattice vectors @xmath9 in vacuum ( of wavenumber @xmath10 , wave speed @xmath11 , and intrinsic impedance @xmath12 . in each unit cell labeled by @xmath9 , there are @xmath13 particles ( indexed by @xmath14 at positions @xmath15 . in the dipolar limit , each artificial particle has an electric dipole moment @xmath16 and a magnetic dipole moment @xmath17 and they are governed by the following equation ( with time dependence factor @xmath18 and s. i. units ) : @xmath19 = { \alpha } _ j \cdot \left [ { { \begin{array}{*{20}c } { { \rm { \bf e}}_{\mbox{local } } \left ( { { \rm { \bf r } } + { \rm { \bf \tau } } _ j } \right ) / \eta _ 0 } \hfill \\ { { \rm { \bf h}}_{\mbox{local } } \left ( { { \rm { \bf r } } + { \rm { \bf \tau } } _ j } \right ) } \hfill \\
\end{array } } } \right],\ ] ] where @xmath20 and @xmath21 are the local electric and local magnetic fields at the center of the @xmath22-th particle .
these local fields include the fields radiated by the external sources and the scattered fields from all the other particles .
the 6-by-6 matrix @xmath23 represents the polarizability of the @xmath22-th particle . in this model
, every artificial particle is replaced by an equivalent dipolar point particle . in order to extract the non - local effective medium of this crystal at an arbitrary wave vector @xmath24 , we drive the crystal by the external fields in the form @xmath25 these external fields are in fact generated by an electric and a magnetic polarization of the same spatial dependence .
the crystal responds to it by generating dipole moments in the form @xmath26 = \left [ { { \begin{array}{*{20}c } { c_0 { \rm { \bf p}}_j } \hfill \\ { { \rm { \bf m}}_j } \hfill \\ \end{array } } } \right]e^{i{\rm { \bf q } } \cdot { \rm { \bf r}}},\ ] ] which satisfies the multiple - scattering equation : @xmath27e^ { - i{\rm { \bf q } } \cdot { \rm { \bf \tau } } _ { j ' } } = \left [ { { \begin{array}{*{20}c } { { \rm { \bf e}}_{\mbox{ext } } / \eta _ 0 } \hfill \\
{ { \rm { \bf h}}_{\mbox{ext } } } \hfill \\ \end{array } } } \right],\ ] ] with @xmath28 being the volume of a single primitive unit cell .
the @xmath29 transition matrix @xmath30 is given by its inverse : @xmath31 where the lattice green s function @xmath32 is defined by @xmath33 the 6-by-6 matrix @xmath34 which relates a dipole moment to its radiation field , together with its fourier transform @xmath35 are given explicitly in appendix [ appendixa ] .
the lattice green s function can be evaluated by the ewald sum technique . for mathematical convenience
, we also define @xmath36,\ ] ] where @xmath37 is the @xmath38 identity matrix . the tensors @xmath39 and @xmath40 are periodic functions in the reciprocal space .
they can be characterized according to the point group @xcite associated with the lattice structure at a particular wave vector @xmath41 together with the requirement that the matrix transpose ( fixed @xmath22 and @xmath42 ) of @xmath40 satisfies @xmath43 which is already implied from eq .
( [ eq6 ] ) with eq .
( [ eqa2 ] ) .
table [ tab1 ] lists the forms of the tensors @xmath39 and @xmath40 for some common single lattices and double lattices .
the matrices are listed with respect to the cartesian basis vectors @xmath44 where @xmath45 is defined to be the direction along @xmath24 in our convention .
@xmath46 & @xmath47 + & at @xmath48 & @xmath49 & 0 + & along [ 111]/[001 ] & @xmath50 $ ] & @xmath51 $ ] + nacl/ cscl & at @xmath48 & @xmath49 & 0 + & along [ 111]/[001 ] & @xmath52 $ ] & @xmath51 $ ] + & at @xmath48 & @xmath49 & 0 + & along [ 111 ] & @xmath52 $ ] & @xmath51 $ ] + & along [ 001 ] : @xmath53 $ ] , @xmath54 $ ] & @xmath55 $ ] & @xmath51 $ ] + in the next step , the dipole moment distributions numerically solved from eq .
( [ eq4 ] ) are then macroscopically averaged .
the procedure we choose here is to apply a low pass filter such that all the fourier components outside the first brillouin zone are eliminated .
in essence , it means that all the microscopic fields are ensemble averaged by translating the whole crystal with an arbitrary distance while the external fields remain unchanged .
thus , the macroscopic polarizations are written as @xmath56 the macroscopic polarizations can now be formally written as @xmath57 = \left\langle { t } \right\rangle \left ( { \rm { \bf q } } \right)\left [ { { \begin{array}{*{20}c } { { \rm { \bf e}}_{\mbox{ext } } / \eta _ 0 } \hfill \\ { { \rm { \bf h}}_{\mbox{ext } } } \hfill \\ \end{array } } } \right],\ ] ] where the averaged transition matrix is given by @xmath58 eq . ( [ eq10 ] ) summarizes the response of the crystal to the external fields .
then , the macroscopic fields are constructed from the sum of the external fields and the secondary radiation by @xmath59 = { g}_0 \left ( { \rm { \bf q } } \right ) \cdot \left [ { { \begin{array}{*{20}c } { c_0 \left\langle { \rm { \bf p } } \right\rangle } \hfill \\ { \left\langle { \rm { \bf m } } \right\rangle } \hfill \\ \end{array } } } \right ] + \left [ { { \begin{array}{*{20}c } { { \rm { \bf e}}_{\mbox{ext } } / \eta _ 0 } \hfill
\\ { { \rm { \bf h}}_{\mbox{ext } } } \hfill \\ \end{array } } } \right].\ ] ] finally , by combining eq .
( [ eq10 ] ) and eq .
( [ eq12 ] ) , we obtain the constitutive relationship @xmath60 = { \left [ { { \begin{array}{*{20}c } { \mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\leftrightarrow$}}\over { \epsilon } } \left ( { \rm { \bf q } } \right ) - \mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\leftrightarrow$}}\over { i } } } \hfill & { \mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\leftrightarrow$}}\over { \xi } } \left ( { \rm { \bf q } } \right ) } \hfill \\ { \mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\leftrightarrow$}}\over { \zeta } } \left ( { \rm { \bf q } } \right ) } \hfill & { \mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\leftrightarrow$}}\over { \mu } } \left ( { \rm { \bf q } } \right ) - \mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\leftrightarrow$}}\over { i } } } \hfill \\ \end{array } } } \right ] } \cdot \left [ { { \begin{array}{*{20}c } { \left\langle { \rm { \bf e } } \right\rangle / \eta _ 0 } \hfill
\\ { \left\langle { \rm { \bf h } } \right\rangle } \hfill \\ \end{array } } } \right],\ ] ] where the four constitutive tensors are given by @xmath61 ^ { - 1 } = \left\langle { t } \right\rangle ^ { - 1}\left ( { \rm { \bf q } } \right ) + { g}_0 \left ( { \rm { \bf q } } \right).\ ] ] in general , apart from the permittivity tensor @xmath62 and the permeability tensor @xmath63 , we also obtain the off - diagonal tensors @xmath64 and @xmath65 which represent the bianisotropy of the metamaterial .
we will see in the next section that the bianisotopy should not be neglected in considering the non - local effective medium .
all the four tensors are functions of the frequency @xmath66 and the wave vector @xmath24 . for convenience
, we will omit writing @xmath66 inside the parenthesis for the constitutive tensors .
moreover , we will further omit the dependence of @xmath24 if the local limit @xmath67 is taken , i.e. @xmath68 is abbreviated as @xmath69 .
we note that in our scheme of defining the non - local effective medium at a particular @xmath24 , a fixed form of external fields ( eq . ( [ eq2 ] ) ) is used .
there are alternative methods which directly deal with the macroscopic fields without considering the external fields .
in this work , we will concentrate on the double lattice structure which is a common type of microstructures for metamaterials in obtaining a negative refractive index .
the double lattice consists of two types of artificial particles .
one type of the particles ( labeled by @xmath70 has a resonating electric response ( e.g. a short wire ) and another ( labeled by @xmath71 has a resonating magnetic response ( e.g. a split - ring ) .
in particular , we would like to study the importance of the electro - magnetic coupling between the electric and the magnetic particles when we have a crystal structure having both electric and magnetic response .
it is a general feature of metamaterials .
first , we would like to restrict our discussion to the class of metamaterials whose local effective medium is free from bianisotropy .
it can be achieved through a careful design of the artificial particles such that the particles are free from cross - coupling .
@xcite moreover , the lattice structure has to be carefully chosen as well .
in fact , all the double lattice structures listed in table [ tab1 ] have vanishing @xmath72 at @xmath73 . from eq .
( [ eq5 ] ) , eq .
( [ eq11 ] ) and eq .
( [ eq14 ] ) , it means that @xmath74 for simplicity , we further assume that the electric particle can only be electrically polarized along the x - direction with polarizability @xmath75 ( @xmath76 ) , the magnetic particle can only be magnetically polarized along the y - direction with polarizability @xmath77 ( @xmath78 ) and all the polarizabilities along the other directions are assumed to be negligible . in this case , the relevant basis vectors can be reduced to @xmath79 in eq .
( [ eq4 ] ) .
( [ eq10 ] ) in the local limit can then be written as @xmath80 = \left [ { { \begin{array}{*{20}c } { e_{\mbox{ext , x } } / \eta _ 0 } \hfill \\ { h_{\mbox{ext , y } } } \hfill \\ \end{array } } } \right],\ ] ] where @xmath81 , \end{array}\ ] ]
is now written in the @xmath82 basis .
the macroscopic electric polarization @xmath83 is only contributed from the first particle while the macroscopic magnetic polarization @xmath84 is only contributed from the second particle .
therefore , the local effective medium is governed by @xmath85 because of the carefully chosen microstructure , the local effective medium has exactly the same form to a single lattice of one type of particles .
@xcite this is the ideal case that the electric and the magnetic particles can not `` see '' each other . here , @xmath86 is the local limit ( @xmath87 ) of @xmath88 defined in eq .
( [ eq7 ] ) .
it is a function of the normalized frequency @xmath89 for the lattice structure and it can be well approximated by a polynomial at small frequencies .
table [ tab2 ] gives the polynomial expansion of @xmath90 for various lattice structures .
note that for cubic lattices , @xmath91 at the long wavelength limit so that eq .
( [ eq18 ] ) returns to the clausius - mossotti equation .
@xcite sc & @xmath92 + fcc & @xmath93 + bcc & @xmath94 + one popular form of the frequency dispersions of the local effective medium is given by @xmath95 here , eq .
( [ eq19 ] ) is still called the lorentz - model dispersion although the resonance term in permeability is modified so that these frequency dispersion forms give us the correct low frequency behavior ( including the frequency regime near resonance ) when there are only one electric resonance and one magnetic resonance .
@xcite one point to note here is that for a passive particle , we have to satisfy @xmath96 for the electric polarizability and also a similar inequality for the magnetic polarizability .
@xcite it guarantees a positive imaginary part for both the local permittivity and the local permeability according to eq .
( [ eq18 ] ) , i.e. @xmath97 or @xmath98 should be zero for a lossless medium or positive for a lossy medium under the dipolar model . in the following , we would like to specify the local effective medium \{@xmath99 , @xmath100 } directly instead of specifying the polarizabilities \{@xmath101 , @xmath77}. as an example , we consider the case that @xmath102 , @xmath103 , @xmath104 , @xmath105 and @xmath106 .
the underlying microstructure is assumed to be a cscl structure of lattice constant @xmath107 .
the polarizabilities can be found by eq .
( [ eq18 ] ) whenever a microscopic description is necessary .
1(a ) and ( b ) show the local effective permittivity and the local permeability .
1(c ) shows the crystal being viewed from the direction [ 001 ] or [ 010 ] . the uniaxial electric particle ( green color )
can be polarized along the [ 100 ] direction and the uniaxial magnetic particle can be polarized along the [ 010 ] direction .
we will start from this configuration in the next section when we consider the non - local effective medium . ) with @xmath108 , @xmath109 , @xmath110 , @xmath105 , @xmath106 and lattice constant a= 5 mm .
( a ) local permittivity ; ( b ) local permeability ; ( c ) the underlying cscl structure viewed along the [ 001 ] and the [ 010 ] direction . the uniaxial electric particle ( green color ) can be polarized along the [ 100 ] direction . the uniaxial magnetic particle ( blue color )
can be polarized along the [ 010 ] direction .
, width=336 ] it is more convenient to specify the metamaterial by its local effective medium directly .
however , a note has to be taken for the freedom of the parameters in the frequency dispersion . in particular , for a non - absorptive particle , the polarizability can be written in terms of the dipolar scattering phase shift @xmath111 ( a real number ) by @xmath112 where @xmath113 for `` e '' wave scattering ( non - zero local electric field with vanishing local magnetic field at the particle ) or `` m '' wave scattering ( non - zero local magnetic field with vanishing local electric field at the particle ) . due to the finite volume of one single unit cell , the phase shift @xmath111 can not vary arbitrarily against frequency .
suppose a sphere of radius @xmath114 can completely enclose one artificial electric particle , the restriction on @xmath115 can be written in terms of the time - averaged total energy @xmath116 within the enclosing sphere ( for a quasi - monochromatic electromagnetic field ) as : @xmath117 where @xmath118 .
see appendix [ appendixb ] for the derivation .
this inequality must be satisfied .
otherwise , we can extract energy from the artificial particle without first pumping energy to it . and
( b ) @xmath119 enclosing the electric particle for `` e '' wave scattering.,width=336 ] for our example , the polarizabilities are obtained through eq .
( [ eq18 ] ) and the time - averaged total energy for `` e '' wave scattering ( in an arbitrary unit ) is obtained from eq .
( [ eq20 ] ) and eq .
( [ eq21 ] ) .
[ fig2 ] shows the results for two different cases .
the solid line shows the case with @xmath120 .
the total energy within the sphere is positive in the whole frequency regime .
the dashed line shows the case with a smaller radius @xmath121 . in this case
, the energy becomes negative which is not valid .
therefore , in order to design our metamaterial , the electric artificial particle must have a certain minimum size . on the other hand ,
if we already know the size of the electric artificial particle , it imposes a maximum on the electric resonance strength @xmath122 ( in eq .
( [ eq19 ] ) ) that the metamaterial can have .
in this section , we investigate how the effective medium changes when the wave vector is deviated from the brillouin zone center .
in fact , the tensor @xmath123 does not vanish in general .
for the double lattice we considered in the last section with @xmath41 along direction [ 111 ] or [ 100]/[010]/[001 ] , eq .
( [ eq10 ] ) can still be written in the form of eq .
( [ eq16 ] ) where @xmath124 . \end{array}\ ] ] this 2-by-2 matrix is in the @xmath125 basis with @xmath45 being the direction along @xmath24 , @xmath126 being the direction the uniaxial electric particle can be polarized and @xmath127 being the direction the uniaxial magnetic particle can be polarized . therefore , by comparing to eq .
( [ eq17 ] ) , the non - local effective medium can be written with respect to the local effective medium by @xmath128^ { - 1 } = \\ & \left [ { { \begin{array}{*{20}c } { \frac{1}{\epsilon _ t - 1 } - \delta \beta _ t \left ( { \omega , { \rm { \bf q } } } \right ) } \hfill & { \gamma _ { 12 } \left ( { \omega , { \rm { \bf q } } } \right ) } \hfill \\ { \gamma _ { 12}^\ast \left ( { \omega , { \rm { \bf q } } } \right ) } \hfill & { \frac{1}{\mu _ t - 1 } - \delta \beta _ t \left ( { \omega , { \rm { \bf q } } } \right ) } \hfill \\
\end{array } } } \right ] , \end{split}\ ] ] where @xmath129 , @xmath130 , @xmath131 and @xmath132 are the elements of the four constitutive tensors in @xmath44 basis : @xmath133;\mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\leftrightarrow$}}\over { \xi } } \left ( { \rm { \bf q } } \right ) = \left [ { { \begin{array}{*{20}c } 0 \hfill & { - \xi \left ( { \rm { \bf q } } \right ) } \hfill & 0 \hfill \\ 0 \hfill & 0 \hfill & 0 \hfill \\ 0 \hfill & 0 \hfill & 0 \hfill \\ \end{array } } } \right ] \\ \mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\leftrightarrow$}}\over { \zeta } } \left ( { \rm { \bf q } } \right ) = \left [ { { \begin{array}{*{20}c } 0 \hfill & 0 \hfill & 0 \hfill \\ { - \zeta \left ( { \rm { \bf q } } \right ) } \hfill & 0 \hfill & 0 \hfill \\ 0 \hfill & 0 \hfill & 0 \hfill \\ \end{array } } } \right];\mathord{\buildrel{\lower3pt\hbox{$\scriptscriptstyle\leftrightarrow$}}\over { \mu } } \left ( { \rm { \bf q } } \right ) = \left [ { { \begin{array}{*{20}c } 1 \hfill & 0 \hfill & 0 \hfill \\ 0 \hfill & { \mu _ t \left ( { \rm { \bf q } } \right ) } \hfill & 0 \hfill \\
0 \hfill & 0 \hfill & 1 \hfill \\
\end{array } } } \right ] \\
\end{array}\ ] ] and @xmath134 is defined by @xmath135 .
@xmath62 in the y / z direction and @xmath63 in the x / z direction have the value one since we have neglected the polarizabilities in these directions for mathematical simplicity .
now , we use eq .
( [ eq23 ] ) to find the non - local effective medium for the previous example shown in fig .
1 . in this case , we plot @xmath129 , @xmath136 and @xmath131 along the @xmath48 to z[001 ] direction in fig .
the profile of @xmath129 and @xmath137 are shown in fig .
[ fig3 ] ( a ) and ( b ) . for every wave vector @xmath24
, we can see that @xmath138 and @xmath130 have the similar frequency dispersion as the local medium correspondence @xmath139 and @xmath100 , only with the resonating frequency being shifted .
this qualitative behavior is expected from eq .
( [ eq23 ] ) by neglecting the presence of @xmath140 .
however , with careful examination , we can still see that when @xmath24 deviates from the @xmath48 point , @xmath129 also diverges at the magnetic resonating frequency .
moreover , we have non - vanishing @xmath131 near the electric or the magnetic resonating frequency .
these result from the presence of the term @xmath140 which means the cross - coupling between the electric and the magnetic field due to spatial dispersion . to the z[001 ]
point for the configuration shown in the caption of fig . 1 : ( a ) @xmath129 ; ( b ) @xmath137 and ( c ) @xmath131 .
white color denotes positive and black color denotes negative values .
, width=336 ] for the current configuration , the effect of @xmath141 is rather weak .
however , it can become more prominent in other cases . as an example , for the same cscl lattice structure , we now consider the wavevector @xmath24 along the @xmath48 to r[111 ] direction and the particles are now orientated so that the uniaxial electric particle can be polarized along the [ -112 ] direction while the uniaxial magnetic particle can be polarized along the [ 1 - 10 ] direction .
this second configuration is shown in fig .
@xmath142 , @xmath130 along the @xmath48 to r[111 ] direction are now plotted in fig .
[ fig5 ] . in fig .
[ fig6](a ) and ( b ) , we also plot @xmath143 and @xmath130 against frequency at exactly the r point . from the results , the resonating frequencies depend on @xmath24 and
both @xmath144 and @xmath130 now clearly diverge at two separate frequencies ( instead of one ) for every non - zero @xmath24 due to the appearance of the term @xmath145 .
the two resonating frequencies @xmath146 ( the lower one ) and @xmath147 ( the upper one ) can be obtained from eq .
( [ eq23 ] ) and are governed by @xmath148 ) with @xmath108 , @xmath109 , @xmath110 , @xmath105 , @xmath106 and lattice constant a= 5 mm .
the underlying cscl structure is viewed along the ( a ) [ 111 ] and ( b ) [ 0 - 10 ] direction . the uniaxial electric particle ( green color ) can be polarized along the [ -112 ] direction . the uniaxial magnetic particle ( blue color ) can be polarized along the [ 1 - 10 ] direction . ,
width=336 ] to the r point for the configuration shown in the caption of fig .
[ fig4 ] : ( a ) @xmath129 ; ( b ) @xmath137 and ( c ) @xmath131 .
white color denotes positive and black color denotes negative values.,width=336 ] : ( a ) @xmath138 ; ( b ) @xmath130 and ( c ) @xmath131.,width=336 ] if the magnetic resonance and the electric resonance are far apart in frequency or the term @xmath140 can be neglected , it decouples to @xmath149 which determines the electric resonating frequency and @xmath150 which determines the magnetic resonating frequency .
it gives us two separate `` bands '' of resonating frequencies .
the term @xmath151 gives us the @xmath24 dependence of the electric or the magnetic resonating frequency ( shown in the previous example in fig .
[ fig3 ] ) . however , in the current example , these two bands hybridize with each other so that @xmath152 or @xmath130 has two resonating frequencies instead of one .
it is called the cross - coupled resonance here .
due to the same reason , the @xmath140 term also causes a structure - induced bianisotropy when we bring the electric and magnetic resonances near to each other in frequency .
this bianisotropy is induced by the structure instead of the cross - coupling effect of the artificial atoms .
the @xmath131 from the @xmath48 to the r point is plotted in fig . [
fig5](c ) .
it also diverges at @xmath153 and @xmath154 .
we have used a lorentz - model dispersion ( eq .
( [ eq19 ] ) ) for the local effective medium .
in fact , @xmath129 , @xmath130 and @xmath131 can be numerically fitted very well by extending the same model : @xmath155 @xmath156 and @xmath157 in the current example at the @xmath158 point , the expressions obtained by setting @xmath159 , @xmath160 , @xmath161 , @xmath162 , @xmath163 , @xmath164 , @xmath165 and @xmath166 can fit the actual results very well .
the fitted result has no noticeable differences from the results shown in fig .
[ fig6 ] .
the extended lorentz - model dispersion is valid even for an absorptive system .
suppose , now we add some absorption to the system by setting @xmath167 and @xmath168 in eq .
( [ eq19 ] ) , the resultant non - local effective medium at the @xmath158 point is shown in fig .
the non - local effective medium can still be fitted by the extended lorentz model but with non - zero @xmath169 and @xmath170 . in this case , we set @xmath171 and @xmath172 while all other parameters remain the same
. the fitted expressions also show no noticeable differences to the results shown in fig .
[ fig7 ] . from fig .
[ fig7 ] , we see that the imaginary part of both @xmath129 and @xmath130 are positive . for a bianisotropic medium , there are no general requirements that the imaginary part of the permittivity and the permeability should be positive .
however , within the dipolar model , from eq .
( [ eq23 ] ) with @xmath173 being real , it is easy to show that both @xmath129 and @xmath130 have positive imaginary part if the local @xmath99 and @xmath100 have positive imaginary part , i.e. @xmath169 and @xmath170 should be positive . in our approach ,
we parameterize the effective medium by using @xmath24 such that @xmath24 is always a real vector within the first brillouin zone .
if , on the other hand , @xmath24 is allowed to be complex ( e.g. traveling along the dispersion of an eigenmode and becoming complex within the resonance gap where no propagating eigenmodes can be found ) , a negative imaginary part in @xmath129 or @xmath137 may appear around a cross - coupled resonance .
