article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
the universe acceleration , shown by several astronomical observations , indicates the existence of a mysterious exotic matter called dark energy ( de ) @xcite . in the classical gravity , whereas the cosmological constant , @xmath0 , is the most prominent candidate of de with the equation of state ( eos ) parameter equals -1 @xcite , there are strong evidence for a dynamical de equation of state .
however , the problem of de , its energy density and eos parameter is still an unsolved problem in classical gravity and may be in the context of quantum gravity we achieve a more inclusive insight to its properties @xcite .
the hde model is an attempt to apply the nature of de within the framework of quantum gravity @xcite .
the holographic principle states that the number of degrees of freedom describing the physics inside a volume ( including gravity ) is bounded by the area of the boundary which encloses this volume and thus relates to entropy scales with the enclosing area of the system @xcite .
since the entropy scales like the area rather than the volume , the fundamental degrees of freedom describing the system are characterized by an effective quantum field theory in a box of size @xmath1 with one fewer space dimensions and with planck - scale uv cut - off @xmath0 @xcite .
as pointed out by the author in @xcite , in quantum field theory , short distance cut - off @xmath0 is related to long distance cut - off @xmath1 due to the limit set by forming a black hole . in other words , if @xmath2 is the quantum zero - point energy density caused by a uv cut - off @xmath0 , the total energy of the system with size @xmath1 should not exceed the mass of the same size black hole , i.e. @xmath3 , where @xmath4 denotes planck mass @xcite . for the largest @xmath1 , holographic energy density is given by , @xmath5 in which c is a free dimensionless parameter and coefficient three is for convenience @xcite .
assuming @xmath1 as the universe current size , the hubble radius @xmath6 , then the de density is close to the observational result , which we identify with the radius of the future event horizon , @xmath7 where @xmath8 is scale factor of the universe @xcite .
choosing the future event horizon as the uv cut - off allows us to construct a satisfactory hde model for the universe accelerating expansion which presents a dynamical view of the de and is consistent with observational data @xcite .
the interacting hde ( ihde ) models have been widely investigated in last few years @xcite . among all , the stability analysis of the ihde solutions and their dynamical characteristics @xcite , ihde due to the bulk - brane interaction @xcite and ihde in non - flat cosmological model @xcite are of particular interest .
in particular , relevant works in the holographic tachyon models in both interacting and non interacting cases can be found in @xcite@xcite@xcite . in here , we would like to extend the previous works carried out in the ihde models , by studying a chameleon - tachyon cosmological model . in our model
the scalar field in the formalism plays two roles ; as a chameleon field interacts with the matter in the universe and as a tachyon field plays the role of de . in the previous works in ihde models ,
the interacting term between ihde and dark matter ( dm ) is not unique and its introduction in the formalism based on phenomenological grounds in studying a suitable coupling between quintessence scalar field and pressureless cold dark matter ( cdm ) to alleviate the coincidence problem @xcite .
however , in our work the interacting term naturally appears in the model from the interaction between chameleon field and matter field in the universe in the formalism .
we consider the chameleon gravity in the presence of cdm with a tachyon potential in the action given by@xcite@xcite , @xmath9\sqrt{-g}d^{4}x,\end{aligned}\ ] ] where @xmath10 is ricci scalar . unlike the usual einstein - hilbert action
, the matter lagrangian @xmath11 is modified as @xmath12 , where @xmath13 is an analytic function of @xmath14 .
this last term in the lagrangian brings about the nonminimal interaction between the matter and chameleon field .
the variation of action ( [ action ] ) with respect to the metric tensor components in a spatially flat frw cosmology yields the following field equations , @xmath15 where @xmath16 . in the above
, we also assumed a perfect fluid filled the universe with the eos @xmath17 .
we can rewrite the above equations as , @xmath18 where @xmath19 and @xmath20 are respectively the energy density and pressure of the tachyon .
the variation of the action ( [ action ] ) with respect to the scalar field @xmath14 provides the wave equation for the scalar field as @xmath21 where prime `` @xmath22 '' indicated differentiation with respect to @xmath14 and @xmath23 . the scalar field as a tachyon field plays the role of ihde and by coupling with the matter field as a chameleon field plays the role of dm .
then , by their interaction , the energy densities no longer satisfy independent conservation laws , rather , they obey : @xmath24 @xmath25 where `` ch '' stands for chameleon , @xmath26 and @xmath27 is the interaction term . in @xmath28 , @xmath29 gauges the intensity of the coupling between matter and chameleon field . for @xmath30 , there is no interaction between dm and ihde .
the @xmath28 term measures the different evolution of the dm due to its interaction with the ihde which gives rise to a different universe expansion .
the remarkable point concerning the interaction term is that in comparison to the other ihde models where the form of the interaction term @xmath28 is not unique and built into the formalism from the outset , in our model the interaction term naturally appears in the model directly as a function of the chameleon coupling function @xmath13 and @xmath31 and indirectly as a function of hubble parameter @xmath32 and @xmath19 . for the decay of hde into dm in the interaction we have to take @xmath33 to be positive similar to the case of @xmath34 in the interaction term @xmath35 introduced in other works @xcite . in order to impose the holographic nature to the tachyon
, one should identify the energy density of tachyon field with the energy density of the cosmological constant . from the definition of the holographic energy density , equation ( [ rol ] ) , and that of the future event horizon ( [ re ] ) we obtain @xcite@xcite , @xmath36 where the derivative in here and from now on is with respect to @xmath37 , also , @xmath38 and @xmath39 .
we can rewrite equation ( [ ct ] ) as @xmath40 from equations ( [ rho ] ) and ( [ ro1 ] ) we drive the dynamical eos parameter as @xmath41 we have to note , however , that although the above eos parameter attributed to the tachyon field , the presence of the @xmath13 in the interaction term is directly related to the chameleon field and dm .
so the situation is somewhat ambiguous due to entanglement of the de and dm . under these conditions
, one can not easily classify the quintessence or phantom unless the interacting term is very weak according to observations . using equations ( [ ro1 ] ) and ( [ omegat ] ) and the definition of @xmath42 we find @xmath43 also , by replacing @xmath44 and @xmath45 into the acceleration equation ( [ fried2 ] ) , one obtains , @xmath46 from equations ( [ hz ] ) and ( [ hz2 ] ) , we drive the evolution equation for @xmath42 in terms of the redshift @xmath47 as , @xmath48,\end{aligned}\ ] ] which controls the dynamics of the ihde model . from equation ( [ hz2 ] )
, the evolution of the hubble parameter @xmath49 with respect to the redshift @xmath47 can be expressed as @xmath50 using equation ( [ fried11 ] ) , and @xmath51 , we derive the interacting holographic tachyon potential , @xmath52 where @xmath53 is the critical energy density of the universe at the present epoch . in addition , using the relation @xmath54 , the evolution of the interacting holographic tachyon scalar field @xmath14 with respect to the redshift z becomes , @xmath55 where the sign is arbitrary and can be changed by a redefinition of the field , @xmath56 .
then , by fixing the field amplitude at the present era to be zero , one may obtain the dynamic of the holographic tachyon field .
it is difficult to solve equations ( [ omegaz])-([phi ] ) analytically , however , the evolutionary form of the interacting holographic tachyon field @xmath14 and @xmath42 can be easily obtained by integrating it numerically from @xmath57 to a given value @xmath47 .
in addition , from the constructed holographic tachyon model , the evolution of @xmath58 with respect to @xmath14 can be determined . in the following numerical computation , we assume that the function @xmath13 behaves exponentially as @xmath59 . since @xmath60 is present in the interaction term @xmath28 , the parameters that directly determine the dynamics of the interaction are @xmath61 and @xmath33 together with @xmath31 and @xmath62 .
note that @xmath63 or @xmath64 leads to the absence of the interaction . in fig . 1 and 2 ,
by solving the equations ( [ omegaz])-([phi ] ) , the reconstructed scalar field @xmath65 and @xmath58 are shown .
the curves are plotted for different values of @xmath61 , @xmath66 , @xmath67 and with the fractional matter density selected to be @xmath68 . from fig .
2 , the reconstructed tachyon potentials are more steep in the early epoch ( @xmath69 ) and tending to be flat in near future when the scalar field @xmath14 is negative .
therefore , the tachyon scalar field rolls down the potential more slowly with the kinetic energy gradually decreasing in near future where from fig .
3a the tachyon eos parameter become tangent to -1 when @xmath70 .
this scenario is similar to the holographic quintessence scenario @xcite .
note that , from fig.1 , the tachyon scalar field increases with @xmath47 , becomes finite at high redshift and decreases as the universe expands .
note also that , due to smallness of interaction term , the behavior of the holographic scalar field is similar in the presence and absence of the interaction term which is compatible with the observational data @xcite . + , for different parameters @xmath61 , @xmath66 , @xmath67 and @xmath68 where @xmath14 is in unit of @xmath71 . ( a )
@xmath72 , ( b ) @xmath73 , ( c ) @xmath74,title="fig : " ] , for different parameters @xmath61 , @xmath66 , @xmath67 and @xmath68 where @xmath14 is in unit of @xmath71 .
( a ) @xmath72 , ( b ) @xmath73 , ( c ) @xmath74,title="fig : " ] , for different parameters @xmath61 , @xmath66 , @xmath67 and @xmath68 where @xmath14 is in unit of @xmath71 . ( a )
@xmath72 , ( b ) @xmath73 , ( c ) @xmath74,title="fig : " ] , for different parameters @xmath61 , @xmath66 , @xmath67 and @xmath68 where @xmath14 is in unit of @xmath71 and @xmath58 in @xmath75 .
( a ) @xmath72 , ( b ) @xmath73 , ( c ) @xmath74,title="fig : " ] , for different parameters @xmath61 , @xmath66 , @xmath67 and @xmath68 where @xmath14 is in unit of @xmath71 and @xmath58 in @xmath75 .
( a ) @xmath72 , ( b ) @xmath73 , ( c ) @xmath74,title="fig : " ] , for different parameters @xmath61 , @xmath66 , @xmath67 and @xmath68 where @xmath14 is in unit of @xmath71 and @xmath58 in @xmath75 . ( a )
@xmath72 , ( b ) @xmath73 , ( c ) @xmath74,title="fig : " ] in addition , the dynamics of the ratios of the energy densities of the hde , @xmath42 , and dm , @xmath76 , to the critical energy density in the universe are shown in fig.3b .
the graph shows that the impact of the interaction between ihde and dm increases at a faster rate compared to the non - interacting case .
it also shows that in the presence of interaction the domination of hde occurs later in @xmath47 . + and @xmath42 with respect to redshift @xmath47 for the holographic tachyon model with and without interaction .
we put @xmath77 and @xmath78,title="fig : " ] and @xmath42 with respect to redshift @xmath47 for the holographic tachyon model with and without interaction .
we put @xmath77 and @xmath78,title="fig : " ] from the above equations one can easily obtain the total eos and deceleration parameters @xmath79 for the model as , @xmath80 in fig .
4 , the deceleration parameter for different parameters @xmath61 , @xmath66 and @xmath67 are shown .
4a ) and 4c ) , for a fixed @xmath66 , show the effect of interaction term on the acceleration of the universe . from these graphs , since in the interaction the ihde decays into dm , the cosmic acceleration starts later with interaction compare to the case without interaction since ihde dominates later .
also for larger interaction , since more ihde decays into dm , the acceleration starts later .
, for different parameters @xmath61 , @xmath66 , @xmath67 and @xmath68 .
( a ) @xmath72 , ( b ) @xmath73 , ( c ) @xmath74,title="fig : " ] , for different parameters @xmath61 , @xmath66 , @xmath67 and @xmath68 . ( a )
@xmath72 , ( b ) @xmath73 , ( c ) @xmath74,title="fig : " ] , for different parameters @xmath61 , @xmath66 , @xmath67 and @xmath68 .
( a ) @xmath72 , ( b ) @xmath73 , ( c ) @xmath74,title="fig : " ] noting that in the constructed holographic model , equations ( [ omegaz])-([phi ] ) satisfy the condition @xmath81 . on the other hand ,
the observational data constrains the @xmath66 value .
however , the latest observational data , including the sne ia , the cmb shift parameter given by the wmap observations , and the bao measurements from sdss , show the possibilities of both @xmath82 and @xmath83 , and their likelihood are almost equal within @xmath84 error range@xcite@xcite .
luminosity distance quantity , @xmath85 , determines de density from observations . in our model with the @xmath49 obtained from numerical calculation , we have @xmath86 the difference between the absolute and apparent luminosity of a distance object called distance modulus , @xmath87 , and is given by @xmath88 . in fig . 5 , we compare our model for the distance modulus with the observational data . here
, we have used recent observational data , including sne ia which consists of 557 data points belonging to the union sample@xcite . + , with respect to the redshift @xmath47 for two different values of @xmath89 , in comparison with the observational data .
we put @xmath90 and @xmath91 the graph shows that for @xmath81 the distance modulus obtained from numerical calculation for the model better fits the observational data .
in this work we have proposed a holographic model of tachyon - chameleon cosmology by taking into account the holographic principle of quantum gravity ; by implementing a combination of holographic de model and interacting tachyon chameleon fields . in @xcite
the holographic dark energy is described by the tachyon field in a certain way .
then a correspondence between the hde and tachyon model of dark energy has been established in two different cases , flat and non - flat frw universes .
the author has shown that the value of @xmath66 is a key point to the feature of the hde and the ultimate fate of the universe as well . in holographic models ,
one considers that the scalar field models of de are effective theories of an underlying theory of de .
in particular , the tachyon scalar field carries some holographic feature that with @xmath92 is capable of realizing the holographic evolution of the universe . in our model , by connection between the tachyon and the holographic de , the potential of the holographic tachyon has been constructed .
although the effect of the interaction term is very small , the possibility of a small , but calculable , interaction in the dark sector remains open , and future investigations , with more realistic models and more observational data are necessary to solve this problem @xcite .
the scalar field in our formalism plays two roles ; as a chameleon field interacts with the matter in the universe and as a tachyon field plays the role of de . in the more familiar models ,
the interacting term between ihde and dm is introduced in the formalism based on phenomenological grounds in studying a suitable coupling between quintessence scalar field and pressureless cdm to alleviate the coincidence problem @xcite . in our work ,
different from other approaches , the interacting term naturally appears in the model from the interaction between chameleon and matter fields in the universe .
the interaction terms is similar to the proposed ones introduced in the ihde models .
a comparison of the model with recent observational data shows that for the parameter @xmath66 greater than one , the model for distance modulus better fits the observational data . in @xcite and
@xcite the author considered @xmath1 measured from the sphere of the horizon , as system s ir cut - off and obtained the equation of state for the interacting holographic energy density in the non - flat universe .
further , in @xcite the effects of the interaction between a brane universe and the bulk and in @xcite , between holographic energy density and cdm are studied . in our paper , the tachyon field in connection with ihde shows a quintessence like equation of state parameter in the past that become tangent to the phantom divide line in future .
the deceleration parameter in the model shows a transition from deceleration to acceleration at about @xmath93 in the past for different model parameters compatible with recent observational data . | we propose in this paper an interacting holographic dark energy ( ihde ) model in chameleon tachyon cosmology by interaction between the components of the dark sectors . in the formalism
, the interaction term emerges from the scalar field coupling matter lagrangian in the model rather than being inserted into the formalism as an external source for the interaction .
the correspondence between the tachyon field and the holographic dark energy ( hde ) densities allows to reconstruct the tachyon scalar field and its potential in a flat frw universe .
the model can show the accelerated expansion of the universe and satisfies the observational data . |
as you know , the universe becomes almost homogeneous and isotropic , when viewed on a sufficiently large scale .
this suggests very strongly that it may be described approximately by a robertson - walker - friedmann metric .
the crucial parameters describing the expansion of the universe are then its density @xmath1 and the curvature of the universe ; @xmath2 or @xmath3 for a critical , closed or open universe , respectively .
if @xmath4 , the density must equal the critical density @xmath5 g/@xmath6 , where @xmath7 is the present hubble expansion rate in units of 100 km / s / mpc .
much of the subsequent cosmological discussion is phrased in terms of the density budget of the universe , expressed as contributions relative to the critical density : @xmath8 .
@xmath9 : inflation suggests that this is practically indistinguishable from unity : @xmath10 @xcite , although there are some models that predict @xmath11 @xcite .
however , the data on the small anisotropies in the cosmic microwave background ( cmb ) @xcite support the inflationary suggestion that @xmath12 and @xmath13 , as summarized in fig . 1 @xcite .
@xmath14 : measurements of the @xmath15 ratio in high - redshift lyman-@xmath16 clouds @xcite correspond to = ( 3.3 0.3)10 ^ -5 [ one ] if this is indeed the correct primordial @xmath15 ratio , big - bang nucleosynthesis calculations suggest that @xcite = ( 5.1 0.3 ) 10 ^ -10 [ two ] corresponding to _
b h^2 = 0.019 0.001 [ three ] using the currently favoured range @xmath17 , we see from ( [ twenty ] ) that @xmath18 , which is insufficient to explain all the matter density in the following paragraph .
@xmath19 : the cluster measurements ( @xmath20 ratio , present and past abundances , cluster dynamics and the baryon fractions inferred from @xmath21-ray measurements ) all suggest @xcite _ m ~0.2 to 0.3 [ four ] as also seen in fig .
moreover , the combination of cmb measurements and high - redshift supernovae @xcite also support independently such a value for @xmath19 .
@xmath22 : the theory @xcite of large - scale structure formation strongly suggests that most of @xmath19 is cold dark matter , so that _ cdm ~_m [ five ] as perhaps provided by the supersymmetric particles discussed later .
@xmath23 : the theory @xcite of structure formation also suggests that the density of hot dark matter @xmath24 .
the present and prospective sensitivities of cosmological data to @xmath25 are shown in fig
. 2 @xcite .
so far , @xmath26 3 ev is excluded by the available upper limit on the density of hot dark matter , whereas the possible comparison of future data on large - scale structure and the cmb are thought to be sensitive to @xmath27 0.3 ev .
this is somewhat above the range @xmath28 0.1 to 0.03 ev favoured by the atmospheric neutrino data , but one should not abandon hope of detecting neutrino masses astrophysically @xcite . as discussed in section 2 ,
the indications of neutrino masses from atmospheric and solar neutrino data can most easily be explained by light neutrinos : @xmath29 0.1 ev , which would make only a small contribution to @xmath9 .
@xmath30 : if one follows the inflationary path supported by the cmb @xcite , so that @xmath12 , and takes at face value the suggestions from cluster measurements that @xmath31 , then the largest fraction of the energy density of the universe may be provided by vacuum energy : @xmath32 .
this scenario is supported by the recent high - redshift supernova data @xcite shown in fig . 3 , which suggest that @xmath33 .
combining this estimate with the suggestion of inflation that @xmath34 , one recovers independently the preference for @xmath35 .
a remarkably consistent picture of the density budget of the universe may be emerging : _ tot 1 = _
m + _ : + _ m ~0.3 , _ ~0.7 [ six ] where _ m = _ cdm + _ + _ b [ seven ] with _
b < 0.1 , _ _ cdm _ m [ eight ] it remains to be seen whether future data confirm this picture . for the moment , let us examine particle candidates for @xmath23 and @xmath22 , emphasizing their cosmic - ray manifestations .
if these are non - zero , laboratory experiments tell us that they must be much smaller than those of the corresponding charged leptons @xcite : m__e 2.5 ev , m _ _ 160 kev , m _ _ 15 ev , [ nine ] so one might think naively that they should vanish entirely . however , theorists believe that particle masses can be strictly zero only if there is a corresponding conserved charge associated with an exact gauge symmetry , which is not the case for lepton number . indeed , non - zero neutrino masses appear generically in grand unified theories ( guts ) @xcite . however , it is not even necessary to postulate new particles to get @xmath36 .
these could be generated by a non - renormalizable interaction among standard model particles @xcite : [ ten ] where @xmath37 is some new , heavy mass scale .
the most plausible guess , though , is that this heavy mass is that of some heavy particle , perhaps a right - handed neutrino @xmath38 with mass @xmath39 . in this case
, one expects to find the characteristic see - saw @xcite form of neutrino mass matrix : ( _ l , _
r ) ( ) ( ) [ eleven ] where the off - diagonal matrix entries in ( [ eleven ] ) break su(2 ) and have the form of dirac mass terms , so that one expects @xmath40 . diagonalizing ( [ eleven ] ) , one finds generically a light neutrino mass m _ [ twelve ] choosing representative numbers @xmath41 10 gev , @xmath42 ev , one finds @xmath43 gev , in the general ballpark of the grand unification scale . as you know
, data on both solar and atmospheric neutrinos favour neutrino oscillations associated with neutrino mass differences : @xmath44 in the solar case and @xmath45 in the atmospheric case .
( both of these can be regarded as cosmic - ray phenomena ! )
there are three possible interpretations of the solar - neutrino data : vacuum oscillations with @xmath46 ev@xmath47 and large mixing , and matter - enhanced msw oscillations with @xmath48 ev@xmath47 and either large or small mixing , as seen in fig
. 4 @xcite .
there is no hint what ( combination of ) other flavours the @xmath49 might be oscillating into . in the atmospheric case ,
@xmath50 to @xmath51 ev@xmath47 and large mixing are required .
the super - kamiokande and chooz data both exclude @xmath52 dominance , and zenith - angle distributions in the super - kamiokande data favour @xmath53 over @xmath54 oscillations into sterile neutrinos @xmath55 .
the past year has witnessed many theoretical studies of neutrino masses @xcite , of which i now pick out just a few key features : + + _ other light neutrinos ? _ we know from the lep neutrino - counting constraint @xcite , that any additional neutrinos must be sterile @xmath55 , with no electroweak interactions or quantum numbers .
but if so , what is to prevent them from acquiring large masses : @xmath56 with @xmath57 , as for the @xmath38 discussed above ? in the absence of some new theoretical superstructure , this is an important objection to simply postulating light @xmath55 or @xmath38 .
+ + _ majorana masses ? _ most theorists expect the light neutrinos to be essentially pure @xmath58 , with only a small admixture @xmath59 of @xmath38 . in this case
, one expects the dominant effective neutrino mass term to be of majorana type @xmath60 , as given by ( [ ten ] ) or ( [ twelve ] ) .
+ + _ large mixing ? _ small neutrino mixing used perhaps to be favoured , by analogy with the cabibbo - kobayashi - maskawa mixing of quarks .
however , theorists now realize that this is by no means necessary . for one thing ,
the off - diagonal entries in ( now considered as a 3@xmath613 matrix ) ( [ twelve ] ) need not be @xmath62 or @xmath63 @xcite . moreover , even if @xmath64 , we have no independent evidence that mixing is small in the lepton sector .
finally , even if @xmath65 were to be approximately diagonal in the same flavour basis as the charged leptons @xmath66 , why should this also be the same case for the heavy majorana matrix @xmath0 @xcite ? + + _ could neutrinos be degenerate ?
are masses @xmath67 2 ev and close to the direct and astrophysical limits allowed @xcite ?
any such scenario would need to respect the stringent constraint imposed by the absence of @xmath68 decay @xcite : _ e c^2_12 c^2_13 e^i + s^2_12c^2_13
e^i + s^2_130.2 ev [ thirteen ] in view of the upper limit on @xmath69 mixing from the chooz experiment @xcite , let us neglect provisionally the last term in ( [ thirteen ] )
. in this case , there must be a cancellation between the first two terms , requiring @xmath70 , and c^2_12-s^2_12 = 2_12 0.1 ^2 2_12 0.99 [ fourteen ] thus maximal @xmath71 mixing is necessary if the neutrinos are heavy and degenerate .
this certainly excludes the small - mixing - angle msw solution and possibly even the large - mixing - angle msw solution , since this is not compatible with @xmath72 ( which would yield a constant energy - independent suppression of the solar neutrino flux ) , and global fits typically indicate that @xmath73 0.97 , as seen in fig
. 5 @xcite .
global fits before the new super - kamiokande data on the energy spectrum indicated that @xmath74 was possible for vacuum - oscillation solutions .
however , the new super - kamiokande analysis of the energy spectrum now indicates @xcite that , if there is any consistent vacuum - oscillation solution at all , it may well have @xmath75 considerably below 1 , providing another potential nail in the coffin of degenerate neutrinos .
the vacuum - oscillation solution would require , moreover , extreme degeneracy : @xmath76 , which is impossible to reconcile with a simple calculation of neutrino mass renormalization in models with degenerate masses at the @xmath77 scale @xcite .
mass - renormalization effects also endanger the large - angle msw solution ( which would require @xmath78 ) , and , in the context of bimaximal mixing models , also generate unacceptable values of the neutrino mixing angles .
these renormalization problems may not be insurmountable @xcite , but they do raise non - trivial issues that must be addressed in models @xcite of ( near-)degenerate neutrino masses .
our provisional conclusion is that @xmath79 2 ev , with the most likely case being @xmath80 0.06 ev , too small to be of much interest for hot dark matter .
the motivation for supersymmetry at an accessible energy is provided by the gauge hierarchy problem @xcite , namely that of understanding why @xmath81 , the only candidate for a fundamental mass scale in physics . alternatively and equivalently
, one may ask why @xmath82 , where @xmath83 is the planck mass , expected to be the fundamental gravitational mass scale . or
one may ask why the coulomb potential inside an atom is so much larger than the newton potential , which is equivalent to why @xmath84 , where @xmath85 are the proton and electron masses .
one might think it would be sufficient to choose the bare mass parameters : @xmath86 .
however , one must then contend with quantum corrections , which are quadratically divergent : m^2_h , w = o ( ) ^2 [ fifteen ] these are much larger than @xmath87 , if the cutoff @xmath88 representing the appearance of new physics is taken to be @xmath89 .
this means that one must fine - tune the bare mass parameter so that it is almost exactly cancelled by the quantum correction ( [ fifteen ] ) in order to obtain a small physical value of @xmath87 .
this seems unnatural , and the alternative is to introduce new physics at the tev scale , so that the correction ( [ fifteen ] ) is naturally small . at one stage , it was proposed that this new physics might correspond to the higgs boson being composite @xcite . however , calculable scenarios of this type are inconsistent with the precision electroweak data from lep and elsewhere .
the alternative is to postulate approximate supersymmetry @xcite , whose pairs of bosons and fermions produce naturally cancelling quantum corrections : m^2_w = o ( ) m^2_b - m^2_f [ sixteen ] that are naturally small : @xmath90 if m^2_b - m^2_f ^2 .
[ seventeen ] there are many other possible motivations for supersymmetry , some of which are discussed below , but this is the only one that gives reason to expect that it might be accessible to the current generation of accelerators and in the range expected for a cold dark matter particle . the minimal supersymmetric extension of the standard model ( mssm ) @xcite has the same gauge interactions as the standard model , and the yukawa interactions are very similar : _
d qd^ch + _ le^ch + _ u qu^c|h + |hh [ eighteen ] where the capital letters denote supermultiplets with the same quantum numbers as the left - handed fermions of the standard model . the couplings @xmath91 give masses to down quarks , leptons and up quarks respectively , via distinct higgs fields @xmath92 and @xmath93 , which are required in order to cancel triangle anomalies
. the new parameter in ( [ eighteen ] ) is the bilinear coupling @xmath94 between these higgs fields , that plays a significant rle in the description of the lightest supersymmetric particle , as we see below .
the gauge quantum numbers do not forbid the appearance of additional couplings @xcite lle^c + ^lqd^c + u^cd^cd^c [ nineteen ] but these violate lepton or baryon number , and we assume they are absent .
one significant aspect of the mssm is that the quartic scalar interactions are determined , leading to important constraints on the higgs mass , as we also see below .
supersymmetry must be broken , since supersymmetric partner particles do not have identical masses , and this is usually parametrized by scalar mass parameters @xmath95 , gaugino masses @xmath96 and trilinear scalar couplings @xmath97 .
these are commonly supposed to be inputs from some high - energy physics such as supergravity or string theory .
it is often hypothesized that these inputs are universal : @xmath98 , but these assumptions are not strongly motivated by any fundamental theory .
the physical sparticle mass parameters are then renormalized in a calculable way : m^2_0_i = m^2_0 + c_i m^2_1/2 , m_a = ( _ a _ gut ) m_1/2 [ twenty ] where the @xmath99 are calculable coefficients @xcite and mssm phenomenology is then parametrized by @xmath100 and @xmath101 ( the ratio of higgs v.e.v.s ) .
precision electroweak data from lep and elsewhere provide two qualitative indications in favour of supersymmetry .
one is that the inferred magnitude of quantum corrections favour a relatively light higgs boson @xcite m_h 200 gev [ twentyone ] which is highly consistent with the value predicted in the mssm : @xmath102 150 gev @xcite as a result of the constrained quartic couplings .
( on the other hand , composite higgs models predicted an effective higgs mass @xmath103 1 tev and other unseen quantum corrections . ) the other indication in favour of low - energy supersymmetry is provided by measurements of the gauge couplings at lep , that correspond to @xmath104 in agreement with the predictions of supersymmetric guts with sparticles weighing about 1 tev , but in disagreement with non - supersymmetric guts that predict @xmath105 to 0.22 @xcite .
neither of these arguments provides an accurate estimate of the sparticle mass scales , however , since they are both only logarithmically sensitive to @xmath106 and/or @xmath107 .
the lightest supersymmetric particle ( lsp ) is expected to be stable in the mssm , and hence should be present in the universe today as a cosmological relic from the big bang @xcite .
this is a consequence of a multiplicatively - conserved quantum number called @xmath108 parity , which is related to baryon number , lepton number and spin : r = ( -1)^3b+l+2s [ twentytwo ] it is easy to check that @xmath109 for all standard model particles and @xmath110 for all their supersymmetric partners .
the interactions ( [ eighteen ] ) conserve @xmath108 , whilst those in ( [ nineteen ] ) would violate @xmath108 , in contrast to a majorana neutrino mass term or the other interactions in minimal @xmath111 or @xmath112 guts .
there are three important consequences of @xmath108 conservation : ( i ) sparticles are always produced in pairs , e.g. , @xmath113 , @xmath114 , ( ii ) heavier sparticles decay into lighter sparticles , e.g. , @xmath115 , @xmath116 , and ( iii ) the lsp is stable because it has no legal decay mode .
if such a supersymmetric relic particle had either electric charge or strong interactions , it would have condensed along with ordinary baryonic matter during the formation of astrophysical structures , and should be present in the universe today in anomalous heavy isotopes .
these have not been seen in studies of @xmath92 , @xmath117 , @xmath118 , @xmath119 , @xmath120 , @xmath121 , @xmath122 , @xmath123 and @xmath124 isotopes at levels ranging from @xmath125 to @xmath126 @xcite , which are far below the calculated relic abundances from the big bang : 10 ^ -6 10 ^ -10 [ twentythree ] for relics with electromagnetic or strong interactions . except possibly for very heavy relics
, one would expect these primordial relic particles to condense into galaxies , stars and planets , along with ordinary bayonic material , and hence show up as an anaomalous heavy isotope of one or more of the elements studied .
there would also be a ` cosmic rain ' of such relics @xcite , but this would presumably not be the dominant source of such particles on earth .
the conflict with ( [ twentythree ] ) is sufficiently acute that the lightest supersymmetric relic must presumably be electromagnetically neutral and weakly interacting @xcite .
in particular , i believe that the possibility of a stable gluino can be excluded .
this leaves as scandidates for cold dark matter a sneutrino @xmath127 with spin 0 , some neutralino mixture of @xmath128 with spin 1/2 , and the gravitino @xmath129 with spin 3/2 .
lep searches for invisible @xmath130 decays require @xmath131 @xcite , and searches for the interactions of relic particles with nuclei then enforce @xmath132 few tev @xcite , so we exclude this possibility for the lsp .
the possibility of a gravitino @xmath129 lsp has attracted renewed interest recently with the revival of gauge - mediated models of supersymmetry breaking @xcite , and could constitute warm dark matter if @xmath133 . in this talk , however , i concentrate on the @xmath134 neutralino combination @xmath135 , which is the best supersymmetric candidate for cold dark matter
. the neutralinos and charginos may be characterized at the tree level by three parameters : @xmath107 , @xmath94 and tan@xmath136 .
the lightest neutralino @xmath135 simplifies in the limit @xmath137 where it becomes essentially a pure photino @xmath138 , or @xmath139 where it becomes essentially a pure higgsino @xmath140 .
these possibilities are excluded , however , by lep and the fnal tevatron collider @xcite . from the point of view of astrophysics and cosmology ,
it is encouraging that there are generic domains of the remaining parameter space where @xmath141 to @xmath142 , in particular in regions where @xmath135 is approximately a @xmath143 gaugino @xmath144 , as seen in fig
. 6 @xcite .
purely experimental searches at lep enforce @xmath145 gev @xcite .
this bound can be strengthened by making various theoretical assumptions , such as the universality of scalar masses @xmath146 , including in the higgs sector , the cosmological dark matter requirement that @xmath147 and the astrophysical preference that @xmath148 . taken together as in fig . 7
, we see that they enforce m _ 50 [ twentyfour ] moreover , lep has already explored almost all the parameter space available for a higgsino - like lsp , and this possibility will also be thoroughly explored by the full running of lep @xcite .
should one be concerned that no sparticles have yet been seen by either lep or the fnal tevatron collider ?
one way to quantify this is via the amount of fine - tuning of the input parameters required to obtain the physical value of @xmath87 @xcite : _ o = max_i [ twentyfive ] where @xmath149 is a generic supergravity input parameter .
the lep exclusions impose @xcite _ o 8 [ twentysix ] although fine - tuning is a matter of taste , this is perhaps not large enough to be alarming , and could in any case be reduced significantly if a suitable theoretical relation between some input parameters is postulated @xcite .
moreover , it is interesting to note that the amount of fine - tuning @xmath150 is minimized when @xmath151 as preferred astrophysically , as seen in fig . 8 @xcite .
this means that solving the gauge hierarchy problem naturally leads to a relic neutralino density in the range of interest to astrophysics and cosmology .
i am unaware of any analogous argument for other particle dark matter candidates such as the neutrino or the axion . for certain ranges of the mssm parameters ,
our present electroweak vacuum is unstable against the development of vev s for @xmath152 and @xmath153 fields , leading to vacua that would break charge and colour conservation . among the dangerous possibilities are flat directions of the effective potential in which combinations such as @xmath154 acquire vev s .
avoiding these vacua imposes constraints that depend on the soft supersymmetry breaking parameters : they are weakest for @xmath155 . fig .
9 illustrates some of the resulting constraints in the @xmath156 plane , for different values of @xmath101 and signs of @xmath94 @xcite .
we see that they cut out large parts of the plane , particularly for low @xmath106 . in combination with cosmology , they tend to rule out large values of @xmath107 , but this aspect needs to be considered in conjunction with the effects of coannihilation , that are discussed in the next paragraph . as @xmath157 increases , the lsp annihilation cross section decreases and hence its relic number and mass density increase
. how heavy could the lsp be ?
until recently , the limit given was @xmath158 gev @xcite .
it has now been pointed out that there are regions of the mssm parameter space where coannihilations of the @xmath135 with the stau slepton @xmath159 could be important , as also seen in fig . 9 @xcite .
these coannihilations would suppress @xmath160 , allowing a heavier neutralino mass , and we now find that @xcite m _ 600 [ twentyseven ] is possible if we require @xmath161 . in the past , it was thought that all the cosmologically - preferred region of mssm parameter space could be explored by the lhc @xcite , as seen in fig .
10 , but it now seems possible that there may be a delicate region close to the upper bound ( [ twentyseven ] ) .
this point requires further study .
one strategy to look for dark matter particles is via their annihilations in the galactic halo ; @xmath162 in the cosmic rays @xcite .
figure 11 shows the current measurements of cosmic - ray @xmath163 s .
the lines indicate the secondary flux expected to be produced by primary matter cosmic rays . some of the earlier measurements were above this conventional expectation , fuelling speculation about possible exotic sources such as @xmath164 annihilation . however , the recent bess data @xcite agree very well with conventional secondary production
. there may still be some scope for exotic sources at low energies @xmath165 300 mev or at higher energies @xmath166 3 gev , and one of the objectives of the ams experiment @xcite is to explore this possibility .
figure 12 shows that some supersymmetric models are already excluded by the bess data @xcite , and indicates how big an opportunity ams may have .
it has been suggested @xcite that there may be an excess of cosmic - ray positrons at energies above 1 gev , as seen in fig .
13 , although the uncertainties in the standard leaky - box model prevent any definite conclusion at this stage .
ams has reported an excess of positrons at lower energies : @xmath167 1 gev , but it seems unlikely that these have an exotic origin .
it may in the future be able to contribute to the clarification of the possible excess at higher energies .
data from egret on @xmath168 rays above 100 mev have been interpreted @xcite as containing an excess from directions centred around the galactic centre
. these could well be due to some unresolved astrophysical sources or some diffuse mechanism such as inverse compton scattering .
if due to dark matter annihilation , they would require @xmath169 to @xmath170 ) cm@xmath47 if the dark matter particles are not clumped .
this range is above that allowed for supersymmetric dark matter : @xmath172 cm@xmath47 .
therefore it has been suggested @xcite that the dark matter may be clumped , as in some models of structure formation .
a phenomenological approach to this possibility is to calculate the maximal clumpiness enhancement allowed for any supersymmetric model , taking into account the experimental upper limits on , e.g. , @xmath163 annihilation products , and then calculate the maximum possible @xmath168 flux .
some representative calculations @xcite are shown in fig .
14 : we see that the @xmath168 fluxes could in principle be orders of magnitude larger than those detectable by glast
. a dark matter passing through the sun ( earth ) may scatter on some nucleus inside , losing source recoil energy which may convert its orbit from hyperbolic to elliptical , with perhelion ( perigee ) below the surface .
then it will scatter repeatedly , eventually settling into a quasi - thermal distribution beneath the surface .
this population of relic particles is controlled by annihilation : @xmath173 , yielding as observable products energetic neutrinos : @xmath174 1 gev .
these may be detected either directly in an underground detector , or indirectly via @xmath94 s produced in material surrounding the detector . as seen in fig .
15 @xcite , a 1 km@xmath47 muon detector would be able to detect quite a number of supersymmetric models that do not produce detectable cosmic - ray @xmath175 fluxes , via either the solar or subterranean cosmic rays they produce .
it has recently been pointed out @xcite that there could be an enhancement of relic annihilations in the earth due to a solar - system population of relic particles that is augmented by jupiter s gravitational field , so the prospects may be even brighter than indicated in fig . 15 .
many experiments around the world are looking for the recoil energy deposited in low - background underground detectors by relic scattering on nuclei .
the typical recoil energy deposited is @xmath176 tens of kev .
the interaction may be mediated by squark , @xmath177 and higgs exchanges , which contribute both spin - dependent and spin - independent matrix elements .
the former are related to the quark contributions to the proton spin @xmath178 and the latter to the quark contributions to the proton mass .
the time - dependent contributions are important for some light nuclei such as fluorine , where they can be calculated quite reliably .
however , the spin - independent contributions are coherent and more important for heavier nuclei @xcite .
figure 16 shows the upper limit for scattering on an individual nucleon , together with the region of parameter space that could not be excluded by an experiment searching for an annual modulation of the detector rate in a particular range of deposited energy @xcite .
although on the large side , the range of cross section in this non - excluded domain is not inconsistent with some model calculations .
future experiments will be able to improve the current upper limits shown in fig .
16 by several orders of magnitude and explore much of the model parameter space .
what if there were a real signal in the non - excluded region in fig .
? it would correspond to a supersymmetric relic weighing @xmath179 50 to 100 gev .
such models might well produce an observable cosmic - ray @xmath163 flux or @xmath94 flux from high - energy solar or subterranean neutrinos @xcite . these searches for astrophysical sparticles must compete with accelerator searches .
lep has almost completed the exploration of its available kinematic reach , and the tevatron has a window for possible sparticle discoveries .
however , the best prospects for the discovery of supersymmetry will be offered by the lhc .
it will benefit from large cross sections for squark @xmath180 and gluino @xmath181 production , and their cascade decays into lighter sparticles offer many opportunities for distinctive signatures . as seen in fig .
10 @xcite , it should be possible to detect @xmath182 2 to 2.5 tev , and some detailed spectroscopic measurements will be possible .
the entire supersymmetric dark matter region were @xmath183 0.3 will be covered by the lhc .
so those looking for supersymmetric dark matter via cosmic rays or other astrophysical signatures should hurry up , and do their best before lhc startup in 2005 !
it has been suggested @xcite that cold dark matter particles should weigh @xmath184 1 tev , as exemplified by the mssm range ( [ twentyseven ] ) .
this expectation is based on the assumption that the cold dark matter particles were at one time in thermal equilibrium
. however , much heavier relic particles are possible if one invokes non - thermal production mechanisms .
non - thermal decays of inflatons in conventional models of cosmological inflation could yield @xmath185 for @xmath186 gev .
preheating via the parametric resonance decay of the inflaton could even yield @xmath185 for @xmath187 gev .
other possibilities include a first - order phase transition at the end of inflation , and gravitational relic production induced by the rapid change in the scale factor in the early universe @xcite .
it is therefore of interest to look for possible experimental signatures of superheavy dark matter .
one such possibility is offered by ultra - high - energy cosmic rays .
those coming from distant parts of the universe @xmath188 are expected to be cut off at an energy @xmath189 gev , because of the reaction @xmath190 @xcite .
however , as discussed extensively here , no such greisen - zatsepin - kuzmin cut - off is seen in the data @xcite the ultra - high - energy cosmic rays must originate nearby , and ( unless the intergalactic magnetic field is unexpectedly high @xcite ) should point back to any point - like sources such as agns @xcite .
however , no such discrete sources have been identified as yet .
could the ultra - high - energy cosmic rays be due to the decays of superheavy relic particles ? these should be clustered in galactic haloes ( including our own ) , and hence give an anisotropic flux @xcite , but there would be no obvious point sources .
there have been some reports of anisotropies in high - energy cosmic rays , but it is not clear whether they could originate in superheavy relic decays .
we analyzed @xcite possible superheavy relic candidates in string @xcite and/or @xmath0 theory .
one expects kaluza - klein states when six excess dimensions are compactified : @xmath191 or @xmath192 , which we call _
hexons_. however , these are expected to weigh @xmath193 gev , which may be too heavy , and there is no particular reason to expect hexons to be metastable . in @xmath0 theory ,
one expects massive states associated with a further compactification : @xmath194 dimensions , which we call _ pentons_. their mass could be @xmath195 gev , which would be suitable , but there is again no good reason to expect them to be metastable .
we are left with bound states from the hidden sector of string/@xmath0 theory , which we call _ cryptons _
these could also have masses @xmath196 gev , and might be metastable for much the same reason as the proton in a gut , decaying via higher - dimensional multiparticle operators .
for example , in a flipped @xmath111 model we have a hidden - sector @xmath197 gauge group , and the former factor confines four - constituent states which we call _ tetrons_. initial studies @xcite indicate that the lightest of these might well have a lifetime @xmath198 , which would be suitable for the decays of superheavy dark matter particles .
detailed simulations have been made of the spectra of particles produced by the fragmentation of their decay products @xcite , and the ultra - high - energy cosmic - ray data are consistent with the decays of superheavy relics weighing @xmath199 gev , as seen in fig .
17 @xcite .
issues to be resolved here include the roles of supersymmetric particles in the fragmentation cascades , and the relative fluxes of @xmath200 and @xmath201 among the ultra - high - energy cosmic rays .
as mentioned in section 1 , data on large - scale structure @xcite and high - redshift supernovae @xcite have recently converged on the suggestion that the energy of the vacuum may be non - zero , as seen in figs . 1 , 3 . in my view , this represents a wonderful opportunity for theoretical physics : a number to be calculated in the theory of everything including quantum gravity .
the possibility that the vacuum energy may be non - zero may even appear more natural than a zero value , since there is no obvious symmetry or other reason known why it should vanish . in the above paragraph
, i have used the term _ vacuum energy _ rather than _ cosmological constant _ , because it may not actually be constant .
this option has been termed _ quintessence _ in @xcite , which discusses a classical scalar - field model that is not strongly motivated by the standard model , supersymmetry or guts , though something similar might emerge from string theory .
i prefer to think that a varying vacuum energy might emerge from a quantum theory of gravity , as the vacuum relaxes towards an asymptotical value ( zero ? ) in an infinitely large and old universe .
we have recently given @xcite an example of one such possible effect which yields a contribution to the vacuum energy that decreases as @xmath202 .
this is compatible with the high - redshift supernova data , and one may hope that these could eventually discriminate between such a possibility and a true cosmological constant .
however , combinations of structure - formation and cosmic microwave - background measurements may be sensitive to smaller neutrino masses : w. hu , d.j .
eisenstein , m. tegmark and m. white , phys .
d59 ( 1999 ) 023512 .
peccei , hep - ph/9906509 , to be published in the proceedings of 8th mexican school of particles and fields ( viii - empc ) , oaxaca de juarez , mexico , 20 - 28 nov 1998 .
r. barbieri , j. ellis and m.k .
gaillard , phys .
lett . 90b ( 1980 ) 249 .
m. gell - mann , p. ramond and r. slansky , proceedings of the stony brook supergravity workshop , new york , 1979 , eds .
p. van nieuwenhuizen and d. freedman ( north - holland , amsterdam ) ; t. yanagida , proceedings of the workshop on unified theories and baryon number in the universe , tsukuba , japan 1979 , eds .
a. sawada and a. sugamoto , kek report no .
79 - 18 .
the literature on the subject is vast . for introductions ,
see : c. d. froggatt and h. b. nielsen , nucl . phys .
b147 ( 1979 ) 277 ; h. fritzsch , phys . lett .
70b ( 1977 ) 436 ; b73 ( 1978 ) 317 ; nucl .
b155 ( 1979 ) 189 ; j. harvey , p. ramond and d. reiss , phys .
b92 ( 1980 ) 309 ; s. dimopoulos , l. j. hall and s. raby , phys .
( 1992 ) 1984 ; c. wetterich , nucl .
b261 ( 1985 ) 461 ; l. ibanez and g.g .
ross , phys .
b332 ( 1994 ) 100 ; g.k .
leontaris and d.v .
nanopoulos , phys .
b212 ( 1988 ) 327 ; y.achiman and t. greiner , phys .
lett . b329 ( 1994 ) 33 ; y. grossman and y. nir , nucl .
b448 ( 1995 ) 30 ; h. dreiner _
et al . _ ,
b436 ( 1995 ) 461 ; p. bintruy , s. lavignac and p. ramond , nucl .
b477 ( 1996 ) 353 ; g.k .
leontaris , _ et al .
d 53 ( 1996 ) 6381 ; p.
b496 ( 1997 ) 3 ; s. lola and j.d .
vergados , progr . part .
40 ( 1998 ) 71 .
for one particular take on this , see : j. ellis , g.k .
leontaris , s. lola and d.v .
nanopoulos , eur . phys . j. c9 ( 1999 ) 389 ; for a recent review , see : g. altarelli and f. feruglio , hep - ph/9905536 .
j. ellis and s. lola , phys .
b458 ( 1999 ) 310 .
l. baudis _
et al . _ ,
83 ( 1999 ) 41 .
r. barbieri , l.j .
hall , g.l .
kane and g.g .
ross , hep - ph/9901228 .
l. maiani , _ proc .
summer school on particle physics _
, gif - sur - yvette , 1979 ( in2p3 , paris , 1980 ) p. 3 ; + g t hooft , in : g t hooft et al .
, _ recent developments in field theories _ ( plenum press , new york , 1980 ) ; + e. witten , nucl . phys .
b188 ( 1981 ) 513 ; + r.k .
kaul , phys .
109b ( 1982 ) 19 . for a review ,
see : e. farhi and l. susskind , phys
. rep .
74c ( 1981 ) 277 ; + s. dimopoulos and l. susskind , nucl .
b155 ( 1979 ) 237 ; + e. eichten and k. lane , phys . lett .
b90 ( 1980 ) 125 ; + m.e . peskin and t. takeuchi ,
d46 ( 1992 ) 381 ; + g. altarelli , r. barbieri and s. jadach , nucl .
b369 ( 1992 ) 3 .
y. okada , m. yamaguchi and t. yanagida , progr .
phys . 85 ( 1991 ) 1 ; + j. ellis , g. ridolfi and f. zwirner , phys .
b257 ( 1991 ) 83 , phys .
b262 ( 1991 ) 477 ; + h.e .
haber and r. hempfling , phys .
lett . 66 ( 1991 ) 1815 ; + r. barbieri , m. frigeni and f. caravaglios , phys .
b258 ( 1991 ) 167 ; + y. okada , m. yamaguchi and t. yanagida , phys . lett .
b262 ( 1991 ) 54 .
j. ellis , s. kelley and d.v .
nanopoulos , phys .
b249 ( 1990 ) 441 and phys .
lett . b260 ( 1991 ) 131 ; + u. amaldi , w. de boer and h. furstenau , phys .
b260 ( 1991 ) 447 ; + p. langacker and m. luo , phys .
d44 ( 1991 ) 817 . j. silk and m. srednicki , phys .
53 ( 1984 ) 624
. s. orito _ et al .
_ , bess collaboration , astro - ph/9906426 .
r. battiston , for the ams collaboration , astro - ph/9907152 .
l. bergstrom , j. edsjo and p. ullio , astro - ph/9906034 , these proceedings , vol .
. s. coutu _ et al .
_ , heat collaboration , astropart .
. 11 ( 1999 ) 429 .
dixon _ et al .
_ , new astron . 3 ( 1998 ) 539 . l. bergstrom , j. edsjo , p. gondolo and p. ullio , phys . rev .
d59 ( 1999 ) 043506 l. bergstrom , j. edsjo and p. gondolo , astro - ph/9906033 , these proceedings , vol .
2 , p.281 .
t. damour and l.m .
krauss , phys .
d59 ( 1999 ) 063509 ; + l. bergstrom , t. damour , j. edsjo , l.m . krauss and p. ullio ,
jhep 08 ( 1999 ) 010 ; + see however : a. gould and s.m .
alam , astro - ph/9911288 .
k. griest , g. jungman and m. kamionkowski , phys .
rept . 267 ( 1996 ) 195 .
r. bernabei _ et al .
_ , dama collaboration , phys .
b450 ( 1999 ) 448 .
a. bottino and n. fornengo , hep - ph/9904469 .
s. dimopoulos , phys .
b246 ( 1990 ) 347 .
kolb , d.j.h .
chung and a. riotto , hep - ph/9810361 and references therein .
e. ahn , g. medina - tanco , p.l .
biermann and t. stanev , astro - ph/9911123 .
the apparent excess of ultra - high - energy cosmic rays is significantly reduced if their sources follow the known inhomogeneous distribution of matter in the nearby universe : g. medina tanco , astro - ph/9905239 , these proceedings , vol . 4 , p.346 . | the density budget of the universe is reviewed , and then specific particle candidates for non - bayonic dark matter are introduced , with emphasis on the relevance of cosmic - ray physics .
the sizes of the _ neutrino masses _ indicated by recent atmospheric and solar neutrino experiments may be too small to contribute much hot dark matter .
my favoured candidate for the dominant cold dark matter is the _ lightest supersymmetric particle _ , which probably weighs between about 50 gev and about 600 gev .
strategies to search for it via cosmic rays due to annihilations in the halo , sun and earth , or via direct scattering experiments , are mentioned .
possible _
superheavy relic particles _ are also discussed , in particular metastable string- or @xmath0-theory _ cryptons _ , that may be responsible for the ultra - high - energy cosmic rays . finally , it is speculated that a non - zero contribution to the _ cosmological vacuum energy _ might result from incomplete relaxation of the quantum - gravitational vacuum
. 0.3 cm cern - th/99 - 358 + astro - ph/9911440 _ contribution to the proceedings of the 26th international cosmic - ray conference , salt lake city , august 1999 _ |
recent technological advancements have enabled increasing use of infrastructure - free wireless communications .
for example , smartphone users can exchange information with each other by exploiting local wi - fi and bluetooth connections , or using the fifth - generation ( 5 g ) cellular device - to - device communications ; and even unmanned aerial vehicles ( uavs ) can directly communicate with nearby ground stations and send back photos and videos in real time .
although these infrastructure - free communication links bring great convenience to our daily lives , they can also be used by malicious users to launch various security attacks .
for instance , terrorists can use peer - to - peer wi - fi connections to communicate and facilitate terror attacks , and criminals can control uavs to spy and collect private information from rightful users . as such malicious attacks are launched via infrastructure - free wireless communications , they are difficult to be monitored by solely using existing information surveillance methods that intercept the communication data at the cellular or internet infrastructures . in response to such new threats on public security , authorized parties such as government agencies should develop new approaches to legitimately surveil these suspicious wireless communication links over the air ( e.g. , via eavesdropping ) to detect malicious attacks , and then intervene in them ( e.g. , via jamming and spoofing ) to quickly defend and disable these attacks .
there have been several recent studies in the literature that investigate the surveillance of wireless communications , where authorized parties efficiently intercept suspicious wireless communication links , extract their exchanged data contents , and help identify the malicious wireless communication links to intervene in .
conventionally , the methods for wireless communications surveillance include wiretapping of wireless operators infrastructures and installation of monitoring software in smartphones .
recently , over - the - air eavesdropping has emerged as a new wireless communications surveillance method . among others ,
passive eavesdropping ( see , e.g. , @xcite ) and proactive eavesdropping @xcite are two approaches implemented at the physical layer , in which authorized parties can deploy dedicated wireless monitors to overhear the targeted wireless communications , especially the infrastructure - free ones .
efficient surveillance can help detect and identify malicious users and their communications .
after that , authorized parties need to quickly respond and defend them via wireless communication intervention .
for example , the security agency may need to disrupt , disable , or spoof ongoing terrorists communications to prevent terror attacks at the planning stage , and it is also desirable to change the control signal of a malicious uav to land it in a targeted location and catch it . in the literature ,
physical - layer jamming ( see , e.g. , @xcite ) is one existing approach that can be employed to intervene in malicious communications , though it was originally proposed for military instead of public security applications . in the physical - layer jamming , the jammer sends artificially generated gaussian noise ( so - called `` uncorrelated jamming '' @xcite ) or a processed version of the malicious signal ( so - called `` correlated jamming '' @xcite ) to disrupt or disable the targeted malicious wireless communications . however , jamming the targeted communications at the physical layer is easy to be detected , and may not be sufficient to successfully intervene in malicious activities .
this is due to the fact that when the targeted communication continuously fails due to the jamming attack , the malicious users may take counter - measures by changing their communication frequency bands or switching to another way of communications .
thus , we are motivated to study a new wireless communication intervention via spoofing at the physical layer , which can keep the malicious communication but change the communicated information to intervene in .
= 1 we investigate the new physical - layer spoofing by considering a fundamental three - node system over additive white gaussian noise ( awgn ) channels . as shown in fig .
[ fig:1 ] , an intermediary legitimate spoofer aims to spoof a malicious communication link from alice to bob , such that the received message at bob is changed from alice s originally sent message to the one desired by the spoofer . under this setup
, we propose a new symbol - level spoofing approach , in which the spoofer designs the spoofing signals via exploiting the symbol - level relationship between each original constellation point of alice and the desirable one of the spoofer , so as to optimize the spoofing performance .
in particular , we consider two cases when alice employs the widely - used binary phase - shift keying ( bpsk ) and quadrature phase - shift keying ( qpsk ) modulations , respectively .-
ary quadrature amplitude modulation ( @xmath0-qam ) and @xmath0-ary phase shift keying ( @xmath0-psk ) with @xmath1 . nevertheless , under these modulation techniques , how to design spoofing signals to optimally solve the average sser minimization problem is generally a more difficult task , since the corresponding sser functions will become very complicated .
] the objective of the spoofer is to minimize the average spoofing - symbol - error - rate ( sser ) , i.e. , the average probability that the symbols decoded by bob fail to be changed as the desirable ones of the spoofer .
the main results of this paper are summarized as follows .
* in the bpsk case ( with the constellation points being @xmath2 ) , the spoofing signals are designed by classifying the symbols into two types . in each of type - i symbols ( see fig . [
fig : bpsk]-(a ) ) , where the original constellation point of alice and the desirable one of the spoofer are identical ( both are @xmath3 or @xmath4 ) , the spoofing signal is designed to _ constructively _ combine with the original signal of alice at bob to help improve the decoding reliability against gaussian noise . in each of type - ii symbols ( see fig . [
fig : bpsk]-(b ) ) , where the original constellation point of alice and the desirable one of the spoofer are opposite ( one is @xmath3 ( or @xmath4 ) but the other is @xmath4 ( or @xmath3 ) ) , the spoofing signal is designed to _
destructively _ combine with the original signal of alice at bob , thus moving the constellation point towards the desirable opposite direction .
we minimize the average sser by optimizing the spoofing signals and their power allocations over type - i and type - ii symbols at the spoofer , subject to its average transmit power constraint .
although this problem is non - convex , we derive its optimal solution .
it is shown that when the transmit power at alice is low or the spoofing power at the spoofer is high , the spoofer should allocate its transmit power to both type - i and type - ii symbols .
otherwise , when the transmit power at alice is high and the spoofing power at the spoofer is low , the spoofer should allocate almost all its transmit power over a certain percentage of type - ii symbols with an `` on - off ''
power control . * in the qpsk case with the constellation points being @xmath5 with @xmath6 , the symbols are further classified into three types , where in type - i , type - ii , and type - iii symbols , the original constellation points of alice and the desirable ones of the spoofer are identical , opposite , and neighboring , respectively , as shown in fig .
[ fig : qpsk ] . for type - i and type - ii symbols ,
the spoofing signals are designed to have equal strengths for the real and imaginary components , such that at the receiver of bob they can be be constructively and destructively combined with the original constellation points by alice , respectively . for type - iii symbols ,
the spoofing signals are designed to have independent real and imaginary components .
under such a design , we formulate the average sser minimization problem by optimizing the spoofing power allocations over symbols , subject to the average transmit power constraint .
though this problem is non - convex and generally difficult , we obtain its optimal solution , motivated by that in the bpsk case .
* numerical results show that for both bpsk and qpsk cases , the symbol - level spoofing scheme with optimized transmission achieves a much better spoofing performance ( in terms of a lower average sser ) , as compared to the block - level spoofing benchmark where the spoofer does not exploit the symbol information of alice , and a heuristically designed symbol - level spoofing scheme .
it is worth noting that in the existing literature there is another type of higher - layer spoofing attack , which can also be utilized for wireless communication intervention ( see , e.g. , @xcite ) .
for example , in the medium access control ( mac ) spoofing @xcite and internet protocol ( ip ) spoofing , a network attacker can hide its true identity and impersonate another user , so as to access the targeted wireless networks .
nevertheless , for these higher - layer spoofing , the network attacker needs to establish new wireless communication links to access the network .
in contrast , our proposed symbol - level spoofing is implemented at the physical layer , which can change the communicated information of _ ongoing _ malicious wireless communications , thus leading to a quicker response and intervention that is also more likely to be covert .
it is also worth comparing our proposed symbol - level spoofing versus the symbol - level precoding ( not for security ) in downlink multiuser multi - antenna systems @xcite . in the symbol - level precoding ,
the transmitter designs its precoding vectors by exploiting the symbol - level relationships among the messages to different receivers , such that the constructive part of the inter - channel interference is preserved and exploited and only the destructive part is eliminated . although the symbol - level spoofing and precoding are based on a similar design principle of exploiting the symbol - level relationship among co - channel signals , they focus on different application scenarios for different purposes , thus requiring different design methods .
the remainder of this paper is organized as follows .
section [ sec:2 ] introduces the system model and formulates the average sser minimization problem .
sections [ sec:3 ] and [ sec : qpsk ] propose the symbol - level spoofing approach and design the spoofing signals and their power allocations for the cases of bpsk and qpsk modulations , respectively . section [ sec:4 ] presents numerical results to evaluate the performance of the proposed symbol - level spoofing design as compared to other benchmark schemes .
finally , section [ sec:5 ] concludes the paper .
as shown in fig . [ fig:1 ] , we consider a fundamental three - node system over awgn channels , where an intermediary legitimate spoofer aims to spoof a malicious wireless communication link from alice to bob by changing the communicated data at the bob side . we consider that the malicious communication employs the bpsk or qpsk modulation techniques , which are most commonly used in existing wireless communication systems . in the @xmath7th symbol of this block , we denote the transmitted signal by alice as @xmath8 , where @xmath9 is the transmit power per symbol at alice , and @xmath10 denotes the message that alice wants to deliver to bob . here , @xmath10 is equally likely chosen from the set of constellation points @xmath11 , where @xmath12 and @xmath13 for the bpsk and qpsk cases , respectively .
therefore , we have @xmath14 .
first , we introduce the receiver model of bob by considering the case without spoofing .
accordingly , the received signal by bob in the @xmath7th symbol is expressed as @xmath15 where @xmath16 denotes the noise at the receiver of bob , which is an independent and identically distributed ( i.i.d . )
circularly symmetric complex gaussian ( cscg ) random variable with zero mean and unit variance . based on the maximum likelihood ( ml ) detection , the decoded message by bob
is expressed as @xmath17 next , we consider the spoofing strategy employed by the spoofer . it is assumed that the spoofer perfectly knows the transmitted symbol information @xmath10 s of alice . here , @xmath10
s can be practically obtained by the spoofer via efficient eavesdropping or wiretapping beforehand .
for example , if alice is an intermediary node of a multi - hop communication link , then the spoofer can obtain @xmath10 s via eavesdropping the previous hops ; if alice gets its transmitted data from the backhaul or infrastructure - based networks , then the spoofer can acquire them via using wiretapping devices to overhear the backhaul communications ; and furthermore , the spoofer can even secretly install an interceptor software ( e.g. , flexispy ) in the alice s device to get @xmath10 s . note that the assumption about the perfect symbol information at the spoofer has been made in the existing correlated jamming literature ( see , e.g.,@xcite ) to improve the jamming performance .
we make a similar assumption here for the purpose of characterizing the spoofing performance upper bound , and leave the details about the symbol information acquisition for future work . based on the information of @xmath10 s ,
the spoofer designs the spoofing signal as @xmath18 in the @xmath7th symbol ( the design details will be provided in the next section ) .
then , the received signal at bob is expressed as @xmath19 with the ml detection , the decoded message by bob is expressed as @xmath20 the spoofer aims to maximize the opportunity of changing the messages of alice to be the desirable ones by itself .
let @xmath21 denote the desirable constellation point for the @xmath7th symbol , which is equally likely chosen from @xmath11 and is independent from the message @xmath10 sent by alice .
nevertheless , due to the limited spoofing power and receiver noise , it is difficult for the spoofer to ensure that all symbols @xmath22 s are successfully changed to be the desirable @xmath21 s . in this case , we define the probability of unsuccessful spoofing in any symbol @xmath7 as the sser , denoted by @xmath23 .
then , the objective of the spoofer is to minimize the average sser , i.e. , @xmath24 , where @xmath25 denotes the statistical expectation over all possible symbols .
suppose that the spoofer is constrained by a maximum average transmit power denoted by @xmath26 , i.e. , @xmath27 .
as a result , the optimization problem of our interest is @xmath28 in the following two sections , we will solve problem ( [ eqn:5 ] ) by considering the bpsp and qpsk modulations , respectively .
= 1 of alice , and the blue circular denotes the desirable constellation point @xmath21 of the spoofer . ( a ) an example of type - i symbols , where @xmath10 and @xmath21 are identical with @xmath29 ; ( b ) an example of type - ii symbols , where @xmath10 and @xmath21 are opposite with @xmath30 and @xmath31.,title="fig:",width=302 ]
in this section , we consider the case with bpsk signaling , i.e. , @xmath12 . in the following ,
we first propose the symbol - level spoofing signals design and then optimally solve the average sser minimization problem ( [ eqn:5 ] ) in this case . to facilitate the description , as shown in the examples in fig .
[ fig : bpsk ] , we classify the symbols over each block into two types as follows based on the relationship between the original constellation point @xmath10 of alice and the desirable one @xmath21 of the spoofer in each symbol @xmath7 . *
_ type - i symbol _ : the symbol @xmath7 is called a type - i symbol if @xmath10 and @xmath21 are identical ( @xmath29 or @xmath32 ) .
we denote the set of all type - i symbols as @xmath33 . * _ type - ii symbol _ : the symbol @xmath7 is called a type - ii symbol if @xmath10 and @xmath21 are opposite ( @xmath30 and @xmath31 , or @xmath34 and @xmath35 ) .
we denote the set of all type - ii symbols as @xmath36 . in the following two propositions , we present the optimal symbol - level spoofing signal design , and obtain the corresponding sser functions . [
proposition : typei ] given any type - i symbol @xmath37 , it is optimal to minimize the conditional sser @xmath38 by designing @xmath39 aligning with @xmath10 , where @xmath40 denotes the spoofing power for this symbol .
accordingly , @xmath38 is given as @xmath41 where @xmath42 is the error function defined as @xmath43 see appendix [ proof : typei ] .
[ proposition : typeii ] given any type - ii symbol @xmath44 , it is optimal to minimize the conditional sser @xmath45 by designing @xmath46 opposite to @xmath10 , where @xmath47 denotes the spoofing power for this symbol .
accordingly , @xmath45 is given as @xmath48 this proposition can be proved by following a similar procedure as for proposition [ proposition : typei ] .
therefore , the details are omitted for brevity .
propositions [ proposition : typei ] and [ proposition : typeii ] are intuitive . in each type - i symbol , proposition [ proposition : typei ] shows that the spoofing signal should be designed such that at the receiver of bob it is _ constructively _ combined with the original signal from alice , thus increasing the received power of the desirable constellation point against gaussian noise . in each type - ii symbol , proposition [ proposition : typeii ] shows that at the receiver of bob the spoofing signal should be _ destructively _ combined with the original signal from alice , so as to move the constellation point towards the desirable opposite direction . based on these two propositions ,
the average sser minimization problem ( [ eqn:5 ] ) is specified as follows by jointly optimizing the spoofing power @xmath40 s over type - i symbols and @xmath47 s over type - ii symbols .
@xmath49 where the term @xmath50 follows from the fact that each of the two symbol sets @xmath33 and @xmath36 on average occupies a half of all symbols over each block .
the spoofing power allocation problem ( [ problem : bpsk ] ) is generally non - convex , since the sser function @xmath51 in the objective is non - convex over @xmath52 ( as will be shown next ) .
therefore , this problem is difficult to solve . in the following ,
we first show some useful properties of the sser functions @xmath53 and @xmath51 , and then present the optimal solution to problem ( [ problem : bpsk ] ) .
first , we have the following lemma for the sser function @xmath53 . [ proposition:1 ]
@xmath53 is monotonically decreasing and convex over @xmath54 .
it is easy to show that over @xmath54 , the first- and second - order derivatives of @xmath53 satisfy that @xmath55 and @xmath56 , respectively .
therefore , this lemma follows .
next , we study the sser function @xmath51 .
[ proposition:2 ] @xmath51 is monotonically decreasing over @xmath57 .
the convexity of @xmath51 is given as follows depending on alice s transmit power @xmath9 .
* _ alice s low transmit power regime ( i.e. , @xmath58 ) _ : @xmath51 is convex over @xmath57 . * _ alice s high transmit power regime ( i.e. , @xmath59 ) _ : @xmath51 is first convex over @xmath60 $ ] , then concave over @xmath61 , and finally convex over @xmath62 , where the two boundary points @xmath63 are given as @xmath64 see appendix [ proof:2 ] . in the alice s high transmit power regime when @xmath59 , we further have the following property for @xmath51 . [ proposition:3 ] when @xmath59 , there exist two points @xmath65 and @xmath66 with @xmath67 and @xmath68 , such that all the points @xmath69 are above the straight line passing through the two points @xmath70 and @xmath71
. see appendix [ proof:3 ] .
note that the two points @xmath65 and @xmath66 can be found by using the iterative computation procedure in appendix [ proof:3 ] . also note that @xmath65 should be strictly positive ( though very small in general ) , since for type - ii symbols and at the zero spoofing power , the marginal sser with respect to the spoofing power is negative infinity ( @xmath72 ) .
for the purpose of illustration , fig .
[ fig : gbn ] shows an example of @xmath51 with @xmath73 , which validates the structural property of @xmath51 in lemmas [ proposition:2 ] and [ proposition:3 ] .
it is observed that for type - ii symbols , when the spoofing power @xmath47 is between @xmath65 and @xmath66 , `` time - sharing '' between the two spoofing powers @xmath65 and @xmath66 can achieve a lower sser ( or equivalently , a better spoofing performance ) than using the spoofing power @xmath47 constantly . for a @xmath74 portion of time , and @xmath66 for the remaining @xmath75 portion of time , where @xmath76 is uniquely chosen such that @xmath77 for any given @xmath78 . ]
this is essential to help derive the optimal power allocation solution to problem ( [ problem : bpsk ] ) , as shown next .
furthermore , fig .
[ fig : spoofingpower ] shows the values of @xmath79 , @xmath80 , @xmath65 , and @xmath66 versus the transmit power @xmath9 at alice .
it is observed that as @xmath9 increases , the values of @xmath80 and @xmath66 increase while those of @xmath79 and @xmath65 decrease . when @xmath81 , the value of @xmath66 is observed to be larger than @xmath9 , while @xmath65 is observed to be close to zero ( though strictly positive ) .
= 1 with @xmath73.,title="fig:",width=302 ] = 1 , @xmath80 , @xmath65 , and @xmath66 under different transmit power @xmath9 at alice.,title="fig:",width=302 ] now , we present the optimal solution to problem ( [ problem : bpsk ] ) by using the properties of @xmath53 and @xmath51 shown above . to help description
, we define a new function @xmath82 : when @xmath83 , we define @xmath82 to be equivalent to @xmath51 , i.e. , @xmath84 ; while when @xmath85 , we define @xmath86 \cup [ \tau_2,+\infty ) \\
cb_n + d,&{\rm if}~b_n\in(\tau_1,\tau_2 ) , \end{array}\right.\label{eqn : barg}\end{aligned}\ ] ] where @xmath87 and @xmath88 . here
, the points @xmath89 correspond to those on the straight line passing through the two points @xmath70 and @xmath71 . based on lemma [ proposition:3 ] ,
it is evident that @xmath90 serves as a lower bound of @xmath51 over @xmath91 , and importantly , @xmath82 is convex over @xmath92 .
accordingly , we define an auxiliary optimization problem @xmath93 which is convex and whose optimal solution is denoted as @xmath94 and @xmath95 . here , since the strict equality @xmath96 should hold at the optimality of problem ( [ problem : bpsk:5 ] ) , @xmath94 and @xmath95 can be obtained by using a simple bisection search .
note that both @xmath94 and @xmath95 should be strictly positive , which is due to the fact that at the zero spoofing power , the marginal ssers with respect to the spoofing power are both negative infinity ( @xmath97 and @xmath72 ) . with the help of @xmath94 and @xmath95
, we have the following proposition .
[ proposition : solution ] the optimal solution of @xmath98 } to problem ( [ problem : bpsk ] ) is given as @xmath99 and that of @xmath100 is given as follows by considering two cases .
* _ when @xmath59 and @xmath101 _ , the spoofer uses time - sharing between the spoofing powers @xmath65 and @xmath66 , i.e. , the spoofer sets @xmath102 over a @xmath74 fraction of the symbols in @xmath36 , and @xmath103 over the remaining @xmath104 fraction in @xmath36 , where @xmath105 is uniquely chosen such that @xmath106 .
* _ otherwise _ , it follows that @xmath107
. see appendix [ proof : solution ] .
.algorithm for solving problem ( [ problem : bpsk ] ) [ cols= " < " , ] [ table : ii ] therefore , problem ( [ eqn : problem : qpsk ] ) is finally solved , and we summarize the algorithm to solve it in table [ table : ii ] .
in this section , we present numerical results to show the performance of our proposed symbol - level spoofing with optimized power allocation , as compared with two benchmark schemes in the following .
* _ block - level spoofing : _ in this scheme , the spoofer is assumed to be not aware of the original symbol information @xmath10 from alice . in this scheme
, the spoofer uses the constant transmit power @xmath26 over all symbols and sets the spoofing signal to be the exact desirable constellation point @xmath21 , i.e. , @xmath108 . * _ heuristic symbol - level spoofing : _ in this scheme , the spoofer designs its spoofing signals by only heuristically exploiting the symbol - level relationship between the original constellation points of alice and the desirable one of the spoofer , but without the sophisticated transmit optimization as in our proposed optimal symbol - level spoofing .
in particular , for type - i symbols , the spoofer does not allocate any spoofing power to them , since the original and desirable constellation points are already identical ; for other symbols ( i.e. , type - ii symbols for the bpsk case , as well as type - ii and type - iii symbols for the qpsk case ) , the spoofer equally allocates its spoofing power to each of them . as a result , in the bpsk case , the spoofing powers allocated for each type - i and type - ii symbols are @xmath109 and @xmath110 , respectively , and the resultant average sser is given as @xmath111 . in the qpsk case ,
the spoofing powers allocated for each type - i , type - ii , and type - iii symbols are @xmath109 , @xmath112 , and @xmath112 , respectively , and the resultant average sser is given to be @xmath113 . here , the average sser @xmath114 for type - iii symbols is obtained by considering the symbols with @xmath115 and @xmath116 , and allocating the spoofing power to the real components only ( as the imaginary components of the original and desirable constellation points are already identical ) .
first , consider the bpsk case , and fig .
[ fig:4 ] shows the optimal spoofing power allocation versus the average spoofing power @xmath26 , where the transmit power at alice is set as @xmath73 .
it is observed that when @xmath26 is small ( i.e. , @xmath117 db ) , almost all the spoofing power is reserved for type - ii symbols to move the constellation points efficiently towards the desirable opposite directions ; whereas when @xmath26 becomes large ( i.e. , @xmath118 db ) , the spoofing power is allocated more fairly between type - i and type - ii symbols .
particularly , in the small @xmath26 regime when @xmath119 db , it is observed that an `` on - off '' time - sharing strategy between the spoofing power @xmath120 and @xmath121 should be employed over type - ii symbols . in other words , in this regime , the spoofer should focus its spoofing power on a certain portion of type - ii symbols .
= 1 , where the transmit power at alice is set as @xmath73.,title="fig:",width=302 ] = 1 .,title="fig:",width=302 ] fig .
[ fig:2 ] shows the average sser performance of the three schemes versus the spoofing power @xmath26 in the bpsk case , where the transmit power at alice is set as @xmath73 .
it is observed that the optimal symbol - level spoofing achieves a better performance ( or equivalently , a lower average sser ) than the block - level spoofing benchmark . in particular , over 3 db performance gain is obtained by the symbol - level spoofing when the average spoofing power @xmath26 becomes large .
it is also observed that the optimal symbol - level spoofing leads to a lower average sser than the heuristic symbol - level spoofing when @xmath119 db and @xmath118 db , and the two schemes have a similar average sser performance when the value of @xmath26 is between 8.366 db and 13 db .
the results can be explained based on the optimal spoofing power allocation shown in fig .
[ fig:4 ] .
when @xmath119 db , the optimal symbol - level spoofing employs an `` on - off '' transmission strategy with time - sharing between the spoofing power @xmath120 and @xmath121 , thus outperforming the heuristic one that uses fixed spoofing power over all type - ii symbols .
when @xmath118 db , the optimal symbol - level spoofing allocates the spoofing power more fairly between type - i and type - ii symbols , thus reducing the average sser as compared to the heuristic one that only allocates the spoofing power to type - ii symbols .
these results show the significance of our proposed optimal spoofing power allocation .
in addition , it is observed that the heuristic symbol - level spoofing performs worse than the block - level spoofing when @xmath26 becomes large , which is due to the fact that the heuristic symbol - level spoofing does not allocate any spoofing power to the type - i symbols , which leads to the average sser floor .
= 1 .,title="fig:",width=302 ] next , consider the qpsk case .
[ fig:222 ] shows the average sser achieved by the three schemes versus the average spoofing power @xmath26 , where the transmit power at alice is set as @xmath73 .
similar to the bpsk case as in fig .
[ fig:2 ] , the optimal symbol - level spoofing is observed to achieve significantly lower sser than the block - level spoofing , and over 5 db average sser reduction is obtained when @xmath26 becomes large .
furthermore , the optimal symbol - level spoofing achieves lower average sser than the heuristic symbol - level spoofing under any value of @xmath26 .
based on these observations , it follows that the optimal spoofing power allocation is more important with higher - order modulations .
this paper proposed spoofing attacks in the physical layer for the legitimate intervention of malicious wireless communications .
we proposed a new symbol - level spoofing approach for a legitimate spoofer to change the messages transmitted in a malicious link .
with knowledge of the original constellation points by alice , the spoofer exploits the correlations between its desirable constellation points and the original ones by alice to improve the spoofing performance .
in particular , we developed optimal spoofing signals design and power allocation in the cases of bpsk and qpsk modulations .
how to extend the symbol - level spoofing into general modulation techniques ( such as @xmath0-psk and @xmath0-qam modulations with @xmath1 ) and practical cases with fading channels and imperfect / partial transmitted message knowledge are interesting problems worth pursuing in the future .
consider a typical type - i symbol @xmath37 with @xmath122 . in this case , the sser @xmath38 is expressed as @xmath123 where ( [ eqn:6 ] ) follows from ( * ? ? ?
* chapter 5 ) based on the ml detection in ( [ eqn : hatx_n ] ) , and ( [ eqn:6:1 ] ) holds from ( [ eqn : y_n ] ) . in this case , to minimize the sser in ( [ eqn:6:1 ] ) , we should set the phase of @xmath18 to be same as that of @xmath10 , and accordingly we have @xmath124 . in this case , note that the term in ( [ eqn:6:1 ] ) is only dependent on the real part of the cscg random variable @xmath16 , which is a real gaussian random variable denoted by @xmath125 with zero mean and variance @xmath50 .
therefore , we further express the function @xmath38 as @xmath126 where ( [ eqn:8 ] ) holds due to the fact that @xmath127 is also a real gaussian random variable with zero mean and variance @xmath50 .
therefore , this proposition is proved .
* in the alice s low transmit power regime ( i.e. , @xmath58 ) , it can be shown that @xmath130 always holds , and therefore , @xmath51 is convex over @xmath57 .
* in the alice s high transmit power regime ( i.e. , @xmath59 ) , the equation @xmath131 ( equivalently , @xmath132 ) has two solutions given as @xmath79 and @xmath80 in ( [ eqn : zeta1 ] ) and ( [ eqn : zeta2 ] ) , respectively .
it is evident that @xmath130 when @xmath60 $ ] and @xmath62 , and @xmath133 when @xmath61 .
therefore , it follows that @xmath51 is convex over @xmath60 $ ] , concave over @xmath61 , and convex over @xmath62 .
we prove this lemma via two steps .
first , we find two points @xmath134 and @xmath135 with @xmath67 and @xmath68 , which satisfy @xmath136 . here , @xmath137 denotes the slope of the straight line passing through @xmath134 and @xmath135 .
then , we show that all the points @xmath69 are above the straight line passing through such two points .
first , we find such two points @xmath134 and @xmath135 with @xmath136 via the following procedure . to start with ,
we set @xmath138 and @xmath139 . since @xmath51 is concave over @xmath140 $ ]
, it is evident that @xmath141 , i.e. , the slope of the line passing through @xmath142 and @xmath143 is between the values of @xmath144 and @xmath145 .
then , we proceed as follows . * in the first step , we decrease the value of @xmath146 to find a new @xmath147 such that @xmath148 .
note that @xmath51 is convex over @xmath149 $ ] , and thus decreasing @xmath146 leads to the decrease of @xmath150 and the increase of @xmath151 . based on this fact together with @xmath72 , such a point @xmath146 can be obtained via bisection . for this newly found @xmath146
, it follows that @xmath152 .
* in the second step , we increase the value of @xmath153 to find a new @xmath154 such that @xmath155 .
note that @xmath51 is convex over @xmath156 , and thus increasing @xmath153 leads to the increase of @xmath157 and the decrease of @xmath151 . based on this fact together with @xmath158
, such a point @xmath153 can be obtained via bisection . for this newly found @xmath153
, it follows that @xmath159 .
* by iteratively implementing the above two steps , @xmath144 is strictly increased and @xmath145 is strictly decreased , while @xmath151 is always between them . note that @xmath51 is continuous and second - order differentiable . by using
this fact together with @xmath72 and @xmath158 , it is evident that there exist two finite extreme points @xmath65 and @xmath66 , such that @xmath160 .
next , we prove that all the points @xmath69 are above the line passing through @xmath134 and @xmath135 .
first , consider the regimes with @xmath161 $ ] and @xmath62 .
since the function @xmath51 is convex over this regime , and @xmath162 , it is evident that over such two regimes , the points @xmath69 are above the line passing through @xmath134 and @xmath135 . then , consider the regime with @xmath163 $ ] . since @xmath51 is concave over this regime , the points @xmath69
are above the line passing through @xmath142 and @xmath143 , and thus are also above that passing through @xmath134 and @xmath135 . to start with
, we define another auxiliary problem @xmath164 which is obtained based on problem ( [ problem : bpsk ] ) by replacing @xmath51 as @xmath82 .
it is evident that the optimal value of problem ( [ problem : bpsk:4 ] ) is a lower bound on that of problem ( [ problem : bpsk ] ) .
therefore , if the objective value of problem ( [ problem : bpsk ] ) achieved by the solution in this proposition is same as the optimal value of problem ( [ problem : bpsk:4 ] ) , then such a solution is optimal for problem ( [ problem : bpsk ] ) .
we prove this proposition based on this observation .
first , we show that the optimal solution to problem ( [ problem : bpsk:4 ] ) is given as @xmath165 and @xmath107 .
note that both @xmath53 and @xmath51 are convex , and therefore , there exists an optimal power allocation solution in which the spoofing power @xmath40 s and @xmath47 s remain constant over @xmath37 and @xmath44 , respectively .
therefore , we can express @xmath166 and @xmath167 .
accordingly , problem ( [ problem : bpsk:4 ] ) is degenerated to be problem ( [ problem : bpsk:5 ] ) . as a result , the optimal solution to problem ( [ problem : bpsk:4 ] ) is @xmath165 and @xmath107 .
next , based on ( [ eqn : barg ] ) and lemma [ proposition:3 ] , it is easy to verify that the objective value of problem ( [ problem : bpsk ] ) achieved by the solution in this proposition is same as the optimal value of problem ( [ problem : bpsk:4 ] ) achieved by @xmath165 and @xmath107 . therefore , this proposition is proved .
consider one particular type - i symbol with @xmath168 .
let the real and imaginary components of the spoofing signal @xmath18 be denoted as @xmath169 and @xmath170 , and those of @xmath171 as @xmath172 and @xmath173 , respectively
. then @xmath172 and @xmath173 are two real gaussian random variables with mean values of @xmath174 and @xmath175 , respectively , as well as variance of @xmath50 . as a result
, the joint pdf of @xmath172 and @xmath173 is given as @xmath176 note that the spoofing is successful when the phase of @xmath171 lies between @xmath109 and @xmath177 ( within the detection regime ) , i.e. , the real and imaginary components of @xmath171 are both positive .
therefore , the conditional sser under given @xmath169 and @xmath170 is given as @xmath178 to minimize the above conditional sser under the given transmit power @xmath40 , i.e. , @xmath179 , it is desirable to set @xmath180 and @xmath181 . as a result ,
obtaining @xmath169 and @xmath170 is equivalent to solving the following problem : @xmath182 note that the function @xmath183 is concave over @xmath184 .
therefore , by substituting @xmath185 and @xmath186 into problem ( [ eqn : problem : tilde : f1 ] ) , we can show that the optimality is obtained as @xmath187 .
therefore , the optimality of the problem ( [ eqn : problem : tilde : f1 ] ) is achieved when @xmath188 . by using this together with ( [ eqn : g1 ] ) , the conditional sser in ( [ eqn : tilde : f1 ] )
is obtained .
therefore , this proposition is proved .
similar to the proof of proposition [ proposition : qpsk:1 ] , the joint pdf of @xmath172 and @xmath173 is given in ( [ eqn : joint : pdf ] ) .
note that the spoofing is successful when the phase of @xmath171 lies between @xmath189 and @xmath190 , i.e. , the real and imaginary components of @xmath171 are both negative . as a result , the conditional sser @xmath191 is given as @xmath192 to minimize the above conditional sser under the given spoofing power @xmath47 , i.e. , @xmath193 , it is desirable to set @xmath194 and @xmath195 . as a result ,
obtaining @xmath169 and @xmath170 is equivalent to solving the following problem : @xmath196 by replacing @xmath169 and @xmath169 as @xmath197 and @xmath198 , problem ( [ eqn : problem : tilde : f2 ] ) is recast as @xmath199 note that when @xmath200 , @xmath201 is concave , and therefore , the optimality of problem ( [ eqn : problem : tilde : f2:2 ] ) is obtained as @xmath202 . as a result ,
the optimality of problem ( [ eqn : problem : tilde : f2 ] ) is achieved when @xmath203 . by using this together with ( [ eqn : g2 ] ) , the conditional sser in ( [ eqn : tilde : f2 ] )
is obtained .
therefore , this proposition is proved .
it is evident that @xmath204 is monotonically decreasing over @xmath205 .
therefore , we only need to show its convexity . the second - order derivative of @xmath204 is given as @xmath206 with @xmath205 , it follows that @xmath207 note that @xmath208 for all @xmath209 , and therefore , @xmath210 . by using this together with ( [ eqn : inequality:1 ] ) and ( [ eqn : inequality:2 ] )
, it follows that @xmath211 .
accordingly , @xmath212 for all @xmath209 . as a result , @xmath204 is a convex function , and this lemma is proved .
first , it is easy to see that @xmath215 as @xmath216 , and @xmath217 .
since @xmath213 is a continuous function , there always exists a positive @xmath218 with @xmath219 , such that over @xmath220 $ ] we have @xmath221 , i.e. , @xmath213 is convex .
next , note that when @xmath222 , it follows that @xmath223 .
also , when @xmath224 , we have @xmath225 . by combining the above two facts , when @xmath226 , it holds that @xmath227 .
accordingly , @xmath221 and @xmath213 is convex .
y. zou , x. wang , and l. hanzo , `` a survey on wireless security : technical challenges , recent advances and future trends , '' to appear in _ proc .
ieee_. [ online ] available : http://arxiv.org/abs/1505.07919 .
y. zeng and r. zhang , `` wireless information surveillance via proactive eavesdropping with spoofing relay , '' to appear in _
ieee j. sel .
topics signal process._. [ online ] available : https://arxiv.org/abs/1606.03851 .
m. alodeh , s. chatzinotas , and b. ottersten , `` constructive multiuser interference in symbol level precoding for the miso downlink channel , '' _ ieee trans .
signal process .
9 , pp . 2239 - 2252 , may 2015 . | with recent developments of wireless communication technologies , malicious users can use them to commit crimes or launch terror attacks , thus imposing new threats on the public security . to quickly respond to defend these attacks , authorized parities ( e.g. , the national security agency of the usa ) need to intervene in the malicious communication links over the air .
this paper investigates this emerging wireless communication intervention problem at the physical layer . unlike prior studies using jamming to disrupt or disable the targeted wireless communications , we propose a new physical - layer spoofing approach to change their communicated information .
consider a fundamental three - node system over additive white gaussian noise ( awgn ) channels , in which an intermediary legitimate spoofer aims to spoof a malicious communication link from alice to bob , such that the received message at bob is changed from alice s originally sent message to the one desired by the spoofer . we propose a new symbol - level spoofing scheme , where the spoofer designs the spoofing signal via exploiting the symbol - level relationship between each original constellation point of alice and the desirable one of the spoofer .
in particular , the spoofer aims to minimize the average spoofing - symbol - error - rate ( sser ) , which is defined as the average probability that the symbols decoded by bob fail to be changed or spoofed , by designing its spoofing signals over symbols subject to the average transmit power constraint . by considering two cases
when alice employs the widely - used binary phase - shift keying ( bpsk ) and quadrature phase - shift keying ( qpsk ) modulations , we obtain the respective optimal solutions to the two average sser minimization problems .
numerical results show that the symbol - level spoofing scheme with optimized transmission achieves a much lower average sser , as compared to other benchmark schemes .
wireless communication surveillance and intervention , symbol - level spoofing , spoofing - symbol - error - rate ( sser ) minimization , power control . [ section ] [ section ] [ section ] [ section ] [ section ] [ section ] [ section ] [ section ] |
this di - hadron and single - proton cross section measurements are based on data samples of @xmath1 and @xmath5 , respectively , collected with the belle detector at the kekb asymmetric - energy @xmath15 ( 3.5 gev on 8 gev ) collider @xcite operating at the @xmath2 resonance ( denoted as on - resonance ) as well as 60 mev below for comparison ( denoted as continuum ) . the belle detector is a large - solid - angle magnetic spectrometer that consists of a silicon vertex detector ( svd ) , a 50-layer central drift chamber ( cdc ) , an array of aerogel threshold cherenkov counters ( acc ) , a barrel - like arrangement of time - of - flight scintillation counters ( tof ) , and an electromagnetic calorimeter comprised of csi(tl ) crystals ( ecl ) located inside a superconducting solenoid coil that provides a 1.5 t magnetic field .
an iron flux - return located outside of the coil is instrumented to detect @xmath27 mesons and to identify muons ( klm ) .
the detector is described in detail elsewhere @xcite .
two inner detector configurations were used .
a 2.0 cm beampipe with 1 mm thickness and a 3-layer svd were used for the first sample of @xmath28 , while a 1.5 cm beampipe , a 4-layer svd and a small - cell inner drift chamber were used to record the remaining @xmath29(@xmath5 for the single - hadron analysis ) @xcite .
the primary light- and charm - quark simulations used in this analysis were generated with pythia6.2 @xcite , embedded into the evtgen @xcite framework , followed by a geant3 @xcite simulation of the detector response .
the various mc samples were produced separately for light ( @xmath30 ) and charm quarks .
in addition , we generated charged and neutral @xmath31 meson pairs from @xmath2 decays in evtgen , @xmath32 pair events with the kkmc @xcite generator and the tauola @xcite decay package , and other events with either pythia or dedicated generators @xcite .
events with at least three reconstructed charged tracks must have a visible energy of charged tracks and neutral clusters above 7 gev ( to remove @xmath32 pair events ) and either a heavy jet mass ( the greater of the two invariant masses of all particles in a hemisphere ) above 1.8 gev / c@xmath33 or a ratio of the heavy jet mass to visible energy above 0.25 . tracks must be within 4 cm ( 2 cm ) of the event vertex along ( perpendicular to ) the positron beam axis .
each must have at least three svd hits and fall within the barrel and full particle - identification ( pid ) polar - angle acceptance of @xmath34 .
the fractional energy of each track must exceed 0.1 .
( note that , in this paper , we study fragmentation functions for @xmath11 above 0.2 ) .
this initial fractional energy selection always takes the nominal hadron mass as given by the pid information into account .
the requirement of @xmath35 therefore safely accommodates pion - kaon misidentification , which is unfolded in the course of this analysis .
illustration of di - hadron fragmentation where the final - state hadrons are depicted as red arrows , the incoming leptons as blue arrows , and the event plane spanned by leptons and thrust axis is depicted as a light blue plane . in this case , both hadrons are found in opposite hemispheres defined by the thrust axis , and generally out of the plane , as indicated by the cones.,width=302 ] in addition , in order to study whether two hadrons have likely emerged from the same parton or different partons , the analysis is performed on several different sets by requiring that both hadrons be in opposite hemispheres , the same hemisphere , or anywhere as depicted in figs . [ fig:(h)(h)x ] and [ fig:(hh)x ] . for the data sets where a hemisphere assignment is required , the hemispheres are defined by the plane perpendicular to the thrust axis and the thrust must satisfy @xmath36 .
illustration of di - hadron fragmentation where the final - state hadrons are depicted as red arrows , the incoming leptons as blue arrows , and the event plane spanned by leptons and initial quarks / thrust axis is depicted as a light blue plane . in this case , both hadrons are found in the same hemisphere as defined by the thrust axis , and generally out of the plane , as indicated by the cones.,width=302 ] to apply the pid correction according to the pid efficiency matrices described in ref .
@xcite , the same selection criteria must be applied to define a charged track as a pion , kaon , proton , electron or muon .
the information is determined from normalized likelihood ratios that are constructed from various detector responses .
if the muon - hadron likelihood ratio is above 0.9 , the track is identified as a muon .
otherwise , if the electron - hadron likelihood ratio is above 0.85 , the track is identified as an electron . if neither of these applies , the track is identified as a kaon by a kaon - pion likelihood ratio above 0.6 and a kaon - proton likelihood ratio above 0.2 .
pions are identified with the kaon - pion likelihood ratio below 0.6 and a pion - proton ratio above 0.2 .
finally , protons are identified with the inverse proton ratios above with kaon - proton and pion - proton ratios below 0.2 .
while neither muons nor electrons are considered explicitly for the single and di - hadron analysis , they are retained as necessary contributors for the pid correction , wherein a certain fraction enter the pion , kaon and proton samples under study .
in the following sections , the di - hadron yields are extracted and , successively , the various corrections and the corresponding systematic uncertainties are applied to arrive at the di - hadron differential cross sections @xmath37 . for the di - hadron cross sections , a ( @xmath25 , @xmath26 ) binning is used .
we forgo a combined @xmath11 and invariant - mass binning of the hadron pair ; the latter , in particular , is relevant in the _
same_-hemisphere topology as an unpolarized baseline to the previously extracted interference fragmentation functions @xcite and would have allowed the extraction of individual fragmentation functions for @xmath38 , @xmath39 , @xmath40 and other resonances .
the @xmath25 and @xmath26 ranges of 0.2 to 1.0 used in this analysis are each partitioned into 16 equidistant bins .
all hadron and charge combinations are treated independently and are merged only after all corrections are applied and after confirming their consistency where applicable ( _ i.e. , _ where the same combinations of fragmentation functions appear , such as @xmath41 and @xmath42 )
. this leaves 16 different charge and type combinations for pions and kaons initially , of which six contain irreducible information .
furthermore , as mentioned in the introduction , three hemisphere combinations are studied : two hadrons in the same hemisphere , two hadrons in opposite hemispheres , and two hadrons irrespective of hemisphere or thrust cut ; these are abbreviated hereinafter as _ same _ , _ opposite _ , and _ any _ , respectively . as in ref .
@xcite , the particle misidentification is corrected via inverted @xmath43 particle - misidentification matrices for the five particle hypotheses ( pions , kaons , protons , muons , and electrons ) for each identified particle , laboratory momentum , and polar angle bin .
these matrices are obtained using decays of @xmath44 , @xmath45 and @xmath46 from data where the true particle type is determined by the charge reconstruction and the invariant mass distribution .
occasionally , when too few events are available in the data , the extracted efficiencies are interpolated and/or extrapolated based on the behavior in the generic mc ; this occurs particularly at the boundaries of the acceptance .
the matrices are calculated for each of the two - dimensional bins in laboratory momentum and polar angle , with the boundaries of the @xmath47 bins in momentum at ( @xmath48 ) gev/@xmath49 and the boundaries of the 9 bins in @xmath50 at ( @xmath51 @xmath52 ) . in this analysis ,
the inverted misidentification matrix is applied for each of the identified hadrons by multiplying the respective weights for each hadron being a pion or kaon to obtain the total weight for the di - hadron and any of the four pion - kaon combinations . to confirm the consistency of this treatment , the @xmath53 branching ratios for the pion - pion and kaon - kaon decay channels to the pion - kaon decay channel
are compared to the pdg @xcite values and found to be consistent .
we confirm that the total yield of particle pairs is unaffected by this treatment .
the corrected yields are distributed among the ( @xmath54 ) bins according to the corresponding hadron masses : one identified hadron pair appears in several @xmath11 bins with the above - determined weights , depending on the particular hadron combination .
the ratios relative to the uncorrected hadron assignment are displayed in fig .
[ fig : beforeafterpid ] , where one can see that the overall corrections are of the order of 20% to 50% .
the uncertainties on the pid matrices are taken into account as uncertainties in the di - hadron yields and propagated through the subsequent corrections . at present , the uncertainties are only assigned individually for each hadron combination , neglecting the correlations between different hadron combinations .
they follow the uncertainties assigned in ref .
@xcite but take into account the additional complication of having two rather than one hadrons to unfold . to obtain the final uncertainties , the asymmetric uncertainties on the inverted pid matrices
are sampled @xmath55 times with a random generator with gaussian distributions around the central value ( separately above and below this value ) . from the resulting yields ,
the 68th percentiles from @xmath55 samples relative to the central values for each ( @xmath56 ) bin are taken as the systematic uncertainties on the pid - corrected di - hadron yields due to the pid matrix evaluation and inversion uncertainties .
the size of the statistical and systematic uncertainties relative to the pid - corrected di - hadron yields is displayed in fig . [
fig : pidsyst ] for selected hadron combinations .
as expected from the overall size of the yields , the statistical precision is best for opposite - sign pion pairs , followed by pion - kaon and then kaon pairs , with the same - sign precision being generally lower .
the pid systematic uncertainties are mostly smaller than the statistical uncertainties in almost any bin , with the exception of the lowest @xmath11 bins .
the reconstructed fractional momentum @xmath11 of each hadron may have been smeared from its actual value and therefore must be corrected . for this purpose ,
the generic mc simulations are used to create two - dimensional histograms with @xmath57 bins of generated and reconstructed ( @xmath56 ) combinations for each of the two hadrons .
only events that are generated and reconstructed within this range of @xmath11 are considered .
events outside this range are treated in the manner described later in the acceptance correction section .
the two - dimensional response histograms are created for each hadron and charge combination and for all topology assignments .
as the pid correction was already applied to the data before the smearing correction , the true particle - type information in the mc is selected for both generated and reconstructed ( @xmath56 ) . in the smearing matrices , the diagonal elements are dominant for all hadron combinations , as can be expected in such a coarse binning . in the case of the pion - kaon and kaon - kaon combinations , the non - diagonal elements are slightly larger than for the pion - pion case , which indicates that the kaon smearing is slightly larger than that of pions .
the small off - diagonal components facilitate the inversion of the smearing matrix significantly and a simple , analytically inverted matrix should be sufficient to unfold the di - hadron yields
. however , the singular value decomposition ( svd ) unfolding method @xcite is used as our default to properly unfold the statistical uncertainties and assign systematic uncertainties due to the limited mc statistics , especially for bins far from the diagonal .
this also takes into account the possible effects of the different @xmath11 distributions in data and mc , which will be discussed later . in the svd unfolding , a regularization parameter @xmath58 accounts for the lack of statistics in the smearing matrix entries and the shape of the mc spectra .
if @xmath58 is not selected properly , the unfolded yields can be either too biased by the mc spectrum ( too small a value for @xmath58 ) or can exhibit large fluctuations ( too large a value for @xmath58 ) . following the procedure of ref .
@xcite , the best regularization parameter is chosen when the index of the regularized vector becomes smaller than unity . initially , the fluctuations observed in the regularization parameter distribution are large , so that the false choice of too - small @xmath58 values leads to spurious discrepancies at very high @xmath11 between distributions of the same physics content that were consistent before unfolding . after smoothing the @xmath58 distributions ,
the expected behavior is markedly better ( _ i.e. , _ a relevant exponentially falling contribution and an irrelevant flat contribution due to mc statistical fluctuations ) and the selection of the regularization parameter is considered more reliable .
the smearing is corrected in all data samples after the pid correction is applied and before the non-@xmath59 events are removed .
the final before / after ratio plots are displayed in fig .
[ fig : beforeaftersmearing ] as a function of ( @xmath56 ) .
apart from the highest ( @xmath56 ) bins , where the corrections get large , the yield ratios are close to unity .
all uncertainties prior to the smearing - unfolding ( pid and statistical uncertainties ) are unfolded as well , resulting in the respective covariance matrices .
the covariance matrix due to the mc statistics itself and the differences with an analytic unfolding ( _ i.e. , _ application of the inverted response matrix ) are assigned as systematic uncertainties related to the unfolding . various qed processes can produce hadronic final states that contribute to our di - hadron yields and must be removed .
apart from particle misidentification , which has been addressed already , @xmath60 and bhabha scattering processes can not contribute to hadronic final states , as has been verified in mc simulations .
the processes that do produce hadron pairs are either qed processes having partons created initially , such as two - photon processes @xmath61 and @xmath62 , or via decays such as from @xmath63 .
hadrons from these processes are not produced directly via @xmath64 and hence should not be included in our extracted di - hadron cross sections .
similarly , resonant @xmath2 production and subsequent decays into neutral or charged @xmath31 meson pairs create pion and kaon pairs that must be removed ( the non - resonant @xmath65 process does not contribute @xcite ) .
the direct production of quark - antiquark pairs in electron - positron annihilation @xmath66 and @xmath67 is treated as signal in this section , while weak decays in these continuum processes will be treated later .
figure [ fig : zqedcontpid0_mix4 ] shows the relative fractions of all these processes for selected hadron pairs in the _ any _ di - hadron topology . due to the large branching fraction of the single - prong @xmath32 decay for at least one of the @xmath32 leptons , @xmath32 processes are the dominant background for pions from small to especially large fractional energies where the single hadron inherits a large fraction of the @xmath32 momentum .
the other non-@xmath59 processes generally play a minor role with contributions less than a few percent , with the exception at high @xmath11 where the two - photon process @xmath68 contributes several percent to pion pairs .
resonant @xmath2 production either in charged or neutral ( @xmath69-mixed ) @xmath31 meson decays contributes a few to about 10% and it vanishes when one fractional energy approaches 0.5 due to the additional decays needed to produce pions and kaons .
the distributions for other di - pion topologies are similar except that the additional thrust requirement removes nearly all @xmath2 decays . for same - sign di - pions , the @xmath32 contribution is substantially smaller as the single - prong decays of oppositely charged @xmath32 create predominantly oppositely - charged pions . for _ same_-hemisphere di - pions , the single - prong @xmath32 decays can not contribute and consequently the relative @xmath32 contributions are below 10 - 20% everywhere . for kaon - related di - hadron combinations , the overall non-@xmath59 contributions are as small as for di - pions , but @xmath70 and @xmath71 are more important .
in addition , the @xmath32 decays do not play a substantial role due to the suppressed kaonic decays .
charm decays generally produce more ckm - favored @xcite kaons than ckm - suppressed pions .
this results in a generally larger fraction of charm events contributing to the pion - kaon and kaon - kaon cross sections : up to 60% for kaon pairs at the lowest @xmath11 , with a similar fall - off as for pion pairs .
similarly , @xmath72 decays favor kaons over pions and thus their fractions are as high as 20% ( summed ) , rapidly disappearing at higher @xmath11 . assuming that the non-@xmath59 and @xmath73 mc reliably describe the data , the background contributions are directly subtracted from the data distributions .
in this way , we avoid introducing further uncertainties due to the shape of the @xmath74 mc .
as all these processes are qed and @xmath2 processes , they are very well understood at the theory level .
the yield uncertainty is 1.4% for the @xmath75 process @xcite but is substantially larger for the two - photon processes due to associated production , which is not taken into account in the current two - photon simulations .
a factor of four relative to the nominal yield has been assumed for the latter @xcite .
for the systematic uncertainties due to the non-@xmath59 background correction , these overall uncertainties as well as the statistical uncertainties in the non-@xmath59 mc , are taken into account .
the preselection and acceptance correction is divided into three separate terms , motivated by the different sources of corrections and to better expose their individual effects .
the first takes into account the effect on the reconstruction within the specified acceptance selection , mostly due to the preselection criteria and decays in flight ; the second treats the losses outside the barrel acceptance ; and the third takes into account potential losses as @xmath76 approaches unity , which are not properly described in the generic mc .
the first correction incorporates generated hadron pairs within the barrel geometry that do not get reconstructed . as particle identification ,
non-@xmath59 removal and smearing have already been applied , the reconstructed events are considered based on the generic reconstructed @xmath74 mc information , but taking the mc - truth particle type and momenta instead of the reconstructed values .
the correction factor is calculated as the ratio of reconstructed to generated events per ( @xmath56 ) bin for each hemisphere assignments , hadron types , and charge combinations .
this correction takes also into account the events that were initially smeared out of or into the @xmath11 range considered for this analysis .
the efficiencies are relatively flat at around 70% and only drop substantially at higher @xmath11 .
this is similar to the behavior noticed in ref .
@xcite , where it was found to be mostly due to the preselection criteria , especially the heavy jet mass restriction that disfavors high-@xmath11 events where the hadrons naturally have to be more aligned with the thrust axis as little other energy remains . also , the minimum track requirement of three disfavors very high @xmath11 hadron pairs , where for the same reason the multiplicity is small . a certain fraction of di - hadrons are not reconstructed because at least one of the hadrons is outside of the barrel acceptance .
this fraction is evaluated by comparing the generated mc within the barrel acceptance ( _ i.e. , _ including the acceptance selection criteria ) with the generated mc without the acceptance requirement .
this acceptance fraction is around 70% and rather flat as a function of ( @xmath56 ) , increasing slightly at very high fractional energies .
the only systematic uncertainties related to these acceptance corrections ( both within and outside the barrel region ) originate from the statistical uncertainty of the fractions within the acceptance .
these uncertainties are rather moderate in comparison to all other systematic and statistical uncertainties that are aggregated in section [ sec : syst ] .
the generated mc does not necessarily reproduce the hadron distributions well for very forward or backward polar angles in the cms .
ideally , the hadron polar angular distributions should resemble those of the initially produced quark - antiquark pairs and thus follow a @xmath77 dependence , neglecting the small linear dependence due to @xmath78 interference .
the lower the fractional energy , the less pronounced this behavior : this is due to the additional smearing by the transverse momentum generated in the fragmentation process .
while such a behavior is roughly visible at smaller polar angles , the distributions rapidly drop off at higher polar angles as if some remaining acceptance cut is still present . as a consequence , the previous acceptance and efficiency corrections are not complete and need to be further corrected for this effect . as the dependence at smaller polar angles
is well described by the expected parabola , this function is used to fit the mc and compare the areas below the fit - result curve and the actual histograms . in principle
, this treatment should be independent for the two hadrons and can be applied by multiplication of the two individual correction factors .
an expected increase of the correction with increasing @xmath26 , due to higher-@xmath11 tracks being more collimated and thus closer to the partonic polar angular dependence , has been confirmed .
the overall effect of this last acceptance correction is on the order of a few percent .
the effect of all three acceptance and efficiency corrections is summarized in fig .
[ fig : acceptance ] , where the ratios of the di - hadron yields before and after the corrections are displayed .
the overall effect amounts to between two times the initial yields at moderate ( @xmath56 ) and more than ten times at very high @xmath11 ( where the event preselection correction dominates ) .
generally , fragmentation functions are only defined for hadrons produced by qcd processes and decays and so any weak decays should be removed . in practice
this is only possible if at all with the help of mc and not entirely reliable .
therefore , many fragmentation results do not exclude weak decays or only those experimentally detectable such as those of @xmath45 baryons and neutral kaons .
the approach taken here is to provide results that either contain all weak decays or completely remove them with the help of mc .
every @xmath67 event undergoes at least one weak decay to produce a pion or kaon .
however , in the fragmentation process , various quark - antiquark pairs are created and consequently pions and kaons can be created that did not originate directly from the decays of charmed hadrons . the only way to separate them is by following the parents of each final state hadron in the mc to either a gluonic string , which corresponds to the absence of a weak decay , or a hadron with a different , non - light valence flavor . in the latter case ,
a weak decay was present and this hadron would have to be removed .
the difficulty is rapidly ( algorithmically ) determining this information for a given hadron type . in the di - hadron analysis
, it can be argued that the chance of at least one of the two hadrons being from a weak decay is much higher for charm events and that removing all charm events is a valid approximation .
however , this needs to be tested .
the mc history of each hadron is studied to find weak decays .
the heaviest flavor of each particle in the decay chain is selected and compared to the mother particles .
if the decay chain ends at a string without a change in its heaviest flavor , no weak decay is present .
if the flavor does change , a few strong decays need to be vetoed before asserting the presence of a weak decay .
examples are various vector mesons and other excited states with nonzero strangeness where the strangeness is retained in a lower mass state , such as @xmath79 . also ,
various @xmath80 and @xmath67 resonances need to be excluded as they also decay strongly despite the zweig rule @xcite . the overall weak- and strong - decay fractions as a function of ( @xmath56 ) are shown in fig .
[ fig : zweakratiop_mix4 ] for the main particle combinations within the _ any _ topology .
similar results are obtained for the other two topologies .
it should be noted that the assumption of charm events creating only weak - decay pions and kaons is almost fulfilled in the procedure mentioned above but that a small fraction of charm di - hadron events nevertheless originates in strong decays .
overall , the fraction of strong decays dominates in all ( @xmath56 ) bins for pion pairs , while the higher fraction of charm events results in larger weak fractions for pion - kaon and is even more pronounced in kaon - kaon combinations . in all cases , the weak fractions drop with @xmath11 as the additional decays
soften the spectrum .
these strong / weak fractions are model - dependent statements as the fragmentation process is only approximated in pythia .
furthermore , the absolute size of weak decays within pythia depends also on the fragmentation settings .
the uncertainties due to these effects are evaluated by comparing the generic belle mc to the pythia default settings in the mc .
the strong fractions for the pythia default mc are given in the plots as well for comparison .
as can be seen , they are rather similar but are generally slightly lower at high @xmath11 where the generally harder spectra in the default pythia settings allow for slightly more weak decays to be present .
the differences are assigned as a systematic uncertainty for the cross sections that have the weak processes removed .
initial - state radiation reduces the cms energy of the produced quark - antiquark pair .
consequently , the fractional energies calculated relative to the nominal cms energy are not correct .
this can alter the shape of the actual @xmath11 dependence of the fragmentation functions and also invalidates pqcd calculations evaluated at the nominal cms energy .
the correction procedure relies on the strategy applied in ref .
@xcite for the single - hadron cross sections .
the events are classified according to their difference from the nominal cms energy ; the events with a cms energy below 99.5% of the nominal energy are removed .
ideally , one would want to observe the initial - state radiation directly in the reconstructed data ; however , most photons are in the very forward and backward regions , outside the belle acceptance .
instead , generated mc data are used to directly identify isr photons and remove their energies from the total cms energy . in the mc , the isr photons are identified by having their mother particles be an initial - state lepton .
the fraction of such events depends on the fractional energies of the two final - state hadrons .
if a large amount of the energy is removed by the photons from the produced quark - antiquark system , very high fractional energies with respect to the nominal @xmath81 are inaccessible .
therefore the fraction of non - isr events ( _ i.e. , _ less than 0.5% cms energy loss ) increases with increasing fractional energies .
this is indeed the case as can be seen in fig .
[ fig : isrzfraction_pid0_mix4 ] for the _ any _ topology hadron pairs ( and similarly for the other two topologies ) .
the events are then corrected by this fraction to obtain the isr - free differential cross sections at the nominal center - of - mass energy . since the isr fraction depends on the fractional energies of the hadrons , the ( @xmath56 ) distribution shape of the mc simulation enters in the isr correction . to address the dependence of the isr correction on the shape in the mc
, an alternative mc is used for comparison and the differences in the extracted di - hadron cross sections is assigned as systematic uncertainties .
these fractions are also shown in fig .
[ fig : isrzfraction_pid0_mix4 ] .
the isr fractions are found to be consistent within the limited precision for both pythia settings .
the total impact of all corrections from the particle misidentification to the correction for isr can be seen successively in fig .
[ fig : zzbeforeall ] for the main hadron combinations without a hemisphere assignment .
the overall correction to the raw yields is substantial , predominantly due to the necessary acceptance corrections .
they are comparable for most ( @xmath56 ) bins but rise at the highest @xmath11 bins due to the acceptance and smearing corrections . to confirm the consistency of the results , various tests are performed .
for example , the dependence on the data - taking periods is studied ; after taking into account variations in acceptance and reconstruction efficiency , the cross sections are consistent within several percent between different periods and no additional systematic uncertainty is assigned . in another study , we compare the data recorded at the @xmath2 resonance with the smaller off - resonance data sample . after removal of the @xmath2 decay contributions in the non-@xmath59 correction ,
the results from both collision energies are consistent . in
yet another set of comparisons with the same physics - related information , such as charge conjugation of both particles ( @xmath82 , etc . ) or ( random ) hemisphere assignments ( @xmath83 ) , no systematic differences beyond the assigned uncertainties are found . [ cols="^,^,^,^,^,^,^",options="header " , ] all diagonal systematic uncertainties are summed in quadrature .
the total relative systematic uncertainties along with the statistical uncertainties are displayed in fig .
[ fig : systall_mix4_sum ] for the relevant hadron pairs without topology assignment for diagonal ( @xmath56 ) bins and in table [ tab : systall ] for the entire measurement range .
this measurement is limited almost everywhere by the systematic uncertainties , for which the dominant contributions arise from the smearing correction except at high @xmath11 where the rapidly falling mc precision contributes comparably . with
increased mc data the systematic uncertainties could be reduced to the level of the statistical uncertainties and be dominated by the smearing correction .
additionally , there are global scale uncertainties due to the luminosity measurement ( 1.4% ) and the track reconstruction ( [email protected]% ) are not shown .
the final cross sections for the main hadron - pair combinations are presented in fig .
[ fig : yieldall_mix4_pid0_1_4_12 ] for all ( @xmath56 ) bins and without topology assignment .
the results shown here and elsewhere include weak decays unless otherwise noted .
as expected , the opposite - sign pion pairs have the largest cross sections at all @xmath11 combinations , followed by the same - sign pion pairs .
however , the oppositely charged pion - kaon and kaon - kaon combinations seem to be of similar magnitude or even larger than the same - sign pions at higher @xmath11 , which might be explained by the potentially larger favored fragmentation combination from strange - quark pairs .
same - sign kaon pairs have the lowest cross sections in general , with the relative differences from the other combinations increasing at increasing @xmath11 .
as at least one kaon in this case needs to be produced from disfavored fragmentation , the additional strangeness suppresses the cross sections beyond that for the disfavored pion fragmentation functions .
the cross sections for di - hadrons in the _ same _ hemisphere are displayed in fig .
[ fig : yieldall_mix1_pid0_1_4_12 ] for all ( @xmath56 ) bins .
the cross sections fall off rapidly and mostly disappear at the boundary @xmath85 , where the total energy of one initial parton is fully contained in the energy of the two hadrons .
a small excess above this limit can be seen .
mc studies show that this excess can be explained qualitatively by a small mis - assignment of hemisphere due to the smearing of the thrust axis relative to the initial quark - antiquark axis .
in addition , hard gluon radiation may create such events .
as several of the uncertainties are common to all charge and hadron combinations , these cancel in ratios and the information about favored and disfavored fragmentation should be more reliable .
for example , ignoring strange - quark fragmentation to pions , the same - sign to opposite - sign pion - pair ratios are simple measures of disfavored _ vs. _ favored pion fragmentation functions for light quarks . as can be seen in the ratios in fig .
[ fig : allratio_pid15_mix16_c8 ] , they show a similar nearly flat behavior at low fractional energies , but deviate substantially from this trend at higher fractional energies .
generally , the opposite - sign pion - kaon and kaon - kaon ratios are more suppressed than the already disfavored same - sign pion pairs , but at the highest ( @xmath56 ) do become comparable . this might be related to the fact that opposite - sign kaons can be favored fragmentation for both hadrons if a @xmath80 was created initially . as the opposite - sign pion - kaon pairs are similar in magnitude , even at high @xmath11 , it can be argued that disfavored fragmentation from strange quarks to pions is not as suppressed .
same - sign kaon pairs , where at least one strange - quark pair needs to be produced in the fragmentation , are always suppressed at least one order of magnitude relative to the opposite - sign pion pairs .
this shows that strangeness produced in fragmentation is indeed strongly suppressed , as is generally assumed in fragmentation models such as those included in pythia .
figure [ fig : xsecs ] displays all six relevant hadron combinations for _ opposite _ hemispheres while fig . [
fig : xsecs2 ] shows the cross sections for hadrons within the _ same _ hemisphere using a thrust requirement of @xmath36 .
note that the requirement of a minimum thrust value is not corrected for in these hemisphere decompositions , which must be taken into account when used for global ff analyses .
as expected , the cross sections at small @xmath11 are of similar magnitude between the _ same _ and _ opposite _ hemispheres , while at higher @xmath11 only _ opposite_-hemisphere pairs remain .
these cross sections with hemisphere assignment can be compared to the cross sections without the hemisphere assignment and without the thrust requirement , as shown in fig .
[ fig : plotmixhemisum_c8 ] for diagonal ( @xmath56 ) bins .
as expected , the _ same _ hemisphere contributions are comparable to those from _ opposite _ hemispheres at very low @xmath11 while they vanish rapidly with increasing @xmath11 .
this is in agreement with the assumption of the _ same _ hemisphere di - hadrons emerging predominantly from the common initial parton .
if so , the sum of the two hadron s fractional energies can not exceed unity and they do indeed drop to zero at ( @xmath56 ) each around 0.5 .
the small difference between the sum of the _ same _ and _ opposite _ hemispheres and the _ any _ topology assignment at low @xmath11 is due to the additional thrust selection used to identify hemispheres .
the small deviations seen occasionally at high @xmath11 are related to variations from the smearing unfolding and are consistent within the uncertainties that are not shown in this figure .
the di - hadron cross sections in the _ any _ topology are compared to various fragmentation settings within the pythia / jetset mc generator .
these are displayed in fig .
[ fig : allratiosum_mix16_c8_32518153 ] for the main hadron combinations and diagonal ( @xmath56 ) bins .
the settings correspond to the default pythia , those currently used in belle for the fully tracked geant simulations , as well as various other settings tuned to specific experiments , collision systems , and energies from the lep / tevatron , aleph , and hermes environments .
generally , all parameterizations agree at very low @xmath11 as the total production yields for certain particles are best known .
however , at very high @xmath11 , the distributions differ greatly .
it appears that for all six hadron and charge combinations the default pythia and the latest belle settings describe the data best even at large fractional energies .
the lep - based tunes generally overshoot the data at high @xmath11 while the older belle and hermes tunes fall off much too rapidly .
in addition to the _ any _ topology combination , which is dominated by the _ opposite _ topology at higher fractional energies , the _ same _ hemisphere combination is compared to these mc tunes . in principle , the different hemisphere combinations are sensitive to different parameters in the pythia settings .
an example of the comparison for @xmath86 pairs is displayed in fig .
[ fig : allxsec_pid16_mix1_c8_32518153 ] ; other hadron combinations are available in the supplementary file .
the overall behavior is similar to fig .
[ fig : allratiosum_mix16_c8_32518153 ] , with the older belle and hermes tunes undershooting the data .
the other parameterizations do not differ as substantially as in the _ any _ hemisphere combinations and they all reproduce the data reasonably well .
the differences in the pythia settings are summarized in table 1 of the supplementary file .
in addition to the di - hadron analysis , the production of single hadrons , especially previously unpublished single protons is considered here .
the single - hadron analysis follows the same procedure as the di - hadron analysis .
the @xmath11 range between 0.1 and 1.0 is divided into 36 bins ; for protons , @xmath87 0.2 is kinematically inaccessible .
the particle misidentification correction is performed as in the di - hadron analysis ( but only for one track ) and the resulting yield modification is shown in fig . [
fig : singlebeforeafter ] . at intermediate @xmath11 , in particular ,
the proton yields are reduced substantially due to proton misidentification .
non-@xmath59 events contribute once again to the pion and kaon distributions but not as much to protons , where predominantly @xmath88 processes at high @xmath11 ( @xmath89 ) and @xmath73 decays at low @xmath11 ( maximally @xmath90 ) are the dominant backgrounds .
all acceptance corrections are only weakly dependent on hadron type and show the same moderate ( substantial ) correction factors at small and intermediate ( high ) @xmath11 ; the high-@xmath11 correction is again dominated by the event preselection efficiencies .
weak decays originate predominantly from charm decays and so are a very small contribution ( @xmath91% ) for protons .
the various correction steps for single - pions , kaons and protons are summarized in fig .
[ fig : singlebeforeafter ] .
the isr correction here is similar to that in the di - hadron analysis . to clarify the correction for the previous single - pion and kaon results @xcite
, we show in fig .
[ fig : singleisr ] the isr and non - isr fractions for single pions and kaons as well as protons . as in the di - hadron analysis ,
the fraction of events with an actual cms energy below 99.5% of the nominal energy is below 30% and decreases with increasing @xmath11 .
( color online ) single - pion ( black circles ) , -kaon ( blue squares ) and -proton ( green triangles ) cross sections from top to bottom , as a function of @xmath11 .
, scaledwidth=45.0% ] the resulting single - pion , -kaon , and -proton cross sections are displayed in fig .
[ fig : singlehadronxsec ] .
while the pion and kaon results are consistent within uncertainties to those published before , the proton results from belle are shown for the first time .
the results are compared with the aforementioned pythia / jetset fragmentation tunes in fig .
[ fig : singleallratio_mix4_c8_32518153 ] .
as has been noted above and in @xcite , the pythia / jetset settings close to the default settings reproduce the pion and kaon cross sections rather well . for the proton cross sections
, no setting describes the entire @xmath11 range , while the aleph and lep / tevatron tunes roughly agree with the data at low @xmath11 and the older belle mc setting is in moderate agreement at high @xmath11 .
in summary , we present @xmath92 differential cross sections in @xmath25 and @xmath26 for pion - pion , pion - kaon and kaon - kaon pairs of the same and opposite charges and in various topologies .
the general expectations of disfavored fragmentation functions being suppressed , especially at large fractional energies , are confirmed within the assumptions mentioned in this article . in particular , the same - sign pion pairs in _ opposite _ hemispheres fall off more rapidly than the opposite - sign pion pairs .
the ordering with additional strangeness is also as expected when taking into account the favored - kaon fragmentation of strange quarks and charm decays .
for example , where strangeness needs to be created in the fragmentation such as for same - sign kaon pairs and , to a lesser extent , the same - sign pion - kaon pairs , the cross sections decrease even more rapidly as the already disfavored same - sign pion pairs .
the vanishing of the _ same_-hemisphere di - hadron cross sections once the sum of the fractional energies of the two hadrons exceeds unity supports the assumption of the _ same _ hemisphere di - hadrons being produced predominantly via single - parton di - hadron fragmentation .
this , in turn , bolsters the interpretation of the _ opposite _ hemisphere hadron pairs as arising from the fragmentation of different partons . as a consequence ,
the inclusion of the _ opposite _ hemisphere di - hadrons in terms of single - hadron fragmentation into a global pqcd fragmentation function analysis should be possible .
the extracted di - hadron cross sections are compared to various pythia mc tunes that were optimized for various other energies and collision systems .
a full optimization at belle energies should be possible based on these results ; nevertheless , both the default pythia fragmentation setting as well as the latest belle fragmentation setting already describe the data reasonably well .
single - proton @xmath4 cross sections differential in @xmath11 are presented in addition to the previously published single - pion and -kaon results .
these are expected to be of use in global analyses of fragmentation functions , including the proton results from babar @xcite .
various pythia tunes are compared with conclusions similar to those in the di - hadron case . however , for the proton cross sections , the agreement is fair at best over the entire @xmath11 range , which suggests room for improvement in the pythia settings to better model the baryon production . with the precision of these measurements and the additional information obtained by the use of di - hadrons , we expect that subsequent global fits to the world data will improve substantially our understanding of fragmentation functions , in particular in terms of the distinction of favored versus disfavored fragmentation .
we thank the kekb group for the excellent operation of the accelerator ; the kek cryogenics group for the efficient operation of the solenoid ; and the kek computer group , the national institute of informatics , and the pnnl / emsl computing group for valuable computing and sinet4 network support .
we acknowledge support from the ministry of education , culture , sports , science , and technology ( mext ) of japan , the japan society for the promotion of science ( jsps ) , and the tau - lepton physics research center of nagoya university ; the australian research council and the australian department of industry , innovation , science and research ; austrian science fund under grant no .
p 22742-n16 and p 26794-n20 ; the national natural science foundation of china under contracts no . 10575109 , no . 10775142 , no .
10875115 , no .
11175187 , and no .
11475187 ; the chinese academy of science center for excellence in particle physics ; the ministry of education , youth and sports of the czech republic under contract no .
lg14034 ; the carl zeiss foundation , the deutsche forschungsgemeinschaft and the volkswagenstiftung ; the department of science and technology of india ; the istituto nazionale di fisica nucleare of italy ; national research foundation ( nrf ) of korea grants no . 2011 - 0029457 , no . 2012 - 0008143 , no . 2012r1a1a2008330 , no .
2013r1a1a3007772 , no .
2014r1a2a2a01005286 , no . 2014r1a2a2a01002734 , no .
2014r1a1a2006456 ; the basic research lab program under nrf grant no . krf-2011 - 0020333 , no .
krf-2011 - 0021196 , center for korean j - parc users , no .
nrf-2013k1a3a7a06056592 ; the brain korea 21-plus program and the global science experimental data hub center of the korea institute of science and technology information ; the polish ministry of science and higher education and the national science center ; the ministry of education and science of the russian federation and the russian foundation for basic research ; the slovenian research agency ; the basque foundation for science ( ikerbasque ) and the euskal herriko unibertsitatea ( upv / ehu ) under program ufi 11/55 ( spain ) ; the swiss national science foundation ; the national science council and the ministry of education of taiwan ; and the u.s.department of energy and the national science foundation .
this work is supported by a grant - in - aid from mext for science research in a priority area ( `` new development of flavor physics '' ) and from jsps for creative scientific research ( `` evolution of tau - lepton physics '' ) .
d. buskulic _ et al . _
[ aleph collaboration ] , z. phys .
c * 66 * , 355 ( 1995 ) .
r. akers _ et al .
_ [ opal collaboration ] , z. phys .
c * 63 * , 181 ( 1994 ) .
p. abreu _ et al .
_ [ delphi collaboration ] , nucl .
b * 444 * , 3 ( 1995 ) .
k. abe _ et al . _
[ sld collaboration ] , phys .
d * 59 * , 052001 ( 1999 ) [ hep - ex/9805029 ] .
m. althoff _ et al . _
[ tasso collaboration ] , z. phys .
c * 17 * , 5 ( 1983 ) .
h. aihara _ et al .
_ [ tpc / two gamma collaboration ] , phys .
* 52 * , 577 ( 1984 ) .
h. schellman _ et al .
_ , phys .
d * 31 * , 3013 ( 1985 ) .
m. derrick _ et al .
_ , phys .
d * 35 * , 2639 ( 1987 ) .
w. braunschweig _ et al . _
[ tasso collaboration ] , z. phys .
c * 42 * , 189 ( 1989 ) .
h. aihara _ et al . _
[ tpc / two gamma collaboration ] , phys . rev .
lett . *
61 * , 1263 ( 1988 ) .
r. itoh _ et al .
_ [ topaz collaboration ] , phys .
b * 345 * , 335 ( 1995 ) [ hep - ex/9412015 ] .
b. a. kniehl , g. kramer and b. potter , nucl .
b * 582 * , 514 ( 2000 ) [ hep - ph/0010289 ] .
s. albino , b. a. kniehl and g. kramer , nucl .
b * 725 * , 181 ( 2005 ) [ hep - ph/0502188 ] .
m. hirai , s. kumano , t .- h .
nagai and k. sudoh , phys .
d * 75 * , 094009 ( 2007 ) [ hep - ph/0702250 ] .
b. i. abelev _ et al .
_ [ star collaboration ] , phys .
d * 80 * , 111108 ( 2009 ) [ arxiv:0911.2773 [ hep - ex ] ] ; + l. adamczyk _ et al .
_ [ star collaboration ] , phys .
d * 89 * , 012001 ( 2014 ) [ arxiv:1309.1800 [ nucl - ex ] ] .
g. agakishiev _ et al . _ [ star collaboration ] , phys .
* 108 * , 072302 ( 2012 ) [ arxiv:1110.0579 [ nucl - ex ] ] .
a. adare _ et al .
_ [ phenix collaboration ] , phys .
d * 91 * , 032001 ( 2015 ) [ arxiv:1409.1907 [ hep - ex ] ] .
b. abelev _ et al .
_ [ alice collaboration ] , phys .
b * 717 * , 162 ( 2012 ) [ arxiv:1205.5724 [ hep - ex ] ] .
d. de florian , r. sassot , m. epele , r. j. hernandez - pinto and m. stratmann , [ arxiv:1410.6027 [ hep - ph ] ] ; + d. de florian , r. sassot and m. stratmann , phys .
d * 75 * , 114010 ( 2007 ) [ arxiv hep - ph/0703242 [ hep - ph ] ] ; + d. de florian , r. sassot and m. stratmann , phys .
d * 76 * , 074033 ( 2007 ) [ arxiv:0707.1506 [ hep - ph ] ] .
d. boer , nucl .
b * 806 * , 23 ( 2009 ) [ arxiv:0804.2408 [ hep - ph ] ] .
r. seidl _ et al .
_ [ belle collaboration ] , phys .
lett . * 96 * , 232002 ( 2006 ) [ arxiv hep - ex/0507063 [ hep - ex ] ] ; + r. seidl _ et al .
_ [ belle collaboration ] , phys .
d * 78 * , 032011 ( 2008 ) [ erratum - ibid .
d * 86 * ( 2012 ) 039905 ] [ arxiv:0805.2975 [ hep - ex ] ] .
a. bacchetta , f. a. ceccopieri , a. mukherjee and m. radici , phys .
d * 79 * , 034029 ( 2009 ) [ arxiv:0812.0611 [ hep - ph ] ] .
a. airapetian _ et al .
_ [ hermes collaboration ] , jhep * 0806 * , 017 ( 2008 ) [ arxiv:0803.2367 [ hep - ex ] ] . c. adolph _ et al . _ [ compass collaboration ] , phys .
b * 713 * , 10 ( 2012 ) [ arxiv:1202.6150 [ hep - ex ] ] .
a. vossen _ et al .
_ [ belle collaboration ] , phys . rev .
* 107 * , 072004 ( 2011 ) [ arxiv:1104.2425 [ hep - ex ] ] .
a. courtoy , a. bacchetta , m. radici and a. bianconi , phys .
d * 85 * , 114023 ( 2012 ) [ arxiv:1202.0323 [ hep - ph ] ] .
s. kurokawa and e. kikutani , nucl .
instr . and
a * 499 * , 1 ( 2003 ) , and other papers included in this volume ; + t.abe _ et al . _ ,
phys . * 2013 * , 03a001 ( 2013 ) and references therein .
z. natkaniec _ et al . _
[ belle svd2 group ] , nucl .
instr . and meth .
a * 560 * , 1 ( 2006 ) .
t. sjstrand , p. edn , c. friberg , l. lnnblad , g. miu , s. mrenna and e. norrbin , comp .
commun . * 135 * , 238 ( 2001 ) .
d. j. lange , nucl .
instr . and meth .
a * 462 * , 152 ( 2001 ) .
r. brun , f. bruyant , m. maire , a. c. mcpherson and p. zanarini , cern - dd / ee/84 - 1 1984 .
s. banerjee , b. pietrzyk , j. m. roney and z. wa , phys .
d * 77 * , 054012 ( 2008 ) [ arxiv:0706.3235 [ hep - ph ] ] ; + s. jadach , b. f. l. ward and z. wa , comp .
commun . * 130 * , 260 ( 2000 ) ; + and soo ryu , private communication .
s. jadach , j.h .
khn , z. wa , comp .
commun . * 64 * , 275 ( 1990 ) .
f. a. berends , p. h. daverveldt , r. kleiss , comp .
commun . * 40 * , 285 ( 1986 ) . | we report the first double differential cross sections of two charged pions and kaons ( @xmath0 ) in electron - positron annihilation as a function of the fractional energies of the two hadrons for any charge and hadron combination .
the dependence of these di - hadron cross sections on the topology ( same , opposite - hemisphere or anywhere ) is also studied with the help of the event shape variable thrust and its axis .
the ratios of these di - hadron cross sections for different charges and hadron combinations directly shed light on the contributing fragmentation functions .
for example , we find that the ratio of same - sign pion pairs over opposite - sign pion pairs drops toward higher fractional energies where disfavored fragmentation is expected to be suppressed .
these di - hadron results are obtained from a @xmath1 data sample collected near the @xmath2 resonance with the belle detector at the kekb asymmetric - energy @xmath3 collider . extending the previously published single - pion and single - kaon cross sections , single - proton ( @xmath4 )
cross sections are extracted from a @xmath5 data sub - sample .
quantum chromodynamics ( qcd ) is generally accepted as the theory of the strong interaction .
it describes successfully many high - energy processes where the strong coupling is small and a perturbative treatment is applicable .
however , the non - perturbative region of hadronic bound states such as the nucleon or the transition from high - energetic partons into confined hadrons can not be described quantitatively so far . despite the recent progress in lattice qcd ,
both parton distribution and fragmentation functions ( ff ) remain quantities that need to be obtained experimentally .
fragmentation functions describe the density of final - state hadrons @xmath6 from an initial parton @xmath7 with a fractional energy @xmath8 at a certain energy scale @xmath9 .
single - hadron , unpolarized fragmentation is described by the function @xmath10 .
the first @xmath11 moment , summed over all final states , corresponds to energy conservation in the transition of the initial - state parton into the total final state .
fragmentation functions can not be directly accessed but can be related to observable quantities whenever hadrons appear in the final state .
the most prominent connection to fragmentation functions can be found in single - hadron inclusive cross sections in electron - positron annihilation .
this process provides very clean access as there are no hadrons in the initial state .
these cross sections can then be related at leading order in the strong coupling , @xmath12 , to fragmentation functions via @xmath13 where the scale @xmath14 is given by the center - of - mass ( cms ) energy .
many such measurements have been performed for light hadrons at a range of cms energies at the b factories @xcite at lep and slc @xcite and other facilities @xcite .
the different energy scales can be related via dglap evolution @xcite .
several global analyses of the @xmath15 fragmentation data have been performed @xcite that provide reasonable precision on the sum of quark and anti - quark fragmentation functions .
the precise belle and babar data allowed to reduce the uncertainties on the gluon fragmentation function to pions , which only enters at the next - to - leading order in @xmath12 .
however , analysis of the single - hadron cross sections from @xmath15 data alone can not distinguish nor flavor - separate quark and anti - quark fragmentation and , in particular , favored _
_ vs . _
_ disfavored fragmentation . favored ( disfavored )
fragmentation describes the fragmentation of a parton into a hadron with ( without ) that parton flavor as valence content , such as @xmath16 ( @xmath17 ) . as the parton distribution functions
are generally well known for up- and down - type flavors , semi - inclusive deep - inelastic scattering ( sidis ) @xcite and hadron - collision @xcite results provide leverage in truly global fits to obtain some flavor and charge separated information on pion and kaon fragmentation functions @xcite .
when selecting hadron pairs in @xmath15 annihilation , the cross section can be expressed at leading order in @xmath12 in terms of products of fragmentation functions @xcite : @xmath18 where it is assumed that both hadrons emerge from different quarks and the scale dependence has been dropped for brevity .
this assumption is strictly valid only at leading order @xcite and for hadrons that are nearly back - to - back . in order to study its validity ,
events are analyzed here in three different topologies .
when both hadrons are in the same hemisphere as defined by the thrust axis defined below , it is more likely that they emerge from the same parton so that a di - hadron fragmentation function should describe the process .
if both hadrons are in opposite hemispheres , the assumption of single - hadron fragmentation for each hadron is more likely . in a third sample ,
all hadron pairs irrespective of topology are considered .
the thrust axis @xmath19 maximizes the thrust @xmath20 @xcite : @xmath21 the sum extends over all detected particles , and @xmath22 denotes the momentum of particle @xmath6 in the cms . if the assumption of single - hadron fragmentation holds as described in eq .
, the cross sections are then sensitive to favored and disfavored fragmentation depending on the charges and hadron types of the two detected hadrons . for pairs of oppositely charged pions ,
either both of the hadrons are produced by favored fragmentation or both are produced by disfavored fragmentation ; for same - sign pion pairs , one is produced from favored and one from disfavored fragmentation .
consequently , the cross section for same - sign pion pairs is smaller than that for opposite - sign pion pairs if disfavored fragmentation functions are smaller , especially at high @xmath11 as found in the global fits and expected in models .
the reason for the different @xmath11 dependence originates in the assumption that more quark - antiquark pairs need to be created to arrive at a disfavored hadron , which reduces its large-@xmath11 possibility .
when neglecting strange and charm fragmentation and assuming @xmath23 isospin symmetry , the ratio of same- over opposite - sign pion pair cross sections reduces , for diagonal @xmath24 elements , to a simple expression of disfavored and favored light - quark fragmentation functions .
strange and charm fragmentation dilute this simple relation for pions but in a global analysis all yields and all flavors can be treated appropriately .
this general idea has been formulated in the context of the collins fragmentation function measurements in @xmath15 and applied there @xcite but has already been considered much earlier in ref .
@xcite . in the case that two hadrons are detected in the same hemisphere
, their production is more likely to arise from the same parton so that di - hadron fragmentation functions ( diff ) should describe their yields theoretically .
the formalism for diffs was developed initially in ref .
@xcite and including dglap evolution @xcite as summarized in ref . @xcite .
their polarized counterparts , sometimes denoted as interference fragmentation functions , have been widely used in sidis experiments @xcite and belle @xcite to access together the quark transverse - spin distribution in the nucleon @xcite . in this paper , the individual @xmath11 dependence of the unpolarized baseline diffs is extracted . it should be noted that the leading - order mapping of single versus di - hadron fragmentation to the opposite versus same hemisphere assignments fails at next - to - leading order , where both types of fragmentation must be considered simultaneously @xcite . due to energy conservation , the momentum of the same - hemisphere di - hadrons should not exceed the total initial momentum of the parton if originating from one parton only . the inclusive cross sections for charged di - hadrons in various topologies as a function of their fractional energies @xmath25 and @xmath26 are extracted in this paper .
to evaluate the role of favored and disfavored fragmentation , the ratios between these cross sections for various charge and hadron type combinations are calculated as well .
the contributions for different topologies are compared to better understand single versus di - hadron fragmentation . finally , the cross sections are compared to various monte carlo ( mc ) simulation tunes optimized for different collision systems and energies .
since the corrections are rather similar , a modified version of the di - hadron analysis code is used to extract single - hadron results as a comparison and cross check to the previously published single - hadron cross sections @xcite . as a new result ,
the previously unpublished single - proton cross sections as a function of @xmath11 are presented here and compared to the aforementioned mc tunes .
this paper is organized as follows : after a short description of the detector in section i , the raw di - hadron measurement is described in section ii before detailing the various corrections necessary to arrive at the final cross sections , their ratios as well as topology dependence . the single - hadron analysis , including the new single - proton results , are presented and compared to mc tunes in section iii .
we conclude with a summary in section iv . |
[ 30 ] e. merzbacher , _ quantum mechanics _ , wiley new york : wiley , 1970 pp 12839 .
z. ahmed , s. kumar , m. sharma and v. sharma , eur .
* 37 * ( 2016 ) 045406 .
z. ahmed , s. pavaskar , d. sharma and l. prakash , arxiv : 1508.00661 [ quant - ph ] . t. kato , perturbation theory of linear operators , springer , new york ( 1980 ) .
z. ahmed , phys .
lett . a * 282 * , 343 ( 2001 ) ; 295 * 287*(2001 ) .
z. ahmed , phys .
a * 364 * , 12 ( 2007 ) . z. ahmed , pramana j. phys * 73 * , 323 ( 2009 ) ; f.m .
fernandez , arxiv:1512.09326 .
[ quant - ph ] . | the sub - barrier pairs of energy levels of a hermitian one - dimensional symmetric double - well potential are known to merge into one , if the inter - well distance ( @xmath0 ) is increased slowly .
the energy at which the doublets merge are the ground state eigenvalues of independent ( @xmath1 ) .
we show that if the double - well is perturbed mildly by a complex pt - symmetric potential , the merging of levels turns into the coalescing of two levels at an exceptional point @xmath2 . for @xmath3 ,
the real part of complex - conjugate eigenvalues coincides with @xmath1 again .
this is an interesting and rare connection between the two phenomena in two domains : hermiticity and complex pt - symmetry . in one dimensional quantum mechanics
there is one to one correspondence between eigenvalues and eigenstates , there is an absence of degeneracy .
so when a parameter of the hamiltonian is varied slowly curves can not cross but they can come quite close and then diverge from each other ( avoided crossing ) . crossings and avoided crossings of levels is commonly observed in the spectra of two or three dimensional systems .
mostly , in one dimensional systems if a parameter of the potential is varied slowly , the eigenvalues increase or decrease monotonically . for particle in an infinitely deep well of width @xmath0 , @xmath4 ) decrease as function of @xmath0 .
for harmonic oscillator potential , @xmath5 increase linearly as function of the frequency parameter @xmath6 .
two levels coming very close may either display merging of two levels or their avoided crossing .
the former is well known to occur in symmetric double - well potentials wherein the sub - barrier doublets of energy levels merge [ 1,2 ] into the levels of the independent wells when the inter - well distance is increased slowly . on the other hand ac
is observed rather rarely in one dimensional systems .
recently , it has been shown [ 2,3 ] that in double - well potentials if the width or depth of the potential is varied slowly very interesting level crossings can be observed .
notably the double - well becomes asymmetric .
for one - dimensional non - hermitian hamiltonions , it is known that two complex eigenvalues may become real at one special value of the parameter(@xmath7 of the potential after this point these two eigenvalues may again be complex .
such special values of the parameter are called exceptional point ( ep ) [ 4 ] . more interestingly , when a potential pt - symmetric ( invariant under the joint action of parity : @xmath8 and time - reversal ( @xmath9 ) , the two discrete eigenvalues make a transition from real to complex - conjugate or _
vice versa_. for instance , in the complex pt - symmetric potential : @xmath10 , @xmath11 is the ep of this potential when @xmath12 eigenvalues are real , otherwise these are complex conjugate pairs[5 ] . ) pt - symmetric potentials . here , @xmath13 .
( a ) : double delta , ( b ) : double delta between two rigid walls , ( c ) : square double - well . , title="fig:",width=4,height=5 ] .5 cm ) pt - symmetric potentials . here , @xmath13 .
( a ) : double delta , ( b ) : double delta between two rigid walls , ( c ) : square double - well . , title="fig:",width=4,height=5 ] .5 cm ) pt - symmetric potentials . here , @xmath13 .
( a ) : double delta , ( b ) : double delta between two rigid walls , ( c ) : square double - well . , title="fig:",width=4,height=5 ] in fig .
2(a , b , c ) , we show the parametric evolution of eigenvalues @xmath14 for various pt - symmetric potentials [ 6,7 ] and for their hermitian counterparts .
[ 6 ] between two rigid walls at @xmath15 , ( b ) : @xmath16 [ 7 ] , ( c ) : @xmath17 [ 7 ] . in all three case @xmath18 .
the eigenvalues of their hermitian counterparts ( when @xmath19 is replaced by @xmath20 ) are shown by dashed lines . in ( b ) and ( c ) they exist for @xmath21 and @xmath22 , respectively .
notice the qualitative dis - similarity between two types of spectra .
however , quantitatively for small values of @xmath19 they do coincide.,title="fig:",width=14,height=6 ] + [ 6 ] between two rigid walls at @xmath15 , ( b ) : @xmath16 [ 7 ] , ( c ) : @xmath17 [ 7 ] . in all three case @xmath18 .
the eigenvalues of their hermitian counterparts ( when @xmath19 is replaced by @xmath20 ) are shown by dashed lines . in ( b ) and ( c ) they exist for @xmath21 and @xmath22 , respectively . notice the qualitative dis - similarity between two types of spectra .
however , quantitatively for small values of @xmath19 they do coincide.,title="fig:",width=14,height=6 ] + [ 6 ] between two rigid walls at @xmath15 , ( b ) : @xmath16 [ 7 ] , ( c ) : @xmath17 [ 7 ] . in all three case @xmath18 .
the eigenvalues of their hermitian counterparts ( when @xmath19 is replaced by @xmath20 ) are shown by dashed lines . in ( b ) and ( c ) they exist for @xmath21 and @xmath22 , respectively . notice the qualitative dis - similarity between two types of spectra .
however , quantitatively for small values of @xmath19 they do coincide.,title="fig:",width=14,height=6 ] of the two levels of the dddp when @xmath0 is varied slowly : ( i ) @xmath23 ( dashed line ) , ( ii ) @xmath24 ( solid line ) , ( iii ) @xmath25 ( dotted line ) . dotted and solid lines represent coalescing of eigenvalues in complex pt - symmetric case .
dashed lines display merging of levels in hermitian case .
the eigenvalues are obtained from eq .
( 1),width=14,height=6 ] of the two levels of the double delta well between two rigid walls ( fig 1(b ) ) , when @xmath0 is varied slowly : ( i ) @xmath23 ( dashed line ) , ( ii ) @xmath24 ( solid line ) , ( iii ) @xmath25 ( dotted line ) . dotted and solid lines represent coalescing of eigenvalues in complex pt - symmetric case .
dashed lines display merging of levels in hermitian case .
not shown are the next two levels which are real without any coalescing .
eigenvalues have been obtained form eq .
( 3),width=14,height=6 ] of the six levels of the square double - well ( fig .
1(c ) ) when @xmath26 is varied : ( i ) @xmath27 ( dashed line ) , ( ii ) @xmath28 ( solid line ) , ( iii ) @xmath29 ( dotted line ) , ( iv ) @xmath30 ( dot - dashed line ) . excepting red dashed lines , others represent coalescing of eigenvalues in complex pt - symmetric case .
red dashed lines present merging of levels in hermitian case .
( a ) : @xmath31 for @xmath32 ( b ) : @xmath33 for @xmath32 and @xmath34 for @xmath35 ( c ) : @xmath36 for @xmath32 , @xmath37 for @xmath35 and @xmath31 for @xmath38
. eq .
( 8) has been used here.,title="fig:",width=14 ] + of the six levels of the square double - well ( fig .
1(c ) ) when @xmath26 is varied : ( i ) @xmath27 ( dashed line ) , ( ii ) @xmath28 ( solid line ) , ( iii ) @xmath29 ( dotted line ) , ( iv ) @xmath30 ( dot - dashed line ) . excepting red dashed lines , others represent coalescing of eigenvalues in complex pt - symmetric case .
red dashed lines present merging of levels in hermitian case .
( a ) : @xmath31 for @xmath32 ( b ) : @xmath33 for @xmath32 and @xmath34 for @xmath35 ( c ) : @xmath36 for @xmath32 , @xmath37 for @xmath35 and @xmath31 for @xmath38 .
eq .
( 8) has been used here.,title="fig:",width=14 ] + of the six levels of the square double - well ( fig .
1(c ) ) when @xmath26 is varied : ( i ) @xmath27 ( dashed line ) , ( ii ) @xmath28 ( solid line ) , ( iii ) @xmath29 ( dotted line ) , ( iv ) @xmath30 ( dot - dashed line ) .
excepting red dashed lines , others represent coalescing of eigenvalues in complex pt - symmetric case .
red dashed lines present merging of levels in hermitian case .
( a ) : @xmath31 for @xmath32 ( b ) : @xmath33 for @xmath32 and @xmath34 for @xmath35 ( c ) : @xmath36 for @xmath32 , @xmath37 for @xmath35 and @xmath31 for @xmath38 . eq . ( 8) has been used here.,title="fig:",width=14 ] for double dirac delta potential(dddp , fig .
1(a ) ) the formula for finding the eigenvalues of bound states [ 2 ] can be written as @xmath39 the solution of schrdinger equation for the double dirac - delta well between two rigid walls ( fig .
1(b ) ) is given by @xmath40 .
@xmath41 @xmath42[v_2-p(1+\coth pd)]=e^{-2pb}[v_1+p(1-\coth pd)][v_2+p(1-\coth pd)].\end{aligned}\ ] ] here , @xmath43 the solution of schrdinger equation for the square double - well potential ( fig.1(c ) )
is given as , @xmath44 and @xmath45 where , @xmath46 @xmath47 @xmath48 @xmath49 for bound states , we demand @xmath50 . from eq.(7 ) we have @xmath51 , where @xmath52 .
finally we get , @xmath53 gives the eigenvalues of bound states of double - well in fig 1(c ) .
using eqs .
( 1,3,6 ) , we obtain eigenvalues of the double - well potentials ( fig .
1(a , b , c ) ) for both cases @xmath54 and @xmath55 . in fig 2 , by solid curves ,
we show the evolution of eigenvalues @xmath14 for three pt - symmetric potentials . by dashed curves
, we show the same for hermitian counterparts of these potentials when @xmath19 is replaced by @xmath20 .
notice the coalescing of eigenvalues at special values of @xmath19 , these special values are called eps . also note that at or around the eps , the evolution of eigenvalues in dashed lines does not show any special feature .
so the spectra of hermitian and their complex pt - symmetric counterparts do not relate to each other well , excepting for very small values of @xmath19 where the eigenvalues of both coincide approximately . in fig .
3 - 5 , we present the variation of eigenvalues when the distance between wells in fig .
1(a , b , c ) is increased slowly .
the blue lines present the coalescing of two levels when the total potential is mildly complex pt - symmetric ( @xmath19 is small ) .
see the dashed red lines for hermitian double - well ( @xmath54 ) representing merging of two levels .
green dotted and blue dot - dashed lines arise when non - hermiticity parameter becomes large , then the coalesced levels are not contained between the merging levels ( red dashed lines ) .
lastly , we conclude that the models discussed here bring the spectral phenomena of coalescing and merging of energy levels closer , however we know that they occur in two different domains : hermitian and complex pt - symmetric .
further investigations in this regard are welcome . |
gan - based semiconductors are widely used for optoelectronic devices@xcite and are being developed for high power applications.@xcite typically wurtzite gan films for these applications are grown in the @xmath0 c - plane orientation . while it was discovered early how to grow high quality films in the c - plane orientation , there has been much progress recently in growth on other surface orientations that offer potential advantages . for light
emitting devices , the intrinsic electric field associated with polar c - plane orientations can limit performance,@xcite and use of non - polar orientations such as @xmath1 m - plane can alleviate this problem.@xcite for high power devices , use of a vertical device geometry involving growth on various surface planes can improve performance.@xcite this motivates our effort to model gan growth on different crystal surface orientations such as c- and m - plane to elucidate the influence of orientation on growth modes and kinetics . in this study
, we use kinetic monte carlo ( kmc ) simulations to observe the effects of surface orientation on atomic - scale mechanisms occurring during homoepitaxy of gan films by typical methods , such as metal - organic vapor phase epitaxy ( movpe ) .
several previous kmc studies of gan growth have been carried out for the @xmath0 c - plane surface .
studies modeling gan growth by molecular beam epitaxy @xcite proposed mechanisms by which ga and n diffusion rates and the ga / n supply ratio influence surface morphology . a previous kmc simulation describing movpe @xcite focused on the chemical reaction , adsorption , and desorption processes without considering surface kinetics such as diffusion of atoms , attachment at step edges , etc . a sequence of studies that focus on surface kinetics @xcite have modeled step morphologies and instabilities as a function of growth conditions . the effect of the ehrlich - schwoebel step - edge barrier @xcite on step instabilities @xcite and growth mode transitions @xcite has been studied .
recent kmc studies @xcite have investigated step instabilities on the @xmath2 surface . here
we develop a model to compare the growth of gan on two different surface orientations , @xmath0 c - plane and @xmath1 m - plane , to observe effects of the crystal lattice structure on atomic - scale mechanisms determining homoepitaxial growth mode boundaries as a function of temperature and growth rate .
we analyze the surface structures observed in reciprocal space , to determine growth mode boundaries and mean island spacings , and to make contact with _ in situ _ surface x - ray and electron scattering studies .
such experiments provide quantitative characterization of atomic - scale surface morphology during growth .
in particular , x - ray methods can penetrate the movpe environment to reveal growth behavior as a function of conditions .
we compare our kmc simulation results with x - ray studies of gan movpe @xcite to fix the relationship between simulation and experimental timescales and provide physical insight into observed behavior .
the plan of this paper is as follows . in section
ii , we describe the features of our kmc model , and their direct implications for step edge energies and diffusion barriers on gan c- and m - plane surfaces . in section iii
, we present growth simulations as a function of temperature and growth rate , and analyze the structures in reciprocal space to obtain island spacings and growth mode boundaries . in section iv
, we compare the results with experiments on gan movpe , and in section v discuss results and conclusions . note that the results of the simulations can also be applied to other hexagonal materials by alternative choices of scaling parameters .
diffusion and chemical reactions of precursor species on the gan surface under movpe conditions are not well understood .
indirect estimates of surface transport rates and mechanisms have been made , e.g. based on the observed temperature and length scale dependence of surface smoothing .
@xcite a number of density functional theory calculations for diffusion and reaction energies and barriers of ga and n species including ammonia have been reported , providing insights into molecular mechanisms of the growth processes.@xcite in general , kmc calculations can take advantage of parameter values based on such dft results
. however , in the absence of a complete set of reliable parameters for different gan orientations , we have been using a more generic energy model that we relate to experimental studies .
while a full atomic - scale description of all processes occuring during movpe is challenging , one aspect of gan growth provides a simplification .
because of the very high equilibrium vapor pressure of n@xmath3 at the ga
/ gan phase boundary at typical growth temperatures,@xcite the nitrogen precursor ( e.g. nh@xmath4 ) is typically provided in large excess .
thus the surface is saturated with respect to nitrogen species ( e.g. nh@xmath5 ) in a dynamic steady state environment , @xcite and the rate - limiting steps for growth involve the deposition and incorporation of ga.@xcite in our model we can thus focus on the behavior of ga atoms , and assume the n structure remains in local equilibrium with the environment .
a version of this assumption has been used in previous simulations . @xcite
while considering only the ga sites involves an approximation , it allows us to investigate the anisotropies on various crystal faces due to the primary underlying crystal symmetry . to simulate movpe of gan
, we use a kmc model based on a crystal lattice of ga atomic sites , where each site is either occupied by an atom or not . the ga sites in the wurtzite gan structure form a hexagonal closed - packed ( hcp ) arrangement , with almost the ideal ratio of the @xmath6 and @xmath7 lattice parameters.@xcite
thus we use an ideal hcp lattice of ga sites in the kmc model .
the use of the hcp lattice ( p6@xmath4/mmc symmetry ) instead of the full wurtzite structure ( p6@xmath4mc symmetry ) does not capture some features of the gan structure , e.g. the asymmetry between @xmath8}$ ] and @xmath9}$ ] due to polarity . an hcp lattice with lattice parameter @xmath7
can be described as an orthorhombic lattice using `` orthohexagonal '' coordinates,@xcite with four sites per unit cell and lattice parameters @xmath7 , @xmath10 , and @xmath11 .
[ fig : planes ] shows the geometry of the ga sites on the c - plane and m - plane gan surfaces .
the three faces of the orthohexagonal unit cell correspond to the @xmath12 a - plane , @xmath13 m - plane , and @xmath14 c - plane surfaces , which are normal to @xmath7 , @xmath15 , and @xmath6 in the @xmath16 , @xmath17 , and @xmath18 directions , respectively . as shown in cross section in fig .
[ fig : planes](b ) , for the c - plane surface , sites 1 and 3 form a layer at fractional coordinate @xmath19 in each unit cell , while sites 2 and 4 form a layer at @xmath20 .
we thus consider each of these layers to comprise a single monolayer ( ml ) , and define the thickness of 1 ml to be @xmath21 for the c - plane .
note that in some of the literature , @xcite this definition of monolayer is denoted as `` bilayer '' , because of the nitrogen site associated with each ga site . for the m - plane surface , with cross section shown in fig .
[ fig : planes](a ) , sites 1 and 2 have similar heights @xmath22 and 1/6 , while sites 3 and 4 have similar heights @xmath23 and 2/3 .
we likewise consider each of these layers to comprise a single ml for the m - plane , with a thickness of @xmath24 .
[ cols="^,^,^,^,^,^,^ " , ] as shown in fig .
[ fig : gm ] , the simulations predict a much larger region of lbl growth for m - plane than c - plane surfaces .
this result is consistent with experiments,@xcite in which a direct transition between sf and 3d growth modes is typically observed on c - plane , while a large intervening region of lbl growth is observed on m - plane .
the observed growth rate at the sf - lbl boundary on m - plane gan@xcite can be described by an arrhenius expression @xmath25 with parameters @xmath26 ev and @xmath27 ( ml / s)@xmath28 = 13.1 \pm 1.4 $ ] . by equating @xmath29 in ml / s with @xmath30 from eqs .
[ eq : sf ] and [ eq : t0 ] , we can obtain expressions for the parameters that give the time scale for gan movpe , @xmath31 @xmath32 using the terrace width @xmath33 nm corresponding to the experiments@xcite and the simulation value of @xmath34 , this gives @xmath35 ev and @xmath36 ( s@xmath37 = 18.5 \pm 1.8 $ ] .
values of @xmath38 estimated using these parameters are given in table [ tab : scales ] . also shown are the calculated surface diffusion coefficients.@xcite using the correspondence between experimental and simulation temperature and time scales given in table [ tab : scales ] , the regions of temperature and growth rate investigated in recent experiments@xcite are shown in fig .
[ fig : gm ] .
the boundaries at fixed experimental growth rates of 0.001 and 2 ml / s appear as diagonal lines when plotted in simulation units of ml / ut , because of the temperature dependence of @xmath38 .
except at low temperature , the region investigated directly in the current simulations corresponds to much higher growth rates than studied in experiments .
the three experimental values of @xmath29 used to obtain the correspondence are likewise plotted in simulation units ( blue dots ) on fig .
[ fig : gm](a ) .
the experimentally observed value of @xmath39 at the lbl-3d transition for m - plane@xcite is plotted ( red dot ) in simulation units on fig .
[ fig : gm](a ) .
it lies very close to the boundary obtained from the simulation results ( red line ) .
this represents a quantitative agreement of the relative positions of the sf - lbl and lbl-3d boundaries between the experiments and the simulations using the nnn model for the m - plane surface .
experiments on c - plane surfaces@xcite show a direct transition between sf and 3d growth modes , with no intervening lbl mode .
this qualitatively agrees with the prediction in fig . [
fig : gm](b ) for the region of experimental conditions .
however , the experimentally observed value for @xmath39 ( red dot ) is significantly higher than the boundary predicted by simulations using the nn model for the c - plane surface ( red line ) .
this quantitative difference may indicate that the effective es barriers produced by the nn model are too large to accurately represent the c - plane , and that behavior intermediate between those of the nn and nnn simulations would agree better with experiment .
however , because we have used results from m - plane to determine the correspondence between experimental and simulation time scales , other effects may also contribute to the difference observed for c - plane .
the simulations presented here of growth on the c- and m - plane surfaces of a hexagonal crystal provide both agreement with , and insight into , experimental results for movpe growth of gan .
we see different behavior on the two surfaces because of the different bonding configurations .
the simulations employ two different models for diffusion kinetics : nn , which allows diffusion jumps only between nearest - neighbor sites ; and nnn , which also allows some next - nearest - neighbor jumps .
these produce higher and lower effective ehrlich - schwoebel step - edge barriers , respectively , and have a significant impact on the crystal growth modes .
we have mapped the transitions as a function of temperature @xmath40 and growth rate @xmath41 among the three homoepitaxial growth modes : three - dimensional ( 3d ) ; layer - by - layer ( lbl ) ; and step - flow ( sf ) .
we have also determined the dependence on @xmath40 and @xmath41 of the island spacing @xmath42 that develops during lbl and 3d growth .
quantitative comparison of these results to experiment allows us to estimate underlying fundamental quantities , such as the surface diffusion coefficients given in table [ tab : scales ] . on both the c- and m - plane surfaces , the island heights found in the simulations during lbl growth are one half of the orthohexagonal unit cell dimension ( @xmath21 and @xmath24 , respectively )
this is in agreement with experimental studies of layer - by - layer growth in gan movpe,@xcite indicating that the basic nearest - neighbor bond counting energetics of the kmc model is generally applicable to this system .
the simulations show that both the typical island shapes at lower temperatures and the growth mode transitions differ between c- and m - plane , and between nn and nnn kinetics .
overall , the nn model with high es barrier provides better agreement with experiment for the c - plane , while the nnn model with low es barrier provides better agreement for m - plane .
in particular , the very narrow region of lbl growth found for c - plane with nn kinetics , and the typical shape of islands on m - plane formed under nnn kinetics ( elongated perpendicular to @xmath43 $ ] and highly correlated parallel to @xmath43 $ ] ) , are both in agreement with x - ray@xcite and afm measurements .
this indicates that the elongated islands observed on m - plane have their origin primarily in anisotropic equilibrium step - edge energies rather than in adatom diffusion kinetics , since the latter is almost isotropic in the nnn model .
similar conclusions were reached in studies of islands on dimerized si ( 001 ) @xcite and ag ( 110 ) @xcite .
our study indicates that the direct transition from sf to 3d growth observed for growth on gan c - plane surfaces can be attributed to a high es barrier due to the bonding arrangement at steps on a close - packed surface .
previous kmc studies on c - plane growth also emphasized the effect of the es barrier on island nucleation , finding that smoother films are obtained when the es barrier is screened.@xcite this agrees with our finding that the lbl range is wider on c - plane when using the nnn model .
our simulations found that the island spacing @xmath42 under lbl growth conditions obeys a negative power - law dependence on growth rate , eq .
[ eq : s ] , with exponent @xmath44 for m - plane and @xmath45 for c - plane .
recent experimental results @xcite for movpe on m - plane gan are in agreement with this value .
growth of anisotropic islands in submonolayer epitaxy has been previously considered in kmc modeling .
@xcite the power law for island density as a function of growth rate shows an exponent smaller than that for the isotropic case , with @xmath46 for a critical nucleus size equal to one .
however , we observe a similar low exponent in our simulations on c - plane surfaces , where the island shapes are isotropic . several topics for future work can be identified .
there is considerable interest in gan growth on surface orientations in addition to c- and m - plane,@xcite and it is straightforward to extend the kmc model and methods developed here to other orientations . while we have estimated the boundary between lbl and sf growth based on island spacing , it will be of interest to explicitly consider growth on vicinal surfaces with steps , which can be done by using helical boundary conditions.@xcite for simplicity we have neglected evaporation from the surface in this work .
inclusion of evaporation will provide a second surface transport mechanism qualitatively different from surface diffusion , that undoubtedly becomes important at higher temperatures.@xcite since the kmc model gives an exact arrangement as a function of time for the atomic positions , it will be very valuable in predicting the results of coherent x - ray experiments ( such as x - ray photon correlation spectroscopy @xcite or coherent diffraction imaging @xcite ) that are becoming feasible .
the model also can be extended to account for full the wurtzite structure of gan , with two atom types and additional parameters derived from experiment or ab initio theory .
this will allow study of the effects of polarity .
finally , the nn and nnn models employed here give fairly extreme values for high and low effective es barriers . the best agreement with experiments
may be obtained from a model between these two extremes .
we gratefully acknowledge support provided by the department of energy , office of science , basic energy sciences , scientific user facilities ( kmc model development ) and materials sciences and engineering ( reciprocal space analysis and experiments ) , and computing resources provided on blues and fusion , high - performance computing clusters operated by the laboratory computing resource center at argonne national laboratory . | the surface orientation can have profound effects on the atomic - scale processes of crystal growth , and is essential to such technologies as gan - based light - emitting diodes and high - power electronics .
we investigate the dependence of homoepitaxial growth mechanisms on the surface orientation of a hexagonal crystal using kinetic monte carlo simulations . to model gan metal - organic vapor phase epitaxy ,
in which n species are supplied in excess , only ga atoms on a hexagonal close - packed ( hcp ) lattice are considered .
the results are thus potentially applicable to any hcp material .
growth behaviors on c - plane @xmath0 and m - plane @xmath1 surfaces are compared .
we present a reciprocal space analysis of the surface morphology , which allows extraction of growth mode boundaries and direct comparison with surface x - ray diffraction experiments . for each orientation
we map the boundaries between 3-dimensional , layer - by - layer , and step flow growth modes as a function of temperature and growth rate .
two models for surface diffusion are used , which produce different effective ehrlich - schwoebel step - edge barriers , and different adatom diffusion anisotropies on m - plane surfaces .
simulation results in agreement with observed gan island morphologies and growth mode boundaries are obtained .
these indicate that anisotropy of step edge energy , rather than adatom diffusion , is responsible for the elongated islands observed on m - plane surfaces .
island nucleation spacing obeys a power - law dependence on growth rate , with exponents of -0.24 and -0.29 for m- and c - plane , respectively . |
of late the rare decays of the b mesons have been recognized as important tools to study the basic structure and validity of the standard model ( sm ) and its extensions . in particular
, the radiative decays , owing to their relative cleanliness as far as experimental signatures are concerned , have attracted a great deal of attention . for a general overview of the kind of issues considered relating to radiative decay modes ,
see @xcite and references therein .
the decays @xmath2 and @xmath3 have been observed @xcite and both these are extensively used and relied upon for constrining the parameters of any new theory or extension of the sm @xcite .
the decay @xmath1 is another potential testing ground for the effective quark level hamiltonian @xmath4 first studied by lin , liu and yao @xcite and pursued further in references @xcite,@xcite and @xcite .
this amplitude has been the focus of considerable research recently , not only for the useful indications it will give to the underlying theories of flavour changing neutral currents , or the possible contributions from loops with supersymmetric partner particles , but for the impending experimental studies of the b - factories in the near future .
as has been previously noted , the @xmath4 amplitude naturally splits into two categories : an irreducible contribution which is well known and usually estimated through basic triangle graphs , and a reducible one , where the second photon is attached to the external quark lines of the @xmath5 amplitude . at the quark level the reducible contribution presents no real problems , however , when we consider an exclusive channel , such as @xmath6 for a specific meson @xmath7 , it becomes more appropriate to consider the second photon as arising from the external hadron legs of the amplitude @xmath8 .
in contrast to the earlier cases of @xmath9 @xcite where the amplitude for a single real photon vanishes identically , resulting in the irreducible diagram to be the sole contributor , the amplitude @xmath3 is non vanishing .
however for the neutral decay mode the second photon can not arise from the resulting @xmath10 , and thus in this case also , it is the irreducible amplitude that stands out , though for a completely different reason .
of course we must also consider the usual long distance contributions , such as the process @xmath11 followed by the decay @xmath12 .
note that for completeness we will also include the @xmath13 contribution , even though the @xmath13 coupling to @xmath14 will be small .
the rate for @xmath15 is anomalously high and many possible mechanisms have been proposed that aim at taking this anomalous production into account@xcite .
however , in the present case there is not enough data corresponding to the @xmath16 channel , and at present only an upper limit on this branching ratio is available .
we therefore tend to remain conservative in the present study regarding this issue and assume that the @xmath17 contribution can be obtained similar to the @xmath18 contribution .
the situation is expected to improve with the availability of more and precise data in this direction .
we therefore include an @xmath17 contribution along the lines of the @xmath18 contribution . in this paper we will estimate the branching ratio for the process @xmath19 by considering the effects of the irreducible triangle diagram contributions in the next section , followed by the resonance contributions in section 3 .
note that in the case of @xmath1 there will only be three sizeable resonance contributions ; @xmath18 , @xmath13 and @xmath20 .
furthermore , each of these contributions will only contribute a narrow peak in the @xmath21 invariant mass spectrum , which is easily separated experimentally .
as such the interference terms for each of these pairs of terms will not be considered here . finally in section 4 we will present our results and analysis .
the irreducible triangle contributions to the process in which we are interested ( @xmath1 ) originate from the quark level process @xmath4 .
the effective hamiltonian for this process is @xcite @xmath22 with @xmath23 the invariant amplitude corresponding to this effective hamiltonian is @xmath24 \bar{u}(p_s ) \left\ { \sum_q a_q j(m_q^2 ) \gamma^{\rho } p_l r_{\mu\nu\rho } \right .
\\ & & + ib\left ( m_s k(m_s^2 ) p_l + m_b k(m_b^2 ) p_r \right ) t_{\mu\nu } \nonumber \\ & & \left .
+ c \left ( -m_s l(m_s^2 ) p_l + m_b l(m_b^2 ) p_r \right ) \epsilon_{\mu\nu\alpha\beta } k_1^{\alpha } k_2^{\beta } \right\ } u(p_b ) \epsilon^{\mu}(k_1 ) \epsilon^{\nu}(k_2 ) , \nonumber\end{aligned}\ ] ] where @xmath25 , & \nonumber\end{aligned}\ ] ] and @xmath26 in the above expressions we introduced the functions @xmath27 where @xmath28 note that to get the @xmath29 invariant amplitude from the quark level amplitude we replace the @xmath30 by @xmath31 for any dirac bilinear @xmath32 . with @xmath33 and following cheng _
@xcite , we parameterize the hadronic matrix elements as @xmath34 , \nonumber\end{aligned}\ ] ] for the functional dependence of various form factors appearing above , we follow @xcite . using these definitions , we determine the irreducible matrix element for the process @xmath35 as @xmath36\ ] ] where @xmath37 \right . \nonumber \\ { \cal m}^{(2)}_{\mu\nu } & = & k_b \left ( k_{1 \mu } k_{2 \nu } - ( k_1.k_2 ) g_{\mu\nu } \right ) ( \epsilon^*_{k^ * } .
{ \cal m}^{(3)}_{\mu\nu } & = & k_c \epsilon_{\mu\nu\alpha\beta } k_1^{\alpha } k_2^{\beta } ( \epsilon^*_{k^ * } .
p_b ) \nonumber\end{aligned}\ ] ] the functions @xmath38 above are defined as @xmath39 \frac{v(q^2)}{m_b + m_{k^ * } } & ; & k_{a2 } = \left [ \sum_q a_q j(m_q^2 ) \right ] \frac{i}{2 } ( m_b + m_{k^ * } ) a_1(q^2 ) \nonumber \\ k_{a3 } = \left [ \sum_q a_q j(m_q^2 ) \right ] \frac{i}{2 } \frac{a_1(q^2)}{m_b + m_{k^ * } } & ; & k_{a4 } = \left [ \sum_q a_q j(m_q^2 ) \right ] \frac{i m_{k^*}^2}{q^2 } \left ( a_3(q^2 ) -
a_0(q^2 ) \right ) \nonumber\end{aligned}\ ] ] and @xmath40 \nonumber \\ k_c & = & - \frac{i c m_{k^*}}{m_b + m_{k^ * } } a_0(q^2 ) \left[m_s l(m_s^2 ) + m_b l(m_b^2 ) \right ] .
\nonumber\end{aligned}\ ] ]
for this process there will be three significant resonance contributions , that from the @xmath18- , @xmath13- and @xmath20-resonances . the @xmath18 contribution to
the decay process comes via the @xmath41-channel decay @xmath42 , with the @xmath18 then decaying into two photons .
the t - matrix element for this process can be written as @xmath43 the amplitude @xmath44 is parameterized as @xcite @xmath45 note that we can determine @xmath46 from the known decay rate : @xmath47 where we have @xmath48 and so @xmath49 the @xmath42 amplitude has been determined in cheng _
@xcite as @xmath50 where @xmath51 note that @xmath52 is defined as @xmath53 . in the above way of parametrizing the @xmath42 ,
there is a lot of model dependence that goe in . since , the branching fractions for this sub - process is known , we can in principle avoid such a model dependnce by writing the amplitude as @xmath54 and determine the effective constant from the corresponding decay rate .
we folow this procedure and therefore try to avoid any model dependence as far as possible .
therefore the total contribution due to the @xmath18 resonances is thus , @xmath55 analogous to the @xmath18 resonance , the @xmath13- and @xmath20-resonance contributions , @xmath56 and @xmath57 , have exactly the same form as equation ( [ etares ] ) with the parameters @xmath46 , @xmath58 and @xmath59 being replaced by their @xmath13- and @xmath20-counterparts respectively .
however in this case we can not define the relative sign of the amplitudes and any of the other components of the amplitude .
the process can also receive additional contribution from the @xmath60 and @xmath61 channels where the b meson decays into a photon and an on - shell @xmath61 or slightly off - shell @xmath60 and then these giving rise to @xmath10 and the second photon .
however , there is no data available at present for either of these and if the widths for the individual channels contributing to the process are significant , the contribution can be sizeable .
however , we expect that these contributions can be eliminated by suitable cuts in the @xmath60- ( or @xmath61- ) photon plane and thus we do not consider them here at all .
the squared amplitude for the process @xmath62 is then ; @xmath63 where the interference terms have not been included here .
the components to the squared amplitude were calculated to be ; @xmath64 where @xmath65 , \nonumber \\ \mathrm{and } \hspace{10pt } \lambda ( s_{\gamma\gamma } , m_{b}^2 , m_{k^*}^2 ) & = & 4 \left ( ( p_b.p_{k^*})^2 - m_{k^*}^2m_b^2 \right ) .\end{aligned}\ ] ] we have similar expressions for the @xmath13 and @xmath20 terms , replacing the parameters @xmath46 , @xmath58 and @xmath59 by their @xmath13 and @xmath20 counterparts .
the irreducible squared matrix element is then ; @xmath66 ^ 2 \frac{q^4}{2 m_{k^*}^2 } \left ( ( p_b.p_{k^*})^2 - m_{k^*}^2m_b^2 \right ) \\ & & \hspace{20pt } \times \left [ \begin{array}{c } \left| k_{a2 } \right|^2 + \left| k'_{a3 } \right|^2 + \left| k'_{a4 } \right|^2 + \left| k_{b } \right|^2 + \left| k_{c } \right|^2 \\ + 2 \mathrm{re}(k_{a5 } ) + 2 \mathrm{re}(k'_{a3 } k^ { * \prime } _ { a4 } ) \\ - 2 \mathrm{re}(k_{a2 } k^ { * \prime}_{a3 } ) - 2 \mathrm{re}(k_{a2 } k^ { * \prime}_{a4 } ) \end{array } \right ] \nonumber\end{aligned}\ ] ] where @xmath67 i m_{k^ * } a_0 ( q^2)\end{aligned}\ ] ] the total decay rate is then given by @xmath68^{1/2 } \sum_{spins } \left| { \cal m } \right|^2 , \ ] ] where @xmath69 is the c.m .
energy of the two photons while @xmath70 is the angle which the decaying b - meson makes with the two photons in the @xmath71 c.m .
our results are presented in figures 1 and 2 ( both plotted with a logarithmic scale on the y - axis ) , where figure 1 shows the differential branching ratio given as a function of the invariant mass of the two photons , for the case of the neutral b meson decay without the interference terms between the resonances and the irreducible background included .
the inclusion of the interference terms will in principle give rise to an interference pattern near the base of the resonance peaks .
since the peaks are narrow and moreover the interference contributions being small ( except the @xmath18-irreducible term ) , we do not include them in the plots .
it is worth mentioning that imposition of suitable cuts in the spectrum to eliminate the resonance contributions will eliminate any such interference patterns also .
the numerical estimate for the branching ratio arising due to all possible contributions is summarized in table 1 .
quite evidently , the largest contribution comes from the @xmath13 resonance mode .
it should be stressed again that using appropriate cuts in the spectrum , the resonances can be completely eliminated and what is left is the background irreducible contribution . in estimating the numerical values for the interference terms
, we have assumed that the relative signs between the terms are such that the @xmath13-irreducible and @xmath17-irreducible interfering contributions add on to the other pieces .
however , because of the smallness of these values , it really makes no significant difference .
contribution & branching ratio @xmath72 + + resonance & + @xmath18 & 4.7 + @xmath13 & 56.9 + @xmath17 & 3.7 + + irreducible & @xmath73 + + interference & + @xmath18-irreducible & 2.6 + @xmath13-irreducible & @xmath74 + @xmath17-irreducible & @xmath75 + + br & 57.5 + in figure 2 , we compare the irreducible contribution with the total contribution to the branching fraction .
clearly , the resonances dominate the results . at this point
, it may be worth mentioning that a quick look at the individual values tabulated in table1 reveal the following .
since the resonances are narrow , one may try to estimate the contributions directly by multiplying the individual branching ratios ie we expect in the narrow width approximation that @xmath76 where @xmath77 denotes any of the resonances .
the numbers quoted clearly show that they are in accord with the expectations .
however , if we had used the form as in eq(17 ) , we might have over- or under - estimated the resonance contributions ( except probably for the @xmath13 mode ) because it is not very clear if such a simple parametrization is the correct one .
we however avoid any such possible conflict by drawing heavily on the experimental values ( upper limit for @xmath20 ) of the various sub - process branching fractions .
the central values of the parameters used in our calculation are shown in the appendix .
we make an attempt to estimate the errors creeping into the numerical calculations due to errors in various input parameters .
the theoretical uncertainties arising out of uncertainties in the parameters are overwhelmingly in the input values of the form factors and the meson decay constants .
the wilson coefficients have been taken to be their nnl values and there are no significant theoretical uncertainties in them . in evaluating the uncertainties of our results ,
it is appropriate to evaluate them separately for the background irreducible contribution and the resonance contributions , since the latter can easily be experimentally separated from the former by suitable cuts in the spectrum . for the model dependent parametrization of eq(17 ) , the theoretical uncertainty in the resonance contribution due to @xmath18 arises mostly because of uncertainties in the ckm parameters , @xmath52 , @xmath46 and @xmath78 ; the wilson coefficients values used are the nnl level values and no comparable uncertainties exist therein .
the ckm parameters relevant to us have an uncertainty of about 10@xmath79 @xcite .
the form factors used are the same as in @xcite where the actual dependence of the form factors as a function of momentum transfer squared are given and hence no errors arise due to parametrization of form factors as a function of @xmath80 .
although this reference does not quote any estimate of the uncertainties in the numbers , a typical uncertainty in this type of calculation based on quark model is given in @xcite and is typically of the order of 15@xmath79 , arising to a great extent due to uncertainty in the strange quark mass .
the rate of the @xmath18 decay into two photons is uncertain by about 40@xmath79 @xcite . a typical estimate of the uncertainty in the value of @xmath52 ( arising mostly again out of uncertainty in the current mass of the s - quark ) has been estimated at about 15@xmath79 in @xcite . combining all this
, we would expect that an estimate of the contribution of the @xmath18 resonance based on such a parametrization to be uncertain by about 50@xmath79 . a similar estimate for the other two resonaces give a somewhat lower value mostly because their decay rates into two photons is better known , to an accuracy of about 10 @xmath79 for @xmath17 and about 5 @xmath79 for the @xmath13 .
the uncertainties in the decay constansts of the @xmath81 and the @xmath13 have been estimated to be about 10@xmath79 @xcite and we estimate the overall uncertainty in our calculation for the @xmath17 and @xmath13 to be about 40@xmath79 and 30@xmath79 respectively .
however , since we have relied on experimental values of rates and branching fractions , the above mentioned uncertainties are significantly reduced and the only source of uncertainty in our estimation is the uncertainty present in sub - process rates . turning to the irreducible contribution ,
the uncertainty arises mostly because of the ckm factors and the form factors .
these combine to give an overall uncertainty of about 20@xmath79 for the irreducible part of the amplitude .
as stressed before , once the suitable cuts are imposed in the two photon spectrum , it is possible to extract the irreducible contribution and here the errors are relatively smaller and are expected to even go down further with more accurate determination of the ckm parameters and the form factors . at the levels reached by the current b - factories , the branching ratios obtained are too low to be observed .
one certainly hopes that in the near future experiments , with better luminosities possible , the numbers obtained will be very useful for confronting theoretical models with experimental data .
as discussed in the text , this decay with two photons depends on the parts of the effective hamiltonian , which the decays with a single photon are not sensitive to and thus provides a more complete test of the underlying theory .
1.5 cm _ * acknowledgements * _ src would like to acknowledge the department of science and technology , government of india for a research grant . + we would like to thank prof .
g.hiller for a communication pointing out the discrepancies between their results and our earlier estimates for the irreducible contribution .
we list the central values of the various parameters entering our numerical estimates : + @xmath82 @xmath83 @xmath84 @xmath85 @xmath86 @xmath87 @xmath88 @xmath89 @xmath90 1 b. grinstein , r. springer and m. wise , nucl .
b339 * , 269 ( 1990 ) ; m. misiak , nucl . phys . * b393 * , 23 ( 1993 ) ; m. ciuchini , e. franco , g. martinelli , l. reina and l. silvestrini , nucl . phys . *
b421 * , 41 ( 1994 ) ; k. adel and y. p. yao , phys . rev . *
d49 * , 4945 ( 1994 ) ; c. greub and t. hurth , phys . rev . *
d56 * , 2934 ( 1997 ) ; a. ali and c. greub , phys .
* b259 * , 182 ( 1991 ) ; k. g. chetyrkin , m. misiak and m. munz , phys . lett . * b400 * , 206 ( 1997 ) ; a. j. buras , m. misiak , m. munz and s. pokorski , nucl . phys . *
b424 * , 374 ( 1994 ) ; a. f. falk , m. luke and m. j. savage , phys . rev . *
d49 * , 3367 ( 1994 ) ; a. khodjamirian , r. ruckl , g. stoll and d. wyler , phys . lett . * b402 * , 167 ( 1997 ) .
cleo collaboration , m. s. alam _ et al _ , phys .
lett . * 74 * , 2885 ( 1995 ) ; cleo collaboration , r. ammar _ et al _ , _ ibid _ * 71 * , 674 ( 1993 ) .
d. pirjol , hep - ph/0207095 ; e. o. iltan , new . j. phys . * 4 * , 54 ( 2002 ) ; k. agashe , n. g. deshpande and g. h. wu , phys .
b514 * , 309 ( 2001 ) ; m. carena , d. garcia , u. nierste and c. e. m. wagner , phys . lett . *
b499 * , 141 ( 2001 ) ; d. chakraverty and d. choudhury , phys . rev . *
d63 * , 075009 ( 2001 ) ; r. barbieri and g. f. giudice , phys . lett . *
b309 * , 86 ( 1993 ) ; j. l. lopez , d. v. nanopoulos , x. wang and a. zichichi , phys .
* d51 * , 147 ( 1995 ) .
g.l.lin , j.liu and y.p.yao , phys.rev.lett .
* 64 * , 1498 ( 1990 ) ; g.l.lin , j.liu and y.p.yao , phys.rev .
* d42 * , 2314 ( 1990 ) .
c.v.chang , g.lin and y.p.yao , phys.lett . *
b415 * , 395 ( 1997 ) .
l.reina , g.riccardi and a.soni , phys.lett . *
b396 * , 231 ( 1997 ) ; phys.rev . *
d56 * , 5805 ( 1997 ) .
s.r.choudhury , g.c.joshi , n.mahajan and b.h.j.mckellar , phys . rev .
* d67 * , 074016 ( 2003 ) ; s.r.choudhury , g.c.joshi , n.mahajan and b.h.j.mckellar , hep - ph/0306293 .
m. r. ahmady , e. kou and a. sugamoto , phys . rev .
* d57 * , 1997 ( 1998 ) ; d. atwood and a. soni , phys . lett . *
b405 * , 150 ( 1997 ) ; i. halprin and a. zhitnisky , phys .
lett . * 80 * , 438 ( 1998 ) ; i. halprin and a. zhitnisky , phys . rev . *
d56 * , 7247 ( 1997 ) ; f. yuan and k. t. chao , phys . rev .
* d56 * , 2495 ( 1997 ) ; w. -s .
hou and b. tseng , phys .
* 80 * , 434 ( 1998 ) .
h.y.cheng , c.y.cheung and c.w.hwang , phys.rev . *
d55 * , 1559 ( 1997 ) .
a. ali , p. ball , l. t. handoko and g. hiller , phys . rev . *
d61 * , 074024 ( 2000 ) .
k. hagiwara _ etal .
_ , particle data group , phys . rev . * d66 * , 010001 ( 2002 ) . m. beneke and m. neubert , hep - ph/0308039 . n. g. deshpande and j. trampetic , phys
b339 * , 270 ( 1994 ) .
chen , h - y .
cheng , b. tsing and k - c .
yang , phys .
rev . * d60 * , 094014 ( 1999 ) .
g. hiller and a. safir , hep - ph/0411344 . | we study the decay of the neutral b meson to @xmath0 within the framework of the standard model , including long distance contributions . * short and long distance contributions to @xmath1 * + + s. r. choudhury + _ department of physics , delhi university , delhi 110007 , india _ , + + a. s. cornell + _ korea institute of advanced study , 207 - 43 cheongryangri 2-dong , _ + _ dongdaemun - gu , seoul 130 - 722 , korea _ , + + namit mahajan + _ harish - chandra research institute , _ + _ chhatnag road , jhunsi , allahabad 211019 , india_. + + - we have corrected a sign error in the numerical program .
the new estimates agree well with the ones given in a recent paper @xcite + |
atomic scale imaging techniques have proven to be useful tools to improve the understanding of atomic processes at surfaces .
prominent examples for studies at metals are the field ion microscopy ( fim ) studies of self diffusion at transition metal surfaces@xcite , the electron microscope studies of step structures and growth on halide , semiconductor , and metal surfaces@xcite , and the scanning tunneling microscopy ( stm ) studies of the epitaxial growth and sputter removal of pt(111).@xcite several observations made in those studies like the reentrant layer - by - layer growth at low temperatures or the temperature variation of the growth form of islands at higher temperatures are not fully understood yet . for other growth phenomena like
the fractal island shape at low temperatures@xcite a satisfying explanation does exist .
understanding the origins for different forms of growth bears technological relevance .
typically it is desirable to have layer - by - layer growth in order to produce high quality films and to achieve this at not too high temperatures.@xcite how a surface develops during growth is a consequence of the microscopic adatom - surface interaction , especially at binding sites and at transition states of surface diffusion.@xcite if the rates for all relevant diffusion processes are known , the evolution of the surface during growth can be calculated.@xcite because of the computational effort required for a quantum - mechanical description of the microscopic interaction , several quasi - classical methods have been used in the past.@xcite however the reliability of these calculations have always been questionable , especially because neither the influence of the kinetic energy operator for the electrons nor self - consistent rearrangements of the electron density were taken into account properly .
the kinetic energy of the electrons largely determines the nature of the chemical bond by splitting the electronic energies into bonding and antibonding levels , or by influencing the charge distribution at metal surfaces , which is governed by the spill out of density into the vacuum and by the reduction of the charge density corrugation ( smoluchowski smoothing @xcite ) .
all quantum - mechanical effects which are considered to be relevant for chemisorption are taken into account in density - functional theory ( dft ) to a high level of accuracy when used together with the local - density approximation ( lda ) of the exchange - correlation functional .
@xcite in this paper we report a rather extensive set of dft - lda calculations of adsorption and diffusion of al adatoms on different surfaces of fcc aluminum.@xcite we even include comprehensively the diffusion at steps in our study of al(111 ) .
the role of steps in determining the growth morphology is known since long time.@xcite we chose al because it is a prototype of a simple @xmath2-@xmath3 metal , hoping that the interpretation of any observation would be particular clear and provides insights which are transferable to other systems . besides the flat ( 111 ) , ( 100 ) , and ( 110 ) surfaces , also the two different close - packed steps on al(111 ) are considered .
these steps are called @xmath4 and @xmath5 according to the step s orientation , which is the @xmath6 direction , and the steepest micro - facet at their edges ( see fig . [
stepscheme]).@xcite the influence of steps is of paramount importance for the description of growth processes.@xcite in particular we like to understand the experimentally established differences between these two sorts of steps on ( 111 ) surfaces of fcc metals .
their slightly different geometry leads to different formation energies,@xcite to different diffusion mechanisms and energy barriers,@xcite and they are found to have different dipole moments.@xcite using the calculated diffusion barriers and estimated diffusion prefactors we estimated the temperature ranges for different growth modes on al(111 ) . to get a detailed description of the temperature dependence of the growth of the al(111 ) surface one should use our results on surface diffusion as input for a theory@xcite which solves for the rate equations which determine the evolving surface morphology during growth
this still remains to be done .
the paper is organized as follows .
first we give a short description of our _ ab - initio _ method and describe some technical aspects which make it particularly efficient for the calculation of large metallic systems .
section [ differ ] describes differences in the formation energy of the two sorts of close - packed steps on al(111 ) . in section [ inddip ]
we discuss the adatom- and step - induced dipole moment on al(111 ) and , connected to that , the work function differences between al(111 ) , al(100 ) , and al(110 ) . in section [ al111 ] the surface self - diffusion is investigated , first on the flat al(111 ) surface , then approaching a step , and finally at the step . also vacancy diffusion on the flat al(111 ) surface is considered .
we compare the self - diffusion on al(110 ) with that on the stepped al(111 ) surface and we compare diffusion at the two different steps on al(111 ) . in section [ unif ]
we describe some uniformity in surface self - diffusion on fcc metals , which help us to estimate diffusion prefactors for the self - diffusion on al surfaces . using the calculated diffusion barriers and the estimated prefactors , we summarize in section [ atom ] our understanding of the temperature dependence of atomic transport processes and of homoepitaxial growth on al(111 ) . in the appendix we also present some results for the self - diffusion on al(100 ) .
the computer code , fhi93cp , used in this study is described in ref . .
we therefore summarize here only some essentials of the method .
we use density - functional theory and treat the exchange - correlation functional in the local - density approximation.@xcite the kohn - sham equations@xcite are solved by a car - parrinello - like iterative scheme @xcite using the steepest descent approach @xcite for the update of the wavefunctions .
we use a fully separable _
ab - initio _
pseudopotential @xcite for al where the @xmath7-potential is treated as local and @xmath2 and @xmath3 potentials are described by projection operators .
the electronic wavefunctions are expanded in a plane - wave basis set with a kinetic energy cut - off of 8ry .
the brillouin zone is sampled at special * k*-points,@xcite which are specified below for the different systems when these are discussed . in the following
we describe in more detail the _ damped newton dynamics _ procedure to relax atoms , the fermi surface smoothing technique , and some technical improvements , which allow us to calculate large systems .
we also describe the geometry of the systems investigated and estimate the numerical accuracy of our results . in our adsorption calculations
we typically allow the al adsorbate and the top two ( 111 ) layers or three ( 110 ) layers to relax until all force components are smaller in magnitude than 0.04ev / .
it was checked that relaxation of an additional layer leaves the adsorption - energy differences practically unchanged .
the most important effect of the adsorbate - induced substrate relaxation is a reduction of barriers for bridge diffusion by [email protected] .
the algorithm to relax the atomic geometry is based on _ damped newton dynamics_. in a finite - difference form the time evolution of any atomic coordinate @xmath9 is given by @xmath10 where @xmath11 is the coordinate at time step @xmath12 and @xmath13 the force on @xmath9 at time step @xmath12 .
the parameters @xmath14 and @xmath15 control the damping and the mass of the coordinate .
the choice of those parameters is guided by the goal that this classical dynamics combines a fast movement of the atoms towards the next local minimum of the born - oppenheimer surface and avoids oscillations around it .
we got fast convergence for our al surface systems with @xmath16 and @xmath17 .
this choice brings the calculations close to the aperiodic limit of a damped oscillator in classical mechanics .
increasing the damping coefficient @xmath14 improves the stability of the atomic relaxation process , reducing it allows for energy barriers to be overcome and so to escape from local minima . in it
s usage of the knowledge of the history of displacements the damped dynamics technique is similar to the conjugate - gradient technique.@xcite obviously the atomic geometry converges faster when larger displacements per time step are executed .
the magnitude of useful displacements is restricted , however , by the efficiency with which the electronic wavefunctions converge to the electronic ground state of the new atomic coordinates after the displacement .
we found it advantageous to have about eight purely electronic iterations after any atomic displacement .
note that the time - consuming calculation of the atomic forces is not done in those purely electronic iterations .
for all systems studied in this paper about ten atomic relaxations are necessary to converge to the desired accuracy . in order to stabilize the self - consistent calculations for the electrons and to improve * k*-space integration , we smear out the fermi surface . for this purpose
the kohn - sham eigenstates of energy @xmath18 are occupied according to a fermi distribution @xmath19 with @xmath20ev . using
the fermi distribution implies that the free energy @xmath21 at the electronic temperature @xmath22 is minimized instead of the total energy @xmath23.@xcite here @xmath24 is the entropy of independent electrons,@xcite @xmath25 \;.\ ] ] this approach may cause some inaccuracies , as we want to get results belonging to @xmath26 . for the free energy at a given geometry the @xmath27 limit can be easily obtained by evaluation of @xmath28.@xcite this value differs from @xmath29 only by terms which are third and higher order in @xmath22 . for the optimization of the geometry the force @xmath30
should be used which is , however , more complicated to evaluate.@xcite for our choice of @xmath20ev the geometries and the total - energy differences are almost not affected .
this was tested for the adsorption of al on al surfaces by using values of 0.05ev and 0.2ev for @xmath31 and an increased number of * k*-points .
a further approach to stabilize the way self - consistency is achieved is to reduce electron transfer between single - particle states in successive iterations . for this purpose
fictitious eigenvalues after pederson and jackson @xcite are introduced .
the occupation numbers are calculated directly from the fictitious eigenvalues according to fermi occupation at @xmath22 .
these fictitious eigenvalues follow the as - calculated eigenvalues in a sort of damped dynamics , so that both sets of eigenvalues will become identical when self - consistency is attained .
this indirect approach of damping charge transfer oscillations is easier to implement than the more obvious one of damping the change in occupation numbers directly .
the reason is that the occupation numbers are constrained to be in the range between 0 and 2 , and their sum has to give the total number of electrons .
for the eigenvalues no such constraints exist .
the computer code is optimized for large systems by using the following concepts . because the evaluation of the non - local part of the pseudopotential dominates the computation time for large systems we utilize the translational symmetry for those atoms sitting on ideal lattice sites .
@xcite without introducing any approximation this optimization typically reduces the number of operations by a factor of ten for these atoms .
one often encountered problem with large systems is the @xmath32 dependence of the electrostatic potential . here
@xmath33 is a reciprocal - lattice vector .
this dependence leads to long - wavelength charge - density oscillations , known as _
charge sloshing _ or _
@xmath32 instability_. see for example ref .. we deal with this problem by starting with a rather good initial density constructed by a superposition of contracted atomic charge densities .
the contraction was done following finnis,@xcite where the radial atomic densities are multiplied by a fermi function .
the contraction anticipates most of the intra - atomic charge transfer that occurs upon building a solid from isolated atoms .
the wavefunctions for the first step of the self - consistent iterations are obtained by diagonalizing of the kohn - sham hamiltonian constructed from this approximate density and within a reduced plane - wave basis ( @xmath34ry , depending on time and memory constraints ) .
then , in the first @xmath35 electronic iterations , the charge density @xmath36 is linearly mixed like in `` standard '' self - consistent calculations : @xmath37 we use a mixing coefficient @xmath38 which increases from 10% to 100% within these first eight time steps . by this procedure
the charge sloshing was not initiated for the systems considered in this paper and so did not cause problems during the rest of the calculations . in calculations of larger cells than those reported here we found the linear mixing in * r*-space insufficient .
the charge density sloshing could however be efficiently suppressed by a mixing in * g*-space with a mixing coefficient @xmath39 which was chosen smaller for smaller @xmath40 . for large systems the required computer memory raises as the square of the number of atoms and
it is largely determined by the number of wavefunction coefficients .
we optimize memory usage in several ways .
a simple steepest - descent update procedure for the wavefunctions is used@xcite so that only the wavefunction coefficients of one iteration need to be stored .
the wavefunction coefficients and most of the other large arrays are stored in single precision ; however , double precision is used for all floating point operations and for storing intermediate results .
our computer code optimizes the data access in computers which use memory of different speed .
the idea is that once data is transferred from slow memory to fast memory , e.g. from the disk to main memory , this data should be used as often as possible before it is moved back to the disk .
this improvement can be accomplished by reordering loops or by blocking techniques.@xcite the most important case where blocking is used is the orthogonalization of the wavefunctions . instead of orthogonalizing always one wavefunction to those with lower index ( the standard gram - schmidt procedure ) , we orthogonalize a block of , say , 30 wavefunctions to those with lower index and then orthogonalize the wavefunctions within the block . in the case where only part of the wavefunctions at one * k*-point fit into main memory this procedure reduces the disk to memory data transfer by a factor up to 30 .
an example of the efficiency of the code is the calculation of an al slab with 350 atom per unit cell in a supercell as large as 560 atomic volumes , and sampling the brillouin - zone at one special * k*-point .
this leads to 28000 plane waves , to 560 involved electronic states , and an overall memory requirement of 200 mb .
this calculation takes about 25h on an ibm/6000 370 risc workstation with 64 mb main memory , if all atoms sit on ideal lattice positions .
if no atoms are sitting on ideal sites , the time increases by a factor of three .
the time spent waiting for disk access during the calculation is below 30% and could even be reduced noticeably by using the faster hard disks available today . in order to describe an adatom on a crystal surface we use a slab in a supercell .
the repeating slabs are isolated by @xmath41 of vacuum spacing . for the purpose of studying practically isolated adsorbates ,
the distance between adatoms in neighboring cells is at least three nearest - neighbor spacings .
this results in an interaction energy below 0.03ev for al adatoms . in order to have more bulk like layers and to avoid artificial adsorbate - adsorbate interaction through the slab , we adsorb the al adatom only on one side .
this approach reduces the slab thickness necessary for the desired degree of accuracy .
@xcite due to the unsymmetrical situation an artificial electric field perpendicular to the slab might arise .
this field is compensated in our calculations as described in ref .
, by introducing a dipole layer in the vacuum region . in the case of an al adsorbate on an al surface
this field is always very small so that even in the uncompensated case the energy differences between different sites are practically unaffected . for the calculation of adsorption on al(111 )
we use typically normally 5-layers - slabs .
calculations with slabs of and 7 layers show that even with a 4 layer slab adsorption energy _
differences _ are accurately given , which means that they change by less than 0.03ev when thicker slabs are used .
for al(100 ) we find that the desired accuracy of 0.03ev requires a slab thickness of at least 6 layers .
the quantity most sensitive to the slab thickness is the energy barrier for exchange diffusion ; for a 5 layer slab this is by 0.25ev or 66% lower than that of the 6- and 7-layers thick slabs ( see appendix [ al100 ] ) .
we use a 4@xmath424 surface cell for the calculations of self - diffusion on al(100 ) . for the al(110 ) surface we used 8 layers and a 3@xmath424 surface cell . in this paper
we treat the two densely packed steps on al(111 ) .
one is called \{111}-faceted , the other \{100}-faceted ( see fig . [ stepscheme ] ) .
the \{111 } and \{100 } micro - facets are the steepest ones and therefore give an unambiguous way of naming the steps.@xcite we shall see , however , that the electronic properties of the \{111}-faceted step are more closely related to the ( 110)-surface .
this similarity was already discovered by nelson and feibelman , @xcite who showed that the relaxation of the al(110 ) surface atoms and the atoms close to the @xmath43\{111 } step is very similar and that the charge densities at the al(110 ) surface and along the \{110 } plane trough a @xmath43\{111 } step are nearly identical .
our calculations confirm this and also show the similarity of chemical properties and of self - diffusion on the al(110 ) surface and at the \{111}-faceted step . in order to analyze the step properties we use three different approaches . in the first approach a slab in ( 111 ) orientation
is put in an orthorhombic supercell
. then half of the atoms of the top layer are removed , so that the remaining grooved surface has two steps , one being \{111}- and the other \{100}-faceted ( see fig . [ stepscheme ] ; the lower part of the supercell , containing the flat termination of the slab and some more vacuum , is omitted in the upper panel of fig .
[ stepscheme ] ) .
we choose different sizes of the surface supercell to study the influence of finite size effects .
the width of the cell in @xmath44 $ ] direction is varied from three to four atoms and the width of the terrace in @xmath45 $ ] direction is three to four atomic rows .
all these systems give results which differ only by @xmath46ev .
one reason for this is the rapid screening of al .
the other reason is that quantum size - effects are often unimportant for total - energy differences on stepped al(111 ) which was already noted by nelson and feibelman .
@xcite the second approach uses a class of systems which are slabs of @xmath47 and of @xmath48 orientation where @xmath49 is an even number .
surfaces with these miller indices are vicinal to the ( 111 ) surface . the @xmath47 surface consist of terraces of ( 111 ) orientation which are @xmath49 atomic rows wide and separated by @xmath50-faceted steps .
the @xmath48 surface has ( 111 ) terraces @xmath51 atomic rows wide , which are separated by \{100}-faceted steps.@xcite the relationship between the miller indices of the vicinal surfaces and the constituent low index facets can be seen by the vector decompositions @xmath52 and @xmath53 .
note that conventionally common factors are removed from miller indices , so that instead of ( 200 ) [ which is the shortest reciprocal lattice vector in the ( 100 ) direction ] the more familiar ( 100 ) is used and for even @xmath49 the common factor two is removed . thus the miller indices are @xmath54 and @xmath55 .
we worked with surfaces belonging to even @xmath49 because they can be accommodated in a monoclinic supercell , whereas for odd @xmath49 a triclinic supercell is required .
the first surface ( i.e. @xmath56 ) of the @xmath47 family is the ( 220)@xmath57(110 ) surface , for which the ( 111 ) terraces are so narrow that no surface atom has a ( 111)-like coordination . after some test calculations with the ( 221 ) surface ( @xmath58 )
, we concentrated on the ( 332 ) surface ( @xmath59 ) for studying the properties of nearly isolated \{111}-faceted steps .
the ( 332 ) surface has ( 111 ) terraces which are six atomic rows wide ( see fig .
[ systeme1 ] ) .
we used a 1@xmath424 surface unit cell , which means that 6@xmath424 atoms are exposed at the surface .
this layer is repeated six times to build a slab containing 144 atoms per cell . for studying the properties of nearly isolated \{100}-faceted steps we used the ( 433 ) surface out of the @xmath48 family , which contains seven atomic rows of ( 111)-orientation .
with a 1@xmath424 surface unit cell , we get 7@xmath424 atoms exposed on each surface ( see fig .
[ systeme2 ] ) .
again , 6 layers were taken which gives a slab containing 168 atoms per cell .
the adsorption calculations at the steps of the ( 221 ) , ( 332 ) , and ( 433 ) surfaces essentially reproduce the results of the grooved surfaces , which reflects the efficient screening at al surfaces .
this agreement of results obtained for different slabs also provides a good test of the convergence of our calculations with respect to system size and * k*-space sampling .
the main advantage of using the vicinal surface systems is that they allow the investigation of long - range adsorbate - step interactions , which were found for example experimentally by wang and ehrlich;@xcite for a given adsorbate - step distance the number of atoms in the cell is only slightly more than half of that required for the grooved slab geometry
. for the purpose of calculating the difference in step formation energy of \{111}- and \{100}-faceted steps a third kind of system has to be used .
it should contain only one kind of step and it should provide a favorable error cancellation when the energy _
difference _ of the \{111}- and \{100}-step systems is determined .
the chosen system consists of a triangular al island on al(111 ) with either \{111}- or \{100}-faceted edges , depending on the triangle s orientation .
note that this way the number of atoms and the * k*-space integration is the same for both step orientations . one new uncertainty in determining the step formation energy difference is now introduced however : also the corners of the triangles contribute to the difference .
the contribution of the corners has to be removed by extrapolating to the thermodynamic limit of infinite island size .
the largest triangles studied consisted of 21 atoms in a @xmath60 al(111 ) surface cell . at a slab thickness of 4 layers
this gives 245 atoms per supercell .
the equilibrium lattice constant for a basis set with 8ry cut - off and a * k*-point sampling like that used for the surface calculations is 3.98 ( including zero - point vibrations),@xcite which is 0.7% smaller than the experimental value .
the cohesive energy is 4.15ev , which is 0.75ev higher than the experimental one of 3.40ev .