@xcite with @xmath174 and @xmath175 : ( a ) @xmath129 ; ( b ) @xmath130 ; ( c ) @xmath176 .
the blue solid line shows the real part and the gray dashed line shows the imaginary part.,width=336 ] the structure - induced bianisotropy @xmath131 appears even if the local effective medium does not suffer bianisotropy .
@xmath131 is generally present and it is purely real for a non - absorptive system in our example .
it disappears in the local effective medium limit only if the particles do not suffer cross - coupling and the particles are placed in a carefully chosen lattice .
for example , if the sublattices of the two types of particles are displaced in the z direction ( along @xmath177 with respect to each other , the 2d symmetry perpendicular to @xmath24 does not change so that the constitutive tensors can still be described by eq .
( [ eq24 ] ) .
however , @xmath178 will not vanish and it is a purely imaginary number . on the other hand , if the lattice is not displaced but the particles suffer cross coupling , @xmath179 also does not vanish and it is purely imaginary .
@xcite to the r point for the configuration shown in the caption of fig .
[ fig4 ] with lattice constant a changed to 2.5 mm : ( a ) @xmath142 ; ( b ) @xmath130 and ( c ) @xmath180 .
white color denotes positive and black color denotes negative values.,width=336 ] the structure - induced bianisotropy and cross - coupled resonance are both due to the term @xmath140 .
in fact , with a smaller lattice constant , the interaction becomes smaller .
[ fig8 ] shows the @xmath142 , @xmath130 and @xmath181 along the @xmath48 to r[111 ] direction when the lattice constant @xmath182 is changed a smaller value 2.5 mm .
the splitting between the two resonating frequencies is smaller than in fig .
[ fig5 ] . on the other hand ,
if the lattice constant becomes larger , the term @xmath183 can not be neglected in finding the effective medium or in finding the dispersion .
direction and the configuration shown in
( b ) fig .
[ fig4 ] along the @xmath184 direction .
the black dashed line represents the corresponding dispersion obtained from the local effective medium .
the blue solid line represents the perturbed dispersion from the local one.,width=336 ] we have investigated the constitutive tensors of the metamaterial . after finding the effective medium
, the dispersion diagram can also be found .
the dispersion relation of the effective medium is governed by @xmath185 after some algebra , we have @xmath186 and the dispersion relation is therefore @xmath187 there is an extra term @xmath188 in the dispersion relation comparing to the one obtained from the local effective medium : @xmath189 unlike the case of the local effective medium which serves as an approximation of the actual band structure , the dispersion relation shown in eq .
( [ eq33 ] ) can be used to find the band structure which shows no noticeable differences to the actual band structure if we solve it directly without first finding the non - local effective medium . for band structure of a single lattice consisting artificial particles of only electric or only magnetic response ,
see ref . and ref .
the dispersion diagram along the direction @xmath190 for the first configuration ( fig . 1 ) and the dispersion diagram along the direction @xmath191 for the second configuration ( fig . [ fig4 ] ) are shown in fig . [ fig9](a ) and ( b ) respectively .
they are plotted in open black circles while the dispersion diagram obtained from the local effective medium is plotted in dashed line . along the @xmath190 direction for the first configuration
, they look very similar especially near the brillouin zone center .
in fact , the formation of the non - local bands can be understood from a perturbation point of view by assuming small @xmath151 and @xmath140 .
for a fixed @xmath24 , we can approximate the change in square frequency from the local to the non - local dispersion by @xmath192 which is evaluated along one of the bands obtained from the local effective medium .
( the proof is not shown here . ) by substituting eq .
( [ eq23 ] ) into it and neglect the second order effect of @xmath88 and @xmath193 , we have @xmath194 this perturbed dispersion is plotted in fig .
[ fig9 ] in solid line .
we can see that it approximates the actual non - local dispersion very well in the @xmath48z direction of the first configuration . in the @xmath48r direction , the first band ( having positive refractive index in the local limit ) and the second band ( having negative refractive index in the local limit ) hybridize with each other to form the actual band . for every mode in the dispersion diagram
, we can characterize it by defining the refractive index @xmath13 of the mode using @xmath195 and defining its characteristic impedance z to be the ratio between the tangential @xmath196 and the tangential @xmath197 field .
they are governed by @xmath198 for example , the @xmath13 and @xmath199 along the band in the @xmath48z direction in fig .
[ fig9 ] are plotted as solid lines in fig .
[ fig10 ] for the double - negative band .
the corresponding values for the local effective medium ( @xmath200,@xmath201 are also shown in dashed lines in the same figure .
we can see that the local effective medium works very well for the propagating bands unless it is near the brillouin zone edge and near the resonance gap .
direction . dashed line :
corresponding values for the local medium.,width=336 ] we have examined the non - local effective medium governing the transverse waves . in fact , if the artificial particles are oriented so that they can only be polarized in a single direction along the z - axis , the metamaterial supports longitudinal waves whose propagating direction is also along the z - axis .
we can do the similar analysis to obtain the longitudinal permittivity @xmath202 and the longitudinal permeability @xmath203 . in this case , the electric and magnetic response are decoupled , @xmath202 and @xmath203 can also be fitted by using a lorentz - model dispersion : @xmath204 and @xmath205 the electric longitudinal mode ( or the electric bulk plasmon with macroscopic e field along @xmath206 ) has a dispersion relation governed by @xmath207 and the dispersion relation for the the magnetic longitudinal mode ( or the magnetic bulk plasmon with macroscopic h field along @xmath206 ) is governed by @xmath208 with spatial dispersion such that @xmath209 and @xmath210 depends on @xmath206 , the band for either the electric or magnetic longitudinal mode becomes dispersive to form a narrow band instead of a flat line in the local medium picture .
we have established a numerical method based on a dipolar model to obtain the non - local effective medium of a metamaterial . in particular ,
we have concentrated on the double lattice structure in which one sublattice holds the electric artificial atoms and another sublattice holds the magnetic artificial atoms . once the dipolar scattering properties of the atoms or the local effective medium parameters are specified , the non - local effective medium with all the four constitutive tensors can be obtained on the whole frequency regime including frequencies near the resonance gap .
we found that the metamaterial should be regarded as bianisotropic near the resonance gap even if the artificial atoms do not suffer cross - coupling . in this case , the cross - coupling comes from spatial dispersion and it also induces a cross - coupled resonance for metamaterials having both electric and magnetic resonances . within our model
, the cross - coupled resonance can still be understood using a lorentz dispersion model .
the effect of the cross - coupling from spatial dispersion depends on the microstructure and it is larger at a higher normalized frequency ( or a larger lattice constant ) .
this work is supported by the croucher foundation fellowship from hong kong .
the radiation field from an electric dipole @xmath211 together with a magnetic dipole @xmath212 at the origin in the vacuum is @xmath213 = { g}_0 \left ( { \rm { \bf r } } \right ) \cdot \left [ { { \begin{array}{*{20}c } { c_0{\rm { \bf p } } } \hfill \\ { \rm { \bf m } } \hfill \\ \end{array } } } \right],\ ] ] where the @xmath214 matrix is defined by @xmath215,\ ] ] with the vacuum green s tensor @xmath216 defined by @xmath217 where @xmath218 is the 3-by-3 identity tensor .
the fourier transform of @xmath34 is given by @xmath219 , \\
\end{split}\ ] ] where @xmath220
here , we will consider the restriction on the polarizability for a lossless classical uniaxial particle which can be described by its permittivity distribution @xmath221 and its permeability distribution @xmath222 . the particle has a finite radius @xmath223 such that it is vacuum outside the sphere of radius @xmath223 centered at the origin . without losing generality
, we consider the situation that `` m '' wave ( with angular momentum @xmath224 and @xmath225 is scattered from the particle which can be magnetically polarized along the z - direction but not the other directions .
we also assume that the particle has no cross - coupling , i.e. the electric dipole moment is zero when the local electric field is zero here . the total wave outside the particle can be written in the following form : with @xmath228 being the out - going spherical hankel function , and @xmath229 being the spherical harmonic .
@xmath230 , the phase shift in scattering , is a real number for a lossless particle and it is related to the magnetic polarizability through the volume and the surface integrals are taken to be the volume and the surface of the sphere of radius @xmath114 which encloses the particle completely . by substituting eq .
( [ eqb1 ] ) into eq .
( [ eqb4 ] ) , we obtain where @xmath234 .
finally , we recognize the l.h.s . of eq .
( [ eqb5 ] ) is just the time - averaged total electromagnetic energy ( @xmath235 within the sphere for a narrow band signal centered at the angular frequency @xmath4 . by substituting eq .
( [ eqb2 ] ) into eq .
( [ eqb5 ] ) , we can express the time - averaged total energy in terms of @xmath230 : which must be a positive number .
the derivation for the electric scattering phase shift is similar .
moreover , the derivation also applies for an isotropic particle instead of an uniaxial particle being discussed here .
2 d. r. smith , d. c. vier , th .
koschny , and c. m. soukoulis , _ phys .
e _ * 71 * , 036617 ( 2005 ) .
j. b. pendry , a. j. holden , d. j. robbins , and w. j. stewart , _ ieee trans .
microwave theory tech .
_ * 47 * , 2075 ( 1999 ) .
d. r. smith , j. gollub , j. j. mock , w. j. padilla and d. schurig , _
phys . _ * 100 * , 024507 ( 2006 ) .
lerat , n. malljac , and o. acher , _ j. appl . phys . _ * 100 * , 084908 ( 2006 ) .
p. a. belov and c. r. simovski , _ phys .
b _ * 73 * , 045102 ( 2006 ) .
p. a. belov , r. marques , s. i. maslovski , i. s. nefedov , m. silveirinha , c. r. simovski , and s. a. tretyakov , _ phys .
b _ * 67 * 113103 ( 2003 )
. m. g. silveirinha , _ phys .
e _ * 73 * , 046612 ( 2006 ) .
m. a. shapiro , g. shvets , j. r. sirigiri and r. j. temkin , _ opt . lett . _ * 31 * , 2051 ( 2006 ) .
f. j. garca de abajo and j. j. senz , phys .
lett . * 95 * , 233901 ( 2005 ) .
r. marques , f. medina , and r. rafii - el - idrissi , _ phys .
b _ * 65 * , 144440 ( 2002 ) . xudong chen , bae - ian wu , jin au kong , and tomasz m. grzegorczyk , _ phys . rev
. e _ * 71 * , 046610 ( 2005 ) .
d. schurig , j. j. mock , and d. r. smith , _ appl .
lett . _ * 88 * 041109 ( 2006 ) .
v. dmitriev , _ prog .
_ * 28 * , 43 ( 2000 ) .
s. ponti , c. oldano , and m. becchi , _ phys .
* 64 * , 021704 ( 2001 ) .
s. ponti , j. a. reyes and c. oldano , _ j. phys .
: cond . matter _ * 14 * , 10173 ( 2002 ) .
juan d. baena , ricardo marqus , francisco medina , and jess martel , _ phys .
b _ * 69 * 014402 ( 2004 ) .
p. a. belov , and c. r. simovski , _ phys .
e _ * 72 * , 026615 ( 2005 ) .
s. a. tretyakov , _ ieee trans .
antennas propag . _ * 51 * 2652 ( 2003 ) .
g. d. mahan , _ phys .
b _ * 74 * , 033407 ( 2006 ) . a. d. boardman and k. marinov , _ phys . rev . b _ * 73 * , 165110 ( 2006 ) .
s. tretyakov , analytical modeling in applied electromagnetics ( artech house , boston 2003 ) . th .
koschny , et.al .
b _ * 71 * , 245105 ( 2005 ) . k. kempa , r. ruppin and j. b. pendry , _ phys . rev .
b _ * 72 * , 205103 ( 2005 ) . | in this paper , we investigate the effect of spatial dispersion on a double - lattice metamaterial which has both magnetic and electric response .
a numerical scheme based on a dipolar model is developed to extract the non - local effective medium of the metamaterial .
we found a structure - induced bianisotropy near the resonance gap even if the artificial particles are free from cross - coupling .
a cross - coupled resonance also results from spatial dispersion and it can be understood using a lorentz frequency dispersion model .
the strength of the cross - coupled resonance depends on the microstructure and it is larger at a higher normalized frequency . |
the results obtained in the last few years by the na48 @xcite and the ktev @xcite collaborations have marked a great experimental achievement , establishing some 35 years after the discovery of cp violation in the neutral kaon system @xcite the existence of a much smaller violation acting directly in the decays : [ eoenew ] ( / ) = \ { ll ( 15.3 2.6)10 ^ -4 & ( na48 ) + ( 20.7 2.8)10 ^ -4 & ( ktev ) . .
the average of these results with the previous measurements by the na31 collaboration at cern and by the e731 experiment at fermilab gives [ eoewa ] ( / ) = ( 17.2 1.8)10 ^ -4 .
while the standard model ( sm ) of strong and electroweak interactions provides an economical and elegant understanding of indirect ( @xmath1 ) and direct ( @xmath2 ) cp violation in term of a single phase , the detailed calculation of the size of these effects implies mastering strong interactions at a scale where perturbative methods break down .
in addition , direct cp violation in @xmath3 decays arises from a detailed balance of two competing sets of contributions , which may hopelessly inflate the uncertainties related to the relevant hadronic matrix elements in the final outcome .
all that makes predicting a complex and challenging task @xcite . just from the onset of the calculation
the presence in the definition of , written as = \ { - } , [ eoedef ] of given ratios of isospin amplitudes warns us of a longstanding and still unsolved theoretical `` problem '' : the explanation of the @xmath0 selection rule .
the @xmath0 selection rule in @xmath3 decays is known since 45 years @xcite and it states the experimental evidence that kaons are 400 times more likely to decay in the @xmath4 two - pion state than in the @xmath5 component ( @xmath6 ) .
this rule is not justified by any symmetry argument and , although it is common understanding that its explanation must be rooted in the dynamics of strong interactions , there is up to date no derivation of this effect from first principle qcd .
given the possibility that common systematic uncertainties may a - priori affect the calculation of and the @xmath0 rule ( see for instance the present difficulties in calculating on the lattice the `` penguin contractions '' for cp violating as well as for cp conserving amplitudes @xcite ) a convincing calculation of must involve at the same time a reliable explanation of the @xmath0 selection rule . both observables indicate the need of large corrections to factorization in the evaluation of the four - quark hadronic transitions . among these corrections
final state interactions ( fsi ) play a substantial role .
however , fsi alone are _ not _ enough to account for the large ratio of the @xmath4 over @xmath5 amplitudes .
other sources of large non - factorizable corrections are therefore needed for the cp conserving amplitudes @xcite , which might affect the determination of as well . as a consequence , a self - contained calculation of
should also address the determination of the @xmath3 rates .
the operator product expansion ( ope ) provides us with a very effective way to address the calculation of hadronic transitions in gauge theories .
the integration of the `` heavy '' gauge and matter fields allows us to write the relevant amplitudes in terms of the hadronic matrix elements of effective quark operators and of the corresponding wilson coefficients ( at a scale @xmath7 ) , which encode the information about those dynamical degrees of freedom which are heavier than the chosen renormalization scale . according to the sm flavor structure the @xmath8 transitions are effectively described by _
s = 1 = v_udv^*_us _ i ( ) . [ leff ]
the entries @xmath9 of the @xmath10 cabibbo - kobayashi - maskawa ( ckm ) matrix describe the flavour mixing in the sm and @xmath11 . for @xmath12 ( @xmath13 ) ,
the relevant quark operators are : lcl .
lcl q_1 & = & ( _ u _ ) _ v - a ( _ d _ ) _ v - a + q_2 & = & ( u ) _ v - a ( d ) _ v - a } & & + .
lcl q_3,5 & = & ( d ) _ v - a _ q ( q ) _ va + q_4,6 & = & ( _ d _ ) _ v - a _ q ( _ q _ ) _ va } & & + .
lcl q_7,9 & = & ( d ) _ v - a _ q _ q ( q ) _ va + q_8,10 & = & ( _ d _ ) _ v - a _ q _ q ( _ q_)_va } & & [ quarkeff ] current - current operators are induced by tree - level w - exchange whereas the so - called penguin ( and `` box '' ) diagrams are generated via an electroweak loop .
only the latter `` feel '' all three quark families via the virtual quark exchange and are therefore sensitive to the weak cp phase .
current - current operators control instead the cp conserving transitions .
this fact suggests already that the connection between and the @xmath0 rule is by no means a straightforward one . using the effective @xmath14 quark hamiltonian we can write as = e^i [ main ] where lcl _ 0 & = & _ i y_i q_i _ 0 ( 1 - _ ib ) + _ 2 & = & _ i y_i q_i _ 2 , [ pi02 ] and @xmath15 .
the rescattering phases @xmath16 can be extracted from elastic @xmath17-@xmath17 scattering data@xcite and are such that @xmath18 and @xmath19 . given that the phase of @xmath1 ( @xmath20 ) is approximately @xmath21 , as well as the difference @xmath22
, the @xmath23 phase @xmath24 turns out to be consistent with zero .
while @xmath25 , @xmath26 , @xmath27 and @xmath28 are precisely determined by experimental data , the first source of uncertainty that we encounter in is the value of @xmath29 , the combination of ckm elements which measures cp violation in @xmath8 transitions .
the determination of @xmath30 depends on b - physics constraints and on @xmath1 @xcite . in turn
, the fit of @xmath1 depends on the theoretical determination of @xmath31 , the @xmath32 hadronic parameter , which should be self - consistently determined within every analysis .
the theoretical uncertainty on @xmath31 was in the past the main component of the final uncertainty on @xmath30 .
the improved determination of the unitarity triangle coming from b - factories and hadronic colliders @xcite has weakened and will eventually lift the dependence of @xmath30 on @xmath31 , allowing for an experimental measurement of the latter from . within kaon physics ,
the decay @xmath33 gives the cleanest `` theoretical '' determination of @xmath34 , albeit representing a great experimental challenge . at present
, a typical range of values for @xmath34 is @xmath35 @xcite .
we come now to the quantities in the square brackets . while the calculation of the wilson coefficients is well under control , thanks primarily to the work done in the early nineties by the munich @xcite and rome @xcite groups , the evaluation of the `` long - distance '' factors in is the crucial issue for the ongoing calculations .
the isospin breaking ( ib ) parameter @xmath36 , gives at the leading - order ( lo ) in the chiral expansion a @xmath37 correction to the @xmath38 amplitude ( proportional to @xmath39 via the @xmath40 mixing ) of about 0.13 @xcite . at the next - to - leading order ( nlo ) the full inclusion of the @xmath41 mixing lift the value of @xmath36 to @xmath42 @xcite .
on the other hand , the complete nlo calculation of ib effects beyond the @xmath41 mixing ( of strong and electromagnetic origin , among which the presence of @xmath43 transitions ) involves a number of unknown nlo chiral couplings and is presently quite uncertain .
dimensional estimates show that ib effects may be large and affect sizeably in both directions @xcite .
although a partial cancellation of the indirect ( @xmath44 ) and direct @xmath45 nlo isospin breaking corrections in may reduce their final numerical impact on , we must await for further analyses in order to confidently assess their relevance . at present
one may use @xmath46 @xcite as a conservative estimate of the ib effects .
the final basic ingredient for the calculation of is the evaluation of the @xmath3 hadronic matrix elements of the quark operators in .
a simple albeit naive approach to the problem is the vacuum saturation approximation ( vsa ) , which is based on two drastic assumptions : the factorization of the four quark operators in products of currents and densities and the saturation of the intermediate states by the vacuum state . as an example
^-|q_6| k^0 & = & 2 ^-|_5 d|0 ^+| u |k^0 - 2 ^+ ^-| d|0 0| _ 5 d |k^0 + & & + 2 ^+ ^-|_5 d |k^0 the vsa does not exhibit a consistent matching of the renormalization scale and scheme dependences of the wilson coefficients and it carries potentially large systematic uncertainties @xcite . on the other hand it provides useful insights on the main features of the problem . a pictorial summary of the relative weights of the contributions of the various operators to , as obtained in the vsa ,
is shown in fig .
[ fig : pie ] .
as we have already mentioned , cp violation involves loop - induced operators ( @xmath47 ) . from fig .
[ fig : pie ] one clearly notices the potentially large cancellation among the strong and electroweak sectors and the leading role played by the gluonic penguin operator @xmath48 and the electroweak operator @xmath49 .
tipical range of values for , obtained using the vsa , are shown in fig .
[ fig : pre ] together with the three most updated predictions available before 1999 ( when the first ktev and na48 results became known ) @xcite .
the fact that the cancellation among the strong and electroweak sectors turns out to be quite effective ( in the vsa ) warns us about the possibility that the uncertainties in the determination of the relevant hadronic matrix elements may be largely amplified in the calculation of .
it is therefore important to asses carefully the approximations related to the various parts of the calculations . in particular , the analysis of the problem suggests that factorization may be highly unreliable .
the dark gray bars in fig . [ fig : pre ] depict the results of three calculations of which are representative of approaches that ( in principle ) allow us to go beyond naive factorization .
they are based from left to right on the large @xmath50 expansion @xcite , on lattice regularization @xcite , and on phenomenological modelling of low - energy qcd ( the chiral quark model ) @xcite .
the experimental and theoretical scenarios have changed substantially after the first ktev data and the subsequent na48 results .
[ fig : post ] shows the present experimental world average for compared with the revised or new theoretical calculations that appeared during the last year . without entering into the details of the results ( for a short summary see @xcite )
they all represent attempts to incorporate non - perturbative information into the calculation of the hadronic matrix elements , whether their are based on the large @xmath50 expansion ( mnchen @xcite , dortmund @xcite , beijing @xcite , taipei @xcite , valencia @xcite ) , phenomenological modelling of low - energy qcd ( dubna @xcite , trieste @xcite , lund @xcite ) , qcd sum rules ( montpellier @xcite ) or , finally , on lattice regularization ( roma @xcite , cp - pacs @xcite , rbc @xcite ) .
overall most of the theoretical calculations are consistent with a non - vanishing positive effect in the sm ( with the exception of the recent lattice results on which i will comment shortly ) . at a closer look however , if we focus our attention on the central values , many of the predictions prefer the @xmath51 regime , whereas only a few of them stand above @xmath52 .
is this just `` noise '' in the theoretical calculations ? without entering the many details on which the estimates are based , most of the aforementioned difference can be explained in terms of a single effect : the different size of the hadronic matrix element of the gluonic penguin @xmath48 as obtained in the various approaches . in turn , this can be understood in terms of sizeable higher order chiral contributions ( nlo in the @xmath53 expansion ) to the @xmath4 amplitudes .
this effect was stigmatized well before the latest experimental round by the work of the trieste group @xcite , and appears clearly in the comparison of the leading @xmath53 and lattice results with the chiral quark model analysis in fig .