@xcite these errors in bulk results are within the expectations for a well - converged dft - lda calculation . the 8ry plane wave cut - off was tested to be sufficient to converge adsorption energy _
differences _ to a numerical accuracy better than 0.02ev ( see also ref . ) . because of the large size of our supercells they comprise , depending on the problem , 140560 atomic volumes one special * k*-point @xcite in the irreducible quarter of the surface brillouin zone , which is always nearly square , is sufficient to give energy differences which are within 0.03ev of those obtained by using two or four times the number of * k*-points .
we also tested the dependence of our results on system size . here the slab thickness as well as the adsorbate - adsorbate and the step - step interactions are relevant .
system size and * k*-space - sampling effects are difficult to separate , because often a change of the size of the system implies different * k*-sampling , and the two effects are about equal in magnitude .
we therefore can not quantify the error introduced by system size effects separately . in order to reduce errors from these two sources
we always quote the mean value of calculations at different * k*-point sampling and system size .
this improves the accuracy because the variations with system size and * k*-space sampling are often oscillatory .
we obtain an overall numerical accuracy of the energy differences given of @xmath61ev , unless a different error margin is stated explicitly .
calculations of the average energy of the two step types can be performed most accurately by investigating the total energy of surfaces with terrace stripes ( see fig . [ stepscheme ] ) , and comparing these results to those of the flat surface .
the mean step formation energy is determined as 0.24ev per step atom .
table [ survacform ] shows that this is about half of the energy required to create the al(111 ) surface per surface atom , and that the step formation energy compares to the difference of the surface energies between al(111 ) and the rougher ( 100 ) or ( 110 ) surfaces per atom .
the energy difference of the two step types can be obtained by investigating triangular islands adsorbed on al(111 ) , as these contain only one type of steps ( see fig.[triangle ] ) . comparing islands with 6 , 10 , 15 , and 21 atoms we can extrapolate to the limit where the influence of the corner atoms is negligible .
table [ triadat ] lists the results for the total - energy differences of two triangles whose orientations differ by @xmath62 , and hence have different step types .
the data show the rapid convergence of this energy difference with island size .
dividing these energy differences into an island size independent contribution from the three corner atoms and a contribution proportional to the number of true edge atoms a fit to our data ( see table [ triadat ] ) gives the following results . triangular islands with \{111}-faceted steps
are more favorable by 0.025 ev per corner and by 0.017 ev per true step atom than islands bounded by \{100}-faceted steps .
the energy differences are almost the same , whether the island atoms are relaxed or not .
this small effect of relaxation shows that the step formation energy difference is an electronic effect and is not determined by a different step - induced atomic relaxation .
it is interesting that our results can not be estimated from simple embedded atom or effective medium theory@xcite or a bond - cutting model.@xcite the reason is that the two different triangular islands have exactly the same number of bonds . from the step formation energy one
can directly obtain the equilibrium shape of larger islands by minimizing the island free energy for a given number of atoms in the island .
the graphical solution of this optimization problem is called the wulff construction.@xcite as a result one obtains hexagonally - shaped islands , where the edges alternate between those with a \{100 } and those with a \{111 } micro - facet .
the calculations predict that in thermodynamic equilibrium the \{111}-faceted edge should be longer with a edge length ratio @xmath63 of @xmath64 .
effects of the vibrational or the configurational entropy on the length ratio , which might be important at higher temperatures , are neglected .
it is interesting to note that such hexagonal islands have been observed experimentally by michely and comsa @xcite in their stm studies of growth and sputter removal of pt(111 ) .
these experiments show that the \{111 } micro - facet is favored which is what we predict for al(111 ) .
there is a quantitative difference , since for pt(111 ) the measured edge - length ratio is 0.66 , i.e. @xmath65 .
the similarity to our results is much more than what one would have expected , as , in general , al and pt behave quite differently .
our calculations of induced surface dipole moments and work function differences at al surfaces give unexpected results .
the first interesting observation is that the al(111 ) surface and the al(110 ) surface have about the same work function @xmath66 .
our best converged results are @xmath67ev,@xcite the experiment finds @xmath68ev and @xmath69ev.@xcite the second result is that steps on al(111 ) do not affect the work function very much .
table [ tabdip ] lists the induced dipole moment @xmath70 per step atom .
the \{111}-faceted step induces practically no dipole ( @xmath70=-0.01debye ) , the \{100}-faceted step has a small dipole moment with the negative end pointing into the vacuum ( @xmath70=0.045debye ) , which means that they increase the work function .
induced dipole moments translate into work function changes @xmath71 according to the helmholtz equation @xmath72 with @xmath70 in debye , @xmath71 in ev , and the area @xmath73 per dipole in bohr@xmath74 .
more noticeable dipole moments are found for 3-fold coordinated al adatoms ( see table [ tabdip ] ) an example is the dipole moment of 0.24debye for an al adatom on the hcp - site . if there was an adlayer of those al adatoms on al(111 ) of , say , 1/10 monolayer coverage , the work function would increase by 0.13ev .
the reported results on induced dipole moments and work function differences contradict the traditional model of charge redistribution at rough metal surfaces and around protrusions on metal surfaces like steps or adatoms.@xcite this model is based on smoluchowski smoothing .
smoluchowski smoothing is driven by the kinetic energy of the electrons which is lower for a less corrugated charge density .
the smoothing of the charge density lowers the work function for rougher surfaces and causes a induced dipole moment with the positive end towards the vacuum for protrusions on the surface .
the smoothing effect is often discussed in a nearly free electron picture .
an example is the calculation by ishida and liebsch@xcite of the induced dipole moment of steps on jellium .
they find that steps reduce the work function .
extrapolating their results for a step on al(111 ) one gets an induced dipole moment of about @xmath75debye per step atom equivalent .
the reason why the smoothing model fails in our calculation of atomic al could be that the smoothing effect , which is certainly there , is ( over- ) compensated by the attraction of electrons towards the deeper potential around surface atoms on al(110 ) , step - edge atoms on al(111 ) , or adatoms on al(111 ) .
these atoms are only 7- or 3-fold coordinated as compared to the 9-fold coordinated surface atoms .
therefore they are less well screened , which leads to the deeper potential in their vicinity .
this effects a net transfer of electrons towards those undercoordinated atoms .
a more thorough discussion of the charge rearrangements in different surface geometries and the resulting induced dipole moments will be published elsewhere.@xcite having a possible explanation why the standard model fails in the case of the simple metal al the remaining puzzle is why it seems to work for the transition metals.@xcite to give an example .
steps on au(111 ) and pt(111 ) show dipole moments between 0.25 debye ( au ) and 0.6 debye ( pt ) per step atom.@xcite a comparison with the jellium calculations in ref .
shows that for those steps the induced dipole moment is larger in magnitude than it would be expected from the smoothing effect of the @xmath2-@xmath3 like electrons only .
the additional negative dipole moment is likely caused by a polarization of the @xmath7-electrons of the step atoms .
this would also explain why au shows a smaller effect than pt .
au has a filled @xmath7-shell , in pt the fermi level cuts the @xmath7-band and therefore it is easier to polarize the @xmath7-states .
we see that for the simple as well as for the transition metals important additions to the smoothing based model for induced surface dipole moments and work function differences have to be made .
this section describes the total - energy surface for an al adsorbate atom on the flat al(111 ) surface and at the \{100}- and \{111}-faceted steps .
this discussion is directly relevant for surface diffusion and crystal growth on al(111 ) . we will study how an al adatom moves across the al(111 ) surface , what happens when the adatom comes close to a step , how it attaches to the step coming from the lower side , and how it incorporates into the step by an atomic replacement process coming from the upper side .
the diffusion energy barriers of an isolated al adatoms on the flat al(111 ) surface is very small ( 0.04ev ; see table [ d111 ] and figs.[diff111s ] and [ diff100s ] ) .
the hcp site is the stable binding site and the energies of bridge and fcc sites are almost degenerate .
the diffusion path between the hcp sites is the direct connection between adjacent hcp , bridge , and fcc sites .
the only marked maximum of the total - energy surface is at the atop site .
the top site is 0.53ev higher in energy than the hcp site .
interesting about the atop site is the low height of the adatom at the atop site which is close to the height in the 3-fold sites ( see table [ d111 ] ) .
this is a consequence of the fact that bond length gets smaller when the coordination is lower ( for the atop site we obtain a bond length of 2.51 which is 6% smaller than for the 3-fold sites ) .
furthermore , we find that the adatom at the atop site introduces a strong substrate relaxation : the substrate atom below the adsorbate is lowered by 0.4 .
a similar substrate relaxation was found in calculations for alkali - metal adsorbates on al(111).@xcite the height of the diffusion energy barrier of 0.04ev is at the lower limit of an energy barrier one would expect for this close - packed surface from experimental data ( see table [ tab1 ] ) .
two features of the total - energy surface are worth discussing further .
we first note that it is unexpected that the hcp site is the equilibrium site for low coverage .
the hcp and fcc sites both provide threefold coordination , but only the fcc site continues the abcabc stacking of the fcc crystal , whereas the hcp site belongs to an abcac stacking .
a simple continuation of the fcc structure would suggest that the fcc site should be occupied .
in fact , the fcc - hcp energy difference is coverage dependent . above 1/4 ml coverage
the fcc site gets more stable . to create a full monolayer of al at the hcp position costs 0.05ev per surface atom
compared to the fcc stacking .
it is interesting that this energy is exactly the mean of the formation energy of the three bulk stacking faults in the @xmath76-direction , which were calculated by hammer et al.@xcite the reason for the different adsorption energy at the hcp and fcc sites at low coverage is not obvious .
half of the difference exists already before the al(111 ) substrate is relaxed , which shows that there is an electronic contribution .
another hint for an understanding might be that the al adsorbate at the fcc site has a larger induced dipole moment than at the hcp site ( see table [ tabdip ] ) .
the second interesting feature of al on al(111 ) is that the three - fold coordinated fcc sites and the two - fold coordinated bridge sites have practically the same total energy .
this is not explainable in a coordination number model.@xcite one might guess that the rather favorable energy of the bridge site is a result of the substrate relaxation .
indeed , on the unrelaxed substrate the bridge site is energetically less favorable ( but only by 0.07ev ) than the fcc site .
if the atomic geometry were not relaxed and all atoms are kept at bulk nearest neighbor spacings , then the fcc site were favored by 0.13ev over the bridge site .
still , this is much less than what a coordination number model would predict .
the energy difference between hcp , bridge , and fcc sites is very small .
it is therefore important to check , if the calculations are sufficiently accurate .
we therefore performed several test calculations , varying carefully all parameters which affect the accuracy .
we used coverages from 1/12 to 1/56 ml , increased the number of * k*-points from 1 to 4 and to 9 , we used from 4 to 7 al(111 ) layers , and we increased the plane wave cut - off .
the energy differences of different sites were very stable and the order of fcc and hcp site never reversed at low coverage .
we expect the difference between hcp and fcc site to be accurate to within 0.02ev .
the relative accuracy for the bridge site is slightly worse .
feibelman s recent lda calculations@xcite found the hcp site more stable than the fcc site by 0.03ev , in good agreement with our results . in order to get an idea of the nucleation probability and the influence of adsorbate - adsorbate interaction on the mobility we calculated the adsorbate - adsorbate interaction between two al adatoms sitting at neighboring hcp - sites .
the energy gain with respect to isolated adatoms is 0.58ev .
we note , however , that only one configuration ( hcp - hcp ) was considered and we did not perform any of our usual checks , so that here an error of @xmath80.2ev might be possible .
table [ d111 ] and figs .
[ diff111s ] and [ diff100s ] show that the al adatom is attracted by the step on the lower as well as on the upper terrace and that this is similar for both sorts of steps .
this attraction leads to an energy gain of about 0.1ev at the threefold sites directly at the upper step edge compared to the flat al(111 ) surface .
the long - range attraction towards the step as experienced by atoms on the lower terrace is weaker , but at closer distance the attraction gets as strong that the last two threefold sites before the fivefold at - step site are not local minima any more and an adsorbate there will be funneled towards the step directly .
this attraction and the funneling of al adsorbates by the lower step edge resembles the behavior found recently by wang and ehrlich for ir on ir(111).@xcite also on pt(111 ) indications of a funneling towards steps exist.@xcite finding the same behavior for different metals suggests that the medium - range to long - range attractive interaction is a more general phenomenon .
therefore , an understanding of the nature of the mechanism behind this attraction of adatoms towards steps is desirable .
we approach the analysis of its origin by considering the following thinkable mechanisms and check whether they are indeed attractive or not .
the mechanisms are dipole - dipole interaction , elastic interaction , and interaction of defect induced surface - states .
the analysis shows that the mechanism is not an attractive dipole - dipole interaction .
the dipole - dipole interaction in the studied cases will be very weak .
even for the largest dipole moments ( those for adatoms at the fcc site and for the \{100 } step , see table [ tabdip ] ) the interaction energy would be below 1mev for distances
larger than one nearest - neighbor spacing .
furthermore we note that the interaction would be _
another possible mechanism responsible for the attractive adatom - step interaction could be the elastic interaction of the adsorbate - induced and the step - induced relaxation fields .
this possibility can easily be checked by a calculation of the energy while keeping the substrate atoms at their ideal positions .
our calculations show that the long - range interaction remains unchanged .
thus , elastic effects can be disregarded as the origin of the attractive interaction between the adatom and the step .
the fact that the adatom - step attraction is not of elastic origin and that it is long range excludes the possibility that it can be reproduced with more simple bonding models like coordination number models@xcite or effective medium and embedded atom.@xcite in consequence there remains only the possibility that the attractive interaction is caused by an interaction of adatom - induced and step - induced surface states or screening charge densities .
this conclusion is supported by recent stm pictures of adsorbate and step induced surface states on cu(111 ) and au(111).@xcite figure [ kugeln ] shows five geometries which we consider to be particularly important for diffusion on the al(110 ) surface and at the \{111}-faceted step on the al(111 ) surface . in the following we will compare the two systems and we will analyze the factors determining the adsorption energies in different bonding geometries .
the results of table [ d110 ] show that at geometries having the same nearest - neighbor environment , the adsorption energies are very similar on the ( 110 ) surface and at the \{111}-faceted step . at the two fivefold sites ( fig .
[ kugeln](a ) ) the adsorption energy is rather large and practically identical in both cases .
the calculated al bulk cohesive energy is only 7% or 0.27ev larger .
the adsorption energy on the threefold equilibrium site on al(111 ) at low coverage ( the hcp site ) is as much as 20% or 0.78ev smaller ( table [ d111 ] ) .
a coordination number model@xcite would give too small an estimate of the adsorption energy .
the high at - step adsorption energy we understand as follows . at both fivefold sites
the adsorbates sit in a valley which should be filled with additional electronic charge due to smoluchowski - smoothing.@xcite additionally the adsorbates sit very close to the surface . on the ( 110 ) surface
the height at the fivefold site is 1.33 relative to the relaxed clean surface .
this should be compared to the hcp site on al(111 ) , where the height of the adsorbate is 2.09 .
the combination of increased charge density and low height provides a good embedding of the adsorbate.@xcite the twofold coordinated bridge sites are possible saddle point configurations for surface self - diffusion .
the energies of the comparable bridge sites on al(110 ) and at the \{111}-faceted step are similar . however , the two different bridges ( see figs.[kugeln](b ) and [ kugeln](c ) ) have a markedly different energy .
this energy difference gives rise to a difference in barrier height of a factor of two ( see table [ d110 ] ) .
what is the reason for this difference of sites with the same coordination ?
our explanation is the same as for the high adsorption energy at the 5-fold site .
the higher adsorption energy for the long bridge is accompanied by a lower distance to the surface . for the long bridge on the ( 110 ) surface
the adatom height is only 1.58 , whereas the height is 2.16 for the short bridge .
another effect should be that smoluchowski - smoothing fills the valleys with electron density taken from the upper part of the rows on the clean surface .
for the short - bridge position , this would reduce the charge density around the adsorbate , while it would rather increase the embedding charge density for the long bridge position . having seen the striking similarity of adsorption energies at the ( 110 ) surface and at the \{111}-faceted step ,
it is no surprise that a system which lies between the two cases , namely the ( 331 ) surface , shows a very similar behavior.@xcite for this system feibelman obtained energy differences between the al adsorption at 2-fold and 5-fold sites which are very close to ours ( see table [ d110 ] ) .
we expect that the agreement with our results would be within 0.05ev if feibelman would have included the adsorbate - induced relaxation of the substrate .
the agreement of both studies is a most demanding test for the numerical accuracy of both calculations , since feibelman used a rather different technique .
we will now show that the just discussed bridge sites are not the lowest energy transition states for diffusion for the surfaces considered .
diffusion takes place preferably via exchange mechanisms . investigating the al(110 )
surface we will first show that the exchange mechanism lowers the energy barriers for diffusion , and this holds even in the so called easy direction , i.e. for diffusion in the channels of this surface .
figure [ kugeln](d ) sketches the symmetric configurations of the exchange paths .
for both the ( 110 ) surface and for the \{111}-faceted step , the geometry may be described as two fivefold coordinated al adatoms which are close to a surface vacancy . at the ( 110 ) surface
the two atoms are displaced from the ideal fcc lattice positions by 0.08 towards the vacancy and they are 0.05 closer to the surface than for the `` normal '' fivefold site . at the step
the corresponding displacements are similar .
to get some insight into the nature of bonding of this configuration , we estimate its energy from the adsorption energy of the `` normal '' fivefold site @xmath78 and the calculated vacancy - formation energies , where the al chemical potential , which defines the energy of the removed al atom , is taken as the bulk cohesive energy , i.e. @xmath79ev . for the vacancy formation energy at the ( 110 ) surface
we then obtain @xmath80ev and at the step the result is @xmath81ev ( see table [ survacform ] ) .
the energy barriers are then estimated as @xmath82ev and @xmath83ev , in close agreement with the full calculation of the exchange - parallel processes in table [ d110 ] .
up to now we have only discussed the symmetric configurations of the exchange diffusion path ( see fig.[kugeln](d ) ) without knowing if these are really saddle point configurations .
additional calculations at the step show that they are _ not _ saddle points , but that they correspond to very shallow minima of the total - energy surface .
the saddle points found in these calculations have an energy about 0.04ev higher than the energy of the symmetric configuration .
this energy difference is about 1/10 of the total energy barrier and also very small compared to the numerical accuracy of our calculations .
it is therefore appropriate to discuss the energy for the symmetric exchange configuration when talking about diffusion barriers .
in addition to the discussed mechanisms a third diffusion mechanism for diffusion along the channels was recently proposed by liu et al.@xcite they argue that the configurations of fig.[kugeln](e ) , which are saddle point configurations for the diffusion perpendicular to the channels , are also approximately saddle point configurations for the diffusion parallel to them .
this would imply that the diffusion parallel and the exchange diffusion perpendicular to the rows would have about the same energy barriers .
this idea is interesting , but we get a higher energy barrier for the perpendicular than for the parallel exchange process ( see next paragraph and table [ d110 ] ) .
thus , for al we can not support the idea @xcite but nevertheless , the proposed process could be the explanation for the near identity of parallel and perpendicular to step diffusion energy barriers on ni , ir , and pt ( see table[tab1 ] ) .
[ diff110cross ] we now discuss the exchange perpendicular to the channels of al(110 ) and across the @xmath77 step on al(111 ) ( see fig . [
kugeln](e ) and [ diff111s ] ) .
again the energies of the saddle point configuration on al(110 ) and at the step are similar ( see table [ d110 ] ) . the slightly higher energy of the saddle point geometry at the step can be understood in terms of the higher vacancy formation energy at the step ( see table [ survacform ] ) and of the threefold rather than fourfold coordination of the upper al atom . on al(110 )
the energy of the exchange configuration can be estimated analogously to our discussion in the last section .
the configuration consists of two neighboring fourfold coordinated atoms close to a vacancy .
we want to approximate the energy of the fourfold coordinated atoms by that of another fourfold atom adsorbed at the same height .
there is however some ambiguity in assigning a height to the two adsorbates .
the possibilities are to take the distance of 0.80 to the ( 110 ) surface or that of 2.05 to the \{111 } micro - facets , which are built by forming a surface vacancy on al(110 ) ( compare fig.[kugeln](e ) ) .
the height above the \{111 } micro - facets is more relevant here than that above the ( 110 ) surface because all substrate nearest neighbors of the two adsorbates belong to these \{111 } micro - facets , and we showed already that the nearest neighbor environment determines the bonding to a large extent as a consequence of the good screening of al .
the adsorption energy can therefore be approximated by @xmath84ev of an al adatom on the fourfold site on al(100 ) , which sits 1.69 above the surface ( see table [ diff100dat ] in the appendix ) .
this gives an estimated energy barrier of @xmath85ev , which equals exactly the energy barrier found in the full calculation .
the success of the simple approach used above to assemble the energy for the exchange configurations from the energies of its constituents exemplifies the role of height and local environment for binding on al surfaces .
this also indicates that the nature of the bonding is very similar for equilibrium adsorption sites and in the examined exchange configurations .
the above studies showed that the interaction of an al adatom with the two kinds of close - packed steps on al(111 ) is at larger distances attractive and practically identical . directly at the step however we identified some important differences in adsorption energies and diffusion barriers and mechanisms .
the results given in table [ d110 ] show that the adsorption energies at the fivefold coordinated at - step sites is nearly the same , but that there is a small preference ( 0.03ev ) for the @xmath86 step .
this energy difference is very small ; however it might obey a general rule and is therefore interesting to analyze .
according to nelson et al.,@xcite it can be explained by the different step formation energies ( see table [ survacform ] ) .
adsorbing , e.g. , an al atom at the \{100}-faceted step creates two \{111}-faceted `` micro - steps '' ( the situation is reversed at the \{111 } step ) .
the creation of \{111}-micro - steps should be favorable , because \{111}-faceted steps are favorable . according to this model
, the adsorption energies should differ by @xmath87ev .
indeed , the calculated at step adsorption energy difference is very close to this value .
it is tempting to apply the same type of discussion also to the creation of vacancies within the step . to form
a vacancy at the \{100}-faceted step creates \{111}-micro - steps and should be favorable .
however , the full calculations show that such analysis is not appropriate ; the vacancy formation energy at the \{111}-faceted step is energetically favorable ( see in table [ survacform ] ) .
the results of table [ d110 ] imply that an al adatom diffusing along the \{100 } step will use the normal bridge hopping mechanism and not the exchange mechanism as at the ( 100 ) surface or at the \{111 } step .
the main reason for this difference between the two steps is that the adsorption energy at the long - bridge position at the \{100 } step is 0.19ev higher .
we understand this as an effect of geometry ( see , e.g. , figs . [ diff100s ] and [ kugeln](b ) ) . whereas at the long - bridge position of the \{111 } step the adsorbate is clearly twofold coordinated , the coordination is nearly fourfold at long - bridge position at the \{100 } step .
the exchange - parallel configurations at the \{111 } step and at the \{100 } step have practically the same energies . as a consequence it follows that the diffusion at the lower terrace parallel to the \{100 } step proceeds via hopping mechanism with an energy barrier about 0.1ev lower than that for the exchange diffusion along the \{111 } step .
the different diffusion mechanisms should lead to different diffusion prefactors @xmath88 .
the impact of this difference for the temperature dependence of crystal growth is discussed below in section vi . for diffusion across the \{111 } and the \{100 } step we obtain in both cases exchange diffusion mechanisms with nearly the same energy barrier heights ( see figs . [ diff111s ] and [ diff100s ] ) . for the diffusion across the step in the descending direction the energy barriers are very small ( 0.06ev and 0.08ev ) ; in fact they are only marginally larger than that for the diffusion on the flat al(111 ) surface .
the exchange path at the \{100}-faceted step is geometrically quite different from that at the \{111}-faceted step,@xcite as it has no mirror symmetry perpendicular to the step .
we mapped out a two - dimensional total - energy surface , varying the @xmath89- and @xmath90-coordinates of the involved step atom ( atom 2 in fig .
[ diff100s ] ) on a 4@xmath424 grid while relaxing all the other coordinates of the adsorbates and the two upper substrate layers .
in addition we calculated the energy for 4 points on the apparent diffusion path .
we then checked if all the atomic configurations were smoothly connected along the diffusion path or if some atomic coordinates change drastically between adjacent configurations .
the results show that all coordinates vary smoothly , which confirms that the described path is physically relevant .
it is interesting to compare our results for al surface self - diffusion to those for other fcc metals .
the results in table [ tab1 ] show that the rule of thumb which says that diffusion barriers scale with the cohesive energy is valid .
we also see , that the energy barriers for surface self - diffusion are clearly lowest for the flattest surface , namely ( 111 ) .
all rougher surfaces have diffusion barriers about 5 times higher than the ( 111 ) surface of the same metal .
the diffusion barriers on these rougher surfaces vary by less than a factor of two between different metals .
important for a description of diffusion and its dependence on the local geometry is the fact that in experiment ( and in theory ) the barriers for diffusion in the channels of the ( 110 ) surface and along the \{111}-faceted steps of the ( 331 ) surface have nearly the same barriers .
the only counter example seems to be ir , but there the barriers for diffusion along the step represent only estimates .
we assume that the diffusion mechanism is the exchange ( see fig.[kugeln](d ) ) in both cases for all metals considered . in agreement with our results for al ,
the experiments show smaller energy barriers for diffusion along the \{100 } step than along the \{111 } step .
one might speculate that this difference of the barriers is connected with the change in diffusion mechanism ; we remind that for al we found that the diffusion along the \{100 } step proceeds by a hopping mechanism while it proceeds via exchange along the \{111 } step .
additional insights about the uniformity of diffusion at fcc(111 ) surfaces can be obtained from the diffusion prefactors @xmath88 . using transition - state theory and semi - empirical calculations ,
@xmath88 was evaluated for a series of metal surfaces.@xcite in all cases the prefactor of an exchange process was larger than that of the normal hopping diffusion ( see also table [ tab2 ] for al ) .
furthermore , we see in table [ tab2 ] that the experimentally determined diffusion prefactors are larger in the cases where we expect exchange diffusion .
thus , from the larger diffusion prefactors we conclude that for rh , pt , and ir there is hopping diffusion on the ( 111 ) surface and along the \{100}-faceted step , and exchange diffusion otherwise ( see table [ tab2 ] ) .
the above discussed results for diffusion mechanisms , energy barriers , and diffusion prefactors will now be used for an examination of the temperature dependence of growth of al(111 ) .
the mode of growth is controlled by the interplay of the deposition rate and of the temperature dependent processes of diffusion and defect creation and annihilation . with the information of previous sections at hand we can estimate the most important features of this temperature dependence .
an extended study with our results and employing a monte - carlo technique@xcite to evaluate the relevant rate equations , would be certainly superior .
here we assume that the temperature dependence of the diffusion constant is given by @xmath91 where @xmath92 is the energy barrier of the considered diffusion process ( see table [ d111 ] and [ d110 ] ) .
the pre - exponential factor @xmath93 can be estimated from the adatom attempt frequency @xmath94 and the distance between neighboring adsorption sites , @xmath95 , as @xmath96 .
the factor @xmath97 equals 2 or 4 , depending on the dimensionality of the diffusion process .
@xmath94 can be calculated within transition - state theory from the the total - energy surface.@xcite on the other hand , @xmath93 can also be taken from experimental results for similar diffusion processes ( see table [ tab2 ] ) .
writing @xmath98 and defining the temperature @xmath99 at which a certain diffusion mechanism becomes active as that at which the adatom will jump at least once per second ( @xmath100s ) , it follows that @xmath101 in table [ tab2 ] we give our choice for @xmath88 , based on theoretical results for the al surface diffusion as well as on experimental data for rh , pt , and ir .
this choice is clearly not unambigous ; one has to consider , however , that an error of a factor of 10 in @xmath93 changes the temperature @xmath99 by less than 10% .
the results of the above procedure to determine onset temperatures for diffusion are given in table [ edtd ] .
for temperatures below 320 k the desorption of adatoms from steps is practically irrelevant ( see , e.g. , fig .
[ diff100s ] and table [ edtd ] ) .
thus adatoms captured at a step edge will stay and the island will grow .
the processes important for growth at temperatures below 320 k are therefore the capture of al adatoms at the steps and the diffusion along the steps and their relative rates .
the following approximate analysis of temperature ranges for different growth modes assumes a deposition rate of @xmath102 ml / s .
the further assumption is that for this deposition rate a jump rate of @xmath103s is large enough in order for a deposited al adatom to diffuse to a more stable site before other deposited adatoms meet the first one and form a nucleus .
if the deposition rate is higher than @xmath102 ml / s then the necessary jump rate to get the same growth mode has to be increased proportionally .
this implies that the temperature has to be increased logarithmically ( see eq . [ it d ] ) .
the analysis implies : * our calculations show that an al dimer on al(111 ) is bound by 0.58 ev and is therefore stable at temperatures below @xmath104k .
if the mobility of the dimer is smaller than that of the single adatom it will serve as a nucleus for the growth of the next layer . in that case
three - dimensional growth would occur whenever two adatoms meet on the upper terrace at a substantial rate.@xcite * at temperatures below 25 k adatoms on the upper terrace are hindered by the step - edge energy barrier to move down ( see table [ edtd ] ) .
we note that island edges are frayed at this temperature ( see next item ) , which may reduce the barrier , and adatoms gain energy by adsorption and by approaching the upper step edge , which might lead to some transient mobility.@xcite furthermore we note that on small and/or narrow islands the attempts to overcome the step - edge barrier are more frequent which increases the chance for a success .
we are therefore not convinced that the three - dimensional growth mode really exists for clean al(111 ) . * for 25 k @xmath105 155 k the energy barriers of diffusion parallel to both close - packed steps will prevent diffusion parallel to the steps .
the diffusion at flat parts of the surface is , however , still easy and therefore the attraction of gas - like adatoms towards the steps edges is still active . as a consequence we expect that islands will be formed in a way which may be described as a `` hit - and - stick '' mechanism .
thus , the edges can not equilibrate and fractal - shaped islands with a layer - by - layer growth mode should result . * for @xmath106 k the step edges will be straight , as diffusion along the step is possible , and therefore the islands will be triangular or hexagonal . according to a simple model by michely _
et al._@xcite the different diffusion properties for atoms at the two kinds of step edges might become important for determination of the detailed growth form of the island .
the model says that the growth perpendicular to the step with the lower adatom mobility will advance fastest .
if the growth speed of one step is faster by more than a factor of two than that of the other step the faster one will eventually disappear . as a consequence
the growth shape of the island would be triangular . for al(111 )
our results imply that at low temperature the diffusion parallel to the \{111 } step is hindered more than that parallel to the \{100 } step because of the higher diffusion barrier . however , because the diffusion mechanism is different at these two steps the prefactor for the diffusion parallel to the \{100 } step is smaller ( see table [ tab2 ] ) . as a consequence , at higher temperature the diffusion coefficient for
the movement parallel to the \{100 } step will be smaller than that for the movement parallel to the \{111 } step.@xcite + the additional assumption in ref .
is that the growth speed of steps with larger diffusion barriers along the step is higher .
this is caused by the less efficient atom transport along the step to corners of the growing island .
atoms moving to the corners contribute to the growth of the adjacent step .
+ we therefore predict that the adatom islands during growth at lower temperature have shorter \{111 } edges and at higher temperature they have shorter \{100 } edges . at a medium temperature the growth shape of islands is hexagonal . * there are two mechanisms for vacancy formation we considered on the al(111 ) surface .
the direct creation of vacancies on flat al(111 ) is expected to happen with a rate above 1/s per surface - atom at temperatures above 730k ( see table [ edtd ] ) . in the presence of al - adatoms
the vacancy creation should already take place at 490k , due to the then reduced barrier ( see table [ edtd ] ) .
adatoms can be provided either from deposition or by desorption from the step .
the relatively high energy barriers to create adatoms lets us assume , that the adatom assisted vacancy formation will not be important before the formation of vacancies on the flat al(111 ) sets in at 730k .
+ the second possibility to create vacancies is at steps . at a \{100}-faceted step
they will be created with rates of @xmath107s and step - atom at a temperature of 320k . at a \{111}-faceted step
this temperature is 380k ( see table [ edtd ] ) .
these vacancies could migrate into the terrace and become `` normal '' surface vacancies .
the barrier for vacancy migration is 0.56ev on the flat al(111 ) , so it becomes active at 240k ( see tables [ tab1 ] and [ edtd ] ) .
the onset temperature for vacancy generation from steps therefore should be 320k . vacancy generation and diffusion into the terrace was also observed in stm experiments on pt(111),@xcite which again shows the similarity of atomic processes on pt(111 ) and al(111 ) .
in conclusion , we have presented results of accurate electronic structure and total - energy calculations which reveal several phenomena directly relevant to the description of self - diffusion at al surfaces and to crystal growth .
the three low index surfaces of al are quite different with regard to surface self diffusion .
the diffusion barriers for al adatoms on al(111 ) [ @xmath92=0.04ev ] are much lower than on al(100 ) and al(110 ) [ @xmath92=0.330.62ev ] . for al(100 ) and al(110 ) atomic exchange mechanisms have lower barriers for surface self diffusion than ordinary hopping .
exchange diffusion was found even in the direction parallel to the atomic rows on al(110 ) .
the diffusion of surface vacancies was studied for the al(111 ) surface ( @xmath92=0.56ev ) .
our calculations predict that al adatoms on al(111 ) are attracted towards the edge of close packed steps by a long - range force which most likely originates from a interaction of adatom and step induced surface states .
adatoms close to the lower step edge are funneled towards the step . the diffusion of an al adatom from the upper to the lower terrace proceeds via an exchange of the on - terrace adatom with an in - step - edge substrate atom .
the barrier for this exchange process is rather low which leads to layer - by - layer growth down to very low temperatures .
corresponding results have been observed in field - ion microscopy studies of w and ir adatoms on ir(111 ) by wang and ehrlich.@xcite on al(111 ) the energy barrier for diffusion of an al at - step adatom parallel to the step is much bigger than that perpendicular to the step in the descending direction . in the temperature range where this energy barrier becomes relevant ,
we expect fractal growth .
the mechanism for diffusion along the two kinds of steps is different .
along the \{111}-faceted steps we find an atomic replacement mechanism similar to that for diffusion parallel to the rows on al(110 ) , along the \{100}-faceted steps the hopping mechanism has the lowest energy barrier .
the differences in energy barrier and diffusion prefactor for diffusion along the two kinds of steps can lead to temperature dependent growth forms of islands .
surface systems where similar growth phenomena were observed experimentally are pt on pt(111)@xcite and au on ru(0001).@xcite a comparison of our results for self diffusion on al surfaces and of other theoretical and experimental results on transition metals indicate that atomic exchange processes in surface self diffusion are more common than assumed previously .
we predict that \{111}-faceted steps on al(111 ) are energetically favorable compared to the \{100}-faceted steps .
a similar energy difference was found for pt(111 ) in experiment.@xcite additionally to the energetics at al surfaces we have discussed the surface dipole moments induced by adatoms and steps on the al(111 ) surface and , related to that , the work function differences of the low index surfaces of al .
our results indicate that the commonly used model based on smoluchowski smoothing alone@xcite has to be modified .
we add here our results for the adsorption and diffusion of al on al(100 ) .
our study repeats that of feibelman on the same system.@xcite the main result of feibelman s paper , which is the favorable energy barrier for an exchange diffusion mechanism , has been questioned recently.@xcite we calculated the adsorption energies at the three sites which are important for the discussion of surface diffusion ( see fig .
[ diff100view ] ) .
our results confirm that the exchange diffusion mechanism has a lower barrier than the bridge diffusion mechanism .
the agreement in adsorption energy differences of our results with those of feibelman is again as excellent as in the other examples in this paper .
however , this agreement with feibelman s results is obtained only , if we use the same slab thickness as he did , i.e. , 5-layers ( see table [ diff100dat ] ) .
the agreement would be even better , if we had not relaxed all atoms but only those which feibelman had relaxed .
however , while the numerical accuracy of both calculations agree , it is most interesting to note that our calculations with 6 and 7 layer slabs show a significant change of the energy of the exchange configuration .
this change increases the barrier for diffusion by nearly a factor of three ( see table [ diff100dat ] ) .
a similar sensitivity of calculated energies with slab thickness was not found for any other system and we do not have an explanation for it .
the energy of the exchange configuration also proved to be especially sensitive to changes in the value of the lattice constant and the * k*-point sampling .
some of the calculations presented here were reported previously in m. scheffler , j. neugebauer , and r. stumpf , j. phys .
matter * 5 * , a91 ( 1993 ) , in r. stumpf and m. scheffler , phys.rev.lett.*72 * , 254 ( 1994 ) , and in r. stumpf and m. scheffler , surf .
sci . * 307 - 309 * , 501 ( 1994 ) .
there exists a more elegant way to smear the occupation numbers around the fermi energy introduced in m. methfessel and a.t .
paxton , phys .
b * 40 * , 3616 ( 1989 ) .
there the total energy is minimized so that the energies and forces do not have to be corrected .
we therefore used this method in later studies .
feibelman , phys .
lett . * 69 * , 1568 ( 1992 ) .
the adsorption energies given in this paper are not directly comparable to ours ( see ref . ) because of the reference energy of the isolated atom .
apparently there is an inconsistency of this reference energy caused by feibelman s greens - function technique ; energy differences are comparable , however , and for these the main ( but small ) differences to our results arise because feibelman did not include the adsorbate - induced relaxation of the al(331 ) substrate .
cohesive and adsorption energies are given with respect to the energy of an isolated al atom calculated in a large cell with the same 8ry cut - off .
adding the spin - polarization energy of 0.15ev gives the cohesive energy as 4.15ev which is 0.75ev higher than the experimental value .
this overbinding is a problem common to converged dft - lda calculations .
it is widely accepted that adsorption energy _
differences _ for different sites are practically unaffected by this problem .
an al atom adsorbed at a kink site gains exactly the cohesive energy .
the embedding charge density is the central quantity in embedded atom and effective medium theory.@xcite the embedding density is closely connected to the coordination number of the adsorbate , so that coordination and embedding may in fact be regarded as different names for the same thing .
`` simple bond - cutting '' models assume that the energy per atom varies linearly with the atom s coordination number .
an improved version which takes the bond saturation into account makes this approach very similar to the effective - medium and embedded - atom methods ( see for example i.j robertson et al . , europhys .
lett . * 15 * , 301 ( 1991 ) , phys .
lett . * 70 * , 1944 ( 1993 ) , and m. methfessel et al .
a * 55 * , 442 ( 1992 ) )
. there could be a nonsymmetric exchange path at the \{111}-faceted step also .
we did not calculate that path as the symmetric one has already such a low barrier which is at the limit of the accuracy of our calculations .
our conclusions therefore would not be affected by an additional diffusion process .
no energetically favorable symmetrical exchange path exists for the \{100}-faceted step , as there the involved step atom would have to go over a top site .
we can estimate the temperature @xmath110 at which diffusion along the two kinds of steps will proceed at the same rate .
transforming eq.[difeq ] to @xmath111 and taking the values for @xmath88 and @xmath92 for the @xmath112\{111}- and @xmath112\{100 } step from tables[tab2 ] and [ edtd ] , we get @xmath113k .
we expect that this value is rather inaccurate however .
cdddd system & & adatom formation & vacancy formation + & ( ev / atom ) & ( ev / @xmath74 ) & ( ev ) & ( ev ) + al(111 ) & 0.48 & 0.070 & 1.05 & 0.67 + al(100 ) & 0.56 & 0.071 & 0.38 & 0.65 + al(110 ) & 0.89 & 0.080 & 0.26 & 0.12 + @xmath112\{111}-step & 0.232 & 0.082 & 0.28 & 0.21 + @xmath112\{100}-step & 0.248 & 0.088 & 0.25 & 0.24 ddddd # atoms & # edge atoms & surf . cell & @xmath114 & @xmath115 + 6 & 6 & @xmath116 & 0.025 & 0.029 + 10 & 9 & @xmath116 & 0.019 & 0.021 + 15 & 12 & @xmath117 & 0.017 & 0.018 + 21 & 15 & @xmath60 & 0.017 & 0.018 + lccdd site & coordination & @xmath120(ev ) & @xmath121( ) + fcc & 3 & @xmath1223.06 & 2.11 + bridge & 2 & @xmath1223.06 & 2.09 + hcp & 3 & @xmath1223.10 & 2.08 + top & 1 & @xmath1222.57 & 2.12 + fcc on @xmath112\{111}-step & 3 & @xmath1223.18 & + hcp on @xmath112\{100}-step & 3 & @xmath1223.18 & l|d|dd|dddd surface & al ( this work ) & & ni@xcite & rh@xcite & pt@xcite & ir@xcite + ( 111 ) & 0.04 & 0.054 & 0.074 & & 0.16 & ( 0.12)&0.27 + vacancy at al(111 ) & 0.56 & & & & & + ( 100 ) & 0.35 & 0.69 & 0.25 & & 0.63 & 0.47 & 0.84 + ( 110)@xmath123 & 0.33 & 0.26 & & ( 0.45)&0.60 & 0.84 & 0.80 + ( 110)@xmath124 & 0.62 & 0.30 & 0.15 & ( 0.45 ) & & 0.78 & 0.71 + @xmath112\{111}-step@xmath123 or ( 332)@xmath123 & 0.42 & 0.27 & 0.24 & ( 0.45)&0.64 & 0.84 & ( 1.05 ) + @xmath112\{100}-step@xmath123 or ( 644)@xmath123 & 0.32 & 0.20 & 0.24 & ( 0.37)&0.54 & 0.69 & ( 0.96 ) + cohesive energy@xcite & & 4.44&5.75 & 5.84 & 6.94 lddddd & & al(331 ) & @xmath112\{111}-step & @xmath112\{100}-step + & @xmath23 & @xmath125 & @xmath23 & @xmath23 & @xmath23 + ( a ) fivefold site & @xmath1223.89 & 1.33 & @xmath1223.68 & @xmath1223.87 & @xmath1223.90 + diffusion & @xmath126 & & @xmath126 & @xmath126 & @xmath126 + ( b ) long bridge & 0.60 & 1.58 & 0.57 & 0.48 & 0.32 + ( c ) short bridge & 1.06 & 2.16 & 1.21 & 1.03 & 1.15 + ( d ) exchange @xmath123 & 0.33 & 1.27 & & 0.39 & 0.44 + ( e ) exchange @xmath124 & 0.62 & 0.80 & & & 0.80 + .diffusion prefactors @xmath88 ( in cm@xmath74/s ) from theory for al ( mean of the two values given in ref . ) and from experiment for rh , pt , and ir surfaces .
values in brackets are considered to be less reliable .
the ( 331 ) surface has \{111}-faceted steps and the ( 311 ) surface has \{100}-faceted steps .
the column `` mechanism '' contains our assumptions about the mechanism of diffusion for every row , and the right column gives the diffusion prefactors which will be used in the temperature dependence of growth of al(111).[tab2 ] [ cols="<,^,^,^,^,^,^,^ " , ] lcddd & configuration & @xmath23 & @xmath128 & @xmath125 + 5-layer ( ref . ) & 4-fold & @xmath1222.93 & & 1.72 + & bridge & @xmath1222.28 & 0.65 & 2.20 + & exchange & @xmath1222.73 & 0.20 & 0.90 + 5-layer & 4-fold & @xmath1223.68 & & 1.58 + & bridge & @xmath1223.05 & 0.63 & 1.91 + & exchange & @xmath1223.55 & 0.13 & 0.74 + 6-layer & 4-fold & @xmath1223.75 & & 1.70 + & bridge & @xmath1223.12 & 0.63 & 2.11 + & exchange & @xmath1223.37 & 0.38 & 0.91 + 7-layer & 4-fold & @xmath1223.75 & & 1.73 + & bridge & @xmath1223.07 & 0.69 & 2.12 + & exchange & @xmath1223.35 & 0.40 & 0.98 + `` average '' & 4-fold & @xmath1223.77 & & 1.69 + & bridge & @xmath1223.12 & 0.68 & 2.09 + & exchange & @xmath1223.42 & 0.35 & 0.90 + | using density - functional theory we investigate several properties of al(111 ) , al(100 ) , al(110 ) , and stepped al(111 ) surfaces .
we report results of formation energies of surfaces , steps , adatoms , and vacancies . for the adsorption and diffusion of al on flat regions of al(111 ) surfaces we find the hcp site
energetically slightly preferred over the fcc site .
the energy barrier for self - diffusion on al(111 ) is very low ( 0.04ev ) .
coming close to one of the two sorts of close packed , monoatomic steps on al(111 ) , labeled according to their \{111 } and \{100 } micro - facets , al adatoms experience an attraction of @xmath0 0.1ev already before direct contact with the edge of the step .
this attraction has a range of several atomic spacings and is of electronic origin . upon arrival at the lower step edge
, the adatom attaches with no barrier at a low energy five - fold coordinated site . coming from the upper terrace
, it incorporates into the step by an atomic exchange process , which has a barrier below @xmath1ev for both sorts of close packed steps .
the barrier for diffusion along the lower edge is 0.32ev at the \{100}-faceted step and 0.39ev at the \{111}-faceted step .
unexpectedly the latter diffusion process proceeds by an exchange mechanism .
diffusion by a very similar exchange mechanism is also found for the `` easy '' direction on the al(110 ) surface , i.e. , along the channels .
we show that al(110 ) is a model system for diffusion at the \{111}-faceted step on al(111 ) because of its very similar local geometry .
our results enable the estimate of temperature ranges for different modes of homoepitaxial growth on al(111 ) .
of particular importance for the growth modes and the resulting surface morphology are the rather low barriers for diffusion across the descending steps and the rather high barriers for diffusion along the steps .
we discuss the shapes of islands on al(111 ) during growth and in thermodynamic equilibrium . during growth ( depending on the temperature )
the shape can be fractal , triangular , or hexagonal and mainly determined by kinetics ; in thermodynamic equilibrium the island shape is hexagonal and determined by the different step formation energies .
many of the phenomena , which we predict for al , were found for other metals in experiment . |
core partitions are combinatorial objects that appear naturally in various type @xmath8 settings .
they are used in the modular representation theory of the symmetric group to describe the blocks @xcite . in the geometry of the affine grassmannian , cores index schubert varieties and related homology classes called @xmath2-schur functions @xcite . at the level of coxeter groups ,
cores correspond to minimal length coset representatives for the parabolic quotient @xmath4 where @xmath5 denotes the affine symmetric group and @xmath6 denotes the finite symmetric group .
the impetus for this paper is combinatorial . in @xcite
, the first and third authors showed that the number of @xmath1-cores with first part equal to @xmath2 is @xmath9 using a bijection @xmath10 between @xmath1-cores with first part @xmath2 and @xmath3-cores with first part less than or equal to @xmath2 .
the fact that such a projection exists is remarkable in part because the coxeter group of affine type @xmath11 is not a parabolic subgroup of @xmath12 . in this work
we review some combinatorial models for cores and interpret the bijection in various guises . in the coxeter system setting , the bijection has a geometric interpretation as a projection from the root lattice of type @xmath7 to an embedded copy of the root lattice of type @xmath13 ; see figure [ f : sl3 ] .
we observe that the bijection reduces the coxeter length of the corresponding minimal length coset representative by exactly @xmath2 .
we also show that the bijection has a natural description in terms of another correspondence between @xmath1-cores and @xmath3-bounded partitions due to lapointe and morse @xcite . in section [
s : intro ] we introduce the bijection @xmath10 in terms of partition diagrams . in section [
s : geometry ] , we review the correspondence between @xmath1-cores and minimal length coset representatives for @xmath4 where @xmath5 denotes the affine symmetric group and @xmath6 denotes the finite symmetric group . in section [
s : geometric_bijection ] , we give a geometric version of the bijection @xmath10 on the root lattice of type @xmath7 . in section [ s :
l - m ] , we show that @xmath10 also has a natural description in terms of bounded partitions using the correspondence @xmath14 due to lapointe and morse @xcite .
let @xmath15 be a partition of @xmath16 and @xmath17 be an integer . throughout this paper ,
all of our partitions are drawn in english notation .
we will use the convention @xmath18 to denote the box which sits in the @xmath19 row and the @xmath20 column of the young diagram of @xmath21 .
we sometimes abuse notation and refer to row @xmath22 in the diagram of @xmath21 as @xmath23 .
@xmath24 will denote the set of all partitions .
the length of a partition @xmath21 is defined to be the number of nonzero parts of @xmath21 and will be denoted @xmath25 . for a fixed choice of @xmath1 , the _ residue _ of box @xmath26
is defined to be the least nonnegative integer @xmath27 congruent to @xmath28 modulo @xmath1 .
we let @xmath29 denote the kronecker delta function .
the _ hook length _ of the @xmath30 box of @xmath21 is defined to be the number of boxes to the right and below the box @xmath30 , including the box @xmath30 itself .
it will be denoted _
@xmath31_. _ _
[ d : lcore ] a partition @xmath21 is an _ @xmath1-core _ if for every box @xmath32 in the young diagram of @xmath21 , we have @xmath33 .
_ _ the set of all @xmath1-cores will be denoted @xmath34 .
the subset of @xmath34 having first part @xmath2 will be denoted @xmath35 and the subset of @xmath34 having first part @xmath36 will be denoted @xmath37 .
see james and kerber s book @xcite for more background on partitions and @xmath1-cores .
definition [ d : lcore ] is most useful for our purposes ; @xmath1-cores are more commonly defined as partitions having no removable @xmath1-rim hook .
the notion of @xmath38-numbers can be found in @xcite .
here we give a modified description of the @xmath38-numbers .
each partition @xmath39 is determined by its hook lengths in the first column , i.e. the @xmath40 . from a sequence @xmath41 of positive decreasing integers
one obtains a partition @xmath42 by requiring that the hook length @xmath43 for @xmath44 .
this gives a bijection between the set of partitions and the set of strictly decreasing sequences of positive integers .
one can generalize this process by looking at the set @xmath45 of infinite sequences @xmath46 of integers .
we give @xmath45 the group structure of component - wise addition .
we define the element * 1 * @xmath47 .
let @xmath48 denote the subgroup generated by * 1 * under addition , so @xmath49 .
a sequence @xmath50 is said to _ stabilize _ if there exists an @xmath16 so that @xmath51 for all @xmath52 .
the set @xmath53 is defined to be the subset of @xmath45 of strictly decreasing sequences that stabilize , modulo the added relation @xmath54 that two sequences are equivalent if their difference is in @xmath48 .
@xmath55 is in @xmath53 . in @xmath53
, we have @xmath56 we define a bijection @xmath38 between the set @xmath24 of partitions and @xmath53 . to a partition @xmath57 of length @xmath58
, we define @xmath59 to be the equivalence class of @xmath60 in @xmath53 .
@xmath61 is the equivalence class of @xmath62 an _ abacus diagram _ is a diagram containing @xmath1 columns labeled @xmath63 , called _
runners_. the horizontal cross - sections or rows will be called _ levels _ and runner @xmath22 contains entries labeled by @xmath64 on each level @xmath58 where @xmath65 .
we draw the abacus so that each runner is vertical , oriented with @xmath66 at the top and @xmath67 at the bottom , with runner 0 in the leftmost position , increasing to runner @xmath68 in the rightmost position .
entries in the abacus diagram may be circled ; such circled elements are called _ beads_. entries which are not circled will be called _
gaps_. the linear ordering of the entries given by the labels @xmath64 is called the _ reading order _ of the abacus and corresponds to scanning left to right , top to bottom .
_ _ _ _ the following abacus diagram has beads in positions ( @xmath69 , -3 , -2 , -1 , 1 , 2 , 4 , 5 , 8) and gaps in positions ( 0 , 3 , 6 , 7 , 9 , 10 , 11 , @xmath70 ) .
level 0 is the row which contains @xmath71 . + + ( 80,120)(0,10 ) ( 42.5,96 ) . ( 42.5,92 ) . ( 42.5,100 ) . ( 72.5,96 ) . ( 72.5,92 ) . ( 72.5,100 ) .
( 11.5,96 ) .
( 11.5,92 ) .
( 11.5,100 ) .
( 9,80 ) -3 ( 39,80)-2 ( 69,80)-1 ( 10,65 ) 0 ( 40,65)1 ( 70,65)2 ( 10,50)3 ( 40 , 50)4 ( 70,50)5 ( 10,35)6 ( 40,35)7 ( 70,35)8 ( 10,20)9 ( 38,20)10 ( 68,20)11 ( 0,105)(1,0)85 ( 0,120)runner ( 30,120)runner ( 60,120)runner ( 10,110)0 ( 40,110)1 ( 70,110)2 ( -40,65)level 0 @xmath72 ( -42,80)level -1 @xmath72 ( -40,50)level 1 @xmath72 ( -40,35)level 2 @xmath72 ( 72.5,53 ) ( 72.5,68 ) ( 72.5,38 ) ( 42.5,68 ) ( 42.5,53 ) ( 72.5,83 ) ( 42.5,83 ) ( 12.5,83 ) ( 42.5,12 ) . ( 42.5,8 ) .
( 42.5,4 ) .
( 72.5,12 ) .
( 72.5,8 ) .
( 72.5,4 ) .
( 11.5,12 ) .
( 11.5,8 ) .
( 11.5,4 ) . a representative @xmath73 of @xmath59 will be called a _ set of @xmath38-numbers _ for @xmath21 .
suppose @xmath74 is a set of @xmath38 numbers for @xmath21 .
an _ abacus _ for @xmath21 is obtained by circling the entries of @xmath73 in an abacus diagram . _ _ the following two diagrams are abaci for @xmath75 , the first comes from the @xmath38-numbers @xmath76 and the second comes from the equivalent @xmath38-numbers @xmath77 .
we list the beads in reverse reading order to be compatible with stability in @xmath53 .
+ + @xmath78 note that an abacus for @xmath21 is not unique because it depends on the set of @xmath38-numbers chosen for @xmath21 .
however , from any abacus of @xmath21 one can obtain the partition @xmath21 by counting the number of gaps before every bead in the abacus in reading order . in the first example above for instance
, we see that @xmath79 since the eight numbers 10,9,8,6,5,3,2,0 are exactly the eight gaps before the bead corresponding to the final bead at position @xmath80 .
we will say that a bead is _ active _ if it occurs in the positions between the first gap and the last bead , in reading order .
the active beads are those that correspond to a nonzero part in a partition . in the left example
above , the bead in spot 11 is active since it corresponds to the part @xmath79 , whereas the bead in spot -1 is not active since it corresponds to @xmath81 .
we define the _ balance number _ of an abacus to be the sum over all runners of the largest level in that runner which contains a bead .
we say that an abacus is _ balanced _ if its balance number is zero . in the example above , the balance number of the first diagram is @xmath82 .
the balance number for the second diagram is @xmath83 , so neither are balanced .
[ r : unique_abacus ] note that there is a unique abacus which represents a given partition for each balance number .
in particular , there is a unique abacus of @xmath21 with balance number 0 .
the balance number for a set of @xmath38-numbers of @xmath21 will increase by exactly @xmath84 when the vector * 1 * is added to the set of @xmath38-numbers . on the abacus picture
, this corresponds to shifting all of the beads forward one entry in the reading order .
a runner is called _ flush _ if no bead on the runner is preceded in reading order by a gap on that same runner .
we say that an abacus is _ flush _ if every runner is flush .
* theorem 2.7.16 , lemma 2.7.38)[t : flush_abacus ] @xmath21 is an @xmath1-core if and only if any ( equivalently , every ) abacus of @xmath21 is flush .
moreover , in the balanced flush abacus of an @xmath1-core @xmath21 , each active bead on runner @xmath22 corresponds to a row of @xmath21 whose rightmost box has residue @xmath22 . in the case that the corresponding abacus is not balanced , the boxes corresponding to the active beads on runner @xmath22 will share the same residue , but the residue may not be @xmath22 . [
52111 ] one can check that the partition @xmath85 is a 4-core .
one set of @xmath38-numbers for @xmath21 is @xmath86 .
this abacus is balanced as @xmath87 .
all of the runners are flush .
the active beads on runner 0 lie in positions 8 , 4 , 0 and these correspond to rows 1 , 2 and 5 of the partition diagram whose final box of residue 0 is highlighted .
llll ( 105,40)(0,80 ) ( 42.5,96 ) .
( 42.5,92 ) .
( 42.5,100 ) .
( 72.5,96 ) .
( 72.5,92 ) .
( 72.5,100 ) .
( 102.5,96 ) .
( 102.5,92 ) .
( 102.5,100 ) .
( 11.5,96 ) .
( 11.5,92 ) .
( 11.5,100 ) .
( 9,80 ) -8 ( 39,80)-7 ( 69,80)-6 ( 99,80)-5 ( 9,65 ) -4 ( 39,65)-3 ( 69,65)-2 ( 99,65)-1 ( 10,50)*0 * ( 40 , 50)1 ( 70,50)2 ( 101,50)3 ( 10,35)*4 * ( 40,35)5 ( 70,35)6 ( 101,35)7 ( 10,20)*8 * ( 40,20)9 ( 68,20)10 ( 99,20)11 ( 72.5,53 ) ( 72.5,68 ) ( 72.5,83 ) ( 102.5,83 ) ( 42.5,53 ) ( 42.5,68 ) ( 42.5,83 ) ( 12.5,23 ) ( 12.5,38 ) ( 12.5,53 ) ( 12.5,68 ) ( 12.5,83 ) ( 42.5,12 ) .
( 42.5,8 ) .
( 42.5,4 ) .
( 72.5,12 ) .
( 72.5,8 ) .
( 72.5,4 ) .
( 102.5,12 ) .
( 102.5,8 ) .
( 102.5,4 ) .
( 11.5,12 ) .
( 11.5,8 ) .
( 11.5,4 ) . & & & + & & & + here we describe the bijection @xmath88 .
given @xmath89 and an abacus for @xmath21 , remove the whole runner which contains the largest bead .
place the remaining runners into an @xmath3 abacus in order . in other words , renumber the runners @xmath90 , keeping the levels of the entries as before .
this will correspond to an @xmath91-core @xmath42 with largest part at most @xmath2 .
then we define @xmath88 to be the map which takes @xmath21 to @xmath42 .
observe that @xmath92 is well defined , independent of the choice of abacus for @xmath21 . to see that @xmath10 is a bijection , observe that the map can be reversed .
starting from an abacus of the @xmath3-core @xmath42 , insert a new flush runner whose largest bead occurs just after the @xmath93 gap in the reading order .
this yields a flush abacus for the @xmath1-core @xmath21 with @xmath94 .
let @xmath95 and @xmath96 .
an abacus for @xmath21 is : ( 80,100 ) ( 42.5,96 ) .
( 42.5,92 ) .
( 42.5,100 ) . ( 72.5,96 ) .
( 72.5,92 ) .
( 72.5,100 ) .
( 102.5,96 ) .
( 102.5,92 ) .
( 102.5,100 ) .
( 11.5,96 ) .
( 11.5,92 ) .
( 11.5,100 ) .
( 9,80 ) -8 ( 39,80)-7 ( 69,80)-6 ( 99,80)-5 ( 9,65 ) -4 ( 39,65)-3 ( 69,65)-2 ( 99,65)-1 ( 10,50)0 ( 40 , 50)1 ( 70,50)2 ( 101,50)3 ( 10,35)4 ( 40,35)5 ( 70,35)6 ( 101,35)7 ( 10,20)8 ( 40,20)9 ( 68,20)10 ( 99,20)11 ( 72.5,53 ) ( 72.5,68 ) ( 72.5,83 ) ( 102.5,83 ) ( 42.5,53 ) ( 42.5,68 ) ( 42.5,83 ) ( 72.5,23 ) ( 72.5,38 ) ( 102.5,68 ) ( 12.5,83 ) ( 42.5,12 ) .
( 42.5,8 ) .
( 42.5,4 ) .
( 72.5,12 ) .
( 72.5,8 ) .
( 72.5,4 ) .
( 102.5,12 ) .
( 102.5,8 ) .
( 102.5,4 ) .
( 11.5,12 ) .
( 11.5,8 ) .
( 11.5,4 ) .
the largest @xmath38-number is 10 .
removing the whole runner containing the 10 , we get the remaining diagram with runners relabeled for @xmath97 ( 80,100 ) ( 42.5,96 ) .
( 42.5,92 ) .
( 42.5,100 ) .
( 72.5,96 ) .
( 72.5,92 ) .
( 72.5,100 ) .
( 102.5,96 ) .
( 102.5,92 ) .
( 102.5,100 ) .
( 11.5,96 ) .
( 11.5,92 ) .
( 11.5,100 ) .
( 9,80 ) -6 ( 39,80)-5 ( 69,80)@xmath98 ( 99,80)-4 ( 9,65 ) -3 ( 39,65)-2 ( 69,65)@xmath98 ( 99,65)-1 ( 10,50)0 ( 40 , 50)1 ( 69,50)@xmath98 ( 101,50)2 ( 10,35)3 ( 40,35)4 ( 69,35)@xmath98 ( 101,35)5 ( 10,20)6 ( 40,20)7 ( 69,20)@xmath98 ( 101,20)8 ( 72.5,53 ) ( 72.5,68 ) ( 72.5,83 ) ( 102.5,83 ) ( 42.5,83 ) ( 42.5,68 ) ( 42.5,53 ) ( 72.5,22 ) ( 72.5,38 ) ( 102.5,68 ) ( 12.5,83 ) ( 42.5,12 ) .
( 42.5,8 ) .
( 42.5,4 ) .
( 72.5,12 ) .
( 72.5,8 ) .
( 72.5,4 ) .
( 102.5,12 ) .
( 102.5,8 ) .
( 102.5,4 ) .
( 11.5,12 ) .
( 11.5,8 ) .
( 11.5,4 ) .
these are a set of @xmath38-numbers for the partition @xmath99 , which is a 3-core with largest part @xmath100 . for the reverse bijection
when @xmath101 , notice that the eighth gap is at entry 7 which dictates where we insert the new runner and beads . also note in this example that the first abacus has balance number -1 while its image has balance number -3 , so balance number is not necessarily preserved . another way to describe
@xmath10 is on the young diagram of @xmath21 .
applying @xmath10 to @xmath21 is the same as removing all of the rows @xmath22 of @xmath21 for which @xmath102 . to illustrate
, we show the bijection on the same example @xmath96 , but performed on a young diagram instead of an abacus .
we start by drawing the young diagram and writing the hooks lengths of the boxes in the first column .
the bijection simply deletes the rows which have a hook length in the first column equivalent to the hook length @xmath103 .
@xmath104 deleting the corresponding rows , we get that @xmath105 .
in this section we recall that the @xmath1-cores index a system of minimal length coset representatives for @xmath4 and describe some associated geometry .
following @xcite , let @xmath106 be an orthonormal basis of the euclidean space @xmath107 and denote the corresponding inner product by @xmath108 .
for @xmath109 , let @xmath110 be the reflection defined by interchanging @xmath111 and @xmath112 ; the reflecting hyperplanes are discussed below
. then @xmath113 are a set of coxeter generators for the symmetric group @xmath6 , which acts on @xmath107 by permuting coordinates in the @xmath111 basis .
let @xmath114 be the affine reflection of @xmath107 defined on @xmath115 by @xmath116 define the _ simple roots _
@xmath117 _ of type @xmath7 _ to be the collection of @xmath68 vectors @xmath118 _ _ _ _ the @xmath119-span @xmath120 of @xmath117 is called the _ root lattice of type @xmath7__. let @xmath121 . observe that each reflection @xmath110 preserves @xmath122 and so @xmath123 are a set of coxeter generators for the affine symmetric group @xmath5 acting on @xmath122 .
for @xmath124 we let @xmath125 denote coxeter length . from now on ,
we restrict our attention from @xmath107 to @xmath122 . _ _ _ in this presentation we see that @xmath6 is a parabolic subgroup of @xmath5 .
we form the parabolic quotient @xmath126 by a standard result in the theory of coxeter groups , this set gives a unique representative of minimal length from each coset @xmath127 of @xmath4 . for more on this construction , see ( * ? ? ?
* section 2.4 ) .
another standard result is that @xmath5 acts on @xmath122 as the semidirect product of @xmath6 and the translation group corresponding to the root lattice @xmath120 .
hence , @xmath4 is also in bijection with @xmath120 and we identify @xmath120 with the translation subgroup @xmath128 of @xmath5 .
let us consider this situation geometrically .
denote the set of _ finite roots _ by @xmath129 .
it is a standard fact that each root @xmath130 can be written as an integral linear combination of the simple roots @xmath117 such that all of the coefficients are positive or all coefficients are negative .
therefore , @xmath131 can be decomposed as @xmath132 . _ _ for each finite root @xmath133 and integer
@xmath2 we can define an affine hyperplane @xmath134 observe that @xmath110 is the reflection over the hyperplane @xmath135 for @xmath136 while @xmath114 is the reflection over @xmath137 where @xmath138 .
let @xmath139 denote the collection of all affine hyperplanes @xmath140 for @xmath141 .
let @xmath142 be the set of all connected components of @xmath143 .
each element of @xmath142 is called an _ alcove__. in particular , @xmath144 is called the _ fundamental alcove _ whose closure is a fundamental domain for the action of @xmath5 on @xmath122 . _ _ _ _ _ ( * ? ? ?
* section 4.5 ) the affine weyl group @xmath5 permutes the collection of alcoves transitively and freely . the closure of @xmath145 is a fundamental domain for the action of @xmath5 on @xmath122 .
define @xmath146 the set @xmath147 contains one alcove for each permutation in @xmath6 and the closure of @xmath147 is a fundamental domain for the action of translation by @xmath120 on @xmath122 .
moreover , for any affine permutation @xmath148 we have that @xmath149 corresponds to the left coset of @xmath6 containing @xmath150 in @xmath4 .
in fact , the set of all @xmath149 is precisely the @xmath5-orbit of @xmath147 .
since length can be computed by counting the minimal number of hyperplanes which must be crossed in a path back to @xmath145 , we have that @xmath151 determines the alcove of @xmath149 that represents @xmath150 and has minimal length : it is the alcove which requires the fewest such hyperplane crossings .
this describes a bijection that we denote @xmath152 in which @xmath153 .
figure [ f : sl3 ] shows the euclidean space @xmath122 associated to type @xmath154 in the context of the bijection @xmath10 .
we now show how an @xmath1-core can be associated to each @xmath155 .
let @xmath156 be a vector in @xmath120 written with respect to the @xmath111 basis , so each @xmath157 and @xmath158 .
we form a balanced flush abacus from @xmath151 by filling the @xmath159 runner with beads from @xmath66 down to level @xmath160 . by remark [ r : unique_abacus ] and theorem [ t : flush_abacus ]
, every @xmath1-core has exactly one balanced flush abacus .
hence , we obtain bijections whose composition we denote by @xmath161 .
@xmath162 because the runners of an abacus are usually labeled by @xmath163 but coordinates of @xmath107 are labeled by @xmath164 we will sometimes coordinatize @xmath120 as @xmath156 or as @xmath165 .
let @xmath95 and let @xmath166 .
then we draw an abacus as shown in figure [ f : ex_pi ] above with beads down to level 2 in runner 0 , level 0 in runner 1 , level 0 in runner 2 , and level -2 in runner 3 . next , we observe that the @xmath1-cores inherit an action of @xmath5 from the bijection @xmath161 .
to describe the action , we draw the diagram of @xmath15 and fill each box @xmath26 with the residue @xmath167 . fix @xmath168 and suppose @xmath169 . for @xmath163 , we say that @xmath110 is an _ ascent _ for @xmath21 if @xmath170 , and we say that @xmath110 is a _ descent _ for @xmath21 if @xmath171 . here
, we interpret @xmath172 as @xmath173 . _ _ _ _ note that the definition for @xmath110 to be a descent ( respectively , ascent , neither ) given above corresponds to @xmath174 ( respectively , @xmath175 , @xmath176 ) .
this corresponds to whether @xmath177 .
the 4-core @xmath178 corresponds to @xmath179 .
hence , @xmath180 and @xmath181 are ascents for @xmath21 and @xmath114 is a descent for @xmath21 . observe that @xmath182 is neither an ascent nor a descent because @xmath183 ceases to be a minimal length left coset representative .
correspondingly @xmath184 .
[ p : action ] let @xmath21 be an @xmath1-core .
if @xmath110 is an ascent for @xmath21 then @xmath110 acts on @xmath21 by adding all boxes with residue @xmath22 to @xmath21 such that the result is a partition .
if @xmath110 is a descent for @xmath21 then @xmath110 acts on @xmath21 by removing all of the boxes with residue @xmath22 that lie at the end of both their row and column so that their removal results in a partition .
if @xmath110 is neither an ascent nor a descent for @xmath21 then @xmath110 does not change @xmath21 .
begin by considering the action of @xmath110 on abaci that comes from the action on @xmath185 .
applying @xmath110 for @xmath186 corresponds to exchanging adjacent runners @xmath187 and @xmath22 in the balanced flush abacus whose runners are labeled @xmath188 .
applying the @xmath114 generator first adds a bead to runner @xmath68 and removes a bead from runner @xmath189 so that they stay flush , and then exchanges the two runners .
since the coordinates of @xmath190 sum to 0 , theorem [ t : flush_abacus ] implies that each active bead in runner @xmath22 of the balanced flush abacus corresponds to a row of @xmath21 whose rightmost box has residue @xmath22 . because the abacus is flush , exchanging runners @xmath22 and @xmath187 either adds some set of boxes with residue @xmath22 to the diagram of @xmath21 in the case that @xmath110 is an ascent , or else removes a set of boxes with residue @xmath22 in the case that @xmath110 is a descent .
the result is again a balanced flush abacus so corresponds to an @xmath1-core .
if @xmath110 is neither an ascent nor a descent then @xmath191 so the abacus remains unchanged .
observe that a box with residue @xmath22 is removable if and only if it lies at the end of its row and column .
this occurs if and only if it corresponds to an active bead on runner @xmath22 with a gap immediately preceding it in the reading order of the abacus .
similarly , a box with residue @xmath22 is addable if and only if it corresponds to a gap on runner @xmath22 with an active bead immediately succeeding it in the reading order of the abacus .
the action of @xmath110 swaps runners @xmath22 and @xmath187 which therefore interchanges all of the @xmath22-addable and @xmath22-removable boxes
. this action can be described as a special case of the action on the crystal graph associated to the irreducible highest - weight representation @xmath192 of @xmath193 that is given by the operators @xmath194 or @xmath195 . for details , see @xcite , @xcite or @xcite .
let @xmath21 be an @xmath1-core .
then we can recursively define a canonical reduced expression for @xmath196 that we denote @xmath197 by choosing @xmath198 where @xmath22 is the residue of the rightmost box in the bottom row of @xmath21 and @xmath199 is the result of applying @xmath110 to @xmath21 as in proposition [ p : action ] . note that @xmath110 is always a descent for this choice of @xmath22 .
the empty partition corresponds to the identity coxeter element .
this reduced expression was previously defined in ( * ? ? ?
* definition 45 ) .
the canonical reduced expression for the 4-core @xmath200 shown in figure [ f : ex_pi ] is @xmath201 .
the first step in reducing this expression to the identity removes the three boxes labeled 0 that lie at the end of their rows which is recorded as the leftmost @xmath114 in the expression .
@xmath202 @xmath203 @xmath204 [ p : redexp ] for @xmath168 we have that @xmath197 is a reduced expression for the minimal length coset representative indexed by @xmath21 .
the coxeter length of @xmath197 is @xmath205 where @xmath206 is the longest row of @xmath21 whose rightmost box has residue @xmath22 .
let @xmath22 be the residue of the rightmost box in the bottom row @xmath207 of @xmath21 .
then the balanced flush abacus configuration corresponding to @xmath21 has an active bead @xmath45 representing @xmath207 on runner @xmath22 by theorem [ t : flush_abacus ] . observe that @xmath45 has a gap immediately preceding it in the reading order because @xmath45 is the first active bead in the reading order and @xmath208 .
since the abacus is flush , every box of @xmath21 with residue @xmath22 that lies at the end of its row corresponds to some active bead @xmath209 on runner @xmath22 of the abacus with a gap immediately preceding @xmath209 in the reading order .
hence , every box of @xmath21 with residue @xmath22 that lies at the end of its row also lies at the end of its column and applying @xmath110 removes every such box by proposition [ p : action ] . iterating this process eventually produces the empty partition , corresponding to the identity coxeter element . at each step
, we have shown that applying @xmath110 removes exactly one box from @xmath206 yielding the length formula .
let @xmath210 and @xmath95 .
then the biggest part ending in residue 1 is @xmath211 , the biggest part ending in residue 3 is @xmath212 , the biggest part ending in residue 0 is @xmath213 , and no part ends in residue 2 .
hence the coxeter length of the minimal length coset representative for this 4-core is @xmath214 .
the canonical minimal length coset representative for this core is @xmath215 .
@xmath216 there is another way to obtain the root vector @xmath217 from the partition @xmath168 due to garvan , kim and stanton @xcite .
region @xmath58 _ of the diagram of @xmath21 is the set of boxes @xmath26 satisfying @xmath218 .
we call a box _ row - exposed _ if it lies at the end of its row
. then , set @xmath219 to be the maximum region of @xmath21 which contains a row - exposed box with residue @xmath22 for @xmath163 .
in particular , we pad @xmath21 with parts of size zero if necessary and label all of the boxes before the @xmath220 column by their residue . in this way @xmath219 is well - defined because column 0 contains infinitely many row - exposed boxes .
we call the vector @xmath221 obtained in this fashion the _ @xmath16-vector of @xmath21 _ and we show that it is the same vector as @xmath222 .
_ let @xmath95 and @xmath223 . from the picture
below , we see that the @xmath16-vector for @xmath21 is @xmath224 .
@xmath225 [ l : n_vector ] let @xmath221 be the @xmath16-vector of an @xmath1-core @xmath21 and let @xmath226 be the result of applying @xmath110 to @xmath21 as described in proposition [ p : action ] .
then , the @xmath16-vector of @xmath226 is @xmath227 for @xmath186 , or @xmath228 if @xmath229 .
denote the @xmath16-vector of @xmath226 by @xmath230 .
it follows from the proof in ( * ? ? ?
* bijection 2 ) that if there exists a row - exposed @xmath22-box in region @xmath58 then there exist row - exposed @xmath22-boxes in each region @xmath231 . observe that a row @xmath232 of @xmath21 has an addable @xmath22-box in region @xmath58 exactly if : 1 .
the row @xmath232 has a row - exposed @xmath233-box in region @xmath234 , and 2 .
the row @xmath235 does not have a row - exposed @xmath22-box , so @xmath236 .
similarly , a row @xmath237 has a removable @xmath22-box in region @xmath238 exactly if : 1 .
the row @xmath237 has a row - exposed @xmath22-box in region @xmath239 , and 2 .
the row @xmath240 does not have a row - exposed @xmath233-box , so @xmath241 . moreover , an @xmath22-box is addable ( removable ) in @xmath21 only if it is removable ( addable , respectively ) in @xmath226 .
therefore , when we apply @xmath110 to @xmath21 we interchange each of the regions @xmath58 and @xmath238 satisfying @xmath242 and @xmath243 hence , we obtain @xmath244 and @xmath245 the @xmath16-vector of @xmath21 is @xmath222 .
the @xmath16-vector of the empty partition is equal to @xmath246 .
the result then follows by induction on the coxeter length of @xmath197 by lemma [ l : n_vector ] .
[ p : k_from_vector ] suppose that @xmath247 .
then we have @xmath248 where @xmath160 is the rightmost occurrence of the largest coordinate in @xmath151 . also , @xmath249 consider the balanced flush abacus corresponding to @xmath21
. then @xmath250 corresponds to the last active bead @xmath45 in the reading order , and @xmath45 lies on runner @xmath187 . in particular
, if there are multiple occurrences of the largest coordinate in @xmath151 then @xmath250 corresponds to the bead on the rightmost runner .
the number of boxes in @xmath250 is the number of gaps prior to @xmath45 in the reading order of the abacus and the second formula follows from counting these gaps , using that the beads are flush on each runner .
since the abacus is balanced , we have that the number of beads strictly below the zero level must be equal to the number of gaps weakly above the zero level .
if the last active bead occurs in level @xmath251 of runner @xmath187 then we could move all of the beads below the zero level to fill in the gaps above the zero level , and so count the gaps starting from entry 0 to the entry that contained @xmath45 as @xmath252 .
this yields the first formula .
the balanced flush abacus corresponding to the 4-core @xmath253 is shown below together with the diagram in which the beads have been moved to calculate @xmath254 as @xmath255 . here , @xmath256 is the largest entry , corresponding to runner 2 .
ccc ( 105,110)(42.5,96 ) .
( 42.5,92 ) .
( 42.5,100 ) .
( 72.5,96 ) .
( 72.5,92 ) .
( 72.5,100 ) .
( 102.5,96 ) .
( 102.5,92 ) .
( 102.5,100 ) .
( 11.5,96 ) .
( 11.5,92 ) .
( 11.5,100 ) .
( 9,80 ) -8 ( 39,80)-7 ( 69,80)-6 ( 99,80)-5 ( 9,65 ) -4 ( 39,65)-3 ( 69,65)-2 ( 99,65)-1 ( 10,50)0 ( 40 , 50)1 ( 70,50)2 ( 101,50)3 ( 10,35)4 ( 40,35)5 ( 70,35)6 ( 101,35)7 ( 10,20)8 ( 40,20)9 ( 68,20)10 ( 99,20)11 ( 5,45.5)(90,0)105 ( 72.5,53 ) ( 72.5,68 ) ( 72.5,83 ) ( 102.5,83 ) ( 42.5,83 ) ( 12.5,38 ) ( 12.5,53 ) ( 12.5,68 ) ( 12.5,83 ) ( 72.5,38 ) ( 72.5,23 ) ( 101.5,68 ) ( 42.5,12 ) .
( 42.5,8 ) .
( 42.5,4 ) .
( 72.5,12 ) .
( 72.5,8 ) .
( 72.5,4 ) .
( 102.5,12 ) .
( 102.5,8 ) .
( 102.5,4 ) .
( 11.5,12 ) .
( 11.5,8 ) .
( 11.5,4 ) . & & ( 105,100)(42.5,96 ) .
( 42.5,92 ) .
( 42.5,100 ) .
( 72.5,96 ) .
( 72.5,92 ) .
( 72.5,100 ) .
( 102.5,96 ) .
( 102.5,92 ) .
( 102.5,100 ) .
( 11.5,96 ) .
( 11.5,92 ) .
( 11.5,100 ) .
( 9,80 ) -8 ( 39,80)-7 ( 69,80)-6 ( 99,80)-5 ( 9,65 ) -4 ( 39,65)-3 ( 69,65)-2 ( 99,65)-1 ( 10,50)0 ( 40 , 50)1 ( 70,50)2 ( 101,50)3 ( 10,35)4 ( 40,35)5 ( 70,35)6 ( 101,35)7 ( 10,20)8 ( 40,20)9 ( 68,20 ) ( 99,20)11 ( 72.5,53 ) ( 72.5,68 ) ( 72.5,83 ) ( 102.5,83 ) ( 42.5,53 ) ( 42.5,68 ) ( 42.5,83 ) ( 12.5,53 ) ( 12.5,68 ) ( 12.5,83 ) ( 101.5,68 ) ( 103.5,53 ) ( 42.5,12 ) .
( 42.5,8 ) .
( 42.5,4 ) .
( 72.5,12 ) .
( 72.5,8 ) .
( 72.5,4 ) .
( 102.5,12 ) .
( 102.5,8 ) .
( 102.5,4 ) .
( 11.5,12 ) .
( 11.5,8 ) .
( 11.5,4 ) .
+ [ c : hyperplane ] for @xmath257 , let @xmath258 denote the affine hyperplane @xmath259 inside @xmath122 , where @xmath260 . then under the correspondence @xmath161 , the @xmath1-cores @xmath21 with @xmath94 all lie inside @xmath261 . we can write @xmath262 uniquely as @xmath263 for @xmath264 and @xmath265 . in this case ,
@xmath266 and @xmath267 .
the result then follows from the first formula of proposition [ p : k_from_vector ] . if @xmath268 then @xmath269 and @xmath270 so the statement holds .
from proposition [ p : k_from_vector ] we see that if @xmath160 is the rightmost occurrence of the largest coordinate in @xmath271 , then @xmath272 .
the next result describes @xmath273 in terms of the root lattice coordinates .
[ t : main ] let @xmath274 be the affine map defined by @xmath275 . then , @xmath276 where @xmath160 is the rightmost occurrence of the largest entry among @xmath277 and the circumflex indicates omission .
. then @xmath21 corresponds to a balanced flush abacus @xmath8 in which the first row of @xmath21 corresponds to the last active bead @xmath45 in the reading order for the abacus , and @xmath45 occurs on runner @xmath187 .
the bijection @xmath279 is defined on @xmath8 by deleting the runner @xmath187 in order to obtain the abacus of an @xmath3-core .
since the original abacus is balanced , when we remove runner @xmath187 from @xmath8 we are left with an abacus @xmath280 in which the balance number is @xmath281 .
applying @xmath282 corresponds to shifting all of the entries of @xmath280 forward one entry in the reading order of @xmath280 , or equivalently adding @xmath283 to the @xmath38-numbers for @xmath21 .
hence , applying @xmath284 to @xmath280 produces a balanced flush abacus for the same partition as @xmath279 .
the geometric interpretation of @xmath10 as pictured in figure [ f : sl3 ] can also be described as follows .
we observed for an @xmath1-core @xmath21 with @xmath94 that @xmath222 lies in the affine hyperplane @xmath285 .
we can identify @xmath286 with a copy of the root lattice of @xmath13 via @xmath287 .
this embedding requires cyclically shifting coordinates as described by @xmath288 and depends on @xmath2 .
[ e : coords ] let @xmath97 .
the affine hyperplane @xmath289 contains @xmath290 .
we decompose @xmath291 as the composition of a translation and a cyclic shift of coordinates .
translating by the vector @xmath292 sends @xmath289 to the linear hyperplane @xmath293 and in particular sends @xmath294 to @xmath295 .
we view @xmath296 as a subspace of @xmath297 with orthonormal basis @xmath298 and an associated root system of type @xmath13 .
hence , we must identify @xmath299 with @xmath300 and @xmath301 with @xmath302 .
therefore , we have @xmath303 corresponding to @xmath304 . as a projection of @xmath305 .
the @xmath149 are hexagons which tile the plane centered at @xmath151 which is darkened , and each @xmath149 contains 6 triangular alcoves .
each alcove corresponds to an element of @xmath5 and each hexagon corresponds to a coset in @xmath4 , so there is a unique 3-core alcove in each hexagon .
these are shaded above .
two alcoves @xmath149 and @xmath306 share the same color only if the first parts of the partitions @xmath307 and @xmath308 agree . ]
if we focus on @xmath25 instead of @xmath250 , all of the @xmath1-cores with fixed length @xmath309 have @xmath222 lying in the affine hyperplane @xmath310 if we drew dotted lines in figure [ f : sl3 ] connecting the @xmath222 with fixed @xmath309 ( instead of those with fixed @xmath311 ) , then the lines would appear to spiral backwards from the direction of those in figure [ f : sl3 ] .
this can be explained by the fact that sending a partition to its transpose corresponds to the transformation of @xmath107 given by @xmath312 as shown in @xcite .
recall from section [ s : phi_on_diagram ] that @xmath10 removes all rows from @xmath21 in the same equivalence class as the first row . in proposition
[ p : redexp ] , we described a canonical reduced expression @xmath197 for @xmath313 . in this construction
, the rows of @xmath21 are partitioned into @xmath1 equivalence classes which we denote by @xmath314 $ ] according to the residue @xmath251 of their rightmost box .
let @xmath22 be the residue of the rightmost box @xmath315 in the last row @xmath316 of @xmath21 .
observe that @xmath315 is a removable @xmath22-box unless @xmath270 , and applying @xmath110 removes one box from each of the rows @xmath317 $ ] .
we claim that two rows of @xmath318 are equivalent if and only if the rows were equivalent in @xmath21 . to see this , consider that there can be no rows @xmath319 $ ] in @xmath21 .
otherwise there exists a box in the same column as @xmath315 whose row is @xmath319 $ ] , and so the hooklength of this box is divisible by @xmath1 which contradicts @xmath21 being an @xmath1-core .
therefore , the rows @xmath319 $ ] in @xmath226 are precisely the rows @xmath317 $ ] in @xmath21 with the possible exception of @xmath316 if @xmath320 . in any case , no other rows change equivalence classes .
suppose @xmath321 is the canonical reduced expression for @xmath21 obtained from proposition [ p : redexp ] , so @xmath322 is the residue of the box @xmath315 and @xmath323 .
working left to right to reduce @xmath197 to the identity , each application of @xmath324 removes a box from every row that is in the same equivalence class as the last row of the intermediate partition .
let @xmath325 be the subset of @xmath326 such that the first row and the last row are not in the same equivalence class in the @xmath1-core @xmath327 .
then the subexpression of @xmath197 corresponding to the indices in @xmath325 gives the canonical minimal length coset representative for the @xmath3-core @xmath328 after relabeling the residues with respect to @xmath68 .
since we remove a box from every row in the equivalence class of the last row at each step , we remove in particular a box from the longest row in that equivalence class .
hence , we see that the positions @xmath329 that are deleted from @xmath197 in the application of @xmath10 correspond with boxes in the first row of @xmath21 , so applying @xmath10 reduces the coxeter length by exactly @xmath2 .
suppose @xmath330 and let @xmath331 .
we label the diagram of @xmath21 as shown in figure [ f : subexp ] by residues with respect to @xmath1 on the bottom of each box , and with respect to @xmath68 on the top of those boxes that are not in rows equivalent to the first row .
we put dots as placeholders in the tops of these boxes .
@xmath332 we see directly from this diagram that @xmath333 where the entries in @xmath325 have been highlighted . for example
, applying @xmath182 to @xmath21 removes the last row as well as the last box from row 5 .
since we did not remove a box from the first row , we have position @xmath334 .
the next step applies @xmath181 to remove the last row of @xmath335 as well as the last box from rows 1 , 2 , and 4 , so position @xmath336 . after shifting the residues in the bold subexpression @xmath337 to be calculated relative to @xmath68
as shown in the diagram , we obtain @xmath338 .
the coxeter length has been reduced by @xmath339 .
in this section , we show that the bijection @xmath10 can be succinctly expressed as a map between @xmath2-bounded partitions ( one whose first part is @xmath36 ) and @xmath340-bounded partitions , using the correspondence @xmath341 of lapointe and morse ( * ? ? ? * section 3 ) to which we refer the reader for more details .
let @xmath342 be defined by the property that the left square in the following diagram commutes , and define @xmath343 to simply delete the first column from the diagram of the partition .
then , we claim that the right square in the diagram also commutes . here , @xmath344 denotes the transpose of a partition .
recall that @xmath346 denotes the hooklength of a box @xmath315 in a partition diagram @xmath21 .
let @xmath347 be an @xmath1-core with first column length @xmath2 .
observe that @xmath342 acts by removing entire _ columns _ from @xmath21 by transposing the explanation in section [ s : phi_on_diagram ] .
row - wise , we can describe @xmath342 as removing the leftmost box @xmath315 in the row together with all of the boxes @xmath348 in the row having @xmath349 .
_ _ as in @xcite , we view @xmath350 as the result of left - justifying all of the rows in a skew - diagram @xmath351 , where @xmath352 consists of the boxes of @xmath21 having hooklength @xmath353 .
we say that the boxes of @xmath21 lying in @xmath351 are the _ skew boxes _ while the boxes of @xmath352 are the _ non - skew boxes__. _ _ _ _ _ suppose @xmath330 and consider the @xmath1-core @xmath354 . in the diagrams below , we have labeled the boxes by their hooklengths .
the skew - boxes of @xmath355 are indicated in boldface , so @xmath356 . the entries that are deleted in the application of @xmath342 are indicated with underline .
@xmath357 from these diagrams , we see that @xmath358 is a 4-core and @xmath359 .
[ l : lm - one - rem ] there is exactly one skew box deleted from each row of @xmath21 in the application of @xmath342 .
also , a box @xmath315 in @xmath21 that is not deleted in the application of @xmath342 is skew with respect to @xmath1 if and only if the corresponding box @xmath366 of @xmath358 is a skew box with respect to @xmath68 .
it suffices to prove these statements for a fixed row .
since the skew boxes are those with hooklength @xmath367 , we delete at most one skew box from each row of @xmath21 when we apply @xmath342 .
next , we show that at least one skew box is deleted .
let @xmath368 be the leftmost box of @xmath21 in row @xmath22 .
since @xmath21 is an @xmath1-core , the partition @xmath369 is also an @xmath1-core .
form an unbalanced abacus @xmath8 of @xmath370 such that the beads of @xmath8 correspond to hooklengths of the first _ row _ of @xmath199 which is the @xmath371 row of @xmath21 .
in particular , those beads corresponding to the boxes with equivalent hooklength to @xmath372 form the rightmost longest runner of the abacus @xmath8 and they are flush . hence , there exists a box @xmath373 in the same row as @xmath368 having hooklength @xmath374 such that @xmath375 , so the corresponding box @xmath373 is a skew box .
the box @xmath373 is deleted from @xmath21 when we apply @xmath342 . _ _ to prove the second statement , suppose @xmath315 is a box in the @xmath22th row of @xmath21 that does not get deleted in @xmath362 .
let @xmath366 be the corresponding box in @xmath362 .
then @xmath315 corresponds to an active bead in the abacus @xmath8 that is not on same runner as @xmath368 .
we have that @xmath315 is skew with respect to @xmath1 if and only if the corresponding bead lies on level 0 of the abacus @xmath8 .
since removing the runner containing @xmath368 does not change any levels of the remaining beads , we have that @xmath366 is skew with respect to @xmath68 if and only if @xmath315 is skew with respect to @xmath1 . c. berg and m. vazirani , ( @xmath377-carter partitions , a generating function , and their crystal theoretic interpretation , _ arxiv mathematics e - prints _ , math.co/0712.2075 .
a. bjrner , f. brenti , combinatorics of coxeter groups , graduate texts in mathematics , springer , berlin ( 2005 ) .
f. garvan , d. kim , and d. stanton , cranks and t - cores , inv .
( 1990 ) , 1 - 17 .
i. grojnowski , affine @xmath378 controls the representation theory of the symmetric group and related hecke algebras , _ arxiv mathematics e - prints _ , math/9907129 humpreys , _ reflection groups and coxeter groups _ , cambridge stud .
math * 29 * cambridge univ . press ( 1990 ) .
g.d.james and a.kerber , _ the representation theory of the symmetric group _ , encyclopedia of mathematics , * 16 * , 1981 .
a. kleshchev , _ linear and projective representations of symmetric groups _
, cambridge tracts in mathematics * 163*. l. lapointe and j. morse , tableaux on k+1-cores , reduced words for affine permutations , and k - schur expansions , _ journal of combinatorial theory _ , series a volume 112 , issue 1 , october 2005 , pages 44 - 81 k.c .
misra and t. miwa , crystal base for the basic representation of @xmath379 , _ commun .
* 134 * ( 1990 ) , 79 - 88 . | let @xmath0 be fixed positive integers . in @xcite , the first and third authors established a bijection between @xmath1-cores with first part equal to @xmath2 and @xmath3-cores with first part less than or equal to @xmath2 .
this paper gives several new interpretations of that bijection .
the @xmath1-cores index minimal length coset representatives for @xmath4 where @xmath5 denotes the affine symmetric group and @xmath6 denotes the finite symmetric group . in this
setting , the bijection has a beautiful geometric interpretation in terms of the root lattice of type @xmath7 .
we also show that the bijection has a natural description in terms of another correspondence due to lapointe and morse @xcite . |
the s - d mixing for @xmath11 and @xmath12 is of great interest in charmonium physics .
if we neglect the charmed meson pair component which is due to coupling to decay channels , @xmath13 and @xmath14 may be approximately expressed as @xmath15 a rough estimate of the s - d mixing angle may be obtained by using the ratio of the observed leptonic decay widths @xcite of @xmath13 and @xmath14 and neglecting the d - wave component contribution to these leptonic decay widths @xmath16 , which results in @xmath17 . however ,
if the d - wave contribution to leptonic decay widths is further included , potential model calculations e.g. in @xcite give two solutions : @xmath18 to @xmath19 or @xmath20 to @xmath21 ( @xmath18 and @xmath22 in @xcite , @xmath23 and @xmath21 in @xcite , and @xmath24 and @xmath25 in @xcite ) .
the small mixing solution ( i.e. @xmath18 ) is compatible with the results obtained in models with coupled decay channels @xcite .
moreover , @xmath18 is favored by the @xmath26 data whereas @xmath27 would lead to @xmath28135 kev due to a large positive s - d interference @xcite , which is higher than its observed value by a factor of 6 and is therefore disfavored .
the s - d mixing may have many interesting phenomenological consequences .
it might slightly @xcite or substantially @xcite affect the @xmath29 decay , which has been recently observed by bes @xcite .
it would also have effects on the @xmath30 search via the @xmath31 decay @xcite .
the s - d mixing could even provide an explanation for the notorious @xmath32 puzzle that the suppression of @xmath33 is due to a destructive interference between the s and d wave states @xcite , and it might also be useful in explaining @xcite the recent observed enhancement of @xmath34 by bes @xcite . moreover , the ratio of @xmath35 may sensitively change from 0.04 ( for @xmath9 ) to 0.22 ( for @xmath18 ) and to 0.06 ( for @xmath20 ) @xcite ) . the experimental examination of these radiative transitions for @xmath4 at bes and cleo - c in the near future will be an interesting test for the s - d mixing .
recently belle collaboration @xcite has observed @xmath4 for the first time in the b meson decay @xmath0 with a branching ratio of @xmath36 , which is comparable to @xmath37 @xcite .
this is quite surprising , since conventionally @xmath4 and @xmath38 are regarded as mainly the @xmath39 and @xmath40 color - singlet @xmath41 states respectively , and the coupling of @xmath39 to the @xmath42 vector current in the weak decay effective hamiltonian is much weaker than that of @xmath40 .
one possible explanation is that the s - d mixing for @xmath4 and @xmath38 is very large , much larger than previously expected .
if in the @xmath3 meson decay we neglect the @xmath39 contribution , which is expected to be much smaller than the @xmath40 contribution , we would get the s - d mixing angle @xmath43 from the observed decay rate ratio @xmath44 although this large mixing angle seems to be not compatible with all our previous knowledge about the s - d mixing , it is still worthwhile to test it with new experiments .
another possible explanation is that the color - octet mechanism in nonrelativistic qcd ( nrqcd ) may play the leading role in the d - wave charmonium production in b meson decays , and it was predicted @xcite that for a pure d - wave state the branching ratio @xmath45 , which is comparable to @xmath46 @xcite ( see also @xcite for similar discussions ) . in a series of papers @xcite it was pointed out that based on the fock state expansion and velocity scaling rules of nrqcd @xcite the production rates of d - wave charmonium states including the @xmath47 , which would predominantly decay to the @xmath48 meson pair , and an expected narrow @xmath49 state , which could have some decay fraction to @xmath50 , would be comparable to that of @xmath51 and @xmath52 in the @xmath53 decay @xcite , the @xmath54 collision at the tevatron @xcite , the b meson decay @xcite , and the fixed target experiments @xcite , whereas in the conventional color - singlet model the d - wave production rates , which are proportional to the squared second derivative of the @xmath41 wave function at the origin , should be greatly suppressed . despite of certain uncertainties associated with the values of color - octet matrix elements ,
the color - octet contributions are expected to be dominant , which are larger than the color - singlet contributions by more than one order of magnitude in those processes .
for instance , in the large @xmath55 charmonium production at the tevatron , the @xmath56 states could have large rates , comparable to @xmath52 , because in both cases the dominant production mechanism is expected to be gluon fragmentation into the color - octet @xmath57 @xmath41 intermediate state which then evolves respectively into the physical color - singlet @xmath56 and @xmath52 states by emitting two soft gluons via double e1 transitions with the same order transition probabilities @xcite .
similarly , the @xmath4 production in the b meson decay could provide another interesting test for the color - octet mechanism in nrqcd . in order to distinguish between the large s - d mixing and the nrqcd color - octet mechanism in the @xmath4 production in the b meson decay , we suggest measuring the @xmath4 production in continuum @xmath5 annihilation at the belle and babar energy @xmath6 gev , and we will give calculations of the production cross sections in the following sections .
we will show that the @xmath4 production in @xmath5 annihilation via double @xmath41 does not receive large contributions from the color - octet mechanism but is very sensitive to the s - d mixing .
@xmath51 inclusive production in @xmath5 annihilation has been investigated within the color - singlet model @xcite and the color - octet model @xcite .
babar @xcite and belle @xcite have measured @xmath51 production rate in continuum @xmath5 annihilations at @xmath58 , which is found to be much larger than the color - singlet prediction .
more interestingly , belle further finds @xmath51 production to be dominated by the double @xmath41 production @xcite .
the measured exclusive cross section for @xmath59 is an order of magnitude larger than the theoretical values@xcite , and the measured inclusive cross section for @xmath60 is more than five times larger than nrqcd predictions .
therefore the inclusive charmonium production via double @xmath41 is particularly interesting and worth investigating . as in the case of inclusive production of @xmath51 , @xmath61 , and @xmath62(j=0 , 1 , 2)@xcite , here
we will also concentrate on the double @xmath63 production for the d - wave charmonium states @xmath64 .
in nrqcd the fock state expansion for the d - wave ( without s - d mixing ) charmonium @xmath68 is @xmath69 following the nrqcd factorization formalism , the scattering amplitude of the process @xmath70 in fig . [ feynman ] is given by @xmath71 where @xmath72 is the @xmath67 pair produced at short distances , which subsequently evolve into a specific charmonium state at long distances , @xmath73 and @xmath74 are the derivatives of the amplitude with respect to the relative momentum between the quark and anti - quark in the bound state .
for the case of color - singlet state , the coefficient @xmath75 can be related to the origin of the radial wave function ( or its derivatives ) of the bound state as @xmath76 the spin projection operators and their derivatives with respect to the relative momentum are @xmath77 @xmath78.\ ] ] @xmath79.\ ] ] the calculation of cross sections for @xmath80 charmonium @xmath81 is straightforward . as in ref .
@xcite we write the differential cross section as @xmath82 where @xmath83 .
the expressions of s(z ) and @xmath84 for @xmath85 are lengthy and will be given in the appendix for @xmath86 . with eq .
( [ cross ] ) we can evaluate the inclusive cross sections for @xmath85 .
the input parameters used in the numerical calculations are @xmath87 and the obtained cross section for the @xmath65 at @xmath6 gev is @xmath88 here we also give the calculated cross section for the @xmath89 at @xmath6 gev @xmath90 we also find that as in the case of other charmonium states @xcite the calculated cross sections for @xmath65 in eq .
( [ dt1 ] ) and @xmath89 in eq .
( [ dt2 ] ) are substantially smaller than those obtained in the fragmentation approximation at @xmath6 gev which would cause a enhancement factor of 1.5 and 2.3 respectively . comparing eq.([dt1 ] ) with @xmath91 calculated for the @xmath51 in @xcite , we see that the inclusive double @xmath41 cross section for the d - wave @xmath92 state is smaller than that for the s - wave @xmath92 states by a factor of 60 .
this illustrates the expectation that within the color - singlet model the suppression of d - wave state production relative to the s - wave state production is usually about two orders of magnitude .
the color - octet contributions to @xmath65 production via double @xmath67 in @xmath5 annihilation come from the feynman diagrams in fig [ feynman ] and fig [ fey2 ] . according to the nrqcd factorization and velocity scaling rules@xcite , the contributions of the second term and third term in the fock state expansion of @xmath65 in eq .
( [ expandfock ] ) are of the same order in the quark relative velocity @xmath93 as the corresponding terms in the fock state expansion of @xmath51 or @xmath52 . as a rough estimate for the nonperturbative matrix elements we may choose @xmath94 @xmath95 @xmath96 @xmath97 here
the color - octet matrix elements for @xmath51 were extracted from the @xmath51 data at the tevatron ( see ref .
@xcite for detailed discussions ) .
there are large uncertainties with @xmath98 and @xmath99 and their combinations , and here we have assumed that they are equal and take the largest fitted values from @xcite to avoid underestimates of the color - octet contributions .
( note that these two matrix elements may be overestimated @xcite . ) in fig .
[ feynman ] , the color - octet contribution can be obtained from the corresponding color - singlet contribution divided by a factor of @xmath100 . with the matrix elements for @xmath101 chosen above
, we find the contributions to @xmath65 production cross section from the color - octet @xmath102 states to be 0.12fb , 1.1fb , 0.29fb , 0.14fb respectively .
the total color - octet contribution to the cross section from fig .
[ feynman ] is 1.65fb . in fig .
[ fey2 ] the color - octet contributions come from four different ( the upper two and the lower two ) diagrams .
the upper diagrams only contribute via the color - octet @xmath57 state , and the differential cross section reads @xmath103 where @xmath104 takes the form @xmath105 + \nonumber\\ & & ( z-2)^2[3\delta^4 - 8\delta^2(3z-4)+32(2+z(z-2))]\nonumber\\ & & { \rm ln}[\frac{z \sqrt{4 + \delta^2 - 4 z}+2\sqrt{(1-z)(z^2-\delta^2)}}{z \sqrt{4 + \delta^2 - 4 z}-2\sqrt{(1-z)(z^2-\delta^2)}}]\}.\end{aligned}\ ] ] the numerical result for the cross section is @xmath106 the lower diagrams in fig .
[ fey2 ] make contributions to @xmath65 production via the color - octet @xmath107 and @xmath108 intermediate states .
we find the following results .
@xmath109 @xmath110 @xmath111 @xmath112 @xmath113 with the chosen parameters mentioned above , the contributions to the @xmath114 production cross section come from the color - octet @xmath115 , @xmath116 and @xmath117 are 0.18fb , 2.7fb and 0.87fb respectively .
the contribution of the color - octet @xmath108 is negligible .
the total color - octet contribution to the cross section from fig .
[ fey2 ] is 8.3fb , and the sum of the color - octet contributions from fig .
[ feynman ] and fig .
[ fey2 ] is 9.9fb .
this number could become substantially smaller if we use the matrix element values given in @xcite . in any case , the color - octet contribution to the d - wave states production cross sections should be of the same order as the color - singlet contribution , because from the @xmath85 fock state expansion in eq.([expandfock ] ) and from fig .
[ feynman ] and fig .
[ fey2 ] it is easy to see that the color - octet and color - singlet contributions are of the same order in both short distance ( @xmath118 ) and long distance ( @xmath119 ) parts .
based on the above calculations including both the color - singlet and color - octet contributions , we think @xmath120 should be a reasonable estimate in nrqcd for the @xmath4 production rate when the s - d mixing is neglected .
the calculated @xmath121 production rate could be significantly enhanced by the s - d mixing if the mixing angle is large . with the calculated @xmath51 production cross section @xmath91 @xcite ,
the @xmath52 production cross section can be approximately obtained by the scale factor @xmath122 , and then we have @xmath123 ( the color - octet contributions to the s - wave charmonia are small and negligible ) . if this estimate for the @xmath52 makes sense , with the large s - d mixing angle @xmath124 we would have @xmath125 this value is much larger than 10 fb , the value in eq.([dt ] ) obtained without s - d mixing in nrqcd . however , as we already mentioned , the belle observed double charm production cross section for the @xmath51 has a much higher value@xcite @xmath126 which is larger than the theoretical expectation in nrqcd by a factor of 6 ( as a rough estimate we neglect the feed down contribution from the @xmath52 and @xmath62 states in the @xmath51 production cross section ) .
if this also happens for the @xmath52 , which is very likely , we would expect @xmath127 to be the observed double charm production cross section for the @xmath52 . then with the large
s - d mixing angle @xmath128 we would have @xmath129 this is more than an order of magnitude larger than the value in eq.([dt ] ) obtained without s - d mixing in nrqcd . in any case
, we see that the large s - d mixing angle @xmath128 would result in a much higher cross section for the @xmath4 double charm production than that without s - d mixing or with small s - d mixing angle like @xmath18 .
this is due to the fact that the double charm production rate for a pure @xmath40 state is much higher than that for a pure @xmath39 state in nrqcd .
we therefore suggest measuring the production cross sections of @xmath130 and @xmath131 in the continuum @xmath5 annihilation at @xmath6 gev by belle and babar .
in contrast , in the @xmath3 inclusive decay to @xmath4 , the decay rate is expected to be insensitive to the s - d mixing in nrqcd , because both the @xmath40 and @xmath39 final states produced in @xmath3 decay are dominated by the color - octet intermediate states with the same quantum numbers in the two cases and thus they may have comparable production rates .
the color - octet dominance in @xmath3 decay relies on two observations .
the first is that in the effective weak interactions the squared short distance coefficient at the @xmath132 quark mass scale for the color - octet part is larger than that for the color - singlet part by more than an order of magnitude .
the second is that the color - singlet s - wave contribution is further suppressed by qcd radiative corrections and therefore negligible ( see @xcite@xcite@xcite for detailed discussions ) .
therefore , despite of certain uncertainties related to the values of color - octet matrix elements and other parameters , the qualitative features for the d - wave charmonium production in @xmath5 annihilation and @xmath3 decay should hold and be tested .
as mentioned already , the small s - d mixing angle like @xmath18 is favored by the observed e1 transition rates of @xmath38 . however , as another independent check for the mixing angle , measurements on the @xmath4 e1 transitions will be also very useful . in ref.@xcite
the e1 transition rates of @xmath4 were calculated for @xmath133 , and the s - d interference effects were found to be significant . based on the same potential model we now estimate the e1 transition widths for @xmath134 and find @xmath135 for j=0,1,2 with @xmath136 ; and @xmath137 for j=0,1,2 with @xmath138 .
we see that the s - d interference effects in the @xmath4 e1 transitions are essential . in particular , the ratio @xmath139 will dramatically change from @xmath8 0.04 ( for @xmath140 ) to @xmath8 200 ( for @xmath10 ) due to the large s - d interference effect .
similar discussions for @xmath141 and @xmath25 can also be found in @xcite .
we hope these measurements can be performed at bes and cleo - c in the near future
. they will be very helpful in clarifying the s - d mixing problem for @xmath4 and @xmath38 .
we first discuss about the uncertainties associated with the color - octet matrix elements with different choices from eqs.(13)-(15 ) . according to the nrqcd velocity scaling rules
, we may have @xmath142 if we use @xmath143 as the input parameter , @xmath144 , we would have the following matrix elements @xmath145 then the color - octet contributions to the cross section @xmath146 can be estimated to be 0.36 fb , 1.1 fb , 0.28 fb , 0.12 fb for @xmath147 color - octet intermediate states respectively from fig . 1 ; and 13.5 fb , 0.17 fb , 2.6 fb , 0.83 fb respectively from fig . 2 .
the total contribution to the cross section @xmath148 is about 20 fb , a factor of 2 larger than 10 fb , the value given in eq.(25 ) obtained by using eqs.(13)-(15 ) .
the above results may demonstrate the possible uncertainties associated with the color - octet matrix elements , which are estimated according to the nrqcd velocity scaling rules but with different choices for the input parameters .
we expect that these uncertainties do not change our main analysis that the @xmath4 production in @xmath5 annihilation receives much smaller contributions from the color - octet channels than from the large s - d mixing .
we suggest performing the measurements on the @xmath52 and @xmath4 production in the continuum @xmath5 annihilation by belle and babar in order to check ( i ) if the @xmath149 cross section is as large as about 0.5pb , which would confirm the s - wave ( not only for 1s but also for 2s ) charmonium production enhancement via double @xmath41 ; ( ii ) if the cross section of @xmath130 is comparable to that of @xmath149 , which would confirm the large s - d mixing angle ; ( iii ) if the cross section of @xmath130 is as small as say 10 fb , which would favor the prediction in nrqcd .
it is very likely that belle and babar will find a strong signal for @xmath149 , because experimentally belle has found a large cross section for the inclusive @xmath150 cross section , which is comparable to that for @xmath151 @xcite .
if a large @xmath149 cross section turned to be the case , then one can easily distinguish between the large s - d mixing and the nrqcd prediction ( together with the small s - d mixing ) by simultaneously measuring the @xmath130 cross section . in conclusion ,
we notice that the large decay rate observed by belle for @xmath0 , which is comparable to @xmath1 , might indicate either an unexpectedly large s - d mixing angle @xmath2 or the leading role of the color - octet mechanism in d - wave charmonium production in @xmath3 decay . by calculating the production rate of @xmath4 in the continuum @xmath5 annihilation at @xmath6 gev with these two possible approaches ( i.e. the large s - d mixing and the color - octet mechanism )
, we show that the measurement for this process at belle and babar may provide a clear cut clarification for the two approaches .
in addition , the radiative e1 transition ratio @xmath7 may dramatically change from @xmath8 0.04 ( for @xmath9 ) to @xmath8 200 ( for @xmath10 ) due to the large s - d interference effect , thus the e1 transition measurement of @xmath4 at bes and cleo - c will also be very useful in clarifying this issue .
we thank z.g.he and y.j.zhang for help in checking some of the calculations . this work was supported in part by the national natural science foundation of china , the education ministry of china , and the beijing electron positron collider national laboratory .
in this appendix , for @xmath65 and @xmath89 we give the expressions of @xmath152 and @xmath84 which are defined in eq .
( [ cross ] ) . @xmath153 + 15\delta^2(z-2)^4{\rm ln}[\frac{z\sqrt{4 + \delta^2 - 4z } + 2\sqrt{(1 - z)(z^2-\delta^2)}}{z\sqrt{4 + \delta^2 - 4z } - 2\sqrt{(1 - z)(z^2-\delta^2)}}]\nonumber\\ & \times&[23040\delta^{12 } - ( 176640\delta^{10 } + 69120\delta^{12})z + ( 519168\delta^8 + 585984\delta^{10 } \nonumber\\ & + & 89856\delta^12)z^2 - ( 608256\delta^6 + 1620480\delta^8 + 799296\delta^{10 } + 63936\delta^{12})z^3 \nonumber\\ & + & ( 442368\delta^4 + 1380864\delta^6 + 1912960\delta^8 + 578944\delta^{10 } + 26880\delta^{12})z^4 \nonumber\\ & -&(724992\delta^4 + 378112\delta^6 + 924480\delta^8 + 227248\delta^{10 } + 6144\delta^{12})z^5 \nonumber\\ & - & ( 1032192\delta^2 + 256000\delta^4 + 1437696\delta^6 + 22176\delta^8 - 46976\delta^{10 } - 1056\delta^{12})z^6 \nonumber\\ & + & ( 2310144\delta^2 + 503808\delta^4 + 1262528\delta^6 + 108816\delta^8 - 11916\delta^{10 } - 252\delta^{12})z^7 \nonumber\\ & -&(229376\delta^2 - 856576\delta^4 - 194944\delta^6 - 29976\delta^8 - 3592\delta^{10 } - 42\delta^{12})z^8 \nonumber\\ & + & ( 294912 - 3428352\delta^2 - 1397248\delta^4 - 571184\delta^6 - 25516\delta^8 - 447\delta^{10})z^9 \nonumber\\ & -&(1392640 - 4259840\delta^2 - 583680\delta^4 - 215072\delta^6 - 3078\delta^8 + 21\delta^{10})z^{10 } \nonumber\\ & + & ( 1998848 - 2720768\delta^2 + 20736\delta^4 - 16052\delta^6 + 345\delta^8)z^{11 } \nonumber\\ & - & ( 1163264 - 1031936\delta^2 + 98816\delta^4 + 3554\delta^6)z^{12 } + ( 284672 - 172800\delta^2 \nonumber\\ & + & 22288\delta^4)z^{13 } - ( 35840 - 5376\delta^2)z^{14 } + 4096z^{15}]\}.\end{aligned}\ ] ] @xmath154 \nonumber\\ & - & 15 \delta^2
( -2 + z)^4{\rm ln}[\frac{z\sqrt{4 + \delta^2 - 4z } + 2\sqrt{(1 - z)(z^2-\delta^2)}}{z\sqrt{4 + \delta^2 - 4z } - 2\sqrt{(1 - z)(z^2-\delta^2 ) } } ] \nonumber\\ & \times&[23040 \delta^{12 } - ( 192000 \delta^{10 } + 69120 \delta^{12 } ) z + ( 519168 \delta^8 + 507648 \delta^{10 } \nonumber\\ & + & 89856 \delta^{12 } ) z^2 - ( 202752 \delta^6 + 1337856 \delta^8 + 533952 \delta^{10 } + 63936 \delta^{12 } ) z^3 \nonumber\\ & + & ( 147456 \delta^4 - 463360 \delta^6 + 1415552 \delta^8 + 245120 \delta^{10 } + 26880 \delta^{12 } ) z^4 \nonumber\\ & - & ( 61440 \delta^4 - 2630912 \delta^6 + 731072 \delta^8 + 13136 \delta^{10 } + 6144 \delta^{12 } ) z^5 \nonumber\\ & - & ( 737280 \delta^2 + 219136 \delta^4 + 4452352 \delta^6 - 169760 \delta^8 + 38144 \delta^{10 } - 1056 \delta^{12 } ) z^6 \nonumber\\ & + & ( 835584 \delta^2 - 724992 \delta^4 + 3959872 \delta^6 - 2256 \delta^8 + 10380 \delta^{10 } - 252 \delta^{12 } ) z^7 \nonumber\\ & - & ( 1179648 - 3260416 \delta^2 - 4187648 \delta^4 + 1798528 \delta^6 - 46024 \delta^8 - 488 \delta^{10 } \nonumber\\ & - & 42 \delta^{12 } ) z^8 + ( 3244032 - 10629120 \delta^2 - 7306752 \delta^4 + 69936 \delta^6 - 32804 \delta^8 \nonumber\\ & - & 465 \delta^{10 } ) z^9 - ( 3031040 - 13373440 \delta^2 - 6077440 \delta^4 - 218592 \delta^6 - 7418 \delta^8 \nonumber\\ & - & 21 \delta^{10 } ) z^{10 } + ( 950272 - 8346624 \delta^2 - 2431744 \delta^4 - 61164
\delta^6 - 345 \delta^8 ) z^{11 } \nonumber\\ & + & ( 57344 + 2462976 \delta^2 + 428416 \delta^4 + 3554 \delta^6 ) z^{12 } - ( 63488 + 259328 \delta^2 \nonumber\\ & + & 22288 \delta^4 ) z^{13 } + ( 17408 - 5376 \delta^2 ) z^{14 } - 4096 z^{15}]\}.\end{aligned}\ ] ] @xmath155\nonumber\\ & + & [ 1843200 \delta^{12 } ( 4 + \delta^2 ) - 614400 \delta^{10 } ( 72 + 89 \delta^2 + 15 \delta^4 ) z \nonumber\\ & + & 61440 \delta^8 ( 1296 + 4228 \delta^2 + 2819 \delta^4 + 342 \delta^6 ) z^2\nonumber\\ & - & 15360 \delta^8 ( 25440 + 42856 \delta^2 + 20593 \delta^4 + 1869 \delta^6 ) z^3\nonumber\\ & + & 7680 \delta^6 ( -25344 + 104048 \delta^2 + 120884 \delta^4 + 48728 \delta^6 + 3391 \delta^8 ) z^4 \nonumber\\ & - & 3840 \delta^6 ( -283392 + 218448 \delta^2 + 197300 \delta^4 + 78807 \delta^6 + 4272 \delta^8 ) z^5 \nonumber\\ & + & 3840 \delta^4 ( 46080 - 725888 \delta^2 + 107264 \delta^4 + 74942 \delta^6 + 44495 \delta^8 + 1901 \delta^{10 } ) z^6 \nonumber\\ & - & 1920 \delta^4 ( 540672 - 2251264 \delta^2 + 11080 \delta^4 - 37482 \delta^6 + 35347 \delta^8 + 1209 \delta^{10 } ) z^7 \nonumber\\ & + & 480 \delta^2 ( -98304 + 5812224 \delta^2 - 9475072 \delta^4 + 31856 \delta^6 - 315944 \delta^8 + 40476 \delta^{10}\nonumber\\ & + & 1147 \delta^{12 } ) z^8 - 480 \delta^2 ( -540672 + 9385984 \delta^2 - 7280768 \delta^4 + 362224 \delta^6 \nonumber\\ & & - 179332 \delta^8 + 9230 \delta^{10 } + 229 \delta^{12 } ) z^9 + 240 \delta^2 ( -2670592 + 20128768 \delta^2 - 8820608 \delta^4\nonumber\\ & + & 831568 \delta^6 - 110696 \delta^8 + 3933 \delta^{10 } + 85 \delta^{12 } ) z^{10 } - 60 \delta^2
( -15597568 + 59650048 \delta^2\nonumber\\ & - & 18438144 \delta^4 + 1779904 \delta^6 - 84800 \delta^8 + 2921 \delta^{10 } + 49 \delta^{12 } ) z^{11 } + 30 \delta^2 ( -30146560 \nonumber\\ & + & 61106176 \delta^2 - 17407488 \delta^4 + 980688 \delta^6 - 26636 \delta^8 + 640 \delta^{10 } + 7 \delta^{12 } ) z^{12}\nonumber\\ & - & 15 \delta^2 ( -41222144 + 41398272 \delta^2 - 13922048 \delta^4 + 210896 \delta^6 - 9404 \delta^8 + 19 \delta^{10 } ) z^{13}\nonumber\\ & - & 15 \delta^2 ( 20840448 - 7954432 \delta^2 + 4165888 \delta^4 + 19040 \delta^6 + 1242 \delta^8 + 7 \delta^{10 } ) z^{14}\nonumber\\ & + & 15 \delta^2 ( 7962624 - 230400 \delta^2 + 829440 \delta^4 + 6508
\delta^6 + 69 \delta^8 ) z^{15 } \nonumber\\ & - & 30 \delta^2 ( 1140736 + 139136 \delta^2 + 48432 \delta^4 + 219 \delta^6 ) z^{16 } + 240 \delta^2 ( 29312 + 4064 \delta^2 \nonumber\\ & + & 321 \delta^4 ) z^{17 } - 1920 \delta^2 ( 488 + 41 \delta^2 ) z^{18 } + 61440 \delta^2 z^{19}]\nonumber\\ & \times&\ln\frac{z\sqrt{4 + \delta^2 - 4 z } + 2 \sqrt{(1 - z)(-\delta^2 +
z^2 ) } } { z\sqrt{4 + \delta^2 - 4 z } - 2 \sqrt{(1 - z)(-\delta^2 + z^2)}}\}.\end{aligned}\ ] ] @xmath156 - [ 614400 \delta^{12 } ( 4 + 3
\delta^2 ) - 614400 \delta^{10 } ( 24 + 31 \delta^2 + 15 \delta^4 ) z \nonumber\\ & + & 61440 \delta^8 ( 432 + 1604 \delta^2 + 909 \delta^4 + 342 \delta^6 ) z^2 - 15360 \delta^8 ( 7968 + 19112 \delta^2 \nonumber\\ & + & 5535 \delta^4 + 1869 \delta^6 ) z^3 + 7680 \delta^6 ( -11520 + 21648 \delta^2 + 68348 \delta^4 + 9120 \delta^6 \nonumber\\ & + & 3391 \delta^8 ) z^4 - 3840 \delta^6 ( -128256 - 39504 \delta^2 + 165980 \delta^4 + 5417 \delta^6 + 4272 \delta^8 ) z^5 \nonumber\\ & + & 3840 \delta^4 ( 39936 - 299136 \delta^2 - 227520 \delta^4 + 146914 \delta^6 - 4783 \delta^8 + 1901 \delta^{10 } ) z^6\nonumber\\ & - & 1920 \delta^4
( 491520 - 740864 \delta^2 - 794120 \delta^4 + 198970 \delta^6 - 13267 \delta^8 + 1209 \delta^{10 } ) z^7 \nonumber\\ & + & 480 \delta^2 ( -98304 + 5386240 \delta^2 - 1953792 \delta^4 - 3425904 \delta^6 + 424008 \delta^8 - 30108 \delta^{10}\nonumber\\ & + & 1147 \delta^{12 } ) z^8 - 480 \delta^2 ( -442368 + 8620032 \delta^2 - 530304 \delta^4 - 2668784 \delta^6 + 175028 \delta^8\nonumber\\ & - & 9350 \delta^{10 } + 229 \delta^{12 } ) z^9 + 240 \delta^2 ( -1753088 + 17685504 \delta^2 - 52352 \delta^4 - 3219728 \delta^6\nonumber\\ & + & 102032 \delta^8 - 2797 \delta^{10 } + 85 \delta^12 ) z^{10 } - 60 \delta^2 ( -7995392
+ 48058368 \delta^2 - 1542144 \delta^4\nonumber\\ & - & 6071104 \delta^6 + 65680 \delta^8 + 135 \delta^{10 } + 49 \delta^{12 } ) z^{11 } + 30 \delta^2 ( -11141120 + 43489280 \delta^2\nonumber\\ & - & 4092416
\delta^4 - 4255696 \delta^6 + 700 \delta^8 + 632 \delta^{10 } + 7 \delta^{12 } ) z^{12 } - 15 \delta^2 ( -8323072\nonumber\\ & + & 26431488 \delta^2 - 3950848 \delta^4 - 2012880 \delta^6 - 5492 \delta^8 + 173 \delta^{10 } ) z^{13 } + 15 \delta^2 ( 196608\nonumber\\ & + & 5984256 \delta^2 - 673536 \delta^4 - 272864
\delta^6 - 182 \delta^8 + 7 \delta^{10 } ) z^{14 } - 15 \delta^2 ( 2064384\nonumber\\ & + & 1408000 \delta^2 + 93184 \delta^4 - 13620 \delta^6 + 69 \delta^8 ) z^{15 } + 30 \delta^2 ( 587776 + 184704 \delta^2\nonumber\\ & + & 24944 \delta^4 + 219 \delta^6 ) z^{16 } - 240 \delta^2 ( 21632
+ 4192 \delta^2 + 321 \delta^4 ) z^{17 } \nonumber\\ & + & 1920 \delta^2 ( 440 + 41 \delta^2 ) z^{18 } - 61440 \delta^2 z^{19}]\nonumber\\ & \times&\ln\frac{z\sqrt{4 + \delta^2 - 4 z } + 2 \sqrt{(1 - z)(-\delta^2 + z^2 ) } } { z\sqrt{4 + \delta^2 - 4 z } - 2 \sqrt{(1 - z)(-\delta^2 + z^2)}}\}.\end{aligned}\ ] ] m.beneke and m.krmer , phys . rev .
* d55 * , ( 1997)5269 .
g.t bodwin et al . , phys
d46 , 3703(1992 ) . l.bergstrom and p. ernstrom , phys .
b328 , 153(1994 ) .
m. beneke et al . , phys .
d59 , 054003(1999 ) . | the large decay rate observed by belle for @xmath0 , which is comparable to @xmath1 , might indicate either an unexpectedly large s - d mixing angle @xmath2 or the leading role of the color - octet mechanism in d - wave charmonium production in @xmath3 decay . by calculating the production rate of @xmath4 in the continuum @xmath5 annihilation at @xmath6 gev with these two possible approaches ( i.e. the large s - d mixing and the color - octet mechanism ) , we show that the measurement for this process at belle and babar may provide a clear cut clarification for the two approaches .
in addition , the radiative e1 transition ratio @xmath7 may dramatically change from @xmath8 0.04 ( for @xmath9 ) to @xmath8 200 ( for @xmath10 ) due to the large s - d interference effect , thus the e1 transition measurement of @xmath4 at bes and cleo - c will also be very useful in clarifying this issue .
pacs number(s ) : 13.66.bc , 14.40.gx , 12.38.bx s - d mixing and @xmath4 production in @xmath5 annihilation and b decay and its radiative transitions + kui - yong liu department of physics , peking university , beijing 100871 , people s republic of china and department of physics , liaoning university , shenyang 110036 , people s republic of china + kuang - ta chao china center of advanced science and technology ( world laboratory ) , beijing 100080 , people s republic of china and department of physics , peking university , beijing 100871 , people s republic of china |
for the general population of vvds galaxies we find @xmath5 roughly constant with a low value of @xmath7 @xmath8 mpc for @xmath9 $ ] , to @xmath10 @xmath8 mpc for @xmath11 $ ] .
this - seemingly surprising - lack of evolution may be understood when we take into account that we are dealing with a mixture of galaxies of different types and luminosities at different redshifts . at low z
we observe mainly a low - luminosity , weakly clustered population . at high z , where we theoretically expect to observe galaxies
to be less clustered than today , we probe only the most luminous , most massive and most clustered objects .
luminosity - dependence of galaxy clustering at @xmath12 is clearly different with what is observed in the nearby universe .
the clustering strength is rising around the characteristic schechter luminosity @xmath13 , with a sharper turn than observed at low redshifts .
additionally , the slope of the correlation function steepens from @xmath14 to @xmath15 .
this is caused by a significant change in the shape of @xmath4 , increasingly deviating from a power - law for the most luminous samples , with a strong upturn at small ( @xmath16 @xmath8 mpc ) scales ( see left panel of figure [ pol1:fig1 ] ) .
this trend , not observed locally , results also in a visible scale - dependence of the relative bias , @xmath17 ( shown in the right panel of figure [ pol1:fig1 ] ) and seems to imply a significant change in the way luminous galaxies trace dark - matter halos at @xmath18 with respect to the present - day universe .
as shown in the left panel of figure [ pol1:fig2 ] , the clustering of galaxies of early spectral types is stronger than that of late - type galaxies at all redshifts up to @xmath19 . at @xmath20
, early - type galaxies display a correlation length @xmath210.9 h@xmath22mpc , while late types have @xmath230.4 h@xmath22mpc .
the relative bias between early- and late - type galaxies remains approximately constant with @xmath24 from @xmath25 to @xmath26 .
similarly , as shown in the right panel of figure [ pol1:fig2 ] , red - sequence galaxies exhibit a larger clustering length than the blue ones with little dependence on redshift .
these results @xcite are consistent with a scenario where the assembly of the most luminous early - type galaxies is practically completed at @xmath12 , with however luminous star - forming objects being nearly as clustered as red galaxies for @xmath27 larger than @xmath28 .
all the above - mentioned measurements are pieces of a puzzle which still have to be fit into the theoretical scheme of gravitational instability scenario , where dark matter halos assemble under the effect of gravity and galaxies form and evolve inside them .
our observations represent an important constraint for models trying to reproduce galaxy clustering , especially at small , non - linear scales .
this research has been developed within the framework of the vvds consortium .
this work has been partially supported by the cnrs - insu and its programme national de cosmologie ( france ) , and by italian ministry ( miur ) grants cofin2000 ( mm02037133 ) and cofin2003 ( num.2003020150 ) .
the vlt - vimos observations have been carried out on guaranteed time ( gto ) allocated by the european southern observatory ( eso ) to the virmos consortium , under a contractual agreement between the centre national de la recherche scientifique of france , heading a consortium of french and italian institutes , and eso , to design , manufacture and test the vimos instrument .
le fvre , o. vettolani , g. , garilli , b. et al . , 2005 , a&a , 439 , 845 le fvre , guzzo , l. , meneux , b. et al . , 2005 ,
a&a , 439 , 877 meneux , b. , guzzo , l. , le fvre et al . , 2006 , a&a , 452 , 387 norberg , p. , baugh , c.m . ,
haekins , e. et al . , 2002 , , 332 , 827 pollo , a. , meneux , a. guzzo , l. , et al . , 2005 , a&a , 439 , 887 pollo , a. , guzzo , l. , le fvre , o. , et al . , 2006 ,
a&a , 451 , 409 white , s.d.m . & rees , m.j . , 1978 , mnras , 183 , 341 zehavi , i. , zheng , z. , weinberg , d.h .
et al . , 2005 ,
apj , 630 , | we discuss the evolution of clustering of galaxies in the universe from the present epoch back to @xmath0 , using the first - epoch data from the vimos - vlt deep survey ( vvds ) .
we present the evolution of the projected two - point correlation function of galaxies for the global galaxy population , as well as its dependence on galaxy intrinsic luminosities and spectral types . while we do not find strong variations of the correlation function parameters with redshift for the global galaxy population , the clustering of objects with different intrinsic luminosities evolved significantly during last 8 - 10 billion years .
our findings indicate that bright galaxies in the past traced higher density peaks than they do now and that the shape of the correlation function of most luminous galaxies is different from observed for their local counterparts , which is a supporting evidence of a non - trivial evolution of the galaxy vs. dark matter bias . according to the current paradigm
, galaxies formed and evolved inside dark matter halos , which merged and grew under the effect of gravity @xcite .
understanding the galaxy clustering may be the key to understand the evolution of the universe itself .
it is therefore an important issue of contemporary cosmology to follow the evolution of galaxy clustering in the past and to understand the bias between the galaxy distribution and the underlying dark matter density field , and how it depends on galaxy types , luminosities and local environment .
we explored these issuees using the unique set of data from the vimos - vlt deep survey , which is strictly selected in magnitude in the range @xmath1 , from a complete deep photometric survey , without any color selection @xcite .
the first - epoch vvds data contain in total 11564 galaxies with measured spectroscopic redshifts up to @xmath2 , making it an ideal tool to investigate the galaxy clustering in a broad range of time .
here we describe briefly our results on the evolution of the clustering properties of the general vvds galaxy population and the dependence of clustering on galaxy luminosities , types and colors .
we measure the galaxy 2-point spatial correlation function @xmath3 and project it along the line of sight to obtainthe function @xmath4 .
the best power - law fit to @xmath4 gives us the correlation length @xmath5 and the slope of a correlation function , @xmath6 ( details in * ? ? ?
* ) . |
the problem of fading and the ways to combat it through spatial diversity techniques have been an active area of research .
multiple - input multiple - output ( mimo ) techniques have become popular in realizing spatial diversity and high data rates through the use of multiple transmit antennas . for such co - located multiple transmit antenna systems low maximum - likelihood ( ml ) decoding complexity space - time block codes ( stbcs )
have been studied by several researchers @xcite-@xcite which include the well known complex orthogonal designs ( cods ) and their generalizations .
recent research has shown that the advantages of spatial diversity could be realized in single - antenna user nodes through user cooperation @xcite,@xcite via relaying .
a simple wireless relay network of @xmath3 nodes consists of a single source - destination pair with @xmath0 relays . for such relay channels , use of cods
@xcite,@xcite has been studied in @xcite .
cods are attractive for cooperative communications for the following reasons : @xmath4 they offer full diversity gain and coding gain , @xmath5 they are ` scale free ' in the sense that deleting some rows does not affect the orthogonality , @xmath6 entries are linear combination of the information symbols and their conjugates which means only linear processing is required at the relays , and @xmath7 they admit very fast ml decoding ( single - symbol decoding ( ssd ) ) .
however , it should be noted that the last property applies only to the decode - and - forward ( df ) policy at the relay node . in a scenario where the relays amplify and forward ( af ) the signal
, it is known that the orthogonality is lost , and hence the destination has to use a complex multi - symbol ml decoding or sphere decoding @xcite,@xcite .
it should be noted that the af policy is attractive for two reasons : @xmath4 the complexity at the relay is greatly reduced , and @xmath5 the restrictions on the rate because the relay has to decode is avoided @xcite . in order to avoid the complex ml decoding at the destination , in @xcite ,
the authors propose an alternative code design strategy and propose a ssd code for 2 and 4 relays .
for arbitrary number of relays , recently in @xcite , distributed orthogonal stbcs ( dostbcs ) have been studied and it is shown that if the destination has the complete channel state information ( csi ) of all the source - to - relay channels and the relay - to - destination channels , then the maximum possible rate is upper bounded by @xmath1 complex symbols per channel use for @xmath0 relays . towards improving the rate of transmission and achieving simultaneously both full - diversity as well as ssd at the destination , in this paper , we study relay channels with the assumption that the relays have the phase information of the source - to - relay channels and the destination has the csi of all the channels .
coding for partially - coherent relay channel ( pcrc , section [ pcrc_sec ] ) has been studied in @xcite , where a sufficient condition for ssd has been presented .
the contributions of this paper can be summarized as follows : * first , a new set of necessary and sufficient conditions for a stbc to be ssd for co - located multiple antenna communication is obtained .
the known set of necessary and sufficient conditions in @xcite is in terms of the dispersion matrices ( weight matrices ) of the code , whereas our new set of conditions is in terms of the column vector representation matrices @xcite of the code and is a generalization of the conditions given in @xcite in terms of column vector representation matrices for cods . * a set of necessary and sufficient conditions for a distributed stbc ( dstbc ) to be ssd for a pcrc is obtained by identifying the additional conditions . using this , several ssd dstbcs for pcrc
are identified among the known classes of stbcs for co - located multiple antenna system . *
it is proved that even if a ssd stbc for a co - located mimo channel does not satisfy the additional conditions for the code to be ssd for a pcrc , single - symbol decoding of it in a pcrc gives full - diversity and only coding gain is lost . *
it is shown that when a dstbc is ssd for a pcrc , then arbitrary coordinate interleaving of the in - phase and quadrature - phase components of the variables does not disturb its ssd property for pcrc . *
it is shown that the possibility of _ channel phase compensation _ operation at the relay nodes using partial csi at the relays increases the possible rate of ssd dstbcs from @xmath1 when the relays do not have csi to @xmath8 which is independent of @xmath0 . *
extensive simulation results are presented to illustrate the above contributions .
the remaining part of the paper is organized as follows : in section [ sec2 ] , the signal model for a pcrc is developed . using this model , in section [ sec3 ] , a new set of necessary and sufficient conditions for a stbc to be ssd in a co -
located mimo is presented .
several classes of ssd codes are discussed and conditions for full - diversity of a subclass of ssd codes is obtained .
then , in section [ sec4 ] , ssd dstbcs for pcrc are characterized by identifying a set of necessary and sufficient conditions .
it is shown that the ssd property is invariant under coordinate interleaving operations which leads to a class of ssd dstbcs for pcrc . the class of rate half cods obtained from rate one real orthogonal designs ( rods ) by stacking construction @xcite is shown to be ssd for pcrc .
also , it is shown that ssd codes for co - located mimo , under suboptimal ssd decoder for pcrc offer full diversity .
simulation results and discussion constitute section [ sec5 ] .
conclusions and scope for further work are presented in section [ sec6 ] .
consider a wireless network with @xmath3 nodes consisting of a source , a destination , and @xmath0 relays , as shown in fig .
all nodes are half - duplex nodes , i.e. , a node can either transmit or receive at a time on a specific frequency .
we consider amplify - and - forward ( af ) transmission at the relays .
transmission from the source to the destination is carried out in two phases . in the first phase ,
the source transmits information symbols @xmath9 in @xmath10 time slots .
all the @xmath0 relays receive these @xmath10 symbols .
this phase is called the _
broadcast phase_. in the second phase , all the @xmath0 relays relays participate in the cooperative transmission .
it is also possible that some relays do not participate in the transmission based on whether the channel is in outage or not .
we do not consider such a partial participation scenario here .
] perform distributed space - time block encoding on their received vectors and transmit the resulting encoded vectors in @xmath11 time slots .
that is , each relay will transmit a column ( with @xmath11 entries ) of a distributed stbc matrix of size @xmath12 .
the destination receives a faded and noise added version of this matrix .
this phase is called the _ relay phase_.
we assume that the source - to - relay channels remain static over @xmath10 time slots , and the relay - to - destination channels remain static over @xmath11 time slots .
the received signal at the @xmath13th relay , @xmath14 , in the @xmath15th time slot , @xmath16 , denoted by @xmath17 , can be written as and @xmath18 denote transpose and conjugate transpose operations , respectively and @xmath19 denotes matrix conjugation operation . ]
@xmath20 where @xmath21 is the complex channel gain from the source @xmath22 to the @xmath13th relay , @xmath23 is additive white gaussian noise at relay @xmath13 with zero mean and unit variance , @xmath24 is the transmit energy per symbol in the broadcast phase , and @xmath25 = 1 $ ] . but
no channel knowledge is assumed at the relays . under the assumption of no csi at the relays ,
the amplified @xmath15th received signal at the @xmath13th relay can be written as @xcite @xmath26 where @xmath27 is the transmit energy per transmission of a symbol in the relay phase , and @xmath28 is the amplification factor at the relay that makes the total transmission energy per symbol in the relay phase to be equal to @xmath27 .
let @xmath29 denote the total energy per symbol in both the phases put together .
then , it is shown in @xcite that the optimum energy allocation that maximizes the receive snr at the destination is when half the energy is spent in the broadcast phase and the remaining half in the relay phase when the time allocations for the relay and broadcast phase are same i.e. , @xmath30 .
we also assume that the energy is distributed equally i.e. , @xmath31 and @xmath32 , where @xmath33 is the number of transmissions per symbol in the stbc .
for the unequal - time allocation ( @xmath34 ) this distribution might not be optimal . at relay @xmath13 , a @xmath35 real vector @xmath36 given by @xmath37^t,\end{aligned}\ ] ]
is formed , where @xmath38 and @xmath39 , respectively , are the in - phase ( real part ) and quadrature ( imaginary part ) components of @xmath40 now , ( [ vhatx ] ) can be written in the form @xmath41 where @xmath42 is the @xmath35 data symbol real vector , given by @xmath43^t,\end{aligned}\ ] ] @xmath44 is the @xmath35 noise vector , given by @xmath45^t,\end{aligned}\ ] ] where @xmath46 , and @xmath47 is a @xmath48 block - diagonal matrix , given by @xmath49 & \cdots & { \bf 0 } \\
\vdots & \ddots & \vdots \\ { \bf 0 } & \cdots & \left [ \begin{array}{cc } h_{sji } & -h_{sjq } \\ h_{sjq } & h_{sji } \end{array}\right ] \end{array}\right ] .\end{aligned}\ ] ] let @xmath50\end{aligned}\ ] ] denote the @xmath12 distributed stbc matrix to be sent in the relay phase jointly by all @xmath0 relays , where @xmath51 denotes the @xmath13th column of @xmath52 .
the @xmath13th column @xmath51 is manufactured by the @xmath13th relay as @xmath53 where @xmath54 is the complex processing matrix of size @xmath55 for the @xmath13th relay , called the _ relay matrix _ and
@xmath56 can be viewed as the column vector representation matrix @xcite for the distributed stbc with the difference that in our case the vector @xmath42 is real whereas in @xcite it is complex .
for example , for the 2-relay case ( i.e. , @xmath57 ) , with @xmath58 using alamouti code , the relay matrices are given by @xmath59 & \mbox { and } & { \bf a}_2 = \left [ \begin{array}{cccc } 0 & 0 & 1 & { \bf j } \\ 1 & -{\bf j } & 0 & 0 \end{array } \right ] .\end{aligned}\ ] ] let @xmath60 denote the @xmath61 received signal vector at the destination in @xmath11 time slots .
then , @xmath60 can be written as @xmath62 where @xmath63 is the complex channel gain from the @xmath13th relay to the destination , and @xmath64 is the awgn noise vector at the destination with zero mean and @xmath65={\bf i}$ ] . substituting ( [ stackx ] ) in ( [ rx1_no_csi ] ) ,
we can write @xmath66 in this subsection , we obtain a signal model for the case of partial csi at the relays , where we assume that each relay has the knowledge of the channel phase on the link between the source and itself in the broadcast phase . that is , defining the channel gain from source to relay @xmath13 as @xmath67
, we assume that relay @xmath13 has perfect knowledge of only @xmath68 and does not have the knowledge of @xmath69 in the proposed scheme , we perform a phase compensation operation on the amplified received signals at the relays , and space - time encoding is done on these phase - compensated signals .
that is , we multiply @xmath70 in ( [ no_comp ] ) by @xmath71 before space - time encoding .
note that multiplication by @xmath71 does not change the statistics of @xmath72 .
therefore , with this phase compensation , the @xmath73 vector in ( [ vhat2x ] ) becomes @xmath74 consequently , the @xmath51 vector generated by relay @xmath13 is given by @xmath75 where @xmath76 is the equivalent weight matrix with phase compensation .
now , we can write the received vector @xmath77 as @xmath78 figure [ fig2 ] shows the processing at the @xmath13th relay in the proposed phase compensation scheme . such systems will be referred as _ partially - coherent relay channels _ ( pcrc ) . a distributed stbc which is ssd for a pcrc
will be referred as ssd - dstbc - pcrc .
the class of ssd codes , including the well known cods , for co - located mimo has been studied in @xcite , where a set of necessary and sufficient conditions for an arbitrary linear stbc to be ssd has been obtained in terms of the dispersion matrices @xcite , also known as weight matrices . in this section , a new set of necessary and sufficient conditions in terms of the column vector representation matrices @xcite of the code
is obtained that are amenable for extension to pcrc .
this is a generalization of the conditions given in @xcite in terms of column vector representation matrices for cods . towards this end , the received vector @xmath60 in a co -
located mimo setup can be written as @xmath79 for co - located mimo with @xmath0 transmit antennas , the linear stbc as given in ( [ rx_colocate ] ) is ssd _ iff _ @xmath80 where @xmath81 , where @xmath82 and @xmath83 are real matrices , and @xmath84 and @xmath85 are block diagonal matrices of the form @xmath86 } _ { { \bf d}_{ij,1}^{(k ) } } & { \bf 0 } & \cdots & { \bf 0 } \\ { \bf 0 } & \underbrace { \left [ \begin{array}{cc } a_{ij,2}^{(k ) } & b_{ij,2}^{(k ) } \\
b_{ij,2}^{(k ) } & c_{ij,2}^{(k ) } \end{array } \right ] } _ { { \bf d}_{ij,2}^{(k ) } } & \cdots & { \bf 0 } \\ \vdots & \vdots & \ddots & \vdots \\ { \bf 0 } & \cdots & \cdots & \underbrace { \left [ \begin{array}{cc } a_{ij , t_1}^{(k ) } & b_{ij , t_1}^{(k ) } \\ b_{ij ,
t_1}^{(k ) } & c_{ij , t_1}^{(k ) } \end{array}\right ] } _ { { \bf d}_{ij , t_1}^{(k ) } } \end{array}\right ] , \end{aligned}\ ] ] where it is understood that whenever the superscript is ( 1 ) as in @xmath87 then @xmath88 _ proof : _ in ( [ rx2_no_csi ] ) , let @xmath89 .
then the ml optimal detection of @xmath42 is given by latexmath:[\ ] ]
we summarize the conclusions in this paper and future work as follows .
amplify - and - forward ( af ) schemes in cooperative communications are attractive because of their simplicity .
full diversity ( fd ) , linear - complexity single symbol decoding ( ssd ) , and high rates of dstbcs are three important attributes to work towards af cooperative communications . earlier work in @xcite has shown that , without assuming phase knowledge at the relays , fd and ssd can be achieved in af distributed orthogonal stbc schemes ; however , the rate achieved decreases linearly with the number of relays @xmath0 .
our work in this paper established that if phase knowledge is exploited at the relays in the way we have proposed , then fd , ssd , and high rate can be achieved simultaneously ; in particular , the rate achieved in our scheme can be @xmath2 , which is independent of the number of relays @xmath0 .
we proved the ssd for our scheme in theorem 2 .
fd was proved in theorem 6 .
rate-1/2 construction for any @xmath0 was presented in theorem 5 .
in addition to these results , we also established other results regarding @xmath4 invariance of ssd under coordinate interleaving ( theorem 4 ) , and @xmath5 retention of fd even with single - symbol non - ml decoding .
simulation results confirming the claims were presented .
all these important results have not been shown in the literature so far .
these results offer useful insights and knowledge for the designers of future cooperative communication based systems ( e.g. , cooperative communication ideas are being considered in future evolution of standards like ieee 802.16 ) . in this work ,
we have assumed only phase knowledge at the relays .
of course , one can assume that both amplitude as well as the phase of source - to - relay are known at the relay .
a natural question that can arise then is ` what can amplitude knowledge at the relay ( in addition to phase knowledge ) buy ? ' since we have shown that phase knowledge alone is adequate to achieve fd , some extra coding gain may be possible with amplitude knowledge .
this aspect of the problem is beyond the scope of this paper ; but it is a valid topic for future work .
v. tarokh , h. jafarkhani , and a. r. calderbank , `` space - time block codes from orthogonal designs , '' _ ieee trans .
theory , _ vol .
1456 - 1467 , july 1999 . o. tirkkonen and a. hottinen , `` square matrix embeddable stbc for complex signal constellations space - time block codes from orthogonal design , '' _ ieee trans .
theory , _ vol .
384 - 395 , february 2002 .
w. su and x .- g .
xia , `` signal constellations for quasi - orthogonal space - time block codes with full diversity , '' _ ieee trans .
theory , _ vol .
2331 - 2347 , october 2004 . c. yuen , y. l. guan , and t. t. tjhung ,
a class of four - group quasi - orthogonal space - time block code achieving full rate and full diversity for any number of antennas , " _ proc .
ieee pimrc2005 , _ vol .
1 , pp . 92 - 96 , september 2005 .
d. n. dao , c. yuen , c. tellambura , y. l. guan , and t. t. tjhung , `` four - group decodable space - time block codes , '' available on line in arxiv:0707.3959v1 [ cs.it ] , 26 july 2007 . also to appear in
_ ieee trans . on signal processing . | space - time block codes ( stbcs ) that are single - symbol decodable ( ssd ) in a co - located multiple antenna setting need not be ssd in a distributed cooperative communication setting .
a relay network with @xmath0 relays and a single source - destination pair is called a partially - coherent relay channel ( pcrc ) if the destination has perfect channel state information ( csi ) of all the channels and the relays have only the phase information of the source - to - relay channels . in this paper , first , a new set of necessary and sufficient conditions for a stbc to be ssd for co - located multiple antenna communication is obtained .
then , this is extended to a set of necessary and sufficient conditions for a distributed stbc ( dstbc ) to be ssd for a pcrc , by identifying the additional conditions . using this , several ssd dstbcs for pcrc
are identified among the known classes of stbcs .
it is proved that even if a ssd stbc for a co - located mimo channel does not satisfy the additional conditions for the code to be ssd for a pcrc , single - symbol decoding of it in a pcrc gives full - diversity and only coding gain is lost .
it is shown that when a dstbc is ssd for a pcrc , then arbitrary coordinate interleaving of the in - phase and quadrature - phase components of the variables does not disturb its ssd property for pcrc .
finally , it is shown that the possibility of _ channel phase compensation _ operation at the relay nodes using partial csi at the relays increases the possible rate of ssd dstbcs from @xmath1 when the relays do not have csi to @xmath2 , which is independent of @xmath0 . 0.25 in 2.00pc 1.85pc |
string theory suggests that our universe is not actually four - dimensional , but in fact a submanifold ( brane ) embedded into a higher - dimensional spacetime ( bulk ) .
braneworld gravity and cosmology , especially based on the proposal by randall and sundrum @xcite , has developed a lot in the literature in the codimension one context @xcite .
string theory ; however , suggests that there are as many as six or seven extra dimensions and thus , one may consider braneworld models with higher codimensions .
recently , codimension two braneworld has been investigated eagerly , because it may give a more _ natural _ resolution of the cosmological constant problem than the codimension one case does , namely the vacuum energy of the brane may affect only the extra dimensions , not the geometry on the brane , see e.g. , @xcite . the basic nature of gravity and cosmology in codimension two braneworld has been reviewed , e.g. , in ref .
@xcite ( see also references therein ) .
the stability of the extra dimensions is one of the most significant issues in braneworld models with two branes .
the interbrane distance between the branes appears as a scalar degree of freedom in the four - dimensional effective theory and affects the geometry and cosmology on the brane .
for instance , in rs - type codimension one brane models , this modulus , called the radion , behaves as a scalar degree of freedom in an effective brans - dicke gravity @xcite .
several stabilization mechanisms have been discussed in the rs model , both by classical dynamics @xcite and quantum corrections to the bulk vacuum state @xcite .
stability of the extra dimensions is also a significant issue in higher codimensions . in this paper
, we focus on the issue of quantum mechanical stabilization of the modulus in codimension two braneworld .
various codimension two braneworld solutions with two branes have been found , where the compact extra dimensions are supported by the dynamics of flux fields @xcite .
solutions with warped compact extra dimensions have also been found @xcite , especially in the context of six - dimensional supergravity theory @xcite . in this paper
, we focus on the solution discussed by aghababaie et al . in @xcite .
similar warped solutions have been recently discussed in @xcite , in the context of six - dimensional einstein - maxwell theory . in these two - brane models
, the magnetic flux plays an essential role in order to obtain regular warped solutions . by
regular " we mean that singularities which are stronger than conical ones are not permitted .
the branes are situated at the conical defects , whose tensions are determined by the conical deficit angles . in warped brane models
the conical defects correspond to the horizons of the bulk geometry and thus , the deficit angles depend on the parameters of the bulk geometry , e.g. , the mass and charge . by the tension - deficit relations ,
if the brane tensions are fixed , then part of bulk geometry is also fixed . in the models discussed in @xcite ,
the bulk geometry is also completely determined .
however , in the model we shall consider here @xcite , which is based on six - dimensional supergravity , only the warp factor is fixed by the classical analysis .
thus , to fix the size of the extra dimensions , quantum corrections of the bulk field should be taken into account . in this paper
, we consider such a supersymmetric warped flux compactification model .
we shall consider the perturbations of a massless scalar field on this brane background and calculate the effective potential of the volume modulus , focusing on whether or not one - loop quantum corrections can stabilize the absolute size of the extra dimensions . due to the scale invariance of the background field theory ,
the form of the modulus effective potential is almost completely fixed and therefore we only need to evaluate the sign of the coefficient of the logarithmic term in the effective potential to investigate the possibility of volume stabilization .
this is directly related to the appropriate heat kernel coefficient , which is composed of contributions from both the bulk and the conical branes .
the paper is organized as follows : in sec .
ii , we present a warped flux compactification model that is based on an einstein - maxwell - dilaton theory with a non - vanishing scalar field potential . in sec .
iii , we consider the perturbation of a massless , minimally coupled bulk scalar field in order to derive the modulus effective potential and then investigate whether or not the effective potential has a minimum . as a result , the stability of the modulus is determined by the appropriate heat kernel coefficient . then , we present results for our analysis of the heat kernel coefficients and we also discuss the _ rigidity _ of the stabilization , i.e. , is the modulus mass stable against kk perturbations . in the actual investigation of stability , rather than the original six - dimensional model ,
we consider a four - dimensional counterpart , because of the difficulties associated with evaluating the uv contributions from the conical branes .
then , we make some suggestions for the original six - dimensional model . in sec v
, we discuss phenomenological implications relating to the hierarchy of the fundamental energy scales and the effective vacuum energy realized on the brane . in sec .
vi , we summarize this paper and discuss possible extensions of our present analyses . in appendix
a , we introduce the four - dimensional version of the warped flux compactification solution and consider the nature of the massless scalar field perturbations on such a background , including quantum effects . in appendix b , we show the conformal invariance of the @xmath0 heat kernel coefficient . in appendix c
we use the conformal invariance of the @xmath1 heat kernel coefficient on the cone to find the smeared coefficient @xmath2 , which is required for the cocycle function . in appendix
d we apply the wkb method to give an estimate for the zeta function in four dimensions , which contributes to the mass scale of the modulus .
we consider a six - dimensional einstein - maxwell - dilaton theory with a non - vanishing scalar potential @xcite as @xmath3 where @xmath4 is a dilaton field , @xmath5 represents a @xmath6 gauge field strength and @xmath7 represents the dilaton potential .
this theory corresponds to the bosonic part of the salam - sezgin , six - dimensional gauged supergravity theory @xcite with vanishing kalb - ramond 2-form and chiral scalar fields .
hereafter we set @xmath8 for simplicity and if needed , we put it back in explicitly . the theory given in eq .
( [ theory6d ] ) has a dilatonic , charged black hole solution @xcite with metric @xmath9 where @xmath10 and @xmath11 are integration constants . at this stage
it will be convenient to perform a _
double _ wick rotation , @xmath12 , @xmath13 with the reparameterizations @xmath14 , @xmath15 @xcite .
this leads to @xmath16 in this scenario , the branes are located at positions which are determined by the _ horizon _
condition @xmath17 ; @xmath18 we can then rewrite @xmath19 and whence we also obtain the following useful relations @xmath20 these give a direct relation between the mass and charge of the magnetic flux . the global period of @xmath21
is called @xmath22 , which is determined by the brane tensions through the tension - deficit relations as will be discussed later .
note that for the @xmath23 limit , we obtain a single brane solution with a naked curvature singularity in the bulk . by changing coordinates to @xmath24 eq .
( [ bulkmetricori ] ) becomes @xmath25\ , .
\label{bulkmetric}\end{aligned}\ ] ] thus , there are two cones at @xmath26 , whose deficit angles are given by @xmath27 respectively . once the brane tensions are fixed , then @xmath28 is fixed .
but the absolute size of the extra dimensions , i.e. , @xmath29 , is not fixed due to scale invariance of the bulk solution , unless quantum corrections are taken into account . in the following ,
instead of fixing @xmath30 , we shall fix @xmath31 and @xmath32 .
the branes are codimension two boundaries , embedded onto conical deficits at @xmath33 , whose actions are given by @xmath34 respectively , where @xmath30 denotes the brane tensions . note that we do not assume any coupling of the dilaton with the brane tensions and such vanishing dilaton couplings are required in order that effective cosmological constant on the branes vanishes completely , at least at the classical level @xcite .
we also assume that there is no ordinary matter on the branes other than their tensions .
the brane tensions are related to the conical deficit angles by @xmath35 we stress that these relations are only valid for sufficiently small brane tensions in comparison with the bulk scale @xmath36 .
then , the angle period @xmath37 is determined by @xmath38 where @xmath39 denotes the surface gravities on each brane : @xmath40 thus , we obtain @xmath41 after eliminating @xmath32 , we obtain @xmath42 we emphasize that , in general , @xmath43 implying that the bulk angular period can be completely determined by the brane tensions . for later convenience , we regard it as a function of @xmath32 and @xmath31 : @xmath44 once the brane tensions , @xmath45 and @xmath46 are fixed , then @xmath32 is also fixed and thus , we now regard the free parameters as @xmath32 and @xmath31 , along with the dilaton bulk coupling @xmath7 . the remaining degree of freedom used to determine the bulk geometry is the absolute size of the bulk , i.e. , @xmath29 . due to the scale invariance of the background brane solution ,
@xmath29 can only be fixed by quantum corrections of bulk fields .
note that there is also magnetic flux constraint given by @xmath47 but the magnetic flux only depends on @xmath32 and @xmath31 and does not contribute to the volume of the bulk .
we consider the modulus dynamics in a four - dimensional model in the context of the _ moduli approximation _ , assuming @xmath29 is a function of @xmath48 .
this approximation may be valid for sufficiently low energy scales compared with the bulk gravitational one @xmath49 . like for the codimension one case
, we expect that this approximation is valid for low energies , namely when the energy scale on the brane is below that of the bulk . in this subsection
, we shall reintroduce @xmath49 in the action in order to clarify the dimensionality .
we regard @xmath29 as a function of the coordinate on the brane @xmath50 in eq .
( [ theory6dmet ] ) and thus , as the volume modulus .
then , we expand the components of the lagrangian density and take the modulus parts out .
it is rather useful to move to the four - dimensional physical frame : @xmath51 \nonumber\\ & = & 2\rho_+(x^{\mu } ) \tilde \rho(z ) \big [ \frac{dz^2}{g_1 ^ 2 ( 1+z^2)^2 } + \frac{g_1 ^ 2 ( 1-r^2)^2}{4 ( 1+r^2 z^2)^2 } z^2 d\theta^2 + \eta_{\mu\nu}dx^{\mu}dx^{\nu } \big ] \nonumber\\ & = & 2 \tilde \rho(z ) \big [ \rho_+(\tilde x^{\mu})\big ( \frac{dz^2}{g_1 ^ 2 ( 1+z^2)^2 } + \frac{g_1 ^ 2 ( 1-r^2)^2 } { 4 ( 1+r^2 z^2)^2 } z^2 d\theta^2
\big ) + \eta_{\mu\nu}d\tilde x^{\mu}d\tilde x^{\nu } \big ] \,,\end{aligned}\ ] ] where @xmath52 has the dimension of physical length .
the metric form is just like a kaluza - klein theory with two compact dimensions , apart from the overall conformal ( warp ) factor @xmath53 .
then , we evaluate the effective action of the volume modulus @xmath54 note that the flux and dilaton potential terms do not give any kinetic term contributions to the moduli kinetic term because the vector potential @xmath55 has no scale dependence . the second term is a total derivative term and does not contribute to the effective action . integrating over the extra dimensions , we obtain @xmath56 after redefining the modulus as @xmath57 we obtain the canonical form of the modulus kinetic term as @xmath58
next , we introduce a massless , minimally coupled scalar field ( not to be confused with the dilaton field that supports the background configuration ) . from this
we can investigate the one - loop effective action for such a scalar field and thus , the effective potential of the volume modulus on the co - dimension two warped brane background .
from now on , we shall work mainly in euclidean space , rather than in the original lorentzian frame .
the action for the massless scalar field perturbations is given by @xmath59 the one - loop effective action for a massless minimally coupled scalar field is defined as @xmath60 where @xmath61 is the six - dimensional laplacian , which is divergent and needs to be regularized and renormalized . for this purpose ,
we define @xmath62 where @xmath63 is the ( integrated ) heat kernel defined by @xmath64 the ( integrated ) zeta function is related to the heat kernel by a mellin transform : @xmath65 and after analytically continuing to @xmath66 we obtain the renormalized one - loop effective action .
the explicit expression for the renormalized effective action is @xmath67 where @xmath68 is euler s constant @xmath69 and the pole term at @xmath70 is removed by renormalization .
thus , after a redefinition of the renormalization scale we obtain the renormalized effective action and the renormalized scalar field effective action can be written as @xmath71 by integrating over the internal dimensions , the four - dimensional effective potential is @xmath72 where @xmath73 has the dimensions @xmath74 . in the next section
, we shall derive the effective potential in a codimension two warped brane model .
for brevity , from now on we shall omit the subscript ren " . in the case of codimension two , the zeta function is given by the summation @xmath75 where the eigenvalues are defined by @xmath76 it is straightforward to show that @xmath77 where @xmath78 is a heat kernel coefficient , defined by the asymptotic expansion of the heat kernel @xcite : @xmath79 in the warped codimension two case , the eigenvalues are unknown and there are not even any analytic solutions for the eigenfunctions so we must resort to using an approximate wkb method to estimate the effective action .
one strategy to evaluate the one - loop effective action and the effective potential is to define a continuous conformal transformation ( parameterized by @xmath80 ) @xmath81 and thus @xmath82 where for @xmath83 we have the original metric , which we shall denote as @xmath84 .
the classical action of this scalar field is changed under a conformal transformation , see eq .
( [ conformal ] ) , by ( here we are considering a massless , minimally coupled bulk scalar field ) @xmath85 where @xmath86 due to technical reasons , which will be explained later , we shall evaluate the zeta function in the non - warped frame @xmath87 . the correction associated with such a conformal transformation
is commonly known as the cocycle function ( obtained from integration along the paths of the conformal transformation ) : @xmath88 furthermore , thanks to the conformal invariance of @xmath89 we arrive at @xmath90 the term @xmath91 is given by the volume integration of linear combinations of cubic order curvature tensors @xcite : @xmath92 \nonumber\\ & & \qquad + ( { \rm contribution\,\ , of\,\ , conical\,\ , branes } ) \big\ } \,.\label{a6}\end{aligned}\ ] ] in order to avoid a volume divergence we shall employ the one - loop effective potential , as given by eq .
( [ effact6d ] ) , to try and stabilize the extra dimensions . from our discussion in the previous subsection ,
the effective potential takes exclusively the following form @xmath93 where we have defined @xmath94 clearly , if @xmath95 , then the modulus effective potential has a minimum at @xmath96 after a redefinition of the modulus , as given by eq .
( [ effrad6d ] ) , the effective potential can be rewritten as @xmath97 the field value at the minimum is then given by @xmath98 and hence , the effective modulus mass becomes @xmath99 thus , the ratio between the stabilized mass and the effective bulk cosmological constant is @xmath100 the expression above allows us to quantify the _ rigidity _ of stabilization ; namely , is the modulus mass stable against kk perturbations
. however , the problem is that the @xmath91 contribution for conical branes has never been formulated as far as the authors are aware .
in this section , in order investigate the one - loop effective potential of the volume modulus , we consider the four - dimensional counterpart of the six - dimensional case in the theory @xcite , @xmath101 where , again , @xmath4 is a dilaton , @xmath5 represents a @xmath6 gauge field strength and @xmath102 represents the dilaton potential ( corresponding to @xmath7 in the original model ) .
hereafter , we set the bulk gravitational scale to @xmath103 for simplicity ; we shall reinsert it if and when it is needed .
the essential properties of this solution is summarized in appendix a. the reason we use this model is that we can evaluate the uv contributions from the conical branes .
we shall extrapolate our results to the original model in six dimensions .
the zeta function and one - loop effective action in four dimensions can be defined similarly to the case of six dimensions , as in eq .
( [ zeta6d ] ) and eq . ( [ effact6d ] ) , just by replacing @xmath104 ( and @xmath105 ) .
thus , the one - loop effective potential of the volume modulus takes the same form as eq .
( [ effact6d ] ) , after the replacement of @xmath106 with @xmath107 .
the effective action also takes the same form as eq .
( [ dec ] ) , given the scale invariant nature of the flux compactification solution , implying the effective potential can be written similarly as @xmath108 where @xmath109 note that the stabilized volume of the internal space is given by @xmath110 again , if @xmath111 , then the modulus effective potential has a minimum at @xmath112 , which is determined by @xmath113 and @xmath114 .
importantly , given the similarity of the spacetime structure in this four dimensional model with the original six dimensional model , the form of the effective modulus field can be estimated by dimensional arguments , in comparison to eq .
( [ effmodu ] ) .
namely , @xmath115 acts as the modulus field in two dimensions . the field value at the minimum is then given by @xmath116 and hence , the effective modulus mass is @xmath117 the _ rigidity _ of the stability is determined by the quantity @xmath118 given the conformal invariance of the heat kernel coefficients , @xmath119 is straightforward to evaluate in the @xmath83 frame . in the original four - dimensional spacetime , including the contribution of the cones @xcite we have @xmath120 \big\},\label{a4}\end{aligned}\ ] ] where @xmath121 runs from @xmath122 to @xmath123 which denotes the two orthonormal directions to the conical branes @xmath124 and @xmath125 is the extrinsic curvature defined by @xmath126 .
here @xmath127 denotes two mutually orthogonal ( inward pointing ) unit normals and @xmath128 and @xmath129 , where @xmath130 is the induced metric on the brane .
the terms @xmath131 are undetermined parameters associated with conformal invariance of the conical heat kernels .
but , as shown in appendix c , the extrinsic curvatures vanish on the cones and we need not specify them .
note that this formula is valid for any deficit angle @xcite .
now we present the results of our analyses on the one - loop effective potential of the modulus . the function @xmath114 that corresponds to @xmath132 in the original six - dimensional model is composed of two parts ; a bulk and conical brane contribution and thus , the total is @xmath133 from the properties of the heat kernel coefficients @xmath114 is invariant under conformal transformations , as is explicitly demonstrated in appendix c. the bulk part of @xmath114 is obtained from eq .
( [ bulk ] ) for @xmath83 and as shown in appendix b , it is in fact independent of @xmath80 .
similarly , the brane contribution to @xmath114 is obtained from eq .
( [ inflationary ] ) . in fig . 1 ,
the bulk contribution to @xmath114 is shown as a function of @xmath134 , for @xmath135 and @xmath136 .
note that in the following plots , which are described as functions of @xmath32 , we have set @xmath136 .
there is no significant dependence on @xmath31 and thus we mainly focus on @xmath32 . for other values of @xmath102 , the behavior of @xmath114
is also almost the same .
the bulk contribution is always positive and as @xmath137 it suffers from a divergence because of the presence of a naked curvature singularity .
is shown as a function of @xmath32 , for @xmath135 . ]
is shown as a function of @xmath32 , for @xmath135 . ] in fig .
2 , the contribution from the branes , to @xmath114 , is shown as a function of @xmath32 . comparing this with fig . 1
, one can readily see that the cone contribution becomes comparable to that of the bulk one but the sign is opposite .
thus , the positivity of the total , @xmath114 , is a rather non - trivial issue . in fig .
3 , the total , @xmath114 , is shown as a function of @xmath32 and we find that the sign of @xmath114 is always positive .
thus , for all values of @xmath32 , the modulus can be stabilized ; this is one of the main results in this article .
though here we discuss the four - dimensional counterpart , we will make suggestions on the more realistic case of six dimensions later .
note that the behavior as shown in fig .
3 is almost the same for other values of the conical deficit angle @xmath31 . )
is shown as a function of @xmath32 for @xmath135 and @xmath136 . in the following all logarithmic plots are to the base @xmath138 . ] ) is shown as a function of @xmath32 for @xmath135 and @xmath136 . in the following all logarithmic plots
are to the base @xmath138 . ]
next , we wish to discuss the minimum potential energy and and thus , the effective mass of the modulus . these can only be considered through the combination @xmath139 , see eq .
( [ exponent ] ) , where @xmath113 corresponds to @xmath140 in the original model .
this ratio should then be compared with the renormalization scale , @xmath141 .
we give an explicit derivation of the cocycle function and the derivative of the zeta function in appendix c and d , respectively .
the term @xmath113 is obtained from eq .
( [ 4dcoef ] ) .
the first term is the derivative of the zeta function in the non - warped conformal frame , @xmath87 , in the context of wkb approximation .
the second term is known as the cocycle function , which corresponds to corrections associated with a conformal transformation . by combining these two contributions we obtain @xmath142
the bulk and brane contribution to the cocycle function , @xmath143 , is given in eq .
( [ inflationary2 ] ) , while the wkb contribution , @xmath144 , in the non - warped frame , is given in eq .
( [ inflationary3 ] ) . actually , the value of @xmath113 itself is not as important as the ratio @xmath145 , because the stabilized size of the extra dimensions is given by eq .
( [ exponent ] ) . in fig . 4
, we show a logarithmic plot of the exponential factor as a function of @xmath32 for @xmath135 and @xmath136 . for all values of @xmath32 , the stabilized size of the extra dimensions is larger than that of the renormalization scale . especially in the limits @xmath137 and @xmath146 ,
much larger extra dimensions are permitted .
furthermore , in order to show how the stability of the extra dimensions are set _ rigid _ , we can consider the ratio between the effective modulus mass and the bulk curvature scales @xmath147 . in fig .
5 , we show logarithmic plots of eq .
( [ massratio ] ) as a function of @xmath32 for various dilaton couplings @xmath102 . ) for @xmath148 is shown as a function of @xmath31 .
the red ( solid ) , green ( dashed , with wider intervals ) and blue ( dashed , with shorter intervals ) curves correspond to @xmath149 , respectively and @xmath150 . ] ) for @xmath148 is shown as a function of @xmath31 .
the red ( solid ) , green ( dashed , with wider intervals ) and blue ( dashed , with shorter intervals ) curves correspond to @xmath149 , respectively and @xmath150 . ] in fig .
6 , we also show them as a function of @xmath31 .
if the ratio is greater than unity , then , the stability of the modulus is generic . from fig . 5 and 6 , the value of eq .
( [ massratio ] ) becomes larger for larger couplings @xmath102 ( in fact the relative mass is almost proportional to @xmath151 ) .
thus , for large dilaton couplings the stability of the modulus is rigid against kk perturbations . in fig . 5
, we also see that for smaller values of @xmath32 , the modulus mass becomes relatively light , whereas for larger @xmath32 it is insensitive to @xmath32 . in the limit @xmath137 the modulus
mass diverges , because a bulk naked singularity is formed . as a final comment , fig .
6 is not particularly sensitive to changes in @xmath31 , but it does exhibit a divergent behavior for @xmath152 .
this is because in such a limit the bulk essentially disappears .
now , we shall make some suggestions on the stability of the volume modulus in six dimensions .
because of our poor knowledge about the contribution from the conical branes we can evaluate only the bulk contribution
. however , we can compare the result of the bulk contribution in six - dimensions to the four - dimensional one discussed previously .
evaluating the bulk part of @xmath132 from eqs .
( [ a6 ] ) and ( [ 6dcoef ] ) leads to @xmath153 where @xmath154 in fig . 7
, we plot eq .
( [ b6conc ] ) as a function of @xmath32 for @xmath155 and @xmath136 ( this behavior is essentially the same for other values of the coupling @xmath7 ) .
the behavior of the bulk part of @xmath132 is similar to that of @xmath114 as shown in fig .
1 . as we mentioned previously , we have no analytic formulae to evaluate the contribution from the conical branes for @xmath78 .
however , by analogy we expect similar behavior for the conical part of the heat kernel coefficient .
it may be worth mentioning that for larger @xmath32 and appropriate values of @xmath7 , @xmath132 is very likely to be positive and thus the modulus can probably be stabilized .
we could also evaluate @xmath156 , via the wkb method , and due to symmetry would follow almost identical steps to the four dimensional case , see appendix c. however , we have no way to include the cocycle correction , at the level of our semi - analytic approach .
is shown as a function of @xmath32 , for @xmath136 .
the red ( solid ) , green ( dashed , with wider intervals ) and blue ( dashed , with shorter intervals ) curves correspond to @xmath149 , respectively . ]
is shown as a function of @xmath32 , for @xmath136 .
the red ( solid ) , green ( dashed , with wider intervals ) and blue ( dashed , with shorter intervals ) curves correspond to @xmath149 , respectively . ]
thus , any further investigation of the conical contribution and the effective mass of the stabilized modulus ( in six dimensions ) is somewhat out of the scope of this article and is left for future work .
finally , before closing this article , we shall discuss some phenomenological implications , i.e. , the hierarchy structure between the fundamental energy scales and the effective vacuum energy on the brane .
even though we are investigating , primarily , the four - dimensional case , we expect similar results to hold in six dimensions .
one of the most longstanding issues in theoretical physics to date , is the problem of the hierarchy structure of the fundamental energy scales , i.e. , the huge gap between the gravitational scale , @xmath157 , and the electroweak scale , @xmath158 . in the original six - dimensional model ,
the _ effective _ four - dimensional planck scale in the four - dimensional model .
] is @xmath159 if we assume a brane localized field whose bare mass is given by @xmath160 on either brane at @xmath161 then the observed mass scales are @xmath162 thus , the mass ratio between the field and the effective planck mass is given by @xmath163 assuming that the factor of @xmath164 takes the optimal value of @xmath165 at the unification of the fundamental scales , the mass ratio becomes @xmath166 where we have used the value of @xmath167 , given by eq .
( [ sta - san ] ) , at stabilization . at present ,
the best we can do is use the results obtained from the four - dimensional model : @xmath168 the results are shown in fig . 8 and
indicate that smaller values of @xmath102 and @xmath32 lead to a large hierarchy on both branes
. however , the mass of the stabilized modulus is perhaps too light to support the realization of the hierarchy at the fundamental scales .
this is somewhat similar to the case of rs codimension one branes @xcite .
another interesting issue is the origin of the observed vacuum energy density , which is much smaller than the field theoretical prediction ; the so - called the cosmological constant problem .
given the background brane solution that we are considering , the brane vacuum energy ( the brane tensions ) only contributes to the bulk geometry and therefore does not act as an energy source for today s accelerated cosmic expansion ( at the background level the branes are to be kept flat ) .
however , in this model , the observed vacuum energy density can arise from one - loop quantum corrections as follows . at the stabilized volume , from eqs .
( [ originaleffpot ] ) and ( [ sta - san ] ) , the effective potential takes the value @xmath169 and hence , the brane realized vacuum energy is almost completely determined by the renormalization scale .
it is then quite natural that the renormalization scale is taken to the unification scale , i.e. , of order @xmath49 .
however , in such a case the exponential factor is also very small and therefore we may be able to obtain a sufficiently small vacuum energy , which acts as a dark energy source . although the vacuum energy is negative definite one might expect some kind of uplifting mechanism , for instance , due to the presence of test branes @xcite . as in the case of the hierarchy problem
, we shall consider the potential energy at the time of the stabilization .
whence , the vacuum energy is given by @xmath170 the coefficient acting on the renormalization scale is very similar to the quantity plotted in fig .
, for smaller @xmath102 and @xmath32 , we can expect smaller vacuum energy densities of the volume modulus .
the mass of the modulus is also very light , but this might be a sign of a dynamically varying vacuum energy .
we have discussed the stability of the volume of a warped codimension two brane model with flux compactification , by taking the one - loop quantum corrections of a massless scalar field into account .
we have also examined the _ rigidity _ of stability by comparing the ratio between comparison the effective modulus mass and the bulk curvature scales .
first , we introduced a warped flux compactification model in six dimensions , where we started from the solution with a spacetime structure similar to the reisner - nortstr@xmath171 m black hole .
after a _ double _ wick rotation , we were able to derive the desired braneworld solution .
the branes are located at the horizon positions and correspond to conical singularities .
for relatively low energies , namely when the brane tensions are less than the bulk gravitational energy scale , the brane tensions are related to the deficit angles directly . once the brane tensions are fixed , via eq .
( [ tendef6d ] ) , the ratio of the brane positions , @xmath28 is also fixed .
but , the volume of the extra dimensions is not completely determined , because of the scale invariance of the solution . to this end , we investigated whether or not this dynamical degree of freedom can be fixed by considering the one - loop quantum effective action for a massless , minimally coupled bulk scalar field .
we investigated the stability of the modulus for given values of the two brane tensions , or equivalently of @xmath32 and @xmath31 ( the deficit angle of the brane located at @xmath29 ) .
the form of the effective potential can be fixed up to overall coefficients , due to scale invariance and dimensional arguments , which depend on @xmath32 and @xmath31 as in eq .
( [ originaleffpot ] ) . from this
the condition for the volume modulus to have a minimum ( and hence be stabilized ) is determined only by the sign of the coefficient in the logarithmic term , i.e. , @xmath132 .
this coefficient is directly related to the zeta function @xmath172 and thus , the heat kernel coefficient @xmath173 . in the actual investigation of stability
, we have used an alternative model in four dimensions that has a very similar spacetime structure to the original six - dimensional model .
this is because the contribution for the cone to the @xmath91 heat kernel coefficient has _
never _ been formulated as far as the authors are aware .
we then showed that the heat kernel coefficient is positive definite for most dilaton couplings , @xmath102 , independently of the choice of @xmath32 and @xmath31 .
the contribution from the bulk is positive , whereas that from the conical branes is negative for smaller @xmath32 ; though they are of the same order .
for all of the cases with @xmath174 , the bulk dominates over the brane parts and thus , the volume is _ stabilized_. however , one might ask whether or not the stability is _ rigid _ against the various kk perturbations .
this rigidity can be determined from the ratio between the effective modulus mass and the stabilized curvature radius of the bulk , as given by eq .
( [ massratio ] ) , implying a knowledge of the ratio of @xmath175 .
if this ratio is less than unity , then the mass of the volume modulus becomes too light and may be destabilized by other kk contributions , whereas if it is greater than unity , the modulus field is not easily perturbed from the stabilized " minimum .
we showed that the mass ratio becomes larger for larger coupling @xmath102 , which is proportional to @xmath151 and therefore , stability is _
rigid_. for larger degrees of warping , i.e. , for smaller value of @xmath32 , other than @xmath176 , we found that the stabilized modulus mass becomes relatively light and that this ratio is not so sensitive to different values of deficit angle @xmath31 .
note that it only exhibits a divergence at @xmath177 , where the bulk essentially vanishes .
then , we made some suggestions for the six - dimensional case and by comparing the bulk part with that of the four - dimensional @xmath178 bulk part we found similar behavior indicating that our conclusion in four dimensions will carry through to the six - dimensional case : the modulus can be stabilized and furthermore our approximate wkb approach suggests that the mass of the modulus leads to rigidity for certain couplings and parameters .
of course , once formal expressions for the six - dimensional conical heat kernel coefficients are found , then we can extend the same analysis as that employed in this paper .
however , a more accurate ( semi - analytical / numerical ) method to evaluate @xmath179 should be found , rather than the approximate wkb eigenvalue method that we used in this article .
we also discussed some phenomenological issues , i.e. , the hierarchy problem of the fundamental energy scales and the extremely tiny value of the observed vacuum energy density .
the six - dimensional results were extrapolated from the corresponding results in four dimensions . we found that in the hierarchy problem , the ratio between the energy scale of the brane localized field and the gravitational scale becomes much smaller than unity , especially for larger degrees of warping @xmath180 .
furthermore , the effective vacuum energy density can also be much smaller than the gravitational scale .
the vacuum energy is negative by definition and we therefore need some kind of uplifting mechanism , to obtain a more realistic cosmology . in these cases , as was mentioned previously , the mass of the volume modulus may also be small and thus , might be destabilized .
it is also interesting to compare our results with the codimension one case .
an important difference is that in the codimension one case all the moduli are not dynamically fixed and should be stabilized by quantum corrections of bulk fields ; however in the codimension two case at least one of the moduli fields is fixed dynamically , because the brane tensions are directly related to the bulk geometry . in terms of heat kernel coefficients another obvious difference is that we are considering even dimensions . in odd dimensions only boundary terms occur and
thus , the renormalization only affects the brane tensions , whereas in even dimensions there is a bulk contribution and hence logarithmic terms appear .
probably the most important difference is that the stabilization of the volume modulus can be realized by the contribution from the bulk , not from the branes , for small warpings . for @xmath180 ( large warpings )
the stabilized modulus effectively behaves as a very light scalar field on the brane , which is somewhat similar to the codimension one case , e.g. , see @xcite . before closing this article we shall refer to some extensions of this work :
one issue is to do with the quantum corrections of other fields , such as the magnetic flux and graviton .
the tensor part of the graviton perturbations should be equivalent to a minimally coupled scalar field , which we investigated in this article and such analyses might give us important insights into warped flux compactifications with self - gravitating branes . besides quantum effects , there are many interesting issues that still remain in codimension two braneworld , e.g. , gravitational waves in the bulk @xcite , non - linear dynamics of the bulk and branes , thick branes and so on
. such work might be key to understanding the self - tuning mechanism of the effective cosmological constant as a dynamical process .
we hope to report on these topics in ongoing work @xcite .
we are grateful to a. flachi and k. uzawa for valuable and interesting discussions .
this work was supported in part by monbukagakusho grant - in - aid for scientific research(s ) no .
14102004 and ( b ) no . 17340075
. m. m. and m. s. were also supported by the japan - france research cooperative program .
in this appendix , we consider an alternative model of a warped flux compactification in a four - dimensional einstein - maxwell - dilaton theory given by eq .
( [ theory ] ) .
this is a toy - model analogue of the original six - dimensional co - dimension two braneworld model with a warped flux compactification and a non - vanishing scalar potential @xcite : hereafter , we set the bulk gravitational scale to @xmath103 for simplicity ; we shall reinsert it if and when it is needed . as for six dimensions ,
we first consider a black hole solution of the theory @xcite . after a _
double _ wick rotation , we obtain @xmath181 in this scenario the branes are located at positions that are determined by the _ horizon _
condition @xmath17 ; @xmath182 we can then rewrite @xmath183 and whence we also obtain the following useful relations @xmath184 this gives a direct relation between the mass and charge to the magnetic flux .
the global period of @xmath21 , denoted @xmath22 , is determined by the brane tensions through the tension - deficit relations as is discussed later .
after a coordinate transformation @xmath185 the bulk metric becomes @xmath186 where our choice of factorization is for later convenience . in this frame
we can now easily investigate the asymptotic behavior of the spacetime near each of the conical branes .
for @xmath187 , there are cones whose deficit angles are obtained from @xmath188 the modulus corresponds to the proper length between two branes ( strings ) , which is characterized by @xmath161 , or equivalently @xmath29 and @xmath32 .
the string tensions are related to the conical deficits by @xmath189 we stress that these relations are only valid for sufficiently small brane tensions in comparison with the bulk scale @xmath190 .
again , @xmath37 is given by a relation identical to eq .
( [ angle period ] ) and thus , @xmath191 after eliminating @xmath32 , we obtain @xmath192 once the brane tensions , @xmath45 and @xmath46 are fixed , then @xmath32 is also fixed and so we now regard the free parameters as @xmath32 and @xmath45 , along with the dilaton bulk coupling @xmath102 . the remaining degree of freedom used to determine the bulk geometry is the absolute size of the bulk , i.e. , @xmath29 .
however , because of the scale invariance , @xmath29 can only be fixed by quantum corrections of bulk fields .
the magnetic flux is given by @xmath193 thus , the gauge field flux does not contribute to the size modulus .
we discuss the basic properties of a massless , minimally coupled scalar field on this background , which is introduced in order to investigate the one - loop quantum effects in the counterpart model discussed in sec iv and v , whose action is given by @xmath194 as before , we consider a continuous conformal transformation of the metric , parameterized by @xmath80 ; @xmath195 and thus @xmath196 where for @xmath83 we have the original metric , which we shall define as @xmath197 .
the classical action changes as @xmath198 where @xmath199
the classical action of a massless , minimally coupled scalar field is changed under this conformal transformation , see eq .
( [ class4d ] ) . in four dimensions ,
including conical branes , the @xmath178 heat kernel coefficient is given by @xmath200 \bigg\}\ , .
\label{a4conf}\end{aligned}\ ] ] the integrand of the bulk part of eq .
( [ a4conf ] ) is @xmath201 where @xmath202 at a glance , eq .
( [ bulk ] ) appears to depend on the parameter @xmath80 ; however , we can show that it is independent of @xmath80 , i.e. , it is conformally invariant .
we can also evaluate the part coming from the cones . starting with the normal derivatives @xmath203 in the bulk we obtain @xmath204 these values are invariant for the limits , @xmath205 . the extrinsic curvature in each direction is given by @xmath206 and we find @xmath207
thus , we see that these terms vanish at the locations of the conical branes : at @xmath208 and @xmath209 . given this fact it is easy to show that @xmath210 is also conformally invariant .
the explicit form of @xmath114 for conical branes is finally obtained as @xmath211 \nonumber\\ & & + \frac{g^2}{288\pi } \frac{\frac{\delta_+}{2\pi}-(1-r)}{r } \frac{r+(1-\frac{\delta_+}{2\pi } ) } { 1-\frac{\delta_+}{2\pi } } \big[4 -\frac{1}{5 } \frac{r^2+(1-\frac{\delta_+}{2\pi})^2 } { ( 1-\frac{\delta_+}{2\pi})^2 } ( 1-\frac{3}{r } ) \big]\,,\label{inflationary } \ ] ] where we have used @xmath212 we have also confirmed that the bulk part of @xmath173 is conformally invariant and we expect a similar behavior for the cone part , as in four dimensions .
in this appendix , we shall give a derivation of the cocycle function for conical boundaries in four dimensions .
in deriving such a term , we shall use the conformal properties of the heat kernel coefficients , see e.g. , @xcite . in general , for a conformal transformation of the metric , given by @xmath213 ,
the heat kernel coefficient in @xmath214-dimensions satisfies the following relation @xmath215 = \frac{1}{d-2k } \delta a^{(d)}_{k } [ e^{2\omega } g_{ij},1]\big|_{\omega=0}\ , .
\label{heatconf}\end{aligned}\ ] ] we shall use this formal relation to derive the _ smeared _ heat kernel coefficients . at a glance , there seems to be a pole at @xmath216 in eq .
( [ heatconf ] ) .
however , because @xmath217 is conformally invariant in four - dimensions ( see appendix b ) , then the leading order pole term vanishes .
if @xmath218 is _ not _ conformally invariant , then we should add the term , @xmath219|_{\omega=0}$ ] , which includes mass and curvature couplings @xcite .
this is in order to keep conformal invariance on the right - hand - side . in our case
@xmath220 is already conformally invariant , because of the @xmath221 term that appears as a result of the conformal transformation of the classical scalar action .
the conical contribution to @xmath210 is given by @xmath222\end{aligned}\ ] ] and in our case the conformal transformation is defined by eq .
( [ confort4d ] ) . for a general @xmath214-dimensional spacetime
, we obtain @xmath223 \,,\end{aligned}\ ] ] where in the second step we have used the conformal scaling property of the laplacian , e.g. , see @xcite .
note that @xmath83 corresponds to the original metric for the codimension two brane .
given the above conformal transformation the ricci scalar , ricci and riemann curvature terms scale as @xcite @xmath224 \nonumber\\ \big(\sum_a r_{aa } -2\sum_{a , b } r_{abab } + \frac{1}{2}\sum_a \kappa^{(a ) } \kappa^{(a)}\big)(\epsilon ) & = & e^{-2(\epsilon-1)\omega } \big [ \sum_a r_{aa } -2\sum_{a , b } r_{abab } + \frac{1}{2 } \sum_{a , b } \kappa^{(a)}\kappa^{(a ) } \nonumber\\ & & -(\epsilon-1 ) \big ( ( d-4 ) \sum_a
\omega_{;ij}n^{(a)i}n^{(a)j } + 2 \delta_2 \omega + ( d-2)\sum_a \kappa^{(a)}\omega_{;i } n^{(a)i } \big ) \nonumber\\ & & + ( \epsilon-1)^2 \big ( \frac{1}{2}(d-4)(d+2 ) \sum_a \omega_{;i}\omega_{;j}n^{(a)i } n^{(a)j } -2(d-4)\omega_{;k}\omega^{;k } \big ) \big ] \nonumber\\ \sum_a \big ( \kappa^{(a)}{}^2 -2 \kappa^{(a)}_{ij}\kappa_{(a)}^{ij } \big ) \big)(\epsilon ) & = & e^{-2(\epsilon-1)\omega } \sum_a \big [ \kappa^{(a)}{}^2 -2 \kappa^{(a)}_{ij}\kappa_{(a)}^{ij } -2(d-4)(\epsilon - 1)\omega_{;i}n^{(a ) i } \kappa^{(a ) } \nonumber\\ & & + ( d-2)(d-4)(\epsilon-1)^2 \omega_{;i } \omega_{;j } n^{(a)i } n^{(a)j}\big]\ , , \label{coldspot}\end{aligned}\ ] ] where @xmath225 denotes the covariant derivative with respect to the metric @xmath83 and curvature tensors which appear on the right - hand - side are also defined for the same metric . here
, @xmath226 represents the laplace operator defined by @xmath227 note that terms including @xmath228 and @xmath229 vanish on the cone in the model which we are considering and thus , we neglect these terms . all the remaining terms proportional to @xmath230 are also proportional to @xmath231 .
the terms which include @xmath232 ( or @xmath233 ) are also vanishing on the cones .
however , in order to see the conformal properties of these terms , in this subsection , we keep them . by multiplying
the conformal transformation of the trace of the brane induced metric , @xmath234 , with these combinations of intrinsic and extrinsic curvature terms , we can expand the terms as a series in powers of ( @xmath231 ) .
for instance , expanding the first term in eq .
( [ a4f1 ] ) we obtain @xmath235 \nonumber\\ & = & \sqrt{h } e^{(d-4)(\epsilon-1)\omega } \big [ \frac{1}{6 } r + \lambda_{1}\sum_a \big(\kappa^{(a)}{}^2 -2 \kappa^{(a)}_{ij}\kappa_{(a)}^{ij}\big ) \nonumber \\ & & + ( d-4 ) \big\ { \frac{1}{6}(\epsilon-1 ) \delta \omega + \frac{1}{12}(\epsilon-1)^2(d-2 ) \omega_{;k}\omega^{;k } + \lambda_1 \big ( -2(\epsilon-1)\sum_a \omega_{;i}n^{(a)i}\kappa^{(a ) } \nonumber\\ & & + ( d-2)(\epsilon-1)^2\sum_a \omega_{;i}\omega_{;j}n^{(a)i}n^{(a)j } \big ) \big\ } \big ] \nonumber\\ & = & \sqrt{h}\big [ \frac{1}{6 } r + \lambda_{1}\sum_a \big ( \kappa^{(a)}{}^2 -2 \kappa^{(a)}_{ij}\kappa_{(a)}^{ij}\big)\big ] \nonumber\\ & + & ( d-4)\sqrt{h } \big [ ( \epsilon-1)\omega \big(\frac{1}{6 } r + \lambda_{1}\sum_a \big ( \kappa^{(a)}{}^2 -2 \kappa^{(a)}_{ij}\kappa_{(a)}^{ij } \big ) \big ) + \frac{1}{6}(\epsilon-1 ) \omega_{;k}{}^{;k } + \frac{1}{12}(d-2)(\epsilon-1)^2 \omega_{;k}\omega^{;k } \nonumber\\ & & + \lambda_1 \big ( -2(\epsilon - 1)\sum_a \omega_{;i}n^{(a ) i } \kappa^{(a ) } + ( d-2)(\epsilon-1)^2 \sum_a \omega_{;i } \omega_{;j } n^{(a)i } n^{(a)j } \big ) \big ] + o((d-4)^2)\,.\end{aligned}\ ] ] similarly , expanding the second piece in eq .
( [ a4f1 ] ) leads to : @xmath236 + o((d-4)^2 ) \,.\end{aligned}\ ] ] now we can construct the cocycle function for the conical boundaries .
the @xmath237 terms can be reduced to @xmath238 and because @xmath239 vanishes on the cone , at @xmath240 , we can discard these terms .
whence @xmath241
\big\}\ , .
\nonumber\end{aligned}\ ] ] finally ,
performing the integration over @xmath80 we obtain the cocycle contribution for the conical branes : @xmath242 \big\ } \nonumber\\ & = & \int d^2 x \big\ { -\frac{g^2(2\pi-\delta_+)}{1440\pi^2 } \int_0 ^ 1 d\epsilon \int^{\infty}_0 dz \big(\frac{1}{2}\ln \big(\frac{2\rho}{\rho_+}\big ) \big ) \frac { z } { ( 1+z^2)(1+r z^2 ) } \frac{\psi_4(1,r,\epsilon , z)}{(1 + r z^2)^4 } \nonumber\\ & -&\frac{g^2}{144\pi } \frac{\delta_+}{2\pi } \frac{2-\frac{\delta_+}{2\pi } } { 1-\frac{\delta_+}{2\pi } } \big[2 \ln(2\rho_+ ) -1+r -\frac{1}{5 } \frac{1+(1-\frac{\delta_+}{2\pi})^2 } { ( 1-\frac{\delta_+}{2\pi})^2 } \big ( \frac{1}{2}\ln(2\rho_+ ) ( 1 - 3r ) + \frac{1-r}{2 } \big ) \big ] \nonumber\\ & - & \frac{g^2}{144\pi } \frac{\frac{\delta_+}{2\pi}-(1-r)}{r } \frac{r+(1-\frac{\delta_+}{2\pi } ) } { 1-\frac{\delta_+}{2\pi } } \nonumber\\ & \times & \big[2\ln(2\rho_- ) -1+\frac{1}{r } -\frac{1}{5 } \frac{r^2+(1-\frac{\delta_+}{2\pi})^2 } { ( 1-\frac{\delta_+}{2\pi})^2 } \big ( \frac{1}{2}\ln(2\rho_- ) ( 1-\frac{3}{r } ) + \frac{1-\frac{1}{r}}{2 } \big ) \big ] \big\ } \,.\end{aligned}\ ] ] combining the above result with the bulk piece , we obtain the total cocycle part of @xmath113 , defined in eq .
( [ 4dcoef ] ) , as @xmath243 \nonumber\\ & - & \frac{g^2}{144\pi } \frac{\frac{\delta_+}{2\pi}-(1-r)}{r } \frac{r+(1-\frac{\delta_+}{2\pi } ) } { 1-\frac{\delta_+}{2\pi } } \nonumber\\ & \times & \big[2\ln(2r ) -1+\frac{1}{r } -\frac{1}{5 } \frac{r^2+(1-\frac{\delta_+}{2\pi})^2 } { ( 1-\frac{\delta_+}{2\pi})^2 } \big ( \frac{1}{2}\ln(2r ) ( 1-\frac{3}{r } ) + \frac{1-\frac{1}{r}}{2 } \big ) \big ] \
, , \label{inflationary2}\end{aligned}\ ] ] where @xmath244 is given in eq .
( [ psiconf ] ) .
in this appendix , we derive the mass spectrum in the conformal frame ( @xmath87 ) : @xmath245 and we consider the corresponding eigenvalue problem @xmath246 where @xmath247 is obtained from eq .
( [ good ] ) and we shall decompose the mass eigenfunction as @xmath248 from now on , we concentrate on the radial mode function and derive the mass spectrum in the context of the wkb approximation . changing variables
to @xmath249 leads to the diagonalized equation of motion in terms of @xmath250 : @xmath251 where @xmath252 to apply the wkb method we require a langer transformation @xcite @xmath253 and a further redefinition of the eigenfunction @xmath254 .
thus , eq . ( [ bulkquintom ] ) becomes @xmath255 where @xmath256 the coordinate @xmath257 runs from @xmath258 to @xmath259 , while the turning points , defined by @xmath260 , are given by @xmath261 where @xmath262 satisfies @xmath263 .
( in the original frame , @xmath83 , the turning point problem is even more difficult to solve . ) furthermore , for the @xmath264 modes the turning points extend to @xmath265 and in such a case the mass spectrum becomes exact in the context of the wkb method . for @xmath266 , turning points appear if and only if @xmath267 the mass spectrum , which will be determined later via the wkb method , is certainly valid for the modes that satisfy the above condition .
however , we will use the wkb result for deriving the full zeta function , because this is the best we can do at present .
nevertheless , it should give a useful order of magnitude approximation to @xmath268 .
in the wkb approximation , the mass spectrum is given by the quantization condition @xmath269 where the integration can be performed as follows : first , we change the integration variable @xmath270 where @xmath271 .
then , we use the fact that @xmath272 in the case that we are now considering , @xmath273\,.\end{aligned}\ ] ] then , we evaluate each term ; @xmath274_{u_+}^{u_- } = -\pi \alpha ( 1+r ) |n|\,,\end{aligned}\ ] ] and @xmath275\,.\label{integral}\end{aligned}\ ] ] unfortunately , in order to find the approximated mass spectrum , we must make one further approximation : @xmath276 .
this limit corresponds to the case where @xmath277 and @xmath278 respectively .
hence , @xmath279 = ( m+\frac{1}{2})\pi\ , .
\label{wkbmass}\end{aligned}\ ] ] the modes with @xmath264 correspond to the limits @xmath280 and @xmath281 .
thus , from eq .
( [ integral ] ) the mass spectrum in the wkb approximation becomes _ exact _ : @xmath282 for these modes we use a single summation in terms of @xmath283 . to investigate the accuracy of wkb method we can compare the result to the non - warped kk limit , where @xmath284
this has an exact analytic solution for the mode functions with @xmath285 where we have made a coordinate transformation from @xmath286 and @xmath287 the exact mass spectrum is then determined by the regularities of the mode functions on both boundaries , at @xmath288 and thus , we find @xmath289 where @xmath290 as before . in order to have positivity of the eigenvalue , @xmath291 , we shall take the latter condition . from this
the _ exact _ mass spectrum is found to be @xmath292 the result of our wkb method , eq .
( [ wkbmass ] ) , agrees with the one above ( when @xmath284 ) except for a constant term in the eigenvalue spectrum .
however , given that our wkb approximation is valid for large quantum numbers this constant term is negligible in such a limit .
given the wkb spectrum derived in the previous subsection , we derive an expression for the zeta function and the associated effective action .
the zeta function is given by the mode sum @xmath293 hereafter , we omit `` wkb '' in the subscript of the zeta function . here
, we shall use the density of states method @xcite , which from eq .
( [ wkbmass ] ) the density of states is easily derived : @xmath294 we can therefore replace the sum over @xmath283 with an integration in terms of @xmath291 @xcite @xmath295 it the following it will be convenient to decompose the zeta function into two pieces , i.e. , the @xmath296 and @xmath264 modes .
the total of the derivative of zeta function is finally given by @xmath297 similar considerations can easily be applied to the six - dimensional case , working in the conformal frame . for the @xmath298 mode , we obtain the following expression @xmath299 where because of the condition that there be two turning points , see eq . ( [ conde ] )
, we must restrict the range of integration .
the @xmath300 denotes that the @xmath264 mode is to be dropped , because the potential for this mode is not oscillator - like . performing the @xmath301-integration first , we obtain @xmath302 where @xmath303 is defined by @xmath304 for sufficiently large @xmath305 , we obtain the following integration formulae @xmath306 \big\ } \ , , \label{yamada}\end{aligned}\ ] ] where again @xmath307 is a hypergeometric function .
note that the first term in eq .
( [ yamada ] ) has been analytically continued by employing the standard duplication formula for the gamma function ; otherwise the result will be ill defined when taking the derivative at @xmath70 . finally , expressing the summations over @xmath308 in terms riemann zeta functions we obtain ( substituting back the expression for @xmath309 , see eq .
( c6 ) ) @xmath310 -1 \big ) \big\}\big]\,,\end{aligned}\ ] ] where @xmath311 is riemann s zeta function .
then , taking the derivative in terms of @xmath312 and an analytic continuation @xmath313 , we find @xmath314 \big\ } \,.\end{aligned}\ ] ] the zeta function for @xmath264 mode is given by @xmath315^{-s } = \frac{\pi ( 2g)^{2 - 2s}}{(s-1 ) } \zeta_{h } ( 2s-2,\frac{1}{2 } ) \,,\end{aligned}\ ] ] where @xmath316 is hurwitz s zeta function . taking the derivative with respect to @xmath312 and continuing to @xmath66
gives @xmath317 s. m. carroll and m. m. guica , arxiv : hep - th/0302067 .
i. navarro , jcap * 0309 * , 004 ( 2003 ) [ arxiv : hep - th/0302129 ] .
y. aghababaie , c. p. burgess , s. l. parameswaran and f. quevedo , nucl .
b * 680 * , 389 ( 2004 ) [ arxiv : hep - th/0304256 ] .
t. tanaka and x. montes , nucl .
b * 582 * , 259 ( 2000 ) [ arxiv : hep - th/0001092 ] . c. csaki , m. graesser , l. randall and j. terning , phys .
d * 62 * , 045015 ( 2000 ) [ arxiv : hep - ph/9911406 ] .
j. garriga , o. pujolas and t. tanaka , nucl .
b * 605 * , 192 ( 2001 ) [ arxiv : hep - th/0004109 ] . j. garriga , o. pujolas and t. tanaka , nucl .
b * 655 * , 127 ( 2003 ) [ arxiv : hep - th/0111277 ] .
g. w. gibbons , r. guven and c. n. pope , phys .
b * 595 * , 498 ( 2004 ) [ arxiv : hep - th/0307238 ] .
y. aghababaie _ et al .
_ , jhep * 0309 * , 037 ( 2003 ) [ arxiv : hep - th/0308064 ] . h. m. lee and c. ludeling , jhep * 0601 * , 062 ( 2006 ) [ arxiv : hep - th/0510026 ] . b. m. n. carter , a. b. nielsen and d. l. wiltshire , arxiv : hep - th/0602086 .
h. nishino and e. sezgin , phys .
b * 144 * , 187 ( 1984 ) .
h. nishino and e. sezgin , nucl .
b * 278 * , 353 ( 1986 ) .
s. mukohyama , y. sendouda , h. yoshiguchi and s. kinoshita , jcap * 0507 * , 013 ( 2005 ) [ arxiv : hep - th/0506050 ] .
h. yoshiguchi , s. mukohyama , y. sendouda and s. kinoshita , jcap * 0603 * , 018 ( 2006 ) [ arxiv : hep - th/0512212 ] .
d. v. vassilevich , phys .
rept . *
388 * , 279 ( 2003 ) [ arxiv : hep - th/0306138 ] .
c. p. burgess and d. hoover , arxiv : hep - th/0504004 .
d. hoover and c. p. burgess , jhep * 0601 * , 058 ( 2006 ) [ arxiv : hep - th/0507293 ] . | we investigate the stability of the extra dimensions in a warped , codimension two braneworld that is based upon an einstein - maxwell - dilaton theory with a non - vanishing scalar field potential .
the braneworld solution has two 3-branes , which are located at the positions of the conical singularities . for this type of brane solution
the relative positions of the branes ( the shape modulus ) are determined via the tension - deficit relations , if the brane tensions are fixed . however , the volume of the extra dimensions ( the volume modulus ) is not fixed in the context of the classical theory , implying we should take quantum corrections into account .
hence , we discuss the one - loop effective potential of the volume modulus for a massless , minimally coupled scalar field . given the scale invariance of the background solution , the form of the modulus effective potential can _ only _ be determined from the sign of the logarithmic term in the effective potential that depends on the renormalization scale .
this term can be evaluated via heat kernel analysis and we show that in most cases the volume modulus is stabilized . in the actual evaluation , due to a lack of knowledge of the uv contributions from the conical branes in a six - dimensional spacetime , we consider its four - dimensional counterpart .
we then go on to discuss the mass scale of the modulus itself and find that it becomes comparable to the gravitational scale for mild degrees of warping , when the renormalization scale is set nominally to the gravitational scale .
then , we make some suggestions on the original six - dimensional model . finally , we close this article , after discussing some phenomenological implications relating to the hierarchy problem of the fundamental energy scales and the smallness of the effective vacuum energy on the brane .
we find that one - loop corrections in this model appear to alleviate some of these problems . |
compressive sensing ( cs ) @xcite is a recently proposed concept that enables sampling below nyquist rate , without ( or with little ) sacrificing reconstruction quality . based on exploiting the signal sparsity in typical domains ,
cs methods can be used in the sensing devices , such as mr imaging @xcite and ad conversion @xcite , where the devices have a high cost of acquiring each additional sample or a high requirement on time .
therefore , as the sparsity level is often not known a priori , it can be very challenging to use cs in practical sensing hardware .
sequential compressive sensing @xcite can effectively deal with the above problems .
sequential cs considers a scenario where the observations can be obtained in sequence , and computations with observations are performed to decide whether these samples are enough .
consequently , it is allowed to recover the signal either exactly or to a given tolerance from the smallest possible number of observations .
there have been several recovery algorithms for sequential cs .
asif @xcite solved the problem by homotopy method .
garrigues @xcite discussed the lasso problem with sequential observations .
this work extends a recent proposed zero - point attracting projection ( zap ) algorithm @xcite to the scenario of sequential cs .
zap employs an approximate @xmath0 norm as the sparsity constraint and updates in the solution space . comparing with the existing algorithms , it needs fewer measurements and lower computation complexity .
therefore the new algorithm can provide a much more appropriate solution for practical sensing devices , which is validated by numerical simulations .
suppose @xmath1 is an unknown sparse signal , which is @xmath2-length but has only @xmath3 nonzero entries , where @xmath4 .
in their ice - breaking contributions , candes et al suggested to measure @xmath1 with under - determined observations , i.e. @xmath5 , where @xmath6 consists of random entries and has much fewer rows than columns .
they also proved that @xmath0 norm or @xmath7 norm constraint optimization can successfully recover the unknown signal with overwhelming probability , @xmath8 there are many methods proposed to solve ( [ problem ] ) , of which concerned in this work is zap .
zap iteratively searches the sparsest result in solution space .
the recursion starts from the least square optimal solution , @xmath9 , where @xmath10 denotes the pseudo - inverse matrix of @xmath11 . in the @xmath12th iteration ,
the solution is first updated along the negative gradient direction of a sparse penalty , @xmath13 where @xmath14 denotes a sparse constraint function and @xmath15 denotes the step size . in the reference , an approximate @xmath0 norm is employed and the corresponding @xmath16th entry of @xmath17 is @xmath18 where @xmath19 is a controlling parameter and it is readily recognized that the penalty tends to @xmath0 norm as @xmath19 approaches to infinity . then @xmath20 is projected back to the solution space to satisfy the observation constraint , @xmath21 where @xmath22 is defined as projection matrix and @xmath23 .
equation ( [ za ] ) appears that an attractor locates at the zero - point is pulling the iterative solution to be sparser , as explains the first part of the algorithm s name .
the last part comes from ( [ projection ] ) , which means that @xmath20 is projected back to the solution space . imagining a scenario that the samples are measured in realtime . at time
@xmath24 , an @xmath24-length measurement vector @xmath25 is collected and utilized to solve the sparsest solution by ( [ problem ] ) .
if the available measurements are not enough to recover the original sparse signal , a new sample @xmath26 is generated at time @xmath27 , where @xmath28 denotes the sampling weight vector .
thus the problem becomes solving @xmath29 , where @xmath30,\qquad { \bf a}_{m+1}=\left[{\bf a}_m \atop { \bf a}_{m+1}^{\rm t}\right].\ ] ] obviously , it is a waste of resources if the recovery algorithm is re - initialized without the utilization of earlier estimate , i.e. the available result at time @xmath24 .
consequently , the basic aim of sequential compressive signal reconstruction is to find an effective method of refining @xmath31 based on the information of @xmath32 .
for conciseness , the detailed iteration procedure of online zap is provided in tab.[tab1 ] .
it can be seen that online zap has two recursions .
the inner iteration is to update the solution by zap with the given measurements .
the outer iteration is for sequential input . in order to improve the performance ,
several techniques are used in online zap and they are discussed in the following subsections .
.the procedure of online zap [ cols= " < " , ] zap works in an iterative way to produce a sparse solution via recursion in the solution space . in the online scenario , the previous estimate can be used to initialize the incoming iteration , i.e. @xmath33 , where @xmath34 denotes the maximum iteration number at time @xmath24 .
the pseudo - inverse matrix @xmath35 plays an important role in the recursion of ( [ projection ] ) . considering the high computational cost of matrix inverse operation , @xmath35 is generally prepared before iterations .
however , in the online scenario , @xmath36 becomes time - dependent and @xmath37 need to be recalculated in each time instant . in order to reduce the complexity ,
the pseudo - inverse matrix is updated iteratively .
define @xmath38 , which is already available after time @xmath24 .
consequently , as the new sample is arriving , using basic algebra one has the recursion @xmath39\left[{\bf a}_m^{\rm t } \ ; { \bf a}_{m+1}\right]\right]^{-1 } = \left[\begin{matrix}{\bf\gamma}_{m}^{-1 } & { \boldsymbol\alpha}_m \\ { \boldsymbol\alpha}_m^{\rm t } & \beta_m\end{matrix}\right]^{-1}\nonumber\\ & = \left[\begin{matrix}{\bf\gamma}_m+\theta_m{\bf\gamma}_m { \boldsymbol\alpha}_m{\boldsymbol\alpha}_m^{\rm t}{\bf\gamma}_m & -\theta_m{\bf\gamma}_m{\boldsymbol\alpha}_m\\ -\theta_m{\boldsymbol\alpha}_m^{\rm t}{\bf\gamma}_m & \theta_m\end{matrix}\right],\label{updategamma}\end{aligned}\ ] ] where @xmath40 as the step size in gradient descent iterations , the parameter @xmath15 controls a tradeoff between the speed of convergence and the accuracy of the solution . in order to improve the performances of the proposed algorithm ,
the idea of variable step size is taken into consideration .
the control scheme is rather direct : @xmath15 is initialized to be a large value after new sample arrived , and reduced by a factor as long as the iteration is convergent .
the reduction is repeated several times until @xmath15 is sufficiently small .
since the algorithm has two recursions , we employ @xmath41 and @xmath42 to denote the decreasing speed of outer and inner iteration , respectively . in addition
, @xmath15 is no longer decreased when the step size is rather small .
there are two kinds of recursions requiring stop rules in the online zap algorithm .
firstly , after the @xmath24th sample arrived , @xmath43 iterates with @xmath12 to produce the best estimate based on the @xmath24 measurements .
the inner iteration should stop after the algorithm reaches steady state , which means the sparsity penalty starts increasing .
consequently , the inner recursion stops ( a ) when the number of reductions of @xmath15 reaches one-@xmath44th of its initial value or ( b ) when the number of iterations reaches the bound @xmath45 . secondly , as soon as the sparse signal is successfully reconstructed , the following samples are no longer necessary and the sensing procedure stops .
therefore , the outer recursion stops when the estimate error is below a particular value @xmath46 .
computer simulation is presented in this section to verify the performance of the proposed algorithm compared with typical sequential cs reconstruction algorithm for solving bpdn problem @xcite , whose matlab code can be downloaded from the website @xcite . in the following experiment ,
the entries of each row of @xmath6 are independently generated from normal distribution .
the locations of @xmath3 nonzero coefficients of sparse signal @xmath1 are randomly chosen with uniform distribution @xmath47 $ ] .
the corresponding nonzero coefficients are gaussian with mean zero and unit variance .
the system parameters are @xmath48 and @xmath49 .
the number of measurements @xmath50 increases form @xmath51 to @xmath52 .
the parameters for bpdn are set as the recommended values by the author .
the parameters for online zap are @xmath53 , @xmath54 , @xmath55 , @xmath56 , @xmath57 , @xmath58 .
the simulation is repeated ten times , then mean square derivation ( msd ) between the original signal and reconstruction signal as well as the average running time calculated .
figure [ varm ] shows msd curve according to @xmath50 . as can be seen ,
the performance of zap is better than that of bpdn .
when the sparse signal is recovered successfully , the number of measurements bpdn needs is larger than @xmath59 , while the number zap algorithm needs is less than 80 .
figure [ time ] demonstrates the cpu running time as @xmath50 increases .
again , zap has the better performance .
the cpu time of bpdn is twice than that of zap for successful recovery ( according to fig.[varm ] , here @xmath50 is chosen as @xmath59 for comparison ) .
we have introduced in this paper a new online signal reconstruction algorithm for sequential compressive sensing .
the proposed algorithm extends zap to sequential scenario . and
in order to improve the performance , some methods , including the warm start and variable step size , are adopted .
the final experiment indicates that the proposed algorithm needs less measurements and less cpu time than the reference algorithm .
d. l. donoho , `` compressed sensing , '' _ ieee trans . on information theory _ , 52(4 ) , pp.1289 - 1306 , april 2006 .
e. candes , j. romberg , and t. tao , `` robust uncertainty principles : exact signal reconstruction from highly incomplete frequency information , '' _ ieee trans .
inform . theory _ , 52:489 - 509 , 2006 . m. lustig , d. donoho , and j. pauly , `` sparse mri : the application of compressed sensing for rapid mr imaging , '' _ magnetic resonance in medicine _ , 58(6 ) pp .