[ fig : pre ] . the chiral quark model approach , together with the fit of the cp conserving amplitudes which normalizes phenomenologically the matching and the model parameters , allows us to carry the calculation of the hadronic matrix elements beyond the leading order in the chiral expansion ( including the needed local counterterms ) .
non - factorizable chiral contributions ( missing in the leading @xmath53 or lattice calculations ) were shown to produce a substantial enhancement of the @xmath4 transitions thus lifting the expectation of at the @xmath52 level . since
then a number of groups have attempted to improve the calculation of @xmath3 matrix elements in a model independent way .
table [ tab : b6b8 ] presents a comparison of different calculations of the relevant matrix elements . due to the leading role played by @xmath48 and @xmath49
we may write a simplified version of , 13 ( ) _ t ^2 , which although `` not be used for any serious analysis '' @xcite gives an effective and practical way to test and compare different calculations .
the b - factors @xmath54 represent a convenient parametrization of the hadronic matrix elements , albeit tricky , in that their values are in general scale and renormalization - scheme dependent , and a spurious dependence on the quark masses is introduced in the result whenever quark densities are involved .
the latter is the case for the @xmath48 and @xmath49 penguins . as a consequence
the vsa normalization may vary from author to author thus introducing systematic ambiguities . by taking the vsa matrix elements at the scale
@xmath55 gev we obtain @xcite lcl@ & & _ vsa = f m_k^4 ( m_s+m_d)^-2 1.1 gev^3 , + & & _ vsa/ _ vsa = -2 ( f_k - f_)/f_-0.63 , [ vsanorm ] where i have used @xmath56 and the chiral value @xmath57 mev for the octet decay constant .
it is known that @xmath58 and @xmath59 are perturbatively very weakly dependent on the renormalization scale @xcite .
therefore it makes sense to compare the @xmath60 s obtained in different approaches , where the matrix elements @xmath61 are computed at different scales .
the results for the relevant penguin matrix elements coming from various approaches are collected in table [ tab : b6b8 ] , paying care to normalizing the data in a homogeneous way ( as far as detailed information on definitions and renormalization schemes was available ) . as a guiding information , taking @xmath62 , the present experimental central value of is reproduced by @xmath63 .
[ tab : b6b8 ] .comparison of various calculations of penguin matrix elements .
the data marked by the star are rescaled by a factor @xmath64 , to account for a different definition of the isospin matrix elements . [ cols="<,<,<,<,<",options="header " , ] the most important fact is the first evidence of a signal in lattice calculations of @xmath65 , obtained by the cp - pacs @xcite and rbc @xcite collaborations .
both groups use the domain wall fermion approach which allows to control the chiral symmetry on the lattice as a volume effect in a fifth dimension .
this approach softens in principle the problem of large power subtractions which affects the lattice extraction of @xmath4 amplitudes ( penguin contractions ) .
still only the @xmath66 transition is computed on the lattice and lo chiral perturbation theory is used to extrapolate it to the physical amplitude .
the two groups obtain comparable values of the @xmath48 and @xmath49 matrix elements leading both to a negative ( and do not agree on the cp conserving @xmath4 amplitude ) . on the other hand
the calculations are at an early stage and do not include higher order chiral dynamics which may be responsible for the enhancement of @xmath4 amplitudes ( as large @xmath50 approaches beyond lo and the chiral quark model strongly suggest ) .
the spqcdr collaboration has reported a result for the @xmath49 matrix element from direct calculation of the @xmath3 amplitude on the lattice .
this result agrees with previous lattice data , albeit it does not yet include the chiral corrections relevant to quenching and to the extrapolation to the physical pion mass @xcite . among the analytic approaches important results
have been obtained using data on spectral functions in connection with qcd sum rules and dispersive relations in the attempt to obtain model independent information on the relevant matrix elements .
these approaches have produced as of today calculations of @xmath49 ( in the chiral limit ) which are subtantially larger than the factorization ( and lattice ) results . while there is still disagreement among the different analysis
, we must await the calculation of the @xmath48 matrix element and a quantitative assessment of chiral breaking effects before drawing conclusions on these as well as lattice results .
calculations which sofar have allowed for the determination of all relevant parameters , based on chiral perturbation theory and/or models of low - energy qcd , have shown the crucial role of higher order non - factorizable corrections in the enhancement of the @xmath4 matrix elements @xcite .
chiral loop corrections drive the final value of in the ballpark of the present data . however the calculation of higher order chiral effects can not be fully accomplished in a model independent way due to the many unknown nlo local couplings . in the chiral quark model approach
all needed local interactions are computed in terms of quark masses , meson decay constants and a few non - perturbative parameters as quark and gluon condensates .
the latter are determined self - consistently in a phenomenological way via the fit of the cp conserving @xmath3 amplitudes , thus encoding the @xmath67 rule in the calculation @xcite .
the analysis shows that the role of local counterterms is subleading to the chiral logs when using the modified minimal subtraction ( as opposed to the commonly used gasser - leutwyler prescription ) .
the phenomenological fit is crucial in stabilizing the numerical prediction @xcite .
the fact that the model parameters ( quark and gluon condesates , constituent quark mass ) turn out to be in the expected range , shows that the explicitly included chiral ( and @xmath68 gluon condensate ) corrections represent the largest non - factorizable effect . among higher order corrections
fsi play a leading role . as a matter of fact
, one should in general expect an enhancement of with respect to the naive vsa due to fsi .
as fermi first argued @xcite , in potential scattering the isospin @xmath4 two - body states feel an attractive interaction , of a sign opposite to that of the @xmath5 components thus affecting the size of the corresponding amplitudes .
this feature is at the root of the enhancement of the @xmath4 amplitude over the @xmath5 one and of the corresponding enhancement of beyond factorization .
an attempt to resum these effects in a model independent way has been worked out by the authors of ref .
@xcite , using a dispersive approach a la omns - mushkelishvili @xcite .
their analysis shows that resummation does not substantially modify the one - loop perturbative result and , as it appears from table [ tab : b6b8 ] , a 50% enhancement of the gluonic penguin matrix element is found over the factorized result .
however , the calculation suffers from a sistematic uncertainty due to the indetermination of the off - shell amplitude which is identified with the large @xmath50 result @xcite .
even when the authors in the most recent work match the dispersive resummation with the on - shell perturbative one - loop calculation , thus including @xmath53 effects , again a systematic uncertainty remains in the unknown polinomial parts of the local chiral counterterms .
therefore , a model - independent complete calculation of chiral loops for @xmath3 is still missing . finally , it has been recently emphasized @xcite that cut - off based approaches should pay attention to higher - dimension operators which become relevant for matching scales below 2 gev and may represent one of the largest sources of uncertainty in present calculations
. the results of refs .
@xcite include these effects .
the calculations based on dimensional regularization may be safe if phenomenological input is used in order to encode in the relevant hadronic matrix elements the physics at all scales ( this is done in the trieste approach ) . in summary
, while model dependent calculations suggest no conflict between theory and experiment for , a precise and `` pristine '' prediction of the observable is still quite ahead of us .
higher - order chiral corrections are taking the stage of @xmath3 physics .
they are needed in order to asses the size of crucial parameters ( as @xmath36 ) and the effect of non - factorizable contributions in the penguin matrix elements .
lattice , as a regularization of qcd , is _ the _ first - principle approach to the problem .
however , lattice calculations still heavily depend on chiral perturbation theory @xcite .
presently , very promising developments are being undertaken to circumvemt the technical and conceptual shortcomings related to the calculation of weak matrix elements @xcite . among those
are the domain wall fermion approach @xcite which allows us to decouple the chiral symmetry from the continuum limit , and the very interesting observation that the maiani - testa theorem @xcite can be overcomed using the fact that lattice calculations are performed in finite volume @xcite , thus allowing for the direct calculation of the physical @xmath3 amplitude on the lattice .
all these developments need a tremendous effort in machine power and in devising faster algorithms .
preliminary results for lattice calculations of both and the @xmath0 selection rule are already available and others are currently under way @xcite . in the meantime analytical and semi - phenomenological approaches
have been crucially helpful in driving the attention of the community on some systematic short - comings of `` first - principle '' calculations .
the amount of theoretical work triggered by the na48 and ktev data promises rewarding and perhaps exciting results in the forthcoming years . | i shortly review the present status of the theoretical calculations of and the comparison with the present experimental results .
i discuss the role of higher order chiral corrections and in general of non - factorizable contributions for the explanation of the @xmath0 selection rule and direct cp violation in kaon decays . still lacking satisfactory lattice calculations , analytic methods and phenomenological approaches
are helpful in understanding correlations among theoretical effects and experimental data .
substantial progress from lattice qcd is expected in the coming years .
= 11.6pt |
the present article proposes a theory that predicts the formation of habit plane normals very close to @xmath0 , observed in steels with low carbon content ( less than @xmath1 @xcite ) .
in fact , _ all _ the predicted habit plane normals are almost exactly with @xmath2 and + @xmath3 , where @xmath4 denotes the euclidean distance . ] @xmath5 . widely accepted models that result in @xmath0 habit planes are double shear theories , e.g. @xcite , and some of the most accurate explanations are due to the algorithm developed by kelly @xcite .
these can be seen as generalisations of the so - called phenomenological theory of martensite most notably developed by wechsler , liebermann & read @xcite to explain the @xmath6 habit planes in plate martensite and bowles & mackenzie @xcite who applied their theory to explain the @xmath7 and @xmath8 habit planes also in plate martensite .
a short - coming of single / double shear theories is the lack of a selection mechanism that picks the right lattice invariant shearing systems ( see e.g. ( * ? ? ?
* table 1 ) ) , in turn leading to a large number of input parameters . to overcome this ,
one approach is to only allow shearing systems that arise from mechanical twinning , cf . . indeed , in the context of single shear theories @xcite and @xcite made this assumption and proposed that the martensite plates consist of a _
`` stack of twin - related laths''_. as pointed out in @xcite , tem investigations in @xcite showed that even though _ `` under the optical microscope there is little sign that this is the case , when such steels are examined with the transmission electron microscope , arrays of very thin @xmath9 twins are indeed found '' _ - marking a significant success from a theoretical prediction to an observed feature in plate martensite .
regarding @xmath0 habit planes in lath martensite , it can be shown that for any reasonable choice of lattice parameters ( see also figure [ figsimplnormals ] ) , a single shear theory with shearing systems arising from twinning in bcc crystals can not give rise to them .
however , in this paper we show that by introducing another level of twinning ( `` twins within twins '' ) we are not only able to explain @xmath10 habit planes but also predict them by showing that it is the only possible family of habit planes that satisfies a condition of maximal compatibility and a condition of small overall atomic movement . under this interpretation
each lath may be seen as a region of twins within twins . in other materials , twins within twins
have commonly been observed purely in martensite @xcite as well as along interfaces with austenite @xcite .
moreover for lath martensite it has been observed in @xcite that _ `` twinning within a lath may be heavy . in any event , whenever an exact twin relationship was identified , it was found to be a result of twinning within a given lath and not of a twin relation between adjacent laths .
it is believed that the existence of heavily twinned local regions of laths , which may appear as separate laths in contrast images , may have caused some misinterpretation in earlier work on lath martensite . '' _ as in single shear theories , twinning of twins is macroscopically equivalent to a simple shear of a simply twinned system ( cf .
text below ) and thus the step from twins within twins is in analogy with the step from a single to a double shear theory .
the strength of the theory presented here is that it enables one to predict @xmath0 habit planes only assuming the lattice parameters of austenite and martensite .
this is particularly striking when compared to the double shear theory in @xcite where the _ `` calculation strategy was to select one of the possible @xmath11 systems and then perform calculations for the @xmath12 systems over a range of values of @xmath13 .
the sign of @xmath13 was selected by trial and error depending on whether the habit plane moved towards or away from @xmath0 . '' _
the theory proposed in this article is derived from the ball - james model @xcite - based on nonlinear elasticity and energy minimisation - and expands on previous work by ball & carstensen @xcite on the possibility of nonclassical austenite - martensite interfaces .
even though lath martensite is commonly associated with a high dislocation density , slip and plasticity the present model does not take such effects into account ( see also section [ secconcl ] ) .
interestingly , the ball - james model recovers the results of the phenomenological theory of martensite , as can be seen through a comparison of the derived formulae for the habit planes between twinned martensite and austenite ( cf .
* eq . ( 5.89 ) ) and ( * ? ? ?
* eq . ( 33)-(34 ) ) ) .
for a self - contained account of the phenomenological theory or of the ball - james model the reader is referred to the monographs @xcite and @xcite , respectively , both addressed to non - specialists . in the ball - james model , which neglects interfacial energy ,
microstructures are identified through minimising sequences @xmath14 , @xmath15 , for a total free energy of the form @xmath16 here , @xmath17 is a region representing the reference configuration of undistorted austenite at the transformation temperature and @xmath18 denotes the deformed position of particle @xmath19 in @xmath17 .
we remark that passing to the limit in these minimising sequences , corresponds in a very precise way to passing from a micro- to a macroscale , so that the limits themselves can be identified with the macroscopic deformations .
the energy density @xmath20 depends only on the deformation gradient @xmath21 , a 3@xmath223 matrix with positive determinant , and the temperature @xmath23 .
also @xmath24 is assumed frame indifferent , i.e. @xmath25 for all rotations @xmath26 - that is , for all 3@xmath223 matrices in @xmath27 and must respect the symmetry of the austenite , i.e. @xmath28 for all rotations @xmath29 leaving the austenite lattice invariant .
for cubic austenite there are precisely 24 such rotations . below the transformation temperature , @xmath24 is minimised on the set @xmath30 of martensitic energy wells , that is @xmath20 is minimal for @xmath31 .
the @xmath32 , positive - definite , symmetric matrices @xmath33 are the pure stretch components of the transformation strains mapping the parent to the product lattice . for example , in the case of fcc to bcc or fcc to bct , these are given by the three bain strains @xmath34 where @xmath35 and @xmath36 .
here @xmath37 is the lattice parameter of the fcc austenite and @xmath38 , @xmath39 are the lattice parameters of the bct martensite ( @xmath40 for bcc ) .
the notation @xmath41 , @xmath42 , @xmath43 has been chosen to emphasise that we are in the bain setting and to stay consistent with the literature .
we remark that the bain transformation @xcite is widely accepted as the transformation from fcc to bct / bcc requiring least atomic movement ; for a rigorous justification see @xcite . a convenient way to understand the relation between microstructures and minimising sequences is illustrated by the following example ( cf .
@xcite ) .
( austenite - twinned martensite interface)[example1 ] + suppose that a region of martensite is occupied by an array of twin related variants @xmath44 and @xmath45 with relative volume fractions @xmath46 and @xmath47 .
the strains @xmath44 and @xmath45 in general can not be invariant plane strains ( ips ) , equivalently , they can not form a fully coherent interface with austenite , represented in this model by the identity matrix @xmath48 . however , for specific volume fractions @xmath49 ( given by for the bain strain ) , the average deformation strain of the twinned region @xmath50 may indeed become an ips . in terms of the nonlinear elasticity model , this inability to form a fully coherent interface at the microscopic level , implies that the austenite - twinned martensite configuration can not exactly minimise the energy .
nevertheless , one can construct an energy minimising sequence @xmath14 , @xmath15 , with gradients @xmath51 as in figure [ fig : laminate ] .
the limit of this sequence is precisely the average strain / total shape deformation @xmath52 that is an ips .
although this average strain does not minimise the energy , it can be interpreted as a minimiser of a corresponding macroscopic energy . at the microscopic level
, one would observe some specific element of the above minimising sequence rather than the limit , due to having neglected interfacial energy .
the set of all matrices that can arise as limits of minimising sequences for the energy is referred to as the quasiconvex hull of the martensitic energy wells @xmath30 , denoted by @xmath53 .
hence , the set @xmath53 corresponds to all possible homogeneous total shape deformations that are energy minimising at the macroscopic level .
then , the requirement that ( up to an overall rotation @xmath26 ) a martensitic microstructure with a total shape deformation @xmath54 is a strain leaving the plane with normal @xmath55 invariant , amounts to finding a vector @xmath56 such that @xmath57 here , @xmath58 denotes the 3@xmath223 matrix @xmath59 . writing @xmath60 , yields the equivalent expression @xmath61 , implying that @xmath62 is the ips given by a shear on the plane with normal @xmath55 , with shearing direction @xmath63 and shearing magnitude @xmath64 . in particular , if the transformation from parent to product phase is volume - preserving
, @xmath65 is a simple shear , corresponding to the vectors @xmath56 and @xmath55 being perpendicular .
we note that equivalently says that @xmath62 can form a fully coherent planar interface with austenite of normal @xmath55 . also by frame - indifference
, the austenite can be represented by any rotation .
hence , a further rotation @xmath66 of the martensite results in @xmath67 , so that the rotated martensite can still form a fully coherent interface with austenite of the same normal @xmath55 . that is , rotations from the left can not change the habit plane normal .
a common feature in both the phenomenological theory and the ball - james model is to construct ( up to an overall rotation @xmath26 ) a total shape deformation @xmath65 that is an ips . in the literature ,
various algorithms have been proposed for the calculation of the corresponding elements of the shear , i.e. the magnitude , direction and normal ( cf . ) .
for example , see @xcite in the context of twinning / single shear theories and @xcite for double shear theories . in general , the problem of finding an overall rotation @xmath26 and shearing elements such that @xmath68 can be simplified by only considering the cauchy - green strain tensor @xmath69 and thus factoring out the overall rotation @xmath26 .
the following proposition ( ( * ? ? ?
* proposition 4 ) ) , allows one to calculate the shearing elements @xmath56 and @xmath55 in terms of the principal stretches and stretch vectors of @xmath65 .
the overall rotation @xmath26 can then be found by substituting @xmath56 and @xmath55 back into equation .
[ propbj ] let @xmath70 be a symmetric @xmath32 matrix with ordered eigenvalues @xmath71
. then @xmath72 can be written as @xmath73 for some @xmath74 if and only if @xmath75 and @xmath76 .
then , there are at most two solutions given by @xmath77 where @xmath78 is a normalisation constant , @xmath79 and @xmath80 are the ( normalised ) eigenvectors of @xmath72 corresponding to @xmath81 and @xmath82 . in the material science literature twins
are often described as two phases related by a specific @xmath83 degree rotation or , equivalently , a reflection . in the mathematical literature a twin is usually characterised by the existence of a rank - one connection between the two deformation strains @xmath44 , @xmath45 corresponding to the two phases , i.e. the existence of vectors @xmath84 and @xmath85 such that @xmath86 a fully coherent interface between the two phases is then given by the plane of normal @xmath85 .
this is because for any vector @xmath87 on that plane , i.e. @xmath88 , we obtain @xmath89 .
also , note that @xmath90 so that the lattice on the one side of the interface can be obtained by shearing the lattice on the other side along the twin plane @xmath91 , in the shearing direction @xmath92 with shearing magnitude @xmath93 .
the latter expression enables one to calculate the vectors @xmath84 and @xmath85 by proposition [ propbj ] through the identification @xmath94 , that is the relative deformation between the two phases is an ips . in view of single shear theories
, the above expression can equivalently be written as @xmath95 and thus @xmath45 can be obtained as a shear of the parent lattice , followed by @xmath44 .
hence , in the case of twins between two martensitic energy wells @xmath96 and @xmath97 one needs to solve the equation @xmath98 for the rotation matrix @xmath66 and the twinning elements @xmath84 and @xmath85 .
if the transformation strains @xmath33 and @xmath99 are related by a @xmath83 rotation , this calculation simplifies significantly by mallard s law ( see ( * ? ? ?
* result 5.2 ) or below ) .
in particular , this assumption holds for @xmath100 and @xmath101 , i.e. for the bain transformation from fcc to bct / bcc .
( mallard s law)[lemmamal ] + let @xmath102 and @xmath103 satisfy @xmath104 for some @xmath83 rotation @xmath29 about a unit vector @xmath105 , i.e. @xmath106 . then the equation @xmath107 admits two solutions given by @xmath108 in each case , @xmath109 .
we conclude this interlude by remarking that twins described by the first solution in mallard s law are _ type i _ twins and the corresponding lattices are related by a 180@xmath110 rotation about the twin plane @xmath111 .
the second solution in mallard s law describes _ type ii _
twins and the lattices are related by a 180@xmath110 rotation about the shearing direction @xmath92 .
it may happen , and it does for the bain strains , that there are two rotations by 180@xmath110 relating @xmath102 and @xmath103 . in this case
, there are seemingly four solutions from mallard s law , however , proposition [ propbj ] says that there can not be more than two .
indeed , the type i solution using one 180@xmath110 rotation is the same as the type ii solution using the other 180@xmath110 rotation and vice versa .
in particular , the lattices on either side of the interface are related by both a 180@xmath110 rotation about the twin plane and a 180@xmath110 rotation about the shearing direction .
solutions of this type are _ compound _ twins .
henceforth , we only consider the bain strains @xmath41 , @xmath42 and @xmath43 for the fcc to bct / bcc transformation in steel .
in single shear theories the total shape deformation is assumed to be decomposable into @xmath112 where @xmath26 is a rotation , @xmath113 is one of the bain strains and @xmath114 is a shear whose specific form varies in the literature . in the ball - james theory ,
the total shape deformation @xmath65 must be macroscopically energy minimising , thus restricting the form of the shear @xmath115 .
the most important case is when @xmath115 arises from twinning . as in example
[ example1 ] , the average strain corresponding to a twinning system between @xmath116 and @xmath117 , satisfying , with volume fractions @xmath46 and @xmath47 , respectively , is given by @xmath118 , for any @xmath119 , where the elements @xmath66 , @xmath84 and @xmath85 can be calculated by mallard s law ( cf .
proposition [ lemmamal ] ) applied to @xmath120 with either @xmath121 or @xmath122 , i.e. the resulting twins are compound . by simple algebraic manipulation
, the average strain can be written as @xmath123 i.e. a single shear @xmath124 , with @xmath125 , of the parent lattice followed by the bain strain @xmath41 . by proposition [ propbj ] , to make the total shape deformation @xmath126 an ips , the volume fraction @xmath127 needs to be chosen such that the middle eigenvalue of @xmath128 is equal to one .
in particular , since one of the eigenvalues must be made equal to one , the expression @xmath129 must vanish , giving rise to the two solutions @xmath49 and @xmath130 , where @xmath131 it is important to then check that it is indeed the middle eigenvalue that is equal to one . for each of the values @xmath132 , @xmath133 ,
we can calculate two habit plane normals according to proposition [ propbj ] .
one of these normals is , up to normalisation , given by @xmath134 , where @xmath135 by considering the remaining normals and all possible pairs of twin related bain strains , one recovers the entire family of normals @xmath136 . in figure [ figsimplnormals ] , the components of one of these normals are plotted for a typical range of lattice parameters @xmath137 , @xmath138 in the fcc to bct / bcc transformation .
we immediately note that for @xmath139 and @xmath140 the predicted habit plane normal arising from simple twinning is almost exactly @xmath141 and for @xmath142 and @xmath140 the habit plane normal is almost exactly @xmath143
. the corresponding ratios of tetragonality are given by @xmath144 and @xmath145 respectively .
this in excellent agreement with the observations in e.g. @xcite of @xmath143 habit planes in steel with carbon content in the range @xmath146 wt-@xmath147 as well as the theoretical and experimental results in e.g. @xcite , @xcite and @xcite of @xmath148 habit planes in highly tetragonal martensite . similarly , in double shear theories the total shape deformation is assumed to be decomposable into @xmath149 where @xmath26 is a rotation , @xmath113 is one of the bain strains and @xmath12 , @xmath11 are two shears . above
we have seen how twinning can be regarded as an instance of a single shear theory . in this section
we show how an additional level of twinning , i.e. twins within twins , results in an instance of a double shear theory , consistent with the ball - james model .
to visualise this type of microstructure , we revisit the construction in example [ example1 ] to construct _ two _ twinning systems one with @xmath116 , @xmath117 and average strain @xmath150 and another one with say @xmath116 , @xmath151 and average strain @xmath152 . for each @xmath127 there
exists a rotation @xmath153 such that @xmath154 and thus the two twinning systems are macroscopically compatible with a fully coherent interface of normal @xmath155 between them ( cf .
figure [ figdouble ] ) .
the elements @xmath153 , @xmath156 and @xmath155 can be calculated by mallard s law ( cf .
proposition [ lemmamal ] ) applied to @xmath157 with @xmath158 , giving rise to a type i and a type ii solution . unlike in the single shear theory ,
these twins are not compound and therefore we distinguish these two solutions by the superscript @xmath159 , @xmath160 for type i and type ii respectively . finally , it can be shown that the volume fractions @xmath127 in each of the twinned regions must necessarily coincide . as in example
[ example1 ] , we can construct an array of twins between the two twinned regions with respective volume fractions @xmath161 and @xmath162 and average strain given by @xmath163 simple algebraic manipulation allows one to write as @xmath164 where @xmath165 and @xmath166 and thus an instance of a double shear theory .
we note that @xmath167 . by proposition [ propbj ] , in order to make the total shape deformation @xmath168 an ips , the volume fractions @xmath127 , @xmath169 need to be chosen such that the middle eigenvalue of @xmath170 is equal to one . solving @xmath171 for each fixed @xmath127 ,
gives rise to a quadratic equation in @xmath169 for each choice of @xmath159 , which can be solved explicitly ( see @xcite for the full details ) .
the expressions are lengthy and we refer to figure [ figdoublelam ] which visualises the dependence @xmath172 for the volume - preserving bain strain .
the endpoints of these curves correspond to the vanishing of one of the twinning systems ( cf . ) and hence to the collapse of the system of twins within twins , to a simple twinning system with volume fractions given by .
the figure remains qualitatively the same for typical lattice parameters .
of @xmath41 , @xmath173 of @xmath42 and @xmath174 of @xmath43 that make the twins within twins an ips for @xmath175 and @xmath176 . ] for each @xmath159 and each admissible pair @xmath177 with corresponding strain @xmath178 , we can calculate two one - parameter families of habit plane normals through proposition [ propbj ] . by considering all possible combinations of twinning systems in our construction ,
we then obtain all crystallographically equivalent normals .
for the volume - preserving bain strain , the habit plane normals that can arise from the family @xmath179 are visualised in figure [ figfirstsolline ] .
however , due to algebraic complexity , it is difficult to write down a formula for the habit plane normals with an explicit dependence on @xmath180 , @xmath181 and @xmath127 .
it is natural to assume that the observed total shape deformation @xmath65 requires small overall atomic movement ( see also @xcite ) relative to the parent phase of austenite .
a measure of this distance is the _ _ strain energy _ _ which yields the same results . ] given by @xmath182 where @xmath183 denotes the frobenius norm and @xmath184 the principal stretches of @xmath65 .
it can be shown that any microstructure with small strain energy must necessarily involve all three bain variants in roughly similar volume fractions .
in particular , this can not be the case for an array of twin related variants and we ought to consider at least twins within twins ( see also fig .
[ figdis ] ) .
although , introducing even further levels of twinning can reduce the strain energy , one could argue that interfacial energy contributions , which are not accounted for in this model , may inhibit such behaviour .
combining the fact that @xmath10 can not result from simple twinning ( cf .
[ figsimplnormals ] ) and that twins within twins are preferable in terms of strain energy , we build a theory that predicts @xmath10 habit plane normals solely based on energy minimisation and geometric compatibility .
firstly , the one - parameter families of habit plane normals obtained from twins within twins ( see section [ secdoubleshear ] ) , contain normals very close to any @xmath10 .
this is at least the case for lattice parameters close to @xmath185 , @xmath186 corresponding to a volume - preserving transformation from fcc to bcc .
this regime of parameters is suitable since @xmath10 habit planes are observed in low - carbon steels where the transformation is very nearly fcc to bcc .
the resulting one - parameter families of habit plane normals , along with their crystallographically equivalent ones , are shown in figure [ fignorm ] .
we stress that the only free parameter in the generation of these normals is @xmath127 which fixes the choice of the shearing systems , based only on the energy minimising property of the microstructure .
secondly , out of these one - parameter families of normals , our theory can identify the @xmath10 habit plane normals as those satisfying a criterion of maximal compatibility . to this end , revisiting our construction of twins within twins there is a choice ( cf . ) of using either @xmath187 as an average strain , corresponding to the type i solution from mallard s law , or @xmath188 corresponding to type ii .
figure [ figdis ] shows the strain energy associated with the two macroscopic strains as a function of @xmath127 .
it is clear that the strain energy of @xmath187 is significantly smaller than that of @xmath188 and is thus preferable .
we also note that , in agreement with the previous section , the strain energies of both @xmath189 and @xmath190 increase rapidly as the volume fraction of @xmath41 approaches @xmath191 or @xmath192 , that is as the microstructure reduces to a single twinning system .
further , we remark that the @xmath188 with minimal strain energy result in habit plane normals which are very nearly @xmath193 ( see also fig . [ fignorm ] ) . nevertheless , the strain energies of any of the @xmath187 that give rise to @xmath10 normals are lower ( cf .
[ figdis ] ) . and @xmath176 .
] even though , the strains resulting in @xmath10 habit plane normals do not minimise the strain energy , they satisfy a strong criterion of compatibility . to understand
this one must think in terms of the dynamic process of nucleation .
as austenite is rapidly quenched , the martensite phase nucleates at various sites .
the strain in a given nucleation site may need to be an ips but otherwise , has no reason to be the same as the strain in any other site . nevertheless , there are essentially only three distinct families of systems of twins within twins and these can be classified by the bain variant which is present in both of the simple twinning systems that comprise the overall microstructure . in figure
[ fignorm ] , these three families are distinguished by colour . as the nuclei grow and approach other nuclei , they need to remain compatible with each other .
remarkably , the only habit plane normals that arise from deformations with low strain energy and can be reached by all three families are @xmath10 . in figure
[ fignorm ] , this can be seen from the fact that all differently coloured curves intersect close to @xmath10 .
for any two such regions of twins within twins with corresponding average strains @xmath194 and @xmath195 , one can see that @xmath196 implying that they can meet along a fully coherent planar interface of normal @xmath0 .
of course , any nucleus interacts with its neighbours faster than it does with distant nuclei . as a result
, blocks of similarly oriented regions of twins within twins ( laths ) may form whose overall orientation may differ from that of other blocks .
the theory proposed here for the prediction of @xmath0 habit plane normals has two possible interpretations . on the one hand , it can be seen as a purely macroscopic theory . in particular , it is an instance of a double shear theory with a precise algorithm to produce the required shears based on energy minimisation , without the need for any further assumptions .
all possible habit plane normals that can arise from the one parameter family ( indexed by @xmath127 ) of macroscopic deformations @xmath179 are shown in figure [ figfirstsolline ] .
table [ tableshear ] then lists the elements of the twinning systems for the values of @xmath127 that produce a near @xmath0 habit plane . with the help of it
is easy to convert between the twinning and shearing systems and thus compute the elements @xmath12 and @xmath11 required in a double shear theory . at this macroscopic level
it is not possible to distinguish between twins within twins , a single twin and one slip system , and a single variant and two slip systems . on the other hand , a physical mechanism for the formation of @xmath0 habit plane normals
is proposed and thus a specific morphology on a microscopic level . according to this interpretation , each lath
may itself be a region of twins within twins with a corresponding lath boundary of normal @xmath0 .
this type of morphology is depicted in figure [ figdouble ] with @xmath197 being a @xmath10 normal and the other elements are as in table [ tableshear ] .
this morphology is a direct consequence of the underlying theory and it would be very interesting if it could be put to experimental scrutiny .
l*4c @xmath127&@xmath198 & @xmath199 & @xmath200 + @xmath84&@xmath201 $ ] & @xmath201 $ ] & @xmath201 $ ] + @xmath85&@xmath202 & @xmath202 & @xmath202 + @xmath203&@xmath204 & @xmath205 & @xmath206 + @xmath156 & @xmath207 $ ] & @xmath208 $ ] & @xmath209 $ ] + @xmath55 & @xmath210 & @xmath210 & @xmath210
the research of a. m. leading to these results has received funding from the european research council under the european union s seventh framework programme ( fp7/2007 - 2013 ) / erc grant agreement n@xmath211 . | a mathematical framework is proposed to predict the features of the ( 557 ) lath transformation in low - carbon steels based on energy minimisation .
this theory generates a one - parameter family of possible habit planes and a selection mechanism then identifies the ( 557 ) normals as those arising from a deformation with small atomic movement and maximal compatibility .
while the calculations bear some resemblance to those of double shear theories , the assumptions and conclusions are different .
interestingly , the predicted microstructure morphology resembles that of plate martensite , in the sense that a type of twinning mechanism is involved .
msc ( 2010 ) : 74a50 , 74n05 , 74n15 keywords : lath martensite , microstructure , twins within twins , ( 557 ) habit planes , double shear theories , energy minimisation , non - classical interfaces maketitle 2em 1.5em .5em [ cols="^ " , ] 1em date 1.5em |
words coding exchange of three intervals represent one of possible generalizations of sturmian words to a ternary alphabet .
an exchange of three intervals is given by a permutation @xmath0 on the set @xmath1 , and a triplet of positive numbers @xmath2 , @xmath3 , @xmath4 , corresponding to lengths of intervals @xmath5 , @xmath6 , @xmath7 , respectively , which define a division of the interval @xmath8 .
in this paper we study infinite words coding exchange of three intervals with the permutation @xmath9 .
such words are called here 3iet words .
properties of 3iet words have been studied from various points of view in papers @xcite .
recently , articles @xcite and @xcite gave a characterization of 3iet words invariant under a substitution .
recall that a similar question for sturmian words ( i.e. words coding exchange of two intervals ) has been partially solved in @xcite .
complete solution to the task was provided by yasutomi @xcite . an alternative proof valid for bidirectional sturmian words
is given in @xcite , yet another proof in @xcite .
one has also asked the question from another angle : what are the substitutions fixing a sturmian word , this problem has been studied in a wider context .
one considers the so - called sturmian morphisms , i.e. morphisms that preserve the set of sturmian words .
the monoid of sturmian morphisms has been described in @xcite .
it turns out that it is generated by three simple morphisms , namely @xmath10 it is known @xcite that a morphism @xmath11 such that @xmath12 is sturmian for at least one sturmian word @xmath13 belongs also to the monoid .
in particular , all morphisms fixing sturmian words are sturmian morphisms .
the aim of this paper is to describe morphisms over the alphabet @xmath14 fixing a 3iet word .
the main tool which we use is a narrow connection between 3iet words and sturmian words over the alphabet @xmath15 by means of morphisms @xmath16 given by @xmath17 in @xcite the following statement is proved .
[ thm:1 ] a ternary word @xmath13 is a 3iet word if and only if both @xmath18 and @xmath19 are sturmian words .
another important statement connecting 3iet words and sturmian words is taken from @xcite .
[ thm:2 ] a non - degenerate 3iet word @xmath13 is invariant under a substitution if and only if both @xmath18 and @xmath19 are invariant under substitution .
the paper is organized as follows . in section [ sec : preli ] we recall the definitions of 3iet words and morphisms and the geometric representation of a fixed point of a morphism . in section [ sec : amicable ] we define a relation on the set of sturmian morphisms with a given incidence matrix , called amicability , and we show how to construct from a pair of amicable morphisms a morphism over the alphabet @xmath14 with a 3iet fixed point ( theorem [ thm : xy ] ) . in section [ sec : ternarizace ] we show , that any morphism @xmath20 fixing a non - degenerate 3iet word ( or its square @xmath21 ) is constructed in this way ( theorem [ thm:001 ] ) .
a transformation @xmath22 of an exchange of three intervals is usually defined as a mapping with the domain @xmath23 , where @xmath24 are arbitrary positive numbers determining the splitting of @xmath8 into three disjoint subintervals @xmath25 .
an infinite word associated to such a transformation is given as a coding of an initial point @xmath26 in a ternary alphabet @xmath14 .
properties of the transformation @xmath27 and the corresponding infinite word do not depend on absolute values of @xmath24 , but rather on their relative sizes . as well , translation of the interval @xmath8 on the real line does not influence the corresponding dynamical system . for the study of substitution properties of 3iet words it proved useful to consider the definition of a 3iet mapping with parameters normalized by @xmath28 and a translation of the interval @xmath8 such that the initial point @xmath29 is the origin .
let @xmath30 be real numbers satisfying @xmath31 the mapping @xmath32 is called exchange of three intervals with permutation @xmath9 .
note that the parameter @xmath33 represents the length of the interval @xmath34 , and @xmath35 corresponds to the length of @xmath36 .
the number @xmath37 is the length of the interval @xmath25 .
the orbit of the point @xmath38 under the transformation @xmath27 of can be coded by an infinite word @xmath39 in the alphabet @xmath14 , where @xmath40 the infinite word @xmath39 is non - periodic exactly in the case that the parameter @xmath33 is irrational .
words coding the orbit of @xmath41 under an exchange of intervals with the permutation @xmath9 and an irrational parameter @xmath33 are called 3iet words . an alphabet @xmath42 is a finite set of symbols . in this paper
we shall systematically use the alphabet @xmath14 for 3iet words , and the alphabet @xmath15 for sturmian words .
a finite word in the alphabet @xmath42 is a concatenation @xmath43 , where @xmath44 for all @xmath45 .
the length of the word @xmath46 is denoted by @xmath47 .
the symbol @xmath48 denotes the set of all finite words over @xmath42 , including the empty word @xmath49 .
equipped with the operation of concatenation , @xmath48 is a monoid .
sequences @xmath50 ,
@xmath51 , @xmath52 are called right - sided , left - sided and bidirectional infinite word , respectively . if for a finite word @xmath53 there exist ( finite or infinite ) words @xmath54 and @xmath55 such that @xmath56 , then @xmath53 is said to be a factor of the ( finite or infinite ) word @xmath46 .
if @xmath54 is the empty word , then @xmath53 is a prefix of @xmath46 , if @xmath57 , then @xmath53 is a suffix of @xmath46 .
the set of all factors of an infinite word @xmath13 is called the language of @xmath13 and denoted @xmath58 .
factors of @xmath13 of length @xmath59 form the set @xmath60 ; obviously @xmath61 .
the mapping @xmath62 given by the prescription @xmath63 is called the factor complexity of the infinite word @xmath13 .
infinite words @xmath13 such that the set @xmath64 is finite for every @xmath65 are called uniformly recurrent .
right - sided sturmian words are defined as right - sided infinite words with factor complexity @xmath66 for all @xmath67 .
bidirectional sturmian words are uniformly recurrent bidirectional infinite words satisfying @xmath66 for all @xmath67 . for the factor complexity @xmath68 of a 3iet word it holds that 1 .
either @xmath69 for all sufficiently large @xmath59 , 2 . or @xmath70 for all @xmath67 .
3iet words with complexity @xmath69 belong to the set of the so - called quasisturmian words , which are images of sturmian words under suitable morphisms .
3iet words with complexity @xmath70 are called non - degenerate 3iet words or regular 3iet words .
the factor complexity of a 3iet word is given by ( i ) or ( ii ) according to the parameters @xmath71 : a 3iet word is non - degenerate if and only if @xmath72:={{\mathbb z}}+\varepsilon{{\mathbb z}}$ ] , see @xcite . a mapping @xmath73 satisfying @xmath74 for all @xmath75 is called a morphism . a morphism is uniquely determined by the images @xmath76 of all letters @xmath77 .
the action of a morphism can be naturally extended to infinite words by @xmath78 with every morphism @xmath11 one can associate a matrix @xmath79 .
the matrix has @xmath80 rows and @xmath81 columns , and @xmath82 a morphism @xmath83 is called primitive if some power of the square matrix @xmath79 has all elements positive . in other words , there exists a positive integer @xmath84 such that for all @xmath85 , the letter @xmath86 is a factor of the @xmath84-th iteration @xmath87 .
an infinite word @xmath13 in @xmath88 , @xmath89 , @xmath90 is said to be a fixed point of a morphism @xmath83 , if @xmath91 .
it is obvious that if @xmath92 is a fixed point of a primitive morphism @xmath11 , then @xmath93 for a non - empty word @xmath53 , and @xmath13 is the limit of finite words @xmath94 , which is usually denoted by @xmath95 .
analogous properties must be satisfied by primitive morphisms fixing left - sided or bidirectional infinite words .
morphisms with the above properties are sometimes called substitutions .
it is quite obvious that the only non - primitive morphism which can fix a 3iet word or a sturmian word is the identity .
therefore it is not misleading not to distinguish between notions of primitive morphism and substitution when speaking about substitution invariant sturmian or 3iet words .
substitution invariance of non - degenerate bidirectional 3iet words has been studied in @xcite .
similarly as in the case of sturmian words , one needs the notion of sturm numbers . the original definition of a sturm number uses continued fractions .
we cite the equivalent definition given in @xcite : a real number @xmath96 is called a sturm number , if it is a quadratic irrational with algebraic conjugate @xmath97 .
let us cite here the characterization of substitution invariant 3iet words from @xcite .
[ thm:7 ] let @xmath13 be a non - degenerate 3iet word with parameters @xmath30 .
then @xmath13 is invariant under a primitive morphism if and only if * @xmath33 is a sturm number * @xmath98 * @xmath99 and @xmath100 , where @xmath101 is the field conjugate of @xmath102 in @xmath103 .
it is useful to reformulate the task of searching for a substitution fixing a given infinite word @xmath104 in geometric terms .
let us associate with letters of the alphabet mutually distinct lengths by an injective mapping @xmath105 . then ,
with the infinite word @xmath13 we associate a strictly increasing sequence @xmath106 such that @xmath107 a number @xmath108 satisfying @xmath109 is called a self - similarity factor of the sequence @xmath106 .
let us suppose that the assignment of lengths @xmath110 and the self - similarity factor @xmath111 satisfy that to every @xmath77 there exists a finite set @xmath112 such that @xmath113\cap \sigma \ = \
\lambda t_n + p_a\qquad\text { for all $ n\in{{\mathbb z}}$ with $ u_n = a$.}\ ] ] it means that the gap between @xmath114 and @xmath115 is after stretching by @xmath111 filled by members of the original sequence @xmath106 in the same way for all gaps corresponding to the letter @xmath86 . an infinite word @xmath13 for which one can find a mapping @xmath110 and a factor @xmath111 with the above described properties is obviously invariant under a substitution @xmath11 , where the image @xmath76 is determined by the distances between consecutive elements of the set @xmath116 .
we call the set @xmath117 with the property the geometric representation of the word @xmath13 with the factor @xmath111 . on the other hand ,
if an infinite word @xmath13 is invariant under a primitive substitution @xmath11 with the matrix @xmath79 , then the eigenvector of @xmath79 corresponding to the dominant eigenvalue @xmath111 is a column of length @xmath80 with all components @xmath118 , @xmath77 , positive , cf .
the correspondence @xmath119 results in a sequence @xmath106 having @xmath111 as its self - similarity factor and satisfying .
therefore the set @xmath120 is the geometric representation of the infinite word @xmath13 with the factor @xmath111 .
we illustrate the concept of the geometric representation in figure [ f ] . fixed by a substitution @xmath20 with the self - similarity factor @xmath111 . in our example , @xmath121 , @xmath122 , @xmath123 .
] in @xcite , the authors derive ( in their corollaries 7.1 and 7.2 ) several properties of matrices of substitutions fixing a 3iet word .
[ thm:8 ] let @xmath13 be a non - degenerate 3iet word with parameters @xmath30 which is invariant under a primitive substitution @xmath20 .
then for the dominant eigenvalue @xmath111 of the matrix @xmath124 one has 1 .
@xmath111 is a quadratic unit ; 2 .
@xmath125 is the right eigenvector of @xmath124 , corresponding to @xmath126 the algebraic conjugate of @xmath111
. item ( 2 ) of the above theorem implies for the matrix @xmath127 that @xmath128 is its right eigenvector corresponding to the dominant eigenvalue @xmath111 . using theorem [ thm:7 ] ,
the parameter @xmath33 is a sturm number , and so @xmath97 .
the vector @xmath129 has thus all components positive or all components negative . in any case , in the geometric representation of the fixed point of the substitution @xmath20 of theorem [ thm:8 ] ,
the length @xmath130 corresponding to the letter @xmath131 is the sum @xmath132 .
the narrow connection of @xmath133iet words and sturmian words and their invariance under morphisms is described in theorems [ thm:1 ] and [ thm:2 ] by means of morphisms @xmath134 , see .
these morphisms also allow us the description of morphisms fixing a 3iet word using sturmian morphisms . for that , several notions need to be defined .
[ d : amicablewords ] let @xmath135 be finite or infinite words over the alphabet @xmath15 .
we say that @xmath13 is amicable to @xmath46 , and denote it by @xmath136 , if there exist a ternary word @xmath53 over @xmath14 such that @xmath137 and @xmath138 .
in such a case we denote @xmath139 and say that @xmath53 is the ternarization of @xmath13 and @xmath46 .
note that the relation @xmath140 is not symmetric .
for example , @xmath141 is amicable to @xmath142 , but not vice versa .
it is also interesting to notice that if two finite words @xmath135 satisfy @xmath136 , then they are of the same length and the number of letters @xmath86 in @xmath13 and @xmath46 are equal for both @xmath143 .
figure [ fff ] illustrates an easy way how to recognize amicability of two words and how to construct their ternarization . according to the definition ,
@xmath136 if @xmath13 can be written as a concatenation @xmath144 and @xmath46 as a concatenation @xmath145 such that for all @xmath146 we have either @xmath147 or @xmath148 or @xmath149 and @xmath150 .
the ternarization @xmath53 is then constructed by associating letters in the alphabet @xmath14 to the blocks , namely it associates @xmath151 , if @xmath147 ; it gives @xmath152 if @xmath148 , and it gives @xmath131 , if @xmath149 and @xmath150 .
@xmath153 w = \quad \end{gathered } { \begin{gathered}\boxed{\:\begin{gathered}0 \\ 0\end{gathered}\:}\\a\end{gathered}}\ { \begin{gathered}\boxed{\:\begin{gathered}1 \\ 1\end{gathered}\:}\\c\end{gathered}}\ { \begin{gathered}\boxed{\:\begin{gathered}0 \\ 0\end{gathered}\:}\\a\end{gathered}}\ { \begin{gathered}\boxed{\:\begin{gathered}0\ 1 \\
1\ 0\end{gathered}\:}\\b\end{gathered}}\ { \begin{gathered}\boxed{\:\begin{gathered}0 \\ 0\end{gathered}\:}\\a\end{gathered}}\ { \begin{gathered}\boxed{\:\begin{gathered}1 \\ 1\end{gathered}\:}\\c\end{gathered}}\ ] ] we introduce the notion of amicability and ternarization also for morphisms .