1182 - 1195 , december 2007 . m. mishali , y. c. eldar , and j. a. tropp , `` efficient sampling of sparse wideband analog signals , '' _ ccit report # 705 _ , october 2008 .
d. malioutov , s. sanghavi , and a. willsky , `` compressed sensing with sequential observations , '' _ icassp _ , pp .
3357 - 3360 , april 2008 .
j. jin , y. gu , and s. mei , `` a stochastic gradient approach on compressive sensing signal reconstruction based on adaptive filtering framework , '' _ ieee journal of selected topics in signal processing _ , 4(2 ) , pp.409 - 420 , 2010 . | sequential compressive sensing , which may be widely used in sensing devices , is a popular topic of recent research .
this paper proposes an online recovery algorithm for sparse approximation of sequential compressive sensing .
several techniques including warm start , fast iteration , and variable step size are adopted in the proposed algorithm to improve its online performance .
finally , numerical simulations demonstrate its better performance than the relative art .
* keywords : * compressive sensing , sparse signal recovery , sequential , online algorithm , zero - point attracting projection |
to present the underlying principle of our method , we first extend the bloch s theorem to open system , then derive the hall conductance for open systems .
take isolated electrons in a potential as an example , the bloch s theorem for a closed system states that the energy eigenstate for an electron in a periodic potential can be written as bloch waves . to extend this theorem from closed to open systems
, we formulate this statement as follows .
consider an electron in a periodic potential @xmath6 with periodicity @xmath7 , i.e. , @xmath8 the one electron schrdinger equation @xmath9 should also have a solution @xmath10 corresponding to the same energy @xmath11 .
namely , @xmath12 . here
, @xmath13 denotes the index for the energy levels , @xmath14 is the mass of electron .
furthermore , the energy eigenstate can be written as , @xmath15 where @xmath16 satisfies @xmath17 are the bloch waves , @xmath18 denotes the bloch vector .
define a translation operator @xmath19 which , when operating on any smooth function @xmath20 , shifts the argument by @xmath7 , @xmath21 this operator can be explicitly written as @xmath22 if @xmath19 is applied to a hamiltonian @xmath23 with periodic potential @xmath6 , the hamiltonian is left invariant , i.e. , @xmath24=0.$ ] now we extend the bloch s theorem from closed to open systems .
suppose that the density matrix @xmath0 of the open system is governed by a master equation @xcite , @xmath25 + { \cal l } ( \rho ) \equiv { \cal p}(\rho ) , \label{mse}\ ] ] where @xmath26 sometimes called dissipator describes the decoherence effect . in the absence of decoherence
, we know that a key ingredient of the bloch s theorem is @xmath24=0.$ ] thus , to preserve the translation invariant of the dynamics , it is natural to restrict the master equation to satisfy @xmath27 which is similar to @xmath28 for a closed system . for a lindblad master equation with decay rates @xmath29 and lindblad operators @xmath30@xcite , @xmath31 + \sum_j \gamma_j(2f_j\rho f_j^\dagger- f_j^\dagger f_j \rho - \rho f_j^\dagger f_j ) , \label{lbmse}\ ] ] eq .
( [ tio ] ) leads to @xmath32=0 $ ] and @xmath24=0 $ ] for any @xmath33 .
consequently , when @xmath34 is a steady state of the system , @xmath35 is also a steady state , since @xmath36 the translation operator satisfying eq .
( [ tio ] ) preserve the decoherence - free subspace(dfs ) @xcite .
dfs has been defined as a collection of states that undergo unitary evolution in the presence of decoherence .
the theory of dfs provides us with an important strategy to the passive presentation of quantum information .
the advantage of this translation - preserved - dfs is its possible applications into quantum information processing in the presence of decoherence . identifying the problem of energy eigenstates in closed system with the problem of steady states in open system
, we formulate the bloch s theorem of open system as follows . for an open system described by eq .
( [ mse ] ) with translation invariant map @xmath37 , its steady state can be written as@xcite , @xmath38 where @xmath39 are bloch waves of the corresponding closed system , and @xmath40 is the vacuum state . the coefficients @xmath41 are independent of position @xmath42 , this fact can lift the limitation on the uniqueness required for steady states @xmath34 . in other words , @xmath43 satisfy naturally in this situation . for the lindblad master equation eq .
( [ lbmse ] ) , @xmath44=0 $ ] yields @xmath45 thus , the lindblad operators @xmath30 conserve the crystalline momentum @xmath18 of the bloch wave .
this does not imply that the steady state has a well - defined crystalline momentum , since the steady state is a convex mixture of well - defined momenta states .
it is worth noticing that the bloch s theorem of open system eq .
( [ obt ] ) relies on a postulate that the number of particles in the system is limited to below 1 . when the number of particles is conserved , and consider the system having only one particle , the last term in eq .
( [ obt ] ) can be omitted . in the following ,
we shall restricted our attention to open systems that possess translation invariance and preserve the ti phase .
for this purpose , we need to specify how the dissipator is realized in physics . in an optical lattice setup ,
such a dissipative dynamics can be engineered by manipulating couplings of the lattice to different atomic species , which play the role of the dissipative bath@xcite . to derive the hall conductance of an open system ,
we first develop a perturbation theory to calculate the steady state of the master equation eq.([mse ] ) .
perturbation theory is a widely accepted tool in the investigation of closed quantum systems . in the context of open quantum systems , however , the perturbation theory based on the markovian quantum master equation is barely developed .
the recent investigation of open systems mostly relies on exact diagonalization of the liouville superoperator or quantum trajectories , this approach is limited by current computational capabilities and is a drawback for analytically understanding open systems .
in a recent work@xcite , we have developed a perturbation theory for open systems based on the lindblad master equation . in this approach ,
the decay rate was treated as a perturbation .
successive terms of those expansions yield characteristic loss rates for dissipation processes . in ref.@xcite , instead of computing the full density matrix , the authors develop a perturbation theory to calculate directly the correlation functions . based on the right and left eigenstates of the superoperator @xmath46 ,
a perturbation theory is proposed@xcite , the non - positivity issue of the steady - state may appear in this method due to truncations . here
, we apply the perturbation theory in ref.@xcite to derive the steady state . instead of treating the decoherence as perturbation , a perturbed term in the hamiltonian
is introduced . to present the main results of our method , we first consider a situation without decoherence , namely , for an open system described by the master equation , @xmath47 + { \cal l } ( \rho ) , \ ] ] we have @xmath48 , where @xmath49 is the first order expansion of steady state
, @xmath50 @xmath51 is the perturbation parameter from @xmath52 , @xmath53 is an eigenstate of @xmath54 with eigenvalue @xmath55 , @xmath56 is the index for the eigenlevels .
the steady state in this situation would be a diagonal matrix in the basis of energy eigenstates due to thermalization , i.e. , @xmath57 with @xmath58 the dirac delta function .
the expansion coefficients then reduce to , @xmath59 obviously , @xmath60 where @xmath61 to shorten the notation , here and hereafter , the perturbation parameter @xmath51 is included in @xmath62 .
namely , @xmath62 here and in the following equals the multiple of @xmath62 and @xmath51 in eq .
( [ vy ] ) . consider a x - direction weak electric field , @xmath63 , simple algebra yields(see methods ) , @xmath64 where @xmath65 is the x - component of @xmath18 , @xmath66 is the charge of electron .
suppose the temperature is zero and the single filled band is the s - th bloch band , i.e. , all @xmath67 except @xmath68 , @xmath69 takes ( @xmath70 runs over the band indices ) , @xmath71 while @xmath72 for other @xmath56 and @xmath33 .
collecting all these results , we have ( @xmath73 , planck constant ) @xmath74.\ ] ] here the fact that the contribution from the filled band is zero has been used .
this is exactly the results in@xcite for closed systems .
next let us consider what happens when there is a single steady band in the presence of decoherence .
we refer the single steady band to that , with a fixed @xmath18 , there is only a single energy eigenstate in the dfs .
we denote this state by @xmath75 . in this case , the operator @xmath30 in eq .
( [ lbmse ] ) may takes , @xmath76 this describes a situation where all bands decay to the s - th band at rates of @xmath29 with preserved momenta @xmath77 , see fig . 1 .
straightforward calculation yields , @xmath78 substituting these equations into eq .
( [ 1thalpha ] ) and using @xmath67 for any @xmath56 and @xmath33 except @xmath79 we arrive at @xmath80 here @xmath81 is defined as @xmath82 and @xmath83 for @xmath84 and @xmath85 . for large energy band gaps , @xmath86 the coefficients approximately take , @xmath87 it is not trivial to extend the case of single steady band to two steady bands , as we shall show below .
denote the two steady bands by @xmath88 and @xmath89 , respectively , a possible realization of the two steady bands is via a dissipator , @xmath90 where we choose @xmath91 and @xmath92 denotes the decay rate . following the same procedure as in the case of single steady band , we find @xmath93 can be written in a form similar to eq .
( [ alphamna ] ) , @xmath94 with @xmath95 defined by , @xmath96 where @xmath97 when @xmath98 or @xmath99 , otherwise it takes 0 . substituting @xmath93 into the hall current and supposing the current is zero in the absence of the external field , we find that the hall current can be separated into two parts . the first part is independent of the decay rates and it can be written in terms of chern number , while the second part takes a different form related closely to the dissipator .
these two parts also manifest in the hall conductivity discussed below , suggesting us to define a _
topological _ value called chern rate for the system . the hall conductivity , defined as the ratio of the hall current density @xmath100 and the electronic field @xmath101 , is therefore given by @xmath102 here @xmath103 to derive these results , @xmath104 has been used .
this is one of the main result of this work .
it is worth pointing out that this result sharply depends on the decoherence mechanism .
in fact , as we will show later in the two - band model , the hall conductivity is not a mixture of hall conductivities for various steady bands .
assume @xmath105 independent of @xmath65 and @xmath106 , the integral on the right hand side of @xmath107 , i.e. , @xmath108 is nothing but the chern number which takes integer values as pointed out in@xcite .
then @xmath107 can be written as @xmath109 @xmath107 is a weighted chern number for the two steady bands .
this term may not be an integer for a general open system , despite its topological origin . for @xmath18-dependent @xmath105 , @xmath110 $ ]
has been defined as the so - called chern value@xcite , which witnesses a topological non - trivial order present in the berry curvature .
it recovers the standard chern number if the steady state is a pure bloch state .
@xmath111 consists of two parts , @xmath112 here , @xmath113 and @xmath114 describe respectively the first order and second order corrections of the decoherence to the hall conductivity . they can not be written in terms of chern number in general , since both @xmath115 and the energy gap depend on band index .
therefore , there is no topological invariance for the open system from the viewpoint of hall conductivity , this is true even when the dissipation rates @xmath29 and the band gaps are independent of band index , @xmath114 can be expressed in terms of chern numbers in this case , but @xmath113 still can not .
the hall current given by @xmath111 characterizes the environmental activation of excited electrons in the bulk , and it is not zero in the regions outside the topological regime , where @xmath116 .
this can be found in eq .
( [ 1thhall ] ) .
these observations motivate us to define a topological value , to which we will refer as chern rate , @xmath117 we adopt terminology chern rate for the following reasons .
firstly , it possesses topological origin ; secondly , it may not take an integer for a general open system ; thirdly , it should differ from the chern value defined in ref.@xcite , and in addition the hall conductance is simply a multiple of the chern rate and @xmath118 . of cause , the chern rate returns back to the chern number when the system is an isolated topological insulator .
it is well known that bloch s waves @xmath119 under time - reversal transformation take @xmath120 , then the berry curvature defined by @xmath121 under the time - reversal transformation satisfies , @xmath122 .
so , for system with time reversal symmetry , @xmath123 is an odd function of @xmath124 . as a consequence ,
the chern number for a time - reversal invariant system is zero , because the integral of an odd function over the whole brillouin zone must be zero .
this is not the case for second line in eq .
( [ 1thhall ] ) that is an even function of @xmath125 .
this fact reflects that the second line in eq .
( [ 1thhall ] ) may not be zero for a time - reversally invariant system , and hence the chern rate loses partially its topological origin in this case .
we will illustrate below that this non - topological term can be eliminated by properly designing @xmath126 and @xmath127 in eq .
( [ h01 ] ) .
we now apply this formalism to derive a formula for hall conductance in a two - band system .
a decoherence mechanism different from this section is considered , namely the decoherence operator @xmath30 in the dissipator is not purely a jordan block .
this difference would manifest in the hall conductivity , for example , the hall conductivity is not a mixture of hall conductivities for various bands .
we can apply the representation to develop a general formula for hall conductance for a two - band system .
let us start with an effective hamiltonian , @xmath128 where @xmath129 is the energy without couplings , it may take @xmath130 for the band electron with effective mass @xmath131 , and @xmath132 with constant @xmath133 and @xmath134 for the surface states of bulk @xmath135@xcite .
@xmath136 are the momentum - dependent coefficients which describe the spin - orbit couplings .
@xmath137 , @xmath138 @xmath139 and @xmath140 .
consider phenomenally a dissipator , @xmath141 where @xmath142 are momentum dependent decay rates , @xmath143 are pauli matrices .
this dissipator describes a decay of the fermion from the spin - up state to the spin - down state with conserved momenta .
it differs from those in the last section at that this dissipator does not describe decays from one band to the other , it instead characterizes the decay of the electron spin states , see fig . 2 .
now we introduce a perturbation @xmath144 to hamiltonian @xmath145 , the total hamiltonian with fixed @xmath146 and @xmath106 is then @xmath147 . up to first order in @xmath51 ,
we write the steady state with fixed @xmath146 and @xmath106 as , @xmath148 tedious but straightforward calculations yield , @xmath149 in the basis spanned by the eigenstates of @xmath145 , we have @xmath150}^2}}},\nonumber\\ { \tau^{(1)}_{21 } } & = & \tau^{(1)*}_{12 } , \label{coefficient}\end{aligned}\ ] ] and @xmath151,\nonumber\\ s_2 & = & e_1 \cos(3\theta ) ( \gamma-2ie_1)[\gamma ( h^{\prime}_{12 } + h^{\prime}_{21})+ 2ie_1 h^{\prime}_{12}],\\ s_3&= & \gamma \sin\theta [ -\gamma^2 + 3i\gamma e_1 + 3e_1 ^ 2+e_1 \cos(2\theta ) ( i\gamma + e_1 ) ] .\nonumber \label{coefficient10}\end{aligned}\ ] ] here , @xmath152 , @xmath153 , @xmath154 are matrix elements of @xmath155 in the basis spanned by the eigenstates of @xmath156 . for more details , see methods . the diagonal elements of @xmath157 is not listed here , since it has no contribution to the conductivity . in weak dissipation limit ,
@xmath158 we can expand @xmath159 in powers of @xmath160 . to first order in @xmath160 ,
@xmath159 can be written as , @xmath161 assuming a weak electric field is applied along the x - direction and the corresponding vector potential is time - dependent , we find by simple algebra that , @xmath162 substituting these equations into the hall conductivity and assuming @xmath160 independent of @xmath18 , we have @xmath163 discussions on the hall conductivity are in order .
the first integral describes a contribution of zeroth order in @xmath160 .
it is different from the usual hall conductivity of tis with a single filled band @xmath164 , the difference comes from the deviation of the steady state from the gibbs states .
note that when @xmath165 , the first integral represents the usual hall conductivity , the second and third integral represent a correction of dissipation to the hall conductivity .
we observe that the third integral vanishes with @xmath165 . in this case , the second integral reduces to , @xmath166 \nonumber\end{aligned}\ ] ] which is exactly the result in the last section for tis with two bands . noting that @xmath167 and @xmath168 can be written in terms of @xmath169 and @xmath170 , @xmath171 , @xmath172 we deduce the hall conductance as , @xmath173 , @xmath174 .
\label{shenhallf } \end{aligned}\ ] ] this equation is available for all two - band system described by the hamiltonian in eq .
( [ h01 ] ) . to be specific ,
we consider a two - dimensional ferromagnetic electron gas with both rashba and dresselhaus coupling , this system can be described by hamiltonian eq .
( [ h01 ] ) with @xmath175 , @xmath176 , and @xmath177 , here the momenta @xmath178 and @xmath179 . using the formula
( [ shenhallf ] ) for the hall conductivity , we calculate the hall conductivity and show the numerical results in fig . 3 .
3(a ) shows the zero - order hall conductivity versus @xmath180 .
the red - solid line is for the closed system , while the blue - dashed line for the open system with @xmath181 .
it is interesting to notice that hall conductivity of the open system with @xmath182 is different from that in closed system .
this is easy to understand , the steady state of an open system is in general a mixed state , even though the decoherence rate is close to zero .
3(a ) shows a phase transition at @xmath183 , when @xmath184 , the chern number of the closed system is 1 , while for @xmath185 , the chern number is -1 . for open system
, the phase transition can still be found from the hall conductivity , even if the absolute value of @xmath186 in the open system is smaller than that in the closed system .
the first - order correction @xmath187 are negative on both sides of @xmath188 , as shown in fig .
3(b ) , where we plot the first - order hall conductivity as a function of @xmath180 and @xmath189 . and @xmath187 as a function of @xmath180 ( mev@xmath190nm/@xmath191 ) and @xmath189 ( in units of mev ) .
parameters chosen are , ( a ) @xmath181 , @xmath192 mev@xmath193nm/@xmath191 , and ( b)@xmath160=0.1 mev , @xmath192 mev@xmath194nm/@xmath191 .
note that @xmath195 is independent of @xmath189 . ]
the second concrete example is bulk @xmath196 .
the low - lying effective model for bulk @xmath196 can be formally diagonalized , which can be interpreted as the @xmath197 and @xmath198 valleys in the graphene@xcite . for the valleys located at @xmath197 ,
the effective hamiltonian takes the same form as in eq .
( [ h01 ] ) but with @xmath199 , @xmath200 , and @xmath201 .
a straightforward calculation shows that the term proportional to @xmath160 in the hall conductivity is zero , this does not mean that the decoherence has no effect on the hall conductivity .
in fact , the decoherence leads the system to a mixed state , yielding the hall conductivity , @xmath202 for @xmath203 and @xmath204 , the hall conductance is zero . for @xmath205 and @xmath204 , @xmath206 , and
@xmath207 when @xmath203 and @xmath208 .
this is different from the results of closed system@xcite . in the third concrete example , we apply the hamiltonian eq .
( [ h01 ] ) to model the two - dimensional lattice in a magnetic field @xcite .
the tight - binding hamiltonian for such a lattice is written as , @xmath209 where @xmath210 is the usual fermion operator on the lattice , @xmath211 and @xmath212 denote the hopping amplitudes along the x- and y - direction , respectively . the first summation is taken over all the nearest - neighbor sites along the x - direction and the second sum along the y - direction .
the phase @xmath213 represents the magnetic flux through the lattice .
when @xmath214 , the single band @xmath215 is doubly degenerate .
the term with @xmath212 in the hamiltonian gives the coupling between the two branches of the dispersion .
consider two branches which are coupled by @xmath216th order perturbation , the gaps open and the size of the gap due to this coupling is the order of @xmath217 .
the effective hamiltonian then take eq .
( [ h01])@xcite with @xmath218 , @xmath219 , and @xmath220 is proportional to ( is the order of ) @xmath221 in terms of @xmath222 , @xmath223 and @xmath224 , the model takes , @xmath225 , @xmath226 , and @xmath227 . as a function of @xmath220 ( in units of mev ) and
@xmath211 ( in units of mev ) .
parameters chosen are @xmath228 ( a ) @xmath229 , and ( b ) @xmath230 .
note that the sign of @xmath231 in figures ( a ) and ( b ) are different .
further numerical simulations show that @xmath231 depends only on the parity of @xmath14 , i.e. , figure ( a ) is for all odd @xmath14 , while figure ( b ) for even @xmath14 . ] when applying the formula to this model , we can prove that @xmath232 and @xmath233 .
this can be done by examining the definition , @xmath234 and replacing @xmath146 in eq .
( [ h01 ] ) by @xmath235 with this observation , the hall conductance reduces to , @xmath236 an interesting observation is that the correction of the decoherence to the hall conductance is zero , this can be understood by examining eq .
( [ shenhallf ] ) , keeping in mind that @xmath169 depends only on @xmath146 while @xmath170 only on @xmath106 .
it is important to point out that the contribution from the steady state in the absence of external field was ignored in this section , this is reasonable that there has no current in the system when it reaches its steady state without external driving fields . in other words
, we here only have interests in the current induced by the external fields , all of other contributions do not concern us .
the dependence of the hall conductivity on @xmath220 and @xmath211 is shown in fig .
we find that @xmath231 change sharply around @xmath237 except at @xmath238 , but there is no phase transition at @xmath237 in the sense that the hall conductance has a same sign for both positive and negative @xmath211 . the topological phase changes with the parity of @xmath14 , when @xmath14 is an odd integer , @xmath239 , whereas for even @xmath14 , @xmath240 .
we have studied the hall conductance of topological insulators in the presence of decoherence . after extending the bloch s theorem from closed to open system ,
we have developed an approach to calculate perturbatively the steady state of the system driven by a perturbation .
then we apply this approach to derive the hall conductance for the open system .
we expand the hall conductance in powers of dissipation rate , and find that the zeroth order covers the usual hall conductance when the open system decays from a band to the others , whereas it can not return to the usual hall conductance with a dissipator in the other form .
the first order gives the correlation of the decoherence to the conductance , which vanishes for the two - dimensional lattice and contributes non - zero value to bulk @xmath241 generally speaking , the hall conductance for open system can not be written as a multiple of a chern number and a constant , or as a weighted sum of chern numbers , in this sense , there is no topological invariant for open systems .
the situation changes when a dissipator keeps the density matrix of the steady state in a diagonal form in a hilbert space spanned by the instantaneous eigenstates of the hamiltonian .
specifically , when the steady state takes , @xmath242 with @xmath243 independent of time , and @xmath244 denotes a wavefunction subject to the hamiltonian , the hall conductivity can be written as a weighted sum of chern numbers .
this is easy to find by expanding @xmath245 up to first order in the field strength and substituting the expansion into the hall conductivity .
an interesting observation of this paper is that by properly designing the hamiltonian , the decoherence effect on the hall conductance can be eliminated in the two - band model .
this observation makes the tis immune to influences of environment and then support its application into quantum information processing .
the kubo formula derived within the framework of linear response theory applies for equilibrium systems .
complementarily , we develop a formalism to explore the linear response of an open system to external field . though we adopt a specific master equation to develop the idea , the general conclusion in this paper should be applicable to other open systems described by various master equations , in particular , for a system not in its equilibrium state . +
this work is supported by the nsf of china under grants no 11175032 .
we start with the master equation eq.([mse ] ) , and introduce a perturbed term @xmath246 to the hamiltonian , @xmath247 when applying the perturbation theory , we may separate the total hamiltonian @xmath248 in such a way that @xmath54 is a proper hamiltonian easy for obtaining the zeroth order steady state , while keep the perturbation part @xmath246 small .
the steady state @xmath34 can be given by solving @xmath249 + { \cal l } ( \rho_{ss})=0.\ ] ] up to first order in @xmath51 , the steady state can be expressed as , @xmath250 the zeroth order steady state @xmath251 is then given by , @xmath252 = { \cal l } ( \rho_{ss}^{(0 ) } ) , \label{steady0}\ ] ] while the first order satisfies , @xmath253+\frac{i}{\hbar}\left [ { h_0,\rho_{ss}^{(1 ) } } \right ] = { \cal l } ( \rho_{ss}^{(1)}).\label{steady1}\ ] ] in a hilbert space spanned by the eigenstates @xmath254 of hamiltonian @xmath54 , @xmath255 the steady state can be written as , @xmath256 substituting this expansion into eq .
( [ steady0 ] ) and eq .
( [ steady1 ] ) , we obtain an equation for the coefficients @xmath257 @xmath258 where @xmath259 and @xmath260 assume the zeroth order steady state is easy to derive , the steady state up to first order in @xmath51 can be given by solving eq .
( [ 1thalpha ] ) . in order to derive the hall conductance as a response to an external field
, we consider the following idealized model : an non - interacting electron gas in an periodic potential @xmath6 . in the presence of a constant electric field @xmath261 and when the field can be represented by a time - dependent vector potential , the system hamiltonian takes@xcite , @xmath262 with @xmath263 taken the electric field in the x - direction , the y - component of the velocity operator in such a case is given by @xmath264 @xcite .
the y - component of the average velocity in the steady state is , @xmath265 up to first order in the perturbation @xmath51 , @xmath266 takes @xmath267 the hall current density is given by , @xmath268 the hall conductivity @xmath231 is defined as the ratio of this current density and the electric field @xmath269 to calculate perturbatively the hall current , we work in the weak field limit , @xmath270 , this allows to use the adiabatic approximation to specify the perturbation hamiltonian @xmath271 induced by the adiabatic change of hamiltonian @xmath272 and calculate the perturbed steady state .
we expand the density matrix in the basis of the energy eigenstates @xmath273\rangle$ ] ( the eigenstates of @xmath274 ) as , @xmath275=\sum_{ij}\alpha_{ij}[\vec{k}(t ) ] this expansion into @xmath276 + { \cal l } ( \rho ) , \ ] ] we have , @xmath277_{ij } } + { [ { \cal l } ( \rho ) ] _ { ij } } - \sum\limits_m { \langle { \phi_i}|{{\dot \phi}_m}\rangle } { \alpha _ { mj}}\nonumber\\ & - & \sum\limits_n { { \alpha _ { in } } } \langle { { \dot \phi}_n}|{\phi_j}\rangle,\end{aligned}\ ] ] where for the sake of simplicity we shorten the notations as @xmath278 $ ] and @xmath279\rangle$ ] .
notice that @xmath280 we obtain the hamiltonian with a perturbation term @xmath271 , @xmath281 where , @xmath282 the hamiltonian in eq .
( [ effh ] ) is the total hamiltonian , which includes a part of zeroth order in @xmath101 and a term of first order in @xmath101 . in the following
, we shall take @xmath101 small such that hamiltonian @xmath271 proportional to @xmath101 can be treated perturbatively . solving the schrdinger equation , @xmath283 with hamiltonian eq .
( [ h01 ] ) , we can obtain the eigenenergies , @xmath284 and the corresponding eigenstates , @xmath285 where , @xmath286 for the sake of simplicity , we transform the formalism into a hilbert space spanned by the eigenstates of @xmath287 . introducing @xmath288
, we find that @xmath289 with @xmath290 define @xmath291 and @xmath153 , the elements of matrix @xmath292 can be expressed as , @xmath293 collecting all these results , the master equation can be re - written as , @xmath294 + \sum_{k_x^0,k_y } \gamma(2f{\rho}{f^\dag } - { f^\dag } f{\rho } - { \rho}{f^\dag } f ) .
\label{rho0}\ ] ] the steady state @xmath295 with fixed @xmath146 and @xmath106 can be given by solving , @xmath296= \gamma(2f\tau^{(0)}{f^\dag } - { f^\dag } f\tau^{(0 ) } - \tau^{(0)}{f^\dag } f ) , \label{rhos}\ ] ] this gives rise to , @xmath297},\nonumber\\ { \tau^{(0)}_{21}}&=&\tau^{(0)*}_{12},\\ { \tau^{(0)}_{22}}&=&1-\tau^{(0)}_{11}. \nonumber \label{coefficient}\end{aligned}\ ] ] here @xmath298 denotes the steady state without perturbations . in weak dissipation limit @xmath181 , we find @xmath299 , and @xmath300 approaches @xmath301 .
obviously , in this limit , @xmath302 when @xmath165 , leading to the thermal state ( ground state ) at zero temperature .
this observation suggests that the steady state under study is in general different from the gibbs states , as a consequence , the hall conductance would be different from that given by the kubo formula . | the hall conductivity given by the kubo formula is a linear response of the quantum transverse transport to a weak electric field .
it has been intensively studied for a quantum system without decoherence , but it is barely explored for systems subject to decoherence . in this paper , we develop a formulism to deal with this issue for topological insulators .
the hall conductance for a topological insulator coupled to an environment is derived , the derivation is based on a linear response theory of open system . as an application ,
the hall conductance of a two - band topological insulator and a two - dimensional lattice is presented and discussed .
topological insulators ( tis ) were theoretically predicted to exist and have been experimentally discovered in @xcite , they are materials that have a bulk electronic band gap like an ordinary insulator but have protected conducting topological states(edge states ) on their surface . in the last decades , these topological materials have gained many interests of scientific community for their unique properties such as quantized conductivities , dissipationless transport and edge states physics@xcite .
although the exploration of topological phases of matter has become a major topics at the frontiers of the condensed matter physics , the behavior of tis subject to dissipative dynamics has been barely explored .
this leads to a lack of capability to discuss issues such as their robustness against decoherence , which is crucial in applications of the materials in quantum information processing and spintronics .
most recently , the study of topological states was extended to non - unitary systems@xcite , going a step further beyond the hamiltonian ground - state scenario .
this first step was taken with specifically designed dissipative dynamics described by a quantum master equation .
such an approach was originally proposed as a means of quantum state preparation and quantum computation@xcite , which relies on the engineering of the system - reservoir coupling . to define the topological invariant for open systems , the authors use a scheme called purification to calculate quantities of quantum system in mixed states . to be specific , for a density matrix @xmath0 in a hilbert space @xmath1
, the density matrix @xmath0 can be purified to @xmath2 by introducing an ancilla acting on a hilbert space @xmath3 such that the tracing over the ancilla @xmath4 yields the density matrix , @xmath5 in other words , mixed states can always be seen as pure states of a larger system ( i.e. , the system plus the introduced ancilla ) , the topological invariant ( called chern value in ref.@xcite ) can then be defined as usual(closed system ) tis .
turn to the topological invariant for closed system in more details .
the topological invariant was first derived by thouless _
et al._@xcite , which provides a characterization of fermionic time - reversal - broken ( trb ) topological order in two spatial dimensions .
this was done by linear response theory in such a way that the hall conductivity is represented in terms of a topological invariant ( or the chern number ) , which is related to an adiabatic change of the hamiltonian in momentum space .
however , the extension of this topological invariant from closed to open systems@xcite is not given in this manner to date , i.e. , it is defined neither via the hall conductance , nor by the linear response theory .
this paper presents a method to extend the topological invariant from closed to open systems .
the scheme is based on a linear response theory developed here for open systems . by calculating the hall conductance as a response to the adiabatic change of the hamiltonian in momentum space , the topological invariant is proportional to the quantized hall conductivity for the system in steady states . |
very recently , a series of charmonium - like resonances has successively been observed by various experimental collaborations , such as @xmath0@xcite , @xmath11@xcite , @xmath12@xcite , and several bottomoniun - like states were also discovered at the belle and babar energy ranges , such as @xmath13@xcite , @xmath14 and @xmath15 @xcite .
it is noted that there is almost no room in the ground state representations of @xmath16 to accommodate those newly observed resonances .
not only their mass spectra , but also their behaviors of production and decay may hint if they are exotic states such as hybrids , molecular states , tetraquarks , or just radial and/or orbital excited charmonia .
it would be an interesting task to determine their inner structures from both experimental and theoretical aspects . as a matter of fact , there are different interpretations for these resonances which are of different constituent compositions from the regular mesons , so that for clarifying its structure , all the interpretations should be studied one by one and then we can see if the predictions are in agreement with data . the most reasonable way is to calculate its mass and decay widths by assuming its structure in certain theoretical frameworks , then a comparison of the prediction with data would confirm or negate the assumption about the identity of the resonance
. among the newly observed resonances , @xmath0 which was found a while ago has caused special interests of experimentalists as well as theorists @xcite .
some authors regard it as a molecular state@xcite , whereas other groups consider it as a tetraquark@xcite . instead
, the recent data on @xmath17 may hint a possible assignment @xmath18 ( i.e. @xmath2 ) for @xmath0 which has inspired theoretical interests@xcite .
the charmonia mass spectrum has been calculated in the potential model and the theoretical prediction for a @xmath18 charmonium is about @xmath19gev@xcite or @xmath20gev@xcite which is 30 - 60 mev lower than the measured value of @xmath0 .
because of uncertainty existing in the calculation of the binding energies of excited states with the potential model such difference may not be too serious , but the difference might imply that our understanding on its assignment might be not completely correct . to determine its structure one needs more information from other sources , especially its decay modes .
the belle and babar collaborations also reported their measurements on the radiative decays of @xmath0@xcite .
the data given by the babar collaboration @xcite are @xmath21 @xmath22 and @xmath23 the belle collaboration also reported their new result as @xmath24 and set an upper bound @xmath25 and @xmath26 because they found no evidence for @xmath27@xcite .
the data from the two collaborations are consistent on @xmath28 but largely apart on @xmath27 .
obviously the further measurement and theoretical study are badly needed . in this work
we investigate the radiative decays of @xmath0 which is supposed to be a @xmath18 charmonium in the light - front quark model .
the results may help us to determine the structure of @xmath0 .
the light - front quark model(lfqm ) is a relativistic model@xcite and in this model wave functions are manifestly lorentz invariant and expressed in terms of the fractions of internal momenta of the constituents which are independent of the total hadron momentum .
this approach has been applied to study many processes and thoroughly discussed in literatures @xcite .
generally the results obtained in this framework qualitatively coincide with the experimental observation on the concerned processes and while taking the error ranges into account ( both experimental and theoretical ) , they can be considered to quantitatively agree with the available data . to evaluate the transition rate in the lfqm one needs to know the wave functions of the parent and daughter hadrons .
the wavefunctions for the s - wave and p - wave were given in ref.@xcite and we studied the wavefunctions of the d - wave in ref.@xcite with which we are able to investigate the transitions involving s- , p- and d - wave mesons .
then for obtaining the transition amplitude , one also needs calculate those form factors in terms of the effective lagrangian . in this work
we will deduce the form factors for the radiative decay of @xmath29 in the covariant light - front quark model . with those formulas we calculate the rate of the radiative decays of @xmath0 which is supposed to be a pure @xmath2 meson .
because of the so - called @xmath30 puzzle , namely the branching ratio of @xmath31 is extremely small while @xmath32 is one of the main decay modes of @xmath33 , it is suggested @xmath8 might not be a pure 2s state ( @xmath34 ) but a mixture of @xmath35 and @xmath36 states ( @xmath37 ) @xcite . this allegation has not been fully confirmed so far , even though it forms a reasonable interpretation for the @xmath30 puzzle and needs more theoretical and experimental tests indeed . in this work ,
we consider the two possibilities and obtain the branching ratio of the decay @xmath38 with and without considering the mixing scenario respectively . in this work after the introduction we derive the form factors for the radiative decay of @xmath39 in the covariant light - front approach in section ii .
then in section iii we present our numerical results about the decay @xmath40 and @xmath27 .
the section iv is devoted to discussions and our conclusion .
in refs.@xcite the authors discussed how to calculate the transition matrix in the covariant light - front approach . following their strategy we formulate the matrix element for @xmath39 .
the relative orbital angular momenta of @xmath2 and @xmath41 are @xmath42(d wave ) and @xmath43(s wave ) respectively .
fist let us list the vertex functions for @xmath2@xcite and @xmath41@xcite states as @xmath44.\end{aligned}\ ] ] where @xmath45 represents the @xmath41 state and @xmath46 .
the amplitude of @xmath18 state decaying into @xmath41 via a photon emission is written as cc + @xmath47{\varepsilon'}^\nu { \varepsilon}^{\alpha\beta},\end{aligned}\ ] ] where @xmath48(\slash\!\!\!p'_1+m_1)\gamma_\mu(\slash\!\!\!p_1+m_1)\}k_\alpha k_\beta,\ ] ] @xmath49(\slash\!\!\!p_1+m_1)\ } k_\alpha k_\beta,\ ] ] @xmath50 , @xmath51 , @xmath52 and @xmath53 .
the momentum @xmath54 is decomposed as ( @xmath55 ) in the light - front frame .
one needs to integrate over @xmath56 by a contour integration@xcite .
here we have set the relations among all relevant momenta as @xmath57 for fig ( a ) ; @xmath58 for fig ( b ) .
then it is easy to find @xmath59 .
the integration contour is closed in the upper plane of the complex @xmath56 for the first term in eq.(2 ) and in the lower plane for the second term .
then the first term after the integration turns into @xmath60 where @xmath61[m_0'^2-(m_1+m_2)^2 ] } } { 12\sqrt{5}m_0'^2\beta'^2}\phi'(ns).\end{aligned}\ ] ] in the formula , @xmath62 and @xmath63 , @xmath64 , @xmath65 , @xmath66 in @xmath67 must be replaced by the appropriate quantities as discussed in ref.@xcite to include the contributions of the zero modes , for example @xmath68 with @xmath69 and @xmath70 being identical to @xmath71 and @xmath72 for the maximally transverse polarized state ( @xmath73 ) and @xmath67 changes into @xmath74 .
more details about the derivation can be found in ref.@xcite .
after the replacement , @xmath74 is decomposed into @xmath75 with @xmath76 where @xmath77 are defined in the appendix .
we define the form factors as following @xmath78 which will be numerically evaluated in next section . with these form factors
the amplitude is obtained as @xmath79{\varepsilon}^\nu{\varepsilon}^{\alpha\beta}.\end{aligned}\ ] ]
with the formulas derived in section ii we can calculate the decay rates for @xmath82 in the light - front quark model .
in fact the main task is to evaluate these form factor @xmath83 , @xmath84 , @xmath85 and @xmath86 at @xmath87 .
we set @xmath88 gev following ref.@xcite at first and will discuss the dependence of our results on the choice in later part of the paper . then we need to fix the parameter @xmath89 in the wavefunction . generally in lfqm , we can fix the @xmath89 value by fitting the decay constant of a vector meson , thus in this work , we employ @xmath33 as the meson . the formula for calculating the decay constant of a vector meson
was given in refs.@xcite : @xmath90,\end{aligned}\ ] ] where @xmath91 and other notations are collected in the appendix .
using the data of @xmath92@xcite we extract the decay constant of @xmath33 : @xmath93 mev by which we fix the model parameter @xmath89 as @xmath94gev .
the masses of @xmath0 , @xmath33 and @xmath34 are chosen according to ref.@xcite .
the numerical values of the form factors for @xmath95 are listed in tab .
i. then we obtain @xmath96gev .
the recent measurement sets only an upper bound on the total width of @xmath0 , so that we can not determine accurate branching ratios for the radiative decays , but only a bound which is listed in table [ tab : decay2 ] .
theoretically , jia @xmath97@xcite studied the same transition within the pnrqcd framework where several phenomenological potentials were adopted , in their work the quantum number of @xmath0 was assigned as @xmath2 .
they obtained @xmath98gev corresponding to different potential models which is in accordance with our result . for comparing theoretical results with the data we present the experimental data and theoretical prediction on the branching ratios @xmath99 in table [ tab : decay2 ] .
the theoretical prediction of @xmath99 is slightly lower than the babar s data but quite close to the belle s data as long as we suppose @xmath0 as a @xmath2 state .
.the form factors @xmath100 and @xmath86 [ cols="^,^,^,^,^",options="header " , ] now we turn to study the dependence of the theoretical predictions on @xmath101 for the three radiative decay modes .
we let @xmath101 vary within a reasonable range from 1.3gev to 1.6gev and see how the widths of @xmath102 , @xmath103 and @xmath104 change .
indeed , it is noted that when we do so , we need to refit the @xmath89 parameter for different @xmath101 values using eq.([26 ] ) .
the dependence of our results on @xmath101 is shown in fig.[mass ] .
for example , when @xmath101 varies from 1.3gev to 1.6gev the theoretical prediction on @xmath104 changes from @xmath105 to @xmath106 .
it is about 33% change , but the order of magnitude remains unchanged .
the main goal of the work is to look for a criterion to determine the spin - orbit - identity of @xmath0 . as aforementioned ,
the babar datum is several orders larger than the prediction , thus the theoretical uncertainty about @xmath101 does not influence the qualitative conclusion .
determining the structures of the newly observed resonances is an interesting and difficult job , but the study is highly rewarding , because one can eventually identify if they are exotic states . retrospecting the history of searching glueballs at the charm energy range , the results are indeed discouraging .
the new discovery offers an opportunity to recognize the hadrons which are not existing in the basic representations of @xmath107 , but as exotic states or even mixtures of glueball and other states . as a matter of fact
, the charm energy range may be an abundant mine for such new exotic resonances .
however , before one can identify any of them as an exotic state , he has to thoroughly investigate if it is an excited state of regular charmonium no matter it is radially or orbitally excited .
@xmath0 is such an unknown resonance .
many authors consider @xmath0 to be a molecular state or tetraquark , instead , it is also suggested that it may be a @xmath2 charmonium . in this work
we calculate the rates of the radiative decays of @xmath0 in the lfqm .
we first deduce the amplitude and corresponding form factors in the amplitude of @xmath108 with the lfqm .
then we suppose the @xmath0 is a pure @xmath2 charmonium and calculate the widths of @xmath40 and @xmath27 .
we obtain @xmath109gev , @xmath110gev , @xmath111gev . our theoretical predictions on @xmath112
is obviously inconsistent with the data of the babar collaboration , but that on @xmath113 is consistent with data within the error tolerance . as a matter of the fact ,
the transition matrix element eventually depends on an overlapping integral of the wavefunctions of the initial and final hadrons .
thus the node structure of the wavefunctions is extremely important .
generally , the state with the principal quantum number @xmath114 possesses @xmath115 nodes , thus the overlapping integral between @xmath0 which is supposed to be an orbital excited state with @xmath116 does not possess a node , and neither @xmath117 , but @xmath34 does have a node .
therefore , the integrand in the overlapping integration of @xmath0 and @xmath118 is always positive , instead , in the overlapping integration between wavefunctions of @xmath0 and @xmath34 , the integrand is positive on the left side of the node of the 2s state and negative on the right side of the node
. thus a cancellation effect exists .
therefore by a common sense , the transition @xmath119 is suppressed and @xmath120 should be smaller than @xmath121 . when the @xmath10 mixing for @xmath8
is considered , the overlapping integration is decomposed into two integrations as @xmath122\ ] ] where @xmath123 , @xmath124 , @xmath125 are the radial parts of the wavefunctions for @xmath34 , @xmath126 which are the ingredients in @xmath8 , and @xmath0 , @xmath127 is the mixing angle of @xmath35 and @xmath36 states in @xmath8 .
there is a node in @xmath123 , but not in @xmath124 , so the first overlapping integration is suppressed as discussed before , but not for the second one .
therefore , the predicted branching ratio is enhanced in comparison with the pure @xmath35 state assignment for @xmath8 .
the data of the babar and belle collaborations on @xmath128 are obviously different , so that more accurate measurements are needed .
our conclusion is that if the branching ratio of @xmath128 is indeed as large as the babar group measured , the @xmath2 charmonium assignment of @xmath0 should be ruled out .
this work is supported by the national natural science foundation of china ( nnsfc ) under the contract no .
11075079 and no . 11005079
; the special grant for the ph.d .
program of ministry of eduction of p.r .
here we list some variables appearing in the context .
the incoming meson in fig .
[ fig : lfqm1 ] has the momentum @xmath129 where @xmath130 and @xmath131 are the momenta of the off - shell quark and antiquark and @xmath132 with @xmath133 and @xmath134 are internal variables and @xmath135 .
99 s. k. choi _ et al . _ [ belle collaboration ] , phys .
lett . * 91 * , 262001 ( 2003 ) [ arxiv : hep - ex/0309032 ] .
k. abe _ et al . _ [ belle collaboration ] , phys .
lett . * 98 * , 082001 ( 2007 ) [ arxiv : hep - ex/0507019 ] .
s. k. choi _ et al . _ [ belle collaboration ] , phys . rev .
* 100 * , 142001 ( 2008 ) [ arxiv:0708.1790 [ hep - ex ] ] .
b. collaboration , arxiv:1105.4583 [ hep - ex ] .
t. mehen and r. springer , phys . rev .
d * 83 * , 094009 ( 2011 ) [ arxiv:1101.5175 [ hep - ph ] ] .
s. narison , f. s. navarra and m. nielsen , phys .
d * 83 * , 016004 ( 2011 ) [ arxiv:1006.4802 [ hep - ph ] ] .
t. h. wang and g. l. wang , phys .
b * 697 * , 233 ( 2011 ) [ arxiv:1006.3363 [ hep - ph ] ] .
m. nielsen and c. m. zanetti , phys .
d * 82 * , 116002 ( 2010 ) [ arxiv:1006.0467 [ hep - ph ] ] .
p. artoisenet , e. braaten and d. kang , phys .
d * 82 * , 014013 ( 2010 ) [ arxiv:1005.2167 [ hep - ph ] ] . t. j. burns , f. piccinini , a. d. polosa , c. sabelli , phys . rev . *
d82 * , 074003 ( 2010 ) .
[ arxiv:1008.0018 [ hep - ph ] ] .
m. b. voloshin , phys .
b * 579 * , 316 ( 2004 ) [ arxiv : hep - ph/0309307 ] . i. w. lee , a. faessler , t. gutsche and v. e. lyubovitskij , phys .
d * 80 * , 094005 ( 2009 ) [ arxiv:0910.1009 [ hep - ph ] ] .
l. maiani , f. piccinini , a. d. polosa , v. riquer , phys .
rev . * d71 * , 014028 ( 2005 ) .
[ hep - ph/0412098 ] .
s. dubnicka , a. z. dubnickova , m. a. ivanov and j. g. korner , phys .
d * 81 * , 114007 ( 2010 ) [ arxiv:1004.1291 [ hep - ph ] ] .
s. dubnicka , a. z. dubnickova , m. a. ivanov , j. g. koerner , p. santorelli and g. g. saidullaeva , arxiv:1104.3974 [ hep - ph ] .
j. s. lange [ belle collaboration ] , arxiv:1010.2331 [ hep - ex ] .
s. kalashnikova and a. v. nefediev , phys .
d * 82 * , 097502 ( 2010 ) [ arxiv:1008.2895 [ hep - ph ] ]
. f. brazzi , b. grinstein , f. piccinini , a. d. polosa and c. sabelli , phys .
d * 84 * , 014003 ( 2011 ) [ arxiv:1103.3155 [ hep - ph ] ] .
m. harada and y. l. ma , prog .
phys . * 126 * , 91 ( 2011 ) [ arxiv:1010.3607 [ hep - ph ] ] .
y. jia , w. l. sang and j. xu , arxiv:1007.4541 [ hep - ph ] . e. eichten , k. gottfried , t. kinoshita , k. d. lane and t. m. yan , phys .
d * 21 * , 203 ( 1980 ) .
w. buchmuller and s. h. h. tye , phys .
d * 24 * , 132 ( 1981 ) .
s. godfrey and n. isgur , phys .
d * 32 * , 189 ( 1985 ) .
b. aubert _ et al . _ [ babar collaboration ] , phys .
d * 74 * , 071101 ( 2006 ) [ arxiv : hep - ex/0607050 ] .
b. aubert _ et al .
_ [ babar collaboration ] , phys .
lett . * 102 * , 132001 ( 2009 ) [ arxiv:0809.0042 [ hep - ex ] ] .
w. jaus , phys . rev .
d * 41 * , 3394 ( 1990 ) . c. r. ji , p. l. chung and s. r. cotanch , phys .
d * 45 * , 4214 ( 1992 ) .
h. y. cheng , c. y. cheung and c. w. hwang , phys .
d * 55 * , 1559 ( 1997 ) [ arxiv : hep - ph/9607332 ] .
w. jaus , phys .
d * 60 * , 054026 ( 1999 ) .
c. w. hwang and z. t. wei , j. phys .
g * 34 * , 687 ( 2007 ) ; c. d. lu , w. wang and z. t. wei , phys
. rev .
d * 76 * , 014013 ( 2007 ) [ arxiv : hep - ph/0701265 ] ; h. w. ke , x. q. li and z. t. wei , phys .
d * 77 * , 014020 ( 2008 ) .
z. t. wei , h. w. ke and x. q. li , phys .
d * 80 * , 094016 ( 2009 ) [ arxiv:0909.0100 [ hep - ph ] ] .
h. m. choi , phys .
d * 75 * , 073016 ( 2007 ) [ arxiv : hep - ph/0701263 ] ; h. w. ke , x. q. li and z. t. wei , phys .
d * 80 * , 074030 ( 2009 ) [ arxiv:0907.5465 [ hep - ph ] ] ; h. w. ke , x. q. li and x. liu , arxiv:1002.1187 [ hep - ph ] .
h. w. ke , x. q. li , z. t. wei and x. liu , phys .
d * 82 * , 034023 ( 2010 ) [ arxiv:1006.1091 [ hep - ph ] ] .
g. li , f. l. shao and w. wang , phys .
d * 82 * , 094031 ( 2010 ) [ arxiv:1008.3696 [ hep - ph ] ] .
z. t. wei , h. w. ke and x. f. yang , phys .
d * 80 * , 015022 ( 2009 ) [ arxiv:0905.3069 [ hep - ph ] ] .
h. w. ke , x. q. li and z. t. wei , eur .
j. c * 69 * , 133 ( 2010 ) [ arxiv:0912.4094 [ hep - ph ] ] ; h. w. ke , x. h. yuan and x. q. li , int . j. mod .
a * 26 * , 4731 ( 2010 ) , arxiv:1101.3407 [ hep - ph ] . | since the discovery of @xmath0 , its structure has been in ceaseless dispute .
the data of @xmath1 suggest that @xmath0 may be a high - spin charmonium - like of @xmath2 . in terms of the light front quark model ( lfqm )
we calculate the rates of the radiative decays @xmath3 supposing @xmath0 to be a @xmath2 charmonium . within this framework ,
our theoretical prediction on @xmath4 is at order of @xmath5 which is slightly lower than the babar s data but close to the belle s .
our prediction on @xmath6 is at order of @xmath7 if @xmath8 is a pure 2s state or @xmath9 if @xmath8 is a @xmath10 mixture , which does not conflict with the upper bound set by the belle collaboration , but is much lower than the babar s data .
thus if the future measurement decides the branching ratio of @xmath6 to be much larger than @xmath9 , the @xmath2 assignment for @xmath0 should be ruled out . |
the relaxation processes in many - body systems can be effectively studied within the kinetic theory , see ref .
@xcite and references therein .
the kinetic approaches operate with the kinetic equation , which is written for the distribution function in phase space .
the advantage of the kinetic approaches is that the kinetic equation can be easily generalized to the case of finite temperatures .
certain difficulties arise when one tries to describe the relaxation and damping effects involving the collision integral in @xmath7-dimension space @xcite . to reduce the kinetic equation for the wigner distribution function
, we will follow the diffuse approach @xcite considering the relaxation on the distorted fermi surface .
the presence of thefermi - surface distortion effects gives rise to some important consequences .
because of the fermi - surface distortion , the scattering of particles leads to the relaxation and the damping .
the purpose of present paper is to study the diffusion and drift terms in a fermi - system which can be applied to the description of the damping of collective and particle - hole excitations as well as the large amplitude dynamics . in sec .
2 we consider the kinetic equation for the wigner distribution function and reduce it applying the diffusion approximation . in sec .
3 we establish the diffusion and drift coefficients taking into consideration the fermi - surface distortion effects .
the discussion of numerical results is presented in sec .
our conclusions are given in sec .
we will restrict ourselves to the born collision approximation in the kinetic equation for wigner distribution function @xmath8 @xcite . introducing the collision integral @xmath9 ,
we will write the kinetic equation as @xmath10where @xmath11 is the driving operator . in the lowest order in @xmath12 the driving operator @xmath11 is given by @xcite @xmath13here ,
the single - particle potential @xmath14 includes , in general , the self - consistent and external fields .
the collision integral @xmath15 in eq .
( [ lv ] ) can be written in the following form @xcite @xmath16 \delta ( \delta \varepsilon ) \delta ( \delta \mathbf{p } ) \nonumber \\ & = & \int \frac{gd\mathbf{p}_{3}}{(2\pi \hbar ) ^{3}}\ \left [ w_{3\rightarrow 1}(\mathbf{p}_{1},\mathbf{p}_{3})\tilde{f}(\mathbf{p}_{1})f(\mathbf{p}% _ { 3})-w_{1\rightarrow 3}(\mathbf{p}_{1},\mathbf{p}_{3})f(\mathbf{p}_{1})% \tilde{f}(\mathbf{p}_{3})\right ] , \label{st1}\end{aligned}\]]where @xmath17 is the spin - isospin degeneracy factor , @xmath18 , @xmath19 , @xmath20 @xmath21 and @xmath22 is the probability of two - body collisions .
the gain and loss terms @xmath23 in eq .
( [ st1 ] ) are given by @xmath24 the transition probability @xmath25 in eq .
( [ st1 ] ) contains the square of the corresponding amplitude of scattering for the direct , @xmath26 , and the reverse , @xmath27 , transitions .
the probability @xmath28 includes also the distribution functions of the scattered particle in the initial and final states .
we will assume that the main contribution to the scattering amplitude is given by the transitions which correspond to a small momentum transfer : @xmath29 , where @xmath30 is the fermi momentum , see also ref .
@xcite . introducing the new variables
@xmath31we will apply the following expansions over small @xmath32 : @xmath33@xmath34and @xmath35note that the spin - averaged probability of two - body collisions @xmath36 in eq .
( [ st1 ] ) can be expressed in term of in - medium scattering cross section @xmath37 as @xmath38 in the case of elastic collisions , the scattering cross section @xmath39 depends on the modulus square of the momentum transfer @xmath32 only . using the expansions eqs .
( [ w13 ] ) and ( [ w31 ] ) , one can reduce the kinetic equation ( [ lv ] ) to the diffusion equation in the following form ( see eq .
( [ stf6 ] ) of ) @xmath40 + \mathbf{\nabla } _ { p}^{2}\,% \left [ f(\mathbf{p})\,d_{p}(\mathbf{p})\right ] , \label{4}\ ] ] where @xmath41 and @xmath42 represent the diffusion and drift terms , respectively .
both kinetic coefficients @xmath41 and @xmath42 are derived by the following relations , see , @xmath43and @xmath44
the obtained expressions ( [ dpdef1 ] ) and ( [ kpdef1 ] ) for the kinetic coefficients imply a smallness of the momentum transfer @xmath32 because of the expansions eqs .
( [ f3 ] ) , ( [ w13 ] ) and ( [ w31 ] ) . to provide a small momentum transfer @xmath32 in eqs .
( [ dpdef1 ] ) and ( [ kpdef1 ] ) we will use the finite - radius inter - particle interaction with the following gaussian form - factor @xmath45 which is appropriate for calculations of the in - medium cross - section within the transport approaches@xcite the differential cross section @xmath37 in the first born approximation is then given by @xcite @xmath46were @xmath47 and @xmath48 are the free parameters . using eqs .
( [ gain1 ] ) and ( [ pot ] ) , we will rewrite the diffuse term @xmath49 of eq .
( [ dpdef1 ] ) as @xmath50integrating in eq .
( [ dp ] ) over @xmath32 , we obtain ( see ) @xmath51 \nonumber \\ & \times & \left\ { ( 1-xy)\ j\left ( 8k^{2}r_{0}^{2}\sqrt{1-x^{2}}\sqrt{1-y^{2}}% /\hbar ^{2}\right ) -\sqrt{1-x^{2}}\sqrt{1-y^{2}}\right . \nonumber \\ & \times & \left .
\widetilde{j}\left ( 8k^{2}r_{0}^{2}\sqrt{1-x^{2}}\sqrt{% 1-y^{2}}/\hbar ^{2}\right ) \right\ } \tilde{f}\left ( \sqrt{k^{2}+p^{2}+2kpx}% \right ) \nonumber \\ & \times & f\left ( \sqrt{k^{2}+p^{2}+2kpy}\right ) , \label{dp8x}\end{aligned}\]]where @xmath52 and @xmath53
. we will apply our consideration to the nuclear fermi - liquid . for the numerical calculations we will adopt the following parameters @xmath54 and @xmath55 , which provide a reasonable value for the in - medium nucleon - nucleon cross section @xmath56 .
we will also use the fermi distribution function @xmath57where @xmath58 is the temperature , @xmath59 $ ] is the chemical potentialand @xmath60 is the fermi energy . on momentum @xmath61 in units of @xmath30 , for temperatures @xmath62 ( solid line ) , @xmath63 (
dashed line ) and @xmath64 @xmath65 ( dotted line ) . ] the results of calculations of the diffusion coefficient @xmath0 accordingly to eq .
( [ dp8x ] ) with the fermi distribution ( [ fd ] ) are presented in .
the calculations were performed for the different values of temperature @xmath58 . as seen from , the momentum dependence of @xmath0 has the clearly observed maximum at the fermi momentum @xmath30 for different temperatures . with an increase of temperature the diffusion coefficient increases as @xmath66 .
the temperature dependence of the diffusion coefficient @xmath67 is shown in as seen from , for a cold fermi system the diffusion coefficient takes the non - zero value @xmath68 @xmath69 which is caused by the relaxation on the distorted fermi - surface in eq .
( [ lv ] ) which exists at @xmath4 also .
for high temperature regime where the fermi statistic comes to maxwell one , the diffusion coefficient @xmath70 behaves as a linear function of @xmath58 in agreement with the einstein s fluctuation - dissipation theorem . , and drift , @xmath71 , coefficients taken at @xmath1 on the temperature @xmath58 . ] using eq .
( [ kpdef1 ] ) and @xmath72 , one can write the drift coefficient @xmath42 as @xmath73where @xmath74for a spherically symmetric distribution @xmath75 the drift coefficient @xmath76 in eq .
( [ k1 ] ) is reduced to the following form ( see ) @xmath77where the first moment function @xmath78 is given by eq .
( [ dk8x ] ) of .
the presence of maximum of @xmath0 at the fermi momentum denotes that @xmath79 and in agreement with eq .
( [ kpp ] ) the drift coefficient @xmath80 is reduced as @xmath81 in we show the dependence of the drift coefficient @xmath2 on the momentum @xmath61 in units of fermi momentum @xmath30 .
one can see that @xmath82 has a minimum localized at @xmath83 @xmath30 which slightly depends on the temperature .
note that for @xmath84 the temperature dependence of @xmath82 is negligible . .
] the numerical estimates of the diffuse coefficient @xmath0 and the drift term @xmath2 near the fermi momentum give @xmath85 @xmath86 and @xmath87 @xmath88 .
both obtained values of @xmath89 and @xmath80 agree with the phenomenological ones used earlier in ref .
evaluating the transport coefficients @xmath0 and @xmath2 caused by the interparticle collisions on the distorted fermi surface , we will restrict ourselves to a fermi system which is homogeneous in @xmath90 -space and omit the driving operator @xmath11 in equation ( [ 4 ] ) . using the diffusion , @xmath0 , and drift , @xmath2 , coefficients from eqs .
( [ dp8x ] ) and ( [ kpp ] ) , and solving the diffusion equation ( [ 4 ] ) with @xmath91 one can evaluate the time evolution of the wigner distribution function @xmath92 .
we will also assume a spherical fermi surface of radius @xmath30 which is derived by the condition for the particle number @xmath93 within a fixed volume @xmath94 @xmath95the diffusion equation ( [ 4 ] ) must be augmented by the initial condition for @xmath92 .
we will consider the time evolution of the initial particle - hole ( @xmath96 ) excitation which is derived at @xmath97 as , see also ref .
@xcite , @xmath98 \left [ 1-\theta ( p - p_{f})\right ] \nonumber \\ & & + \left [ 1-\theta ( p - p_{2})\right ] \theta ( p - p_{1})\theta ( p - p_{f } ) .
\label{fin}\end{aligned}\]]the distribution @xmath99 of eq .
( [ fin ] ) means the particle located at @xmath100 and the hole excitation at @xmath101 for fixed @xmath102 and @xmath103 , respectively .
the intervals @xmath104 and @xmath105 are derived from the conditions @xmath106 in we have plotted the time evolution of the wigner distribution function @xmath92 for the initial particle - hole excitation in the fermi - system with @xmath5 .
we used the initial distribution @xmath107 given by eq .
( [ fin ] ) and assumed the initial excitation energy @xmath108 @xmath65 . ) in momentum space ( in units of the fermi momentum @xmath30 ) with @xmath109 , @xmath110 , @xmath111 , @xmath112 , which corresponds to the initial excitation energy @xmath108 mev .
the solid line is the equilibrium distribution @xmath113 . ]
one can see from that the momentum distribution @xmath92 evolves to the fermi equilibrium limit @xmath113 of eq .
( [ fd ] ) with temperature @xmath114 , where @xmath115 is the level density parameter @xcite and @xmath116 is the single - particle level density at the fermi energy**. * * the corresponding relaxation time @xmath117 for the diffusion process in can be obtained considering the time evolution of the deviation @xmath118 of the distribution function @xmath92 from its equilibrium limit @xmath113 .
we introduce the mean square deviation @xmath119}^{2}-\left [ \int d% \mathbf{p\ } \delta f(p\mathbf{,}t\mathbf{)}\right ] ^{2}=\int d\mathbf{p\ [ } % \delta f(p , t)\mathbf{]}^{2}. \label{deltat}\]]the time dependence of @xmath120 for the distribution function @xmath92 from is plotted in .
the function @xmath120 can be fitted by the exponential dependence @xmath121 @xmath122 which is shown in as the dashed line . from eq .
( [ deltat ] ) on the time ( solid line ) .
the dashed line is the mean square fit to the exponential function @xmath123 . ]
both curves in are normalized to the initial mean square deviation @xmath124}^{2},\]]where @xmath125 .
the corresponding relaxation time @xmath117 is estimated as @xmath126 .
the obtained value of @xmath117 agrees with a typical collisional relaxation time @xmath127 in a nucleus @xmath128 bert78,shko05,khol82,wegm74 .
we note also that considering the time evolution of the distribution function to the equilibrium ( solid line in ) , we reach the regime where the further change of distribution function is negligible .
that means that the collision integral in equation ( [ 4 ] ) should be zero in the case of equilibrium . to verify this fact we consider the time evolution of the collision integral ( [ stf6 ] ) . in
we show the ratio * * * * @xmath129}^{2}% } { \int d\mathbf{p\ [ } \mathrm{st}\{f(t=0)\mathbf{\}]}^{2 } } , \label{rst}\]]which represents the time dependence of the mean square of collision integral @xmath130 normalized to the initial value of @xmath131 . .
] as seen from , the collision term * * * * @xmath132 * * * * ( [ stf6 ] ) approaches to zero with time as it should be in the equilibrium limit
we have considered the kinetic approach with the collision integral for the wigner distribution function using the diffusion approximation .
we have demonstrated that the collisional landau - vlasov kinetic equation can be reduced to the form of the diffusion equation .
we have established the expressions for both the diffusion coefficient @xmath0 and the drift term @xmath2 for the fermi - systems where the relaxation processes occur on the distorted fermi surface .
we have found out that the simplest isotropic assumption for the collision probability in the collision integral is insufficient for the correct description of the kinetic coefficients .
the forward scattering should be ensured for the inter - particle interaction .
for the description of the inter - particle interaction we have used the finite - range potential .
such kind of inter - particle interaction provides the necessary smallness of the momentum transfer and thereby the forward scattering of particles .
for the nuclear fermi - liquid we have calculated dependencies of the diffuse coefficient @xmath0 and the drift term @xmath2 on the momentum @xmath61 and the temperature @xmath58 . for the diffuse coefficient @xmath0 the momentum dependence possesses the clearly observed maximum of order of @xmath133 near the fermi momentum for different temperatures . with an increase of temperature the value of the diffusion coefficient increases also . excluding the small momentum values
, the drift term @xmath2 has negative sign and the minimum of order of @xmath134 .
inside the fermi sphere for @xmath135 the drift term @xmath2 shows the strong dependence on the temperature .
for @xmath84 the temperature dependence of @xmath82 is practically negligible .
we have established the finite value of the diffusion coefficient in a cold fermi system with @xmath68 @xmath136 which is caused by the relaxation on the distorted fermi - surface at @xmath4 .
the diffusion process was investigated numerically assuming the initial non - equilibrium @xmath96-excitation for a finite fermi - system with number of particles @xmath5 .
it was shown that the @xmath137-excitation relaxes to the equilibrium fermi distribution .
the numerical estimate for the relaxation time @xmath117 gives @xmath138 which is close to the corresponding estimates in a nuclear fermi - liquid obtained within the kinetic theory .
applying the expansions of @xmath139 and @xmath140 over small @xmath141 ( see eqs .
( [ w13 ] ) and ( [ w31 ] ) ) , the collision integral ( [ st1 ] ) can be written as @xmath142 .
\label{stf1}\end{aligned}\ ] ] using then eqs .
( [ gains ] ) , ( [ losses ] ) and ( [ wdsdo ] ) , we obtain @xmath143 and @xmath144 the collision integral of eq .
( [ stf1 ] ) is transformed to the following form @xmath145 introducing the two first moments @xmath146 and @xmath147 of scattering probability @xmath148 as @xmath149the collision integral ( [ stf2 ] ) is reduced as@xmath150 . \label{stf4}\]]using the relations @xmath151we rewrite the collision integral ( [ stf4 ] ) as @xmath152 \nonumber \\ & + & \nabla _ { p_{1},\nu }
\nabla _ { p_{1},\mu } ( f(\mathbf{p}_{1})b_{\nu\mu } ( % \mathbf{p}_{1 } ) ) .
\label{stf5}\end{aligned}\ ] ] finally , we obtain @xmath153 \nonumber \\ & + & \nabla_{p_{1},\nu}^{2}\,\left [ f(\mathbf{p}_{1})\,d_{p}(\mathbf{p}_{1})% \right ] .
\label{stf6}\end{aligned}\ ] ]
integrating in eq . ( [ dp ] ) over @xmath32 , one obtains @xmath154 the argument of the delta function in eq .
( [ dp1 ] ) can be written as @xmath155 where @xmath156 .
we will introduce the new variables @xmath157 and @xmath158 which allows one to rewrite eq .
( [ dp1 ] ) in the following form @xmath159 using the spherically symmetric distribution function @xmath160 and the relation @xmath161we will rewrite eq .
( [ dp3 ] ) as @xmath162integrating in eq .
( [ dp4y ] ) over @xmath163 , we obtain @xmath164 using the addition theorem for spherical harmonics , we can write in an arbitrary spherical coordinate system @xmath165taking into account the spherically symmetry of the distribution functions @xmath166 and using @xmath167we obtain @xmath168we will also introduce the following notations @xmath169 @xmath170which are attributed as functions of the argument @xmath171 . using @xmath172 and @xmath173 , one can write the angular integrals in eq .
( [ dp7x ] ) as @xmath174 finally , we obtain the diffuse coefficient @xmath0 in the form given by eq .
( [ dp8x ] ) .
the drift coefficient @xmath42 of eq .
( [ kpdef1 ] ) can be reduced similarly to above described procedure . using derivation of eq .
( a1 ) , we will reduce @xmath175 to the following form @xmath176 or @xmath177finally , we obtain @xmath178 | we suggest the diffuse approach to the relaxation processes within the kinetic theory for the wigner distribution function .
the diffusion and drift coefficients are evaluated taking into consideration the interparticle collisions on the distorted fermi surface . using the finite range interaction ,
we show that the momentum dependence of the diffuse coefficient @xmath0 has a maximum at fermi momentum @xmath1 whereas the drift coefficient @xmath2 is negative and reaches a minimum at @xmath3 . for a cold fermi system
the diffusion coefficient takes the non - zero value which is caused by the relaxation on the distorted fermi - surface at temperature @xmath4 .
the numerical solution of the diffusion equation was performed for the particle - hole excitation in a nucleus with @xmath5 . the evaluated relaxation time @xmath6 is close to the corresponding result in a nuclear fermi - liquid obtained within the kinetic theory . |
before we investigate the transport properties of a luttinger liquid , we study the role of umklapp scattering in an anisotropic fermi liquid .
the arguments given here will be published in a more extended version in ref .
@xcite .
we consider an anisotropic metal close to half filling with a clearly defined most - conducting axis in @xmath2-direction .
it is assumed that two well defined fermi sheets perpendicular to this axis exist ( see fig .
[ figfs ] ) .
the curvature of those sheets is not required to be extremely small , for our argument it is sufficient it is less than e.g. a quarter of the width of the brillouin zone ( i.e. within the shaded area of fig .
[ figfs ] ) .
we consider a rather arbitrary lattice , assuming only the existence of a translation vector @xmath3 of the underlying lattice in x - direction .
umklapp processes lead to a decay of any macroscopic momentum .
for example , close to @xmath4 filling a low - energy umklapp processes ( type ii in fig . [ figfs ] ) with
a momentum transfer @xmath5 relaxes the momentum in @xmath2 direction , where @xmath6 is a reciprocal lattice vector .
however , a pseudo momentum @xmath7 can be defined , which is conserved by two - particle scattering processes close to the fermi surface as we will show in the following @xmath8.\label{definep}\end{aligned}\ ] ] the pseudo momentum @xmath9 is measured with respect to the line @xmath10 ( dashed line in fig . [ figfs ] ) . here , we concentrate on systems close to half filling for a quasi 1d system close to a different commensurate filling , other pseudo - momenta are more relevant @xcite
. to check to what extent @xmath7 is conserved , we calculate the commutator of @xmath7 with a generic 2-particle scattering term : @xmath11 [ { \tilde{p}_x},h_2]&=&\sum_{{\mbox{1 .
bz } } } c^\dagger_{\vec{k}_1 } c^\dagger_{\vec{k}_2 } c_{\vec{k}'_2 } c_{\vec{k}'_1 } v_{\vec{k}_1 \vec{k}_2,\vec{k}'_1 \vec{k}'_2 } ( \delta k_{1x}+\delta k_{2x}-\delta k'_{1x}-\delta k'_{2x } ) \nonumber \\ & & \times \sum_{\vec{g_{\vec{n } } } } \delta(\vec{k}_1+\vec{k}_2-\vec{k}'_1-\vec{k}'_2 -\vec{g}_{\vec{n } } ) \label{comm}\end{aligned}\ ] ] where the @xmath12 are reciprocal lattice vectors .
it is easy to check , that all those terms on the right - hand side of ( [ comm ] ) vanish if all four momenta are in the shaded region of fig , [ figfs ] , i.e. @xmath13 , @xmath14 , @xmath15 , @xmath16 . for `` forward scattering '' processes of type i in fig , [ figfs ] , @xmath17 ,
both momentum @xmath18 and pseudo momentum @xmath7 are conserved as for all 4 momenta have the same sign .
while `` umklapp '' processes of type ii pick up a lattice momentum @xmath5 , this is exactly compensated by the fact that two electrons are moving from the right to the left fermi surface due to the term @xmath19 g_{1x}/4 $ ] in the definition of @xmath9 ( [ definep ] ) and @xmath7 is again conserved .
the pseudo momentum can only decay by high - energy processes far from the fermi surface , e.g. iii in fig .
[ figfs ] . in a two - particle scattering event ( e.g. in 2nd order perturbation theory )
such a scattering process is exponentially suppressed even at moderate temperatures . in high orders of perturbation theory ,
however , low - energy contribution can result from these high - energy processes due to virtual excitations . or to put it in the language of renormalization group : @xmath20-particle interactions are generated .
will they relax @xmath7 ? from the definition of the pseudo momentum
it is clear that @xmath7 is conserved modulo @xmath21 .
therefore any @xmath20-particle scattering event will change @xmath7 by a multiple of @xmath21 .
accordingly , a relaxation of @xmath7 is not possible if all @xmath22 pseudo momenta @xmath23 involved in the scattering process are smaller than @xmath24 . at low @xmath25
, a decay of @xmath7 by @xmath20-particle collision can only happen for @xmath26 where @xmath27 is the maximal distance of the fermi surface from the plane @xmath28 ( dashed line in fig .
[ figfs ] ) . at sufficiently high temperatures ,
the broadening of the fermi - surface and the thermal excitation of states with higher energy will favor decay channels of the pseudo momentum with smaller @xmath20 ( this effect can crudely be described by adding @xmath29 to @xmath27 in ( [ minn ] ) ) .
the temperature dependence of the decay - rate @xmath30 of @xmath7 at low @xmath25 in the fermi liquid regime is determined by the usual phase - space arguments : a particle of energy @xmath31 decays in @xmath32 particle and hole excitation , one of the energies is fixed by energy conservation , the remaining @xmath33 energies each have a phase - space of order @xmath34 and therefore @xmath35 where the integer @xmath20 is the smallest value consistent with ( [ minn ] ) .
the prefactor in ( [ gammap ] ) depends in a rather delicate way on the strength and range of the interaction , the screening and the band - curvature .
note , that a _ local _ @xmath20-particle interaction will give no contribution for @xmath36 due to the pauli - principle .
therefore the scattering rate is strongly suppressed for weakly coupled chains with well - screened interactions .
furthermore , additional logarithmic temperature dependences of the scattering vertices are expected even in the fermi - liquid regime as it is well known from fermi liquid theory .
we want to stress that the analysis given above is valid for interactions of arbitrary strength as long as a fermi liquid description is possible ( for strong interaction , one should , however , consider the pseudo momentum of quasi - particles which slightly differs from the pseudo momentum of the bare electrons , defined above ) .
in the preceding section we have established that the pseudo momentum @xmath7 will decay very slowly in a quasi one - dimensional metal .
how does this influence the optical conductivity @xmath37 ?
we will argue in the following , that the pseudo momentum conservation leads to well defined peak in the optical conductivity at zero frequency and we will show how its weight can be calculated reliably . furthermore , the @xmath25-dependence of the dc - conductivity is primarily given by the decay rate of @xmath7 ( [ gammap ] ) at sufficiently low temperatures .
we will first try to derive our results using rather simple hand - waving arguments which will be substantiated by some rigorous results derived many years ago @xcite .
we consider the following gedankenexperiment : at time @xmath38 we prepare a state with a finite current @xmath39 . as the current
is not conserved , it will decay rather fast by two - particle collisions ( @xmath40 at low @xmath25 ) .
typically , the initial state with finite current will also have a finite pseudo momentum @xmath41 which will decay much slower than the current ; @xmath42 .
the important point is now to realize that any state with finite pseudo momentum will typically carry a finite current @xmath43 .
accordingly , a finite fraction of @xmath44 will not decay with the fast rate @xmath45 but with the much smaller rate @xmath30 as is shown schematically in fig .
[ figjdecay ] .
how large is the fraction of the current which decays slowly ? to answer this question , we have to consider two separate questions : first , how large is @xmath46 , and second , how much current does a state with finite @xmath7 carry ? at
@xmath38 one obtains in linear response theory the ratio of the expectation values of the operators @xmath47 in terms of the corresponding susceptibilities ( defined as usual ) . to answer the second question
, we consider a situation where a field conjugated to @xmath7 is applied and find in analogy to the argument above @xmath48 .
therefore , we expect ( we will discuss later under what condition this statement is rigorously true ) that the low - energy peak in the optical conductivity carries a fraction @xmath49 of the total weight @xmath50 with @xmath51 .
note , that due to the cauchy - schwartz inequality the relative weight @xmath52 is always smaller than 1 as it has to be . for lower and lower temperature
the peak in @xmath37 gets sharper . according to ( [ druderel ] ) ,
the total weight of the peak is @xmath53 @xmath54 should not be confused with the zero - temperature drude weight as we have not included frequencies of the order of the short - time decay rate @xmath40 in the definition of its weight .
it is obvious that the weight of the low - frequency peak in @xmath37 ( shaded area in fig .
[ figjdecay ] ) can rigorously be extracted only in the limit @xmath55 , i.e. in a situation where @xmath7 is exactly conserved and @xmath54 is really a finite - temperature drude weight .
in this situation , mazur @xcite has derived long ago an exact inequality for certain correlation functions which in our context reads @xmath56 it is valid if all @xmath57 are conserved and @xmath58 for @xmath59 .
the importance of mazur s inequality has been recently emphasized by zotos _
@xcite ( see below ) .
furthermore , suzuki @xcite showed that the inequality in ( [ mazurinequ ] ) can be replaced by an equality if the sum includes _ all _ conservation laws ! if therefore @xmath7 is the _ only _ ( approximately ) conserved quantity in the system with a finite overlap to the current @xmath60 ( i.e. if @xmath7 is the slowest `` current - like '' mode in the system and @xmath30 the smallest decay rate ) then ( [ drude ] ) is exact as was tested numerically for a simple model in @xcite .
@xmath61 can easily be calculated at low @xmath25 within fermi liquid theory following standard text books @xcite .
the result will in general depend on the details of the momentum dependence of the effective interactions and the band - structure .
we assume a quasi 1d system with a fermi velocity @xmath62 parallel to the most conducting axis and for simplicity
completely local interactions characterized by two fermi liquid parameter @xmath63 and @xmath64 in the spin - singlet channel to describe the interactions of two density excitations @xmath65 on the same fermi sheet or on different sheets , respectively . with @xmath66 , the relative weight of the low frequency peak in the optical conductivity for low @xmath25 is given by @xmath67 @xmath68 is defined as an average over the fermi sheet , for example @xmath69 , where @xmath70 is the x - component of the fermi momentum on the right fermi sheet .
note that due to luttinger s theorem @xmath71 , where @xmath72 is the deviation of the electron - density from half filling .
if the interactions are sufficiently weak so that no phase transition is induced , the low - frequency weight @xmath54 vanishes close to half filling with @xmath73 where @xmath74 is the renormalized fermi energy .
we expect that the low - frequency weight @xmath54 decreases with increasing temperature , mainly due to the thermal broadening of @xmath75 .
leading finite-@xmath25 corrections to ( [ drudefl ] ) or ( [ drudeflhalf ] ) are of order @xmath76 .
the physics of approximate conservation laws and their influence on transport discussed in the previous section is even more important for exactly one - dimensional systems . surprisingly , little
is known about the conductivity @xmath77 of an ideal one - dimensional wire in the presence of umklapp scattering induced by a periodic potential . for a long time , there has not even an agreement whether @xmath0 is finite or infinite at @xmath78 for generic systems @xcite .
for example , t. giamarchi @xcite found a finite conductivity for @xmath78 within a certain perturbation theory .
however , using a luther - emery transformation of the dominant umklapp process , he concluded that the conductivity is actually infinite @xcite .
furthermore , castella , zotos and others @xcite were able to calculate the drude weight for a number of exactly solvable , integrable lattice models and they also found an infinite conductivity for @xmath78 . from our discussion in [ conduct.section
] this is not too surprising as any integrable model is characterized by an infinite number of conservation laws this point of view has been emphasized by zotos _
@xcite , who proposed to use ( [ mazurinequ ] ) to obtain an estimate for the drude weight in these models . on general grounds ,
however , one would expect a finite conductivity in any _ generic _ lattice model .
most of the above discussed discrepancies can be resolved from an analysis of the approximate conservation laws of the appropriate low - energy theories as has been realized recently by the author and n. andrei @xcite . within a certain hydrodynamic theory of the approximately conserved quantities ,
the main characteristics of the optical conductivity can be calculated reliably .
the topology of the fermi surface of a @xmath79 metal determines its low - energy excitations .
two well defined fermi - points exist at momenta @xmath80 , allowing us to define left and right moving excitations , to be described by @xmath81 ( see fig .
[ umklapp.fig ] ) .
we shall include in the fields momentum modes extending to the edge of the brillouin zone , usually omitted in treatments that concentrate on physics very close to the fermi - surface .
the hamiltonian , including high energy processes , is @xmath82 @xmath83 is the well - known luttinger liquid hamiltonian capturing the low energy behavior @xcite , @xmath84 \nonumber\end{aligned}\ ] ] @xmath85 is the fermi velocity , @xmath86 measures the strength of interactions , @xmath87 is the sum of the left and right moving electron densities . in the second line
we wrote the bosonized @xcite version of the hamiltonian . here
@xmath88 , @xmath89 are the spin and charge velocities , and the interactions determine the luttinger parameters @xmath90 with @xmath91 , @xmath92 , @xmath93 .
the high energy processes are captured in the subsequent terms which are formally irrelevant at low energies ( we consider only systems away from a mott transition , i.e. away from half filling ) . some of them , however , determine the low - frequency behavior of the conductivity at any finite @xmath25 , since they induce the decay of the conserved modes of @xmath83 ( they are `` dangerously irrelevant '' ) .
we classify these irrelevant terms with the help of two operators which will play the central role in our discussion .
the first one is the translation operator @xmath94 of the right- and left - moving fields , the second one , @xmath95 , is the difference of the number of right- and left - moving electrons , and is up to @xmath85 , the charge current of @xmath83 : @xmath96 \\ \label{j0 } j_0&=&n_r - n_l= \sum_{\sigma } \int dx \left [ \psi_{r \sigma}^\dagger \psi_{r\sigma}-\psi_{l\sigma}^\dagger \psi_{l\sigma } \right]\end{aligned}\ ] ] the linear combination @xmath97 can be identified with the total momentum of the full hamiltonian @xmath98 .
we proceed to the classification of the formally irrelevant terms in the hamiltonian .
this classification allows us to select all those terms ( actually few in number ) that determine the current dynamics .
@xmath99 includes all terms in @xmath100 which commute with both @xmath94 and @xmath101 , such as corrections due to the finite band curvature , due to finite - range interactions and similar terms .
we will not need their explicit form .
the umklapp terms @xmath102 ( @xmath103 ) convert @xmath104 right - movers to left - movers ( and vice versa ) picking up lattice momentum @xmath105 , and do not commute with either @xmath94 or @xmath101 ( see fig .
[ umklapp.fig ] ) .
leading terms are of the form , @xmath106 with momentum transfer @xmath107 .
the important step is to realize that certain pseudo momenta , @xmath108 , defined as linear combinations of @xmath101 and @xmath94 @xmath109 are approximately conserved . for @xmath110 , @xmath111 is nothing but the usual momentum , for @xmath112 we recover the pseudo momentum discussed in section [ pseudofermiliquid ] . if the hamiltonian includes only a _
single _ type of umklapp process @xmath113 , then the pseudo momentum is _ exactly _ conserved @xmath114=0.\end{aligned}\ ] ] even in the presence of band - curvature and similar terms .
therefore we can expect an infinite conductivity in such a model .
at least two independent umklapp terms are required to lead to a complete decay of the current and a finite conductivity ( an exception are exactly commensurate systems , see @xcite ) . in any generic lattice model ,
certainly all types of umklapp processes are present or are `` generated '' in the language of renormalization group .
it is clear that in such a situation , the strongest umklapp process will not determine the @xmath25 dependence of the dc - conductivity as the associated conservation law prohibits the decay of the current .
instead , the second strongest umklapp determines the decay rate of the current as we have argued in section [ conduct.section ] . as
the second strongest umklapp process is typically far from the fermi surface , the conductivity can be very large .
for lowest temperatures , we find in @xcite close to a commensurate filling @xmath115 , where @xmath116 and @xmath20 are integers with with @xmath36 @xmath117\end{aligned}\ ] ] while @xmath118 $ ] at typical incommensurate fillings .
details and omitted prefactors can be found in ref .
the calculation of the conductivity in ref .
@xcite is based on the memory matrix formalism @xcite . for a certain classical model of charge excitations of a weakly doped mott insulator
these effects have been calculated numerically in @xcite .
these numerical calculations have served as a test for the analytical methods used in @xcite .
if one analyses various scattering processes in ( quasi ) one - dimensional materials , one realizes that many of them can be characterized by approximate conservation laws .
we argued above , that umklapp scattering @xmath119 conserves pseudo - momentum but also scattering from low - energy phonons or slowly varying potential fluctuations are characterized by well - known conserved quantities as is shown in table [ t1 ] .
the consequences can be that e.g. in a situation with very strong umklapp scattering the @xmath25 dependence of the conductivity is determined by phonons , while in a regime where phonons dominate , the power - laws associated to umklapp processes show up in @xmath77 .
.[t1 ] scattering processes and the associated conservation laws [ cols="<,<",options="header " , ] this analysis applies both to one- and quasi one - dimensional materials independent of whether the system is better described by a luttinger liquid , a fermi liquid or something else ( however , the @xmath25 dependence of @xmath0 will be very different ) .
generically , the low - energy behavior of exactly one - dimensional systems is characterized by an integrable fixed - point hamiltonian with an _ infinite _ number of conservation laws .
it is therefore important to analyze whether these more complicated conservation laws influence measurable quantities like the conductivity for _ generic _ models . for incommensurate systems it is easy to check that at low temperature it is sufficient to keep track of the pseudo - momentum conservation as the decay rate of the relevant @xmath120 is exponentially small in temperature while other more complicated quantities decay with algebraically small rates
the situation is more complicated for an exactly commensurate filling as @xmath121 vanishes with exponential precision if @xmath122 @xcite . in this situation
one has to analyze in more detail the influence of other approximate conservation laws .
preliminary results suggest that for 1d mott insulators with a gap @xmath123 , and for temperatures @xmath124 , more complicated conservation laws of the associated fixed point hamiltonian ( the sine - gordon model in this case ) are indeed important .
this implies , that the optical conductivity will develop a well defined low - frequency peak as sketched in fig .
[ figjdecay ] associated to slowly decaying modes which have their origin in the structure of conservation laws of the sine - gordon model .
in this paper , we have discussed various approximate conservation laws which determine the low - frequency conductivity of clean ( quasi ) one - dimensional materials . for a large class of situations ,
the dominant scattering process leads not to a decay of the current due to the presence of a slowly decaying mode .
in this situation the second strongest scattering process determines the dc - conductivity .
the most important signature of this type of physics is a well - defined low frequency peak in the optical conductivity as is shown schematically in fig .
[ figjdecay ] .
the weight of such a peak can be calculated from ( [ drude ] ) . in experiments on bechgaard salts
@xcite indeed a well defined low - frequency peak with a small @xmath125 weight have been found . presently , it is however not clear whether one of the conservation laws discussed in this paper is at the origin of this feature .
j. slyom : the fermi gas model of one - dimensional conductors , adv . phys . * 28 * , 201303 ( 1979 ) ; v.j .
emery in _ highly conducting one - dimensional solids _ , eds . j. devreese _ et al . _
( plenum , new york , 1979 ) , p. 247
v. vescoli _
et al . _ : dimensionality - driven insulator - to - metal transition in the bechgaard salts , science , * 281 * , 1188 ( 1998 ) ; a. schwartz _
et al . _ : on - chain electrodynamics of metallic ( tmtsf)@xmath126x salts : observation of tomonaga - luttinger liquid response , phys . rev .
b * 58 * , 1261 ( 1998 ) a. rosch and n. andrei : conductivity of a clean one - dimensional wire , phys . rev
* 85 * , 10921096 ( 2000 ) m. garst and a. rosch : transport in a classical model of an one - dimensional mott insulator : influence of conservation laws , preprint , cond - mat/0102109 a. rosch and n. andrei , to be published p. mazur : non - ergodicity of phase functions in certain systems , physica * 43 * , 533545 ( 1969 ) m. suzuki : ergodicity , constants of motion and bounds for susceptibilities , physica 51 , 277289 ( 1971 ) d. pines and p. nozires , _ the theory of quantum liquids : volume 1 _ , benjamin ( new york 1966 ) t. giamarchi : umklapp process and resistivity in one - dimensional fermion systems , phys .
b * 44 * , 29052913 ( 1991 ) t. giamarchi and a.j .
millis : conductivity of a luttinger liquid , phys .
b * 46 * , 93259331 ( 1992 ) s. fujimoto and n. kawakami : exact drude weight for the one - dimensional hubbard model at finite temperatures j. phys .
a * 31 * , 465474 ( 1998 ) x. zotos : finite temperature drude weight of the one - dimensional spin-@xmath4 heisenberg model , phys .
. lett . * 82 * , 1764 ( 1998 ) ; h. castella , x. zotos , and p. prelov@xmath127ek : integrability and ideal conductance at finite temperature , phys .
lett . * 74 * , 972 ( 1995 ) x. zotos , f. naef , and p. prelov@xmath127ek : transport and conservation laws , phys .
b * 55 * , 11029 ( 1997 ) s. kirchner _
et al . _ ,
phys . rev .
b * 59 * , 1825 ( 1999 ) ; s. sachdev and k. damle , phys . rev
. lett . * 78 * , 943 ( 1997 ) ; v.v .
ponomarenko and n. nagaosa , phys .
lett . * 79 * , 1714 ( 1997 ) ; a.a .
odintsov , y. tokura , s. tarucha , phys . rev .
b * 56 * , 12729 ( 1997 ) ; m. mori , m. ogata , h. fukuyama , j. phys .
j. * 66 * , 3363 ( 1997 ) . k. le hur , cond - mat/0001439 d. forster , _ hydrodynamic fluctuations , broken symmetry , and correlation functions _ , ( benjamin , massachusetts , 1975 ) | the interplay of umklapp scattering from a periodic potential and other scattering processes determine the conductivity of ( quasi ) one - dimensional metals . we show that the transport at finite temperature is qualitatively and quantitatively strongly influenced by a number of approximate conservation laws .
typically , not the strongest but the second strongest scattering mechanism determines the dc - conductivity .
we discuss the optical conductivity both of strongly anisotropic , quasi one - dimensional fermi liquids and of luttinger liquids .
thanks to the miracles of bosonization , conformal field theory , bethe ansatz and renormalization group , the theoretical description of one - dimensional systems is one of the best developed areas in the theory of strongly correlated systems @xcite .
the enormous theoretical advance is contrasted by failure to understand certain experimental quasi one - dimensional systems like the bechgaard salts even on a qualitative level @xcite . in recent years
, carbon nanotubes have proven to be an almost ideal one - dimensional system where theoretical concepts can be tested in some detail .
astonishingly , the theory of finite temperature transport in clean one - dimensional ( 1d ) systems is not developed very far .
the technical reason for this is that in a pure luttinger - liquid the conductivity is infinite and @xmath0 is therefore a singular function of irrelevant operators and naive perturbation theory can be problematic and misleading . in ref .
@xcite , we argued that main features of the optical conductivity @xmath1 can be understood from the analysis of certain approximate conservation laws and proposed to calculate @xmath0 from a hydrodynamic theory of the corresponding slowly decaying modes .
the main result is , that in many experimentally relevant situations , the strongest scattering process alone can not lead to a relaxation of the current due to some conservation law . as a consequence , temperature dependence of the dc - conductivity
is often determined by the second strongest scattering processes .
this point of view has been tested numerically in @xcite .
we will first discuss in some detail the role of approximately conserved quantities for a quasi one - dimensional fermi liquid , afterwards we will shortly review the results for umklapp scattering in a luttinger liquid and discuss how the results generalize for other scattering mechanisms . |
the power of optical - infrared interferometry to obtain information about the astronomical source morphology ( including the angular size ) is now well - established . to properly determine the source properties ,
the quality of the measurements is an issue , which is still the subject of active research . as with any other measuring apparatus ,
the absolute calibration of the instrument ( including atmosphere ) requires careful attention .
derived from the measurement of the mutual degree of coherence of the incident radiation field , on spatial frequencies sampled by the aperture - array configuration , the final interferometric observables are non - linear mixes of noisy quantities , and of parameter estimates with their own uncertainties . in this paper , we propose to revisit and extend the existing data processing and calibration methods , in the aim to obtain reliable estimates and robust uncertainties for calibrated measurements of visibility and complex triple product .
the careful reduction process , described in the present paper , has been elaborated for the scientific study of a sample of 18 bright cool giant and supergiant stars ( see table [ tabscience ] ) .
measurements were obtained with the vlti / amber facility at medium spectral resolution ( @xmath0=1500 ) in the @xmath1 band , using triplets of 1.80-m auxiliary telescopes .
observations have been conducted during 15 observing nights between may 2009 and december 2010 ( under belgian visa guaranteed time ) .
the aim of the present work is to test the data reduction and calibration software , on data with various qualities , that we started to develop in 2006 .
.science targets of our programme measured with vlti / amber .
hipparcos parallaxes are in mas .
last column gives the calibrator(s ) associated with each science target . [ cols="<,^,^,^,^",options="header " , ] when the positive and negative values of @xmath2 along the spectral domain almost mutually compensate ( as seen on fig .
[ csplambda ] for tx psc and w ori ) , the spectral integration produces a nearly null global closure phase . as mentioned above , this can be taken as a hint for centro - symmetry .
however , @xmath2 is clearly non - null in some parts of the spectrum , and this is a hint for a deviation from centro - symmetry . to rule out the contradiction , and
so as to consider this latter possibility , we introduce a new estimator , that we call the _ centro - symmetry parameter _ ( csp ) , similar to the global closure phase , but using instead the absolute value of @xmath2 in the numerator , and the modulus of @xmath3 in the denominator , @xmath4 as for the global closure phase , a small csp value ( less than , for our sample ) suggests a high degree of centro - symmetry .
significantly high csp values , as judged from their uncertainties ( bootstrap method again ) , require to use asymmetric brightness distribution models , for the fitting on the spi data .
the values of the global csp in the @xmath1-band , and their uncertainties , are given in the second column of table [ tabcspclos ] .
, the global closure phase in @xmath1 , for the science targets .
the red inset in the upper right corner enlarges the group of points located at the bottom , around the null value of the global closure phase , pertaining to the targets @xmath5 oph , @xmath6 car , l@xmath7 pup , @xmath8 cet , @xmath9 ara , @xmath6 tra , and @xmath6 hya .
the dash - dot - dash diagonals trace the relation csp = @xmath10 . ] to illustrate the difference between the global closure phase and the csp , we plot in fig .
[ csplambda ] three quantities , for 3 scientific targets showing a global closure phase close to zero , 1 . in the panels on the left
is shown the imaginary part of the triple product @xmath11 , divided by the spectral mean ] of the real part @xmath12 .
note that the spectral mean of this ratio is equal to the tangent of the global closure phase , as defined in eq .
( [ psifact ] ) ; 2 . in the panels forming the central column
is shown the absolute value of the imaginary part of the triple product @xmath13 , divided by the spectral mean of the modulus @xmath14 .
note that the spectral mean of this ratio is equal to the sine of the global csp , as defined in eq .
( [ csptrip ] ) ; 3 .
in the panels on the right is shown the absolute value of the real part of the triple product @xmath15 , divided by the spectral mean of the modulus @xmath14 .
the three panels on the left show three typical behaviours for @xmath11 : 1 . uniformly close to zero ( @xmath5 oph ) ; 2 . decreasing symmetrically around zero ( tx psc ) ; 3 .
increasing symmetrically around zero ( w ori ) .
for each of these situations , the integration over the spectrum produces a nearly null result ( hint for centro - symmetry ) .
however the csp allows to detect deviations from centro - symmetry ( see fig .
[ fig : psicsp ] ) .
if the real and imaginary parts of the triple product are wavelength - independent , for the observation spectral range , one can show that @xmath16 .
if not , no analytical relation linking the two quantities can be derived from eqs .
( [ psifact ] ) and ( [ csptrip ] ) , because of the integrals in the numerators and the denominators .
figure [ fig : psicsp ] shows the csp in @xmath1 displayed versus the global closure phase in the same spectral band , for each science target , choosing the ob which gives the smallest uncertainty on the csp .
almost all the stars fall along the diagonals of the ( @xmath17 , csp ) diagram , which trace the relation csp = @xmath18 .
one ( w ori ) , however , is flagged as asymmetrical with the csp indicator , but not with @xmath17 .
this is precisely why we favour the csp over @xmath17 , as discussed above .
t cet has the largest values for csp and @xmath17 as well , indicating large deviation from centro - symmetry ( may be due to strong asymmetries at the wavelengths of the co bands or even suggestive of the presence of a binary companion ) .
the second largest csp value pertains to r scl , recently revealed as a wide binary by @xcite , using the alma array at millimetric wavelengths . in addition
, the last column of table [ tabcspclos ] shows that , except for t cet , the snr for the csp is comparable or larger than the snr for the global closure phase .
in order to interpret the final calibrated spi observables , _ spidast _ provides a fitting routine , based on the modified gradient - expansion algorithm , very similar to the algorithm invented by @xcite , and improved by @xcite .
the fit applies on any calibrated spi measurements ( to be chosen between visibility , flux , coherent flux , bispectrum , triple product , and closure phase ) , using a library of single - component or composite parametric chromatic models , characterised by the fourier spectrum of their intensity distribution , and the associated first - order partial derivatives w.r.t .
the model parameters . figure [ figdelophfit ] shows examples of visibility - fitting results obtained with _
spidast _ , using two models ( ud and ld - marcs ) .
this fitting routine , applied on spectro - photometric data , is used to determine the calibrator angular diameters from the marcs + turbospectrum synthetic spectra ( see subsect .
[ subseccaldiam ] ) .
note that the _ litpro _ software , developed by the jmmc working group `` model fitting '' , uses a set of elementary geometrical and center - to - limb darkening functions , as well @xcite .
however , it does not offer , contrary to _ spidast _ , the possibility to fit stellar - disk models with synthetic tabulated radiance data ( or exitance , for fits on spectro - photometric measurements ) , such those produced by the marcs + turbospectrum code .
since the uncertainties of the final calibrated data are not normally distributed , the covariance matrix , that comes out of the chi - squared fit , can not be used to infer the parameter uncertainties @xcite . to compute the uncertainties , _ spidast _ uses the residual - bootstrap method , described in detail in @xcite : 1 .
`` synthetic '' datasets are produced from random sampling with replacement of the pearson residuals ( difference between calibrated and fitted values divided by the uncertainty of the observed value ) , added to the initial fitted values ; 2 .
fitting the model on these new datasets produces a set of chi - squared minima , following a probability distribution , from which we extract the boundaries of the confidence interval , with a given confidence level ; 3 .
the parameter values associated with these chi - squared boundaries give the upper and lower limits of the parameter estimates , leading to asymmetric uncertainties .
the whole procedure allows the computing of reliable estimates of the parameter uncertainties .
in the present paper , we introduce the new _ spidast _ software , developed since 2006 with the aim to reduce , calibrate , and interpret the visibility and triple product measurements obtained with the vlti / amber facility .
_ spidast _ contains a whole set of modules , which can be launched separately or in an automatic batch file , and summarised herafter : 1 . the raw data reduction used by _
spidast _ computes the weighted average of non - aberrant data , at spectral - channel level , providing estimates of the spi observables using an automated procedure , while the method presently used with _
amdlib _ selects the `` best '' frames according to a quality threshold determined _ a posteriori _ , after several trials .
2 . the wavelength calibration procedure performed by _
spidast _ provides spectral shifts following a polynomial law , tracing them at the channel level .
this is done by computing the cross - correlation of the measured spectra of the calibrators with their synthetic spectra produced by the marcs model , while _ amdlib _ only provides a constant spectral shift over the spectrum , from the correlation with the telluric lines .
3 . for selected calibrators
not included in the calibrator catalogues , _ spidast _ provides several routines for estimation of the angular diameter with indirect methods , the most accurate being the fit of stellar - atmosphere model spectra given by marcs on spectro - photometric data .
the calibrator synthetic observables are derived using clv functions produced by marcs .
4 . to obtain an accurate interferometric calibration ( via an automated procedure ) , _ spidast _ : ( 1 ) divides the calibrator raw data with the associated synthetic observables in each spectral channel , which gives the instrumental response function in squared visibility and triple product ; ( 2 ) interpolates or fits the response at the time of each exposure on the science target ; ( 3 ) divides the science raw data with the interpolated / fitted response , which gives the science calibrated observables for each exposure ; ( 4 ) computes their weighted average over each ob . at each processing step , the uncertainties are computed thanks to the bootstrap method applied on the weighted means .
5 . using the real and imaginary parts of the calibrated triple product , _ spidast _
measures the deviation from centro - symmetry of the brightness distribution of each science target in the observation spectral band , thanks to a new parameter , more sensitive to asymmetries than the global closure phase .
6 . finally , _
spidast _ proposes a complete fitting tool , using a set of parametric and chromatic models , and accepting input tables of flux / intensity synthetic data .
such a careful calibration process of spi data is a crucial step for their trustworthy astrophysical exploitation , which is the topic of two associated papers @xcite .
parameter extraction using non - linear fits of source model , as well as aperture - synthesis image reconstruction , need reliable estimates of the calibrated observables , with robust uncertainties .
our reduction , calibration , and fitting routines also apply to any other spectral datasets , including spectroscopic data .
we made the _
spidast _ software public : the source code of any program of our software suite can be obtained by sending an e - mail to the first author of the present paper .
the authors thank the eso - paranal vlti team for supporting their amber observations , especially the night astronomers a. mrand , g. montagnier , f. patru , j .- b .
le bouquin , s. rengaswamy , and w.j .
de wit , the vlti group coordinator s. brillant , and the telescope and instrument operators a. and j. cortes , a. pino , c. herrera , d. castex , s. cerda , and c. cid .
they also thank the programme national de physique stellaire ( pnps ) for supporting part of this collaborative research .
is grateful to b. plez , k. eriksson , and t. masseron for their ongoing support on the use of the marcs code .
s.s . was ( partly ) supported by the austrian science fund through fwf project p19503-n16 ; a.c . is post - doctoral fellow from f.r.s .- fnrs ( belgium ; grant 2.4513.11 ) ; e.p . is supported by prodex .
the data were partially reduced using the publicly available data reduction software package _ amdlib _
, kindly provided by the jean - marie mariotti center .
this study used the simbad and vizier databases at the cds , strasbourg ( france ) , and nasas ads bibliographic services .
s. , mella g. , bonneau d. , et al .
, 2010 , in w. c. danchi et al . ed . , optical and infrared interferometry ii vol .
7734 of spie conf .
, building the jmmc stellar diameters catalog using searchcal .
san diego , ca s. , mrand a. , le bouquin j .- b .
, et al . , 2011 , in the origin and fate of the sun : evolution of solar - mass stars observed with high angular resolution problems of the amber wavelength calibration . to be published , garching b. mnchen , germany i. , tallon m. , thibaut e. , et al . , 2008 , in optical and infrared interferometry vol
7013 of spie conf .
, litpro : a model fitting software for optical interferometry .
marseille , france , pp 70131j-170131j-9 | extracting stellar fundamental parameters from spectro - interferometric ( spi ) data requires reliable estimates of observables and with robust uncertainties ( visibility , triple product , phase closure ) .
a number of fine calibration procedures is necessary throughout the reduction process .
testing departures from centro - symmetry of brightness distributions is a useful complement . developing a set of automatic routines , called _ spidast _
( made available to the community ) to reduce , calibrate and interpret raw data sets of instantaneous spectro - interferograms at the spectral channel level , we complement ( and in some respects improve ) the ones contained in the _ amdlib _ data reduction software .
our new software _
spidast _ is designed to work in an automatic mode , free from subjective choices , while being versatile enough to suit various processing strategies .