[ d : amicablemorphisms ] let @xmath154 be two morphisms .
we say that @xmath155 is amicable to @xmath156 , and denote it by @xmath157 , if the three following relations hold @xmath158 the morphism @xmath159 given by @xmath160 is called the ternarization of @xmath155 and @xmath156 and denoted by @xmath161 . as an example , consider two basic sturmian morphisms @xmath155 , @xmath156 from , @xmath162 it can be easily checked that @xmath157 and that their ternarization @xmath163 is of the form @xmath164 from the definition of amicability of words it follows that if @xmath136 and @xmath165 then for their concatenation we have @xmath166 . as a simple consequence of this idea , we have the following lemma . [
lem : amicable - induction ] let @xmath135 be two ( finite or infinite ) words over @xmath15 such that @xmath136 , and let @xmath154 be two morphisms such that @xmath157
. then @xmath167
. moreover , if @xmath168 , then @xmath169 , where @xmath163 . note that if @xmath157 and @xmath163 , then @xmath170 [ thm : xy ] let @xmath154 be two primitive sturmian morphisms having fixed points such that @xmath157 .
then the morphism @xmath171 given by @xmath163 has a 3iet fixed point .
the first step is to prove that a fixed point of @xmath155 , say @xmath13 , is amicable to a fixed point of @xmath156 , say @xmath46 .
we prove the statement for right - sided words only , the proof for left - sided and bidirectional fixed points follows the same lines .
we will discuss two separate cases .
_ case a. _ let there exists a letter @xmath172 such that @xmath173 starts with @xmath174 and @xmath175 starts with @xmath174 .
primitivity of @xmath155 and @xmath156 implies that both @xmath173 and @xmath175 have at least two letters .
therefore @xmath176 since @xmath177 we have @xmath136 by lemma [ lem : amicable - induction ] .
_ case b. _ let the negation of case a hold . a. let @xmath178 start with @xmath179
. then necessarily @xmath180 starts with @xmath41 which is in contradiction with @xmath181 .
b. let @xmath178 start with @xmath41 .
since @xmath155 has a fixed point , @xmath182 must start with @xmath41 .
thus @xmath183 does not start with @xmath41 , which implies that @xmath180 starts with @xmath179 since @xmath156 also has a fixed point .
+ consider @xmath184 and @xmath185 .
clearly , @xmath186 starts with @xmath41 and @xmath187 starts with @xmath179 .
moreover , since @xmath188 , the word @xmath184 must have the prefix @xmath189 and the word @xmath185 must have the prefix @xmath190 .
therefore @xmath191 and @xmath192 . now @xmath193 and therefore by lemma [ lem : amicable - induction ]
, it follows that @xmath136 .
we have shown in all cases that the fixed points @xmath135 of the sturmian morphisms @xmath157 satisfy @xmath136 .
moreover , lemma [ lem : amicable - induction ] implies that if @xmath168 , then @xmath194 , i.e. @xmath53 is the fixed point of the ternarization of @xmath155 and @xmath156 .
but since @xmath195 , @xmath196 are fixed points of primitive sturmian morphisms , they are sturmian words , and therefore the infinite word @xmath53 must be a 3iet word , as follows from theorem [ thm:1 ] .
the aim of this section is to prove the following theorem .
[ thm:001 ] let @xmath20 be a primitive substitution fixing a non - degenerate 3iet word @xmath13 .
then there exist sturmian morphisms @xmath155 and @xmath156 having fixed points , such that @xmath157 and @xmath20 or @xmath21 is equal to @xmath197 .
the proof will combine results of papers @xcite concerning substitution invariance of non - degenerate 3iet words and of the paper @xcite which solves the same question for sturmian words .
we shall study infinite words defined by under a transformation @xmath27 from where parameters @xmath71 satisfy additional conditions @xmath198={{\mathbb z}}+\varepsilon{{\mathbb z}}\,.\ ] ] these conditions guarantee that the corresponding infinite 3iet word is non - degenerate . according to theorem [ thm:1 ] ,
the images of a 3iet word @xmath13 under morphisms @xmath199 , @xmath200 are sturmian words .
let us determine parameters of the sturmian words @xmath18 , @xmath19 ( i.e. the corresponding exchanges of two intervals ) , provided that the parameters of @xmath13 are @xmath30 . the procedure is illustrated in figure [ ff ] . and sturmian words @xmath18 , @xmath19 . ]
define the mapping @xmath201 by @xmath202 comparing @xmath203 and @xmath27 we obtain ( see figure [ ff ] ) @xmath204 for @xmath205 we have @xmath206 therefore @xmath18 is the infinite word coding the orbit of 0 under the exchange @xmath203 of intervals with lengths @xmath33 and @xmath35 .
such a word is a sturmian word of the slope @xmath33 and intercept @xmath207 ( i.e. the distance of the initial point of the orbit and the left end - point of the interval @xmath208 which is the domain of @xmath203 ) . in a similar way
, we derive that the infinite word @xmath19 is the coding of the orbit of 0 under the exchange of two intervals @xmath209 .
in particular , it is a sturmian word of the slope @xmath33 and intercept @xmath210 .
let us cite the result characterizing substitution invariant sturmian words .
comparing @xcite and @xcite we obtain that a right - sided sturmian word with the slope @xmath2 and intercept @xmath3 is substitution invariant if and only if the bidirectional sturmian word with the same slope and intercept is substitution invariant .
[ thm:9 ] let @xmath211 be irrational and @xmath212 .
a sturmian word with the slope @xmath2 and intercept @xmath3 is invariant under a primitive morphism if and only if 1 .
@xmath2 is a sturm number ; 2 .
@xmath213 ; 3 .
@xmath214 , where @xmath215 , @xmath216 denote the field conjugates of @xmath2 , @xmath3 in @xmath217 . note that the inequalities in item ( 3 ) are satisfied for @xmath216 if and only they are satisfied replacing @xmath216 by @xmath218 .
knowing the slope and intercept of sturmian words @xmath18 , @xmath19 we can deduce from theorem [ thm:7 ] the statement of theorem [ thm:1 ] , namely that a non - degenerate 3iet word is invariant under a primitive substitution if and only if both sturmian words @xmath18 , @xmath19 are substitution invariant
. we will now put into relation the substitutions fixing infinite words @xmath13 , @xmath18 , and @xmath19 .
first we consider the self - similarity factors and geometric representations of these substitutions .
[ l:9 ] let @xmath20 be a primitive substitution over the alphabet @xmath14 having as its fixed point a non - degenerate 3iet word @xmath13 .
let us denote its parameters @xmath30 .
denote by @xmath111 the dominant eigenvalue of the matrix @xmath127 and by @xmath219 its positive right eigenvector corresponding to @xmath111 .
if @xmath220 , then there exist substitutions @xmath221 fixing @xmath18 , @xmath19 , respectively , and such that @xmath111 is the dominant eigenvalue of @xmath222 and @xmath223 , and @xmath224 is their common right eigenvector corresponding to @xmath111 .
moreover , @xmath225 .
theorems [ thm:7 ] and [ thm:8 ] imply that @xmath226 is a right eigenvector of @xmath127 corresponding to @xmath111 . recall that @xmath18 is the sturmian word of the slope @xmath33 and intercept @xmath207 , and @xmath19 the sturmian word of the slope @xmath33 and intercept @xmath210 . by theorem [ thm:2 ]
they are invariant under substitutions , say @xmath227 , @xmath156 . since @xmath33 and @xmath35 are densities of letters @xmath41 and @xmath179 respectively , the substitution matrices @xmath222 and @xmath223 must have the eigenvector @xmath228 . obviously @xmath225 .
it remains to show that @xmath229 can be chosen so that the dominant eigenvalue of @xmath20 , i.e. @xmath111 , is also the dominant eigenvalue of @xmath222 , @xmath223 . as a consequence of the equality @xmath230 and the fact that @xmath111 is a quadratic unit
, we have @xmath231 = \lambda { { \mathbb z}}[\varepsilon ' ] = \lambda'{{\mathbb z}}[\varepsilon']$ ] , which , after conjugation , gives @xmath232 = \lambda ' { { \mathbb z}}[\varepsilon ] = \lambda{{\mathbb z}}[\varepsilon]\,.\ ] ] proposition 5.6 of @xcite ( see also remarks 5.7 and 6.5 ibidem ) implies that @xmath233 $ ] and @xmath234 $ ] .
this , together with gives @xmath235 ) & = c+{{\mathbb z}}[\varepsilon]\,,\\ \lambda'(c+l-1+\varepsilon+{{\mathbb z}}[\varepsilon])&= c+l-1+\varepsilon + { { \mathbb z}}[\varepsilon]\ , . \end{split } \ ] ] note that the assumption @xmath220 is required in order that we can use results from @xcite .
realize that substitution invariance of @xmath18 and @xmath19 implies by theorem [ thm:9 ] that their parameters satisfy @xmath236 these inequalities , together with , already imply that there exist substitutions @xmath155 and @xmath156 with factor @xmath111 ( see proof of proposition 5.3 in @xcite ) .
the dominant eigenvalue of the matrix @xmath127 is a quadratic unit @xmath111 .
if @xmath220 , we shall prove the statement for @xmath20 .
if @xmath237 , we will consider the second iteration @xmath21 .
therefore we consider without loss of generality @xmath220 . with the help of geometric representation of infinite words we will show that morphisms @xmath229 found by lemma [ l:9 ] are amicable , i.e. @xmath157 , and that @xmath20 is their ternarization .
we use the fact that all of the considered substitutions , @xmath238 and @xmath156 have the same factor @xmath111 .
the idea of the proof is illustrated in figure [ ffff ] .
ternary substitution @xmath239 , @xmath240 , @xmath241 and its fixed point @xmath13 , @xmath18 and @xmath19 , and the substitutions @xmath20 , @xmath155 , @xmath156 ( all with the same self - similarity factor @xmath111 ) fixing them .
we have @xmath242 and @xmath163.,title="fig : " ] 0.5 cm sturmian substitution @xmath243 and its fixed point @xmath18 , @xmath18 and @xmath19 , and the substitutions @xmath20 , @xmath155 , @xmath156 ( all with the same self - similarity factor @xmath111 ) fixing them .
we have @xmath242 and @xmath163.,title="fig : " ] 0.5 cm sturmian substitution @xmath244 and its fixed point @xmath19 , @xmath18 and @xmath19 , and the substitutions @xmath20 , @xmath155 , @xmath156 ( all with the same self - similarity factor @xmath111 ) fixing them .
we have @xmath242 and @xmath163 . ]
let @xmath13 be a fixed point of @xmath20 and let @xmath245 be the geometric representation of the substitution @xmath20 with the dominant eigenvalue @xmath111 and the right eigenvector @xmath219 for which @xmath246 morphisms @xmath155 and @xmath156 are sturmian substitutions with fixed points @xmath18 , @xmath19 , respectively .
the geometric representation of the infinite word @xmath18 is @xmath247 and the geometric representation of the infinite word @xmath19 is @xmath248 if @xmath249 , i.e. @xmath250 , then the segment in @xmath245 between @xmath251 and @xmath252 ( both in @xmath245 ) contains points ordered according to @xmath253 .
and the segment in @xmath254 between @xmath251 and @xmath252 ( both in @xmath254 ) contains points ordered according to @xmath255 .
analogically , for @xmath59 such that @xmath256 , points in @xmath254 between @xmath251 and @xmath252 are ordered according to @xmath257 . from what was said above
it is obvious , that the substitution @xmath155 with factor @xmath111 fixing the sturmian word @xmath18 must be of the form @xmath258 in a similar way , we can deduce that the substitution @xmath156 under which the infinite word @xmath19 is invariant is of the form @xmath259 by definition [ d : amicablewords ] , we have that @xmath260 and @xmath181 , and that @xmath261 , @xmath262 . in order to complete the proof of the theorem
, we have to show that @xmath188 and @xmath263 .
for that , consider @xmath264 such that @xmath265 , i.e. @xmath266 .
the segment between @xmath251 and @xmath267 in the geometric representation @xmath254 of @xmath18 contains the points arranged according to @xmath268 .
similarly , the segment between @xmath269 and @xmath270 contains the points arranged according to @xmath271 .
of course , the segment between @xmath251 and @xmath252 in the geometric representation @xmath245 of the original infinite word @xmath13 is arranged according to @xmath272 .
altogether , we have @xmath273 analogously , @xmath274 this means that @xmath188 , and the word @xmath272 is the ternarization of words @xmath184 and @xmath185 .
consequently , @xmath155 is amicable to @xmath156 , and the substitution @xmath20 is the ternarization of @xmath155 and @xmath156 .
we acknowledge financial support by the grants msm6840770039 and lc06002 of the ministry of education , youth , and sports of the czech republic . | we consider words coding exchange of three intervals with permutation ( 3,2,1 ) , here called 3iet words . recently
, a characterization of substitution invariant 3iet words was provided .
we study the opposite question : what are the morphisms fixing a 3iet word ?
we reveal a narrow connection of such morphisms and morphisms fixing sturmian words using the new notion of amicability . |
an accurate measurement of the hyperfine splitting ( hfs ) of the @xmath0 ground state of hydrogen - like highly charged ions ( hci ) is a good test of quantum electrodynamics ( qed ) in the limit of strong electric fields ( @xmath1 v / cm @xcite ) .
such strong fields can not be produced using conventional laboratory techniques , but naturally exist close to the stripped nuclei of heavy elements like pb , bi or u. by detecting the fluorescence from the laser - excited upper hyperfine state of such a trapped and cold ion , a high - precision measurement of the hyperfine splitting can be made .
hci with extremely high charge states ( _ e.g. _ pb@xmath2 or u@xmath3 ) and relativistic energies ( 400 mev / u ) are created in the heavy ion facility ( sis ) at gsi in darmstadt .
these hci are injected into the experimental storage ring ( esr ) and can be decelerated ( down to 4 mev / u ) forming 1 @xmath4s long bunches containing about @xmath5 ions , arriving every 10 seconds . in the hitrap project @xcite these bunches will be extracted from the esr , decelerated by linear ( ih - linac ) and radiofrequency ( rfq ) stages , trapped and cooled in a penning trap ( cooler trap ) , and made available for experiments . in the esr at gsi , previous measurements of the hfs were made on bunches of relativistic hci such as @xmath6bi@xmath7 @xcite and @xmath8pb@xmath2 @xcite .
hfs measurements were also made at the superebit on @xmath9ho@xmath10 @xcite , @xmath11re@xmath12 @xcite and @xmath13tl@xmath14 @xcite .
the resolution obtained in the above experiments is mainly limited by the doppler effect .
a measurement of the @xmath0 ground state hfs of trapped cold hci using laser spectroscopy should be even more accurate due to a cryogenic uhv environment , high ion cloud density , the absence of a large doppler shift and virtually unlimited measurement time .
laser spectroscopy offers the possibility of high - accuracy measurements of transition wavelengths in the visible region @xcite . in hci ,
electronic transitions are generally in the far uv or x - ray regions of the spectrum . however , since the @xmath0 groundstate hfs scales with the atomic number @xmath15 as @xmath16 , the ground state hfs of hydrogen - like hci can move into the visible spectrum for @xmath17 @xcite .
the lifetime of this transition falls as @xmath18 and is of the order of milliseconds for @xmath17 .
a measurement of this transition wavelength gives information on the qed corrections to the hfs or on the spatial distribution of the nuclear magnetisation ( bohr - weisskopf effect ) , which is affected by core polarisation and is not really well understood @xcite .
its measurement thus allows for critical tests of nuclear models . from a comparison of measurements of the hfs of hydrogen - like and lithium - like hci the nuclear effects
can be eliminated so that an accurate measurement of the qed effects can be made @xcite .
there are several candidate systems that can be studied at hitrap , including radioactive isotopes
. an interesting first challenge would be to measure the hfs of the @xmath0 ground state @xmath19 transition of @xmath8pb@xmath2 ions ( @xmath20 nm @xcite ) by means of laser spectroscopy . in order to reach the necessary accuracy ,
the hci will be trapped in a cryogenic uhv environment .
once trapped , the ions can be easily stored for long times , therefore the lifetime @xmath21
ms @xcite of the upper hyperfine state is not a problem .
electron capture ( neutralisation ) by collisions is strongly reduced by operating the trap at cryogenic temperatures ( 4 k ) under uhv conditions ( below @xmath22 mbar ) .
ground states of @xmath23 nuclei plotted versus magnetic field : a ) hydrogen atom @xmath24h@xmath25 , b ) highly charged lead ion @xmath8pb@xmath2 .
( the plots are drawn with different scales.),width=283 ]
the high precision measurements of the hfs will be performed on hci in a penning trap , but the strong magnetic field will shift and split the hyperfine levels and thus the natural hfs .
this ` zeeman effect ' @xcite is schematically illustrated in figure 1a ) for the @xmath0 ground state hfs in hydrogen , which has nuclear spin @xmath23 . for hydrogen
the hfs can be easily calculated using the famous breit - rabi formula @xcite .
the zero - field splitting @xmath26 is modified by the interaction between the magnetic moment @xmath4 of the ion and the magnetic field strength @xmath27 . in the simple hydrogen case at a field of 1 t , the zeeman splitting of about @xmath28 hz is more than one order of magnitude larger than the hfs in hydrogen ( @xmath29 hz or 21 cm ) .
however , for hci the hfs dominates the zeeman splitting by several orders of magnitude .
this is indicated in figure 1b ) for the @xmath0 ground state hfs of @xmath8pb@xmath2 , which also has nuclear spin @xmath23 .
ground state hyperfine splittings in the visible spectrum for atomic number @xmath15 ranging from 50 to 100.,width=264 ] in figure 2 the calculated wavelengths of the @xmath0 ground state hyperfine transition ( neglecting qed and the bohr - weisskopf effect ) are plotted for hci with atomic number @xmath15 ranging from 50 to 100 .
most of these heavier elements have their hfs in the visible region of the spectrum and are thus easily excited by conventional tunable laser systems .
the hci will be extracted from the hitrap cooler trap in a long narrow bunch with an average kinetic energy of a few ev and containing about @xmath5 ions .
the radial energy spread will be reduced so as to maintain a parallel beam @xcite .
the hci will then be loaded into the spectroscopy trap , which is a cylindrical open - endcap penning trap @xcite with compensation electrodes to create a nearly perfect quadrupole potential at the trap centre .
figure 3 schematically shows the expected layout of the electrodes , the tank circuit for resistive cooling @xcite and the trap loading scheme : the hci enter from the right , are reflected by the left capture electrode , enclosed by the right one , localised near the trap centre and finally cooled and compressed .
the principle of resistive cooling is fairly straightforward and well - known : the axial motion of the hci in the trap induces image charges in the open - endcap electrodes , which can readily be detected by a frequency resonant electronic tank circuit ( lcr - filter ) attached to the electrodes .
if the tank circuit is at cryogenic temperatures ( 4 k ) , kinetic energy of the hci is dissipated by the circuit , thus effectively cooling the ions .
since the cooling time varies with the charge state @xmath30 of the ions as @xmath31 , this cooling scheme is very effective for hci @xcite .
after the cooling procedure , the hci will be focussed into a smaller cloud using the rotating wall technique @xcite .
if the coupling of an ion plasma in a penning trap is strong enough , the density of the ion cloud depends on its global rotation frequency . by applying rf voltages to a segmented ring electrode ,
the radial electric quadrupole field will rotate and drive the ion cloud .
there are two limits for the rotation frequency @xmath32 .
the lower limit , which leads to the minimal density , is reached when @xmath32 equals the single - ion magnetron frequency @xmath33 . at the upper limit , when @xmath32 is set to half the cyclotron frequency @xmath34 , the density is maximum ( brillouin limit ) .
( see _ e.g. _
@xcite for the equations of motion of an ion in a penning trap . ) realistic values for the penning trap are : a magnetic field of 6 t , an applied potential of 500 v , an ion cloud temperature of 4 k , and a trap parameter @xmath35 mm ( see ref @xcite ) .
these values lead to the trap frequencies @xmath36 khz , @xmath37 mhz , and an axial frequency @xmath38 mhz .
if we set the rotation frequency to @xmath39 mhz , the number density of a cloud of @xmath5 @xmath8pb@xmath2 ions is @xmath40 @xmath41 ( the brillouin limit is @xmath42 @xmath41 ) .
such an ion cloud has a length of 6.8 mm and a diameter of 0.75 mm , which leads to an aspect ratio of 9.0 .
the debye length of this cloud is 250 nm .
the dependence of the ion cloud parameters on @xmath32 is shown in figure 4 .
@xmath8pb@xmath2 ions , plotted as a function of the rotation frequency @xmath32 .
( @xmath43 mm),width=283 ] a spectrum of the ground state hyperfine transition can be obtained by scanning the laser wavelength across the resonance while recording the fluorescence from the trapped decaying hci .
repeated measurements on a single cloud of ions can be made , as detection is not destructive . to obtain the highest spectral resolution
, the laser beam must be parallel to the magnetic field ( trap axis ) , so as to avoid the large doppler shift associated with the global rotation of the ion cloud .
laser excitation along the trap axis , on an ion cloud smaller than the laser beam , will also help to ensure that the transition is fully saturated .
an ion temperature of 4 k corresponds to a doppler width of 30 mhz . for @xmath8pb@xmath2 ions
, the wavelength of 1020 nm corresponds to an upper state lifetime of 50 ms and a linewidth of 3 hz . the laser power to fully saturate the doppler - broadened transition is roughly 3 mw and is easily reached by _
e.g. _ a ti : sapphire or ar@xmath44 pumped dye laser . if the experiment is run in a continuous mode , assuming an overall detection efficiency of @xmath45 , we expect to detect about @xmath46 photon emissions per second for a completely saturated ion cloud
. it should be possible to reduce the background signal to less than @xmath47 counts per second , yielding a typical s / n ratio of 40 .
alternatively , if the laser excitation is pulsed with a duty cycle of 200 ms ( @xmath48 ) , the signal is about @xmath49 counts per second , without any background from scattered laser light .
these values are high enough to allow easy detection and measurement of the transition wavelength .
once the signal is seen , this will allow a wavelength determination to an accuracy which far exceeds the theoretical uncertainties .
the linewidth of _
e.g. _ a ti : sapphire laser is 1 mhz at a wavelength of 1 @xmath4 m , which is much smaller than the doppler broadening of 30 mhz .
therefore the accuracy of the hfs measurement is estimated to be about @xmath50 , which is 3 orders of magnitude better than the current accuracy @xcite .
this work is supported by the european commission within the rtd programme fp5 ( hpri - ct-2001 - 50036 hitrap ) .
jrcp acknowledges the support by conacyt , sep and the ors awards scheme . | laser spectroscopy studies are being prepared to measure the @xmath0 ground state hyperfine splitting in trapped cold highly charged ions .
the purpose of such experiments is to test quantum electrodynamics in the strong electric field regime .
these experiments form part of the hitrap project at gsi .
a brief review of the planned experiments is presented . , ,
highly charged ion , hyperfine splitting , laser spectroscopy 32.30.jc , 12.20.fv , 21.10.ky |
the nature of transverse momentum@xcite spectra is quite interesting for the following reasons : firstly , in the form of invariant cross section in terms of @xmath0 , it offers a ready , reliable and very basic observable for both measurements , and for theorization to be attempted .
secondly , the @xmath0-spectra throw light on the particle production mechanism on the whole ; they have also impact and indirect reflections on the predicted features of what is normally projected as the signatures of the conjectured quark - gluon plasma(qgp ) .
thirdly and finally , the expressions for invariant @xmath0-spectra lead us to define some other physically significant observables in the domain of particle physics , like the average multiplicity , average transverse momentum etc .
after pions , kaons of all varieties ( with @xmath4 and @xmath5 ) constitute the second most abundant species among the particle secondaries which are produced in multiparticle processes of hadron - hadron , hadron - nucleus or nucleus - nucleus interactions at high and very high energies .
secondly , kaons are the lightest ` strange ' particles for which they have a special status . besides
, kaons are supposed to bear , as told above , some strong reflections on what are viewed as the qgp diagnostics .
lastly , the kaon - pion ratios and their energy - dependences play a very significant role in both the particle physics and non - accelerator physics sectors .
so , analyzing the kaon production characteristics , especially the @xmath0-spectra , in the light of any model assumes a high degree of importance and comes as a challenging task as well .
furthermore , the production of proton - antiprotons , the first family - set of baryonic category , is also important from various physical considerations .
the nature of the ratios @xmath6 has a strong bearing on astroparticle physics . besides , the generalized antiparticle to particle ratios are of importance in the study of matter - antimatter symmetry in the universe , since they provide the relative abundance of antiparticle productions@xcite .
very recently , spurts of data on the various aspects of particle production from the cern - sps and the rhic - bnl , studying high energy interaction properties in the collisions between @xmath7 and @xmath8 respectively , are available . and providing interpretations for both the previous measurements and these oncoming blizzards of new data is really a very significant work .
our objective here is to try to describe and/or explain the nature of the transverse momentum ( @xmath0)-spectra of kaons , protons and antiprotons in a large variety of proton - induced and nucleus - induced high energy interactions with the help of a combinational approach(ca ) , called also the grand combination of models ( gcm ) .
the present work is prompted by our previous successes in analyzing the @xmath0-spectra and rapidity - spectra@xcite of pions alone in a host of nuclear collisions at high and superhigh energies .
in fact , taking up the task of interpreting the @xmath0-spectra of a set of non - pion secondaries produced in nuclear collisions becomes the logical imperative and the natural follow - up in maintaining a sequence of our studies and with the model of our choice in a self - complete way . in real terms , while proceeding with the present work , we would follow the trajectory of the ref.@xcite here more closely than the other one , as in the former the approach to the studies on @xmath0-spectra has been rationalized in a better way and much more consistent manner .
we have consciously and carefully avoided here to adopt operationally any of the known popular / standard models and have also tried to escape , as far as possible , the buzzwords related to them .
and we have singularly been guided by the results of experimental measurements on the nature of @xmath0-spectra alone in a very dispassionate and objective manner .
this paper is organized here as follows . in the next section ( section ii )
we give the outline of the combinational approach which is to be taken up for this study and the sketch of the physical perspective which prompted us to proceed in the stated direction . in section
iii we present the essential steps of the methodology of our work , the results of model - based calculations and some brief discussion on the results obtained by this combinational approach .
the last section is reserved for summing up the conclusions and pointing out the shortcomings of the approach adopted here .
following the suggestion of faessler@xcite and the work of peitzmann@xcite and also of schmidt and schukraft@xcite , we propose here a generalized empirical relationship between the inclusive cross - section for production of @xmath4 , @xmath9 and @xmath10 in nucleon(n)-nucleon(n ) collision and that for nucleus(a)-nucleus(a ) collision as given below :
@xmath11 where q stands for @xmath4 , @xmath9 , @xmath10 and @xmath12 could be expressed in the factorization form , @xmath13 .
while investigating a specific nature of dependence of the two variables(@xmath14 and @xmath0 ) , either of them is assumed to remain averaged or with definite values .
speaking in clearer terms , if and when @xmath15-dependence is studied by experimental group , the rapidity factor is integrated over certain limits and is absorbed in the normalization factor.so , the formula turns into @xmath16 the main bulk of work , thus , converges to the making of an appropriate choice of form for @xmath17 .
and the necessary choices are to be made on the basis of certain premises and physical considerations which do not violate the canons of high energy particle interactions .
let us now tentatively propose that the expressions for inclusive cross - section of non - pion secondaries in proton - proton scattering at high energies in eqn.(2 ) could also be chosen in the same form as that was suggested for pions by hagedorn@xcite : @xmath18 where @xmath19 is the normalization constant , and @xmath20 , @xmath21 are interaction - dependent chosen phenomenological parameters for which the values are to be obtained by the method of fitting .
their @xmath22-dependences are here proposed to be given by the following formulations : @xmath23 and @xmath24 where @xmath25 , @xmath26 , @xmath27 and @xmath28 are four constants .
the @xmath22-dependence of @xmath29 and @xmath21 would be shown later diagrammatically in the text and their data - base would also be indicated .
the nature and significance of these parameters could be appreciated from the work of hagedorn@xcite , and those of bielich et al@xcite and albrecht et al@xcite .
the final working formula for the nucleus - nucleus collisions is now being proposed here in the form given below :
@xmath30 with @xmath31 , wherein this quadratic form of parametrization is suggested here tentatively by us for analyzing data and testing its efficacy , if any and is hereafter called de - bhattacharyya parametrization(dbp ) . in the above expression @xmath32 is the normalization term which has a dependence either on the rapidity or on the rapidity - density of the produced secondary ; @xmath33 , @xmath34 and @xmath35 are constants for a specific set of projectile and target .
earlier experimental works@xcite showed that @xmath17 is less than unity in the @xmath0-domain , @xmath36 gev / c . besides
, it was also observed that the parameter @xmath33 , which gives the value of @xmath17 at @xmath37 , is also less than one and this value differs from collision to collision .
the other two parameters @xmath34 and @xmath35 essentially determine the nature of curvature of @xmath17 .
however , in the present context , precise determination of @xmath33 is not possible for the following understated reasons : \(i ) to make our point let us recast the expression for ( 6 ) in the form given below : @xmath38 quite obviously , we have adopted here the method of fitting .
now , in eqn.(7 ) one finds that there are two constant terms @xmath32 and @xmath33 which are neither the coefficients nor the exponent terms of any function of the variable , @xmath0 . and as @xmath33 is a constant for a specific collision at a specific energy , the product of the two terms @xmath32 and @xmath39 appears as just a new constant . and , it will not be possible to obtain fit - values simultaneously for two constants of the above types through the method of fitting .
\(ii ) from eqn.(2 ) the nature of @xmath17 can easily be determined by calculating the ratio of the logarithm of the ratios of nuclear - to-@xmath1 collision and the logarithm of the product @xmath40 .
thus , one can measure @xmath33 from the intercept of @xmath17 along y - axis as soon as one gets the values of @xmath41 for both @xmath40 collision and @xmath1 collision at the same c.m . energy . in the present study we have tried to consider the various collision systems in as many number as possible . to do so
, we have to consider the data on normalized versions of @xmath41 for many collision systems for which clear @xmath41-data were not available to us . furthermore , from these normalized versions we can / could not extract the appropriate values of @xmath41 as the normalization terms , total inclusive cross - sections@xmath42 etc .
, for these collision systems can not always be readily obtained .
besides , it will also not be possible to get readily the data on inclusive spectra for @xmath1 collisions at all c.m.energies , like e.g. , at @xmath43(c.m .
energy of @xmath44 collision ) . in order to sidetrack these difficulties and also to build - up an escape - route , we have concentrated almost wholly here to the values of @xmath34 and @xmath35 for various collision systems ; and the effects of @xmath32 and @xmath33 have been compressed to a single constant term @xmath45 .
hence , the final expression becomes , @xmath46 with @xmath47 .
the exponent term @xmath48 obviously represents here @xmath49 $ ] instead of @xmath17 alone .
thus , after obtaining fit - values of @xmath34 and @xmath35 , if @xmath49 $ ] are plotted for various collision systems , all the curves would originate from a single point , i.e. origin ; and the systems and processes are then really comparable . in other words , in this convenient way we could study and check the scaling characteristics of @xmath17 with respect to the collision systems .
the expression(8 ) given above is the physical embodiment of what we have termed to be the grand combination of models(gcm ) or the combinational approach(ca ) that has been applied here .
firstly , the results of @xmath1 scattering are obtained in the above on the basis of an assumption of validity of eqn.(3 ) for even the non - pion secondaries , though originally the hagedorn s model(hm ) was proposed only for pion secondaries .
secondly , the route for converting the results for the non - pion secondaries produced in @xmath50 to @xmath51 or @xmath40 collisions is built up by the peitzmann s approach(pa ) represented by expression(2 ) .
thirdly , the further input is the de - bhattacharyya parametrization for the nature of the exponent .
thus , the gcm or the ca provides the combination of hm , pa and the dbp , and thus constitutes a single whole . and
this whole is both useful and economical ; it is economical in the sense that it can accommodate vast wealth of data with only three arbitrary parameters .
compared to the theoretical apparatus provided by the standard model(sm)@xcite , this number is quite low .
and the choice of this quadratic form is not altogether a coincidence . in dealing with the emc effect in the lepton - nucleus collisions ,
one of the authors here(sb)@xcite , made use of a polynomial form of @xmath52-dependence with the variable @xmath53(feynman scaling variable ) .
this gives us a clue to make a similar choice with both @xmath0 and @xmath54 variable(s ) in each case separately .
in the recent times , de - bhattacharyya parametrization is being extensively applied to interpret the measured data on the various aspects@xcite of the particle - nucleus and nucleus - nucleus interactions at high energies . in the recent past
al.@xcite also made use of this sort of relationship in a somewhat different context .
the underlying physics implications of this parametrization stem mainly from the expression(6 ) which could be identified as a clear mechanism for switch - over of the results obtained for nucleon - nucleon(@xmath1 ) collision to those for nucleus - nucleus interactions at high energies in a direct and straightforward manner .
the polynomial exponent of the product term on @xmath40 takes care of the totality of the nuclear effects .
the individual model(s ) , seen in a split manner , is(are ) certainly not new ; but the combination with the proposed two - factor quadratic @xmath0-dependence of the exponent ( called dbp in the text ) for @xmath40 or @xmath55 interactions at high energies is testably a new proposition offered by us . besides , the probing of the role of the hagedorn s model in understanding the production of non - pion secondaries in high energy collisions is another virgin feature in our work .
the testing of this mechanism suggested by us is with the views of ( i ) presenting an integrated approach toward production of various kind of particle - secondaries in high energy collisions and ( ii ) providing a unified outlook to particle production in all varieties of particle - particle , particle - nucleus or nucleus - nucleus interactions at high energies .
quite understandably , only by intensive analysis of data on several / all sets of collisions somewhat successfully such claims could , at all , be made , and/or later be justified .
the trends that emerge from this extensive data - analysis with this grand combination of models(gcm ) has been given in the next section . obviously , in its approach and method
, this work is undoubtedly a continuation of one of our previous works@xcite .
but , even at the risk of being a bit repetitive , we have to proceed and carry on . because , without these special efforts at studying on a case - to - case basis for each of the specific variety of the secondaries , we can not reasonably arrive at any definitive conclusion and also can not confirm the physical basis , if any , of our work . and
the two points made very precisely in the preceding paragraph represent the strong and crucial aspects of our physics motivation behind this proposed formalism .
a point is to be made .
our intention here , is surely not just piling up of large bulk of data in a purposeless way ; rather we would like to make the fullest possible utilization of them for checking the proposed gcm as extensively as possible with the widest range of available data on diverse set of reactions . because , such rigorous and intensive checkings only could lead to meaningful and valid conclusions , if any . and , in our opinion
, this has helped us to substantiate , to a considerable degree , the functional efficacy and strength of the proposed phenomenological model which is called here the combinational approach .
in fact , the revelation of the potency and strength of this phenomenological approach is the other most striking feature which we would surely like to emphasize here .
it is obvious that there are two phenomenological parameters in @xmath17 in expression(8 ) which need to be physically explained and/or identified . in compliance with this condition
we offer the following physical explanations for the occurrence of all these factors .
the particle - nucleus or nucleus - nucleus collisions at high energies subsequently gives rise to an expanding blob or fireball with rising temperature . in real and concrete terms
this stage indicates the growing participation of the already - expanded nuclear blob .
as temperature increases at this stage , the emission of highly energetic secondaries(which are mostly peripheral nucleons or baryons ) with increasing transverse momentum is perfectly possible .
the coefficient @xmath34 addresses this particularity of the natural event ; and this is manifested in the enhancement of the nuclear contribution with the rise of the transverse momentum .
thereafter , there is a turnabout in the state of reality .
after the initial fractions of seconds , the earlier - excited nuclear matter starts to cool down and there is a clear natural contraction at this stage as the system suffers a gradual fall in temperature .
finally , this leads to what one might call ` freeze - out ' stage , which results in extensive hadronization , especially in production of hadrons with very low transverse momentum . in other words , the production of large-@xmath0 particles at this stage is lowered to a considerable extent .
this fact is represented by the damping or attenuation term for the production of high-@xmath0 particles .
the factor @xmath35 with negative values takes care of this state of physical reality .
thus the function denoted by @xmath17 symbolizes the totality of the features of the expansion - contraction dynamical scenario in the after - collision stage .
this interpretation is , at present , only of very suggestive nature .
however , let us make some further clarifications .
the physical foundation that has here been attempted to be built up is inspired by thermodynamic pictures , whereas the quantitative calculations are based on a sort of pqcd - motivated power - law formula represented by eqn.(6 ) .
this seems to be somewhat paradoxical , because it would be hard to justify the hypothesis of local thermal equilibrium in multihadron systems produced by high energy collisions in terms of successive collision of the qcd - partons(like quarks and gluons ) excited or created in the course of the overall process .
except exclusively for central heavy ion collisions , a typical parton can only undergo very few interactions before the final - state hadrons ` freeze out ' , i.e. escape as free particles or resonances .
the fact is the hadronic system , before the freeze - out starts , expands a great deal both longitudinally and transversally while these very few interactions take place@xcite .
but the number of parton interactions is just one of the several other relevant factors for the formation of local equilibrium . of equal importance
is the parton distribution produced early in the collision process .
this early distribution is supposed to be a superposition of collective flow and highly randomized internal motions in each space cell which helps the system to achieve a situation close to the equilibrium leading to the appropriate values of collective variables including concerned and/or almost concerned quantities .
the parameter @xmath34 in expression(6 ) is somehow related to the measure of the ratio of the net binary collision number to the total permissible number among the constituent partons in the pre - freeze out expanding stage identified to be a sort of explosive ` detonation'@xcite stage .
this is approximated by a superposition of collective flow and thermalized internal motion , which is a function of hadronic temperature manifested in the behaviour of the average transverse momentum .
the post freeze - out hadron production scenario is taken care of by the soft interaction which is proportional@xcite to the number of participant nucleons , @xmath56 , according to almost any variety of wounded nuclear model .
the factor @xmath35 , we conjecture , might have a relationship with the ratio of the actual participating nucleons to the total number of maximum allowable(participating ) nucleons .
in fact , this sort of physical explanations is reproduced here from some of our previous works@xcite .
at the very start we study the @xmath0-spectra for @xmath57 , @xmath9 and @xmath10 inclusive production in @xmath1 collisions at several energies and try to fit the expression in the formula(3 ) given in the previous section .
the graphs are shown in the several diagrams in fig.1 for kaons , protons and antiprotons .
the fits for the average kaon yield and antiproton production in @xmath2 reactions have also been obtained(fig.2 ) on the basis of hagedorn s model(eqn.(3 ) ) . the same
could not be done for proton production in @xmath2 reactions due to unavailability of the measured experimental data in this particular collisions .
the obtained fitted - values of the parameters , @xmath29 and @xmath21 , for different secondaries produced in @xmath58 collisions at different energies have been displayed in table - i - table - v .
the graphs for @xmath22-dependences of @xmath29 and @xmath21 , as proposed in eqn.(4 ) and eqn.(5 ) , have been depicted in fig.3 and the necessary values of the parameters , @xmath25 , @xmath27 , @xmath26 and @xmath28 have been presented in table - vi . while plotting graphs we have obviously assumed that at very high energies the proton - proton and proton - antiproton collisions could be treated at par and with the acceptance of equivalence between each other .
hereafter , in studying the nature of @xmath0-spectra in all nucleon - nucleus and nucleus - nucleus collisions , the interaction energy in all cases of nucleus - induced reactions is invariably converted first into the c.m .
system values , that is expressed in @xmath59 , then the values of @xmath29 and @xmath21 are picked up from the graphical plots drawn already and shown by fig.3 . while analyzing nucleus - dependence with expression(8 ) and trying with the fit parameters @xmath45 , @xmath34 and @xmath35 , we have inserted these extracted values of @xmath29 and @xmath21 for @xmath1 cross section term occurring in the expression(8 ) given above .
however , the plots presented in fig.3 deserve special mention from phenomenological points of view .
the @xmath29 and @xmath21 values for production of the same secondary at various energies ( @xmath22-values ) show very slow fall - off with energy in these graphs .
the nearness of values of @xmath60 signals a march towards a steady state of hadronization .
once the natures are established by the theoretical procedures adopted for either @xmath58 reactions at high or very high energies , we would be compelled , for the sake of consistency , to reduce / minimize the degree of arbitrariness in choosing values of @xmath29 and @xmath21 for nucleon - nucleus or nucleus - nucleus collisions . as soon as the values of @xmath61 is given for a specific interaction , the values of @xmath29 and @xmath21
are to be obtained from the theoretical plots of them shown in fig.3(a ) , fig.3(b ) and fig.3(c ) for different secondaries and/or for different collision(s ) .
the diagrams given in fig.4 to fig.8 do actually provide the nature of collected data and also the theoretical descriptions for diverse reactions indicated by the labels in each diagram .
the solid curves represent the results arrived at on the basis of the new combination of models(ncm ) and the use of de - bhattacharyya parametrization(dbp ) . and
the totality of this combination of ncm and dbp is called grand combination of models(gcm ) . and the tables ( displayed by table - vii to table - ix )
provide the necessary parameter values that have entered into each calculation of the theoretical solid curves in all the diagrams of fig.4-fig.8 for various reactions . in some cases , due to lack of availability of clear proton data ,
we have used the net proton(@xmath62 ) data with the assumption of the near equivalence and as an approximate measure .
the diagrams presented in fig.9(a ) to fig.9(c ) are full of physical import and convey the message of mass number scaling of nuclear reactions .
the deviations , whenever and wherever if any , from this observation could be attributed , in the main , to the following few points : ( i ) limits of uncertainties in the measurements ; ( ii ) some unavoidable approximations in theoretical calculations ; and ( iii ) availability , in some cases , of only too sparse and rare data on which reliability is certainly questionable . obviously , we present here the gcm - based analyses of extensive sets of data on some of the observables measured by the various experimental groups in high energy particle and nuclear physics . the study has , thus , given rise to some crucial observations with revelations of following systematic trends : ( a ) for the non - pion secondaries produced in nucleus - nucleus collisions at high energies the nature of fits is not as good as in @xmath1 reaction , especially at low momentum region , and especially for @xmath63 vs. @xmath0 plots .
the reason(s ) for this discrepancy would be discussed later in the last section .
( b ) the fit parameters @xmath29 and @xmath21 for production of secondary kaon(s ) and protons are , by ascription , different in the basic @xmath1 reactions . and
this is consistent with the theoretical expectations .
the factor @xmath60 has , we recognize , a relationship with the slope parameter of the @xmath0-spectra . and the slope parameters for the lighter and heavier particles , one knows , are not the same .
but , quite interestingly , the nature of @xmath22-dependences of @xmath29 and @xmath21 values of all three species , kaons , protons and antiprotons , as shown in fig.3(a)-3(b ) , are qualitatively quite alike.(c ) it is quite evident from the tables vii , viii , and ix that the values of @xmath34 and @xmath35 are too close .
this , in our opinion , reflects the basic fact that the nuclear effects are quite finite and certainly not too pronounced even with the heaviest of the nuclei and the highest available energies .
in the end let us summarize our main findings from the present study in the following way : 1 . quite naturally and predictably , the @xmath64 values shown in table - vii - ix present , at times , sharply contrasting nature , because of either very sparse available data and that too for a narrow region of @xmath0-values , or large ranges of errors and uncertainties in the measurements . 2 .
still , on the whole , the approach describes modestly well the large bulk of data on kaon - antikaon and proton - antiproton production phenomena in high energy nucleon - nucleus and nucleus - nucleus collisions .
3 . related with production of @xmath57 and @xmath65 , the c.m .
energy - dependences of the parameter values occurring in the basic @xmath1 reactions are of the same hyperbolic nature as in the cases of pions .
this element of observation on similarity aspect of @xmath4 and @xmath65 with pions is certainly a new finding here made by us , as hagedorn s early work@xcite was made exclusively on pions .
4 . besides , the entire energy - dependences , even in purely nuclear collisions , are manifested by only the basic @xmath1 interaction ; the nuclear geometry exhibits no separate energy behaviour .
this is quite significant in the sense that it might be a great hint to and confirmation of the enormous importance of @xmath1 interaction even in understanding the nature of heavy ion collisions .
the observation of the close values of @xmath34 in various sets of nuclear collisions with varying pair of projectile and target and at different energies is another interesting finding .
it indicates that enhancement due to nuclear contribution could never assume too large values , even with the heaviest of the nuclei , though the parameters are uniformly non - zero for each and every collision and the values are limited .
this does , in effect , imply the very finite degree of the nuclear effects .
the constraint would appear more severe and acute , if the diminutive role of the @xmath35 factor is reckoned with .
this might be attributed to the large baryon stopping effect in heavy ion collisions .
+ besides , the small positive @xmath34-values in any collission represent physically the nuclear enhancement though certainly not of ` anomalous ' nature called cronin effect .
secondly , the negative @xmath35-values might explain the suppressions of productions of hadrons at relatively large values of the transverse momenta .
even with some heaviest nuclei , the observations from fig.9(a ) to fig.9(c ) on the validity of a sort of mass - number(@xmath52 ) scaling for production of both strange and non - strange variety of hadrons constitute an interesting and probeworthy point .
physically , this might signal the limit to the finite degree of impact parameter dependence of nuclear collisions .
the near - saturation of the parameter value at and after certain magnitude of @xmath40 indicates the stringent limit on both the number of binary collisions and the number of participating nucleons in the colliding nuclei .
the reason for observation of departure of the fits from the data on @xmath0-spectra produced in nuclear collisions , especially at low @xmath0 region , could be ascribed , in the main , to the following reasons .
firstly , because for the nucleus - involved collisions , the measurements in the experiments on purely soft collisions are very difficult , as hard collision effects are almost unavoidably present there .
secondly , the measurements of the factor @xmath63 through the array of several detectors have always an intrinsic simulation - based standard - model ( mainly the hijing version ) component , for which the obtained data are not the pure products of experimental measurements alone .
but the present model can , in no way , accommodate such sorts of superposition effects .
the property of ` universality ' of hadronic , hadronuclear and nuclear reactions at very high energies is also vindicated by the approach adopted here . the implications of all these are obvious .
with the series of such modestly successful ventures , the combinational approach attains a plateau and assumes a potential to claim the status of an alternative approach to understand functionally some aspects of the heavy ion collisions in general , and the nature of @xmath0/rapidity - spectra in @xmath51 or @xmath40 collisions in particular . however , we must not gloss over the gross limitations of the present approach which are as follows : ( i ) it is a fact that we could , so far , offer no measure to estimate quantitatively the values of @xmath34 and @xmath35 .
the hints are , uptil now , of only very qualitative nature .