_ spidast _ performs the following automated operations : weighting of non - aberrant spi data ( visibility , triple product ) , fine spectral calibration ( sub - pixel level ) , accurate and robust determinations of stellar diameters for calibrator sources ( and their uncertainties as well ) , correction for the degradations of the interferometer response in visibility and triple product , calculation of the centro - symmetry parameter ( csp ) from the calibrated triple product , fit of parametric chromatic models on spi observables , to extract model parameters .
_ spidast _ is currently applied to the scientific study of 18 cool giant and supergiant stars , observed with the vlti / amber facility at medium resolution in the k band . because part of their calibrators have no diameter in the current catalogs , _
spidast _ provides new determinations of the angular diameters of all calibrators .
comparison of _ spidast _ final calibrated observables with _ amdlib _
determinations shows good agreement , under good and poor seeing conditions .
[ firstpage ] methods : data analysis techniques : interferometric stars : late - type |
ruthenium oxide based perovskites are presently of great interest due to the range of magnetic and strongly correlated electronic behavior they exhibit , often breaching the limits of current condensed matter theory .
layered sr@xmath0ruo@xmath5 , for example , is a superconductor at @xmath6 1 k,@xcite and is postulated to have _
p_-wave and spin - triplet pairing in the superconduting state.@xcite of particular interest is the delicate balance many ruthenate perovskites display between ferromagnetism and exotic electronic states .
when superconducting sr@xmath0ruo@xmath5 is doped by small amounts of nonmagnetic ti@xmath7 on the ruthenium site , there is a cross - over from superconductivity to short range ferromagnetic ordering.@xcite in the same structural system , sr@xmath8ru@xmath0o@xmath3 shifts from a paramagnetic , strongly correlated fermi - liquid to a ferromagnet under applied pressure.@xcite also , at fields above 7 t the compound undergoes a metamagnetic quantum transition to a high moment state.@xcite finally , ti@xmath7 doping on the ruthenium site of caruo@xmath8 induces ferromagnetism in what was originally a paramagnetic material.@xcite as much of the recent work focuses on perovskite based structures , the question follows as to whether the observed balance between itinerant and ferromagnetic behavior is characteristic of ruthenium oxides in general , or specific to perovskites .
therefore we have investigated another common oxide structure , the pyrochlore .
the pyrochlore oxide bi@xmath0ru@xmath0o@xmath3 , which exhibits temperature independent pauli paramagnetism and is weakly metallic , affords an excellent opportunity to test the generality of this balance . like the perovskite
, the pyrochlore structure , a@xmath0m@xmath0o@xmath3 , contains a three dimensional network of corner - sharing mo@xmath9 octahedra ( fig .
[ structure ] ) . in the perovskite
these are regular mo@xmath9 octahedra with 160 - 180@xmath10 m - o - m bonds , forming a three dimensional square net of metal atoms . however , in the pyrochlore the corner - sharing results in tetrahedrally related metal atoms , and coordination is such that the m - o - m bond angle must be @xmath11 130@xmath10 .
the a cation in the pyrochlore occupies hexagonal channels along @xmath6011@xmath12 and equivalent directions .
three dimensionally , the cations form an a@xmath5o sublattice of corner sharing tetrahedra .
view of ruo@xmath9 network in a ) perovskite along @xmath6100@xmath12 and b ) pyrochlore along @xmath6011@xmath12.,width=288 ] previous experimental work has probed differences between bi@xmath0ru@xmath0o@xmath3 , y@xmath0ru@xmath0o@xmath3 , and ln@xmath0ru@xmath0o@xmath3 , the latter two being electronically insulating although isoelectronic to bi@xmath0ru@xmath0o@xmath3.@xcite one proposed reason for the difference is the change in ru - o - ru angle with a cation size.@xcite it is also postulated that in the compounds y@xmath0ru@xmath0o@xmath3 and ln@xmath0ru@xmath0o@xmath3 the cation valence orbitals are energetically prohibited from contributing to fermi level bonding , resulting in narrower localized bands.@xcite reports are conflicting as to whether bismuth _ s _ and _ p _ orbitals hybridize to some extent at the fermi level , participating in conduction.@xcite experimental work has followed the metal to insulator transition in several of these solid solutions.@xcite here we report the properties of bi@xmath0@xmath1@xmath2m@xmath2ru@xmath0o@xmath3 , m = ( mn , fe , co , ni , cu ) , observing that the first row transition metal d - orbitals are proximate to the energy of the fermi level of bi@xmath0ru@xmath0o@xmath3 . thus it is likely that the overall properties of these doped materials will be markedly different from those previously described in the literature .
suitable stoichiometric amounts of ruo@xmath0 , bi@xmath0o@xmath8 , co@xmath8o@xmath5 , cuo , mno@xmath0 , fe@xmath0o@xmath8 , and ni@xmath0o@xmath8 were thoroughly ground in an agate mortar and pressed into pellets .
ruo@xmath0 was dried for 2 hours at 900 @xmath10c before use .
the pellets were successively heated in air at 750 @xmath10c for 1 day , 900 @xmath10c for 1 day , 950 @xmath10c for 3 days , and 975 @xmath10c for 1 day , with several intermediate grindings .
iron and nickel samples were heated for an additional day at 1020 @xmath10c .
sample purity was monitored using room temperature powder x - ray diffraction employing cu - k@xmath13 radiation .
magnetic properties and electrical resistivity were evaluated using a quantum design physical property measurement system ( ppms ) .
for all samples , the suseptibilty was measured from 300 to 5 k in an applied field of 1 t. data were also collected from 2 to 30 k in a field of 1 t after zero - field cooling , to assess any low temperature magnetic transitions .
magnetic hysteresis loops were performed at 5 k in the range of -9 to 9 t , and the data were fit to the paramagnetic brillouin function @xmath14 where @xmath15.@xcite resistivity measurements were performed on sintered powder pellets , which were cut into bars of approximately [email protected]@xmath163.5 mm , and measured with the standard 4-point ac method .
seebeck coefficient measurements employed a commercial apparatus ( mmr technologies ) within the temperature range 200 - 500 k.
substitution of first row transition metals on the ruthenium site of bi@xmath0ru@xmath0o@xmath3 was not possible within the synthetic conditions explored .
the first row transition metals mn , fe , co , ni , and cu instead substitute on the bismuth site of the pyrochlore .
we report complete structural characterization of bi@xmath17cu@xmath18ru@xmath0o@xmath3 , bi@xmath17co@xmath18ru@xmath0o@xmath3 , and bi@xmath0ru@xmath0o@xmath3 elsewhere,@xcite by refinement of neutron powder diffraction data .
this type of substitution is not unprecedented , as synthesis of bi@xmath19cu@xmath2ir@xmath20ru@xmath21o@xmath3 was reported , although the compound was not fully characterized.@xcite cubic lattice parameter , determined by powder x - ray diffraction , for bi@xmath0@xmath1@xmath2m@xmath2ru@xmath0o@xmath3 ( m = cu , ni , co ) as a function of doping level.,width=288 ] figure [ lattice ] plots lattice parameters obtained from powder x - ray diffraction as a function of x for bi@xmath0@xmath1@xmath2m@xmath2ru@xmath0o@xmath3 ( m = cu , ni , co ) .
to a first approximation the lattice parameters decrease linearly .
this is consistent with the smaller radii of the first row transition metals in comparison to that of bismuth . for copper , nickel , and cobalt dopants ,
the solid solution exists up to x=0.5 : at larger nominal concentrations the lattice parameters remain relatively constant . in several samples ,
ruo@xmath0 is found as a very minor impurity .
for example , the cobalt - doped sample with x=0.4 , contains 0.5% ruo@xmath0 by weight , as determined by powder neutron diffraction refinement.@xcite as ruo@xmath0 is non - magnetic , and the amount present is negligible , its presence will not affect physical property characterization . for iron and manganese - doped series the substitution limit is x=0.1 , resulting in unit cells of a=10.282(1 ) and 10.279(2 ) respectively .
the magnetic suseptibility ( @xmath22 ) of bi@xmath0ru@xmath0o@xmath3 between 300 and 5 k is plotted in figure [ chi_cu ] .
it is essentially temperature measured magnetic suseptibility as a function of temperature for bi@xmath0@xmath1@xmath2cu@xmath2ru@xmath0o@xmath3.,width=288 ] independent with a relatively small magnitude of 2 @xmath16 10@xmath23 emu / oe mol f.u . , consistent with pauli paramagnetism .
also plotted in figure [ chi_cu ] is the suseptibility of bi@xmath17cu@xmath18ru@xmath0o@xmath3 and bi@xmath24cu@xmath25ru@xmath0o@xmath3 .
it can be seen that x=0.4 copper - doping increases the magnitude of the temperature independent suseptibility by a factor of two .
this may be due to an enhancement of the metallic density of states .
there is not any significant indication of moment localization for either copper or ruthenium in these samples .
each does display a small curie tail at low temperature , likely due to the presence of microscopic amounts of impurity spins .
seebeck coefficients ( s ) for bi@xmath0@xmath1@xmath2cu@xmath2ru@xmath0o@xmath3 are plotted in figure [ s_cu ] .
( small peaks evident in the data around 300 k are seebeck coefficients for bi@xmath0@xmath1@xmath2m@xmath2ru@xmath0o@xmath3 ( m = cu , ni ) as a function of temperature.,width=288 ] instrumental . )
it can be seen that values of s for undoped bi@xmath0ru@xmath0o@xmath3are between -11 and -4 @xmath26v / k , indicating that the dominant carriers are electrons .
ten percent copper doping ( x=0.2 ) increases s , though the values remain negative , between -3 and -1 @xmath26v / k .
however with twenty percent ( x=0.4 ) doping , the seebeck coefficients become positive , and the dominant carriers cross over to hole - like .
resistivity data are presented in figure [ r_cu ] .
the resistivity of undoped bi@xmath0ru@xmath0o@xmath3 is largely temperature independent and on the order of 1 m@xmath27 cm , consistent with the description of a poor metal .
copper doping slightly decreases the magnitude of the resistivity but maintains a similar temperature dependence compared to the undoped sample .
resistivity as a function of temperature for bi@xmath0@xmath1@xmath2m@xmath2ru@xmath0o@xmath3 ( m = cu , ni).,width=288 ] magnetic suseptibility data for nickel - doped samples are shown in figure [ chi_ni ] .
the top panel plots measured suseptibility as a function of temperature .
in contrast to copper doping , these top panel : measured magnetic suseptibility as a function of temperature for bi@xmath0@xmath1@xmath2ni@xmath2ru@xmath0o@xmath3 .
bottom panel : inverse magnetic suseptiblity ( points ) and curie - weiss fits ( solid lines ) for bi@xmath0@xmath1@xmath2ni@xmath2ru@xmath0o@xmath3 .
fitted temperature independent terms ( @xmath28 ) are subtracted from measured @xmath22.,width=288 ] samples begin to display local moment behavior as nickel is substituted .
the curie - tail at low temperature that is visible on the undoped pyrochlore becomes more pronounced with 5% ( x=0.1 ) nickel doping .
by 20% ( x=0.4 ) nickel doping , the magnetic suseptibility scales with the curie - weiss equation for localized paramagnetic moments . data from nickel - doped samples for x=0.3 and x=0.4 were fit between 150 - 300 k to the function @xmath29 where @xmath28 is the sum of all temperature independent terms .
values for the effective moment and curie - weiss temperature ( @xmath30 ) were extracted , with the results summarized in table [ fits ] .
moments are between 2.1 and 2.3 @xmath31/ni , only slightly less than the theoretical spin - only value of 2.82 for ni@xmath4 .
therefore we assume that the observed local moments are accounted for by the nickel dopant , and that no local moment has been induced at ruthenium centers .
further , values of @xmath30 are very small , indicating weak coupling between the localized spins in all cases . .[fits ] calculated magnetic constants from curie - weiss fitting . [ cols="^,^,^,^ " , ] the bottom panel of figure [ chi_ni ] plots inverse magnetic suseptibility , 1/(@xmath32 ) , for the nickel - doped samples .
experimental points are overlayed by curie - weiss fits ( solid lines ) for x=0.3 and x=0.4 .
the overlays highlight evident deviation from ideal behavior below 100 k for the two samples . and in fact , at nickel doping levels of x=0.2 and lower , deviation from linear behavior is such that curie - weiss fitting is not applicable .
thus the magnetism can be classified as intermediate to localized and itinerant behavior .
seebeck coefficients ( s ) for bi@xmath0@xmath1@xmath2ni@xmath2ru@xmath0o@xmath3 are presented in figure [ s_cu ] . at 200 k
, the magnitude of the coefficient goes from -10 in undoped bi@xmath0ru@xmath0o@xmath3 to + 10 @xmath26v / k in the 10@xmath33 nickel - doped sample .
the nickel - doped samples have higher s values than the copper - doped samples , indicating that there is a higher concentration of hole carriers in the former .
resistivity ( @xmath34 ) behavior is also different for the copper and nickel - doped samples .
figure [ r_cu ] illustrates that while copper - doping decreases the magnitude of the resistivity , nickel - doping increases it . as nickel
is substituted on the lattice of bi@xmath0ru@xmath0o@xmath3 the resistivity becomes more like that of a degenerate semi - conductor , where @xmath34 begins to increase slightly with decreasing temperature .
cobalt doping has a dramatic effect on the magnetic properties of bi@xmath0ru@xmath0o@xmath3 .
samples of bi@xmath0@xmath1@xmath2co@xmath2ru@xmath0o@xmath3 display curie - weiss behavior for x=0.1 to 0.5 in the temperature range 300 - 5 k and in 1 t .
plots of 1/@xmath35 versus temperature ( fig .
[ chi_co ] ) demonstrate the inverse magnetic suseptibility ( points ) and curie - weiss fits ( solid lines ) for bi@xmath0@xmath1@xmath2co@xmath2ru@xmath0o@xmath3 as a function of temperature .
fitted temperature independent terms ( @xmath28 ) are substracted from measured @xmath22 .
insets plot low temperature magnetic suseptibility down to 2 k.,width=288 ] goodness of fit , where calculated values ( solid lines ) overlay experimental points .
the effective moment normalized to cobalt content ( table [ fits ] ) remains essentially constant across the doping series , indicating that the curie moments are due solely to unpaired cobalt d electrons . again
no local moment is induced on ruthenium centers with doping .
calculated moments are between 4.1 and 4.4 @xmath31/co , which are slightly greater than the spin - only theoretical value of 3.87 for co@xmath4 but consistent with previously reported values for high spin co@xmath4 of 4.1 to 5.2 @xmath31/co.@xcite values of theta are close to zero for all cobalt doped samples , signifying that spin - coupling is very weak .
below 5 @xmath10k , small peaks in magnetic suseptibility are observed for x=0.4 and x=0.5 samples , possibily due to spin - glass transitions ( fig .
[ chi_co ] , inset ) .
the fact that the transitions are at such low temperatures is consistent with weak coupling of the spins .
additionally , the peak maximum increases slightly from samples x=0.4 to x=0.5 .
in fact , although there is not a distinct suseptibility peak for sample x=0.3 , there are significant deviations from curie - weiss behavior below 5 k , signifying that some degree of interaction is present already at this doping level .
the doped pyrochlores have the two characteristics common to spin - glass systems ; atomic disorder and frustrated geometry .
magnetization ( m ) versus field ( h ) curves are presented in figure [ loop_co ] for bi@xmath0@xmath1@xmath2co@xmath2ru@xmath0o@xmath3 , where data are collected at 5 k. no magnetic hysteresis is present for any doping level .
the curves are not magnetization as a function of field for bi@xmath0@xmath1@xmath2co@xmath2ru@xmath0o@xmath3 .
inset : expansion of measured magnetization versus field data for undoped bi@xmath0ru@xmath0o@xmath3 in low temperature region .
units are : @xmath22 ( emu / oe mol f.u . ) , t ( k ) .
, width=288 ] linear , but instead scale with the paramagnetic brillion function , as m@xmath31h is on the order of k@xmath36 t at this temperature and field range . although m vs h for undoped
bi@xmath0ru@xmath0o@xmath3 does appear to be flat and linear at this scale , the figure inset shows the true shape of the curve . fitting the high field data for undoped bi@xmath0ru@xmath0o@xmath3to the brillouin function with g=2 , the number of localized spin 1 moments is approximated to be 0.002 spins / mol f.u .
for bi@xmath0@xmath1@xmath2co@xmath2ru@xmath0o@xmath3 the magnetization increases with the amount of cobalt dopant as is expected .
however , attempts to fit these curves to the brillion function were unsuccessful .
this is due to the fact that the data were collected at 5 k , where an assumption of completely localized non - interacting moments is inappropriate for these samples .
figure [ ht_co ] plots magnetization as a function of h / t for magnetization as a function of h / t for bi@xmath17co@xmath18ru@xmath0o@xmath3 , with data sets collected at three different temperatures.,width=288 ] sample bi@xmath17co@xmath18ru@xmath0o@xmath3 at different temperatures .
if the data follow the paramagnetic brillouin function , then plots taken at different temperatures should superimpose . in figure 9
it is evident that plots for 25 k and 55 k data essentially overlay one another .
however , data taken at 5 k deviates significantly .
therefore the cobalt moments are truly independent at higher temperatures , but by 5 k , some degree of interaction is present between the cobalt centers .
this is supported by our suseptibility data , where possible spin - glass ordering is present below 5 k for the more heavily doped samples .
it is surprising that such a large number of magnetic co@xmath4 atoms are accommodated in the pyrochlore without inducing magnetism on the ruthenium sublattice , and that the co@xmath4 spins are so weakly interacting .
seebeck coefficients and resistivity for bi@xmath0@xmath1@xmath2co@xmath2ru@xmath0o@xmath3 are plotted in figures [ s_co ] and [ r_co ] respectively .
seebeck coefficients increase seebeck coefficients as a function of temperature for bi@xmath0@xmath1@xmath2co@xmath2ru@xmath0o@xmath3.,width=288 ] resistivity as a function of temperature for bi@xmath0@xmath1@xmath2co@xmath2ru@xmath0o@xmath3.,width=288 ] with the amount of cobalt doping , reaching a maximum of + 18 @xmath26v / k for x=0.4 . once again , the first row transition metal is hole - doping the pyrochlore .
the slope of the temperature dependence of s also becomes flatter with increasing cobalt .
the resistivity behavior of bi@xmath0@xmath1@xmath2co@xmath2ru@xmath0o@xmath3 is similar to that of bi@xmath0@xmath1@xmath2ni@xmath2ru@xmath0o@xmath3 .
as dopant is introduced , the magnitude of the resistivity increases , and the temperature dependence becomes more semi - conductor like .
the increased resistivity may be due to several factors , one of which is the scattering of electrons by the localized dopant moments .
this explanation is supported by the fact that in copper - doped samples , where there are no localized spins , the resistivity decreases slightly , instead of increasing . however , the resisitivity of bi@xmath17co@xmath18ru@xmath0o@xmath3 is unaffected by a field of 9 t ( figure [ cmr_co ] ) .
this is not consistent with the description of spin scattering , where at 9 t the spins should align and thus decrease the resistance .
therefore the origin of resistivity behavior in copper and nickel doped bi@xmath0ru@xmath0o@xmath3 is unresolved .
resistivity as a function of temperature for bi@xmath17co@xmath18ru@xmath0o@xmath3 at 0 tesla and 9 tesla.,width=288 ] inverse suseptibility ( points ) and curie - weiss fits ( solid lines ) for bi@xmath37mn@xmath38ru@xmath0o@xmath3 and bi@xmath37fe@xmath38ru@xmath0o@xmath3 as a function of temperature .
fitted temperature independent terms ( @xmath28 ) are subtracted from measured @xmath22 .
insets plot measured magnetic suseptibility down to 2 k. units are : @xmath22 ( emu / oe mol f.u . ) , t ( k ) . , width=288 ] for dopants fe and mn , the substitution limit is 5@xmath33 ( x=0.1 ) .
inverse suseptibility for these two samples is plotted in figure [ chi_mn ] , with overlayed curie - weiss fits ( solid lines ) .
both bi@xmath37fe@xmath38ru@xmath0o@xmath3 and bi@xmath37mn@xmath38ru@xmath0o@xmath3 obey the curie - weiss law in the temperature range 300 - 5 k. magnetic constants are listed in table [ fits ] .
fitted magnetic moments are 4.9 and 5.1 @xmath31/x for fe and mn respectively .
these values are in good agreement with the spin - only theoretical value of 4.89 @xmath31 for s=2 , again indicating that the observed localized moments are due to the dopants .
values of @xmath30 are close to zero , indicating that interaction between the magnetic centers are negligible .
in fact , the manganese spins are truly independent even down to low temperature .
this is illustrated in figure [ b_mn ] , where magnetization is plotted as a function of field for bi@xmath37mn@xmath38ru@xmath0o@xmath3 .
the high field data were fit to the paramagnetic brillouin function with set values g=2 and j=2 , and with only n ( number of spins per mole ) allowed to vary . assuming that all localized spins originate from manganese , the resulting value of x found for bi@xmath37mn@xmath38ru@xmath0o@xmath3 was 0.10 , in agreement with the nominal concentration .
experimental magnetization vs. field data for bi@xmath37mn@xmath38ru@xmath0o@xmath3 and paramagnetic brillouin function ( dotted line).,width=288 ] resistivity ( fig .
[ r_mn])and seebeck coefficients ( fig
. [ s_mn ] ) for mn and fe - doped samples follow similar trends as those seen for ni and co. the introduction of dopant increases the magnitude of the resistivity , and samples become more semi - conducting . for fe and mn - doping
, seebeck coefficients also increase relative to undoped bi@xmath0ru@xmath0o@xmath3 indicating that the transition metals are once again hole doping the pyrochlore .
resistivity as a function of temperature for bi@xmath0@xmath1@xmath2m@xmath2ru@xmath0o@xmath3 ( m = fe , mn).,width=288 ] seebeck coefficient as a function of temperature for bi@xmath0@xmath1@xmath2m@xmath2ru@xmath0o@xmath3 ( m = fe , mn).,width=288 ]
the pyrochlore structure can be described as two interpenetrating networks ; one of corner - sharing ruo@xmath9 polyhedra and one of corner - sharing bi@xmath5o tetrahedra .
our magnetic data on transition metal doped bi@xmath0ru@xmath0o@xmath3 indicates that the two networks are , surprisingly , magnetically independent .
we successfully substituted the first row transition metal series mn@xmath39 , fe@xmath4 , co@xmath4 , ni@xmath4 , and cu@xmath4 on the bismuth site of bi@xmath0ru@xmath0o@xmath3 .
each of these dopants has unpaired valence shell d - electrons . however , the ruthenium moments in the pyrochlore are unaffected by bi - site doping .
this is in strong contrast to results for the doped perovskites caru@xmath40ti@xmath2o@xmath8 and sr@xmath0ru@xmath40ti@xmath2o@xmath5 .
although we find no ruthenium local moments , the unpaired magnetic suseptbilibty at 300 k ( top panel ) and 5 k ( bottom panel ) as a function of dopant for copper , nickel , and cobalt doped samples.,width=288 ] dopant electrons diverge from localized to itinerant magnetic behavior across the first row transition metal series . first , in bi@xmath37mn@xmath38ru@xmath0o@xmath3 the substituted manganese moments are localized and essentially noninteracting , even down to low temperature . in bi@xmath0@xmath1@xmath2co@xmath2ru@xmath0o@xmath3 the cobalt moments are also localized , with suseptibility data fitting the curie - weiss law between 5 - 300 k. however , at low temperature the spins are no longer completely isolated .
in fact , spin - glass transitions are present below 5 k for bi@xmath17co@xmath18ru@xmath0o@xmath3 and bi@xmath41co@xmath42ru@xmath0o@xmath3 . in nickel - doped samples the unpaired nickel spins display behavior intermediate to localized and itinerant .
finally , in copper - doped samples , there is no evidence for moment localization , and only the metallic temperature independent paramagnetism is enhanced with doping .
these magnetic trends are illustrated in figure [ c1 ] , where the measured suseptibility at 5 k and 300 k is plotted as a function of dopant .
for localized moments , the magnetic suseptibility at low temperature increases observably as a function of x , such as in the case of cobalt .
however , for itinerant moments as in copper , @xmath43 remains relatively flat with increasing x. for nickel , with atomic number intermediate to co and cu , the behavior is a combination of localized and itinerant .
up to x=0.1 @xmath43 remains flat , while above x=0.1 @xmath43 increases with x , although not as steeply as @xmath43 in the case of cobalt . across the series ,
the d - orbitals of the first row transition metal become more proximate in energy to the valance orbitals of oxygen .
thus copper hybridizes more strongly with oxygen than cobalt or nickel , resulting in itinerant behavior .
seebeck coefficients as a function of dopant concentration for copper , nickel , and cobalt doped bi@xmath0ru@xmath0o@xmath3.,width=288 ] the transport properties of bi@xmath0@xmath1@xmath2m@xmath2ru@xmath0o@xmath3 are also evaluated .
figure [ c2 ] displays s as a function of dopant concentration for co , ni , and cu doped samples . at x=0.4 , values of s
are all positive .
thus in each series the dominant carriers cross - over from electrons to holes upon doping .
the absolute value of the seebeck coefficient , @xmath44s@xmath44 , is proportional to log(n)-log(n ) where n is the concentration of carriers and n the available density of states . if n is essentially constant , then copper doping results in an increased concentration of carriers ( n ) relative to undoped bi@xmath0ru@xmath0o@xmath3 . this interpretation agrees with resistivity data , where the magnitude of @xmath34 decreases as copper is substituted on the pyrochlore lattice .
conversely , seebeck coefficients for cobalt and nickel samples increase more steeply with x than those for copper . at x=0.4 for these two dopants , the concentration of carriers ( n ) is decreased relative to undoped bi@xmath0ru@xmath0o@xmath3 .
in summary , we report magnetic and transport properties of bi@xmath0@xmath1@xmath2m@xmath2ru@xmath0o@xmath3(m = mn , fe , co , ni , cu ) . in no case did ferromagnetism or other localized ruthenium moment behavior occur upon doping .
however , in such previously reported cases as caru@xmath40ti@xmath2o@xmath8 and sr@xmath0ru@xmath40ti@xmath2o@xmath5 , the dopant was substituted on the ruthenium site .
thus those substitutions could be expected to strongly perturb the ru - o network . in the present study ,
the transition metal dopants preferentially replace bismuth in bi@xmath0ru@xmath0o@xmath3 .
our magnetic data indicates that the ruthenium and bismuth sublattices in bi@xmath0ru@xmath0o@xmath3 are magnetically independent , and therefore the metal substitutions represent a second - order perturbation of the ru - o lattice , which was not effective in causing a crossover to the ferromagnetic state . in caruo@xmath8 ,
conversely , a - site substitution does induce ferromagnetism even though it is not a direct perturbation of the ru - o lattice.@xcite therefore ruthenium oxides in the pyrochlore geometry appear to be far from the ferromagnetic instability observed in the perovskites , and such instability is not a broadly general characteristic of ruthenium oxide compounds .
this work was supported by the national science foundation grant no .
the work at argonne national laboratory was supported by the us department of energy , office of basic energy science , contract no .
w-31 - 109-eng-38 . | we report physical property characterization of bi@xmath0@xmath1@xmath2m@xmath2ru@xmath0o@xmath3 pyrochlores , including magnetic suseptibility , resistivity , and seebeck coefficients .
the solid solution exists up to x=0.5 for ( m = cu , ni , co ) and up to x=0.1 for ( m = fe , mn ) .
none of the doped materials exhibit ferromagnetism or any localized ruthenium moment behavior .
instead we find the ru - o and bi - o sublattices to be essentially independent , with any magnetism resulting from the unpaired transition metal dopant spins .
cobalt substitution for bismuth results in localized co@xmath4 , and low temperature spin - glass transitions in several cases .
nickel moments on the pyrochlore lattice display properties intermediate to localized and itinerant .
finally , copper doping results in only an enhancement of the pauli metallic density of states . |
by an anick space we mean a homotopy cw complex @xmath1 which occurs in a fibration sequence . @xmath2^{\pi_n } & s^{2n-1}\ar@{->}[r ] & t_{2n-1}\ar@{->}[r ] & \omega s^{2n+1 } } \ ] ] where the composition @xmath3^{\pi_n } & s^{2n-1}\ar@{->}[r]^{e^2 } & \omega^2s^{2n+1 } } \ ] ] is homotopic to the @xmath4th power map on @xmath5 .
we will say that a space is a homotopy - abelian @xmath0-space if it has a homotopy associative and homotopy commutative @xmath0-space structure . throughout this work
we will assume that all spaces are localized at @xmath6 unless otherwise indicated .
there is a homotopy - abelian anick space for any @xmath7 and @xmath8 .
we will write @xmath9 for the @xmath10 dimensional moore space @xmath11 . from ( [ eq1.1a ] )
we see that @xmath1 is @xmath12 connected and there is a @xmath10-equivalence @xmath13 for any homotopy - abelian @xmath0-space @xmath14 , let @xmath15_h$ ] be the abelian group of homotopy classes of based @xmath0-maps from @xmath1 to @xmath14 .
let @xmath16 @xmath15_h\cong \underset{\leftarrow}{\lim } g_k(z)$ ] where @xmath17 = \pi_{2n}(z;z / p^r)$ ] and there are exact sequences : @xmath18&p_k(\omega z)\ar@{->}[r]^{{\scalebox{.9}{${e}$}}}&g_k(z)\ar@{->}[r]^{{\scalebox{.9}{${r}$}}}&g_{k-1}(z)\ar@{->}[r]^{{\scalebox{.9}{${\beta}$}}}&p_k(z ) . }
\ ] ] in particular , if @xmath19 , there is an isomorphism @xmath20\approx[t_{2n-1 } , z]_h\ ] ] given by the restriction @xmath21 several examples are given in chapter [ chap7 ] .
in particular given two homotopy - abelian anick spaces for the same values of @xmath22 , @xmath23 and @xmath24 , there is an @xmath0 map between them which is a homotopy equivalence . in any homotopy - abelian @xmath0-space structure on an anick space
, the identity map has order @xmath4 .
. then analogous results to theorems a and b are well known ( @xcite , @xcite ) when @xmath24 , and in particular , any map @xmath26 corresponds to a unique @xmath0-map @xmath27 such that the diagram @xmath28^{\alpha}\ar@{->}[d]&p^{m+1}(p^r)\ar@{->}[d]\\ t_n\ar@{->}[r]^{\widehat{\alpha}}&t_m } \ ] ] commutes up to homotopy , for any @xmath22 and @xmath29 .
this corollary was the object of the conjectures in @xcite .
in developing these results , several new techniques of geometric homotopy theory are introduced .
these may be of some use in other problems .
a summary of some of these techniques can be found in section [ subsec1.4 ] . in an appendix
, we show that these results do not generally hold if @xmath30 , and we treat the special case when @xmath31 .
the author would like to thank joseph neisendorfer for many helpful conversations during this work .
a map @xmath32 with the property that the composition @xmath3^{\pi_n}&s^{2n-1}\ar@{->}[r]^{e^2}&\omega^2s^{2n+1 } } \ ] ] is homotopic to the @xmath4th power map was first discovered when @xmath33 by cohen , moore and neisendorfer ( @xcite ) and played a crucial role in determining the maximal exponent for the torsion in the homotopy groups of spheres . in @xcite
, the authors raised the question of whether a fibration such as [ eq1.1a ] could exist .
the feasibility of constructing a secondary ehp sequence refining the secondary suspension ( @xcite , @xcite ) together with a theory of compositions ( @xcite ) was studied in @xcite , @xcite .
this led the author to conjecture the existence of anick spaces with theorems a and b and corollaries c and d. at about the same time , david anick was studying the decomposition of the loop space on a finite complex ( @xcite ) .
his intention was to find a list of indecomposable spaces which , away from a few small primes , could be used for decomposition .
this led to the construction of a sequence of spaces for @xmath6 .
the limit of this sequence is the space sought after in @xcite .
this work of anick was published in a 370-page book ( @xcite ) . in @xcite
, the authors showed that the anick space so constructed admitted an @xmath0-space structure when @xmath6 .
they also proved a weaker version of theorem b. they showed that if @xmath34 for all @xmath35 , an extension to an anick space existed , but there was no indication that the extension would be an @xmath0-map or that it would be unique .
at that time it was thought that the torsion condition was a peculiarity of the approach and , it was conjectured that , as in the case of @xmath36 , this requirement was unnecessary . in @xcite ,
the author asserted theorem a and corollary c and that for each homotopy abelian @xmath0-space @xmath14 , there is an equivalence @xmath37\simeq[t_{2n-1},z]_h$ ] .
this , however , is not consistent with the results in @xcite , where counter examples are provided with @xmath14 is as eilenberg
maclane space .
the assertions in @xcite depend on the author s theorem 2.1 which is quoted as `` to appear in topology '' in the author s bibliography .
this , however , did not appear .
many of the results in the author s section 5 are inconsistent with results we obtain here .
the main result of the author s section 4 is valid and we give a much simplified proof of it here ( 2.9 ) . in @xcite a much simpler construction of the anick spaces was obtained which worked for all @xmath33 .
this result replicated the results of @xcite and @xcite and extended them to the case @xmath30 .
furthermore they showed that the homotopy type of an anick space that supports an @xmath0-space structure is uniquely characterized by @xmath22 , @xmath38 and @xmath23 . in @xcite , a proof was given that if @xmath39 for all @xmath35 , there is at most one extension of a map @xmath40 to an @xmath0-map @xmath41 and examples were presented to show that this torsion condition is necessary .
the proof we give here is entirely different . in @xcite
, the authors constructed the ehp sequences conjectured in @xcite .
@xmath42^{e}&\omega t_{2n}\ar@{->}[r]^h & bw_n\\ & t_{2n}\ar@{->}[r]^-<<<<{e}&\omega t_{2n+1}\ar@{->}[r]^{h}&bw_{n+1 } } \end{aligned}\ ] ] ( where @xmath36 ) , and @xmath43 lies in a fibration sequence : @xmath44^{e^2}&\omega^2s^{2n+1}\ar@{->}[r]^{\nu}&bw_n . } \ ] ] more details appear in the next section .
the only remaining unsettled conjecture in @xcite for the anick spaces is the following : there is a homotopy equivalence @xmath45 where @xmath46 is the anick space with @xmath47 . in @xcite ,
theriault constructed @xmath48 for @xmath49 , but there is no @xmath0-space structure in this case . throughout this work
we will fix @xmath22 and abbreviate @xmath1 as @xmath50 if this will not lead to confusion .
the construction in @xcite begins with a fibration sequence : @xmath51&s^{2n-1}\ar@{->}[r]^{{\scalebox{.9}{${e^2}$ } } } & \omega^2s^{2n+1}\ar@{->}[r]^{{\scalebox{.9}{${\nu}$}}}&bw_n } \ ] ] where @xmath52 is the double suspension map ( @xcite ) .
the authors construct a factorization of the map @xmath53 : @xmath3 ^ ->>>>{{\scalebox{.9}{${\omega\partial}$ } } } & \omega s^{2n+1}\{p^r\}\ar@{->}[r]^-<<<{{\scalebox{.9}{${h}$}}}&bw_n } \ ] ] where @xmath54 occurs in the fibration sequence defining @xmath55 , the fiber of the degree @xmath4 map on @xmath56 . for any choice of @xmath0
, there is a homotopy commutative diagram : @xmath57\ar@{->}[d]_{\pi_n}&\omega^2s^{2n+1}\ar@{->}[d]_{p^r}\\ s^{2n+1}\ar@{->}[r]\ar@{->}[d]&\omega^2s^{2n+1}\ar@{->}[r]^{\nu}\ar@{->}[d]_{\omega \partial}&bw_n\ar@{=}[d]\\ t\ar@{->}[r]^->>>>>>e\ar@{->}[d]&\omega s^{2n+1}\{p^r\}\ar@{->}[r]^->>>h\ar@{->}[d]&bw_n\\ \omega s^{2n+1}\ar@{=}[r]&\omega s^{2n+1 } } \end{split}\ ] ] and consequently for each choice of @xmath0 , the fiber of @xmath0 is an anick space by ( [ eq1.1a ] ) . in @xcite , the authors proceed to show that for any such choice , the anick space admits an @xmath0-space structure such that the fibration [ eq1.1a ] is an @xmath0-fibration . furthermore , they prove fix @xmath22 , @xmath23 and @xmath33 . then any two anick spaces which admit an @xmath0-space structure are homotopy equivalent .
the @xmath0-space structure is constructed in @xcite as follows .
the authors construct a splitting : @xmath58 where @xmath59 is atomic and @xmath60 is a @xmath61 connected wedge of moore spaces .
a map @xmath62 is constructed from ( [ eq1.5new ] ) and the adjoint of the map @xmath63 in ( [ eq1.3new ] ) @xmath64&\sigma t\ar@{->}[r]^->>>>{{\scalebox{.9}{${\widetilde{e}}$}}}&s^{2n+1}\{p^r\ } } \ ] ] and a space @xmath63 is constructed as the fiber of @xmath65 : @xmath66 ^ -<<<<{\partial } & e\ar@{->}[r]&g\ar@{->}[r]^->>>>{{\scalebox{.9}{${\varphi}$}}}&s^{2n+1}\{p^r\}. } \ ] ] the @xmath0-space structure arises from a factorization of @xmath0 through the space @xmath63 : @xmath67\ar@{->}[d]_{{\scalebox{.9}{${h}$}}}&\omega g\ar@{->}[d]\\ t\ar@{->}[r]^->>>>{{\scalebox{.9}{${e}$}}}\ar@{->}[d]&\omega s^{2n+1}\{p^r\}\ar@{->}[r]^-<<<{{\scalebox{.9}{${h}$}}}\ar@{->}[d]_{{\scalebox{.9}{${\partial'}$}}}&bw_n\ar@{=}[d]\\ r \ar@{->}[r]\ar@{->}[d]&e\ar@{->}[r]^{{\scalebox{.9}{${\nu_{\infty}}$}}}\ar@{->}[d]_{{\scalebox{.9}{${\pi}$ } } } & bw_n\\ g\ar@{=}[r]&g } \end{split}\ ] ] @xmath68 has a right homotopy inverse @xmath69 defined by the splitting ( [ eq1.5new ] ) since @xmath50 is atomic .
thus @xmath50 inherits an @xmath0 space structure as a retract of @xmath70 .
the @xmath0-space structure depends on @xmath68 and consequently on the choice of @xmath71 .
the splitting ( [ eq1.5new ] ) and map @xmath71 are defined inductively .
the space @xmath59 is filtered by subspaces @xmath72 where @xmath73 and the fibration @xmath74^{{\scalebox{.9}{${\pi}$}}}&g}$ ] is the union of the induced fibrations @xmath75 the splitting ( [ eq1.5new ] ) is approximated by a sequence of splittings : @xmath76 and @xmath71 is simultaneously constructed by induction over the restriction of ( [ eq1.6new ] ) to @xmath72 .
the map @xmath71 is then the limit of maps @xmath77 the extension theorem @xcite is applied which guarantees that any map @xmath78 extends to @xmath79 .
an arbitrary choice is made for each @xmath80 .
it seems likely that the number of choices for @xmath71 is uncountable .
the basis of this paper is to modify and sharpen the construction in @xcite as described in section [ subsec1.3 ] .
as explained there , the @xmath0-space properties that the anick space inherits depend on a choice of a map @xmath81 in chapter [ chap2 ] , we introduce maps @xmath82 and prove that if we choose @xmath83 such that @xmath84 is null homotopic for each @xmath35 , the induced @xmath0-space structure will be homotopy - abelian .
we also recall , at this point , various facts about the anick spaces that were developed in @xcite which will be needed in the sequel . in chapter [ chap3 ] , we recall from @xcite the construction of a whitehead product pairing @xmath85\times[h , x]\to [ g\circ h , x]\ ] ] where @xmath59 and @xmath0 are simply connected co-@xmath0 spaces and @xmath86 is a new simply connected co-@xmath0 space .
this generalizes the classical pairing : @xmath87\times [ \sigma b , x]\to [ \sigma a\wedge b , x].\ ] ] we also generalize neisendorfer s theory of relative whitehead products and @xmath0-space based whitehead products in the @xmath88 homotopy of a principal fibration @xcite by replacing moore spaces with arbitrary co-@xmath0 spaces .
we then reduce the question of whether @xmath84 is null homotopic to whether a sequence of iterated @xmath0-space based and relative whitehead products @xmath89 are annihilated by @xmath83 for all @xmath90 . in chapter [ chap4 ]
we construct @xmath91 homotopy classes @xmath92 and @xmath93 in @xmath79 for @xmath94 and a @xmath95 homotopy class @xmath96 which will play a key role .
we also introduce `` index @xmath38 approximation '' and show that the iterated whitehead product under investigation can be approximated by iterated whitehead products in homotopy groups with coefficients in @xmath97 for @xmath98 .
this approximation excludes the case @xmath31 which is handled in the appendix .
these classes are the obstructions and we seek to choose @xmath83 which annihilates them .
when @xmath80 , the obstructions actually belong to two classes , @xmath99 and @xmath100 depending on whether they involve the @xmath92 or the @xmath93 .
( it turns out that any obstruction involving both @xmath92 and @xmath101 is automatically in the kernel of @xmath83 . ) in [ subsec5.1 ] we simplify the procedure by defining a quotient space @xmath102 of @xmath79 which is universal for annihilating the classes as @xmath100 .
@xmath102 is a principal fibration over a space @xmath103 which is a quotient , defined differently , occurs in the original construction of anick ( @xcite ) .
the results of @xcite are obtained by replacing @xmath103 by @xmath72 .
] of @xmath72 and we seek a factorization of @xmath83 : @xmath104^{{\scalebox{.9}{${\tau_k}$ } } } & j_k\ar@{->}[r]^->>>>{{\scalebox{.9}{${\gamma_k}$ } } } & bw_n } \index{tauk@$\tau_k$|lb}\ ] ] where @xmath105 annihilates the obstructions in @xmath99 . in section [ subsec5.2 ]
we introduce a congruence relation among homotopy classes and the relative whitehead products and @xmath0-space based whitehead products have better properties in the congruence homotopy category .
this allows for a further reduction in obstructions to a collection of @xmath88 homotopy classes . in chapter [ chap6 ]
we introduce the controlled extension theorem ( [ theor6.1 ] ) .
this is a modification of the extension theorem in @xcite which allows maps defined on the total space of an induced fibration of a principal fibration over a subspace to be extended over the total space under certain conditions . in the controlled extension theorem ,
conditions are given for the extension to annihilate certain maps maps @xmath106 .
this is immediately applied to the case @xmath107 where we construct @xmath108 by induction over the skeleta of a space @xmath109 .
a complexity arises because for each @xmath80 , there are level @xmath35 obstructions in infinitely many dimensions .
when we modify @xmath83 to eliminate these obstructions , we ca nt assume that it will be an extension of @xmath110 and consequently the level @xmath111 obstructions may reappear .
a separate argument ( [ lem6.40 ] ) dispenses with this issue . in section [ subsec6.2 ] we introduce the inductive hypothesis ( [ theor6.7 ] ) and a space @xmath112 is analyzed to prepare for the inductive step .
this is accomplished in section [ subsec6.3 ] . in chapter [ chap7 ] , we discuss the universal properties of the anick spaces . from the fibration sequence ( [ eq1.6new ] )
we extract the following fibration sequence @xmath113 ^ -<<<{*}&\omega r\ar@{->}[r]&\omega g\ar@{->}[r]^{h}&t\ar@{->}[r]^{*}&\mbox { } } \ ] ] which we think of as a presentation of @xmath50 .
the proof of theorem b depends on an understanding of the map @xmath114 . from @xcite we know that for any choice of @xmath71 , @xmath115 is a wedge of moore spaces .
certain of these moore spaces are needed to resolve the relationship between @xmath116 , which has infinitely many generators and @xmath117 which has two generators .
these are the classes @xmath92 and @xmath93 .
the others are necessary to enforce the homotopy commutativity in @xmath118 .
these are either whitehead products or generalized whitehead products of order @xmath119 could not be a classical whitehead product for @xmath80 .
it is defined as a composition @xmath120&g_k\circ g_k\ar@{->}[r]^-<<<<{{\scalebox{.9}{${w}$}}}&g_k } \ ] ] where @xmath60 is a generalized whitehead product . ] defined by co-@xmath0-spaces in @xcite .
the obstruction to extensions depend on certain homotopy classes @xmath121 which must be annihilated in order for an extension to proceed . in the appendix
we discuss the case @xmath31 and the case @xmath30 .
all spaces will be localized at a prime @xmath33 and usually we will assume @xmath6 .
@xmath122 and @xmath123 will designate the @xmath124 homology and cohomology . if other coefficients are used (
usually @xmath125 ) they will be specified in the usual way .
we write @xmath126 for the moore space . throughout we will fix @xmath8 and we will always have @xmath127 .
we will abbreviate @xmath128 simply as @xmath129 .
we will write @xmath130 and @xmath131 for the usual maps @xmath132 ^ ->>>>{\iota_{m-1}}&p^m(p^s)\ar@{->}[r]^{\pi_m}&s^m . } \ ] ] we designate the symbols @xmath133 , @xmath134 , @xmath135 for the maps @xmath136 with @xmath137 , @xmath138 and @xmath139 .
these symbols will not be indexed by the dimension and can be composed , so that we have formulas @xmath140 where @xmath38 is the degree @xmath38 self map .
we write @xmath141 and will frequently use the cofibration sequence @xmath142 ^ ->>>>>>{-\delta_t\vee \rho^t}&p^m(p^{s+t})}\\ \xymatrix@c=33pt{\ar@{->}[r]^-<<<<{p^s}&p^m(p^{s+t})\ar@{->}[r]^->>>>>{\sigma^t\vee\sigma^t\beta}&p^m(p^s)\vee p^{m+1}(p^s ) } \end{gathered}\ ] ] especially when @xmath143 and @xmath144 . we will write @xmath145for the largest exponent of @xmath38 that divides @xmath29 and often set @xmath146 . for any map
@xmath147 we will write @xmath148 for either its left or right adjoint , if there is no possibility of confusion . by a diagram of fibration sequences ,
we will mean a diagram in which any sequences , either vertical or horizontal are fibration sequences up to homotopy .
we begin by reviewing some material about principal fibrations . in section [ subsec2.1 ] ,
we recall the construction of @xmath43 ( @xcite ) and the extension theorem ( @xcite ) , and for certain principal fibrations we construct a natural map @xmath149 in section [ subsec2.2 ] .
@xmath150 is the lynchpin for generalizing neisendorfer s @xmath0-space based whitehead products . in [ prop2.3 ]
we give a short proof of a result of theriault giving a criterion for an @xmath0-space structure to be homotopy - abelian .
we use this to relate the map @xmath150 to the obstructions for the induced @xmath0-space structure on the fiber being homotopy abelian .
we conclude with proposition [ prop2.12 ] which presents the conditions we will establish in the next 4 chapters .
finally , we recall some results from @xcite that will be used in the sequel . in @xcite ,
a clutching construction was described for hurewicz fibrations in case that the base is a mapping cone .
this construction is particularly simple in the case of a principal fibration .
suppose @xmath151 .
we describe a principal fibration @xmath152^{i}&e\ar@{->}[r]^{\pi}&b } \ ] ] where @xmath153 , where @xmath154 is the space of paths @xmath155 with @xmath156 . in case @xmath157
, we have a pair of principal fibrations : @xmath158\ar@{=}[r]&\omega x\ar@{->}[d]\\ & e_0\ar@{->}[r]\ar@{->}[d]_{\pi_0}&e\ar@{->}[d]_{\pi}\\ a\ar@{-->}[ur]\ar@{->}[r]^{\theta}&b_0\ar@{->}[r]&b } \ ] ] clearly @xmath159 lifts to a map @xmath160 .
we assert that there is a lifting @xmath161 such that the composition : @xmath162^{{\scalebox{.9}{${\overline{\theta}}$ } } } & e_0\ar@{->}[r]&e } \ ] ] is null homotopic . for if the composition @xmath162^{\theta'}&e_0\ar@{->}[r]&e } \ ] ] is essential , it factors through @xmath163 up to homotopy , and we can use the principal action @xmath164 ^ ->>>>{a}&(e , e_0 ) } \ ] ] to define a different lifting @xmath161 of @xmath159 in which the composition into @xmath63 is null homotopic . in particular , the composition @xmath165^{\overline{\theta}}&(e , e_0)\ar@{->}[r]^{\pi}&(b , b_0 ) } \ ] ] induces an isomorphism in homology . [ prop2.1 ] suppose @xmath166^{{\scalebox{.9}{${\pi}$}}}&(b , b_0)}\ ] ] is a hurewicz fibration with fiber @xmath167 where @xmath157 .
then there is a map @xmath168
^ -<<<{{\scalebox{.9}{${\varphi}$}}}&(e , e_0 ) , } $ ] and a pushout diagram @xmath169 ^ -<<<{\varphi}&e\\ f\times a\ar@{->}[r]^{\varphi}\ar@{->}[u]&e_0\ar@{->}[u ] } \ ] ] where @xmath170 is the projection onto @xmath171 .
in particular @xmath172 in case that @xmath173 is a principal fibration , we can take @xmath65 to be the composition : @xmath174^{1\times\overline{\theta } } & \omega x\times ( e , e_0 ) \ar@{->}[r]^-<<<<<{a } & ( e , e_0 ) } \ ] ] where @xmath175 is the principal action map .
the following result ( @xcite ) is a simple application of [ prop2.1 ] .
[ prop2.3 ] localized at a prime @xmath176 , there is a fibration sequence @xmath3 ^ ->>>{\partial}&bw_n\times s^{4n-1}\ar@{->}[r]&s^{2n}\ar@{->}[r]^->>>>>{e}&\omega s^{2n+1 } } \ ] ] where @xmath54 factors @xmath177^{\nu}&bw_n\ar@{->}[r]^->>>{i_1}&bw_n\times s^{4n-1}}$ ] and the homotopy fiber of @xmath53 is @xmath178 @xmath44^{e^2}&\omega^2s^{2n+1}\ar@{->}[r]^{\nu}&bw_n . }
\ ] ] furthermore if @xmath6 , @xmath43 has a homotopy - abelian @xmath0-space structure and @xmath53 is an @xmath0-map .
using the work of cohen , moore and neisendorfer , theriault has proved [ prop2.4 ] @xmath43 has @xmath0-space exponent - space has @xmath0-space exponent @xmath179 if the @xmath180 power map is null homotopic in some association . ]
@xmath38 .
an important application of [ prop2.1 ] is the extension theorem [ theor2.5 ] suppose @xmath99 is a co-@xmath0 space such that the map @xmath161 is divisible by @xmath179 in the co-@xmath0-space structure , and @xmath14 is a connected @xmath0-space with @xmath0-space exponent @xmath179 .
then the restriction @xmath181\to [ e_0,z]\ ] ] is onto .
this is a powerful tool and is the key to showing that each map @xmath182 extends to a map @xmath183 in @xcite . in section [ subsec6.1 ] we will enhance this to the controlled extension theorem , which will allow for a good choice of extension .
whitehead and samelson products in homotopy groups with coefficients were introduced by neisendorfer ( @xcite ) .
these included relative whitehead products and in @xcite , he introduced @xmath0-space based whitehead products
. these will be useful in chapter [ chap3 ] where we will generalize these constructions to classes defined on co-@xmath0 spaces . at this point
we will introduce a basic construction and show how it is related to the question of homotopy - abelian @xmath0-space structures .
we begin with the construction of a map @xmath184 we use the decomposition @xmath185 where @xmath100 is the cone functor with vertex at @xmath186 and define @xmath187 by @xmath188 ^ -<<<{\pi_2}&c(\omega v)\ar@{->}[r]^-<<<{\epsilon}&v \ar@{->}[r]&u\vee v } \\ & \xymatrix { c(\omega u)\times \omega v\ar@{->}[r]^-<<<{\pi_1}&c(\omega u)\ar@{->}[r]^-<<<{\epsilon}&u \ar@{->}[r]&u\vee v}\end{aligned}\ ] ] where @xmath189 is an evaluation .
[ prop2.6 ] @xmath187 lifts to a homotopy equivalence with the homotopy fiber of the inclusion @xmath190 .
the homotopy fiber is the union of the parts over @xmath191 and @xmath192 ; i.e. , @xmath193 for any space @xmath60 , we define a homotopy equivalence of pairs @xmath194 by @xmath195 . according to @xcite ,
the pair @xmath196 is an ndr pair , so the induced map @xmath197 is a homotopy equivalence .
now suppose that @xmath152^{i}&e\ar@{->}[r]^{\pi}&b } \ ] ] is a principal fibration induced by a map @xmath151 where @xmath198 is an @xmath0-space .
we can and will assume that the multiplication on @xmath198 has a strict unit ( @xcite ) .
[ prop2.7 ] there is a strictly commutative diagram , @xmath199_{\omega}\ar@{->}[r]^-<<<{\gamma}&e\ar@{->}[d]_{\pi}\\ b\vee b\ar@{->}[r]^{\nabla}&b } \ ] ] where @xmath200 is the folding map .
furthermore , @xmath150 is natural with respect to the data .
the square @xmath201^{\nabla}\ar@{->}[d]&&b\ar@{->}[d]_{\varphi}\\ b\times b\ar@{->}[r]^{\varphi\times \varphi}&x\times x\ar@{->}[r]^{\mu}&x } \ ] ] is strictly commutative and induces a map of the induced principal fibrations .
@xmath150 is the composition of @xmath202 with the map @xmath203 given by the formula @xmath204 where @xmath205 .
recall ( for example @xcite ) that there is a natural homotopy equivalence @xmath206 such that the diagram @xmath207_{-\tau}&{}\simeq{}&x*y\ar@{->}[d]_{\tau}\\
\sigma(y\wedge x)&{}\simeq{}&y * x } \ ] ] commutes up to homotopy , where the maps labeled @xmath208 are the transposition maps .
[ prop2.8 ] suppose that @xmath198 is homotopy commutative .
then there is a homotopy commutative diagram : @xmath209_{-\tau}&{}\simeq{}&\omega b*\omega b\ar@{->}[dr]^{\gamma}\ar@{->}[dd]_{\tau}&\\ & & & e\\ \sigma(\omega b\wedge \omega
b)&{}\simeq{}&\omega b*\omega b\ar@{->}[ur]_{\gamma } } \ ] ] by @xcite , we can assume that the homotopy of commutation @xmath210 is stationary on the axes , where @xmath211 and @xmath212 .
we then define @xmath213 and use this to define @xmath214 , a homotopy between @xmath150 and @xmath215 . the map @xmath216 plays a role in a useful condition for an @xmath0-space to have a homotopy - abelian structure .
[ prop2.9 ] suppose that @xmath217^{h}&f\ar@{->}[r]^{i}&e\ar@{->}[r]^{\pi}&b } \ ] ] is a fibration sequence in which @xmath218 is null homotopic .
suppose that there is a lifting @xmath219 of @xmath220 in the diagram : @xmath221 ^ -<<<{\overline{\omega}}\ar@{->}[d]_{\omega}&e\ar@{->}[d]_{\pi}\\ b\vee b\ar@{->}[r]^{\nabla}&b } \ ] ] then the @xmath0-space structure defined on @xmath167 by any right inverse @xmath222 of @xmath68 defines a homotopy - abelian @xmath0-space structure
. for any pointed space @xmath14 let @xmath223 $ ] and @xmath224 $ ] .
then @xmath59 is a group which acts on @xmath198 via the action map @xmath225 ^ -<<<{a}&f . } \ ] ] since @xmath68 has a right homotopy inverse , the orbit of @xmath226 is all of @xmath198 .
the adjoint of the composition @xmath227 ^ -<<<{\omega}&b\vee b\ar@{->}[r]^-<<<{\nabla}&b } \ ] ] is well known to be homotopic to the commuter map @xmath228 ^ -<<<{c}&\omega b. } \ ] ] ( see , for example , @xcite ) .
consequently , the existence of @xmath229 implies that every commutator in @xmath59 acts trivially on @xmath226 ; i.e. , @xmath230 . let @xmath231 be the stabilizer of @xmath232 .
then @xmath233 is a normal subgroup since if @xmath234 , @xmath235 .
consequently @xmath236 is a quotient group of @xmath59 .
it is abelian since @xmath230 .
this group structure on @xmath224 $ ] is natural for maps in @xmath14 .
apply this in case @xmath237 and @xmath238 to construct a homotopy - abelian @xmath0-space structure on @xmath167 .
now recall that we have fixed @xmath72 and @xmath239 and we set @xmath240 .
let @xmath241 .
consider the commutative diagram : @xmath242\\ \omega g * \omega g\ar@{->}[d]_{\omega}\ar@{->}[r]^{\gamma}&e\ar@{->}[d]\\ g\vee g\ar@{->}[r]^{\nabla}&g } \end{split}\ ] ] [ prop2.11 ] if the composition : @xmath243 ^ -<<<{{\scalebox{.9}{${\gamma}$}}}&e\ar@{->}[r]^->>>>{{\scalebox{.9}{${\nu_{\infty}}$}}}&bw_n } \ ] ] is null homotopic , the induced @xmath0-space structure on @xmath50 is homotopy abelian .
compare ( [ eq2.10 ] ) with ( [ eq1.6new ] ) and apply ( [ prop2.9 ] ) . in the sequel
we will construct maps @xmath183 such that the composition : @xmath244 ^ -<<<{{\scalebox{.9}{${\gamma_k}$ } } } & e_k\ar@{->}[r]^->>>>{{\scalebox{.9}{${\nu_k}$}}}&bw_n } \ ] ] is null homotopic .
we begin with an arbitrary choice of @xmath83 as in @xcite and modify it using the controlled extension theorem ( [ theor6.1 ] ) . having
done this , we have no reason to assume that the composition : @xmath245^{{\scalebox{.9}{${e_k}$ } } } & e_k\ar@{->}[r]^->>>>{{\scalebox{.9}{${\nu_k}$}}}&bw_n } \ ] ] is homotopic to @xmath110 .
however , all modifications occur in dimensions @xmath246 and larger , so we can assume these maps agree up to dimension @xmath247 .
the proof of theorem a will then follow from [ prop2.12 ] suppose we can construct maps @xmath248 for each @xmath249 such that : @xmath177 ^ ->>>>{{\scalebox{.9}{${\omega\partial}$}}}&\omega s^{2n+1}\{p^r\}\ar@{->}[r]^-<<<<{{\scalebox{.9}{${\partial'}$}}}&e_k\ar@{->}[r]^->>>{{\scalebox{.9}{${\nu_k}$}}}&bw_n}$]induces an isomorphism in @xmath250 the composition : @xmath251 ^ -<<<{{\scalebox{.9}{${\gamma_k}$}}}&e_k\ar@{->}[r]^->>>>{{\scalebox{.9}{${\nu_k}$}}}&bw_n } \ ] ] is null homotopic for each @xmath249
. the restrictions of @xmath252 and @xmath110 to the @xmath247 skeleton of @xmath253 are homotopic . then there is a map @xmath254 such that @xmath255 is null homotopic and the induced @xmath0-space structure on @xmath50 is homotopy abelian .
since the inclusion : @xmath256 is a homotopy equivalence , we can define @xmath257 which restricts to @xmath83 on @xmath258 . since @xmath259 factors through @xmath79 for each @xmath35
, we have a homotopy commutative diagram : @xmath260\ar@{->}[dd]&\omega s^{2n+1}\{p^r\}\ar@{->}[dr]^{{\scalebox{.9}{${h}$}}}\ar@{->}[dd]&\\ & & bw_n\\ e_k\ar@{->}[r]&e\ar@{->}[ru]_{{\scalebox{.9}{${\nu_k}$ } } } & } \ ] ] from the case @xmath107 , we see that the composition : @xmath261 ^ ->>>>{{\scalebox{.9}{${\omega\partial}$}}}&\omega s^{2n+1}\{p^r\}\ar@{->}[r]^-<<<<{{\scalebox{.9}{${h}$}}}&bw_n } \ ] ] induces an isomorphism in @xmath250 . using cup product and the bockstein ,
we can conclude that this composition is an epimorphism in all dimensions and consequently the fiber of this composition is @xmath178 .
now compare with ( [ eq1.3new ] ) to see that the fiber of @xmath0 is an anick space .
in the diagram @xmath262\ar@{->}[d ] & e_k^{2np^{k+1}-2 } \ar@{->}[r]^{{\scalebox{.9}{${\nu_k}$ } } } \ar@{->}[d ] & bw_n \ar@{=}[d]\\ \omega g\ast \omega g\ar@{->}[r]^{{\scalebox{.9}{${\gamma}$}}}&e\ar@{->}[r]^{{\scalebox{.9}{${\nu_{\infty}}$}}}&bw_n , } \ ] ] the upper composition is null homotopic for each @xmath35 , so the lower composition is null homotopic as well , thus the result follows from [ prop2.11 ] .
we now recall , for future use , the properties of @xmath59 and @xmath50 that we will be using in the sequel .
[ theor2.14 ] for @xmath33 , @xmath8 and @xmath7 , there is an anick space @xmath50 ; i.e. , there is a fibration sequence @xmath3^{\pi_n}&s^{2n-1}\ar@{->}[r]&t\ar@{->}[r]&\omega s^{2n+1 } } \ ] ] such that the composition @xmath44^{e^2}&\omega^2s^{2n+1}\ar@{->}[r]^{\pi_n}&s^{2n+1 } } \ ] ] has degree @xmath263 .
furthermore we have the following properties : there exists a space @xmath59 and maps @xmath264 such that the compositions @xmath265^{f}&\sigma t\ar@{->}[r]^{\widetilde{g}}&g } \\ & \xymatrix { t\ar@{->}[r]^{g}&\omega g\ar@{->}[r]^{h}&t } \end{aligned}\ ] ] are homotopic to the identity .
both @xmath50 and @xmath59 are atomic and @xmath38-complete .
the homotopy type of @xmath50 and @xmath59 are unique satisfying these conditions .
@xmath266 is a wedge of moore spaces . for any @xmath0-space structure defined by the maps @xmath267 in
, there is an @xmath0-map @xmath268 such that @xmath269 .
@xmath270 is generated by classes @xmath271 for each @xmath272 subject to the relations @xmath273 and @xmath274 .
@xmath275\otimes \bigwedge(u)$ ] where @xmath276 and @xmath277 .
furthermore @xmath278 .
@xmath279 @xmath280 and @xmath281 are each homotopy equivalent to a wedge of moore spaces .
any choice of @xmath183 in ( [ eq1.6new ] ) has a right homotopy inverse .
@xmath115 is homotopy equivalent to a wedge of @xmath282 moore spaces for @xmath127 .
@xmath283 is homotopy equivalent to a wedge of moore spaces most of these are restatements of results in @xcite .
properties ( a ) , ( b ) , ( c ) , ( d ) , ( e ) , and ( f ) are respectively 4.4 , 4.7 , 4.9 , 4.3(m ) , 4.6 , and 4.1 of @xcite .
property ( g ) follows immediately by applying the serre spectral sequence to 1.1 .
properties ( h ) and ( i ) are 4.3(c ) and 4.5 respectively . for ( j ) , a right homotopy inverse is given by the composition @xmath284 constructed from the proof of [ prop3.5 ] .
property ( k ) is 4.8 . for ( l )
note that @xmath285 since @xmath286 .
thus @xmath287 where @xmath288 is a wedge of moore spaces by ( i ) and ( k ) .
thus @xmath289 and @xmath290 which is in a wedge of moore space by property ( d ) .
in this chapter we will review and extend some results in @xcite which generalize the notion of whitehead products .
in particular , @xmath59 and @xmath0 will designate an arbitrary co-@xmath0 space .
then we will return @xmath59 as to designating the co-@xmath0 space corresponding to @xmath50 in future chapters . ]
given two simply connected co-@xmath0 spaces @xmath59 and @xmath0 , we will construct a new co-@xmath0 space @xmath86 and a cofibration sequence @xmath291^{w } & g\vee h\ar@{->}[r ] & g\times h. } \ ] ] composition with @xmath60 defines a more general notion of whitehead products . in section [ subsec3.1 ]
we will review the material in @xcite . in section [ subsec3.2 ]
we will discuss relative whitehead products and @xmath0-space based whitehead products which have been developed for homotopy groups with coefficients in @xmath292 by neisendorfer @xcite .
we will define these in the total space of a principal fibration using co-@xmath0 spaces in place of moore spaces . in section [ subsec3.3 ]
we will use these products to decompose @xmath293 and @xmath294 when @xmath59 and @xmath0 are co-@xmath0 spaces and decompose the map @xmath150 from section [ subsec2.2 ] as a wedge of iterated whitehead products when the base is a co-@xmath0 space . in section [ subsec3.4 ]
we recall and generalize slightly the results of neisendorfer in the case that @xmath59 and @xmath0 are moore spaces . given two simply connected co-@xmath0 spaces @xmath295 , we introduce a new co-@xmath0 space @xmath86 together with a cofibration sequence : @xmath291 ^ -{w}&g\vee h\ar@{->}[r]&g\times h. } \ ] ] to do this , suppose that @xmath59 and @xmath0 are given co-@xmath0 space structures by constructing right inverses to the respective evaluation maps : @xmath296 ^ -{\nu_1}&\sigma\omega g\ar@{->}[r]^-{\epsilon_1}&g}\\ \xymatrix{h\ar@{->}[r]^-{\nu_2}&\sigma\omega h\ar@{->}[r]^-{\epsilon_2}&h.}\end{gathered}\ ] ] we define a self map @xmath297 as the composition @xmath298
^ ->>>>{\epsilon_1\wedge 1 } & g\wedge \omega h \ar@{->}[r]^->>>>{\nu_1\wedge 1 } & \sigma(\omega g\wedge\omega h ) \ar@{->}[r]^->>>>{1\wedge \epsilon_2 } & \omega g\wedge h \ar@{->}[r]^->>>>{1\wedge \nu_2 } & \sigma(\omega g\wedge\omega h ) ; } \ ] ] @xmath86 is then defined as the telescopic direct limit of @xmath299 .
we then have : [ prop3.1 ] the identity map of @xmath86 factors : @xmath291 ^ -{\psi } & \sigma(\omega g\wedge\omega h)\ar@{->}[r]^-{\theta } & g\circ h .}\ ] ] furthermore , if @xmath300 and @xmath301 are co-@xmath0 maps , there are induced co-@xmath0 maps so that the diagram @xmath302_{f\circ g}\ar@{->}[r]^-{\psi}&\sigma(\omega g\wedge \omega h)\ar@{->}[r]^-{\theta}\ar@{->}[d]_{\sigma(\omega f\wedge \omega g)}&g\circ h\ar@{->}[d]_{f\circ g}\\ g'\circ h'\ar@{->}[r]^-{\psi'}&\sigma(\omega g'\wedge \omega h')\ar@{->}[r]^-{\theta'}&g'\circ h ' } \ ] ] commutes up to homotopy . since @xmath86 is the limit of the telescope defined by @xmath299 , @xmath303 , so the composition
^ -{\psi}&\sigma(\omega g\wedge \omega h)\ar@{->}[r]^-e & \sigma(\omega g\wedge \omega h)\ar@{->}[r]^-{\theta}&g\circ h } \ ] ] is homotopic to the identity .
the map @xmath299 , however , is a composition of 4 maps between co-@xmath0 spaces , and thus @xmath86 is a retract of 3 different co-@xmath0 spaces and one of them , @xmath304 , in two potentially distinct ways .
this provides 4 potentially distinct co-@xmath0 space structures on @xmath86 .
we choose the structure defined by @xmath305 and @xmath159 ; viz .
, @xmath291 ^ -{\psi}&\sigma(\omega g\wedge\omega h)\ar@{->}[r]^-{\sigma\widetilde{\theta}}&\sigma\omega(g\circ h ) } \ ] ] or equivalently @xmath306 ^ -{\psi}&\sigma(\omega g\!{}\wedge{}\!\omega h ) \ar@{->}[r]&\sigma(\omega g\!{}\wedge\omega h)\vee\sigma(\omega g\!{}\wedge{}\ ! \omega h)}\hspace*{-6pt } \xymatrix@c=26pt { \ar@{->}[r]^->>>>>{\theta\!{}\wedge{}\!\theta}&g\!{}\circ{}\ ! h\!{}\vee{}\ !
! h } \end{gathered}\ ] ] where @xmath307 is the adjoint of @xmath159 .
[ prop3.2 ] there are co-@xmath0 equivalences , @xmath308 which are natural for co-@xmath0 maps in @xmath59 and @xmath0 and continuous maps in @xmath198 .
[ prop3.3 ] there is a natural cofibration sequence @xmath291 ^ -w&g\vee h\ar@{->}[r]&g\times h } \ ] ] where @xmath60 is the composition : @xmath291 ^ -<<<<{\psi}&\sigma(\omega g\wedge\omega h)\ar@{->}[r]^->>>>{\omega}&g\vee h .}\ ] ] [ def3.4 ] let @xmath309 and @xmath310 .
we define the whitehead product rather than the usual @xmath311 $ ] since in an important application we need to make a distinction . that is the case when @xmath59 and @xmath0 are both moore spaces . in this case
@xmath86 is a wedge of two moore spaces . by choosing the higher dimensional one ,
neisendorfer @xcite defines internal whitehead products in homotopy with coefficients in @xmath292 .
this is denoted @xmath311 $ ] , while @xmath312 is the `` external '' whitehead product . ] @xmath313 as the composition @xmath291 ^
-w&g\vee h\ar@{->}[r]^-{\alpha\vee\beta}&x .}\ ] ] [ prop3.5 ] each whitehead product @xmath314 factors through the `` universal whitehead product '' @xmath315 the right hand square in the following diagram determines a map from the upper row which is a cofibration sequence to the lower row which is a fibration sequence @xmath291 ^ -w\ar@{->}[d]_{\xi}&g\vee h\ar@{->}[r]\ar@{->}[d]_{\alpha\vee\beta}&g\times h\ar@{->}[d]\\ \sigma(\omega x\wedge\omega x)\ar@{->}[r]^-{\omega}&x\vee x\ar@{->}[r ] & x\times x } \ ] ] and @xmath316 is unique up to homotopy since @xmath86 is a co-@xmath0 space and @xmath317 .
it follows that @xmath318{}\{\alpha,\beta\}=\nabla(\alpha\vee\beta)w\sim \nabla\omega\xi = w\xi.\makebox[78pt]{}\rlap{\qed}\ ] ] in this section we will discuss @xmath0-space based whitehead products and relative whitehead products . in the case that @xmath59 and
@xmath0 are moore spaces , this material is covered in @xcite , and what we present is a mild generalization .
we need to consider whitehead products instead of their adjoints the samelson products ( which neisendorfer considered ) since the domains are not necessarily suspensions .
we also consider principal fibrations , so these products occur in the total space rather than the fiber of a fibration as in neisendorfer s version .
we wish to thank joe neisendorfer for several interesting conversations during the development of this material .
we begin with a principal fibration @xmath152 ^ -{i}&e\ar@{->}[r]^-{\pi}&b } \ ] ] induced by a map @xmath151 .
the ( external ) relative whitehead product then is a pairing @xmath319\times[h , e]\to [ g\circ h , e].\ ] ] in the case that @xmath198 is a homotopy commutative @xmath0-space with strict unit , we also define the @xmath0-space based whitehead product . it is a pairing @xmath319\times[h , b]\to [ g\circ h , e].\ ] ] suppose we are given maps : @xmath64^{{\scalebox{.9}{${\alpha}$}}}&b,\ h\ar@{->}[r]^{{\scalebox{.9}{${\beta}$}}}&b,\ g\ar@{->}[r]^{{\scalebox{.9}{${\gamma}$ } } } & e,\ h\ar@{->}[r]^{{\scalebox{.9}{${\delta}$ } } } & e. } \ ] ] we will use the notation @xmath320\ ] ] for the relative whitehead product and @xmath321 \index{\{\ , \
\}$_{\times}$|lb}\ ] ] for the @xmath0-space based whitehead product .
these products and the absolute whitehead product are related by the following formulas to be proved : @xmath322 we begin with the @xmath0-space based whitehead product .
these are defined using the map @xmath150 from ( [ prop2.7 ] ) .
the product @xmath323 is defined as the homotopy class of the upper composition in the diagram : @xmath291 ^ -<<<{\psi}\ar@{->}[dr]_w & \sigma(\omega g\wedge\omega h)\simeq\omega g * \omega h\ar@{->}[d]_{\omega}\ar@{->}[r]&\omega b*\omega b\ar@{->}[r]^->>>>>{\gamma}\ar@{->}[d]_{\omega } & e\ar@{->}[d]_{\pi}\\ & g\vee h\ar@{->}[r]^-{\alpha\vee\beta}&b\vee b\ar@{->}[r]^{\nabla}&b.}\ ] ] [ prop3.6 ] given @xmath324 and @xmath325 , the homotopy class of the @xmath0-space based whitehead product @xmath326 depends only on the homotopy classes of @xmath327 and @xmath133 .
furthermore if @xmath328 and @xmath329 are co-@xmath0 maps , @xmath330 given an induced fibration @xmath331 ^ -{\widetilde{\xi}}\ar@{->}[d ] & e\ar@{->}[d]\\ b'\ar@{->}[r]^-{\xi } & b\ar@{->}[r]^-{\varphi } & x } \ ] ] and @xmath332 , @xmath333 , we have @xmath334 @xmath335 .
suppose @xmath336 is a strict @xmath0-map and we have a pointwise commutative diagram @xmath337 ^ -{\xi}\ar@{->}[d]_{\varphi } & b'\ar@{->}[d]_{\varphi'}\\ x\ar@{->}[r]^-{\eta } & x ' } \ ] ] which defines a map of principal fibrations : @xmath338 ^ -{\widetilde{\xi}}\ar@{->}[d ] & e'\ar@{->}[d]\\ b\ar@{->}[r]^-{\xi } & b ' .}\ ] ] then @xmath339 @xmath340 .
these all follow directly from the definition except for . to prove this we apply to the diagram @xmath338 ^ -{\pi}\ar@{->}[d]_k & b\ar@{->}[d]_{\varphi}\\ px\ar@{->}[r]^-{\epsilon } & x } \ ] ] where @xmath341 .
give @xmath154 the @xmath0-space structure of pointwise multiplication of paths in @xmath198 .
then @xmath189 is a strict @xmath0-map .
this gives a map of principal fibrations : @xmath342{}\xymatrix { \widetilde{e}\ar@{->}[r]^-{\widetilde{\pi}}\ar@{->}[d]_e & e\ar@{->}[d]_{\pi}\\ e\ar@{->}[r]^-{\pi } & b. } \\[-15pt ] & \makebox[341pt]{}\qed\end{aligned}\ ] ] by we have @xmath343 it suffices to show that @xmath344 by applying part to the left hand fibration .
the space @xmath345 can be described as follows @xmath346 with @xmath347 where @xmath348 and @xmath349 where @xmath350 .
define @xmath351 by @xmath134 .
the result then follows from : [ hrl3.7 ] suppose @xmath352 and @xmath353 . then there is a homotopy @xmath354 such that @xmath355 the left side and the bottom of the square are mapped to the basepoint . by rotating from the top to the right hand side pivoting at the point @xmath356
, we obtain the required homotopy .
we now describe the relative whitehead product .
we assume a principal fibration @xmath152^{i}&e\ar@{->}[r]^{\pi}&b } \ ] ] induced by a map @xmath151 ( as in section [ subsec2.1 ] ) , but we wo nt assume an @xmath0-space structure on @xmath198 .
define @xmath357 by the second component , so @xmath358 . the principal action @xmath359&e}\ ] ] is defined by the formula @xmath360 where @xmath361 is given by@xmath362 we then describe a strictly commutative diagram of vertical fibration sequences @xmath217^{d}\ar@{->}[d]&\omega x\ar@{->}[d]_{i}\\ \omega b\times e\cup pb\ar@{->}[r]^-<<<<{\gamma ' } \ar@{->}[d]_{\epsilon\vee\pi_2 } & e\ar@{->}[d]_{\pi}\\ b\vee e\ar@{->}[r]^{1\vee \pi}&b } \ ] ] where @xmath363 is to be considered as a subspace of @xmath364 .
the map @xmath365 is endpoint evaluation and @xmath366 is defind by the formula @xmath367 where\ar@{->}[r]^{f } & x}$ ] to the real line by @xmath368 for @xmath369 and @xmath370 for @xmath371 . ]
@xmath372 , and the map @xmath373 is given by @xmath374 .
the left hand fibration is the principal fibration induced by the projection @xmath375 .
observe that @xmath376 we record an important commutative diagram @xmath377\ar@{->}[r]^{\omega \varphi\times 1}&\omega x\times e\ar@{->}[d]_{a}\\ \omega b\ltimes e\simeq \omega b\times e\cup pb\ar@{->}[r]^-<<<<{\gamma'}&e } \end{split}\ ] ] which will be useful in evaluating the relative whitehead products in homology .
consider the strictly commutative square @xmath378\ar@{=}[r]&b\vee e\ar@{->}[d]_{\pi_1}\\ b\times e\ar@{->}[r]^{\pi_1}&b . } \ ] ] taking homotopy fibers vertically , we obtain a diagram of principal fibrations @xmath379\ar@{->}[r]^{\omega \pi_1}&\omega b\ar@{->}[d]\\ \omega b\times pe\cup pb\times \omega e\ar@{->}[r]^-<<<{\zeta}\ar@{->}[d]&\omega b\times e\cup pb\ar@{->}[d]\\
b\vee e\ar@{=}[r]&b\vee e } \ ] ] the map @xmath380 is defined by the formula @xmath381^{1\times \epsilon}&\omega b\times e\\ pb\times \omega e\ar@{->}[r]^-<<<<<{\pi_1}&pb } \ ] ] combining these diagrams , we obtain a strictly commutative diagram @xmath382_{w } & \hspace*{-75pt}\llap{\makebox[0pt]{$\simeq$}}\hspace*{-75pt } & \omega b\times pe\cup pb\times \omega e\ar@{->}[r]^-<<{\zeta}\ar@{->}[d ] & \omega b\times e\cup pb\ar@{->}[r]^-<<<{\gamma'}\ar@{->}[d ] & e\ar@{->}[d]_{\pi}\\ & & b\vee e\ar@{=}[r ] & b\vee e\ar@{->}[r]^{1\vee\pi}&b . } \end{split}\ ] ] for @xmath324 and @xmath383 , we define the relative whitehead product @xmath384 as the composition @xmath385 ^ ->>{\psi}&\omega g*\omega h\ar@{->}[r]&\omega b * \omega e\ar@{->}[r]^->>>{\zeta}&\omega b\times e\cup pb\ar@{->}[r]^-<<{\gamma'}&e } \ ] ] and , analogous to [ prop3.6 ] , we have [ prop3.11 ] the homotopy class of the relative whitehead product @xmath386 depends only on the homotopy classes of @xmath327 and @xmath387 . furthermore if @xmath388 and @xmath329 are co-@xmath0 maps , then @xmath389 given an induced fibration @xmath390\ar@{->}[r]^-{\widetilde{\xi}}&e\ar@{->}[d]\\ b'\ar@{->}[r]^-{\xi}&b\ar@{->}[r]^-{\varphi}&x } \ ] ] and classes @xmath332 , @xmath391 , we have @xmath392 @xmath393 .
suppose we have a strictly commutative diagram @xmath337^{\xi'}\ar@{->}[d]_{\varphi}&b'\ar@{->}[d]_{\varphi'}\\ x\ar@{->}[r]^{\eta}&x ' } \ ] ] inducing a map between principal fibrations : @xmath338^{\xi'}\ar@{->}[d]_{\pi}&e'\ar@{->}[d]_{\pi'}\\ b\ar@{->}[r]^{\xi}&b ' } \ ] ] then @xmath394 .
@xmath395 .
all parts except ( e ) follow directly from the definitions . for part ( e )
we construct a map of principal fibrations exactly as in [ prop3.6](e ) : @xmath396^{\widetilde{\pi}}\ar@{->}[d]_e&e\ar@{->}[d]_{\pi}\\ e\ar@{->}[r]^{\pi}&b } \ ] ] recall that @xmath397 , and since @xmath154 is contractible , both @xmath299 and @xmath398 are homotopy equivalences .
choose @xmath399 such that @xmath400 .
then by part ( d ) we have @xmath401 by part ( c ) .
however this is homotopic to @xmath402 . at this point
we will discuss the compatibility of the @xmath0-space based whitehead product and the relative whitehead product .
[ prop3.12 ] suppose @xmath198 is an @xmath0-space with strict unit and we are given @xmath324 , @xmath383
. then @xmath403 to prove this we will combine two homotopies .
the first homotopy will replace the sequential composition of paths in the definition of the action map @xmath175 and @xmath366 with a blending of the homotopies using the @xmath0-space structure in @xmath198 .
the second homotopy will apply the homotopy rotation lemma ( [ hrl3.7 ] ) .
recall the map @xmath404 with the property that @xmath405 @xmath406{}\xymatrix { e\ar@{->}[r]^{\pi}\ar@{->}[d]_{k}&b\ar@{->}[d]\\ px\ar@{->}[r]^{\epsilon}&x } \\*[-16pt ] & \makebox[341pt]{}\qed\end{aligned}\ ] ] [ lem3.13 ] there is a homotopy @xmath407 with @xmath408 and @xmath409 given by the formula @xmath410 and a compatible homotopy @xmath411 with @xmath412 and @xmath413 given by the formula @xmath414 recall that any map @xmath415\to x$ ] is to be extended to a map @xmath416 by defining @xmath417 if @xmath369 and @xmath418 if @xmath371 . then , for example , @xmath419 since @xmath420 , the unit for @xmath421 .
we define @xmath422 where @xmath423 we define @xmath424 and @xmath425 in case @xmath426 . using @xmath413 we consider the composition @xmath427 ^ ->>>>>{\zeta}&\omega b\times e\cup pb\ar@{->}[r]^-<<<<{\gamma'_0}&e . } \ ] ] using the identification , we have the following formula for this composition @xmath428 & ( \omega_1,\omega_2(1))\ar@{->}[r ] & ( \pi\omega_2(1),\mu(\varphi\omega_1(t),k(\omega_2(1))(t ) ) ) . } \ ] ] we now apply [ hrl3.7 ] to the homotopy @xmath429 given by @xmath430 to obtain a homotopy @xmath431 with @xmath432 from this we construct a homotopy @xmath433 given by @xmath434 then @xmath435 and @xmath436 .
we have proved [ prop3.15 ] if @xmath198 is an @xmath0 space with a strict unit , there is a homotopy commutative diagram : @xmath437{}\xymatrix { \omega b\times e\cup pb\ar@{->}[r]^-<<<<<{\gamma'}&e\\ \omega b * \omega e\ar@{->}[r]^{1*\omega\pi}\ar@{->}[u]^{\zeta}&\omega b*\omega b\ar@{->}[u]^{\gamma } } \\*[-14pt ] & \makebox[341pt]{}\qed\end{aligned}\ ] ] clearly [ prop3.12 ] follows from [ prop3.15 ] .
we next describe a simplification of the relative whitehead product in case @xmath438 .
[ prop3.16 ] suppose @xmath439 and @xmath383 .
then the relative whitehead product @xmath440 ^ -<<<<<<{\{\alpha,\delta\}_r}&e } \ ] ] is represented by the composition @xmath441^{\theta}&a\ltimes h\ar@{->}[r]^->>>>>{\widetilde{\alpha}\ltimes\delta}&\omega b\ltimes e\ar@{->}[r]^-<<<<{\gamma'}&e } \ ] ] where @xmath159 is a right homotopy inverse to the projection will depend on the co-@xmath0 structure of @xmath0 . ] which pinches @xmath0 to a point . to construct the map
@xmath159 we need to generalize the context in which the map @xmath380 was defined in [ eq3.9 ] .
the homotopy fiber of the map @xmath442 ^ -<<<<{\pi_2}&y } \ ] ] which pinches @xmath443 to the basepoint is of the form @xmath444 using the homotopy equivalence @xmath445 ( see [ prop2.6 ] ) , we get a homotopy equivalent fibration sequence @xmath446 ^ ->>>>{\zeta}&\parbox[b]{63pt}{\mbox{}\newline \vspace*{4pt } $ \underset{\rotatebox{90}{\makebox[21pt]{$\simeq$}}}{cx\cup x\times y}$}\ar@{->}[r]^-<<<{\pi_2}&y\\ & x\ltimes y & } \ ] ] furthermore , the composition @xmath447 ^ -<<<<{\zeta}&x\ltimes y\ar@{->}[r]&x\wedge y } \ ] ] collapses @xmath448 to a point , so there is a commutative square @xmath447^{\zeta}\ar@{->}[d]_{\simeq}&x\ltimes y\ar@{->}[d]\\ x\wedge \sigma\omega y\ar@{->}[r]^{1\wedge \epsilon}&x\wedge y. } \ ] ] the relative whitehead product in [ prop3.16 ] is given by the upper composition in the homotopy commutative diagram @xmath449 ^ ->>>>{\zeta}&\omega b\!{}\times{}\ ! e\!{}\cup{}\ ! pb\ar@{->}[r]^-<<{\gamma'}&e\\ ( \sigma a)\!{}\circ{}\!h\ar@{->}[r]&\sigma(\omega\sigma a\!{}\wedge{}\!\omega h ) & \kern-201pt\llap{\makebox[0pt]{$\!{}\simeq{}\!$}}\kern-201pt & \omega\sigma a*\omega h\ar@{->}[r]^->>{\zeta}\ar@{->}[u]^{\omega \alpha*\omega \delta}&\omega\sigma a\!{}\times{}\ !
! p\sigma a\ar@{->}[u]\\ a\!{}\wedge{}\ !
h)\ar@{->}[u ] & \kern-201pt\llap{\makebox[0pt]{$\!{}\simeq{}\!$}}\kern-201pt & a*\omega h\ar@{->}[u]\ar@{->}[r]^->>>>{\zeta}&a\!{}\times{}\ !
! ca\ar@{->}[u ] & \kern-201pt\llap{\makebox[0pt]{$\!{}\simeq{}\!$}}\kern-201pt & a\!{}\ltimes{}\!h } \ ] ] where the lower composition is the map @xmath159 .
the right hand vertical map is the composition @xmath450
^ ->>>>>>{\widetilde{\alpha}\ltimes \delta}&\omega b\ltimes e\simeq \omega b\times e\cup pb . } \ ] ] by the homotopy commutative square above , @xmath159 has a right homotopy inverse since @xmath451 , and @xmath159 projects trivially to @xmath0 since @xmath452 is null homotopic .
[ cor3.18 ] suppose @xmath439 and @xmath383 .
then for any ring @xmath115 , the homomorphism @xmath453 is given by the composition @xmath454 ^ -<<<<<{(\widetilde{\varphi \alpha}\times \delta)_*}&h_*(\omega x\times e;r)\ar@{->}[r]^-<<<<<{a_*}&h_*(e;r ) } \ ] ] apply [ eq3.8 ] and [ prop3.16 ] .
we need , also , to discuss iterated whitehead products .
suppose @xmath455 for @xmath456 .
we define the iterated whitehead product @xmath457 as @xmath458 . in case @xmath459 for each @xmath218
we define @xmath460}=g\circ g^{[n-1]}.\ ] ] we also define @xmath461}h^{[j]}$ ] as @xmath462}h^{[j]}_*)$ ] when @xmath463 and as @xmath464}$ ] when @xmath465 .
suppose now that @xmath466 and @xmath467 are the inclusions .
we then consider @xmath468}h\to g\vee h.\ ] ] given a principal fibration @xmath152&e\ar@{->}[r]&b } \ ] ] and maps @xmath469 , @xmath470 , we define @xmath471 as @xmath472 by an iterated application of [ prop3.11](c ) , we have @xmath473 now consider the principal fibration @xmath474 ^ ->>>>{i}&\omega g\ltimes
h\ar@{->}[r]^{\pi}&g\vee h. } \ ] ] let @xmath475}h$ ] .
let @xmath476 be the inclusion of the second factor .
[ prop3.20 ] the maps @xmath477}h\to\omega g\ltimes h$ ] define a homotopy equivalence @xmath478}h\vee h\to \omega g\ltimes h.\ ] ] according to @xcite , such a homotopy equivalence exists where the maps @xmath479}h\to \omega g\ltimes h\ ] ] are chosen so that @xmath480 . however by ( [ eq3.19 ] ) , @xmath481 ; since the map @xmath482 is null homotopic , @xmath483 .
we now consider the principal fibration @xmath484&\omega g*\omega h\ar@{->}[r]&g\vee h. } \ ] ] using the map @xmath485 we define iterated relative whitehead products : @xmath486}g^{[j]}(g\circ h)\to \omega g*\omega h\ ] ] [ prop3.21 ] the maps @xmath487 for @xmath272 , @xmath488 define a homotopy equivalence @xmath489}g^{[j]}(g\circ h)\to \omega g*\omega h.\ ] ] consider the diagram of principal fibrations @xmath490
^ ->>>>>{\iota_2}\ar@{->}[d]&\omega g\times \omega h\ar@{->}[d]\\ \omega g*\omega h\ar@{=}[r]\ar@{->}[d]_{\zeta}&\omega g*\omega h\ar@{->}[d]_{\pi}\\ \omega g\ltimes h\ar@{->}[r]&g\vee h } \ ] ] the maps @xmath491}(g\circ h)\to \omega g\ltimes h$ ] defined by [ prop3.20 ] lift to @xmath293 since @xmath380 has a right homotopy inverse .
since the liftings project by @xmath173 onto the maps : @xmath492}(g\circ h)\to g\vee h,\ ] ] these liftings are homotopic to the relative whitehead product defined by @xmath173 .
however @xmath294 is homotopy equivalent to @xmath493 so the maps @xmath494}k\to \omega g*\omega h\ ] ] define a homotopy equivalence @xmath495}k\to \omega g * \omega h.\ ] ] furthermore @xmath496}(g\circ h)\ ] ] and the relative whitehead products defined by the left hand fibration are mapped to the corresponding relative whitehead products in the right hand fibration .
thus we have @xmath497{}\xymatrix{\raisebox{-8pt}{$\smash{\bigvee\limits_{\substack{i\geqslant 0\\ j\geqslant 0}}h^{[i]}g^{[j]}(g\circ h)}$}\ar@{->}[r]^-<<<<{\simeq}&\raisebox{-7pt}{$\smash{\bigvee\limits_{i\geqslant 0}h^{[i]}k}$}\ar@{->}[r]^->>>>{\simeq}&\omega g*\omega h } } \makebox[38.5pt]{}\rlap{\qed}\ ] ] [ theor3.22 ] suppose @xmath152^{i}&e\ar@{->}[r]^{\pi}&g } \ ] ] is a principal fibration induced by a map @xmath498 where @xmath198 is an @xmath0-space with strict unit .
. then the composition @xmath500 ^ -<<<{\gamma}&e\ar@{->}[r]^{\nu}&z } \ ] ] is null homotopic iff the compositions @xmath501}\to e\to z\ ] ] are null homotopic for each @xmath272 , where @xmath502 is the identity map . in this case @xmath503 and the map of principal fibrations @xmath504\ar@{->}[r]&\omega
x\ar@{->}[d]\\ \omega g*\omega g\ar@{->}[d]\ar@{->}[r]&e\ar@{->}[d]\\ g\vee g\ar@{->}[r]^{\nabla}&g } \ ] ] maps @xmath505}g^{[j]}(g\circ h)$ ] to @xmath506}(g\circ g)$ ] which only depends on @xmath507 . in the case that the co-@xmath0 spaces are moore spaces , the resulting whitehead products occur in the homotopy groups with coefficients .
the adjoint theory of samelson products is due to neisendorfer @xcite , and was crucial in the work of @xcite .
this theory has been further developed in @xcite where @xmath0-space based whitehead products were introduced .
we need to make a mild generalization of this in that we must consider the case where @xmath508 in this case @xmath509 this splitting is not unique and we must choose a splitting .
choose a map @xmath510 so that the diagram @xmath511 ^ -{\delta}\ar@{->}[d]_{\pi_{m+n}}&p^m(p^s)\wedge p^n(p^s)\ar@{->}[d]\\ s^{m+n}\ar@{->}[r]^-{\simeq}&s^m\wedge s^n } \end{split}\ ] ] commutes up to homotopy .
such a choice is possible when @xmath512 for @xmath38 odd and is unique up to homotopy .
neisendorfer @xcite has produced internal whitehead and samelson products for homotopy with @xmath97 coefficients .
the whitehead product of @xmath513 and @xmath514 is an element @xmath515\in\pi_{m+n+1}(x;z / p^s)\ ] ] defined as the homotopy class of the composition : @xmath516 ^ ->>>>{{\scalebox{.9}{${\sigma\delta}$}}}&\sigma p^m(p^s)\wedge p^n(p^s)}\\ & = \xymatrix{p^{m+1}(p^s)\circ p^{n+1}(p^s)\ar@{->}[r]^-<<<{{\scalebox{.9}{${\{x , y\}}$}}}&x}\notag\end{aligned}\ ] ] as we will need to consider such pairings with different coefficients , suppose @xmath517 and @xmath518 .
we can still form the external whitehead product : @xmath519 ^ -<<<{{\scalebox{.9}{${\{x , y\}}$}}}&x.}\ ] ] since the map of degree @xmath520 on @xmath128 is null homotopic , there is a splitting : @xmath521 we now choose an explicit splitting . recall ( [ eq1.5new ] ) @xmath522 .
[ prop3.25 ] there is a splitting of @xmath523 defined by the two compositions : @xmath524 ^ -{{\scalebox{.9}{${\delta}$}}}&p^m(p^r)\wedge p^n(p^r)\ar@{->}[r]^{{\scalebox{.9}{${1\wedge \rho^t}$ } } } & p^m(p^r)\wedge p^n(p^{r+t})}\\ \xymatrix{p^{m+n-1}(p^r)\ar@{->}[r]^-{{\scalebox{.9}{${\delta}$}}}&p^m(p^r)\wedge p^{n-1}(p^r)\ar@{->}[r]^-{{\scalebox{.9}{${1\wedge \delta_t}$ } } } & p^m(p^r)\wedge p^n(p^{r+t})}\end{gathered}\ ] ] @xmath525 induces a @xmath526 homology isomorphism , so @xmath527 induces a homology monomorphism .
the second composition factors @xmath528 ^ -{{\scalebox{.9}{${\delta}$}}}&p^m(p^r)\wedge p^{n-1}(p^r)\ar@{->}[r]^-{{\scalebox{.9}{${1\wedge \pi_{n-1}}$}}}&p^m(p^r)\wedge s^{n-1}}\\ \xymatrix@c=35pt{\ar@{->}[r]^-{{\scalebox{.9}{${1\wedge \iota_{n-1}}$}}}&p^m(p^r)\wedge p^{n}(p^{r+t})}\end{gathered}\ ] ] and the composition of the first two maps is a homotopy equivalence .
since the third map induces a @xmath526 homology monomorphism , this composition does as well .
counting ranks , we see that the two maps together define a homotopy equivalence : @xmath529 { } \xymatrix{e\colon p^{m+n}(p^r)\vee p^{m+n-1}(p^r)\ar@{->}[r]^-{{\scalebox{.9}{${\simeq}$}}}&p^m(p^r)\wedge p^n(p^{r+t } ) } \makebox[33pt]{}\rlap{\qed}\ ] ] we apply this to the internal whitehead product ( [ eq3.24 ] ) to get [ prop3.26 ] @xmath530 { } \{x , y\}e=[x , y \rho^t]\vee[x , y\delta_t]\colon p^{m+n}(p^r)\vee p^{m+n-1}(p^r)\to x. \makebox[21pt]{}\rlap{\qed}\ ] ] [ prop3.26 ] resolves the external whitehead product with different coefficients into internal whitehead products with coefficients in @xmath292 as considered by neisendorfer .
suppose now that we are given a principal fibration @xmath531 classified by a map @xmath151 where @xmath198 is a homotopy commutative @xmath0-space with strict unit and we are given classes @xmath532 and @xmath533 .
then we have [ prop3.27 ] @xmath534_{\times}\vee[u , v\delta_t]_{\times}\colon p^{m+n}(p^r)\vee p^{m+n-1}(p^{r+t})\to e.\ ] ] both @xmath535 and @xmath536_{\times}\vee[u , v\delta_t]_{\times}$ ] are the images under @xmath150 of maps : @xmath537 which are homotopic after projection @xmath538 ^ -{{\scalebox{.9}{${\omega}$}}}&b\vee b}\ ] ] by [ prop3.26 ] . since @xmath539 has a left homotopy universe , these maps are homotopic .
composing with @xmath540 finishes the proof .
similar to [ prop3.27 ] , we have [ prop3.28 ] @xmath541_r\vee[x , u\delta_t]_r$ ] where @xmath542 and @xmath543 .
there is one special case of this that we will need in section [ subsec6.3 ] .
this involves relative whitehead products @xmath544_r$ ] when @xmath545{}\!&e}$ ] and @xmath546{}\!&\nobreak b}\!\!$ ] . in this case
@xmath547_r=\{x , u\}_r\colon \xymatrix{p^m\circ s^n\ar@{->}[r]&e}.\ ] ] [ prop3.29 ] @xmath548_r=[x , u]_r\colon \xymatrix{p^{m+n-1}\ar@{->}[r]&e}$ ] .
@xmath549_r\vee 0 $ ] since @xmath550 .
consequently we have a homotopy commutative diagram @xmath551 ^ -e\ar@{->}[dr]_-{[x , u\pi_n]_r\vee 0\makebox[3pt ] { } } & p^m\circ p^n\ar@{->}[r]^-{1\circ \pi_n}\ar@{->}[d]^-<<<<{\{x , u\pi_n\}_r } & p^m\circ s^n\ar@{->}[dl]^-{[x , u]_r } & \hspace*{-28pt}\llap{{\scalebox{.9}{${\simeq{}}$ } } } p^{m+n-1}\\ & e & } \ ] ] where the upper composition is homotopic to projection onto the first factor .
suppose then we are given a principal fibration @xmath152 ^ -{{\scalebox{.9}{${i}$ } } } & e\ar@{->}[r]^-{{\scalebox{.9}{${\pi}$ } } } & b } \ ] ] induced by a map @xmath151 where @xmath198 is a homotopy commutative @xmath0-space with strict unit .
suppose we are given classes @xmath552 recall that by using the map @xmath553 we define the internal @xmath0-space based whitehead product @xmath554_{\times}=\{\alpha,\beta\}_{\times}\delta\in \pi_{m+n-1}(e;z / p^r)\ ] ] and internal relative whitehead product @xmath555_r=\{\alpha,\gamma\}_r\delta\in \pi_{m+k-1}(e;z / p^r).\ ] ] these are related as in [ prop3.6 ] , [ prop3.11 ] and [ prop3.12 ] .
[ prop3.30 ] 1 .
@xmath556_{\times}=[\alpha,\beta]\in\pi_{m+n-1}(b;z / p^r)$ ] 2 .
@xmath557_{\times}=[\gamma,\delta]\in\pi_{k+\ell-1}(e;z / p^r)$ ] 3 .
@xmath558_r=[\alpha,\pi_*\delta]_{\times}\in\pi_{m+\ell-1}(e;z / p^r)$ ] 4 .
@xmath559_r=[\alpha,\pi_*\delta]\in\pi_{m+\ell-1}(b;z / p^r)$ ] 5 .
@xmath560_r=[\gamma,\delta]\in\pi_{k+\ell-1}(e;z / p^r)$ ] according to neisendorfer @xcite , we also have standard whitehead product formulas : [ prop3.31 ] the following identities hold : @xmath311_{\times}=-(-1)^{(m+1)(n+1)}[\beta,\alpha]_{\times}$ ] @xmath561_{\times}=[\alpha_1,\beta]_{\times}+[\alpha_2,\beta]_{\times}$ ] @xmath562_{\times}=[[\alpha,\beta],\eta]_{\times}+(-1)^{(m+1)(n+1)}[\beta,[\alpha,\eta]]_{\times}$ ] + for @xmath563 @xmath564_r]=[[\alpha,\beta],\gamma]_r+(-1)^{(m+1)(n+1)}[\beta,[\alpha,\gamma]_r]_r$ ] @xmath565_{\times}=[\beta^{(r)}\alpha,\beta]_{\times}+(-1)^{m+1}[\alpha,\beta^{(r)}\beta]_{\times}$ ] where @xmath566 is the bockstein associated with the composition @xmath567 for appropriate @xmath35 .
see @xcite .
neisendorfer considers the adjoint samelson products , so there is a dimension shift .
the goal of this chapter is to replace the co-@xmath0 spaces @xmath568}$ ] from [ theor3.22 ] by a finite wedge of moore spaces in case @xmath569 .
the iterated whitehead products involving @xmath72 are then replaced by iterated whitehead products in @xmath282 homotopy , which are more manageable . in [ subsec4.1 ] , we construct certain @xmath91 homotopy classes @xmath92 and @xmath93 for @xmath94 . this is a refinement of a similar construction in @xcite , and leads to a ladder of cofibration sequences . in [ subsec4.2 ] , we construct new co-@xmath0 spaces @xmath570 when @xmath569 , and introduce index @xmath38 approximation . using this
we exploit the fact ( @xcite ) that the identity map of @xmath43 has order @xmath38 to reduce the size of the set of obstructions .
this allows for the replacement of the iterated relative and @xmath0-space based whitehead products based on @xmath72 with iterated relative and @xmath0-space based whitehead products in the @xmath282 homotopy groups for @xmath571 .
the case @xmath31 is simpler and we show that @xmath50 is homotopy - abelian in the appendix .
nevertheless , the constructions in chapters [ chap4 ] , [ chap5 ] and [ chap6 ] will be used in chapter [ chap7 ] in case @xmath31 as well . in this section
we will assume an arbitrary @xmath0-space structure on the anick space as given in @xcite and use its existence to develop certain maps @xmath572 and @xmath573 for @xmath574 .
we begin with a strengthening of @xcite ) .
[ prop4.1 ] there is a map @xmath575 which induces an epimorphism in @xmath124 homology in dimensions @xmath246 and @xmath576 .
furthermore the composition of @xmath299 with the map @xmath577 ^ ->>>{\widetilde{e}}&s^{2n+1}\{p^r\ } } \ ] ] is null homotopic , where @xmath578 is adjoint to the map @xmath63 of [ theor2.14](e ) .
recall by [ theor2.14](g ) @xmath579\otimes\lambda(u)\ ] ] where @xmath580 , @xmath581 and @xmath582 . using some @xmath0-space structure map @xmath421
we consider the hopf construction : @xmath583 note that in homology @xmath584 if @xmath585 and @xmath586 .
we now define homology classes @xmath587 by the formulas @xmath588 @xmath589 \left(h(\mu)\right)_*\alpha&=-\sigma\otimes v^{p^k}\\ \left(h(\mu)\right)_*\beta&=\sigma\otimes uv^{p^k-1}.\end{aligned}\ ] ] also @xmath590 and @xmath591 are both nonzero . by [ theor2.14](d ) , @xmath592 is a wedge of moore spaces ; consequently there are maps @xmath593 such that @xmath327 is in the image of @xmath594 and @xmath133 is in the image of @xmath595 . combining these
we get a map @xmath299 @xmath596 ^ -{{\scalebox{.9}{${a\vee b}$}}}&\sigma(\gamma\wedge t)\ar@{->}[r]^-{{\scalebox{.9}{${h(\mu)}$}}}&\sigma t}\ ] ] such that @xmath597 and @xmath598 are in the image of @xmath599 . from this
we see that there is a homotopy commutative diagram @xmath600 ^ -{e}\ar@{->}[d]_{b\vee a}&\sigma t^{2np^k}\ar@{->}[d]\\ \sigma(t\wedge t)\ar@{->}[r]^-{h(\mu)}&\sigma t } \end{split}\ ] ] where @xmath299 induces an epimorphism in @xmath526 homology in dimensions @xmath246 and @xmath576 .
it remains to show that the composition @xmath601
^ -{{\scalebox{.9}{${h(\mu)}$ } } } & \sigma t\ar@{->}[r]^-{{\scalebox{.9}{${\widetilde{e}}$ } } } & s^{2n+1}\{p^r\ } } \ ] ] is null homotopic .
however , since @xmath602 is an @xmath0 map by [ theor2.14](e ) , there is a homotopy commutative diagram : @xmath603 ^ ->>>>>>>{\sigma(e\wedge e)}\ar@{->}[d]_{h(\mu)}&\sigma(\omega s^{2n+1}\{p^r\}\wedge \omega s^{2n+1}\{p^r\})\ar@{->}[d]^{h(\mu')}\\ \sigma t\ar@{->}[r]^->>>>>>>>>>>{\sigma e}&\sigma\omega s^{2n+1}\{p^r\ } .}\ ] ] where @xmath604 is the loop space structure map on @xmath605 .
since @xmath605 is a loop space , the right hand map is part of the classifying space structure @xmath606_{h(\mu')}\ar@{->}[r]&\dots\ar@{->}[r]&e^{\infty}\ar@{->}[d]\\ \sigma\omega s^{2n+1}\{p^r\}\ar@{->}[r]&\dots\ar@{->}[r]&s^{2n+1}\{p^r\ } } \ ] ] where @xmath607 is contractible and the bottom horizontal map is the evaluation map .
the result follows since @xmath578 is the composition : @xmath608 { } \xymatrix { \sigma t\ar@{->}[r ] & \sigma\omega s^{2n+1}\{p^r\}\ar@{->}[r]^-{{\scalebox{.9}{${ev}$ } } } & s^{2n+1}\{p^r\}. \makebox[64pt ] { } \rlap{\qed } } \ ] ] now recall from [ theor2.14](a ) the maps @xmath609^{g}&\omega g\ar@{->}[r]^{h}&t } \ ] ] with @xmath286 .
restricting we get @xmath610^{{\scalebox{.9}{${g_k}$ } } } & \omega g_k}$ ] .
let @xmath611 be the restriction of @xmath65 to @xmath72 .
then we have @xmath612_{g } & t^{2np^k } \ar@{->}[l ] \ar@{->}[d]^{g_*}\\ \omega g \ar@{=}[r ] \ar@{->}[d]_{h } & \omega g \ar@{->}[d]_{\omega\varphi } & \omega g_k \ar@{->}[l ] \ar@{->}[d]_{\omega\varphi_k}\\ t \ar@{->}[r]^->>>>>{e } & \omega s^{2n+1}\{p^r\ } \ar@{=}[r ] & \omega s^{2n+1}\{p^r\ } } \ ] ] where the left hand square commutes up to homotopy by [ theor2.14](e ) .
since @xmath613 , we get the homotopy commutative square : @xmath614^{g_k}\ar@{->}[d]&\omega g_k\ar@{->}[d]_{\omega \varphi_k}\\ t\ar@{->}[r]^->>>>>{e}&\omega s^{2n+1}\{p^r\}. } \ ] ] since @xmath299 factors through @xmath615 , we combine this with [ prop4.1 ] to see that the central composition @xmath616\\ p^{2np^k}(p^{r+k-1})\vee p^{2np^k+1}(p^{r+k-1})\ar@{-->}[rru]\ar@{->}[r]\ar@{->}[dr]^{e}&\sigma t^{2np^k}\ar@{->}[r]\ar@{->}[d]&g_k\ar@{->}[d]^{\varphi_k}\\ & \sigma t\ar@{->}[r]^->>>>>{\widetilde{e}}&s^{2n+1}\{p^r\ } } \ ] ] factors through @xmath79
. we state this as [ prop4.3 ] for any @xmath0-space structure on @xmath50 with corresponding maps @xmath68 and @xmath617 , there is a lifting of @xmath618 to @xmath79 @xmath619 ^ -{a(k)\vee c(k)}\ar@{->}[d]_e&e_k\ar@{->}[d]_{\pi_k}\\ \sigma t^{2np^k}\ar@{->}[r]^-{\widetilde{g}_k}&g_k .}\ ] ] in the diagram below , the left column is a standard cofibration sequence ( [ eq1.9 ] ) and the right column is a fibration sequence defined by the projection @xmath173 : @xmath620_{p^{r+k-1 } } \ar@{->}[rr]^{\theta_1 } & & \omega p^{2np^k\!{}+{}\!1}(p^{r+k } ) \ar@{->}[d ] \\
p^{2np^k}(p^{r+k } ) \ar@{->}[rr]^{\theta_2 } \ar@{->}[d]_{\sigma\vee \sigma \beta } & & j \ar@{->}[d ] \\
p^{2np^k}(p^{r+k-1})\!{}\vee{}\ ! p^{2np^k+1}(p^{r+k-1 } ) \ar@{->}[d]_{-\delta_1\vee \rho } \ar@{->}[r]^-{a(k)\vee c(k ) } & e_k \ar@{->}[r]^-{\pi_k } & g_k \ar@{->}[d]_{\pi } \\
p^{2np^k+1}(p^{r+k } ) \ar@{=}[rr ] & & p^{2np^k\!{}+{}\!1}(p^{r+k } ) . }
\ ] ] the homological properties of @xmath299 and [ prop4.3 ] imply that the bottom region commutes up to homotopy since @xmath621 has a right homotopy inverse .
the maps @xmath622 and @xmath623 are induced from this region in the standard way .
for dimensional reasons @xmath623 factors through @xmath624 and since @xmath245\ar@{->}[d]&g_{k-1}\ar@{->}[d]\\ e_k\ar@{->}[r]&g_k } \ ] ] is a pullback diagram , @xmath623 factors through @xmath253 .
@xmath622 factors through @xmath625 also for dimensional reasons .
we obtain [ theor4.4 ] there is a homotopy commutative ladder of cofibrations : , @xmath572 and @xmath573 for the composition @xmath626 , @xmath627 , and @xmath628 if it will not lead to confusion . ]
@xmath629_{p^{r+k-1 } } \ar@{=}[rr ] & & p^{2np^k}(p^{r+k } ) \ar@{->}[d]_{\alpha_k}\\ p^{2np^k}(p^{r+k } ) \ar@{->}[r]^-{\beta_k } \ar@{->}[d]_{\sigma\vee \sigma \beta } & e_{k-1 } \ar@{->}[r]^-{\pi_{k-1 } } \ar@{->}[d ] & g_{k-1 } \ar@{->}[d]\\ p^{2np^k}(p^{r+k-1})\!{}\vee{}\
! p^{2np^k+1}(p^{r+k-1 } ) \ar@{->}[r]^-{a(k)\vee c(k ) } \ar@{->}[d]_{-\delta_1\vee \rho } & e_k \ar@{->}[r]^-{\pi_k } & g_k \ar@{->}[d]_{\pi}\\ p^{2np^k+1}(p^{r+k } ) \ar@{=}[rr ] & & p^{2np^k+1}(p^{r+k } ) .}\ ] ] furthermore , for any choice of retraction @xmath630 , we can construct @xmath572 and @xmath573 so that @xmath631 .
we need only demonstrate the last statement .
suppose we are given a map @xmath632 so that the diagram commutes up to homotopy .
given a retraction @xmath633&bw_n}$ ] , we get a splitting @xmath634 by [ theor2.14](j ) where @xmath635 is the fiber of @xmath110 .
we can then write @xmath636 where @xmath189 is the component of @xmath637 that factors through @xmath638 and @xmath96 factors through @xmath639 . since each map
@xmath640 has order @xmath38 , @xmath189 has order @xmath38 and thus @xmath641 as @xmath642 .
thus the upper region commutes up to homotopy when @xmath637 is replaced by @xmath96 . since @xmath643
, @xmath189 factors @xmath644 ^ -{\sigma\vee\sigma\beta } & p^{2np^k}(p^{r+k-1})\vee p^{2np^k+1}(p^{r+k-1})\ar@{->}[r]^-{\epsilon ' } & e_{k-1 } } \ ] ] as the left hand column is a cofibration sequence .
we now redefine @xmath572 and @xmath573 by subtracting off the appropriate components of @xmath645 and the middle region now commutes up to homotopy . since this alteration of @xmath572 and @xmath573 factors through @xmath253 , the projections to @xmath72 vanish when projected to @xmath646 so the bottom region also commutes up to homotopy . during the inductive construction of @xmath83
we will be assuming that @xmath647 is defined for @xmath648 and the alterations in [ theor4.4 ] have been made so that @xmath96 is in the kernel of @xmath110 .
the goal of this section is to replace the co-@xmath0 space @xmath72 by a sequence of approximations
. the end result will be to replace @xmath568}$ ] by a wedge of @xmath649 moore spaces for @xmath98 .
we begin with a cofibration sequence based on the ladder [ theor4.4 ] . throughout this section we will exclude the case @xmath31 . the case
@xmath31 is dealt with in the appendix .
[ prop4.5 ] for @xmath574 there is a cofibration sequence @xmath650 ^ -{\xi_k}&l_k \ar@{->}[r]^-{\zeta_k}&g_k\ar@{->}[r]^-{\pi ' } & p^{2np^k+1}(p^{r+k } ) } \]]where @xmath651 , @xmath652 is induced by the inclusion of @xmath653 and the maps @xmath627 and @xmath628 , and @xmath654 .
this is a standard consequence of a ladder in which each third rail is an equivalence ( as in the usual proof of the mayer
vietoris sequence ) .
[ prop4.6 ] if @xmath569 , there is a unique co-@xmath0 space structure on @xmath570 so that the cofibration in [ prop4.5 ] is a cofibration of co-@xmath0 maps .
let @xmath655 be the pullback in the diagram : @xmath656\ar@{->}[d]&g_k\vee g_k\ar@{->}[d]\\ l_k\times l_k\ar@{->}[r ] & g_k\times g_k . } \ ] ] there is a map @xmath657 which projects to the diagonal map on @xmath570 and the composition : @xmath658 ^ -{\zeta _
g_k . } \ ] ] we first assert that @xmath659 is unique up to homotopy . since @xmath570 is a wedge of moore spaces , it suffices to show that if @xmath660 projects trivially to @xmath661 and @xmath662 , it is itself trivial .
now the homotopy fiber of the map @xmath663 is the same as the homotopy fiber of @xmath664 , i.e. , @xmath665 ; we conclude that @xmath189 must factor through , and that the composition @xmath658 ^ -{\epsilon'}&\sigma(\omega g_k\wedge \omega g_k)\ar@{->}[r ] & g_k\vee g_k } \ ] ] is null homotopic .
this implies that @xmath645 is null homotopic and hence @xmath189 is as well since @xmath666 has a right homotopy inverse .
the map @xmath667 is a @xmath668 equivalence since @xmath669 is @xmath668 connected .
since @xmath570 and @xmath72 are both @xmath670 connected , this implies that the composition @xmath671 & \sigma(\omega l_k\wedge \omega g_k)\ar@{->}[r ] & \sigma(\omega g_k\wedge \omega g_k ) } \ ] ] is a @xmath672 equivalence .
now consider the diagram of vertical fibrations : @xmath671\ar@{->}[d ] & \sigma(\omega g_k\wedge \omega g_k)\ar@{->}[r]^-{\simeq}\ar@{->}[d ] & \sigma(\omega g_k\wedge \omega g_k)\ar@{->}[d]\\ l_k\vee l_k\ar@{->}[r]\ar@{->}[d]&p\ar@{->}[r]\ar@{->}[d]&g_k\vee g_k\ar@{->}[d]\\ l_k\times l_k\ar@{=}[r]&l_k\times l_k\ar@{->}[r]&g_k\times g_k . } \ ] ] from this we see that the map @xmath673 is a @xmath672 equivalence . since @xmath569 , that implies that there is a unique lifting of @xmath657 to @xmath674 , which defines a co-@xmath0 space structure on @xmath570 such that @xmath667 is a co-@xmath0 map .
similarly , we observe that the @xmath675 skeleton of the fiber of the map @xmath676 is @xmath677 , so the composition @xmath650
^ -{\xi_k}&l_k\ar@{->}[r]&l_k\vee l_k } \ ] ] factors through @xmath677 and such a factorization defines a co-@xmath0 space structure on @xmath646 .
that structure , of course , is unique .
so this @xmath678 is a co-@xmath0 map with the suspension structure .
warning : @xmath570 does not split as a co - product of co-@xmath0 spaces . in particular , the inclusion @xmath679 is not a co-@xmath0 map . if it were , the map @xmath680 ^ -{c(k)}&e_k\ar@{->}[r]^-{\pi_k}&g_k } \ ] ] would be a co-@xmath0 map , contradicting @xcite .
write @xmath681\colon \sigma x\to \sigma x$ ] for the @xmath35-fold sum of the identity map .
[ def4.7 ] suppose @xmath682^{{\scalebox{.9}{${f}$}}}&h\ar@{->}[r]^{{\scalebox{.9}{${g}$}}}&\sigma k}$ ] is a cofibration sequence of co-@xmath0 spaces and co-@xmath0 maps .
we will say that @xmath683 is an index @xmath38 approximation if there is a co-@xmath0 map @xmath684 such that @xmath685 factors @xmath686 ^ -{g'}&\sigma^2k\ar@{->}[r]^-{[p]}&\sigma^2k\\ } \ ] ] up to homotopy .
@xmath687 will be called an iterated index @xmath38 approximation if @xmath683 is homotopic to a composition @xmath688 where each map @xmath689 is an index @xmath38 approximation .
thus , for example , @xmath690 is an index @xmath38 approximation .
[ prop4.8 ] suppose that @xmath687 is an iterated index @xmath38 approximation and @xmath691 .
then @xmath53 is null homotopic iff @xmath692 is null homotopic .
we will only consider the case when @xmath683 is an index @xmath38 approximation , as the general result follows by an easy induction .
suppose then that @xmath687 is an index @xmath38 approximation and @xmath685 factors up to homotopy : @xmath686 ^ -{g'}&\sigma^2 k\ar@{->}[r]^-{[p]}&\sigma^2k . } \ ] ] assume that @xmath692 is null homotopic , so we can factor @xmath53 as @xmath693 ^ -{g}&\sigma k\ar@{->}[r]^-{\nu'}&bw_n . } \ ] ] consider the diagram : @xmath694 ^ -{\omega[p ] } & \omega \sigma^2k\ar@{->}[r]^-{\omega\sigma\nu ' } & \omega \sigma bw_n\ar@{->}[r ] & bw_n\\ h\ar@{->}[r]^-{g}\ar@{->}[u]^{\widetilde{g}'}&\sigma k\ar@{->}[r]^-{\nu'}\ar@{->}[u]&bw_n\ar@{=}[ur]\ar@{->}[u ] } \ ] ] since @xmath43 is homotopy associative ( @xcite ) , the upper composition is an @xmath0-map .
this composition is thus inessential if its restriction to @xmath695 is inessential .
however this restriction factors through @xmath696\colon\sigma k\to \sigma k$ ] .
since @xmath43 has @xmath0-space exponent @xmath38 ( @xcite ) , we conclude that the upper composition is inessential .
this @xmath697 is inessential as well .
[ lem4.9 ] there is a homotopy commutative diagram @xmath698 ^ -{[p]\circ 1 } \ar@{}[d]^{\rotatebox{90}{$\simeq$ } } & ( \sigma^2h)\circ k \ar@{}[d]^{\rotatebox{90}{$\simeq$ } } \\
( \sigma h)\wedge k \ar@{->}[r]^-{[p]\wedge 1 } \ar@{}[d]^{\rotatebox{90}{$\simeq$ } } & ( \sigma h)\wedge k \ar@{}[d]^{\rotatebox{90}{$\simeq$ } } \\ \sigma^2(h\circ k ) \ar@{->}[r]^-{[p ] } & \sigma^2(h\circ k ) } \ ] ] where the equivalences are co-@xmath0 equivalences .
the vertical equivalences follow from [ prop3.2 ] .
these equivalences are natural for co-@xmath0 maps .
however @xmath696\colon \sigma h\to \sigma h$ ] is a co-@xmath0 map since @xmath0 is a co-@xmath0 space .
[ lem4.10 ] suppose @xmath699 ^ -{{\scalebox{.9}{${\alpha}$}}}&g_2\ar@{->}[r]^-{{\scalebox{.9}{${\beta}$}}}&g_3}$ ] is a cofibration sequence of co-@xmath0 spaces and co-@xmath0 maps .
then for each co-@xmath0 space @xmath0 , @xmath700 ^ -{\alpha\circ 1 } & g_2\circ h \ar@{->}[r]^-{\beta\circ 1 } & g_3\circ h\\ h\circ g_1 \ar@{->}[r]^-{1\circ \alpha } & h\circ g_2 \ar@{->}[r]^-{1\circ \beta } & h\circ g_3 } \ ] ] are both cofibration sequences . in the extended cofibration sequence ,
the composition of two adjacent maps is null homotopic @xmath701 ^
-{\alpha } & g_2\ar@{->}[r]^-{\beta } & g_3\ar@{->}[r]^-{\gamma } & \sigma g_1\ar@{->}[r]^-{\sigma\alpha } & \sigma g_2\ar@{->}[r]^-{\sigma\beta } & \sigma g_3\ar@{->}[r]&\dots } \ ] ] and all maps are co-@xmath0 maps .
it follows that the same is true for the sequence : @xmath702 ^ -{\alpha\circ 1 } & g_2\circ h \ar@{->}[r]^-{\beta\circ 1 } & g_3\circ h \ar@{->}[r]^-{\gamma\circ 1 } & ( \sigma g_1)\circ h\ar@{->}[r]&\dots } \ ] ] where @xmath703 . since this sequence also induces an exact sequence in homology it is a cofibration sequence .
the other case is similar .
[ prop4.11 ] if @xmath687 is an index @xmath38 approximation , the maps @xmath704 and @xmath705 are index @xmath38 approximations as well .
factor @xmath617 as @xmath686 ^ -{g'}&\sigma^2k\ar@{->}[r]^-{[p]}&\sigma^2k } \ ] ] and consider the diagram @xmath706 ^ -{g\circ 1}&(\sigma^2k)\circ l & \hspace*{-16pt}\llap{{\scalebox{.9}{${\simeq{}}$ } } } \sigma^2(k\circ l)\\ ( \sigma h)\circ l \ar@{->}[r]^-{g'\circ 1 } \ar@{=}[u]^{= } & ( \sigma^2k)\circ l \ar@{->}[u]^{[p]\circ 1 } & \hspace*{-16pt}\llap{{\scalebox{.9}{${\simeq{}}$ } } } \sigma^2(k\circ l ) \ar@<8pt>[u]^{[p]}}\end{aligned}\ ] ] where the right hand square commutes by [ lem4.9 ] .
the map @xmath707 is the cofiber of @xmath708 , and @xmath709 is a co-@xmath0 map since @xmath710 is a co-@xmath0 map .
thus @xmath708 is an index @xmath38 approximation .
the other case is similar .
[ cor4.12 ] suppose @xmath682 ^ -{{\scalebox{.9}{${f}$}}}&h}$ ] is an index @xmath38 approximation .
then @xmath711}\!\colon g^{[i]}\to h^{[i]}\ ] ] is an iterated index @xmath38 approximation .
we first observe that @xmath712}\ar@{->}[r]^-{{\scalebox{.9}{${f\circ 1}$}}}&h\circ h^{[j]}=h^{[j+1]}}\ ] ] is an index @xmath38 approximation by [ prop4.11 ] .
we then see by induction that @xmath713}h^{[j]}=g\circ ( g^{[i-1]}h^{[j]})\to g\circ(g^{[i-2]}h^{[i+1]})=g^{[i-1]}h^{[j+1]}\end{aligned}\ ] ] is an index @xmath38 approximation .
finally @xmath714}\to g^{[i-1]}h\ ] ] is an iterated index @xmath38 approximation by induction since it factors as @xmath714}=g\circ(g^{[i-1]})\to g\circ(g^{[i-2]}h)=g^{[i-1]}h.\ ] ] consequently @xmath714}\to g^{[i-1]}h\to g^{[i-2]}h^{[2]}\to \dots\to g\circ h^{[i-1]}\to h^{[i]}\ ] ] is an iterated index @xmath38 approximation .
[ theor4.13 ] suppose @xmath715 is a principal fibration classified by a map @xmath498 where @xmath198 is an @xmath0-space with strict unit .
suppose @xmath716 is an index @xmath38 approximation .
then , for any map @xmath717 the compositions
@xmath718 ^ -{{\scalebox{.9}{${\gamma}$ } } } & e\ar@{->}[r]^-{{\scalebox{.9}{${\nu}$ } } } & bw_n } \ ] ] is null homotopic iff the composition @xmath719
^ -{{\scalebox{.9}{${\sigma(\omega f\wedge\omega f)}$ } } } & \sigma(\omega g\wedge \omega g)\ar@{->}[r]^-{{\scalebox{.9}{${\gamma}$ } } } & e\ar@{->}[r]^-{{\scalebox{.9}{${\nu}$ } } } & bw_n } \ ] ] is null homotopic .
suppose the composition @xmath720&\sigma(\omega
g\wedge \omega g)\ar@{->}[r]^-{{\scalebox{.9}{${\gamma}$ } } } & e\ar@{->}[r]^-{{\scalebox{.9}{${\nu}$ } } } & bw_n } \ ] ] is null homotopic .
let @xmath502 be the identity map and @xmath721 .
since @xmath716 is a co-@xmath0 map , there is a homotopy commutative diagram : @xmath722}\ar@{->}[r]^-{ad_r^{i-2}(\beta)(\{\beta,\beta\}_{\times } ) } \ar@{->}[d]_{f^{[i ] } } & \sigma(\omega h\wedge \omega h ) \ar@{->}[d]\\ g^{[i ] } \ar@{->}[r]^-{ad^{i-2}(\alpha)(\{\alpha,\alpha\}_{\times } ) } & \sigma(\omega g\wedge \omega g ) } \ ] ] by [ prop3.6](d ) and [ prop3.11](d ) .
since @xmath683 is an index @xmath38 approximation , @xmath723}$ ] is an iterated index @xmath38 approximation by [ cor4.12 ] ; thus the compositions @xmath724}\ar@{->}[r]^-{{\scalebox{.9}{${ad^{i-2}(\alpha)(\{\alpha,\alpha\}_{\times})}$ } } } & \sigma(\omega g\wedge \omega g)\ar@{->}[r]^-{{\scalebox{.9}{${\gamma}$ } } } & e\ar@{->}[r]^-{{\scalebox{.9}{${\nu}$ } } } & bw_n } \ ] ] are null homotopic for all @xmath90 .
the result then follows from [ theor3.22 ] .
we will use this result to transfer conditions on @xmath83 to the composition : @xmath671 & \sigma(\omega g_k\wedge \omega g_k)\ar@{->}[r]^-{{\scalebox{.9}{${\gamma_k}$ } } } & e_k . } \ ] ] we need to iterate this .
we have to consider the issue that for @xmath725 to be a co-@xmath0 map , we need to use an exotic co-@xmath0 space structure on @xmath570 .
we will show that the triviality of the composition above does not depend on the co-@xmath0 space structure of @xmath570 .
to see this , recall that the map@xmath726 was defined in section [ subsec2.2 ] based on the fact that @xmath79 was defined by a principal fibration @xmath727 classified by a map @xmath728 where @xmath55 is an @xmath0-space with @xmath0-space structure map chosen to have a strict unit .
the fact that @xmath72 is a co-@xmath0 space was not used . for any space @xmath198 and map @xmath729
, we can construct the pullback @xmath730 \ar@{->}[d ] & \omega s^{2n+1}\{p^r\}\ar@{->}[d]\\ e(x)\ar@{->}[r]\ar@{->}[d]&e_k\ar@{->}[d]\\ x\ar@{->}[r]^-{\zeta}&g_k } \ ] ] which is induced by the composition @xmath731 .
consequently there is a map @xmath732 and a strictly commutative diagram : @xmath733\ar@{->}[d]_{\gamma(x ) } & \sigma(\omega g_k\wedge \omega g_k)\ar@{->}[d]_{\gamma_k}\\ e(x)\ar@{->}[r]&e_k . }
\ ] ] consider the homotopy equivalence @xmath734 where we give @xmath735 the split co-@xmath0 space structure , so this map is not a co-@xmath0 map .
nevertheless , we have a strictly commutative diagram @xmath736 ^ -{\simeq}\ar@{->}[d]_{\gamma(x_k ) } & \sigma(\omega l_k\wedge\omega l_k)\ar@{->}[r]\ar@{->}[d]_{\gamma(l_k ) } & \sigma(\omega g_k\wedge\omega g_k)\ar@{->}[d]_{\gamma_k}\\ e(x_k)\ar@{->}[r]^-{\simeq}\ar@{->}[d]&e(l_k)\ar@{->}[r]\ar@{->}[d]&e_k\ar@{->}[d]_{\pi_k}\\ x_k\ar@{->}[r]^-{\simeq}&l_k\ar@{->}[r]^-{\zeta_k}&g_k } \ ] ] since @xmath667 is an iterated index @xmath38 approximation , we have [ prop4.14 ] for any map @xmath737 , @xmath738 is null homotopic iff the composition @xmath739 ^ -{{\scalebox{.9}{${\gamma(x_k)}$ } } } & e(x_k)\simeq e(l_k)\ar@{->}[r]&e_k\ar@{->}[r]^-{{\scalebox{.9}{${\nu}$ } } } & bw_n } \ ] ] is null homotopic .
we now define spaces with split co-@xmath0 space structures ( coproducts in the category of co-@xmath0 spaces ) : @xmath740 consequently we have homotopy equivalences @xmath741 which are not co-@xmath0 equivalences , and index @xmath38 approximations @xmath742 ^ -{{\scalebox{.9}{${\zeta_j\vee 1}$ } } } & g(j , k).}\ ] ] this leads to a chain : @xmath743 [ theor4.15 ] suppose @xmath569 .
then for any given map @xmath737 the composition @xmath744 ^ -{{\scalebox{.9}{${\gamma_k}$ } } } & e_k\ar@{->}[r]^-{{\scalebox{.9}{${\nu}$ } } } & bw_n } \ ] ] is null homotopic iff the composition @xmath745 ^ -{{\scalebox{.9}{${\sigma(\omega f\wedge\omega f)}$ } } } & \sigma(\omega g_k\wedge\omega g_k ) } \hspace*{-6pt}\xymatrix{\mbox{}\ar@{->}[r]^-{{\scalebox{.9}{${\gamma_k}$ } } } & e_k\ar@{->}[r]^-{{\scalebox{.9}{${\nu}$ } } } & bw_n } \ ] ] is null homotopic , where @xmath746 and the map @xmath747 is defined by the inclusion of @xmath748 and the maps @xmath749 and @xmath750 for @xmath751 : @xmath752 { } & \xymatrix{p^{2np^i}(p^{r+i-1})\ar@{->}[r]^-{{\scalebox{.9}{${a(i)}$}}}&e_i\ar@{->}[r]^-{{\scalebox{.9}{${}$ } } } & e_k\ar@{->}[r]^-{{\scalebox{.9}{${\pi_k}$}}}&g_k}\\ & \xymatrix{p^{2np^i+1}(p^{r+i-1})\ar@{->}[r]^-{{\scalebox{.9}{${c(i)}$}}}&e_i\ar@{->}[r]^-{{\scalebox{.9}{${}$ } } } & e_k\ar@{->}[r]^-{{\scalebox{.9}{${\pi_k}$}}}&g_k . }
\makebox[53.5pt ] { } \rlap{\qed}\end{aligned}\ ] ] observe the diagram @xmath753\ar@{->}[d]&e_k\ar@{->}[d]\\ \sigma(\omega g(0,k)\wedge \omega g(0,k))\ar@{->}[r]\ar@{->}[ru]^{\gamma}&g(0,k)\ar@{->}[r]^-{f}&g_k } \ ] ] where @xmath150 is @xmath754 composed with @xmath755 . since @xmath756 is a wedge of moore spaces , the components of @xmath150
are @xmath0-space based whitehead products as defined by neisendorfer @xcite .
this will be studied in the next chapter .
in this chapter we work with the obstructions obtained in chapter [ chap4 ] .
these are @xmath282 homotopy classes for @xmath98 . in section [ subsec5.1 ]
, we define a quotient space @xmath103 of @xmath72 and a corresponding principal fibration @xmath102 over @xmath103 .
this has the property that roughly half of the obstructions vanish in @xmath102 ; we then seek a factorization of @xmath83 through @xmath102 . in section [ subsec5.2 ]
we introduce a congruence relation on homotopy classes , and show that we need only consider the obstructions up to congruence homotopy .
this leads to a shorter list of obstructions .
congruence homotopy will also be useful in chapter [ chap6 ] , since the properties of relative whitehead products are simpler up to congruence homotopy .
the inductive hypothesis ( [ theor6.7 ] ) in the next chapter is a strengthening of proposition [ prop2.12 ] , so in particular , we will be assuming that we have constructed a retraction @xmath78 such that @xmath757 is null homotopic .
in section [ subsec4.1 ] we constructed classes @xmath92 , @xmath93 , and @xmath758 for @xmath94 and in section [ subsec4.2 ] we reduced the constraints on the construction of @xmath83 to a condition involving the maps @xmath92 and @xmath93 .
some of the material in this section and section [ subsec6.2 ] can be found at arxiv:0804.1896 .
we now make a further simplification by burying the classes @xmath93 in the base space .
specifically , we define a map @xmath759 by the compositions @xmath760 ^ -{{\scalebox{.9}{${c(i)}$ } } } & e_i\ar@{->}[r ] & e_k , } \ ] ] and define were first defined in @xcite , but were abandoned in @xcite as the related spaces @xmath72 have better properties . as we will see , the spaces @xmath103 have smaller homotopy which is useful in our analysis . ]
@xmath103 by a cofibration @xmath761 ^ -{{\scalebox{.9}{${\pi_kc}$ } } } & g_k\ar@{->}[r ] & d_k . } \ ] ] [ prop5.1 ] there is a homotopy commutative diagram of cofibration sequences @xmath762
^ -{\rho } & p^{2np^k+1}(p^{r+k})\ar@{->}[r]^-{\sigma^{r+k-1 } } & p^{2np^k+1}(p)\\ c_k\ar@{->}[r]^-{\pi_kc}\ar@{->}[u ] & g_k\ar@{->}[r]\ar@{->}[u]^{\pi } & d_k\ar@{->}[u]\\ c_{k-1}\ar@{->}[r]^-{\pi_{k-1}c}\ar@{->}[u]&g_{k-1}\ar@{->}[r]\ar@{->}[u]&d_{k-1}\ar@{->}[u ] } \ ] ] and @xmath763 the composition @xmath680 ^ -{{\scalebox{.9}{${c(k)}$ } } } & e_k\ar@{->}[r]^-{{\scalebox{.9}{${\pi_k}$ } } } & g_k\ar@{->}[r]^-{{\scalebox{.9}{${\pi}$ } } } & p^{2np^k+1}(p^{r+k } ) } \ ] ] is @xmath135 by [ theor4.4 ] .
the homology calculation is immediate .
since @xmath764 is null homotopic , we can extend @xmath765 to a map @xmath766 any such extension defines a diagram of vertical fibration sequences : @xmath767 ^ -{\tau_k } \ar@{->}[d]_{\pi_k } & j_k \ar@{->}[r]^-{\eta_k } \ar@{->}[d]_{\xi_k } & f_k \ar@{->}[d]_{\sigma_k } \\
g_k\ar@{->}[r]\ar@{->}[d]_{\varphi_k } & d_k\ar@{=}[r]\ar@{->}[d]_{\varphi'_k } & d_k\ar@{->}[d]\\ s^{2n+1}\{p^r\}\ar@{->}[r ] & s^{2n+1}\{p^r\}\ar@{->}[r ] & s^{2n+1 } } \end{split}\ ] ] [ prop5.3 ] we can choose an extension @xmath768 of @xmath769 in such a way that the composition @xmath680 ^ -{{\scalebox{.9}{${c(k)}$ } } } & e_k\ar@{->}[r]^-{{\scalebox{.9}{${\tau_k}$ } } } & j_k } \ ] ] is null homotopic .
we begin by defining @xmath770 by a pushout square : @xmath771&d\\ g_{k-1}\ar@{->}[u]\ar@{->}[r]&d_{k-1}\ar@{->}[u ] } \ ] ] using the lower right hand square in [ prop5.1 ] and the pushout property , we see that there is a cofibration @xmath680 ^ -{{\scalebox{.9}{${\alpha}$ } } } & d\ar@{->}[r ] & d_k } \ ] ] where @xmath327 is the composition : @xmath680 ^ -{{\scalebox{.9}{${c(k)}$ } } } & e_k\ar@{->}[r]^-{{\scalebox{.9}{${\pi_k}$ } } } & g_k\ar@{->}[r ] & d. } \ ] ] we use the pushout property to construct @xmath772 by @xmath769 on @xmath773 and @xmath765 on @xmath72 .
we seek a map @xmath768 in the diagram @xmath104\ar@{->}[d]_{\pi_k}&j\ar@{->}[r]\ar@{->}[d]&j_k\ar@{->}[d]\\ g_k\ar@{->}[r]\ar@{->}[d]_{\varphi_k}&d\ar@{->}[r]\ar@{->}[d]_{\varphi}&d_k = d\cup_{\alpha}cp^{2np^k+1}(p^{r+k-1})\ar@{->}[d]_{\varphi'_k}\\ s^{2n+1}\{p^r\}\ar@{=}[r]^{}&s^{2n+1}\{p^r\}\ar@{=}[r]^{}&s^{2n+1}\{p^r\ } } \ ] ] we assert that we can choose @xmath768 so that the composition @xmath680 ^ -{{\scalebox{.9}{${c(k)}$ } } } & e_k\ar@{->}[r ] & j\ar@{->}[r ] & j_k } \ ] ] is null homotopic .
note that @xmath327 is homotopic to the composition @xmath680
^ -{{\scalebox{.9}{${c(k)}$ } } } & e_k\ar@{->}[r ] & j\ar@{->}[r ] & d. } \ ] ]
the assertion then follows from [ lem5.4 ] suppose @xmath774 ^ -{{\scalebox{.9}{${\pi}$}}}&d}$ ] is a principal fibration induced by a map @xmath775 .
suppose @xmath776 and @xmath777 is the mapping cone of @xmath778 .
then there is a map @xmath779 with homotopy fiber @xmath780 such that the composition @xmath781&j\ar@{->}[r]&j'}$ ] is null homotopic in the diagram : @xmath782 ^ -{c}&j\ar@{->}[r]\ar@{->}[d]_{\pi}&j'\ar@{->}[d]\\ & d\ar@{->}[r]\ar@{->}[d]_{\varphi}&d'\ar@{->}[d]_{\varphi'}\\ & s\ar@{=}[r]^{}&s . } \ ] ] @xmath783 so @xmath784 has components @xmath785 where @xmath786 , @xmath787 with @xmath788 and @xmath789 .
write @xmath790 with @xmath186 at the vertex of the cone and @xmath791 .
now define @xmath792 by @xmath793 for @xmath794 and @xmath795 .
this is well defined and we can define a homotopy @xmath796 by the formula @xmath797 where @xmath798 is the path defined as @xmath799 .
this proves the lemma and hence the proposition
. now assume @xmath569 and define @xmath800 so @xmath801 , and we have a homotopy commutative square @xmath802\ar@{->}[d]&g_k\ar@{->}[d]\\ u_k\ar@{->}[r]^-{a}&d_k } \ ] ] where the left hand vertical map is the projection and @xmath175 is defined on@xmath803 as the composition : @xmath804 ^ -{{\scalebox{.9}{${a(i)}$ } } } & e_k\ar@{->}[r]^-{{\scalebox{.9}{${\tau_k}$ } } } & j_k\ar@{->}[r]^-{{\scalebox{.9}{${\xi_k}$ } } } & d_k . } \ ] ] from this we construct homotopy commutative diagram : @xmath805\ar@{->}[d ] & \sigma(\omega g_k\wedge\omega g_k)\ar@{->}[r]^-{\gamma_k}\ar@{->}[d ] & e_k\ar@{->}[d]_{\tau_k}\\ \sigma(\omega u_k\wedge \omega u_k)\ar@{->}[r ] & \sigma(\omega d_k\wedge \omega d_k)\ar@{->}[r]^-{\overline{\gamma}_k } & j_k } \end{split}\ ] ] [ prop5.7 ] suppose that @xmath569 and there is a retraction @xmath806 such that the compositions @xmath807}\ar@{->}[r]^-{{\scalebox{.9}{${\{a,\dots , a,\{a , a\}_{\times}\}_r}$ } } } & j_k\ar@{->}[r]^-{{\scalebox{.9}{${\gamma_k}$ } } } & bw_n } \ ] ] are null homotopic for each @xmath808
. then the composition @xmath809 ^ -{{\scalebox{.9}{${\gamma_k}$ } } } & e_k\ar@{->}[r]^-{{\scalebox{.9}{${\tau_k}$ } } } & j_k\ar@{->}[r]^-{{\scalebox{.9}{${\gamma_k}$ } } } & bw_n } \ ] ] is null homotopic . by ( [ eq5.6 ] ) and [ theor4.15 ]
, it suffices to show that the composition @xmath810 & \sigma(\omega d_k\wedge \omega d_k)\ar@{->}[r]^-{{\scalebox{.9}{${\overline{\gamma}_k}$ } } } & j_k\ar@{->}[r]^-{{\scalebox{.9}{${\gamma_k}$ } } } & bw_n } \ ] ] is null homotopic .
define @xmath811 as a pullback : @xmath812\ar@{->}[d]&j_k\ar@{->}[d]\\ u_k\ar@{->}[r]^-{a}&d_k } \ ] ] then by naturality , it suffices to show that the composition @xmath813
^ -{{\scalebox{.9}{${\gamma(u_k)}$ } } } & e(u_k)\ar@{->}[r ] & j_k\ar@{->}[r]^-{{\scalebox{.9}{${\gamma_k}$ } } } & bw_n } \ ] ] is null homotopic .
but since @xmath814 is a co-@xmath0 space , we can apply [ theor3.22 ] to finish the proof .
now write @xmath815 where @xmath816 so @xmath817}_k=\sigma p_k\wedge\dots\wedge p_k$ ] by [ prop3.2 ] . using the splitting of @xmath818 into @xmath819 moore spaces , we obtain [ prop5.8 ] suppose @xmath569 .
then the map @xmath820}=\sigma p_k^{(j)}\to j_k\ ] ] when restricted to one of the @xmath821 iterated smash products of moore spaces is an iterated external whitehead product of the form @xmath822 where each @xmath823 is either @xmath824 for @xmath751 or the inclusion @xmath825 . by applying [ prop3.26 ] , we resolve these external whitehead products into internal whitehead products .
[ theor5.9 ] suppose @xmath569 .
then the restriction of the map @xmath826 to any moore space in any decomposition of @xmath827 is homotopic to a linear combination of weight @xmath828 iterated internal @xmath0-space based whitehead products @xmath829_{\times}]_r\ ] ] where each @xmath823 is one of the following : @xmath830 , @xmath831 , @xmath421 , @xmath53 for @xmath751 and for appropriate values of @xmath832 .
this is an easy induction on @xmath828 using [ prop3.26 ] .
the results of section [ subsec5.1 ] , and in particular [ theor5.9 ] , indicate that the obstructions to constructing a suitable retraction @xmath833 are @xmath834 homotopy classes in @xmath102 for @xmath98 .
these obstructions are iterated compositions of relative and @xmath0-space based whitehead products .
however , they are not whitehead products of maps into @xmath102 in general . since @xmath43 is an @xmath0-space , any whitehead products of classes in @xmath102 will be annihilated by any such map @xmath105 .
we are led to a coarser classification .
we introduce a congruence relation among homotopy classes so that it is only necessary to annihilate a representative of each congruence class .
a remarkable and useful feature is that the congruence homotopy of @xmath102 is a module over the symmetric algebra on @xmath835 .
[ def7.1 ] two maps @xmath836 will be called congruent ( written @xmath837 ) if @xmath838 and @xmath685 are homotopic in @xmath839 $ ] .
we write @xmath840 $ ] for the set of congruence classes of pointed maps : @xmath841 and @xmath842.\ ] ] clearly congruence is an equivalence relation and composition is well defined on congruence classes .
this defines the congruence homotopy category .
it is easy to prove [ prop5.11 ] suppose @xmath843 and @xmath844 where @xmath14 is an @xmath0-space . @xmath845 and @xmath846 are homotopic .
consequently , it is sufficient to classify the iterated @xmath0-space based whitehead products of [ theor5.9 ] up to congruence . we will also need to consider congruence in a different way in section [ subsec6.3 ] . in constructing @xmath105
we will make alterations in dimensions where obstructions of level @xmath111 may resurface .
this is a delicate point which has needed much attention .
for this reason we need to develop some deeper properties of congruence homotopy theory .
[ prop5.12 ] the inclusion @xmath847 ^ -{{\scalebox{.9}{${i}$}}}&y_1\times y_2}$ ] defines a map @xmath848\to e[x , y_1\times y_2]$ ] .
furthermore , if @xmath59 is a co-@xmath0 space @xmath849 $ ] is a abelian group and @xmath850\cong e[g , y_1\times y_2]\cong e[g , y_1]\oplus e[g , y_2].\ ] ] suppose @xmath851 and the compositions : @xmath852 ^ -{{\scalebox{.9}{${\sigma f}$}}}&\sigma(y_1\vee y_2)\ar@{->}[r]^-{{\scalebox{.9}{${\sigma i}$}}}&\sigma(y_1\times y_2)}\\ \xymatrix{\sigma x\ar@{->}[r]^-{{\scalebox{.9}{${\sigma g}$}}}&\sigma(y_1\vee y_2)\ar@{->}[r]^-{{\scalebox{.9}{${\sigma i}$}}}&\sigma(y_1\times y_2)}\end{aligned}\ ] ] are homotopic
. since @xmath853 has a left homotopy inverse , @xmath838 and @xmath685 are homotopic .
the co-@xmath0 space structure on @xmath59 defines a multiplication on @xmath854 $ ] and the map @xmath85\to [ g,\omega\sigma x]\ ] ] is multiplicative .
however @xmath855 $ ] is a abelian group by a standard argument .
since @xmath849 $ ] is a subgroup of @xmath855\cong [ \sigma g,\sigma x]$ ] , it also is abelian . finally observe that the composition @xmath856\oplus e[g , y_2]\to e[g , y_1\vee y_2]\to e[g , y_1\times y_2]\to e[g , y_1]\oplus e[g , y_2]\ ] ] is the identity where the first and last maps are defined by naturality .
thus the composition of the first two is . but
this composition is also onto since any element of @xmath857 $ ] is represented by a map @xmath858 so all these maps are isomorphisms .
[ prop5.13 ] suppose @xmath59 and @xmath0 are co-@xmath0 spaces .
then composition defines a homomorphism : @xmath859\otimes e[h , x]\to e[g , x].\ ] ] the only issue is the distributive law @xmath860 for @xmath861 and @xmath862 . to prove this we show that the following diagram commutes up to congruence : @xmath863 ^ -{\alpha}\ar@{->}[d]\ar@{}[dr]|{\equiv}&h\ar@{->}[d]\\ g\vee g\ar@{->}[r]^-{\alpha\vee\alpha}&h\vee h. } \ ] ] this certainly commutes after the inclusion of @xmath864 .
thus it commutes up to congruence by [ prop5.12 ]
. this will be useful when @xmath59 and @xmath0 are moore spaces .
[ cor5.14 ] the category of co-@xmath0 spaces and congruence classes of contiuous maps is an additive category .
@xmath865 is both a product and co - product by [ prop5.12 ] and composition is bilinear by [ prop5.13 ]
. the following result will be needed in [ prop5.19 ] and in section [ subsec6.3 ] .
[ theor5.15 ] suppose @xmath866 has order @xmath4
. then there are maps @xmath867 and @xmath868 such that @xmath869 according to @xcite or @xcite , there is a fibration sequence @xmath870 ^ -{{\scalebox{.9}{${\partial}$ } } } & s^{2m-1}\{p^{r+t}\}\ar@{->}[r ] & w\ar@{->}[r]^-{{\scalebox{.9}{${\pi}$ } } } & p^{2m}(p^{r+t } ) } \ ] ] where @xmath60 is a @xmath871 connected wedge of moore spaces and @xmath173 is an iterated whitehead product on each factor . in particular
, @xmath173 is null congruent .
a right homotopy inverse for @xmath54 is constructed as follows .
given any map @xmath872 , there is a natural map from the fiber of @xmath159 to the loop space on the cofiber : @xmath873 this defines a map @xmath874 and @xmath875 is a homology equivalence since @xmath876 is atomic .
this defines a splitting of @xmath877 and we have a direct sum decomposition @xmath878\oplus [ \sigma x , s^{2m-1}\{p^{r+t}\}]\ar@{->}[r ] & [ \sigma^2x , p^{2m}(p^{r+t})]\\ ( \alpha,\beta)\ar@{<->}[r]&\varphi=\pi\alpha+\widetilde{\phi\beta } } \ ] ] where @xmath879 is the adjoint of @xmath880 .
since @xmath65 has order @xmath4 , both @xmath327 and @xmath133 have order @xmath4 .
since @xmath173 is null congruent , we have @xmath881 now consider the diagram of fibration sequences : @xmath882 ^ -{f}\ar@{->}[d]&s^{2m-1}\{p^{r+t}\}\ar@{->}[r]^-{p^r}\ar@{->}[d]&s^{2m-1}\{p^{r+t}\}\ar@{->}[d]\\ s^{2m-1}\{p^r\}\ar@{->}[r]\ar@{->}[d]_{*\sim p^{r+t}}&s^{2m-1}\ar@{->}[r]^-{p^r}\ar@{->}[d]_{p^{r+t}}&s^{2m-1}\ar@{->}[d]_{p^{r+t}}\\ s^{2m-1}\{p^r\}\ar@{->}[r]&s^{2m-1}\ar@{->}[r]^-{p^r}&s^{2m-1}. } \ ] ] from this we see that @xmath883 .
we choose a splitting of @xmath167 as follows : define a map @xmath884 by the diagram of vertical fibration sequences : @xmath885 ^ -{\rho^t}\ar@{->}[d ] & s^{2m-1}\{p^{r+t}\}\ar@{->}[d]\\ s^{2m-1}\ar@{=}[r]\ar@{->}[d]_{p^r } & s^{2m-1}\ar@{->}[d]_{p^{r+t}}\\ s^{2m-1}\ar@{->}[r]^-{p^t } & s^{2m-1}. } \ ] ] the map @xmath884 factors through @xmath683 and defines a splitting .
thus the composition
@xmath886 ^ -<<<{\simeq } & f\ar@{->}[r]^-{f}&s^{2m-1}\{p^{r+t}\ } } \ ] ] is homotopic to the map @xmath887 ^ -{{\scalebox{.9}{${\rho^t\times\delta_t}$ } } } & s^{2m-1}\{p^{r+t}\}\times s^{2m-1}\{p^{r+t}\}\ar@{->}[r]^-{{\scalebox{.9}{${\mu}$}}}&s^{2m-1}\{p^{r+t}\}}\end{gathered}\ ] ] where @xmath888 is the composition @xmath889 ^ -{{\scalebox{.9}{${\omega\pi}$ } } } & \omega s^{2m-1}\ar@{->}[r]^-{{\scalebox{.9}{${\iota}$ } } } & s^{2m-1}\{p^{r+t}\}. } \ ] ] since @xmath133 has order @xmath4 , it factors through @xmath167 and we conclude that @xmath133 is homotopic to a composition : @xmath890 & s^{2m-1}\{p^r\}\times \omega s^{2m-1}\{p^r\ } } \\
\xymatrix@c=37pt { \ar@{->}[r]^-{{\scalebox{.9}{${\rho^t\times \delta_t}$ } } } & s^{2m-1}\{p^{r+t}\}\times s^{2m-1}\{p^{r+t}\ } \ar@{->}[r]&s^{2m-1}\{p^{r+t}\}. } \end{gathered}\ ] ] this map is homotopic to the sum of the two compositions @xmath891 & s^{2m-1}\{p^r\}\ar@{->}[r]^-{{\scalebox{.9}{${\rho^t}$ } } } & s^{2m-1}\{p^{r+t}\ } } \\ & \xymatrix { \sigma x\ar@{->}[r ] & \omega s^{2m-1}\{p^r\}\ar@{->}[r ] & \omega s^{2m-1}\ar@{->}[r ] & s^{2m-1}\{p^{r+t}\}. } \end{aligned}\ ] ] by the naturality of @xmath892 , @xmath893 is homotopic to the sum of the maps @xmath894 ^ -{{\scalebox{.9}{${\widetilde{\varphi}_1}$ } } } & \omega p^{2m}(p^r)\ar@{->}[r]^-{{\scalebox{.9}{${\omega\rho^t}$}}}&\omega p^{2m}(p^{r+t})}\\ & \xymatrix@c=35pt{\sigma x\ar@{->}[r]^-{{\scalebox{.9}{${\widetilde{\varphi}_2}$ } } } & \omega s^{2m-1}\ar@{->}[r]^-{{\scalebox{.9}{${\omega\iota_{2m-1}}$}}}&\omega p^{2m}(p^{r+t}).}\end{aligned}\ ] ] thus @xmath895 which is homotopic to @xmath896 . at this point
we will examine relative whitehead products in congruence homotopy .
this will be applied to determining the congruence homotopy classes of the obstruction in theorem [ theor5.9 ] .
we define a @xmath292 module @xmath897 by @xmath898 [ prop5.16 ] the relative whitehead product induces a homomorphism : @xmath899 which commutes with the hurewicz homomorphism ; i.e. , the diagram @xmath900\ar@{->}[d]&e\pi_{m+k}(e;z / p^r)\ar@{->}[d]\\ h_m(\omega b;z / p^r)\otimes h_k(e;z / p^r)\ar@{->}[r]^-<<<<{a_*}&h_{m+k}(e;z / p^r ) } \ ] ] commutes where @xmath594 is induced by the principal action . to see that the pairing is well defined , it suffices to show that if @xmath901 , then @xmath902_r\sim*$ ] .
recall that @xmath558_r$ ] is given in [ prop3.16 ] by the composition : @xmath903 ^ -{{\scalebox{.9}{${\delta}$ } } } & p^m\wedge p^k\ar@{->}[r]^-{{\scalebox{.9}{${\theta}$ } } } & p^{m}\ltimes p^k\ar@{->}[r]^-<<<<{{\scalebox{.9}{${\widetilde{\alpha}\ltimes
\delta}$ } } } & \omega b\ltimes e\ar@{->}[r]^-{{\scalebox{.9}{${\gamma'}$ } } } & e. } \end{gathered}\ ] ] however , since there is a natural homeomorphism @xmath904 @xmath902_r$ ] factors through the map @xmath905 ^ -{{\scalebox{.9}{${\widetilde{\alpha}\ltimes \sigma\delta}$}}}&\omega b\ltimes \sigma e.}\ ] ] this map is null homotopic since it factors through @xmath906 up to homotopy .
clearly the hurewicz map factors through congruence homotopy and the homology calculation follows from [ cor3.18 ] .
[ def5.17 ] @xmath907 is the graded symmetric algebra generated by @xmath897 .
[ theor5.18 ] the relative whitehead product induces the structure of a graded differential @xmath907 module on @xmath908 . by [ prop5.16
] , there is an action of @xmath909 on @xmath908 and hence an action of the tensor algebra .
it suffices to show that @xmath910_r]_r\equiv ( -1)^{mn}[\beta,[\alpha,\gamma]_r]_r\ ] ] where @xmath911 and @xmath912 .
however we have @xmath913_r]_r&=[\alpha,[\beta,\pi_*(\gamma)]]_{\times}\\ & = [ [ \alpha,\beta],\pi_*(\gamma)]_{\times}+(-1)^{mn}[\beta,[\alpha,\pi_*(\gamma)]]_{\times}\end{aligned}\ ] ] by [ prop3.30](c ) , ( d ) and [ prop3.31](c ) . but
@xmath311=\pi_*[\alpha,\beta]_{\times}$ ] by [ prop3.30](a ) , so we have @xmath914,\pi_*(\gamma)]_{\times}&=[\pi_*([\alpha,\beta]_{\times}),\pi_*(\gamma)]_{\times}\\ & = [ [ \alpha,\beta]_{\times},\gamma]\end{aligned}\ ] ] by [ prop3.30](b ) . since @xmath915_{\times},\gamma]$ ] is a whitehead product of classes in @xmath916 , @xmath915_{\times},\gamma]\equiv 0 $ ]
. thus @xmath917 { } [ \alpha,[\beta,\gamma]_r]_r&\equiv(-1)^{mn}[\beta,[\alpha,\pi_*(\gamma)]]_{\times}\\ & = ( -1)^{mn}[\beta,[\alpha,\gamma]_r]_r .
\makebox[84pt ] { } \rlap{\qed}\end{aligned}\ ] ] since the action of @xmath907 is associative , we will also use the notation @xmath918_r\ ] ] to simplify the notation and distinguish this action from composition .
[ prop5.19 ] suppose @xmath919 , @xmath383 and @xmath920
. then @xmath921 .
suppose @xmath922 , @xmath923 and @xmath924 .
then there is a splitting @xmath925 such that @xmath926_r\equiv[\alpha,\delta]_r(\sigma^m\xi)\vee[\beta(\alpha),\delta]_r(\xi')\ ] ] where @xmath927 .
these relative whitehead products are not homotopic in general unless @xmath316 is a co-@xmath0 map .
using [ prop3.16 ] , we construct the following diagram : @xmath928^{\theta}&p\ltimes h\ar@{->}[r]^{\widetilde{\alpha}\ltimes\delta}&\omega b\ltimes e\ar@{->}[dr]^{\gamma'}\ar@{=}[dd]&\\ & & & e\\ p\wedge h'\ar@{->}[uu]^{1\wedge\xi}\ar@{->}[r]^{\theta } & p\ltimes h'\ar@{->}[r]^{\widetilde{\alpha}\ltimes \delta\xi } \ar@{->}[uu]^{1\ltimes\xi}&\omega b\ltimes e\ar@{->}[ur]_{\gamma ' } & } \ ] ] the two right hand regions are commutative .
we claim that the left hand region commutes up to congruence .
since @xmath929 is a co-@xmath0 space , there is a homotopy equivalence @xmath930 given by the sum of the map pinching @xmath0 to a point and the projection onto @xmath0 . since the composition @xmath931^{\theta}&p\ltimes h\ar@{->}[r]&p\wedge h } \ ] ] is a homotopy equivalence while the composition @xmath931^{\theta}&p\ltimes h\ar@{->}[r]&h } \ ] ] is null homotopic , both compositions in the left hand square become homotopic when composed with the map @xmath932 ^ ->>>>{\simeq}&p\wedge h\vee h\ar@{->}[r]&(p\wedge h)\times h. } \ ] ] the result then follows from [ prop5.12 ] . for part @xmath933 , we split @xmath934 by the composition @xmath935 ^ ->>>>>>>{\delta\vee \delta}&p^m\wedge p^k\vee p^{m-1}\wedge p^k\ar@{->}[r]^-<<<<<<<{1\vee \beta\wedge 1}&p^m\wedge p^k } \ ] ] ( compare to [ prop3.25 ] ) @xmath936_r$ ] is given by precomposing @xmath937 with @xmath553 . the composition @xmath938 ^ ->>>>{\delta}&p^m\wedge p^{\ell } \ar@{->}[r]^->>>>>{1\wedge
\xi}&p^m\wedge p^k\simeq p^{m+k}\vee p^{m+k-1 } } \ ] ] is congruent to a map with components @xmath939 and @xmath940 .
the result then follows from the splitting . in case @xmath569 , we now apply the @xmath941 module structure to the study of the congruence classes of the obstructions in theorem [ theor5.9 ]
. we will actually only consider the subalgebra of @xmath941 generated by @xmath942 and @xmath943 .
these elements generate a subalgebra @xmath944\otimes\wedge(\mu)\subset a_*(d_k)\ ] ] and @xmath945 is a module over this algebra .
we define classes @xmath946_{\times}\\ \overline{b(0)}&=[\mu,\mu]_{\times}\end{aligned}\ ] ] for @xmath751 .
[ theor5.20 ] in case @xmath569 , the collection of congruence classes of the set of obstructions listed in theorem [ theor5.9 ] is spanned , as a module over @xmath947\otimes\wedge(\mu)$ ] by the classes @xmath948 and @xmath949 of weight @xmath808 for @xmath950 , where the weight of @xmath948 and @xmath949 are both one , except when @xmath951 where the weight is two .
we first consider internal @xmath0-space based whitehead products of weight @xmath952 .
recall that by [ prop3.30](b ) @xmath953_{\times}=[\gamma,\delta]$ ] which is null congruent , so we need only consider weight @xmath952 products @xmath954_{\times}$ ] in which at least one of @xmath955 is not in the image of @xmath678 . by [ prop3.31](a )
, we will assume that @xmath956 or @xmath53 .
this gives the following possibilities for weight @xmath952 .
@xmath957_{\times},\ [ \mu,\xi_k\overline{a(i)}]_{\times},\ [ \nu,\xi_k\overline{b(i)}]_{\times},\ [ \mu,\xi_k\overline{b(i)}]_{\times}\ ] ] for @xmath751 . applying [ prop3.30](e ) again
, we see that for @xmath958 the class of @xmath829_{\times}]_r\ ] ] is null congruent if @xmath959 is in the image ov @xmath678 .
thus each of @xmath960 must be either @xmath421 or @xmath53 , and this class is in the @xmath947\otimes\wedge(\mu)$ ] submodule generated by @xmath961_{\times}$ ] .
we will refer to the submodule generated by @xmath962 and @xmath963 as the level @xmath35 obstructions . in case
@xmath107 we have some simple relations : [ prop5.21 ] @xmath964 and @xmath965 .
consequently the submodule generated by @xmath966 and @xmath967 has a basis consisting of @xmath968 and @xmath969 for @xmath249 .
[ cols= " < " , ] + using [ prop3.31](a ) and ( c ) we get @xmath970_{\times}=[[\mu,\nu],\mu]_{\times } + [ \nu,[\mu,\mu]]_{\times}=-[\mu,[\nu,\mu]]_{\times } + [ \nu,[\mu,\mu]]_{\times},\ ] ] so @xmath971_{\times}=[\nu,[\mu , u]]_{\times}\equiv \nu\cdot\nobreak[\mu,\mu]_{\times}$ ] .
because of these relations we define is the adjoint of the similarly named class in @xcite . ]
@xmath972_{\times}$ ] and @xmath973_{\times}$ ]
. then @xmath974 and @xmath975 .
[ prop5.22 ] in case @xmath569 , the level @xmath186 congruence classes are generated by @xmath976 and @xmath977 for @xmath808 with the relations @xmath978 and @xmath979 .
furthermore @xmath980 and @xmath981 .
@xmath982_{\times}\equiv\nu^{k-1}\cdot y_2=\nu\cdot y_k$ ] .
@xmath983_{\times}\equiv 0 $ ] by [ prop5.21 ] .
these relations imply that the @xmath984 and @xmath985 are linear generators .
we will see in section [ subsec6.1 ] that they are actually linearly independent .
@xmath986_{\times}\equiv(k-2)\mu\cdot x_{k-1}+2y_k\equiv ky_k$ ] .
@xmath987_{\times}\equiv 0 $ ] .
in this chapter we assemble the material in the previous chapters and construct a proof of theorem a. in [ subsec6.1 ] we introduce the controlled extension theorem and apply this in case @xmath107 . this case is simpler since @xmath988 , and serves as a model for the more complicated case when @xmath80 .
we recall a space @xmath109 ( called @xmath167 in @xcite ) and construct @xmath989 over the skeleta of @xmath109 as in @xcite .
however we choose @xmath989 so that it annihilates @xmath990 and @xmath991 for each @xmath992 , and conclude that @xmath993 is null homotopic .
in [ subsec6.2 ] we establish an inductive hypothesis and derive important properties from the inductive assumption . in particular , we construct @xmath112 and calculate its homology .
this allows for a secondary induction over the skeleta of @xmath112 . in [ subsec6.3 ]
the calculations are made to construct @xmath994 so that it annihilates the level @xmath35 obstructions . at this point
it is necessary to show that the level @xmath111 obstructions do not reappear .
this requires a careful analysis of the congruence class of the level @xmath111 obstructions .
theorem [ theor6.28 ] and corollary [ cor6.37 ] provide the necessary decompositions , and the issue is resolved by [ lem6.40 ] .
the induction is completed by [ theor6.41 ] and [ theor6.43 ] in this section we will introduce the controlled extension theorem and apply it to the simplest case : the construction of a retraction map @xmath995 such that the composition @xmath996 ^ -{\gamma_0 } & e_0\ar@{->}[r]^-{\nu_0 } & bw_n } \ ] ] is null homotopic .
this case is considerably simpler than the case @xmath80 and will serve as a model for the later cases .
the controlled extension theorem is an enhancement of the extension theorem ( [ theor2.5 ] ) , and our construction of @xmath989 is a controlled version of the construction of @xmath997 of section 3 of @xcite . [ theor6.1 ] suppose that all spaces are localized at @xmath176 and we have a diagram of principal fibrations induced by a map @xmath998 : @xmath999\ar@{->}[d ] & \omega x\ar@{->}[d]\\ e_0\ar@{->}[r]\ar@{->}[d]&e\ar@{->}[d]_{\pi}\\ b\ar@{->}[r]&b\cup_{\theta}e^m . } \ ] ] suppose @xmath1000 and we are given a map @xmath1001 with @xmath1002 such that @xmath1003 induces an isomorphism in @xmath526 cohomology in dimension @xmath29 .
suppose also that we are given a map @xmath1004 .
then : 1 . there is an extension of @xmath1005 to @xmath1006 .
2 . suppose also that we are given a map @xmath106 and a subspace @xmath1007 such that the composition @xmath1008&p\ar@{->}[r]^-{u}&e\ar@{->}[r]^-{\gamma'}&bw_n } \ ] ] is null homotopic and such that the quotient map @xmath1009 factors up to homotopy @xmath656 ^ -{u}&e\ar@{->}[r]^-{q'}&e
/ e_0\ar@{->}[r]^-{\xi}&p / p_0 } \ ] ] for some map @xmath316 .
then there is an extension @xmath1010 of @xmath1005 such that @xmath1011 .
clearly the map @xmath1012 induces an isomorphism in homology , so the existence of @xmath1013 follows from the extension theorem ( [ theor2.5 ] ) . to prove part ( b ) , we suppose @xmath1013 is given and we construct @xmath1010 as the composition : @xmath1014 { } \xymatrix@c=30pt { e\ar@{->}[r]^-{\delta}&e\!{}\times{}\ ! e / e_0\ar@{->}[r]^-{\gamma'\!{}\times{}\!\xi}&bw_n\!{}\times{}\!p / p_0\ar@{->}[r]^-{1\!{}\times{}\ ! \eta}&bw_n\!{}\times{}\!bw_n\ar@{->}[r]^-{\div}&bw_n } \makebox[0pt]{}\ ] ] where @xmath1015 is defined by the null homotopy of @xmath1016 and @xmath1017 is the @xmath0-space division map . clearly @xmath1018 . to study @xmath1019 , consider the diagram @xmath1020 ^ -{\delta } & e\!{}\times e / e_0\ar@{->}[r]^-{\gamma'\!{}\times{}\ ! \xi } & bw_n\!{}\times{}\ !
p / p_0\ar@{->}[r]^-{1\!{}\times{}\ ! \eta } & bw_n\!{}\times{}\ !
bw_n\ar@{->}[r]^-{\div}\ar@{=}[d ] & bw_n\ar@{=}[d]\\ p\ar@{->}[r]^-{\delta}\ar@{->}[u]^{u } & p\!{}\times{}\ ! p\ar@{->}[u]^{u\!{}\times{}\ ! q'u}\ar@{->}[r]^-{\eta q\!{}\times{}\
! q } & bw_n\!{}\times{}\ !
p / p_0\ar@{->}[u]\ar@{->}[r]^-{1\!{}\times{}\!\eta } & bw_n\!{}\times{}\ ! bw_n\ar@{->}[r]^-{\div } &
bw_n } \ ] ] where the lower composition of the first 3 maps factors through the diagonal map of @xmath43 , so the lower composition is null homotopic .
now we will apply this in the case @xmath107 .
recall the spaces ( [ eq5.2 ] ) . in this case @xmath1021 and
@xmath1022 : @xmath1023 & e_0\ar@{->}[r]^-{\tau_0}\ar@{->}[d ] & f_0\ar@{->}[r]\ar@{->}[d ] & \omega s^{2n+1}\ar@{->}[d]\\ & p^{2n+1}\ar@{=}[r]\ar@{->}[d ] & p^{2n+1}\ar@{->}[r]\ar@{->}[d ] & ps^{2n+1}\ar@{->}[d]\\ & s^{2n+1}\{p^r\}\ar@{->}[r ] & s^{2n+1}\ar@{->}[r]^-{p^r } & s^{2n+1}. } \ ] ] these spaces were introduced in @xcite where @xmath1024 is called @xmath1025 and @xmath109 is called @xmath1026 .
[ prop6.2 ] @xmath1027 this is immediate from consideration of the cohomology serre spectral sequence of the middle fibration , which is induced from the path space fibration on the right .
all differentials are controlled by the path space fibration .
we filter @xmath109 by setting @xmath1028 to be the @xmath1029 skeleton of @xmath109 and define @xmath1030 to be the pullback over @xmath1028 @xmath1031\ar@{->}[d ] & \omega^2s^{2n+1}\ar@{->}[d]\\ e_0(m)\ar@{->}[r]\ar@{->}[d ] & e_0\ar@{->}[d]_{\eta_0}\\ f_0(m)\ar@{->}[r]&f_0 } \ ] ] since @xmath1032 , this fibration in case @xmath1033 is the fibration which defines @xmath43 ( see [ prop2.3 ] ) @xmath3&s^{4n-1}\times bw_n\ar@{->}[r]&s^{2n}\ar@{->}[r]&\omega s^{2n+1}. } \ ] ] we consequently define @xmath1034 by retracting onto @xmath43 . clearly @xmath1035 is null homotopic since @xmath1036 and @xmath43 is @xmath1037 connected .
we will use [ theor6.1 ] to construct @xmath1038 such that @xmath1039 and @xmath1040 are null homotopic for @xmath1041 .
[ prop6.3 ] for each @xmath992 , there is a retraction @xmath1042 extending @xmath1043 such that @xmath1044 annihilates the classes @xmath990 and @xmath1045 from [ prop5.22 ] .
the proof of this result will depend on two lemmas .
[ lem6.4 ] the composition @xmath1046
^ -{x_j } & e_0\ar@{->}[r]^-{\eta_0 } & f_0 } \ ] ] induces a cohomology epimorphism . to study @xmath1047
, we use the principal fibration : @xmath1048 & f_0\ar@{->}[r ] & p^{2n+1 } } \ ] ] clearly @xmath1049 lifts to a map @xmath1050 which induces a cohomology epimorphism .
then @xmath1051_r$ ] for each @xmath808 so we can apply [ cor3.18 ] to evaluate @xmath976 in cohomology .
[ lem6.5 ] the composition @xmath1052 ^ -{y_{j+1 } } & e_0(j)\ar@{->}[r ] & e_0(j)/e_0(j-1 ) } \\ \xymatrix { \omega^2s^{2n+1 } \ltimes s^{2nj } \ar@{->}[r]^-{\xi } & s^{2nj}\vee s^{2n(j+1)-1 } } \end{gathered}\ ] ] induces an integral cohomology epimorphism . since @xmath1053 , @xmath1054 by the clutching construction ( [ eq2.2 ] ) . from the homotopy commutative square @xmath1055\ar@{->}[d]&e_0(j)/e_0(j-1)\ar@{->}[d]\\ f_0(j)\ar@{->}[r]&s^{2nj } } \simeq s^{2nj}\ltimes \omega^2s^{2n+1}\ ] ] we see that the composition @xmath1056 ^ -{x_j}&e_0(j)\ar@{->}[r]&e_0(j)/e_0(j-1 ) } \\ \xymatrix { \simeq \omega^2s^{2n+1 } \ltimes s^{2nj } \ar@{->}[r]^-{\xi } & s^{2nj}\vee s^{2n(j+1)-1}}\end{gathered}\ ] ] is an integral cohomology epimorphism .
since the action of @xmath5 on @xmath1057 corresponds with the action of @xmath5 on @xmath1058 we can apply [ cor3.18 ] to see that the composition in question @xmath1059 & \omega^2s^{2n+1 } \ltimes s^{2nj } } \ ] ] induces an epimorphism in integral cohomology .
we apply [ theor6.1 ] to the diagram @xmath1031\ar@{->}[d ] & \omega^2s^{2n+1}\ar@{->}[d]\\ e_0(m-1)\ar@{->}[r]\ar@{->}[d ] & e_0(m)\ar@{->}[d]\\ f_0(m-1)\ar@{->}[r ] & f_0(m)&{\mbox{\hspace*{-31pt}$=f_0(m-1)\cup_{\theta}e^{2mn}$ } } } \ ] ] where @xmath159 is the attaching map of the @xmath1029 cell .
let @xmath1060 .
choose an extension @xmath1013 of @xmath1061 .
let @xmath1062 and @xmath1063 ^ -{{\scalebox{.9}{${x_m\vee y_{m+1}}$}}}&e_0(m);}\ ] ] let @xmath1064 .
the composition @xmath1065&s^{2mn-1}\ar@{->}[r]&p^{2mn}\ar@{->}[r]^-{x_m}&e_0(m ) } \ ] ] is @xmath1066 by [ prop5.22 ] . since @xmath1067 is null homotopic , the composition @xmath1065 ^ -{\beta}&p^{2mn}\ar@{->}[r]^-{x_m}&e_0(m)\ar@{->}[r]^-{\gamma'}&bw_n } \ ] ] is null homotopic . @xmath133
factors : @xmath1068 , so the composition @xmath1069&p^{2mn}\ar@{->}[r]^-{x_m}&e_0(m)\ar@{->}[r]^-{\gamma'}&bw_n } \ ] ] is divisible by @xmath4 .
however @xmath1070 , so this composition is null homotopic . similarly , since @xmath1071 , the composition @xmath1072 & p^{2(m+1)n-1}\ar@{->}[r]^-{y_{m+1 } } & e_0(m)\ar@{->}[r]^-{\gamma ' } & bw_n } \ ] ] is null homotopic . thus the composition @xmath1073&p\ar@{->}[r]^-{x_m\vee y_{m+1 } } &
e_0(m)\ar@{->}[r]^-{\gamma'}&bw_n } \ ] ] is null homotopic .
however @xmath1074 ^ -{x_m\!{}\vee{}\!y_{m+1 } } & e_0(m)\ar@{->}[r ] & } \\ \xymatrix { e_0(m)/e_0(m-1 ) \cong \omega^2s^{2n+1 } \!{}\ltimes{}\ ! s^{2mn } } \end{gathered}\ ] ] induces an integral cohomology epimorphism by [ lem6.4 ] and [ lem6.5 ] .
let @xmath316 be the composition @xmath1075&s^{2mn}\vee s^{2(m+1)n-1 } } \ ] ] where the last map is obtained by evaluation on the double loop space .
clearly the conditions of [ theor6.1 ] are satisfied , so we can choose an extension @xmath1076 of @xmath1043 such that @xmath1077 and @xmath1078 .
this completes the induction .
[ cor6.6 ] there is a retraction @xmath1079 such that the composition @xmath1080
^ -{\gamma_0}&e_0\ar@{->}[r]^-{\nu_0}&bw_n } \ ] ] is null homotopic . by [ prop6.3 ]
@xmath1081 and @xmath1082 for @xmath992 . by [ theor5.9 ] and [ prop5.22 ] , @xmath1083&e_0\ar@{->}[r]^-{{\scalebox{.9}{${\nu_0}$}}}&bw_n}\ ] ] is null homotopic for all @xmath808 .
the conclusion follows from [ prop5.7 ] . at this point
we present the inductive hypothesis . as pointed out in section [ subsec2.4 ] , this will be stronger than proposition [ prop2.12 ] given there .
this strengthening will be a factorization of the map @xmath83 through the map @xmath1084 . in case @xmath107 , @xmath1085 and @xmath1086 is the identity map , so this alteration only applies when @xmath80 .
[ theor6.7 ] there is a map @xmath1087 such that the composition @xmath1088 ^ -<<<{\gamma_k } & e_k\ar@{->}[r]^{\tau_k}&j_k\ar@{->}[r]^->>>>{\gamma_k } & bw_n } \index{gammak@$\gamma_{k}$|lb } \index{gammak@$\gamma_{k}$|lb}\ ] ] is null homotopic and such that the compositions @xmath1089&e_{k-1}\ar@{->}[r]^{e_k}\index{$e_k$|lb}&e_k\ar@{->}[r]^{\tau_k}&j_k\ar@{->}[r]^->>>>{\gamma_k}&bw_n } \\ \xymatrix { e^{2np^k-1}_{k-1}\ar@{->}[r]&e_{k-1}\ar@{->}[r]^{\tau_{k-1}}&j_{k-1}\ar@{->}[r]^->>>>{\gamma_{k-1}}&bw_n } \end{gathered}\ ] ] are homotopic , where @xmath1090 as constructed in [ cor6.6 ] . by [ prop2.3 ] , the composition @xmath3&e_0(1)\ar@{->}[r]^{\nu_0(1)}&bw_n } \ ] ] is homotopic to @xmath53 .
consequently [ theor6.7 ] implies [ prop2.12 ] in case @xmath107 .
we will assume that we have constructed @xmath1091 for @xmath648 , having @xmath1091 we obtain @xmath1092 and construct @xmath572 , @xmath573 and @xmath96 as in [ theor4.4 ] with @xmath631 .
we then define @xmath103 and @xmath102 and proceed to construct @xmath105 .
the construction is completed with [ theor6.43 ] .
the procedure in section [ subsec6.1 ] is a model for the inductive step . to proceed
, we will first need to prove : @xmath1093 we will then use an inductive procedure over the skeleta of @xmath112 as in section [ subsec6.1 ] .
the proof of [ eq6.8 ] will be by induction on @xmath35 .
the case @xmath107 is [ prop6.2 ] . at this point
we will assume [ eq6.8 ] in case @xmath111 .
[ prop6.9 ] let @xmath1094 be the fiber of @xmath1095 .
then we have a homotopy commutative diagram of vertical fibration sequences @xmath1096\ar@{->}[d ] & t\ar@{->}[r]\ar@{->}[d ] & \omega s^{2n+1}\ar@{->}[d]\\ r_{k-1}\ar@{->}[r]\ar@{->}[d ] & w_{k-1}\ar@{->}[r]\ar@{->}[d ] & f_{k-1}\ar@{->}[d]^-{\sigma_{k-1}}\\ g_{k-1}\ar@{->}[r ] & d_{k-1}\ar@{=}[r ] & d_{k-1 } } \ ] ] and two diagrams of fibration sequences @xmath1097\ar@{->}[d ] & \omega^2s^{2n+1}\ar@{->}[r]\ar@{->}[d ] & bw_n\ar@{=}[d]\\ w_{k-1}\ar@{->}[r]\ar@{->}[d ] & j_{k-1}\ar@{->}[r]^-{\gamma_{k-1}}\ar@{->}[d ] & bw_n\\ f_{k-1}\ar@{=}[r ] & f_{k-1 } } \\ \mbox{}\\ \mbox{}\\ \xymatrix { s^{2n-1}\ar@{->}[r]\ar@{->}[d ] & t\ar@{->}[r]\ar@{->}[d ] & \omega s^{2n+1}\\ w_{k-1}\ar@{=}[r]\ar@{->}[d ] & w_{k-1}\ar@{->}[d]&\\ f_{k-1}\ar@{->}[r ] & d_{k-1}\ar@{->}[r]&s^{2n+1 } } \end{gathered}\ ] ] we define @xmath110 to be the composition @xmath245 ^ -{\tau_{k-1 } } & j_{k-1}\ar@{->}[r]^-{\gamma_{k-1 } } & bw_n . } \ ] ] from this it follows that we have a commutative diagram of fibration sequences @xmath1098\ar@{->}[d ] & e_{k-1}\ar@{->}[r]^-{\nu_{k-1}}\ar@{->}[d]_{\tau_{k-1 } } & bw_n\ar@{=}[d]\\ w_{k-1}\ar@{->}[r ] & j_{k-1}\ar@{->}[r]^-{\gamma_{k-1 } } & bw_n . } \ ] ] consequently the square @xmath1098\ar@{->}[d ] & w_{k-1}\ar@{->}[d]\\ g_{k-1}\ar@{->}[r ] & d_{k-1 } } \ ] ] is the composition of two pullback squares @xmath1098\ar@{->}[d ] & e_{k-1}\ar@{->}[r]^-{\pi_{k-1}}\ar@{->}[d]_{\tau_{k-1 } } & g_{k-1}\ar@{->}[d]\\ w_{k-1}\ar@{->}[r ] & j_{k-1}\ar@{->}[r]^-{\xi_{k-1 } } & d_{k-1 } , } \ ] ] so it is a pullback and the first diagram commutes up to homotopy .
the second diagram follows from the definition of @xmath1094 and the third is a combination of the first two .
[ prop6.10 ] @xmath1099 . extending the third diagram of [ prop6.9 ] to the left yields a diagram @xmath1023\ar@{->}[d ] & s^{2n-1}\ar@{->}[d]\\ { * } \ar@{->}[r]\ar@{->}[d ] & w_{k-1}\ar@{->}[d]\\ \omega s^{2n+1}\ar@{->}[r ] & f_{k-1 } ; } \ ] ] both of the horizontal maps have degree @xmath4 in dimension @xmath10 , so @xmath1094 is @xmath1100 connected and the map @xmath1101 is null homotopic . from this
it follows that @xmath1102 .
[ prop6.11 ] the homomorphism @xmath1103 is onto and @xmath1104 consider the serre spectral sequence for the @xmath38-local homology of the fibration @xmath1048
^ -{\delta_{k-1 } } & f_{k-1}\ar@{->}[r]^-{\sigma_{k-1 } } & d_{k-1 } } \ ] ] since @xmath1105 is only nonzero when @xmath38 and @xmath179 are divisible by @xmath10 , @xmath1106 .
we assume the result ( [ eq6.8 ] ) for the case @xmath111 by induction .
since @xmath1107 has finite order when @xmath1108 and @xmath1109 is free , all extensions are nontrivial .
let @xmath1110 be the generator dual to the @xmath1111 power of a chosen fixed generator in @xmath1112 , so @xmath1113 using the nontrivial extensions in the serre spectral sequence , we can choose generators @xmath1114 so that @xmath1115 since @xmath1116 is a monomorphism , it is easy to check that @xmath1117 it now follows from the @xmath38-local cohomology serre spectral sequence for the fibration @xmath44&w_{k-1}\ar@{->}[r]&f_{k-1 } } \ ] ] that @xmath1118 from this we can read off the cohomology of @xmath1094 . [ prop6.12 ] the homomorphism @xmath1119&h^{2np^k}(t;z_{(p ) } ) } \ ] ] is nontrivial of order @xmath38 . from [ prop6.9 ]
we have a homotopy commutative square @xmath609\ar@{->}[d]&\omega s^{2n+1}\ar@{->}[d]_{\delta_{k-1}}\\ w_{k-1}\ar@{->}[r]&f_{k-1 } } \ ] ] to which we apply cohomology @xmath1120\\ h^{2np^k}(w_{k-1})\ar@{->}[u ] & h^{2np^k}(f_{k-1})\ar@{->}[u]^{\delta^*_{k-1 } } \ar@{->}[l ] } \ ] ] which we evaluate @xmath1121\\ z/ p^{r+k}\ar@{->}[u ] & z_{(p)}\ar@{->}[l]\ar@{->}[u]^{p^{r+k-1 } } } \ ] ] where the two horizontal arrows are epimorphisms .
the result follows .
[ prop6.13 ] the map @xmath1122 extends to a map @xmath1123 such that the composition @xmath1124 & t/ t^{2np^k-2}\ar@{->}[r ] & r_{k-1}\ar@{->}[r ] & r_k } \ ] ] is null homotopic . since the fibration @xmath609&r_{k-1}\ar@{->}[r]&g_{k-1 } } \ ] ] is induced from the fibration over @xmath72 , we have a homotopy commutative square @xmath1125\ar@{->}[d ] & r_{k-1}\ar@{->}[d]\\ t/\omega g_{k}\ar@{->}[r ] & r_k . } \ ] ] by [ theor2.14](a ) the inclusion @xmath1126&t}$ ] factors through @xmath1127 , this gives a homotopy commutative square @xmath1128\ar@{->}[d ] & r_{k-1}\ar@{->}[d]\\ t / t^{2np^{k+1}-2}\ar@{->}[r ] & r_k } \ ] ] the result follows by restriction to @xmath1129 .
[ prop6.14 ] let @xmath1130 be the composition of the first two maps in [ prop6.13 ] .
then the composition @xmath650
^ -{\widetilde{\alpha}_k } & r_{k-1}\ar@{->}[r ] & w_{k-1 } } \ ] ] is nonzero in @xmath38 local cohomology .
this follows from [ prop6.12 ] using the diagram @xmath650\ar@{->}[dr]^-{\widetilde{\alpha}_k}&t / t^{2np^k-2 } \ar@{->}[d]&t\ar@{->}[d ] \ar@{->}[l ] \\ & r_{k-1}\ar@{->}[r]&w_{k-1 } } \ ] ] where the three spaces on the top have isomorphic cohomology in dimension @xmath246 .
we assume the result for @xmath1131 by induction .
since @xmath112 is the total space of a principal fibration over @xmath1132 whose restriction to @xmath773 is @xmath1131 , we have by [ eq2.2 ] @xmath1133 and consequently we have a short exact sequence @xmath18&h_{2nm}(f_{k-1};z_{(p)})\ar@{->}[r]&h_{2nm}(f_k;z_{(p)})\ar@{->}[r]&z / p\ar@{->}[r]&0 } \ ] ] for @xmath1134 , while @xmath1135 for @xmath1136 .
we will prove that the extension is nontrivial .
it suffices to show that @xmath1137 since the module action of @xmath1138 on both @xmath1109 and @xmath1139 implies the result for all @xmath1140 .
if this failed we would conclude that @xmath1141 .
this would imply that the homomorphism @xmath1142&h^{2np^k}(f_{k-1};z_{(p ) } ) } \ ] ] is onto .
we will show that this is impossible .
suppose then that this homomorphism is onto and consider the homotopy commutative diagram @xmath1143_{\delta_k}\\ p^{2np^k}(p^{r+k})\ar@{->}[rrrru]^-{l}\ar@{->}[r]^->>>>{\widetilde{\alpha}_k } & r_{k-1}\ar@{->}[r]\ar@{->}[d ] & w_{k-1}\ar@{->}[r]\ar@{->}[d ] & f_{k-1}\ar@{->}[r]\ar@{->}[d]_{\sigma_{k-1 } } & f_k\ar@{->}[d]_{\sigma_k}\\ & g_{k-1}\ar@{->}[r ] & d_{k-1}\ar@{=}[r ] & d_{k-1}\ar@{->}[r ] & d_k . } \ ] ] the map @xmath1144 exists since the composition into @xmath103 factors through the composition @xmath1098&r_k\ar@{->}[r]&g_k\ar@{->}[r]&d_k ; } \ ] ] thus this composition is null homotopic by [ prop6.13 ] .
now if @xmath1145 & h^{2np^k}(f_{k-1};z_{(p)})}\ ] ] is onto then the entire horizontal composition @xmath1146&h^{2np^k}(p^{2np^k}(p^{r+k});z_{(p ) } ) } \ ] ] is nonzero by [ prop6.11 ] and [ prop6.14 ] . but @xmath1147 factors : @xmath1048 ^ -{\delta_{k-1}}&f_{k-1}\ar@{->}[r]&f_k } \ ] ] and @xmath1148 is divisible by @xmath1149 .
consequently the image of @xmath1150 is divisible by @xmath1149 ; since @xmath1151 is nonzero and @xmath1152 we conclude that @xmath1153 is onto .
this is impossible for then the composition @xmath650
^ -{l}&\omega s^{2n+1}\ar@{->}[r]^-{h_{p^{k-1}}}&\omega s^{2np^{k-1}+1 } } \ ] ] would be onto , where @xmath1154 is the james hopf invariant .
but there is never a map @xmath1155 which is onto in cohomology when @xmath1156 since the adjoint @xmath1157&\omega^2 s^{2m+1 } } \ ] ] would also be onto .
such a map would not commute with the bockstein .
consequently the extension is nontrivial and the cohomology is free .
[ cor6.15 ] the induced homomorphism @xmath1158&h_{2ni}(f_k;z_{(p ) } ) } \ ] ] is an isomorphism when @xmath1159 and has degree @xmath38 if @xmath1160 .
furthermore the principal action defines an isomorphism @xmath1161 & h_{2np^k+2ni}(f_k;z_{(p ) } ) } \\*[-18pt ] \makebox[\textwidth]{\hfill\qed}\end{gathered}\ ] ] this completes the first task of this section .
our second task will be to give a sharper understanding of the spaces @xmath639 and , in particular , @xmath1094 . recall that by [ theor2.14](a ) , @xmath653 is a retract of @xmath1162 .
consequently we have a sequence of induced fibrations from [ prop6.9 ] @xmath1096\ar@{->}[d]&t\ar@{=}[r]\ar@{->}[d]&t\ar@{->}[d]\\
r_{k-1}\ar@{->}[r]\ar@{->}[d]&q_{k-1}\ar@{->}[r]\ar@{->}[d]&r_{k-1}\ar@{->}[d]\\ g_{k-1}\ar@{->}[r]&\sigma t^{2np^k-2}\ar@{->}[r]&g_{k-1 } } \ ] ] from which we see that @xmath639 is a retract of @xmath1163 . using the clutching construction ( [ prop2.1 ] )
, we see that @xmath1163 is homotopy equivalent to a pushout with @xmath1164 and @xmath1165 @xmath609&q_{k-1}\\ t \times t^{2np^k-2}\ar@{->}[r]\ar@{->}[u]^{\varphi}&t\times ct^{2np^k-2}\ar@{->}[u ] } \ ] ] where @xmath65 is the restriction of the action map : @xmath1166&t\times\omega g_{k-1}\ar@{->}[r]^-{a}&t . }
\ ] ] since the inclusion @xmath1167 factors through @xmath1127 , the composition @xmath1168&t\ar@{->}[r]&q_{k-1 } } \ ] ] is null homotopic ; by applying [ eq2.2 ] we have an equivalence @xmath1169 .
however , from the pushout diagram , we have @xmath1170 restricting to the @xmath247 skeleton , we get @xmath1171 so @xmath1172 .
now @xmath1173 is a wedge of moore spaces by [ theor2.14](d ) in section [ subsec2.3 ] and only has cells in dimensions congruent to @xmath1174 , @xmath186 , or @xmath1175 @xmath1176 .
consequently @xmath1177 is a wedge of moore spaces , and the largest exponent is the same as the largest exponent in @xmath1162 , which is @xmath1149 .
since @xmath639 is a retract of @xmath1163 , we have proved [ prop6.16 ] @xmath1178 is a wedge of @xmath834 moore spaces @xmath1179 for @xmath1180 .
there are no spheres in this wedge as there are no moore spaces in @xmath1181 of dimension @xmath668 .
[ prop6.17 ] the homomorphism in integral homology @xmath1182&h_i(w_{k-1};z_{(p ) } ) } \ ] ] is onto for all @xmath218 and split for @xmath1183 . since @xmath773 is the mapping cone of the composition @xmath1184
^ -{c}&e_{k-1}\ar@{->}[r]^-{\pi_{k-1}}&g_{k-1 } , } \ ] ] we apply [ eq2.2 ] to the fibrations in [ prop6.9 ] @xmath1096\ar@{->}[d]&t\ar@{->}[d]\\ r_{k-1}\ar@{->}[r]\ar@{->}[d ] & w_{k-1}\ar@{->}[d]\\ g_{k-1}\ar@{->}[r ] & d_{k-1 } } \ ] ] and we can then describe @xmath1094 by a pushout diagram @xmath1185&w_{k-1}\\ t\times c_{k-1}\ar@{->}[u]\ar@{->}[r]&r_{k-1}.\ar@{->}[u ] } \ ] ] this leads to a long exact sequence @xmath1186 & \widetilde{h}_i(t\times c_{k-1};z_{(p)})\ar@{->}[r ] & \widetilde{h}_i(r_{k-1};z_{(p)})\otimes \widetilde{h}_i(t;z_{(p)})}\\ \xymatrix { \ar@{->}[r ] & \widetilde{h}_i(w_{k-1};z_{(p)})\ar@{->}[r ] & \widetilde{h}_{i-1}(t\times c_{k-1};z_{(p ) } ) . }
\end{gathered}\ ] ] we assert that the homomorphism @xmath1187 & \widetilde{h}_{i-1}(t\times c_{k-1};z_{(p ) } ) } \ ] ] is trivial . by [ prop6.11 ] , @xmath1188 is only nontrivial when @xmath1189 for some @xmath1190 . but @xmath1191 since there are no cells in these dimensions .
now since @xmath1192 is onto in homology , we conclude that @xmath1193&h_i(w_{k-1};z_{(p)})}$ ] is onto . to show that this is split when @xmath1183 , we note that since @xmath1194 is cyclic by [ prop6.11 ] , it suffices to show that the exponent of @xmath1195 is not larger than the exponent of @xmath1196 for @xmath1183 . by [ prop6.11 ]
, we have @xmath1197 but @xmath1198 when @xmath1199 .
however @xmath1200 { } ip^rh_{2ni-1}(\sigma t\wedge t;z_{(p)})&=0\\ \intertext{and } p^{r+s-1}h_{2np^s-1}(\sigma t\wedge t;z_{(p)})&=0 .
\makebox[92.5pt ] { } \rlap{\qed}\end{aligned}\ ] ] [ prop6.18 ] @xmath1201 is a wedge of moore spaces .
since @xmath1202 is split onto , we can find a moore space in the decomposition of @xmath1178 for each @xmath218 representing a given generator .
this constructs a subcomplex of @xmath1178 which is homotopy equivalent to @xmath1201 .
[ cor6.19 ] @xmath1203 where @xmath1204 in this section we perform the inductive step of constructing a retraction @xmath1087 for @xmath574 . as in the proof of [ prop6.3 ]
, we will apply [ theor6.1 ] to the fibration @xmath3&j_k\ar@{->}[r]&f_k } \ ] ] and do an induction over the cells of @xmath112 . at each stage in this secondary induction we will make choices to eliminate the obstructions from [ theor5.20 ] . we will construct a map @xmath1087 which will annihilate the level @xmath35 obstructions .
however , @xmath1095 is not homotopic to the composition @xmath1205&j_k\ar@{->}[r]^->>>>{\gamma_k}&bw_n } \ ] ] so we will need an extra argument to show that @xmath105 annihilates the obstructions of level less than @xmath35 .
this is accomplished by some general results ( [ theor6.28 ] and [ cor6.37 ] ) which decompose certain relative whitehead products .
this is applied in [ lem6.40 ] to control the obstructions of a lower level .
we presume that @xmath1095 has been constructed such that the composition @xmath1206
^ -{\gamma_{k-1 } } & e_{k-1}\ar@{->}[r]^-{\tau_{k-1 } } & j_{k-1}\ar@{->}[r]^-{\gamma_{k-1 } } & bw_n } \ ] ] is null homotopic .
this defines the fiber @xmath639 of @xmath1207 and we construct @xmath96 , @xmath572 and @xmath573 in accordance with [ theor4.4 ] , and @xmath103 , @xmath102 and @xmath112 as in [ eq5.2 ] .
we next construct a modification of [ theor4.4 ] in this context .
[ prop6.20 ] there is a homotopy commutative ladder of cofibration sequences : @xmath1208\ar@{<->}[dddd]_{= } & p^{2np^k}(p^{r+k})\ar@{->}[r]^-{\sigma}\ar@{->}[d]_{\beta_k } & p^{2np^k}(p^{r+k-1})\ar@{->}[r]\ar@{->}[d]_{a(k ) } & p^{2np^k+1}(p)\ar@{<->}[dddd]_{=}\\ & e_{k-1}\ar@{->}[d]_{\tau_{k-1 } } & e_k\ar@{->}[d]_{\tau_k}&\\ & j_{k-1}\ar@{->}[r]^-{\iota}\ar@{->}[d]_{\eta_{k-1 } } & j_k\ar@{->}[d]_{\eta_k}&\\ & f_{k-1}\ar@{->}[r]\ar@{->}[d]_{\sigma_{k-1 } } & f_k\ar@{->}[d]_{\sigma_k}&\\ p^{2np^k}(p)\ar@{->}[r ] & d_{k-1}\ar@{->}[r ] & d_k\ar@{->}[r ] & p^{2np^k+1}(p)\\ } \ ] ] the upper central square commutes up to homotopy by [ theor4.4 ] and [ prop5.3 ] and the lower central squares follow from ( [ eq5.2 ] ) . by a cohomology calculation , the right hand square commutes up to homotopy . for the left hand region , observe that the @xmath246 skeleton of the fiber of the inclusion of @xmath773 into @xmath103 is homotopy equivalent to @xmath1209 ; a standard argument with cofibration sequences shows that the left hand vertical map can be taken to be the identity .
[ cor6.21 ] the compositions @xmath1210 ^ -{a(k)}&e_k\ar@{->}[r]^-{\tau_k}&j_k\ar@{->}[r]^-{\eta_k}&f_k } \\ & \xymatrix { p^{2np^k}(p^{r+k-1})\ar@{->}[r]^-{\beta_k}&e_{k-1}\ar@{->}[r]^-{\tau_{k-1}}&j_{k-1}\ar@{->}[r]^-{\eta_{k-1}}&f_{k-1 } } \end{aligned}\ ] ] induce integral cohomology epimorphisms .
the first composition is handled by applying integral cohomology to the right hand region of [ prop6.20 ] .
for the second composition we consider the upper two parts of the middle region .
the map @xmath134 has degree @xmath38 in @xmath1211 as does the map @xmath1212 by [ cor6.15 ] .
since @xmath1213 , this is enough to imply the result .
[ prop6.22 ] @xmath1214 is a wedge of moore spaces . by [ prop6.18 ]
, @xmath1201 is a wedge of moore spaces . by [ prop6.11 ] , it suffices to show that @xmath96 factors through @xmath1094 : @xmath650 ^ -<<<{{\scalebox{.9}{${\beta_k}$ } } } & w^{2np^k}_{k-1 } } \ ] ] is an epimorphism .
but by [ cor6.21 ] , the composition @xmath650 ^ -{\beta_k}&w^{2np^k}_{k-1}\ar@{->}[r]&f_{k-1 } } \ ] ] induces an epimorphism in @xmath38-local cohomology .
we now filter @xmath112 by skeleta and apply [ eq6.8 ] . as in section [ subsec6.1 ] , let @xmath1215 be the @xmath1029 skeleton of @xmath112 , so @xmath1216 let @xmath1217 be the pullback of @xmath102 to @xmath1215 , so we have a map of principal fibrations @xmath1218\ar@{->}[d ] & \omega^2s^{2n+1}\ar@{->}[d]\\ j_k(m-1)\ar@{->}[r]\ar@{->}[d]&j_k(m)\ar@{->}[d]\\ f_k(m-1)\ar@{->}[r]&f_k(m ) } \end{split}\ ] ] and using the clutching construction ( [ prop2.1 ] ) we see that @xmath1219 the obstructions that we need to consider at level @xmath35 are the elements @xmath1220 and @xmath1221 for @xmath1222 where @xmath1223 and @xmath1224 generate @xmath1225 .
( see [ theor5.18 ] and [ theor5.20 ] ) .
[ prop6.24 ] the compositions @xmath1226
^ -{\nu^i\cdot\overline{a(k)}}&j_k\ar@{->}[r]^-{\eta_k}&f_k } \\ & \xymatrix@c=46pt { p^{2np^k+2ni-1}\ar@{->}[r]^-{\mu\nu^{i-1}\cdot\overline{a(k)}}&j_k(p^k+i-1)\ar@{->}[r]^-q & \omega^2s^{2n+1 } \ltimes s^{2(p^k+i-1)n } } \end{aligned}\ ] ] induce integral cohomology epimorphisms where @xmath179 is the quotient map
. the first composition is evaluated by [ cor6.21 ] when @xmath951 . in case @xmath1227 ,
we use induction on @xmath218 .
we apply [ prop3.11](d ) to the diagram @xmath1228 ^ -{\eta_k}\ar@{->}[d]&f_k\ar@{->}[d]\\ d_k\ar@{=}[r]\ar@{->}[d]_{\varphi'_k}&d_k\ar@{->}[d]\\ s^{2n+1}\{p^r\}\ar@{->}[r]&s^{2n+1 } } \ ] ] to see that @xmath1229_r$ ] .
the result then follows from [ cor3.18 ] and [ cor6.15 ] .
the second composition is evaluated by using [ cor3.18 ] again since @xmath1230_r$ ] .
[ cor6.25 ] the composition @xmath1231 ^ -{\nu^i\cdot\overline{a(j)}}&w_j\ar@{->}[r]&w_{k-1 } } \ ] ] induces an integral cohomology epimorphism when @xmath1232 and @xmath1233 ; likewise the composition @xmath1234
^ -{\mu\nu^i\cdot\overline{a(j ) } } & w_j\ar@{->}[r]&w_{k-1 } } \ ] ] is nonzero in @xmath526 cohomology in dimension @xmath1235 . by the induction hypothesis ,
@xmath1236 and @xmath1237 are in the kernel of @xmath1238 for @xmath1233 , so they factor through @xmath1239 .
we then construct the diagram @xmath1240 ^ -{\nu^i\cdot\overline{a(j)}}&w_j\ar@{->}[r]\ar@{->}[d]&w_{k-1}\ar@{->}[d]\\ & f_j\ar@{->}[r]&f_{k-1 } } \ ] ] when @xmath1241 .
since the map @xmath1242 induces an isomorphism in cohomology in dimensions less than @xmath1243 , the first result follows from [ prop6.24 ] .
the second result follows directly from the first since there is a map of fibrations @xmath1244 { } \xymatrix@c=30pt { s^{2n-1}\ar@{->}[r]\ar@{->}[d]&\omega^2s^{2n+1}\ar@{->}[d]\\ w_{k-1}\ar@{->}[r]\ar@{->}[d]&j_{k-1}\ar@{->}[d]\\ f_{k-1}\ar@{=}[r]&f_{k-1 } } \\[-17pt ] & \makebox[323pt ] { } \rlap{\qed } \end{aligned}\ ] ] at this point we introduce a simplified notation analogous to the notation in case @xmath107 .
we define @xmath1245 homotopy classes @xmath1246 for @xmath90 by the formulas @xmath1247 consequently , if @xmath1248 , @xmath1249 .
we will often not distinguish between @xmath1250 and its composition with @xmath1251 for @xmath1252 .
however @xmath1253 we will write @xmath1254 for @xmath1255 with @xmath1256 unspecified but @xmath1257 , so @xmath1258 and similarly for @xmath1259 .
[ cor6.27 ] the compositions @xmath1260 ^ -{x_i(k)}&j_k\ar@{->}[r]&f_k } \\ \xymatrix { p^{2ni-1}\ar@{->}[r]^-{y_i(k)}&j_k(i-1)\ar@{->}[r ] & j_k(i-1)/j_k(i-2 ) \simeq \omega^2s^{2n+1}\ltimes s^{2n(i-1 ) } } \end{gathered}\ ] ] induce integral cohomology epimorphisms for all @xmath90 .
this follows from [ cor6.25 ] and [ prop6.11 ]
. the main technical tool in relating the @xmath1261 and @xmath1262 with @xmath1263 and @xmath1264 will be the following [ theor6.28 ] suppose @xmath1265 with @xmath1266 and the composition @xmath1267
^ -{\varphi}&w_{k-1}^{2np^k-2}\ar@{->}[r]&j_{k-1 } } \ ] ] has order @xmath4 .
then @xmath1268 is congruent to the sum @xmath1269 where @xmath1270 and @xmath1271 .
there are several steps in the proof of [ theor6.28 ] . under the inductive hypothesis , @xmath1263 and @xmath1264 factor through @xmath1094 .
[ prop6.29 ] the map @xmath1272 ^ -{{\scalebox{.9}{${y_i(k-1)\vee x_i(k-1)}$ } } } & w^{2np^k-2}_{k-1}}\ ] ] induces a monomorphism @xmath526 cohomology .
@xmath1273 is trivial unless @xmath1274 or @xmath1275 for @xmath1276 , in which case it is @xmath1277 by [ cor6.19 ] .
each of these classes is nontrivial under either @xmath1263 or @xmath1264 .
we seek to compare the maps @xmath1263 and @xmath1264 to a natural basis for @xmath1201 .
choose maps @xmath1278 for @xmath1279 which define the splitting of [ cor6.19 ] .
@xmath1280 ^ -{{\scalebox{.9}{${\cong}$}}}&w^{2np^k-2}_{k-1}}\ ] ] where @xmath1281 if @xmath1282 and @xmath1283 if @xmath1284 . now define a map @xmath1285 ^ -{}&w^{2np^k-2}_{k-1}}\ ] ] with components @xmath1286 and @xmath1287 .
[ prop6.30 ] suppose @xmath1288 has order @xmath4 . then there is a congruence @xmath1289 for some maps @xmath1290 and @xmath1291 .
this follows directly from [ prop5.12 ] and [ theor5.15 ] , since @xmath1201 is a wedge of even dimensional moore spaces by [ cor6.19 ] .
[ cor6.31 ] if @xmath569 , there is a diagram @xmath1292_{f } \ar@{->}[dr]^-{\xi}&\\ \ar@{}[r]|(.3){\equiv } & w^{2np^k-2}_{k-1}\\ \bigvee\limits^{p^k-1}_{i=2}p^{2ni}\vee s^{2ni-1}\ar@{->}[ru]^{\lambda } & } \ ] ] which commutes up to congruence , for some map @xmath167 . if @xmath569 , @xmath1293 is a double suspension whose identity map has order @xmath4
. thus [ cor6.31 ] follows from [ prop6.30 ] .
in particular , we obtain a congruence formula by restricting [ cor6.31 ] to @xmath1294 : @xmath1295 for some maps @xmath1296 and @xmath1297 . actually , the coefficient of @xmath1298 in this formula is a unit by a cohomology calculation .
we can safely assume it is the identity by adjusting the basis @xmath1299 .
we intend to use this formula to replace the term @xmath1298 in [ prop6.30 ] by @xmath1263 plus lower dimensional terms .
this is a matter of linear substitutions , and we explain this more clearly in a general context .
observe that all the spaces in these formulas are co-@xmath0 spaces and [ cor5.14 ] applies .