( ii ) quite admittedly , the physical explanations outlined here for @xmath34 and @xmath35 are also of tentative nature ; they have , still , not been identified concretely with any of the known physical observables in the domain of collision dynamics for multiple production of hadrons and of nuclear geometry as well .
( iii ) the calculations for particle - nucleus or nucleus - nucleus interaction cases are absolutely dependent on the apriori and data - based knowledge of the production characteristics of the same particle species in @xmath1 reactions at various energies . without it ,
the method is ineffective and that is a major handicap .
( iv ) the approach fails to respond to the very basic query on actual mechanisms for particle production processes in any collision .
( v ) and just because of it , the approach is insensitive to the charge - state of the specific secondary produced in any high energy interaction and that is certainly a great difficulty .
so , unless these problems could be remedied to a great extent , we accept that we have no reasons for complacency . *
r. c. hwa and c. b. yang : nucl - th/0301004(02 january 2003 ) .
y. zhang et al : phys . rev .
* c 65 * ( 2003 ) 034903 .
p. staszel(brahms collaboration ) : act .
phys . pol .
* b 33 * ( 2002 ) 1387 .
bhaskar de , s. bhattacharyya and p. guptaroy : jour .
g 28*(2002)2963 .
bhaskar de , s.bhattacharyya and p.guptaroy : int .
phys . * a 17*(2002)4615 .
m.a.faessler : phys . rep . *
115*(1984 ) 1 .
t.peitzmann : phys . lett . * b 450*(1999 ) 7 .
h.r.schmidt and j.schukraft : j. phys . * g 19*(1993 ) 1705 .
r.hagedorn : riv .
cim . * 6*(1983 ) 1 .
j.s.bielich et al : nucl .
phys . * a 705*(2002 ) 494 .
r.albrecht et al(wa80 collaboration ) : eur . phys
* c 5*(1998 ) 255 .
d.antreasyan et al : phys . rev .
* d 19*(1979 ) 764 .
m.m.aggarwal et al(wa98 collaboration ) : eur . phys
* c 23 * ( 2002 ) 225 .
j.ellis : int .
phys . * a 17*(2002 ) 3284 .
s.bhattacharyya : lett .
* 44*(1985)119 .
r.c.hwa et .
phys . rev . * c 64*(2001)054611 .
l.van-hove : z. phys .
* c 21*(1983)93 .
s.y.li and x.n.wang : phys . lett .
* b 527*(2002)85 .
bhaskar de , s.bhattacharyya and p.guptaroy : eur .
jour . * a 16*(2003)415 ; bhaskar de , s.bhattacharyya and p.guptaroy : int .
* a 13*(2002)2367
. b. alper et al : nucl .
* b 100*(1975 ) 237 .
t. alexopoulos et al(e735 collaboration ) : phys .
d 48*(1993 ) 984 .
j. baechler et al : z. phys .
* c 58*(1993 ) 367 .
t. alber et al(na35 collaboration ) : eur .
* c 2*(1998)643 .
h. boggild et al(na44 collaboration ) : phys . rev .
* c 59*(1999 ) 328 .
i. g. bearden et al(na44 collaboration ) : phys .
* c 57*(1998 ) 837 .
i. g. bearden et al(na44 collaboration ) : phys
* c 66*(2002 ) 044907 . t. alber et al(na35 collaboration ) : phys . lett . * b 366*(1996)56 .
j. velkovska : nucl . phys . * a 698*(2002 ) 507 . vs. @xmath0 for various secondaries produced in @xmath67 collisions at @xmath68 gev .
the various experimental points are taken from ref.@xcite .
the solid curves provide the fits on the basis of eqn.8.,width=264 ] vs. @xmath0 for various secondaries produced in @xmath67 collisions at @xmath68 gev .
the various experimental points are taken from ref.@xcite .
the solid curves provide the fits on the basis of eqn.8.,width=264 ] | in continuation of our perusal of the studies on transverse momentum spectra for the main varieties of secondaries from a consistent and comprehensive phenomenological approach , we propose to take up here after a successful completion of reporting in detail the results ( ref.@xcite in the text ) on our analyses of the @xmath0-spectra of pions the studies specially on production of kaons , protons and antiprotons in several proton - induced and nucleus - involved collisions at high energies . the measured data on inclusive cross sections of kaons , protons and antiprotons from the various published sources have here been assorted first .
next , data on the @xmath0-spectra of the specific secondaries produced in @xmath1 and @xmath2 reactions have been scanned and analyzed with the help of hagedorn s model(hm ) .
thereafter a connector , named here the combinational approach(ca ) , has been constructed and used to analyze the data on @xmath0-spectra of some major category of non - pion secondaries produced in nucleus - nucleus(@xmath3 ) collisions at high energies . and these exercises have , finally , led to the modestly successful interpretations of a wide band of data with the revelation of some insightful physical aspects of high energy interactions .
the limitations of the approach have also been precisely pointed out in the end . |
in recent years the representation theory of the braid groups @xmath7 in @xmath1 strands has attracted attention due to two ground breaking developments . one of them is in the work of bigelow and krammer @xcite , who managed to resolve the long standing problem of the linearilty of the braid groups by showing that a two - parameter generalization of the classical burau representation is faithful .
the second development is the emergence of vast families of braid group representations that are constructed from quantum algebras and conformal field theories .
intriguing relationships between these seemingly very different approaches have been discovered and they remain a fascinating area of study . in this article
we give an explicit construction and proof of an isomorphism between the faithful representation @xmath8 of @xmath0 considered by bigelow and krammer and the submodule of the _ r - matrix _ representations on @xmath9 for the generic verma module @xmath10 of the quantum group @xmath11 .
following @xcite we will refer in this article to @xmath8 as the _ lawrence - krammer - bigelow representation _ or lkb representation .
it is , by construction , defined over the ring @xmath12\,$ ] of two variable laurent polynomials with integral coefficients .
the parameters @xmath13 and @xmath14 are associated to deck transformations of a covering @xmath15 , where @xmath16 is the two - point configuration space on a disc with @xmath1-punctures , and @xmath17 ( see @xcite ) .
the first obstacle in finding such an isomorphism is that the braid group representations obtained from quantum groups are originally defined over the complex numbers rather than integral two - variable laurent polynomials . to this end
we will define @xmath11 as an algebra @xmath18 over @xmath19 $ ] , and introduce generic verma modules over @xmath20\,$ ] , where @xmath21 may be thought of as the exponential highest weight @xmath22 .
the braid group action commutes with the @xmath18-action so that the highest weight spaces @xmath23 of @xmath18 , corresponding to weights @xmath24 , are again @xmath0-invariant .
we prove that the @xmath25 are free @xmath26-modules , and construct explicit bases @xmath27 such that @xmath25 is the @xmath26-span of @xmath28 .
[ thm - freemodule ] the highest weight space @xmath29 is a free module over @xmath20\,$ ] with explicitly given basis @xmath30 as in ( [ eq - basis ] ) .
hence , for each @xmath31 we obtain a representation of the braid group @xmath0 in @xmath32-strands given by a homomorphism as follows : _ n , l : b_ngl ( , ) _ ( _
n , l ) the identification of the quantum representation on @xmath33 from theorem [ thm - freemodule ] with the lkb representation @xmath8 further requires an identification of parameters which we give by the following monomorphism between laurent polynomials . @xmath34\,\longrightarrow\,{\mathbb
z}[{s}^{\pm 1},{q}^{\pm 1}]=\rqs\;:\;\ ; \begin{array}{rcl } { \tt q } & \mapsto & s^{2}\\ { \tt t}&\mapsto & -q^{-2}\rule{0mm}{6 mm } \end{array}\ ] ] consider also the involutory automorphism @xmath35 of @xmath0 defined on the generators by @xmath36 ( given by switching all crossing or reflection at the plane of projection of a braid ) , and denote by @xmath37 the representation given by pre - composing the action on @xmath38 with @xmath35 .
with these conventions the main result of this article can be formulated as follows : [ thm - main ] for every @xmath39 there is a isomorphism of @xmath0-representations over @xmath40 _
n,2_n,2^_which maps the basis @xmath41 to the fork basis from @xcite
. in @xcite zinno manages to find a different identification of the lkb representation with a quantum algebraic object , namely the quotient of the birman - wenzl - murakami algebra similarly defined over @xmath12\,$ ] .
this representation can , by @xcite , be understood as the one arising from the quantum orthogonal groups @xmath42 acting on the @xmath32-fold tensor product of the fundamental representation .
since the representation in @xcite is irreducible this implies that @xmath8 and hence @xmath5 are irreducible for all @xmath39 .
and @xmath43 these representation are clearly reducible for all @xmath44 we generalize this result in our case to obtain [ thm - irred ] for all @xmath45 and @xmath46 the @xmath0-representation @xmath3 is irreducible over the fraction field @xmath47
. faithfulness of @xmath3 for @xmath48 is still an open question , as are identifications of these representation with geometrically constructed ones analogous to theorem [ thm - main ] .
obvious candidates for a generalization of theorem [ thm - main ] are the @xmath0-representations constructed by lawrence in @xcite .
the starting point there is again the configuration space @xmath49 of @xmath50 points in the plane with @xmath51 holes .
the braid group action is then naturally defined on @xmath52 , where @xmath53 is the canonical cover of @xmath54 with covering group @xmath55 .
the latter makes the representation spaces into @xmath12$]-modules .
[ conj - lawrence - iso ] the spaces @xmath56 are free @xmath12$]-modules which carry an ( irreducible ) action of @xmath0 as defined in @xcite .
they are isomorphic to the representations of @xmath0 on weight spaces @xmath3 over @xmath40 after appropriate identifications of parameters in the laurent polynomial rings .
the first obvious piece evidence for this conjecture is that it holds for @xmath57 .
indeed , for @xmath58 both @xmath59 and @xmath60 can be readily identified with the classical burau representation of @xmath0 . for more details
see the beginning of section [ sec - lkb ] .
for @xmath61 the lawrence representation @xmath62 is identical with @xmath8 so that in this case conjecture [ conj - lawrence - iso ] coincides with theorem [ thm - main ] .
furthermore , lawrence @xcite identifies the representations over the reduced ring @xmath63={\mbox{\raise .15em\hbox{$\mbox{\footnotesize${\mathbb z}[{\tt q}^{\pm 1},{\tt t}^{\pm 1}]$}$}\!\big/\!\raise -.18em\hbox{$\mbox{\footnotesize $ ( { \tt t}+{\tt q}^{-1})$}$}}}$ ] with the representations of the hecke algebra @xmath64 associated to two - row young diagrams .
the fact that this corresponds to a temperly lieb algebra quotient corroborates the quantum-@xmath65 content of these representations suggested in conjecture [ conj - lawrence - iso ] .
theorem [ thm - main ] as well as its generalization in conjecture [ conj - lawrence - iso ] are inspired by @xcite and @xcite where quantum-@xmath65 actions on the homology of local systems over similar configuration spaces are constructed .
* acknowledgments : * the second author thanks giovanni felder for very useful discussions about @xcite which motivated this article .
in this section we review the basic definitions and constructions of quantum @xmath66 which lead to the relevant representations of the braid groups . we will start from the framework of quasi - triangular topological hopf algebras due to drinfeld @xcite over rings of power series .
an exposition and further development of drinfeld s theory can be found in kassel s textbook @xcite which we will use as main reference .
we start with the definition of the algebra @xmath67 over a power series ring @xmath68 $ ] where @xmath69 is some commutative ring containing the rational numbers @xmath70 .
the indeterminate is related to @xmath71 used in @xcite by @xmath72 .
the generators of @xmath67 are @xmath73 , @xmath74 , and @xmath75 with relations @xmath76 & = & 2e\\ \,[h , f ] & = & -2f\rule{0mm}{7 mm } \end{array } \qquad\quad [ e , f ] = \frac { \sinh ( \hbar h ) } { \sinh ( \hbar ) } \;\;.\ ] ] the algebra @xmath67 is given by formal power series @xmath77 where each coefficient @xmath78 is a finite combination of monomials in the generators @xmath73 , @xmath74 , and @xmath75 over @xmath79 .
it is easy to see that the expression for @xmath80 $ ] can indeed be written in this way .
in addition , the comultiplication on @xmath67 is defined by @xmath81 formally , the coproduct is a homomorphism @xmath82 , where the tensor completion is described in section xvi.3 of @xcite .
we introduce the usual set of notations for @xmath83-numbers , @xmath83-factorials , and @xmath83-binomial coefficients : @xmath84_q&=&\displaystyle \frac { q^n - q^{-n}}{q - q^{-1}}=\frac { \sinh\left ( { { \hbar } n}\right ) } { \sinh\left ( { { \hbar } } \right)}\rule{0mm}{8 mm } \end{array } \qquad \begin{array}{rcl } \,[n]_q!&=&[n]_q[n-1]_q\ldots[2]_q[1]_q\\ { \bigl[{\begin{matrix}\raise 0.32ex \hbox{{\mbox{\fontsize{10}{1}$n$}}}\vspace*{-2.5 mm } \\ { \mbox{\fontsize{10}{1}$j$}}\end{matrix}}\bigr]}_q&=&\displaystyle \frac { [ n]_q!}{[n - j]_q![j]_q!}\rule{0mm}{8 mm } \end{array}\ ] ] note that all of these quantities are invertible in @xmath68 $ ] for @xmath85 .
a _ universal r - matrix _ for @xmath67 is now given as in theorem xvii.4.2 of @xcite is by @xmath86_q ! } e^n\otimes f^n \bigr ) \quad\in\ ; \alg_{\hbar}\widetilde\otimes \alg_{\hbar}\ ] ] drinfeld s construction from @xcite as described in @xcite implies that the _ r_-matrix from ( [ eq - univrmat ] ) makes @xmath67 into a quasi - triangular topological hopf algebra .
particularly , this implies that @xmath87 obeys the _ yang - baxter _ relation given as an equation in @xmath88 by @xmath89 moreover , @xmath90 fulfills the usual commutation relation in @xmath91 given by @xmath92 in order to construct representations of the braid groups we will need to consider first representations of @xmath67 . instead of distinguishing many representations by their highest weights we consider only one representation and `` absorb '' the highest weight as a parameter in the underlying coefficient ring as follows . in @xcite
the coefficient ring was chosen as @xmath93 , yet all calculations and statements there clearly also apply for any other choice of @xmath94 . for our purposes
we will choose the coefficient ring to be @xmath95 $ ] , that is , the polynomial ring with rational coefficients in one indeterminate @xmath96 which may be thought of as a generic highest weight .
@xmath67 is thus an algebra over @xmath97[[\hbar]]$ ] the ring of power series in @xmath98 whose coefficients are rational polynomials in @xmath96 . in this setting
@xmath67 admits a special highest weight module over the same ring described as follows .
consider the @xmath97$]-module @xmath99 freely generated by an infinite sequence of vectors denoted by @xmath100 .
the _ generic verma module _
$ ] is then the associate topologically free module by in the sense of section xvi.2 of @xcite .
the action of @xmath67 on @xmath102 is given by @xmath103_q\cdot[\lambda - j]_q\,v_{j+1}\nonumber\end{aligned}\ ] ] note here that indeed @xmath104_q=\frac { \sinh(\hbar(\lambda - j ) ) } { \sinh(\hbar ) } \in\,\mathbb q[\lambda][[\hbar]]\,$ ] .
the module is similar to the standard highest weight module obtained as the induced representation associated to the one - dimensional representation of the borel algebra generated by @xmath73 and @xmath75 acting on @xmath105 .
it is , however , not equivalent to this module since the elements @xmath106_q$ ] are not invertible in @xmath107[[\hbar]]\,$ ] .
( ring evaluations @xmath108 create additional highest weight vectors for the traditional verma module , but additional lowest weight vectors for the representation in ( [ eq - haction ] ) .
the modules are equivalent ( and irreducible ) only for evaluations @xmath109 .
as described in the end of sections xvi.4 any topological @xmath67-module such as @xmath102 now entails a solution to the yang - baxter equation on @xmath110 by ( [ eq - ybr ] ) , which commutes with the action of @xmath67 on the same space by ( [ eq - ybcomm ] ) . as an endomorphism on @xmath111 we define the action of a braid group generator by @xmath112 here
@xmath113 acts as an element of @xmath114 on @xmath111 , and @xmath115 denotes the usual transposition @xmath116 .
we also multiply the map by the unit @xmath117[[\hbar]]\,$ ] which also yields a solution to the yang baxter equation since this relation is homogeneous . the braid group @xmath118 is now represented on @xmath119 by the assignment @xmath120 the goal of the following constructions is to identify a sublattice in @xmath121 which is invariant under this action of the braid group @xmath0 . this lattice will be a free module over a subring @xmath122[[\hbar]]\,$ ] which is characterized as follows .
consider first the following ring homomorphism from the two - variable laurent polynomials to the power series ring : @xmath123\,\longrightarrow\,\mathbb
q[\lambda][[\hbar]]\;:\quad \begin{array}{rcl } q&\mapsto & e^{\hbar}\\ s&\mapsto & e^{\hbar\lambda}\;. \end{array}\ ] ] it is clear that @xmath124 is well defined by inspection of the power series expansion in @xmath125 , and that @xmath124 is a monomorphism since @xmath126 has dense image in @xmath127 .
we will thus denote the image of @xmath124 also by @xmath128\subset \,\mathbb q[\lambda][[\hbar]]\,$ ] with the identification of parameters as prescribed in ( [ eq - laurent2power ] ) . in order to find a suitable subalgebra over this ring we define next a set of special generators in @xmath67 by @xmath129_q!}$}}f^n\;.\ ] ]
the following relations readily follow from the ones given in ( [ eq - hrels ] ) : @xmath130}_q f^{(n+m)}\ , , \quad&{\rm and}\quad [ e , f^{(n+1)}]\;=\;f^{(n ) } ( q^{-n}k - q^{n}k^{-1})\;. \end{split}\ ] ] let now @xmath131 be the subalgebra over @xmath40 generated by the set of elements @xmath132 . as a sublattice @xmath18 is the free @xmath40-module spanned by the pbw basis @xmath133 .
in fact , @xmath18 is isomorphic to the algebra over @xmath40 defined abstractly by generators @xmath134 and the relations given in ( [ eq - urels ] ) .
the coproduct and antipode evaluated on the generators of @xmath18 are readily computed : @xmath135 these formulas immediately imply that the coproduct is in fact a map @xmath136 with @xmath137 taken over @xmath40 .
consequently , @xmath18 is a hopf subalgebra of @xmath67 , and thus a hopf algebra over @xmath40 by itself .
next , let @xmath138 be the free @xmath40-module generated by the basis vectors @xmath100 .
that is , an element in @xmath10 is given by @xmath139 with @xmath140 $ ] and only finitely many @xmath141 are non - zero .
the actions of the generators of @xmath18 on the basis vectors @xmath142 are easily worked out from the action of @xmath67 to be the following : @xmath143}_q \prod_{k=0}^{n-1 } ( sq^{-k - j}-s^{-1}q^{k+j})}\right ) v_{j+n}. \end{split}\ ] ] observe that all coefficients in these formulae lie in the subring @xmath144 $ ] and contain only a finite number ( one ) of vectors .
this immediately implies the following : [ eq - u - lat ] the subspace @xmath145 is invariant under the action of the subalgebra @xmath146 .
this also implies that the natural actions of @xmath18 as well as @xmath147 on @xmath119 map the respective subspace @xmath148 to itself .
the main observation of this section is that the same is true for the braid group action .
[ lm - r - lat ] the map @xmath149 , as defined in ( [ eq - braidrept ] ) , maps the subspace @xmath150 to itself .
we first note that the map @xmath151 can be written as the composite of three maps @xmath152 where @xmath115 is the usual transposition as in ( [ eq - braidrept ] ) .
the operator @xmath153 is given by the action of the factor @xmath154 from the expression in ( [ eq - univrmat ] ) for the universal @xmath155-matrix multiplied by the extra term @xmath156 that occurs in ( [ eq - braidrept ] ) . finally , @xmath157 is given by application of the remaining summation in parentheses in ( [ eq - univrmat ] ) .
we prove that each of these three operators in ( [ eq - r = tcp ] ) preserves @xmath158 as a subspace .
this is trivially true for @xmath115 . for the action of @xmath153
we compute @xmath159 thus @xmath160 and the claim follows for @xmath153 .
for @xmath157 we first rewrite the summation expression for the universal @xmath155-matrix in ( [ eq - univrmat ] ) in terms of the generators of @xmath18 . @xmath161_q ! } e^n\otimes f^n\ , \;=\ ;
\sum_{n=0}^{\infty}q^{\frac { n(n-1)}2 } e^{n}\otimes f^{(n)}\;.\ ] ] the fact that the action of @xmath73 on @xmath10 is locally nilpotent together with the observation that any finite truncation of the summation in ( [ eq - pexp ] ) yields an element in @xmath162 imply the claim for @xmath157 .
more specifically , the action of @xmath157 can be worked out explicitly to be the following .
@xmath163}_q \prod_{k=0}^{n-1 } ( sq^{-k - j}-s^{-1}q^{k+j } ) \,v_{i - n}\otimes v_{j+n}.\ ] ] since the summation is a finite one and all coefficients are in @xmath164 $ ] we can now infer that @xmath165 .
consequently , all three operators @xmath115 , @xmath153 , and @xmath157 map the subspace @xmath166 to itself , which proves the lemma . for future use
let us also record the explicit formula for the action of @xmath151 on @xmath158 .
@xmath167}_q \prod_{k=0}^{n-1 } ( sq^{-k - j}-s^{-1}q^{k+j } ) \,v_{j+n } \otimes v_{i - n}\ , .
\end{split}\ ] ] let us conclude this section with a summary of our findings . the maps @xmath168 , with the @xmath169 as in ( [ eq - intrmat ] ) , define a representation of the braid group @xmath0 on @xmath170 , as a free @xmath171$]-module .
the maps @xmath172 also commute with the action of @xmath18 on @xmath170 and preserve the natural @xmath173 grading . the fact that the maps @xmath172 preserve @xmath170 is immediate from lemma [ lm - r - lat ] .
they fulfill the braid group relations since they are restrictions of the maps @xmath174 from ( [ eq - sigmatilde ] ) which fulfill these relations by construction .
moreover , these maps commute with the action of @xmath67 and hence also with the action of @xmath18 .
the main purpose of this section is to prove the assertion in theorem [ thm - freemodule ] , namely , that the highest weight spaces are free @xmath40-modules . in order to define these highest weight spaces
let @xmath175 be the weight space corresponding to the weight @xmath176 .
we recall that @xmath177 acts on @xmath178 by @xmath179 . by and
, @xmath180 is the @xmath40-span of the vectors @xmath181 where @xmath182 .
we now define @xmath183 the space @xmath3 is the so - called _ highest weight space _ corresponding to the weight @xmath176 .
since the representation of @xmath0 on @xmath9 commutes with the @xmath18-action , we see that both @xmath180 and @xmath3 are also @xmath0 representations .
let us also define @xmath184 for @xmath185 by @xmath186 and @xmath187 we immediately see that @xmath188 . given a multi - index @xmath189 for some @xmath190 such that @xmath191
, we can define an element of @xmath192 by [ abasis ] a _ = v_0^(j-2 ) v_1 v _ where @xmath193 .
clearly , letting @xmath194 vary among all such multi - indices gives the basis @xmath192 of @xmath195 .
[ eonb ] for all @xmath196 and @xmath185 , the map @xmath197 is an @xmath40-linear isomorphism . to show that @xmath197 is surjective we need to show that for every @xmath198 there is some @xmath199 such that @xmath200 .
we proceed by induction on @xmath201 , where @xmath202 is the first nonzero entry in the multi - index @xmath203 .
the initial case , when @xmath204 , occurs when @xmath205 and is handled simply by observing that @xmath206 . to prove the induction step
let us take @xmath207 such that @xmath208 .
setting @xmath209 we see that @xmath199 and e.b = ( ) v _ + ( ) where the first nonzero index in each of the other terms is @xmath210 . hence
, the other terms satisfy the induction hypothesis and so are in the image of @xmath73 . from this
it follows that @xmath211 is in the image of @xmath73 . to show that @xmath212 has no kernel take some @xmath213 .
then @xmath214 will have some minimal term in its expression , namely , some @xmath215 where @xmath216 for all @xmath217 , @xmath218 , and if @xmath219 is in the expression for @xmath214 then @xmath220 for all @xmath221 and either @xmath222 or @xmath223 . then comparing the terms in the expression for @xmath224 we see that it is impossible to cancel out the term @xmath225 .
hence , @xmath226 .
when @xmath185 we define an @xmath40-linear map @xmath227 by defining it on basis elements as follows : @xmath228 the coefficients are given by [ c ] b _ , k = ( -1)^k-1 s^(k-1)(j - n-1 ) q^(k-1)(2l - k -2 ) .
notice that when @xmath229 in we have a multiple of the term @xmath230 .
by @xmath231 we mean the unique element @xmath232 such that @xmath233 .
such an element @xmath234 exists and is unique because of lemma [ eonb ] .
[ lm - phiiso ] we have @xmath235 so that @xmath236 is an automorphism of @xmath180 . clearly , we have @xmath237 for @xmath238 . for @xmath239
we have @xmath240 so that @xmath241 and hence @xmath242 .
the nilpotency relation immediately implies that @xmath243 is an inverse . under the change of basis on @xmath180 given by @xmath236 the operator @xmath73 has a simple form . [ lm - ephi - form ] for all @xmath196 and @xmath185
the composite @xmath244 vanishes on @xmath195 and is injective on @xmath245 with @xmath246 this implies that the following is an isomorphism of @xmath40-modules : @xmath247 the first half of the action in ( [ eq - decephi ] ) is to show @xmath244 is zero on any element @xmath248 which we verify by explicit computation : @xmath249 here we use that implies the recursion @xmath250 now ( [ eq - decephi ] ) follows from ( [ map ] ) , where @xmath236 is defined to be identity on @xmath251 . by lemma [ eonb ]
we have that @xmath212 is injective so that @xmath252 . since , by lemma [ lm - phiiso ] , @xmath236 is an automorphism of @xmath40-modules this implies ( [ eq - phiisoaw ] ) .
let us also describe the case @xmath58 more explicitly .
a basis of @xmath253 is given by @xmath254 the subspaces defined in ( [ eq - adef ] ) and ( [ eq - bdef ] ) are defined slightly different for @xmath58 , namely @xmath255 in this setting we have @xmath256 .
formula ( [ map ] ) thus yields a basis for @xmath60 given by vectors @xmath257 with these conventions it is easy to see that all previous lemmas in this section also apply to the case @xmath58 ( and trivially so to the case @xmath258 ) .
we are now in a position to prove theorem [ thm - freemodule ] , namely , that the highest weight spaces are free @xmath40-modules .
since ( [ eq - phiisoaw ] ) is an isomorphism of @xmath40-modules and @xmath195 is clearly a free module , also @xmath3 has to be a free @xmath40-module .
the rank is given by the number of vectors in the set of spanning vectors given in ( [ eq - adef ] ) , which is given by @xmath259 .
since the generators @xmath260 , as defined in ( [ eq - sigmatilde ] ) , map ( by @xmath18-equivariance ) each @xmath3 subspace to itself , lemma [ lm - ephi - form ] implies that the conjugate maps @xmath261 map @xmath195 to itself .
thus the representation of @xmath0 over @xmath40 given by @xmath262 is equivalent to the representation given by the maps @xmath263 .
suppose @xmath264 is the projection of @xmath180 onto @xmath195 along @xmath245 .
observe also that @xmath265 .
this yields the basic but useful formula : @xmath266 implicit to this formula is the method of calculating the action of a braid generator @xmath267 on a particular basis vector : 1 . for a basis vector @xmath268
determine @xmath269 by ( [ map ] ) .
2 . use ( [ eq - intrmat ] ) and ( [ eq - sigmatilde ] ) to determine the image of @xmath270 .
3 . write @xmath270 in the standard basis @xmath271 and eliminate the components of @xmath272 leaving an @xmath40-linear combination of vectors from @xmath192 . in the following we also consider the action of @xmath0 directly on @xmath273 .
a natural basis is given by @xmath274 . by construction the explicit action of the braid group generators @xmath275 in this basis is exactly the same as the action of the @xmath267 in the basis @xmath192 so that the computations remain the same .
here we prove that the representation of @xmath0 on @xmath5 is isomorphic the lkb representation which was recently shown in @xcite and @xcite to be faithful . as preparation let us show first that the representation @xmath60 is isomorphic to the classical , reduced burau representation over @xmath276\,$ ] .
the formula for the @xmath155-matrix in ( [ eq - intrmat ] ) implies @xmath277 , @xmath278 , and @xmath279 . applied to the basis @xmath280 of @xmath253 from ( [ eq - c - bas ] )
this implies the following action of @xmath0 on @xmath281 : @xmath282 using the rescaled basis @xmath283 and with a substitution of parameter @xmath284 the action from ( [ eq - sbur ] ) turns out to yield exactly the _ unreduced _ burau representation @xmath285 of dimension @xmath1 as described , for example , in ( 3 - 23 ) of @xcite .
thus we have by identification of basis vectors that @xmath286 now , the basis for @xmath60 from ( [ eq - defwl=1 ] ) may also be rescaled as @xmath287 recall that the _ reduced _ burau representation @xmath288 of dimension @xmath289 is given by the kernel of the map @xmath290:\,d_j\mapsto { \tt t}^j\,$ ] .
clearly , the basis described in ( [ eq - ubas ] ) is thus a basis also of @xmath288 and we obtain the following relation .
@xmath291 let us now turn to the @xmath61 case .
the basis @xmath292 from is given by elements @xmath293 correspondingly , the @xmath294 consists of the following elements : @xmath295 the basis @xmath41 for @xmath5 is given by application of the map in to @xmath292 which yields the following set of elements : @xmath296 the action of the braid group @xmath0 on these vectors is now computed using the step by step procedure following ( [ basiscalc ] ) .
in addition to the formulae in the previous paragraph this also involves calculating expressions for @xmath297 with @xmath298 . in each of these expressions
only the coefficients of the @xmath299-term needs to be considered since the contributions of the @xmath300-terms and @xmath301-terms will be projected out by @xmath302 .
the relevant relations are thus the following : @xmath303 applying ( [ eq - r2ord ] ) to the elements in ( [ eq - defaij ] ) and ( [ eq - defbk ] ) , and combining expressions in ( [ eq - defwij ] ) we can compute the action of @xmath0 on the basis vectors in @xmath41 according to the procedure given at the end of the previous section .
the resulting formulae for the generators of @xmath0 are listed next where we assume that @xmath304 : @xmath305 for comparison we consider the explicit lawrence krammer bigelow representation @xmath306 of @xmath0 as given in section 5.2 of @xcite .
( note that the representation given in @xcite contains a sign error which is corrected in @xcite ) .
there the space @xmath306 is described as the free @xmath307$]-module spanned by basis elements @xmath308 . from the formulae in @xcite the actions of the inverses of the generators of @xmath0 on @xmath306
are readily worked out to be as follows : @xmath309 let us define the map @xmath310 by @xmath311 where @xmath312 $ ] , @xmath313 , and @xmath314 is the ring homomorphism given in ( [ eq - deftheta ] ) .
it follows now by direct computation from the equations in ( [ eq - sigma.w ] ) and ( [ eq - sigma.f ] ) that @xmath315 where @xmath35 is the involution described in the introduction .
hence @xmath316 is @xmath0-equivariant by definition .
since it also maps basis vectors to basis vectors of free modules and @xmath314 is a monomorphism , @xmath317 can be considered a @xmath0-submodule of @xmath5 whose @xmath40-span is again @xmath5 .
this implies the isomorphism in theorem [ thm - main ] .
in this section we look more closely at the structure and decomposition of the verma module representations @xmath180 . more specifically , we look at eigenspace decompositions of @xmath180 under the operators @xmath318 . the main purpose of these decompositions is to allow us to prove the irreducibility of the highest weight representations in the next section .
recall that we have previously defined @xmath319 , the fraction field of @xmath40 . in
what follows , we will often speak of @xmath180 as a vector space over @xmath320 .
of course what we really mean is @xmath321 , but we will usually make no distinction .
we could , in the interest of generality , carry out our calculations over a smaller ring , essentially inverting only those elements of @xmath40 that are necessary , but this level of generality adds little to the discussion at hand .
the weight space @xmath180 splits as a @xmath322$]-module into a direct sum of highest weight spaces : _
n , l = _ k=0^l
f^(k ) _ n , l - k _ k=0^l _
n , l - k .
we already know that @xmath323 as @xmath40-modules .
so @xmath180 does decompose into a direct sum of highest weight spaces @xmath324 .
this decomposition does not preserve the braid group action , however . to prove the decomposition @xmath325 we proceed by induction on @xmath326 .
for @xmath258 we have an obvious identity .
suppose now that @xmath325 and take @xmath327 for some @xmath328 .
we apply @xmath318 to @xmath329 to obtain @xmath330_q e^t f^{(t+k ) } w \lb{irredeqn } \\ & = \bigl[\begin{matrix}t+k\vspace*{-1.6 mm } \\k\end{matrix}\bigr]_q f^{(k ) } \big ( \prod_{j=1}^t ( q^{j - k - t } k - q^{k+t - j } k^{-1 } ) \big ) w \notag \\ & = \bigl[\begin{matrix}t+k\vspace*{-1.6 mm } \\k\end{matrix}\bigr]_q \mu_{t , k}^{n , l } \ ; v \notag\end{aligned}\ ] ] where @xmath331 is the nonzero constant given by _
t , k^n , l= _ j=1^t ( s^n q^-2l+k - t+j - s^-n q^2l - k+t - j ) .
thus , in particular @xmath332_q \mu_{1,k}^{n , l } v$ ] .
since the constants @xmath333_q \mu_{1,k}^{n , l } \in \rqs$ ] are distinct for distinct @xmath334 , we see that the decomposition @xmath325 is the eigenspace decomposition of the transformation @xmath335 .
the eigenvalues @xmath333_q \mu_{1,k}^{n , l}$ ] are each nonzero , so we see that the map @xmath336 is injective .
the image of this map ( over the fraction field ) is @xmath337 and it is clear from that @xmath338 . counting dimensions , we see that @xmath339 .
having obtained a decomposition of @xmath180 , we would now like to obtain a similar decomposition of the highest weight spaces @xmath3 by restricting the braid action .
consider the @xmath340-action on @xmath341 .
the map @xmath342 defined by @xmath343 gives us an inclusion @xmath344 . in the standard basis of @xmath341
the elements of @xmath3 correspond to the vectors @xmath345 where @xmath346 for @xmath347 ( see and ) .
we also have the inclusion @xmath348 that takes @xmath349 to @xmath350 . with this identification
the inclusion @xmath344 is @xmath351-equivariant .
the quotient @xmath352 is isomorphic to @xmath353 as an @xmath354$]-module .
the isomorphism is given by ( a _ ) v_. let @xmath355 be the composition of the quotient map @xmath356 with the isomorphism given in .
we seek a splitting of @xmath357 .
let @xmath358 be recursively defined by setting @xmath359 and c_k , j+1 = c_k , j for each @xmath360 we define a map @xmath361 by _ k : w _
j=0^k c_k , j f^(k - j ) ( v_j w ) .
let us take @xmath328 and compute the action of @xmath73 on @xmath362 : @xmath363 thus , @xmath202 actually maps @xmath364 into @xmath365 . in the standard basis of @xmath341
the element @xmath366 corresponds , modulo @xmath3 , to a multiple of @xmath367 . to be more precise : _ k
w = _ k ( v_1 f^(k-1)w ) _ n , l where @xmath368 thus , using the identification @xmath369 given by lemma [ decomp ] we see that @xmath370 acts on @xmath364 as multiplication by the nonzero constant @xmath371 . define a map @xmath372 by = _
k. the previous discussion yields the following : the map @xmath373 defines a @xmath351-equivariant splitting of the map @xmath355 .
this gives a decomposition as @xmath351-modules _
n+1,l = _ k=0^l _
n , l - k .
in this last section we wish to prove theorem [ thm - irred ] , namely that the highest weight representations @xmath3 are irreducible over the fraction field @xmath320 .
the proof makes use of the decompositions of the previous section and proceeds by induction on @xmath1 .
notice that , in the general case , if @xmath374 is a @xmath0-submodule , then as a @xmath375-module it must decompose into a direct sum of lower degree submodules following the decomposition @xmath376 by the induction hypothesis , each of these summands is an irreducible representation of @xmath375 so that @xmath377 must be a direct sum of some collection of these @xmath378 ( for more detail see the proof at the end of the section ) . in what follows
we give explicit computations of the action of @xmath379 on certain elements of these components .
these computations show that we must , in fact , have @xmath380 for all @xmath381 , thus proving the theorem . for any @xmath386
we apply @xmath387 to @xmath329 to obtain @xmath388 so that we can solve recursively for @xmath385 : w_t = ( e^t v - _ t,1^n , l - t f^(1 ) w_t+1 - - _ t , l - t^n , l - t f^(l - t ) w_l ) .
proceeding by induction , we see that we must have w_t = _ i=0^l - t z_t , i^n , l f^(i ) e^t+i v for some coefficients
we see from that @xmath390 and an induction argument shows that , in general , z_t , i^n , l|_q=1 = ( -1)^i z_t+i,0^n , l|_q=1 = ( -1)^i ( s^n - s^-n)^-t - i . in particular , the coefficients @xmath391 are never zero .
[ example ] let us define @xmath392 . then , as above , _
j= w_j,0 + f^(1 ) w_j,1 +
+ f^(j ) w_j , j for some @xmath393 .
let us use to define _
j w_j,0 = _
i=0^j z_0,i^n , j f^(i ) e^i _ j. in other words , @xmath394 is the first term of @xmath395 in the decomposition @xmath396 . since @xmath397
, we see from that @xmath398 for all @xmath381 .
also , from we see that @xmath399 .
thus , equation can be written as _
i=0^j s^(n-1)i z_i,0^n , j f^(i ) _ j - i . and then defining for @xmath409 _ j , k , r = ( [ k - r]_q s _ j , k-1,r - [ j+1]_q ( s^2 q^-j - q^j ) _
j+1,k-1,r-1 ) .
the verification of this fact follows by an induction argument from the identity _
j , k = f^(1 ) _ j , k-1 -(s^2 q^-j - q^j ) _ j+1,k-1 .
using the @xmath229 case , equation becomes @xmath410_q f^{(k - r+t ) } \omega_{j+r - t } \\ & = \sum_{i=0}^{j+k } \gamma_{j , k , i } f^{(i ) } \omega_{j+k - i}\end{aligned}\ ] ] where @xmath411_q.\ ] ] now , we would like to know something about these coefficients @xmath412 .
at least we would like to know that they are nonzero .
the relation in along with an easy induction argument show that @xmath413 is a polynomial in @xmath21 of the form @xmath414 also , from we have @xmath415 so for each @xmath416 with @xmath417 we see that setting @xmath43 in @xmath418_q$ ] will indeed give us a laurent polynomial in @xmath21 with smallest degree term @xmath419 since the degree of this term is positively related to @xmath420 , the overall smallest degree term of @xmath412 will occur when @xmath420 is as small as possible . for @xmath421 , the smallest @xmath420 may be is 0 and in this case we also have @xmath422 . for @xmath423
the smallest @xmath420 may be is @xmath424 and in this case we have @xmath425 .
this proves and the lemma .
suppose we take @xmath328 to be the basis vector given by @xmath426 for some @xmath427 and where the coefficients @xmath428 are given as in .
we then have @xmath429 let us ( temporarily ) set @xmath430 we act on @xmath366 by @xmath431 and compute _ 1 ( _ k w ) = _ h=0^ _ j=0^k _
t=0^l - k c_k , j b_t d_h , j , t f^(k - j ) ( f^(h ) v_t
e^h v_j e^t-1u ) .
recall that @xmath355 is the map that first mods out by @xmath3 , then projects to @xmath432 , then removes the leading @xmath433 component in the tensor product .
so applying @xmath357 to @xmath434 , the only terms to survive are those for which @xmath435 , and we obtain ( _ 1 ( _ k w ) ) = _ j=0^k f^(k - j)(_j v_j u + b_0 _ j v_j-1 e^-1u ) . where the coefficients @xmath436 and @xmath437 are calculated to be @xmath438
let us take @xmath445 and define @xmath446 where @xmath447_q \mu_{1,l - k-1}^{n-1,l - k-1}}.\ ] ] then @xmath448 , which follows by the computations found in and the proof of theorem [ thm - freemodule ] , and comparing the expressions of @xmath449 and @xmath450 in the standard basis we see that @xmath450 is a nonzero multiple of @xmath449 .
hence if we prove the lemma for @xmath449 then it will also follow for @xmath450 .
notice that the discussion following lemma [ splitting ] will apply formally to @xmath449 if we make the substitutions @xmath451 and @xmath452 .
thus in the present case becomes ( _ 1 ( _ k ) ) = _ j=0^k _ j f^(k - j)(v_j u ) + _
j=1^k x_0 _ j f^(k - j)(v_j-1 f^(1 ) u ) . we have @xmath453_q f^{(k - j+1)}(v_{j-1 } \otimes u ) \nonumber \\ & \phantom{adad } - s q^{-2(j-1 ) } ( s q^{1-j } s^{-1 } q^{j-1 } ) [ j]_q f^{(k - j)}(v_j \otimes u)\end{aligned}\ ] ] which allows us to write as ( _ 1 ( _ k ) ) = _ j=0^k _ j f^(k - j)(v_j u ) where _ j = _
j + x_0 s q^-2(j-1 ) ( _ j+1 q^-2 [ k]_q - _ j ( s q^1-j - s^-1 q^j-1 ) [ j]_q ) which makes sense for all @xmath454 so long as we define @xmath455 .
we apply lemma [ lem - irr ] to obtain ( _ 1 ( _ k ) ) = _ j=0^k _ i=0^j+l - k-1 _ j _ j , l - k-1,i _ q f^(k - j+i ) _ j+l - k-1-i .
thus , if @xmath456 where @xmath457 , then for all @xmath458 we will have _
r = _ i=0^r _ k - r+i _ k - r+i , l - k-1,i _ q _ l-1-r .
thus , to complete the proof we need only show that the coefficient in is nonzero for all @xmath459 .
but we can check that @xmath460 is a laurent polynomial in @xmath21 having smallest degree term @xmath461 compare this to the minimum degree term of @xmath462 given via and we see that , after multiplying by an appropriate constant not depending on @xmath463 , the smallest degree term of @xmath464 is a strictly decreasing function of @xmath463 ( details are left to reader ) . thus , @xmath465_q$ ] must be nonzero .
we give the standard basis of @xmath3 the following `` lexicographical '' ordering : @xmath470 where by @xmath471 we mean the number of components in the multi - index ( that is , if @xmath472 then @xmath473 ) and by @xmath474 we mean the usual lexicographical ordering on ordered tuples of integers that have the same number of components .
the element @xmath475 is the maximal element in this ordering and the braid @xmath439 acts on @xmath475 as the invertible constant @xmath476 now take @xmath477 . a simple case - by - case analysis shows that there is a polynomial @xmath478 of degree at most 2 such that @xmath479 either belongs to @xmath480 or is a sum of strictly higher order terms . for instance
, suppose @xmath481 and consider @xmath482 for some @xmath483 and some @xmath484
. then we compute @xmath485 for some @xmath486 . when @xmath487 we have @xmath488 for a nonzero constant @xmath489 so that we will have @xmath490 similar arguments apply in the remaining cases .
we proceed by induction on @xmath1 .
the base case when @xmath491 is trivial since the dimension of @xmath492 is 1 for all @xmath46 .
suppose now the theorem is true for all @xmath493 .
suppose @xmath374 is a @xmath0-submodule . as a @xmath375-module
we have seen that @xmath3 decomposes into a direct sum @xmath376 by the induction hypothesis , each of these summands is an irreducible representation of @xmath375 and since the dimensions are different they are inequivalent .
( note , in case n=3 the summands are isomorphic , but the braid action is given by distinct eigenvalues on distinct summands . ) thus @xmath377 must be a direct sum of some collection of these @xmath378 . hence , to complete the proof all that remains is to show that @xmath500 .
but by lemma [ wmax ] any @xmath501 can be multiplied by a polynomial @xmath502 to obtain a multiple of @xmath475 . for instance , if we take @xmath503 to be the image under @xmath373 of the maximal basis element in @xmath504 , then @xmath505 will be a nonzero multiple of @xmath475 .
but @xmath506 so we see that @xmath377 must equal @xmath3 . | we construct representations of the braid groups @xmath0 on @xmath1 strands on free @xmath2$]-modules @xmath3 using generic verma modules for an integral version of @xmath4 .
we prove that the @xmath5 are isomorphic to the faithful lawrence krammer bigelow representations of @xmath0 after appropriate identification of parameters of laurent polynomial rings by constructing explicit integral bases and isomorphism .
we also prove that the @xmath0-representations @xmath3 are irreducible over the fractional field @xmath6 . |
Subsets and Splits
No saved queries yet
Save your SQL queries to embed, download, and access them later. Queries will appear here once saved.