[ lem6.32 ] in an additive category , the formulas @xmath1300 imply that there is a formula : @xmath1301 use downward induction beginning with replacing @xmath1302 with @xmath1303 . comparing [ prop6.30 ] with the formula for @xmath1263 above and applying [ lem6.32 ] , we get [ cor6.33 ] suppose @xmath1304 has order @xmath4 and @xmath569
. then @xmath65 is congruent to a sum @xmath1305 where @xmath1306 and @xmath1307 .
we apply [ prop5.19](a ) to [ cor6.33 ] to obtain a congruence @xmath1308 but by [ prop3.29 ] , @xmath1309 where @xmath1310 substituting we get @xmath1311 we also apply [ cor6.33 ] with @xmath1312 to get @xmath1313 and apply [ prop5.19 ] , we get @xmath1314 we now apply [ lem6.32 ] to [ eq6.34 ] and [ eq6.35 ] with @xmath1315 , @xmath1316 and @xmath1317 to obtain @xmath1318{}\{\zeta,\varphi\}_r\equiv\sum^{p^k-1}_{i=2}\{\zeta , y_i-(k-1)\}_r\alpha_i+\{\zeta , x_i(k-1)\}_r\beta_i .
\makebox[20pt ] { } \rlap{\qed}\ ] ] in case @xmath1319 we can precompose with @xmath553 @xmath1320&p^m\wedge p^{\ell}\ar@{->}[r]^-<<<<<{\{\zeta,\varphi\}_r}&j_{k-1 } } \ ] ] to obtain @xmath1321_r\equiv\sum^{p^k-1}_{i=2}\{\zeta , y_i(k-1)\}_r\alpha'_i+\{\zeta , x_i(k-1)\}_r\beta'_i\ ] ] and apply [ prop5.19](b ) to obtain [ cor6.36 ] suppose @xmath1322 and @xmath1323 .
then @xmath1324_r$ ] is congruent to a sum @xmath1325_r\alpha_i+[\beta(\zeta),y_i(k-1)]_r\beta_i+[\zeta , x_i(k-1)]_r\gamma_i+[\beta(\zeta),x_i(k-1)]_r\delta_i\ ] ] for some maps @xmath1326 [ cor6.37 ] suppose @xmath1327
. then @xmath1328 and @xmath1329 .
in case @xmath1330 we apply [ cor6.36 ] .
the formula simplifies since @xmath1331 and @xmath1332 , while @xmath1333 and @xmath1334 . in case
@xmath1335 we apply induction and [ prop5.19](b ) with @xmath1336 and either @xmath1337 we will use [ cor6.37 ] to compare the obstructions at adjacent levels . recall ( [ theor4.4 ] ) , the map @xmath1338 ^ -{{\scalebox{.9}{${\beta_k}$}}}&r_{k-1}\ar@{->}[r]^-{{\scalebox{.9}{${}$}}}&w_{k-1};}\ ] ] @xmath96 induces a cohomology epimorphism by [ cor6.21 ] .
we apply [ theor6.28 ] where @xmath65 is one of the two maps : @xmath1339 the maps @xmath1340 and @xmath1341 are uniquely defined as maps to @xmath1094 since each term lies in the kernel of @xmath1095 .
the fact that they factor through @xmath1342 follows from [ prop6.22 ] and [ prop6.24 ] ( in the case @xmath1343 apply [ lem6.4 ] and [ lem6.5 ] in place of [ prop6.24 ] ) .
note that by [ prop6.20 ] @xmath1344 where @xmath1345 , so @xmath1346 thus we have @xmath1347 we will filter @xmath102 by spaces @xmath1217 as in [ eq6.23 ] and construct maps @xmath1348 by induction first on @xmath35 and then on @xmath29 . by design
the map @xmath1349 will annihilate the classes @xmath1350 and @xmath1351 .
however @xmath1349 will not be an extension of @xmath1352 and we need to know that the classes @xmath1353 and @xmath1354 are also in the kernel . for this purpose
we establish the following [ lem6.40 ] suppose @xmath1134 and we have constructed @xmath1355 such that the kernel of @xmath1356 contains the classes @xmath1357 and @xmath1358 for @xmath1359 and @xmath1360
. then the kernel of @xmath1356 also contains the classes @xmath1361 when @xmath1362 .
let @xmath1363 .
then we have @xmath1364 by [ eq6.26 ] , [ eq6.39 ] and [ cor6.37 ] .
since the identity map of @xmath43 has order @xmath38 and each of the classes @xmath1365 and @xmath1366 is equal to @xmath1367 and @xmath1368 respectively with @xmath1360 , we can conclude that @xmath1369 .
similarly , @xmath1370 by [ cor6.37 ] . consequently @xmath1371 . likewise , by [ eq6.39 ] we obtain @xmath1372 which lies in @xmath1373 when @xmath1374 .
similarly @xmath1375 which is in @xmath1373 when @xmath1374 .
[ theor6.41 ] there are maps @xmath1087 such that @xmath105 restricted to @xmath1376 is homotopic to the restriction of @xmath1095 and such that the compositions @xmath1377 ^ -<<<{a(k)}&e_k\ar@{->}[r]^{\tau_k}&j_k\ar@{->}[r]^->>>>>{\gamma_k}&bw_n\qquad(k\geqslant 1)\\ p^{2ni}\ar@{->}[r]^{x_i(s)}&j_s\ar@{->}[r]&j_k\ar@{->}[r]^{\gamma_k}&bw_n\\ p^{2ni-1}\ar@{->}[r]^{y_i(s)}&j_s\ar@{->}[r]&j_k\ar@{->}[r]^{\gamma_k}&bw_n } \end{gathered}\ ] ] are null homotopic for @xmath90 and @xmath1378 .
recall ( [ eq6.23 ] ) that @xmath1379 where @xmath1380 and @xmath1381 when @xmath1382 .
we have induced principal fibrations : @xmath1031\ar@{->}[d ] & \omega^2s^{2n+1}\ar@{->}[d]\\ j_k(m-1)\ar@{->}[r]\ar@{->}[d]&j_k(m)\ar@{->}[d]\\ f_k(m-1)\ar@{->}[r]&f_k(m ) } \ ] ] we will proceed by induction first on @xmath35 and then on @xmath29 .
the result follows from [ prop6.3 ] when @xmath107 .
suppose we have constructed @xmath1095 and @xmath1383 is defined agreeing with @xmath1095 on @xmath1384 and such that the classes @xmath1357 and @xmath1358 are in the kernel of @xmath1383 when @xmath1359 and @xmath1378 . by [ prop6.24 ]
the composition @xmath1385 ^ ->>>>>>>{x_m(k)}&j_k(m)\ar@{->}[r]^{\eta_k}&f_k(m ) } \ ] ] induces an isomorphism in @xmath526 cohomology in dimension @xmath1029 .
we apply [ theor6.1 ] to construct an extension @xmath1386 of @xmath1383 .
we first consider the case @xmath1387 .
since @xmath1013 extends @xmath1388 , the classes @xmath1357 and @xmath1358 are in the kernel of @xmath1013 when @xmath1360 and @xmath1159 . by [ lem6.40 ] , the kernel of @xmath1013 also contains the classes @xmath1389 , @xmath1390 , @xmath1391 , @xmath1392 and @xmath1393 .
now let @xmath1394 and @xmath1395 be given by @xmath1396 .
let @xmath1397 .
we next show that the composition @xmath1398&p\ar@{->}[r]^->>>>{u}&j_k(p^k)\ar@{->}[r]^{\gamma'}&bw_n } \ ] ] is null homotopic . since @xmath1399 and @xmath96 is in the kernel of @xmath1095 , @xmath1400 .
this implies that the composition @xmath1401&p^{2np^k}(p^{r+k-1})\ar@{->}[r]^-<<<<{a(k)}&j_k(p^k)\ar@{->}[r]^{\gamma'}&bw_n } \ ] ] is divisible by @xmath1402 . since @xmath1070 ,
this composition is null homotopic .
similarly @xmath1403=-p^{k-1}\mu\cdot b(k)\ ] ] which is in the kernel of @xmath1013 .
thus the composition [ eq6.42 ] is null homotopic and we can apply [ theor6.1 ] to construct a differnt extension @xmath1404 of @xmath1405 .
we use [ prop6.24 ] to verify that the composition @xmath1406&j_k(p^k)\ar@{->}[r]&j_k(p^k)\!{}\big/{}\!j_k(p^k\!{}-{}\!1)\simeq
\omega^2s^{2n+1 } \!{}\ltimes{}\ !
s^{2np^k } \ar@{->}[r]^->>{\xi}&s^{2np^k}\!{}\vee{}\ !
s^{2n(p^k+1)-1 } } \ ] ] induces an integral cohomology epimorphism .
thus after composing with a homotopy equivalence on the wedge of spheres , we see that it is homotopic to the quotient map @xmath1407 .
we apply [ theor6.1 ] to construct @xmath1408 with @xmath1409 and @xmath1410 in the kernel .
apply [ lem6.40 ] again , this time to @xmath1408 to see that the classes @xmath1411 and @xmath1390 are in the kernel .
then we repeatedly apply [ lem6.40 ] to restrictions of @xmath1408 to @xmath1412 for @xmath1413 to see that @xmath1414 and @xmath1415 are in the kernel for @xmath1360 .
the case @xmath1416 is similar . since @xmath1013 extends @xmath1383 , the classes @xmath1357 and @xmath1358 are in the kernel of @xmath1013 when @xmath1359 and @xmath1360 . by [ lem6.40 ]
the kernel of @xmath1013 also contains the classes @xmath1417 , @xmath1354 and the classes @xmath1418 and @xmath1419 when @xmath1362 .
we now define @xmath655 and @xmath1420 @xmath1421 and we calculate @xmath1422 @xmath1423 is in the kernel of @xmath1383 and hence in the kernel of @xmath1013 and @xmath1424 is in the kernel since @xmath1425 . similarly @xmath1426 is in @xmath1427 . as before , this implies that the restriction of @xmath1428 to @xmath1429 is null homotopic and we can construct the required map @xmath316 satisfying [ theor6.1 ] .
this allows for the construction of @xmath1349 which annihilates @xmath1350 and @xmath1351 . as before ,
we apply [ lem6.40 ] to conclude that all classes @xmath1357 and @xmath1358 are annihilated by @xmath1349 when @xmath1430 and @xmath1378 .
[ theor6.43 ] suppose @xmath569 .
then the composition @xmath809 ^ -<<<<{\gamma_k}&e_k\ar@{->}[r]^{\eta_k}&j_k\ar@{->}[r]^->>>>{\gamma_k}&bw_n } \ ] ] is null homotopic by [ theor6.41 ] @xmath105 annihilates @xmath1431 for each @xmath992 and @xmath94 . by [ lem6.40 ]
@xmath105 annihilates @xmath1432 for each @xmath488 .
the result follows [ prop5.7 ] , [ theor5.9 ] and [ theor5.20 ] . as in the proof of [ prop2.12 ]
, we have a homotopy equivalence @xmath1433 and consequently we can construct @xmath1434 and redefine @xmath105 as the restriction of @xmath1435 to @xmath102 . similarly we define @xmath1436 and we have [ theor6.44 ] there are maps @xmath105 and @xmath83 such that @xmath1437 and @xmath1438 . furthermore , there are homotopy commutative diagrams of fibration sequences @xmath1439\ar@{->}[d]_{h_k } & \omega g_k\ar@{->}[d]_{\omega\varphi_k}\\ t\ar@{->}[r]\ar@{->}[d]&\omega s^{2n+1}\{p^r\}\ar@{->}[r]^-<<<{h}\ar@{->}[d]&bw_n\ar@{=}[d]\\ r_k\ar@{->}[d]\ar@{->}[r]&e_k\ar@{->}[d]\ar@{->}[r]^{\nu_k}&bw_n\\ g_k\ar@{=}[r]&g_k } & \xymatrix{\omega d_k\ar@{=}[r]\ar@{->}[d]_{h'_k}&\omega d_k\ar@{->}[d]\\ t\ar@{->}[d]\ar@{->}[r]&\omega s^{2n+1}\{p^r\}\ar@{->}[d]\ar@{->}[r]^-<<<{h}&bw_n\ar@{=}[d]\\ w_k\ar@{->}[r]\ar@{->}[d]&j_k\ar@{->}[d]\ar@{->}[r]^{\gamma_k}&bw_n\\ d_k\ar@{=}[r]&d_k } \end{tabular*}\ ] ] where the left hand diagram maps into the right hand diagram .
the maps @xmath1440 and @xmath1441 are the restrictions of maps @xmath1442 and @xmath1443 and there are compatible maps @xmath1444 with @xmath1445 homotopic to the inclusion and compatible maps @xmath1446 with @xmath1447 homotopic to the identity .
the aim of this chapter is to prove theorem b , corollaries c and d , and to discuss some applications .
we describe an obstruction theory for the existence and uniqueness of extensions @xmath27 and evaluate the obstructions in some cases .
we have , however , no example of a homotopy abelian @xmath0-space and map @xmath1448 for which no extension to an @xmath0-map @xmath41 exists . in this section
we describe the basic result relating to the existence and uniqueness of an extension of a map @xmath40 to an @xmath0-map @xmath41 .
the proofs of these results are reserved for the next two sections .
we note , however , that a different argument for the obstruction to uniqueness was obtained in @xcite .
we begin with some notation . throughout this chapter
, @xmath14 will be an arbitrary homotopy abelian @xmath0-space .
we will call an @xmath0-map @xmath1449 proper if the compositions : @xmath1450 ^ -<<<<<<{{\scalebox{.9}{${a(i)\vee c(i)}$}}}&\omega g_i\ar@{->}[r]&\omega g_k\ar@{->}[r]^{{\scalebox{.9}{${\alpha}$}}}&z } \ ] ] are null - homotopic for each @xmath218 , @xmath751 .
let @xmath1451 be the abelian group of all homotopy classes of proper @xmath0-maps @xmath1449 , where @xmath1452 ( @xmath1453 ) .
let @xmath1454 by which we mean the subgroup of all elements of @xmath1455 which are divisible by @xmath1149 .
let @xmath1456_h$ ] be the abelian group of all @xmath0-maps from @xmath1457 to @xmath1458 . clearly @xmath1459=\pi_{2n}(z;z / p^r)\ ] ] [ theor7.2 ] @xmath1460_h$ ] .
[ theor7.3 ] there is an exact sequence : @xmath18&p_k(\omega z)\ar@{->}[r]^{{\scalebox{.9}{${e}$}}}&g_k(z)\ar@{->}[r]^{{\scalebox{.9}{${r}$}}}&g_{k-1}(z)\ar@{->}[r]^{{\scalebox{.9}{${\beta}$}}}&p_k(z ) } \ ] ] we will see by example that this sequence is not exact on the right .
in fact , we have no example in which @xmath1461 .
but there are examples in which @xmath1462 . in this section
we will prove theorem [ theor7.2 ] .
it is a consequence of the following propositions : [ prop7.4 ] @xmath1463 [ prop7.5 ] @xmath1464_h\cong g(z)$ ] .
we first establish that for any space @xmath198 , @xmath1465\cong [ g , x]\ ] ] the argument here is a special case of the results in @xcite .
first we observe that the restrictions define an epimorphism : @xmath1466\ar@{->}[r]&\underset{\leftarrow}{\lim } [ g_k , x]}\ ] ] by the homotopy extension property applied inductively for each @xmath35 .
suppose , however , that @xmath1467 $ ] lies in the kernel ; i.e. , the restrictions : @xmath1468&g\ar@{->}[r]^-{{\scalebox{.9}{${\alpha}$}}}&x}\ ] ] are all null homotopic .
we construct a homotopy commutative diagram in which the horizontal sequence is a cofibration sequence : @xmath1469&g\ar@{->}[r]\ar@{->}[d]_{{\scalebox{.9}{${\alpha}$}}}&{\mathcal{c}}(g)\ar@{->}[r]\ar@{-->}[dl]^{{\scalebox{.9}{${\alpha'}$}}}&\raisebox{-6pt}{$\smash{\bigvee\limits_{k\geqslant 0}\sigma g_k}$}\ar@{->}[r]&\sigma g\\ & x & & & } \ ] ] however by [ theor2.14](i ) , @xmath280 is a wedge of moore spaces , so the map @xmath1470 has a right homotopy inverse .
this implies that @xmath1471 and consequently the map @xmath1472 is null homotopic .
it follows that @xmath327 is null homotopic . to complete the proof of [ prop7.4
] , consider the diagram : @xmath1473_h\ar@{->}[r]^->>>>{{\scalebox{.9}{${l}$}}}\ar@{->}[d]_{{\scalebox{.9}{${\varphi}$}}}&\underset{\leftarrow}{\lim}[\omega g_k , z]_h\ar@{->}[d]_{{\scalebox{.9}{${\varphi}$}}}\\ [ g,\sigma z]\ar@{->}[r]^->>>>>{{\scalebox{.9}{${\cong}$}}}\ar@{->}[d]_{{\scalebox{.9}{${\psi}$}}}&\underset{\leftarrow}{\lim}[g_k,\sigma z]\ar@{->}[d]_{{\scalebox{.9}{${\psi}$}}}\\ [ \omega g , z]_h\ar@{->}[r]^->>>>>{{\scalebox{.9}{${l}$}}}&\underset{\leftarrow}{\lim}[\omega g_k , z]_h } \ ] ] where @xmath1474 and @xmath1475 . clearly @xmath1476 .
the middle horizontal homomorphism is an isomorphism by [ eq7.6 ] . since @xmath305 is an epimorphism
, @xmath1144 is an epimorphism and since @xmath65 is a monomorphism , @xmath1144 is a monomorphism . clearly proper @xmath0-maps in @xmath1477_h$ ] correspond to proper @xmath0-maps in @xmath1478_h$ ] for each @xmath35 .
the proof of [ prop7.5 ] will depend on an analysis of @xmath115 .
we define spaces @xmath99 and @xmath100 and maps @xmath175 and @xmath1479 as follows : @xmath1480 by the maps @xmath572 and @xmath573 from ( [ theor4.4 ] ) on the respective factors . the maps @xmath175 and@xmath1479 factor through @xmath115 by [ prop5.3 ] and [ theor6.41 ] .
[ prop7.7new ] @xmath1481 where the inclusion of @xmath1482 in @xmath115 factors through @xmath1483 . given an @xmath0-map @xmath1484 , the composition @xmath1485 : @xmath474^{{\scalebox{.9}{${h}$}}}&t\ar@{->}[r]^{{\scalebox{.9}{${\alpha}$}}}&z } \ ] ] is a proper @xmath0-map since @xmath68 is proper .
we construct an inverse : @xmath609^{{\scalebox{.9}{${g}$}}}&\omega g\ar@{->}[r]^{{\scalebox{.9}{${\beta}$}}}&z } \ ] ] however , since @xmath617 is not an @xmath0-map , we need an extra argument to show that @xmath1486 is an @xmath0-map .
consider the diagram : @xmath1487 ^ ->>>>>{{\scalebox{.9}{${g\times g}$}}}\ar@{->}[d]_{{\scalebox{.9}{${\mu}$}}}&\omega g\times \omega g\ar@{->}[r]^{{\scalebox{.9}{${h\times h}$}}}\ar@{->}[d]_{{\scalebox{.9}{${\mu}$}}}&t\times t\ar@{->}[d]_{{\scalebox{.9}{${\mu}$}}}\\ t\ar@{->}[r]^{{\scalebox{.9}{${g}$}}}&\omega g\ar@{->}[r]^{{\scalebox{.9}{${h}$}}}&t } \ ] ] in which the left - hand square is not homotopy commutative .
since the right - hand square and the rectangle are homotopy commutative , the difference between the two sides of the left - hand square @xmath1488 factors through the fiber of @xmath68 : @xmath1489&\omega g\ar@{->}[r]^{{\scalebox{.9}{${h}$}}}&t\ar@{->}[r]&r\ar@{->}[r]&g . } \ ] ] however since @xmath133 is proper and @xmath14 is homotopy abelian the composition : @xmath1489&\omega g\ar@{->}[r]^{{\scalebox{.9}{${\beta}$}}}&z } \ ] ] is null homotopic by [ prop7.7new ] thus @xmath1490 is null homotopic and @xmath1486 is an @xmath0-map .
the remainder of this section will be devoted to a proof of [ prop7.7new ] we begin by clarifying the relationship between @xmath115 and @xmath60 . [ prop7.8new ] @xmath1491 . by [ prop6.17 ] ,
the homomorphism @xmath1492 is a split epimorphism .
since @xmath115 is a wedge of moore spaces by [ prop6.16 ] and @xmath60 is a wedge of moore spaces by [ cor6.19 ] , the map @xmath1493 has a right homotopy inverse . in the proof of [ prop6.17 ]
, a map of fibration sequences was studied ; in the limit , this is of the form @xmath1096\ar@{->}[d]&t\ar@{->}[d]\\
r\ar@{->}[r]\ar@{->}[d]&w\ar@{->}[d]\\ g\ar@{->}[r]&d } \ ] ] where @xmath770 is the mapping cone of the map @xmath1494 .
it follows from [ prop2.1 ] that there is a homotopy pushout diagram @xmath1495&w\\ t\times c\ar@{->}[u]\ar@{->}[r]&r\ar@{->}[u ] } \ ] ] where @xmath1496 is the cone on @xmath100 .
since the inclusion of @xmath50 in @xmath115 is null homotopic , there is an induced map @xmath1497 whose cofiber is @xmath60 ; @xmath1498 since the map @xmath1493 has a right homotopy inverse , the result follows .
[ prop7.5old ] there is a homotopy commutative diagram @xmath1499\ar@{->}[d]_{g\ltimes c}&r\ar@{->}[d]\\ \omega g\ltimes e\ar@{->}[r]^-<<<<{\pi\gamma'}&g . }
\ ] ] the map @xmath1500 is given as follows .
the restriction to @xmath1501 comes from a trivialization of the pullback over @xmath100 in the diagram @xmath1096\ar@{->}[d]&\ar@{->}[d]\ar@{=}[r]t&t\ar@{->}[d]\\ t\times c\ar@{->}[r]\ar@{->}[d]&r\ar@{->}[r]\ar@{->}[d]&\ar@{->}[d]j\\ c\ar@{->}[r]^{c}&\ar@{->}[r]g&d } \ ] ] so the composition @xmath1502 is given by the map @xmath1503 ^ -<<<{\pi_2}&c\ar@{->}[r]^{c}&g . }
\ ] ] to understand the map @xmath1504 , consider a general fibration @xmath882&e\ar@{->}[r]^{\pi}&b . }
\ ] ] using the homotopy lifting property , one can construct a map @xmath1505^{l}&(e , f ) } \ ] ] extending the connecting map @xmath1506 and such that @xmath1507 is endpoint evaluation .
in our case we use the map @xmath1508 together with @xmath1509 to obtain the composition @xmath1510 ^ ->>>{g}&(c\omega g,\omega g)\ar@{->}[r]^{\xi}&(pg,\omega g)\ar@{->}[r]^{l}&(r , t ) . } \ ] ] this defines the map @xmath1504 and the composition @xmath1511&r\ar@{->}[r]&g } \ ] ] is given by @xmath1511&{{\mathcal{c}}}(\omega g)\ar@{->}[r]^{\epsilon}&g . } \ ] ] according to the definition of @xmath366 ( see [ eq3.8 ] ) , the map @xmath1512 ^ -<<<{\gamma'}&e\ar@{->}[r]^{\pi}&g}\ ] ] is given by @xmath1513 ^ -<<<{\pi_2}&e\ar@{->}[r]^{\pi}&g } \\
\xymatrix{pg\ar@{->}[r]^{\epsilon}&g}\\[-15pt ] \makebox[\textwidth]{\hfill\qed}\end{gathered}\ ] ] [ cor7.6 ] @xmath1514 and the composition @xmath1515^{\zeta}&t\ltimes c\ar@{->}[r]&r\ar@{->}[r]&g } \ ] ] factors through @xmath1516 ^ -<<<{\gamma}&e\ar@{->}[r]^{\pi}&c . }
$ ] since @xmath100 is a suspension , @xmath1514 . by [ prop7.5old ] , the composition in question factors up to homotopy as @xmath1517&t\ltimes c\ar@{->}[r]^->>>>>{g\ltimes c}&\omega g\ltimes e\ar@{->}[r]^-<<<<{\gamma'}&e\ar@{->}[r]^{\pi}&g } \ ] ] in the diagram below @xmath1518 factors through @xmath1519 since the middle row is a fibration sequence by [ eq3.17 ] : @xmath1515\ar@{->}[d]&t\ltimes c\ar@{->}[r]\ar@{->}[d]_{g\ltimes c}&c\ar@{->}[d]_{c}\\ \omega g*\omega e\ar@{->}[r]\ar@{->}[d]&\omega g \ltimes e\ar@{->}[r]\ar@{->}[d]_{\gamma'}&e\\ \omega g * \omega g\ar@{->}[r]^-<<<<{\gamma}&e\ar@{->}[d]_{\pi}&\\ & g & } \ ] ] since the lower square commutes up to homotopy by [ prop3.15 ] . at this point we have @xmath1520 , where @xmath1521 factors through @xmath500 ^ -<<<{\gamma}&e\ar@{->}[r]^{\pi}&g . }
\ ] ] the remainder of this section will be focused on proving [ prop7.7 ] @xmath1522 where @xmath1523 furthermore , there is a factorization : @xmath1524\ar@{->}[d]&r\ar@{->}[d]\\ \omega g * \omega g\ar@{->}[r]^-<<<{\gamma}&g } \ ] ] the main ingredient for the proof of [ prop7.7 ] is the following result : [ prop7.8 ] suppose @xmath1525 and @xmath146.then there is a map @xmath1526 such that the composition @xmath1527 ^ -<<<{{\scalebox{.9}{${f(m)}$}}}&j_k\ar@{->}[r]^{{\scalebox{.9}{${\eta_k}$}}}&f_k } \ ] ] induces a cohomology epimorphism . there is a factorization : @xmath1528 ^ -<<<{f(m)}\ar@{->}[d]_{w}&j_k\\ \omega g_k * \omega g_k\ar@{->}[r]^-<<<{\gamma_k}&e_k\ar@{->}[u]^{\tau_k } } \ ] ] suppose @xmath1525 . by [ prop7.8 ] , @xmath1529 is null , so there is a factorization : @xmath1530^{\eta_k}&\\ p^{2mn}(p^{r+s})\ar@{->}[ru]^{f(m)}\ar@{->}[rd]_{\overline{f}(m)}&&f_k\\ & w_k\ar@{->}[uu]\ar@{->}[ur ] & } \ ] ] by [ prop6.11 ] , @xmath1531 induces an isomorphism in integral homology in dimension @xmath1532 . assembling the maps @xmath1533 together for all @xmath35 together with the maps @xmath572
we get a map @xmath1534&w } \ ] ] which induces an isomorphism in homology .
thus @xmath1522 and @xmath1535 factors through @xmath1536 by [ prop7.8 ] . the construction of @xmath1533 when @xmath1537 is immediate .
we simply set @xmath1538 . in case
@xmath1539 we need to construct relative whitehead products using @xmath72 . in the special case that @xmath1540 we will need to use an @xmath0-space based whitehead product .
the proof of [ prop7.8 ] will rely on four lemmas .
we will call an integer @xmath29 acceptable if there is a map @xmath1541 with @xmath1542 , satisfying [ prop7.8](a ) and ( b ) .
[ lem7.9 ] suppose @xmath1525 and @xmath1543 is a map such that the composition @xmath1544
^ -<<<{x}&e_k\ar@{->}[r]^{\tau_k } & j_k\ar@{->}[r]^{\eta_k}&f_k } \ ] ] induces an integral cohomology epimorphism .
suppose @xmath1545 .
then there is a map @xmath1546 satisfying [ prop7.8 ] and .
consequently if @xmath1547 , @xmath1548 is acceptable .
this applies in particular when @xmath1416 is acceptable with @xmath1549 .
write @xmath1550 and let @xmath1551 be the inclusion when @xmath94
. we will define @xmath1552 using the relative whitehead product : @xmath1553 ^ -<<<<<<{\{\lambda_i , x\}_r}&e_k\ar@{->}[r]^{\tau_k}&j_k\ar@{->}[r]^{\eta_k}&f_k . } \ ] ] our first task will be to evaluate this in @xmath1554 . recall ( [ eq3.10 ] ) that @xmath1555 is the composition : @xmath1556 ^ -{\psi}&\sigma(\omega g_i*\omega \sigma p)\!{}\simeq{}\!\omega g_i\!{}*{}\!\omega \sigma p\ar@{->}[r]^-{\zeta}&\omega g_i\!{}\ltimes{}\ ! \sigma p\ar@{->}[r]&\omega g_k\!{}\ltimes{}\ !
e_k\ar@{->}[r]^-{\gamma'}&e_k } \ ] ] for the first part of this composition , consider the diagram @xmath1557 ^ ->>>{\psi } & \sigma \omega g_i\wedge \omega\sigma p\\ \mbox{}&\mbox{}\\ g_i\wedge p \ar@{->}[r]^->>>>>{\nu\wedge 1 } \ar@{->}[uu]^{\simeq } & \sigma \omega g_i\wedge p \ar@{->}[uu]_{1\wedge i } } & \parbox[t]{9pt}{\mbox{}\newline\vspace*{-24pt } $ \simeq{}$\newline \vspace*{37pt } $ { } \simeq{}$ } & \xymatrix@c=23pt@r=21.25pt { \omega g_i*\omega\sigma p \ar@{->}[rd]^{\zeta } & \\ & \omega g_i\ltimes \sigma
p\\ \omega g_i*p \ar@{->}[uu]_{1*i } \ar@{->}[ur]_{\zeta ' } } \end{tabular*}\ ] ] where @xmath53 is the co-@xmath0 space structure map on @xmath1558 .
the left hand square is homotopy commutative by [ prop3.2 ] and the right hand triangle defines @xmath1559 ( see footnote to [ eq3.17 ] ) .
choose a generator @xmath1560 and let @xmath1561 be the image of this generator under @xmath1562 .
choose a generator @xmath1563 .
then by the above diagram we have @xmath1564 by applying the commutative square in the proof of [ eq3.17 ] .
we now construct a diagram where the two right hand squares are homotopy commutative by [ eq3.8 ] , [ prop3.11](b ) and [ prop3.11](d ) : @xmath1565^{\omega\lambda_i\ltimes x}&\omega g_k\ltimes e_k\ar@{->}[r]^{\gamma ' } \ar@{->}[d]\ar@{->}[r]^{\gamma'}&e_k\ar@{->}[d]^{\eta_i\tau_k}\\ & \omega d_k\ltimes f_k\ar@{->}[r]^-<<<<<<<{\gamma'}&f_k\\ & \omega d_k\times f_k\ar@{->}[u]\ar@{->}[r]\ar@{->}[u]&\omega s^{2n+1}\times f_k\ar@{->}[u]^{a } } \end{split}\ ] ] to evaluate @xmath1566 in @xmath124 homology we observe that the image of @xmath1567 is @xmath1568 where @xmath1569 is the image of @xmath1570 under the homomorphism induced by the composition @xmath1571 & \omega g_k\ar@{->}[r ] & \omega d_k\ar@{->}[r ] & \omega s^{2n+1 } } \ ] ] and @xmath133 is the image of @xmath1572 under the homomorphism @xmath1573^{x}&e_k\ar@{->}[r]^{\eta_k\tau_k}&f_k . } \ ] ] by hypothesis @xmath1574 is a generator .
[ lem7.12 ] the image of @xmath1575 under the homomorphism @xmath1576 is a unit multiple of the generator . by [ eq5.2 ] ,
the composition in question is homotopic to the composition : @xmath1577 ^ ->>>{\omega\varphi_k}&\omega s^{2n+1}\{p^r\}\ar@{->}[r]&\omega s^{2n+1}. } \ ] ] by [ theor6.44 ] , this factors as @xmath1577^{h_k}&t\ar@{->}[r]&\omega s^{2n+1}. } \ ] ] however the composition @xmath771^{\nu}&\sigma\omega
g_k\ar@{->}[r]^{\sigma h_k}&\sigma t } \ ] ] induces a monomorphism in @xmath124 homology , so @xmath1578 is a nonzero generator which is mapped to a unit multiple of @xmath1579 under the map @xmath1580\omega s^{2m-1}$ ] .
we now complete the proof of [ lem7.9 ] .
since @xmath1581 , we can find a map @xmath1582 which induces an isomorphism in @xmath124 homology in dimension @xmath1583 using [ theor2.14](i ) and [ prop3.25 ] .
then let @xmath1552 be the composition : @xmath1584&g_i\circ p^{2mn}(p^{r+t})\ar@{->}[r]^-<<<<{\{\lambda_i , x\}_r}&e_k\ar@{->}[r]^{\tau_k}&j_k . }
\ ] ] by [ eq7.10 ] , [ eq7.11 ] , and [ lem7.12 ] , @xmath1585 induces an epimorphism in @xmath124 cohomology and consequently in integral cohomology as well . by [ prop3.15 ]
, we have a homotopy commutative diagram @xmath1586^{\zeta}\ar@{->}[d]&\omega g_i\ltimes \sigma p\ar@{->}[d]\\ \omega g_k*\omega e_k\ar@{->}[r]^{\zeta}\ar@{->}[d]&\omega g_k\ltimes e_k\ar@{->}[r]\ar@{->}[d]_{\gamma'}&\omega d_k\ltimes
j_k\ar@{->}[d]_{\gamma'}\\ \omega g_k*\omega g_k\ar@{->}[r]^-<<<<<{\gamma}&e_k\ar@{->}[r]&j_k } \ ] ] so @xmath1552 factors through @xmath150
. we need to consider the case @xmath1540 separately .
we will construct @xmath120 ^ -<<<<{f(2p^k)}&w . } \ ] ] it is the first example of a mod @xmath119 moore space in @xmath60 by [ cor6.19 ] .
[ prop7.13 ] @xmath1587 is acceptable .
we will construct @xmath1588 as an @xmath0-space based whitehead product .
let @xmath1589 be the @xmath246 skeleton of @xmath72 .
then @xmath1590 has dimension @xmath1591 .
[ lem7.14 ] @xmath1590 is a wedge of moore spaces .
consider the cofibration sequence @xmath1592&g_k\circ g_k\ar@{->}[r]^->>>>>{q}&s^{2np^k+1}\circ g_k\simeq \sigma^{2np^k}g_k . } \ ] ] since @xmath1593 is a wedge of moore spaces by [ theor2.14](l ) and [ prop3.21 ] , it suffices to show that this cofibration splits .
choose a basis @xmath1594 for @xmath1595 with @xmath950 and @xmath1596 .
consider the classes @xmath1597 corresponding to @xmath1598 .
we can construct a basis for @xmath1599 which contains the elements @xmath1597 and @xmath1600 .
since @xmath1593 is a wedge of moore spaces we can construct maps of moore spaces into @xmath1593 realizing these basis elements and from this a right homotopy inverse to @xmath1601 .
g_k\simeq\overline{g}_k\circ g_k\vee \sigma^{2np^k}g_k.\makebox[98pt]{}\qed\ ] ] now consider the principal fibration sequence : @xmath1603 & \omega \overline{g}_k * \omega g_k\ar@{->}[r ] & \overline{g}_k\vee g_k . }
\ ] ] we will study the integral homology serre spectral sequence of this fibration .
the principal action defines a module structure @xmath1604 for @xmath1605 , let @xmath1606 be the image of @xmath1607 in the @xmath1111 axis .
let @xmath1608 be the desuspended image under @xmath1562 of @xmath1607 , and @xmath1609 the image of @xmath1610 in the @xmath1611 axis in @xmath1612 . using the universal coefficient theorem
, we define a monomorphism @xmath1613 then @xmath1614 .
define @xmath1615 by @xmath1616 @xmath316 has order @xmath119 and @xmath1617 .
since @xmath1618 when @xmath1619 , @xmath316 survives to an element of order @xmath119 in @xmath1620 . by [ lem7.14 ] , [ prop3.21 ] and [ theor2.14](l ) ,
@xmath1621 is a wedge of moore spaces and by an easy homology calculation , every element in the homology has order at most @xmath119 .
consequently @xmath316 converges to an element @xmath1622 $ ] of order @xmath119 in @xmath1623 .
since @xmath1624 is a wedge of moore spaces , there is a map @xmath1625 whose homology image contains a class @xmath1626 with @xmath1627 equal to the @xmath124 reduction of @xmath1622 $ ] .
now let @xmath1010 be the composition : @xmath1628&\omega g_k*\omega g_k\ar@{->}[r]^-<<<{\gamma_k}&e_k\ar@{->}[r]^{\tau_k}&j_k\ar@{->}[r]^{\eta_k}&f_k } \ ] ] [ lem7.15 ] the composition @xmath120^{\varphi}&\omega\overline{g}_k*\omega g_k\ar@{->}[r]^-<<<{\gamma}&f_k } \ ] ] induces an isomorphism in @xmath124 homology in dimension @xmath1591 .
this implies that @xmath1629 induces an epimorphism in integral cohomology .
consider the two diagrams of principal fibrations @xmath1630\ar@{->}[d ] & \omega\overline{g}_k\times \omega g_k\ar@{->}[d]\\ l\ar@{->}[r]\ar@{->}[d]&\omega\overline{g}_k*\omega g_k\ar@{->}[d]\\ g_{k-1}\vee g_{k-1}\ar@{->}[r]&\overline{g}_k\vee g_k } & \xymatrix { \omega^2s^{2n+1}\ar@{=}[r]\ar@{->}[d ] & \omega^2s^{2n+1}\ar@{->}[d]\\ f_{k-1}\ar@{->}[r]\ar@{->}[d]&f_k\ar@{->}[d]\\ d_{k-1}\ar@{->}[r]&d_k } \end{tabular*}\ ] ] where @xmath1144 is the total space of the induced fibration .
the homotopy commutative square @xmath1631\ar@{->}[d]&d_k\ar@{->}[d]\\ \overline{g}_k\times g_k\ar@{->}[r]&s^{2n+1 } } \ ] ] induces a map from the left hand pair of fibrations to the right hand pair .
we apply [ prop2.1 ] to obtain the following homotopy commutative diagram : @xmath1632^{\gamma}\ar@{->}[d]_{\simeq}&f_k\big / f_{k-1}\ar@{->}[d]_{\simeq}\\ ( s^{2np^k}\vee p^{2np^k+1}(p^{r+k}))\rtimes \omega \overline{g}_k\times \omega g_k\ar@{->}[r]^-<<<<{\gamma'}&\omega s^{2n+1}\ltimes p^{2np^k+1}(p ) } \ ] ] let @xmath1633 be the image of @xmath1634 and @xmath1635 be the image of @xmath1636 under the quotient map @xmath1631&\overline{g}_k\vee g_k\big / g_{k-1}\vee g_{k-1}\simeq s^{2np^k}\vee p^{2np^k+1}(p^{r+k } ) . } \ ] ] choose @xmath617 with @xmath1637
. then the image of @xmath1622 $ ] under the left hand equivalence is @xmath1638 .
the image of @xmath659 is a class @xmath1639 such that @xmath1640 at this point in the bockstein spectral sequence there are very few classes left and the only possibility for @xmath1639 is @xmath1641 where @xmath1642 .
the components of the map @xmath1013 are the inclusion @xmath1643&p^{2np^k+1}(p)}$ ] and @xmath1644&p^{2np^k+1}(p).}\ ] ] by [ lem7.12 ] , the image of @xmath1645 is @xmath1646 .
we conclude that the image of @xmath1639 is nonzero @xmath1647^{\varphi}&\omega\overline{g}_k*\omega g_k\ar@{->}[r]^-<{\gamma}&f_k\ar@{->}[r]&f_k\big / f_{k-1}\simeq p^{2np^k+1}(p)\rtimes \omega s^{2n+1 } } \ ] ] in @xmath526 homology , from which the conclusion follows .
let @xmath1588 be the composition @xmath120 ^ ->>>{\varphi}&\omega\overline{g}_k*\omega g_k\ar@{->}[r]^-<<<{\gamma_k}&e_k\ar@{->}[r]^{\eta_k}&j_k . } \ ] ] the result follows from [ lem7.15 ] .
write @xmath1648 where @xmath1649 .
let @xmath1650 bet the number of coefficients @xmath1651 which are nonzero .
we first deal with the case @xmath1652 . then @xmath1653 with @xmath1654 since @xmath1525 .
the case @xmath1655 is [ prop7.13 ] . if @xmath1656 we apply [ lem7.9 ] with @xmath1657 and @xmath1658 , to establish this case by induction . in case @xmath1659
, we first consider the case @xmath1660 for @xmath648 . in case
@xmath1661 we apply [ lem7.9 ] with @xmath1662 and @xmath1663 . in case
@xmath1664 we apply [ lem7.9 ] with @xmath1665 . in this case @xmath1666 .
we now consider the general case with @xmath1659 . in this case
@xmath1667 with @xmath648 .
we do this by induction on @xmath1651 with @xmath1668 as before .
the general case is by induction on @xmath1650 and then induction on the coefficient of the least power of @xmath38 in the expansion using [ lem7.9 ] repeatedly . by [ prop7.8new ] and [ cor7.6 ] @xmath1669 where @xmath1518 is a wedge of moore spaces which factors through @xmath150 . by [ prop7.7 ]
@xmath1522 where @xmath1535 is a wedge of moore spaces which factors through @xmath150 .
set @xmath1670 . in this section
we will define the homomorphisms @xmath299 , @xmath23 , and @xmath133 and prove theorem [ theor7.3 ] .
we define @xmath1671 as follows .
let @xmath1672 , and extend the adjoint @xmath1673 to an @xmath0-map @xmath1674 since @xmath1675 is divisible by @xmath1149 , so is @xmath1676 and consequently the composition @xmath1577
^ ->>>>{{\scalebox{.9}{${\omega \pi'}$}}}&\omega p^{2np^k+1}(p^{r+k})\ar@{->}[r]^->>>>{{\scalebox{.9}{${\widehat{\phi}}$}}}&z } \ ] ] is a proper @xmath0-map .
we define @xmath1677 .
@xmath299 is clearly a homomorphism . to see that @xmath299 is a monomorphism
, we suppose @xmath1678 is null homotopic .
since @xmath1679 is a co-@xmath0 map , we have a homotopy commutative diagram : @xmath1680^{{\scalebox{.9}{${\nu}$}}}\ar@{->}[dd]_{{\scalebox{.9}{${\pi'}$}}}&\sigma \omega g_k\ar@{->}[dr]^{{\scalebox{.9}{${\sigma e(\phi)}$}}}\ar@{->}[dd]_{{\scalebox{.9}{${\sigma\omega\pi'}$}}}&\\ & & \sigma z\\ p^{2np^k+1}(p^{r+k})\ar@{->}[r]&\sigma \omega p^{2np^k+1}(p^{r+k})\ar@{->}[ru]_{{\scalebox{.9}{${\sigma\widehat{\phi}}$ } } } & } \ ] ] since @xmath1678 is null homotopic , the upper composition is null homotopic , so the lower composition factors over the cofiber of @xmath1679 : @xmath771 ^ ->>>>{\pi'}&p^{2np^k+1}(p^{r+k})\ar@{->}[r]&\sigma g_{k-1}\ar@{->}[r]&\sigma g_k . } \ ] ] but since @xmath1681 splits as a wedge of moore spaces , the map @xmath1682 is null homotopic .
it follows that the composition : @xmath1683&\sigma \omega p^{2np^k+1}(p^{r+k})\ar@{->}[r]^-<<<<{{\scalebox{.9}{${\sigma\widehat{\phi}}$}}}&\sigma z } \ ] ] is null homotopic . since @xmath14 in an @xmath0 space ,
we conclude that @xmath650&\omega p^{2np^k+1}(p^{r+k})\ar@{->}[r]^-<<<<{{\scalebox{.9}{${\widehat{\phi}}$}}}&z } \ ] ] is null homotopic .
since @xmath1676 is an @xmath0-map , @xmath1684 is null homotopic and consequently @xmath1675 is as well .
thus @xmath299 is a monomorphism .
the map @xmath1685 is given by restriction . clearly @xmath1686 .
suppose @xmath1687 for some proper @xmath0-map @xmath1449 .
we construct an extension @xmath1688 in the diagram : @xmath1689\ar@{->}[d]_{{\scalebox{.9}{${\nu}$}}}&g_k\ar@{->}[r]\ar@{->}[d]_{{\scalebox{.9}{${\nu}$}}}&p^{2np^k+1}(p^{r+k})\ar@{-->}[ddl]_{{\scalebox{.9}{${\overline{\phi}}$}}}\\ \sigma \omega g_{k-1}\ar@{->}[r]\ar@{->}[dr]_{{\scalebox{.9}{${\ast}$ } } } & \sigma\omega g_k \ar@{->}[d]_{{\scalebox{.9}{${\sigma\alpha}$ } } } & \\ & \sigma z & } \ ] ] we include the loops on the right - hand triangle into the diagram : @xmath1690 ^ -<<<<{{\scalebox{.9}{${\omega\pi'}$}}}\ar@{->}[d]_{{\scalebox{.9}{${\omega\nu}$}}}&\omega p^{2np^k+1}(p^{r+k})\ar@{->}[ddl]_{{\scalebox{.9}{${\omega\overline{\phi}}$}}}\\ \omega g_k\ar@{=}[ru]\ar@{->}[d]_{{\scalebox{.9}{${\alpha}$}}}&\omega\sigma\omega g_k\ar@{->}[l]_{{\scalebox{.9}{${\omega\epsilon}$}}}\ar@{->}[d]_{{\scalebox{.9}{${\omega \sigma \alpha}$}}}&\\ z&\omega \sigma z\ar@{->}[l]_{{\scalebox{.9}{${\mu}$ } } } & } \ ] ] to see that @xmath327 is homotopic to @xmath1691 where @xmath1676 is the composition @xmath1692 . restricting @xmath1676 to
@xmath646 defines @xmath1684 whose adjoint is @xmath1693 to see that @xmath1675 is divisible by @xmath1149 , it suffices to show that @xmath1688 is divisible by @xmath1149 .
however since @xmath327 is proper , the upper composition in the diagram : @xmath1694 ^ -<<<<<<<<{{\scalebox{.9}{${a(k)\vee c(k)}$}}}\ar@{->}[dr]&\omega g_k\ar@{->}[r]^{{\scalebox{.9}{${\alpha}$}}}\ar@{->}[d]_{{\scalebox{.9}{${\omega \pi'}$}}}&\ar@{->}[d]z\\ & \omega p^{2np^k+1}(p^{r+k-1})\ar@{->}[r]^-<<<{{\scalebox{.9}{${\omega\overline{\phi}}$}}}&\omega \sigma z } \ ] ] is null homotopic .
consequently the lower composition and its adjoint are null homotopic : @xmath1695 ^ -<<<<<<{{\scalebox{.9}{${-\delta_1\vee p}$}}}&p^{2np^k+1}(p^{r+k})\ar@{->}[r]^-<<<<<<{{\scalebox{.9}{${\overline{\phi}}$}}}&\sigma z } \ ] ] it follows from [ eq1.9 ] that @xmath1688 is divisible by @xmath1149 .
thus @xmath1684 and @xmath1675 are divisible as well .
finally we define @xmath1696 as the composition @xmath1697
^ -<<<{{\scalebox{.9}{${\widetilde{\beta_{k+1}}}$}}}&\omega e_k\ar@{->}[r]&\omega g_k\ar@{->}[r]^-<<<{{\scalebox{.9}{${\alpha}$}}}&z } \ ] ] using the @xmath0-space structure on @xmath14 we see that @xmath133 is a homomorphism . [ theor7.20 ]
the composition @xmath1698 is divisible by @xmath119 .
this will follow from a sequence of lemmas [ lem7.17]the homomorphism induced by the inclusion @xmath1699 is a monomorphism . by [ prop6.11 ]
, @xmath1700 has one cell in each dimension of the form @xmath1701 or @xmath1702 for each @xmath90 ; consequently @xmath1703 by [ cor6.27 ] , the maps @xmath1704 and @xmath1705 are nonzero in @xmath124 homology in dimensions @xmath1701 and @xmath1702 respectively . by [ theor6.41 ] ,
the maps @xmath1261 and @xmath1262 factor through @xmath1700 up to homotopy when @xmath90 .
the result follows
. we will write @xmath1706 and @xmath1707 for the images of the generators in the homology of the respective moore spaces .
note that in the congruence homotopy of @xmath102 we have @xmath1708 when @xmath1160 by [ eq6.26 ] .
the action of the principal fibration defines an action @xmath1709 let @xmath1710 be the hurewicz images of @xmath53 and @xmath421 ( see [ theor5.20 ] )
. then @xmath1711\otimes \lambda(u)\subset h_*(\omega d_k)$ ] acts on @xmath1712 .
[ lem7.18 ] if @xmath1160 , @xmath1713 and @xmath1714 .
apply [ prop5.16 ] .
[ lem7.19 ] there are maps @xmath1715 and @xmath1716 which are nonzero in @xmath124 homology in dimensions @xmath1717 and @xmath1718 respectively and whose image in @xmath102 factors through the map @xmath1719 ^ -<<<{\gamma_k}&e_k\ar@{->}[r]^{\tau_k}&j_k . } \ ] ] both @xmath1720 and @xmath1721 satisfy the homological condition of [ lem7.19 ] and factor through @xmath1700 .
however @xmath1722 does not factor through @xmath754 since it has order @xmath1723 , and there are no elements in the homotopy of @xmath1724 in that dimension of that order .
let @xmath1725 be the map constructed in [ prop7.8 ] .
apply [ lem7.9 ] with @xmath1726 to construct @xmath1727 the class @xmath1728 in the homology image of @xmath23 since @xmath23 factors through @xmath1700 and projects to a nonzero class in @xmath1729 . to construct @xmath1730 , we return to the class @xmath147 above and observe that the homomorphism induced by @xmath147 : @xmath1731 contains both @xmath1732 and @xmath1733 in its image since @xmath147 factors through @xmath1700 which has @xmath1734 as a retract .
consequently @xmath1735 has @xmath1733 in its homology image .
now consider @xmath1736_r\colon g_k\circ p^{2np^k(p-1)-1}(p^{r+k})\to j_k\ ] ] and apply [ eq7.10 ] and [ eq7.11 ] .
it follows that @xmath1737 is in the homology image by [ lem7.12 ] .
we then choose a map @xmath1738 which induces an isomorphism in dimension @xmath1718 to construct @xmath1730 . since @xmath60 is a retract of @xmath115 by [ prop7.8new ] we can assume that @xmath23 and @xmath1730 factor through @xmath115 .
choose units @xmath1739 and @xmath1740 so that the maps @xmath1741 are trivial in @xmath124 homology .
this can be done since the relevant factor of @xmath1700 is @xmath1742 and both @xmath1720 and @xmath23 are nontrivial in dimension @xmath1717 while @xmath1721 and @xmath1730 are nontrivial in dimension @xmath1718 .
it follows from [ prop7.7new ] that both @xmath1340 and @xmath1341 factor through @xmath1743 since @xmath327 is proper , we conclude that the compositions @xmath1744 ^ -<<<<{\widetilde{\delta}_1}&\omega r_k\ar@{->}[r]&\omega g_k\ar@{->}[r]^{\alpha}&z } \\ & \xymatrix { p^{2np^{k+1}-2}(p^{r+k})\ar@{->}[r]^-<<<<{\widetilde{\delta}_2}&\omega r_k\ar@{->}[r]&\omega g_k\ar@{->}[r]^{\alpha}&z } \end{aligned}\ ] ] are both null homotopic .
since the maprs @xmath23 and @xmath1730 factor through @xmath1724 , these terms are null homotopic and we conclude that the compositions @xmath1745 ^ -<<<<{\widetilde{\beta}_{k+1}\rho}&\omega r_k\ar@{->}[r]&\omega g_k\ar@{->}[r]^{\alpha}&z } \\ & \xymatrix@c=30pt { p^{2np^{k+1}-2}(p^{r+k})\ar@{->}[r]^-<<<<{\widetilde{\beta}_{k+1}\delta_1}&\omega r_k\ar@{->}[r]&\omega g_k\ar@{->}[r]^{\alpha}&z } \end{aligned}\ ] ] are null homotopic .
now consider the cofibration sequence @xmath1746 ^ -<<<<{-\delta_1\vee \rho}&p^{2np^{k+1}-1}(p^{r+k+1})\ar@{->}[r]^{p^{r+k}}&p^{2np^{k+1}-1}(p^{r+k+1 } ) . }
\end{gathered}\ ] ] from this we see that the composition @xmath1747
^ -<<<{\widetilde{\beta}_{k+1}}&\omega r_k\ar@{->}[r]&\omega g_k\ar@{->}[r]^{\alpha}&z } \ ] ] is divisible by @xmath119 .
[ prop7.24 ] @xmath1748 since @xmath327 is proper , the composition on the right in the diagram : @xmath1749 ^ -<<<<<{{\scalebox{.9}{${\sigma\vee\sigma\beta}$}}}\ar@{->}[d]^{{\scalebox{.9}{${\widetilde{\beta}_k}$}}}&p^{2np^k-1}(p^{r+k-1})\vee p^{2np^k}(p^{r+k-1})\ar@{->}[d]^{{\scalebox{.9}{${a(k)\vee c(k)}$}}}&\\ \omega g_{k-1}\ar@{->}[r]&\omega g_k\ar@{->}[r]^{{\scalebox{.9}{${\alpha}$}}}&z } \ ] ] is null homotopic .
the diagram commutes up to homotopy by [ theor4.4 ] from which the result follows .
[ prop7.25 ] if @xmath1750 , @xmath1751 for some proper @xmath0-map @xmath1752 . by [ theor4.4 ] ,
since @xmath59 is a retract of @xmath1754 , @xmath653 and @xmath72 are co-@xmath0 spaces and there is a homotopy commutative diagram @xmath1689\ar@{->}[d]&g_k\ar@{->}[d]\\ g_{k-1}\vee g_{k-1}\ar@{->}[r]&g_k\vee g_k } \ ] ] the map @xmath1755 factors through the fiber of the lower horizontal map in the diagram and defines a homotopy equivalence with the @xmath246 skeleton of the fiber of that map .
however the composition @xmath650 ^ -<<<{\alpha_k}&g_{k-1}\ar@{->}[r]&g_{k-1}\vee g_{k-1 } } \ ] ] factors through this fiber , from which we get a homotopy commutative square @xmath1756^{\alpha_k}\ar@{->}[d]&g_{k-1}\ar@{->}[d]\\ p^{2np^k}(p^{r+k})\vee p^{2np^k}(p^{r+k})\ar@{->}[r]^-<<<<<<{\alpha_k\vee\alpha_k } & g_{k-1}\vee g_{k-1 } ; } \ ] ] that is , @xmath1757 is a co-@xmath0 map .
consequently there is also a homotopy commutative diagram : @xmath650 ^ -<<<<<<{\alpha_k}\ar@{->}[d]&g_{k-1}\ar@{->}[d]_{\nu}\\ \sigma \omega p^{2np^k}(p^{r+k})\ar@{->}[r]&\sigma \omega g_{k-1 } } \ ] ] however , the composition on the left and the bottom is @xmath1758 , where @xmath1759 is the adjoint of @xmath1757 .
let @xmath1760 . by [ theor4.4 ] ,
@xmath1761 is homotopic to @xmath1757 .
this leads to a homotopy commutative diagram : @xmath1762\ar@{->}[d]_{\alpha_k}&p^{2np^k}(p^{r+k})\ar@{->}[r]^{p^{r+k-1}}\ar@{->}[d]_{\sigma\widetilde{\alpha}_k}&p^{2np^k}(p^{r+k})\ar@{->}[d]_{\sigma \widetilde{b}}\\
g_{k-1}\ar@{->}[r]^{\nu}&\sigma \omega g_{k-1}\ar@{->}[r]&\sigma \omega g_{k-1 } } \ ] ] taking cofibers vertically , we get a composition @xmath1763&\sigma(\omega g_{k-1}\cup_{\widetilde{\alpha}_k}cp^{2np^k-1}(p^{r+k}))\ar@{->}[r]&\sigma(\omega g_{k-1}\cup_{\widetilde{b}}cp^{2np^k-1}(p^{r+k } ) ) . } \ ] ] by hypothesis , @xmath327 extends to a map @xmath1764 ^ -<<<<{\overline{\alpha}}&z . } \ ] ] composing these maps together defines a map @xmath1765 @xmath771&\sigma(\omega g_{k-1}\cup_{\widetilde{b}}cp^{2np^k-1}(p^{r+k})\ar@{->}[r]^-<<<<{\sigma\overline{\alpha}})&\sigma z } \ ] ] whose restriction to @xmath653 is the composition @xmath1766^{\nu}&\sigma\omega
g_{k-1}\ar@{->}[r]^{\sigma\alpha}&\sigma z. } \ ] ] we now form the homotopy commutative diagram @xmath1767^{\omega \alpha''}&\omega\sigma z\ar@{->}[r]^{\mu}&z\\ \omega g_{k-1}\ar@{->}[r]^->>>>>>>{\omega \nu}\ar@{->}[u]&\omega\sigma \omega g_{k-1}\ar@{->}[r]^{\omega \epsilon}\ar@{->}[u]^{\omega\sigma\alpha}&\omega g_{k-1}\ar@{->}[u]^{\alpha } } \end{split}\ ] ] where the lower composition is homotopic to the identity .
the upper composition is an @xmath0-map extending @xmath327 .
we will modify this slightly to satisfy our requirements . in the diagram below , the left hand triangle is homotopy commutative due to [ eq7.16 ] and the left hand square follows from [ theor4.4 ] .
@xmath1768_{\widetilde{\beta}_k } \ar@{->}[r]^-<<<<{\sigma\vee \sigma \beta } & p^{2np^k-1}(p^{r+k-1})\!{}\vee{}\ ! p^{2np^k}(p^{r+k-1 } ) \ar@{->}[r]^-<<<<{-\delta_1\vee \rho } \ar@{->}[d]_{(\omega\pi_k)(\widetilde{\alpha(k)}\!{}\vee{}\!\widetilde{c(k ) } ) } & p^{2np^k}(p^{r+k } ) \ar@{-->}[dddl]^{\epsilon}\\ \omega g_{k-1 } \ar@{->}[rdd]_{\makebox[46pt]{$\alpha$ } } \ar@{->}[r ] & \omega g_k \ar@{->}[d]_{\omega\alpha''}&\\ & \omega\sigma z \ar@{->}[d]_{\mu}&\\ & z&\\ } \ ] ] by hypothesis @xmath1769 is null homotopic .
since the upper horizontal sequence is a cofibration sequence , there is an extension @xmath1770 extend @xmath189 to an @xmath0-map @xmath1771 and define @xmath1772 where @xmath1679 is the projection of @xmath72 onto @xmath669 .
since @xmath14 is homotopy - abelian and @xmath1773 is the difference between two @xmath0-maps , @xmath1773 is an @xmath0-map .
since the restriction of @xmath1774 to @xmath1127 is null homotopic , @xmath1773 extends @xmath327 by [ eq7.16 ] . from [ theor4.4 ]
we construct a homotopy commutative square : @xmath1775^{i}&\omega p^{2np^k+1}(p^{r+k})\\ p^{2np^k-1}(p^{r+k-1})\vee p^{2np^k}(p^{r+k-1})\ar@{->}[r]^-<<<<<<<<<<<<{(\omega\pi_k)(\widetilde{a(k)}\vee \widetilde{c(k)})}\ar@{->}[u]^{-\delta_1\vee \rho}&\omega g_k\ar@{->}[u]^{\omega \pi ' } } \ ] ] consequently , @xmath1776{}\alpha'(\omega \pi_k)(\widetilde{a(k)}\vee \widetilde{c(k)}&\sim(\mu\omega\alpha''-\epsilon'\omega\pi')(\omega\pi_k)(\widetilde{a(k)}\vee \widetilde{c(k)})\\ & \sim\epsilon ( \delta_1\vee \rho)-\epsilon'i(-\delta_1\vee \rho)\sim * \makebox[56pt]{}\qed\end{aligned}\ ] ] this completes the proof of [ theor7.3 ] in this section we will discuss various applications of the results developed in the previous sections .
[ prop7.27 ] suppose @xmath1777 .
then there is a natural exact sequence : @xmath1778&p^r\left[p^{2np}(p^{r+1}),z\right]\ar@{->}[r]^-<<<<{{\scalebox{.9}{${e}$}}}&[t , z]_h \ar@{->}[r]^->>>>{{\scalebox{.9}{${r}$}}}&\left[p^{2n}(p^r),z\right ] } \\ \xymatrix { \ar@{->}[r]^->>>>{{\scalebox{.9}{${\beta}$}}}&p^r\left[p^{2np-1}(p^{r+2}),z\right ] . }
\end{gathered}\ ] ] in particular , if @xmath1779 for @xmath1780 , @xmath23 is an isomorphism .
we are not asserting that @xmath133 is onto .
in fact , we have no examples where @xmath1461 .
this is immediate from [ theor7.2 ] and [ theor7.3 ] since @xmath34 and @xmath1781= 0 $ ] for all @xmath1782 .
[ cor7.28 ] suppose @xmath50 and @xmath1783 are two homotopy abelian anick spaces for the same values of @xmath22 , @xmath23 and @xmath24 .
then there is a homotopy equivalence via an @xmath0-map and the @xmath0-space expondent is @xmath4 .
we apply [ prop7.27 ] with @xmath1784 .
it suffices to show that @xmath1785 for @xmath1780 .
we apply the fibration sequence : @xmath1786&t'_{2n-1}\ar@{->}[r]^->>>>{{\scalebox{.9}{${e}$}}}&\omega t_{2n}(p^r)\ar@{->}[r]^-<<<{{\scalebox{.9}{${h}$}}}&bw_n } \ ] ] from @xcite . according to @xcite , @xmath1787 and according to @xcite @xmath1788 , so @xmath1789 . however , since @xmath33 , @xmath1790 when @xmath1791 , so @xmath1785 in this range .
thus @xmath1792_h\simeq\left[p^{2n}(p^r),t'\right]=z / p^r\ ] ] and the result follows . in particular , @xmath1793 for all @xmath218 as a consequence .
[ cor7.29 ] suppose @xmath1794 with @xmath1795 then there is a unique @xmath0-map @xmath27 such that the diagram : @xmath1796^{{\scalebox{.9}{${\widehat{\alpha}}$}}}&t_{2m-1}(p^s)\\ p^{2n}(p^r)\ar@{->}[r]^{{\scalebox{.9}{${\alpha}$}}}\ar@{->}[u]^{{\scalebox{.9}{${i}$}}}&p^{2m}(p^s ) \ar@{->}[u]^{{\scalebox{.9}{${i}$}}}}\ ] ] homotopy commutes . in case @xmath1797 , this result was the original motivation for these conjectures , leading to a secondary composition theory @xcite .
this follows from [ prop7.27 ] and [ cor7.28 ] .
[ prop7.30 ] there is an @xmath0-map @xmath1798 which induces a homomorphism of degree @xmath4 in @xmath1799 .
furthermore , the map @xmath299 in [ prop7.27 ] , evaluated on on @xmath1800 is the consideration : @xmath1796 ^ ->>>>{{\scalebox{.9}{${\theta_1}$}}}&t_{2np-1}(p^{r+1})\ar@{->}[r]^-<<<<{{\scalebox{.9}{${\widehat{f}}$}}}&z } \ ] ] where @xmath1801 is the unique extension of @xmath683 to an @xmath0-map . by [ cor7.28 ] , @xmath1802 , so we may apply [ prop7.27 ] with @xmath1803 and use naturality under @xmath1801 .
this leads to a commutative square @xmath18&p^r\left[p^{2np}(p^{r+1}),z\right]\ar@{->}[r]^{{\scalebox{.9}{${e}$}}}&[t , z]_h\\ 0\ar@{->}[r ] & p^r\left[p^{2np}(p^{r+1}),t_{2np-1}(p^{r+1})\right]\ar@{->}[r]^-<<<<{{\scalebox{.9}{${e}$ } } } \ar@{->}[u]_{{\scalebox{.9}{${\widehat{f}_*}$ } } } \ar@{}[d]^{\rotatebox{90}{$\approx$ } } & \left[t , t_{2np-1}(p^{r+1})\right]_h \ar@{->}[u]^{{\scalebox{.9}{${\widehat{f}_*}$}}}\\ & z / p & } \ ] ] the image of the generator in the lower left hand corner under @xmath299 is an @xmath0-map @xmath622 which is nonzero and of order @xmath38 . to evalueate @xmath1804 in cohomology ,
use the diagram : @xmath1796 ^ ->>>>>{{\scalebox{.9}{${\theta_1}$}}}&t_{2np-1}(p^{r+1})\\ \omega g_1\ar@{->}[r]^->>>>>>{{\scalebox{.9}{${p^r(\omega\pi')}$}}}\ar@{->}[u]^{{\scalebox{.9}{${h_1}$}}}&\omega p^{2np+1}(p^{r+1})\ar@{->}[u ] } \ ] ] based on the definition of @xmath299 .
note that a similar construction can be made in case @xmath1805 . in this case
@xmath1806_h\simeq g_k(z)\ ] ] and @xmath1807\ar@{->}[r]&g_k(z)}\]]can be evaluated on @xmath1808 as a composition : @xmath1809 ^ ->>>>{{\scalebox{.9}{${\theta_k}$}}}&t_{2np^k-1}\ar@{->}[r]^-<<<{{\scalebox{.9}{${\widehat{f}}$}}}&z } \ ] ] where @xmath1810 is an @xmath0-map of order @xmath38 inducing @xmath1149 in @xmath1211 . in section [ subsec1.5 ] , certain coefficient maps were labeled for use : @xmath1811 analogs of these maps were implicitly defined and used in section [ subsec5.2 ] : @xmath1812 and one can easily define @xmath133 as the compositions : @xmath1813 using [ cor7.29 ] , we can define @xmath1814 we apply [ prop7.27 ] to construct @xmath135 , but it is not unique in general @xmath1815 [ prop7.31 ] there is a split short exact sequence : @xmath1816&p_1(\omega t_{2n-1}(p^{r+1})\ar@{->}[r]&\left[t_{2n-1}(p^r),t_{2n-1}(p^{r+1})\right]_h\ar@{->}[r]^-<<<{{\scalebox{.9}{${e}$}}}&z / p^r\ar@{->}[r]&0 } \ ] ] and @xmath1817\oplus\left[p^{2np+2}(p^{r+1}),s^{2n+1}\right]\right\}\ ] ] @xmath1818 is known to be nonzero when @xmath4 divides @xmath22 ( @xcite ) and is known to be zero when @xmath1819 ( @xcite ) . in order to establish this exact sequence
we show that the map @xmath133 in fact is zero in this case . since @xmath1820 factors through @xmath1024 , by [ theor4.4 ] , the composition : @xmath1821^{{\scalebox{.9}{${\beta_1}$}}}&p^{2n+1}(p)\ar@{->}[r]^{{\scalebox{.9}{${\varphi_0}$}}}&s^{2n+1}\{p^r\ } } \ ] ] is null homotopic .
let @xmath828 be composition @xmath1822^{{\scalebox{.9}{${\varphi_0}$}}}&s^{2n+1}\{p^r\}\ar@{->}[r]^->>>>{{\scalebox{.9}{${\sigma}$}}}&s^{2n+1}\{p^{r+1}\}=t_{2n}(p^{r+1 } ) ; } \ ] ] then @xmath1823 is homotopic to the composition : @xmath1824 ^ ->>>>{{\scalebox{.9}{${\omega \rho}$}}}&\omega p^{2n+1}(p^{r+1})\ar@{->}[r]^{{\scalebox{.9}{${h_0}$}}}&t_{2n-1}(p^{r+1})\ar@{->}[r]^{{\scalebox{.9}{${e}$}}}&\omega t_{2n}(p^{r * } ) } \ ] ] since both compositions are @xmath0-maps which agree on @xmath9 . in the diagram below ,
the upper composition is null homotopic and the lower sequence is a fibration sequence : @xmath1825 ^ ->>>>{{\scalebox{.9}{${\widetilde{\beta_1}}$}}}\ar@{-->}[d]_{{\scalebox{.9}{${\xi}$}}}&\omega p^{2n+1}(p^r)\ar@{->}[r]^{{\scalebox{.9}{${\omega j}$}}}\ar@{->}[d]^{{\scalebox{.9}{${h_0\omega \rho}$}}}&\omega t_{2n}(p^r)\ar@{=}[d]\\ w_n\ar@{->}[r]&t_{2n-1}(p^{r+1})\ar@{->}[r]^{{\scalebox{.9}{${e}$}}}&\omega t_{2n}(p^{r+1 } ) . } \ ] ] it follows that the map @xmath316 exists forming a homotopy commutative square . but
@xmath1826=*,\ ] ] so the composition : @xmath1827^{{\scalebox{.9}{${\widetilde{\beta_1}}$}}}&\omega p^{2n+1}(p^r)\ar@{->}[r]^->>>>{{\scalebox{.9}{${\omega \rho}$}}}&\omega p^{2n+1}(p^{r+1})\ar@{->}[r]^{{\scalebox{.9}{${h_0}$}}}&t_{2n-1}(p^{r _ * } ) } \ ] ] is null homotopic .
however @xmath1828 generates @xmath1829_h\cong\left[p^{2n}(p^r),t_{2n-1}(p^{r+1})\right]\cong z / p^r\ ] ] and consequently @xmath1830 .
finally @xmath1831.\ ] ] but since @xmath1832=0=[p^{2np}(p^{r+1}),bw_n]$ ] , @xmath1833&=\left[p^{2np}(p^{r+1}),\omega t_{2n}(p^{r+1})\right]\\ & = \left[p^{2np+1}(p^{r+1}),s^{2n+1}\{p^{r+1}\}\right]\\ & = \left[p^{2np+1}(p^{r+1}),s^{2n+1}\right]\oplus\left[p^{2np+2}(p^{r+1}),s^{2n+1}\right],\end{aligned}\ ] ] so @xmath1834\oplus\left[p^{2np+2}(p^{r+1}),s^{2n+1}\right]\right\}.\ ] ] these groups are stable and trivial if @xmath1819 .
however , if @xmath4 divides @xmath22 , there is an element of @xmath1835 of order @xmath1836 and consequently @xmath1837 in this case .
the exact sequence is split since @xmath1838 has exponent @xmath4 . finally we note that for every choice of @xmath135 , @xmath1839 , @xmath1840 and @xmath1841 , as in section [ subsec1.5 ] .
[ prop7.32 ] there is a unique @xmath0-map @xmath1809 ^ ->>>>{{\scalebox{.9}{${f}$}}}&\omega^2p^{2n+2}(p^r ) } \ ] ] up to a unit whose double adjoint has a right homotopy inverse . by @xcite , @xmath1842
, so we apply [ prop7.27 ] .
we have @xmath1843&=p^r\left[p^{2np+2}(p^{r+1}),p^{2n+2}(p^r)\right]\\ p^r\left[p^{2np-1}(p^{r+1}),\omega^2p^{2n+1}(p^r)\right]&=p^r\left[p^{2np+1}(p^{r+1}),p^{2n+2}(p^r)\right]\end{aligned}\ ] ] however @xmath1844 for @xmath1845 .
since @xmath1846 is a retract of @xmath1847 , the double adjoint has a right homotopy inverse .
[ cor7.33 ] if @xmath14 is @xmath0-equivalent to the loop space on an @xmath0-space , every map @xmath1848 ^ ->>>{{\scalebox{.9}{${\alpha}$}}}&z}$ ] has an extension to an @xmath0-map @xmath1849^{{\scalebox{.9}{${\widehat{\alpha}}$}}}&z}$ ] . if @xmath1850 the adjoint of @xmath327 extends @xmath1851^{{\scalebox{.9}{${\alpha'}$}}}&\\ p^{2n+1}(p^r)\ar@{->}[r]^-<<{{\scalebox{.9}{${\widetilde{\alpha}}$}}}\ar@{->}[u]&w } \ ] ] and we construct @xmath1852 as the composition : @xmath1796&\omega^2p^{2n+2}(p^r)\ar@{->}[r]^-<<<{{\scalebox{.9}{${\omega \alpha'}$}}}&z } \ ] ] using [ prop7.32 ] .
in section [ subsec4.2 ] , we applied index @xmath38 approximation to reduce the obstructions to a homotopy - abelian @xmath0-space structure to a family of elements in the homotopy of @xmath79 with @xmath282 coefficients .
this reduction only works when @xmath569 .
we will use a different method in this case . however
, the material in sections [ subsec5.1 ] and [ subsec6.2 ] on @xmath103 , @xmath102 and @xmath112 does not depend on [ subsec4.2 ] , and we can still construct @xmath1853 ( see for example @xcite ) . let @xmath1854 be a generator and @xmath1855 .
let @xmath198 be the homotopy fiber of @xmath1856 .
then we have a homotopy commutative diagram of fibration sequences @xmath152&s^3\ar@{->}[r]^->>>>{\omega \kappa}&k(z;3)\\ s^3\{p^r\}\ar@{->}[r]\ar@{->}[u]^{\gamma}&s^3\ar@{->}[r]^{p^r}\ar@{=}[u]&s^3\ar@{->}[u]_{\omega e } } \ ] ] with @xmath1010 uniquely determined . using @xmath1010
, we construct a diagram of vertical fibration sequences : @xmath1857\ar@{->}[d]&\omega^2s^3\ar@{->}[r]\ar@{->}[d]&s^1\ar@{->}[d]\\ t_1\ar@{->}[r]\ar@{->}[d]&\omega s^3\{p^r\}\ar@{->}[d]\ar@{->}[r]^{\omega \gamma}&\omega^2x\ar@{->}[d]\\ \omega s^3\ar@{=}[r]&\omega s^3\ar@{=}[r]&\omega s^3 } \ ] ] since the upper horizontal composition is a homotopy equivalence , @xmath1858 .
note that the right hand fibration is an @xmath0-fibration and is an anick fibration . for any homotopy associative space @xmath50
, there is a map : @xmath1861 building the third stage of the classifying space construction ( @xcite,@xcite ) .
the mapping cone @xmath198 of this map has the cohomology of the bar construction on the homology of @xmath50 through dimension @xmath1862 .
in particular the @xmath526 cohomology of the @xmath1863 skeleton of @xmath198 has as a basis , classes @xmath1428 , @xmath1864 , @xmath1865 , @xmath1866 , @xmath1867 where @xmath1868 and @xmath1869 .
the @xmath1863 skeleton of the subspace @xmath1870 has cohomology generated by @xmath1428 , @xmath1864 , @xmath1865 , @xmath1866 .
now since the map @xmath114 is a retract of the map @xmath1871 , the @xmath1872 skeleton of @xmath1870 contains the @xmath1872 skeleton of @xmath1873 as a retract ( see the proof of [ cor6.15 ] ) . but
@xmath1874^{4n}=p^{2n+1}\cup_{x_2}cp^{4n}$ ] ; consequently @xmath1875 note that @xmath1876 generates @xmath1877 .
since @xmath1878 is inessential we can pinch the middle cells to a point after one suspension , and obtain a space with cell structure @xmath1879 with @xmath1880 .
however , @xmath1881 is decomposable unless @xmath1882 .
furthermore , the decomposition of @xmath1883 by secondary operations ( @xcite ) implies that if @xmath569 , we must have @xmath47 .
note that such a space for @xmath569 would imply that the `` @xmath1884 arf invariant class '' survives the adams spectral sequence .
this does happen when @xmath1885 with @xmath1886 , but not when @xmath1887 .
f. r. cohen , j. c. moore , and j. a. neisendorfer , _ decompositions of loop spaces and applications to exponents _ , algebraic topology , aarhus 1978 @xmath1888proc .
aarhus , aarhus , 1978@xmath1889 , lecture notes in math .
, 763 , springer , berlin , 1979 , pp . | anick spaces are closely connected with both ehp sequences and the study of torsion exponents .
in addition they refine the secondary suspension and enter unstable periodicity . in this work
we describe their @xmath0-space properties as well as universal properties .
techniques include a new kind on whitehead product defined for maps out of co - h spaces , calculations in an additive category that lies between the unstable category and the stable category , and a controlled version of the extension theorem of @xcite . |
the classical robinson - schensted algorithm establishes a bijection between permutations @xmath1 and ordered pairs of same - shape standard young tableaux of size @xmath2 .
this map has proven particularly well - suited to certain questions in the representation theory of both @xmath3 and the semisimple lie groups of type @xmath4 .
for instance , kazhdan - lusztig cells as well as the primitive spectra of semisimple lie algebras can be readily described in terms of images of this correspondence .
other sometimes more elementary representation - theoretic information requires more work to extract from standard young tableaux .
for instance , in independent work , a. reifegerste @xcite and j. sjstrand @xcite developed a method for reading the value of the sign representation of a permutation @xmath1 based on two tableaux statistics .
let @xmath1 and write @xmath5 for its image under the classical robinson - schensted map .
if we write @xmath6 for the number of squares in the even - indexed rows of @xmath7 , let @xmath8 be the sign of a tableau @xmath9 derived from its inversion number , and let @xmath10 be the usual sign representation on @xmath3 , then @xmath11 the focus of our paper is to extend this result to the complex reflection groups @xmath12 its two main ingredients generalize readily to this setting .
first , the robinson - schensted algorithm admits a straightforward extension mapping each @xmath13 to a same - shape pair of @xmath14-multitableaux , see r. stanley @xcite and l. iancu @xcite . at the same time , the sign of a permutation in @xmath3 extends to a family of @xmath14 one - dimensional representations of @xmath0 . after defining new spin and sign statistics on @xmath14-multitableaux ,
we prove the following extension of the result of a. reifegerste and j. sjstrand : let @xmath13 and write @xmath15 for its image under the generalized robinson - schensted map . given a primitive @xmath16 root of unity @xmath17 and the associated family @xmath18 of representations of @xmath0 , we have @xmath19 where @xmath20 is the total sum of the lengths of the even - indexed rows of the component tableaux of @xmath21 .
to prove this , we define a particular set of equivalence classes on @xmath22 and show that the formula either holds or fails for all members of a given class . by construction , each class contains a representative for which the theorem is easy to verify by a direct appeal to the original symmetric group formula , thereby completing the proof . for the classical weyl groups , all of which appear among the above complex reflection groups , the mbius function of the bruhat order can be expressed in terms of the values of the sign representation @xcite .
its use is ubiquitous in kazhdan - lusztig theory , and the above formulas allow the values of the mbius function to be read off from the images of the appropriate robinson - schensted map .
this is especially relevant in light of the tableaux classification of kazhdan - lusztig cells . in type @xmath4 ,
left kazhdan - lusztig cells consist of those permutations whose recording tableaux agree , see a. joseph @xcite or s. ariki @xcite .
a similar result holds for the so - called asymptotic left cells in the iwahori - hecke algebras in type @xmath23 , this time in terms of @xmath24-multitableaux @xcite .
after defining the family of complex reflection groups @xmath0 and describing their one - dimensional representations , we define multipartitions , a generalization of the robinson - schensted algorithm , and tableaux statistics that we will use to describe these representations
. consider positive integers @xmath14 , @xmath25 , and @xmath2 with @xmath25 dividing @xmath14 and let @xmath26 .
we define the complex reflection groups @xmath0 as subgroups of @xmath27 consisting of matrices such that * the entries are either @xmath28 or powers of @xmath17 , * there is exactly one non - zero entry in each row and column , * the @xmath29-th power of the product of all non - zero entries is @xmath30 .
together with thirty - four exceptional groups , the groups @xmath0 account for all finite groups generated by complex reflections @xcite , and include among them all the classical weyl groups . in our work
the parameter @xmath14 will generally be fixed allowing us to write simply @xmath31 for the group @xmath32 in order to establish succinct notation , we will write @xmath33\ ] ] for the matrix whose non - zero entry in the @xmath34th column is @xmath35 and appears in row @xmath36 . utilizing this notation ,
define the set @xmath37 where @xmath38 \text { , and } \\ s_i & = [ 1,2 , \ldots , i-1 , i+1 , i , i+2 , \ldots , n].\end{aligned}\ ] ] further , let @xmath39 the set @xmath40 generates @xmath31 with presentation given as @xmath41 $ } \rangle.\ ] ] subject to similar relations , @xmath42 generates a subgroup @xmath0 of @xmath31 of index @xmath25 , see s. ariki @xcite .
there are exactly @xmath43 one - dimensional representations of @xmath31 ; they divide naturally into two families .
[ def : one - diml - representations ] for each integer @xmath34 between @xmath28 and @xmath44 , we define representations @xmath36 and @xmath45 of @xmath31 by specifying their values on the generating set @xmath40 .
let @xmath46 and define @xmath47 and @xmath48 .
each becomes a representation of the subgroup @xmath0 by restriction .
we write a partition @xmath49 of an integer @xmath50 as a nonincreasing sequence of positive integers @xmath51 and define its rank as @xmath52 .
young diagram _
@xmath53 $ ] of @xmath49 is a left - justified array of boxes containing @xmath54 boxes in its @xmath34th row .
the shape of a young diagram will refer to its underlying partition . with the integer @xmath14 fixed , a _ multipartition of rank @xmath2 _ is an @xmath14-tuple @xmath55 of partitions the sum of whose individual ranks equals @xmath2 . the _ young diagram _
$ ] of @xmath57 is the @xmath14-tuple @xmath58 , \ldots , [ \lambda^{r-1}])$ ] .
we refer to @xmath57 as the _ shape _ of the diagram @xmath56 $ ] and define @xmath59 .
we will follow a convention of denoting objects derived from multipartitions in boldface while writing those derived from single partitions using a normal weight font . a _ standard young tableaux of shape _
@xmath57 is the young diagram @xmath56 $ ] of rank @xmath2 together with a labeling of each of its boxes of with the elements of @xmath60 in such a way that each number is used exactly once , and the labels of the boxes within each component young diagram @xmath61 $ ] increase along its rows and down its columns . remembering that @xmath14 is fixed , we will write @xmath62 for the set of all standard young tableaux of rank @xmath2 whose shape is a multipartition with @xmath14 components . a tableau @xmath63 will be called _ ascending _ if for each @xmath64 , the elements of the set of labels of the component tableau @xmath65 are pairwise smaller than the elements of the set of labels of @xmath66 .
take @xmath67 .
the following standard young tableau @xmath68 has shape @xmath69 , is of rank @xmath70 , and is ascending : @xmath71 following @xcite and @xcite , we define a map from @xmath31 to same - shape pairs of @xmath14-tuples of standard young tableaux .
consider @xmath72 \in w_n$ ] and define ordered sets @xmath73 for @xmath74 .
let @xmath75 be the image of the sequence @xmath76 under the usual robinson - schensted map , labeling squares of @xmath77 according to the relative positions of @xmath78 within @xmath79 , and define @xmath80 the multitableaux robinson - schensted map is defined by @xmath81 it maps @xmath31 onto the set of same - shape pairs of elements of @xmath62 and is in fact a bijection .
our goal is to describe values of the sign representations on @xmath31 under the above generalization of the robinson - schensted map .
to do so , we rely on a few statistics that can be readily computed from multitableaux .
an _ inversion _ in a young tableau @xmath9 is a pair @xmath82 with @xmath83 for which the box labeled by @xmath34 is contained in a row strictly below the box labeled @xmath84 .
let @xmath85 be the set of inversions in @xmath9 , and write @xmath86 for its cardinality .
if @xmath87 is a multitableau , we extend this notion and define : @xmath88 where @xmath89 we will be mainly interested in the parity of the size of this set and define @xmath90 for a young tableau @xmath9 , write @xmath91 for the total number of boxes in its rows of even index . for a multitableau @xmath87 , we write @xmath92 for the shape of the young diagram underlying @xmath93 and define the statistics @xmath6 and @xmath94 as follows : @xmath95 the @xmath94 statistic provides a simple description of the image of the subgroup @xmath0 under the @xmath14-multitableaux robinson - schensted map
. @xmath96 if and only if @xmath97 write @xmath98 \in g(r , p , n)$ ] and for each @xmath99 , set @xmath100 . if @xmath101 , then @xmath102 , and hence it follows that @xmath103 equals @xmath104 . since @xmath105 , we conclude that @xmath106 finally , since @xmath13 if and only if @xmath107 , it follows that @xmath13 if and only if @xmath108 as claimed .
we define a sequence of functions on @xmath31 . in the next section
we will show that they coincide with the sign representations on @xmath31 .
again , for @xmath109 , let @xmath101 .
for @xmath110 , we will write @xmath111 [ running_example ] consider @xmath112 $ ] in @xmath113 recalling the notation in section [ subsection : multitableaux ] , we have @xmath114 , @xmath115 , @xmath116 , and @xmath117 .
further , @xmath118 from these we construct the robinson - schensted image of @xmath79 : @xmath119 we read off @xmath120 , @xmath121 , @xmath122 , and @xmath123 .
hence @xmath124 which coincides with @xmath125 .
the aim of this section is to verify the formulas for the one - dimensional representations of @xmath31 given above .
we reduce the problem to the setting of the classical robinson - schensted map by considering a particular equivalence relation on the elements of @xmath31 .
recall the notation from section [ subsection : multitableaux ] and write @xmath109 in one - line notation as @xmath126.\ ] ] we will say that @xmath79 is _ ascending _ if the exponents @xmath127 increase weakly and for every integer @xmath128 all elements of @xmath129 are smaller than the elements of @xmath130 .
the following fact is immediate : we proceed to define equivalence classes on @xmath31 , each with an ascending representative . adopting the notation from the definition above ,
consider @xmath133 and define @xmath134 whenever @xmath135 and @xmath136 whenever @xmath137 , @xmath138 , and @xmath139 .
we will call operators of this form _ left _ and _ right admissible _ , respectively .
to define an equivalence relation , we let @xmath140 and @xmath141 for every @xmath34 and take its transitive closure .
a right admissible operator permutes adjacent entries of @xmath79 , while each left admissible operator exchanges some @xmath142 and @xmath143 which have different exponents in @xmath79 and satisfy @xmath144 while preserving the exponents in each position . thus by successive application of right admissible operators , the entries of @xmath79 can be arranged to have weakly increasing exponents .
furthermore , by successive application of left admissible operators , the @xmath142 can be permuted so that for every integer @xmath128 all entries with coefficient @xmath145 are smaller than those with coefficient @xmath146 .
we have shown the following : applying a right admissible transformation @xmath150 to @xmath79 does not change the relative order of the entries sharing the coefficient @xmath151 .
thus as ordered sets , we have @xmath152 for all @xmath153 and the first part of the proposition follows .
suppose that @xmath142 and @xmath143 have different exponents in @xmath79 and satisfy @xmath144 .
exchanging the two while preserving the exponents in each position preserves the order new boxes are appended in constructing the component tableaux of @xmath154 , thus a left admissible operation fixes @xmath155 , as claimed .
the multitableau @xmath156 is obtained from @xmath155 by exchanging boxes with labels @xmath34 and @xmath157 .
the result is standard as the boxes labeled @xmath34 and @xmath157 lie in different component tableaux of @xmath155 .
furthermore , @xmath158 equals @xmath159 with @xmath34 and @xmath157 interchanged and the inversion @xmath160 either added or removed . in particular , a right admissible transformation changes @xmath161 by one while fixing @xmath162 for each @xmath153 .
similar reasoning can be applied to @xmath163 , thereby proving : [ proposition : inversions ] a right admissible transformation changes @xmath161 by one while fixing @xmath162 for each @xmath153 , while a left admissible transformation changes @xmath164 by one while fixing @xmath165 for each @xmath153 .
first note that @xmath167 for @xmath168 and consequently for each @xmath34 , @xmath169 since @xmath6 and @xmath94 are shape - based statistics and admissible transformations preserve the shapes of the underlying tableaux , the only possible difference between @xmath170 and @xmath171 lies in the signs of their left and right tableaux .
this is addressed explicitly in proposition [ proposition : inversions ] .
we obtain @xmath172 and the lemma follows .
[ main_theorem ] let @xmath109 and write @xmath173 for its image under the generalized robinson - schensted map .
given a primitive @xmath16 root of unity @xmath17 and the associated family @xmath174 of representations of @xmath31 , we have @xmath175 by lemma [ lemma : ascending ] , it is enough to verify the above formula for ascending elements of @xmath31 . consequently , consider an ascending @xmath109 , writing it in one - line notation as @xmath176.$ ] let @xmath177 and @xmath178 be the component tableaux of @xmath179 and @xmath180 .
since the latter tableaux are both ascending and of the same shape , for each @xmath153 , @xmath181 and @xmath77 are labeled by the same set of integers , say @xmath182 .
let @xmath183 be the group of permutations of @xmath182 and define @xmath184 for the natural extension of the classical robinson - schensted map to the ordered set @xmath182 .
let @xmath185 if @xmath186 . also let @xmath187 \in w_n$ ] .
note that we can view @xmath183 as a subgroup of @xmath188 for each @xmath153 . from this perspective , since @xmath179 and @xmath180 are ascending , we have @xmath189 and consequently @xmath190 where @xmath10 is the usual sign of a permutation .
we would like to reduce this expression to the formula proposed by the theorem .
the first step invokes the original result of a. reifegerste @xcite and j. sjstrand @xcite .
for @xmath191 , we have @xmath192 where @xmath193 .
furthermore , @xmath194 is easily computed to be @xmath195 and consequently we have : @xmath196 to further simplify the above expression we note that @xmath197 , by definition .
furthermore , if @xmath198 is the cardinality of the set of entries of @xmath79 with exponent equal to @xmath153 , then @xmath199 finally , since both @xmath179 and @xmath180 are ascending tableaux , the sets of inversions @xmath200 and @xmath201 are both empty for @xmath202 .
consequently , @xmath203 and @xmath204 are just the sums of the numbers of inversions in their component tableaux .
the theorem now holds by direct substitution . continuing with @xmath79 from example [ running_example ]
, we have the ascending representative @xmath205 $ ] .
its robinson - schensted image is : @xmath206 in contrast to the behavior of the robinson - schensted map in the setting of classical weyl groups , the fact that @xmath21 and @xmath207 coincide does not necessarily imply that the corresponding group element is an involution , as occurs here .
we decompose @xmath147 as @xmath208 where @xmath209 and : @xmath210 \in s_{\{1,2,3,4\ } } & u_1 & = [ 6,5 ] \in s_{\{5,6\}}\\ & u_2 = [ 7,8 ] \in s_{\{7,8\ } } & u_4 & = [ 1 , 2 , 3 , 4 , \zeta^1 \ , 5 , \zeta^1 \ , 6 , \zeta^2 \ , 7 , \zeta^2 \ , 8 ] \in w_8.\end{aligned}\ ] ] since @xmath211 , @xmath212 , and @xmath213 , we compute @xmath214 . furthermore , since @xmath215 , @xmath216 , and @xmath217 , we compute @xmath218 , which agrees with @xmath219 , as expected .
it is also possible to approach the present problem by examining the coplactic classes on @xmath31 defined by the robinson - schensted map instead of the equivalence classes used here . in lieu of an ascending class representative
, one can use the unique element in each coplactic class whose left multitableau has no inversions . using the generalized knuth relations for @xmath31 described in @xcite
, one can then produce the same formula by following their action on @xmath45 , albeit with a little more work .
the advantage of the equivalence classes described in this paper is that they reduce the problem to the symmetric group setting as completely as possible .
a. joseph . a characteristic variety for the primitive spectrum of a semisimple lie algebra .
short version in : _ non - commutative harmonic analysis _ ( actes colloq . , marseille - luminy , 1976 ) , , 587:102118 , spinger , berlin , 1977 . | we present a formula for the values of the sign representations of the complex reflection groups @xmath0 in terms of its image under a generalized robinson - schensted algorithm . |
observations and modelling of the circumnuclear regions ( cnrs ) of barred galaxies give important clues about the nature of active galactic nuclei ( agn ) , ( circum)nuclear starbursts and the gas dynamics in the central kpc ( e.g. knapen et al .
1995a , b ; buta & combes 1996 ; elmegreen et al .
1997 ; quillen et al . 1999
; regan & mulchaey 1999 ; see also reviews by knapen 1999 and shlosman 1999 ) .
the presence of non - axisymmetries , such as oval distortions or bars , in the gravitational potential of a disc galaxy can lead to the infall of gaseous material , from galactocentric radii of a few kpc , into the central kpc .
the nonaxisymmetry facilitates the removal of angular momentum from the gas , and directs it into the cnr or possibly the nucleus ( shlosman et al .
1990 ; athanassoula 1992 ; phinney 1994 ) .
this gas fuelling process is expected to induce morphological signatures in the bars and cnrs , which are useful tools in the understanding of the dynamics and kinematics of galaxies .
dust lanes and nuclear rings appear to be the dominant circumnuclear features . [
cols="^,^,>,>,>,^,^,^,^ " , ] dust lanes have been interpreted as the location of shocks in the gas flow .
thus , the morphology of the dust distribution can reveal characteristics of the dynamics of these galaxies ( athanassoula 1992 ) and of the kinematics of the gas . according to athanassoula s models ,
the degree of curvature of the dust lanes is a direct indicator of the strength of the bar for bars which are not very strong .
nuclear rings , which are usually sites of active star formation ( sf ) , occur at the location of strong density enhancements in the gas , where the bar - driven inflow of gas slows down in the vicinity of inner lindblad resonances ( ilrs ; athanassoula 1992 ; heller & shlosman 1994 ; buta & combes 1996 ; shlosman 1999 ) .
sf may result from the gas becoming gravitationally unstable ( elmegreen 1994 , 1997 ) , or from triggering in miniature spiral arms ( knapen et al .
1995b , 2000 ) .
whereas imaging in the blue , ultraviolet , or in spectral lines like h@xmath2 traces the massive sf ( e.g. sersic & pastoriza 1967 ; pogge 1989a , b ; knapen et al . 1995a ; maoz et al . 1996 ; colina et al .
1997 ) , near - infrared ( nir ) imaging offers very important advantages .
firstly , the presence of substantial and mostly unquantified amounts of extinguishing ( absorbing and scattering ) dust in the cnrs hampers the interpretation of imaging at especially the uv and optical wavelengths .
nir , particularly @xmath3-band , emission is much less susceptible to extinction by dust ( a factor of 10 when comparing @xmath3 to @xmath4 ) , whereas colour index images such as @xmath5 ( knapen et al .
1995a ) or @xmath6 are clear morphological dust indicators due to the relatively modest changes in those colours caused by different stellar populations .
secondly , nir observations are of particular importance to the study of the cnrs because they allow a better determination of the mass distribution ( excluding dark matter ) .
nir imaging is more sensitive to light primarily from cool giants and dwarfs which dominate the mass or are at least directly proportional to it .
however , as shown in the literature and by us in this paper , there can be significant , or even dominant , contributions from young stars to the nir emission even in the @xmath3-band , so for objects with strong sf any estimate of the mass distribution made from nir maps must take the varying @xmath7 ratio into account ( knapen et al .
1995a , b ; elmegreen et al .
1997 ; wada , sakamoto & minezaki 1998 ; ryder & knapen 1999 ) .
the atlas of images presented in this paper ( paper i ) represents the first results of a programme which studies the circumnuclear structures that appear at small scales ( a few hundred parsecs to a kpc ) and their connection with the global disc structure in a sample of 12 barred spiral galaxies .
we present @xmath8 and @xmath3-images at subarcsec resolution for all 12 sample galaxies , and _ hubble space telescope _
( _ hst _ ) archive @xmath9-band images for 10 of them . in paper ii ( prez ramrez & knapen 2000a ) , we present an accompanying optical data set of broad - band and h@xmath2 images of the complete discs of all our sample galaxies , while in paper iii ( prez ramrez & knapen 2000b ) the morphological information is interpreted in terms of the structure and evolution of cnrs of barred galaxies . in section 2
, we describe the observations and data reduction techniques .
the imaging data are presented as a series of multi - panel figures , which show broad - band images , colour index images highlighting dust and sf features against the dominant stellar luminosity , and radial profiles of ellipticity , position angle , surface brightness and colour .
the results are summarized in section 3 for individual galaxies , and briefly discussed in section 4 , while concluding remarks are given in section 5 .
the main selection criterion for our sample is the presence of a bar and evidence of some circumnuclear structure associated with it , such as rings , nuclear bars or regions of sf .
furthermore , the sample galaxies should be nearby , bright and observable from the northern hemisphere .
some of the sample galaxies are taken from the lists published by sersic & pastoriza ( 1967 ) and pogge ( 1989a , b ) .
the sample can be considered more anecdotal than complete in any sense .
however , it does significantly increase the sample size when compared to published nir studies of cnrs in barred galaxies , which rarely include more than three objects and usually discuss only one ( e.g. knapen et al .
1995a , b ; elmegreen et al .
1997 ; laine et al .
1999 ; ryder & knapen 1999 ) . with our new imaging , we also improve the spatial resolution , and show different colour index maps , @xmath10 ( primarily outlining dust extinction ) and @xmath11 ( additionally indicating the possible emission due to hot dust in the case of the more active sample galaxies ) .
regan & mulchaey ( 1999 ) show _ hst _
1.6@xmath12 m and optical - infrared ( basically @xmath13 ) colour index images of a sample of seyfert ( sy ) galaxies , including a number of cnr galaxies .
two of the galaxies in their sample are also included in our sample ( ngc 3516 and ngc 3982 ) .
the @xmath13 colour , however , is much more sensitive to changes in stellar populations than the nir colours used in our paper . for 10 of our sample galaxies ,
we have retrieved , in addition to our own data , @xmath9-band images from the _ hst _ archive ( fig . 1 ) , and we comment on the high - resolution nir morphology of those objects . in table 1
we give some information about the classification of the galaxies of our sample , the predominant circumnuclear feature , as determined from the data , and the type of nuclear activity , if known .
we have obtained nir broad - band images in the @xmath14 ( 1.25@xmath12 m ) , @xmath9 ( 1.65 @xmath12 m ) and @xmath3 ( 2.2 @xmath12 m ) bands of the 12 barred galaxies that make up our sample , during three observing runs : 1994
september 26 , 1995 november 56 , and 1996 february 35 .
the observations were made with the 3.6 m canada france hawaii telescope ( _ cfht _ ) .
we used the montreal nir camera ( monica ; nadeau et al .
1994 ) , equipped with a 256 x 256 pixel hgcdte array detector with a projected pixel size of 0.248 .
since the sky background in the nir is bright and changes rapidly , it is essential to obtain sky frames frequently .
we observed the sky usually just before and after each galaxy observation .
frames typically consisted of four co - added individual exposures .
separate frames with the nucleus of the galaxy at slightly offset positions were co - added in the reduction process to produce the final mosaic images .
the weather was clear and generally photometric thoughout these nights , and the seeing values as measured from our final images were 0.7 1.0 .
the main steps in the reduction of the nir data include subtracting the sky background , interpolating across known bad pixels , and registering and combining the images .
the data reduction was done partly with private programmes and partly with standard iraf routines .
we first combined four sky exposures taken at different positions on the sky immediately before and after a series of galaxy exposures by median - averaging them , while iteratively rejecting those values in the averaged sky exposure that deviate by more than 3@xmath15 from the average value in that frame .
this procedure allows the automated rejection of any star images that might have been present in the sky exposures , which were observed by offsetting the telescope blindly to a position a few arcmin from the galaxy centre .
we then subtracted the averaged sky image from the relevant galaxy images .
a close inspection of the sky - subtracted galaxy images of the 1996 run showed the presence of periodic , horizontal lines which could be traced back to electronic crosstalk in the system .
this raises the noise level in these images . in order to correct the problem
, we used a fourier transform technique to locate the frequencies of the maximum intensity in the images , and filtered them out .
this resulted in a considerable improvement in most images , although in a few images interference stripes are still notable at low levels .
after flat - fielding with dome flats , and masking the unreliable pixels in the array , we combined the sky - subtracted and , where appropriate , de - striped images .
ideally , one would like to locate one or more common field stars in the images to achieve the most accurate alignment possible .
unfortunately , the field of view of our images is small and in most cases we could only use the nucleus of the galaxy itself to determine the reference position .
the individual images were thus registered and shifted to a common position ( in all cases accurate to a fraction of the pixel size ) and averaged . in the process
, the pixel size was halved to increase sampling ( reduced from 0@xmath16 to 0@xmath17 ) . finally , we rotated the images by an angle of 88.8@xmath18 in order to obtain the correct north ( up ) , east ( left ) orientation .
although we subtracted the background sky as described in section 2.3 , we can not check on how accurate our sky subtraction is since the sky is variable , and the frame is small compared with the galaxy .
this makes it very hard to estimate to what level the flux registered in the photometric aperture is influenced by any residual background emission .
for that reason , we have used aperture photometry from the literature . by calibrating with multiple aperture measurements we can better estimate the value of the sky and the efficiency of the system .
if only one useful aperture is available , we can use only the relatively bright part of the image .
we were able to find the aperture photometry in @xmath14 , @xmath9 and @xmath3-bands for all but one of the galaxies , ngc 1530 .
for this galaxy , we used the average magnitude offset as determined for the rest of the galaxies .
literature photometry sources include glass ( 1976 ) , aaronson ( 1977 ) , mcalary , mclaren & crabtree ( 1979 ) , balzano & weedman ( 1981 ) , willner et al .
( 1985 ) and spinoglio et al .
( 1995 ) . to check the quality of our photometry
, we also made comparisons with data from the literature .
we selected values from glass ( 1984 ) , cidziel , wynn - williams & becklin ( 1985 ) , devereux , becklin & scoville ( 1987 ) and hunt et al .
the average absolute difference between our results and the results from different authors was 0.12 mag .
direct comparison of the surface brightness profile in the @xmath9-band and the @xmath11 colour profile plots was possible for two of our sample galaxies , ngc 3516 and ngc 3982 , which are in common with the peletier et al .
( 1999 ) sample .
the photometric calibration procedures used were slightly different , but the agreement in both cases was good .
colour index maps were created after assuring that the relevant pairs of images were at the same pixel scale and orientation and at comparable resolution before combining them .
after calibrating the images , we used galphot ( jrgensen , franx & kjaergaard 1992 ) to fit ellipses to the images .
the centre position , ellipticity and the position angle of the fitted ellipses were allowed to change freely as a function of radius for the @xmath9 passband .
the values for these parameters were then used as input for the fits to the images in other passbands , in order to ensure that we produce reliable colour profiles . in the case of ngc 2903 ,
the position angle and ellipticity were kept constant because the large amount of structure present in the core did not allow a meaningful fit to these parameters .
we thus produced plots ( shown in fig .
2 ) of the surface brightness of the @xmath9-band as a function of radius , and colour profiles in @xmath10 and @xmath11 for all the galaxies which were imaged in these bands .
errorbars in these plots indicate the uncertainty in the determination of the sky background . we have estimated these uncertainties using the procedure described by peletier et al .
( 1999 ) , where 1@xmath15 errorbars correspond to @xmath19= 21.5 mag arcsec@xmath20 , @xmath21= 21.2 mag arcsec@xmath20 and @xmath22= 20.5 mag arcsec@xmath20 ( about 0.1% of the sky background in @xmath9 and @xmath3 and 0.3 % in @xmath14 ) . in fig . 2
, we show for each galaxy ( except ngc 3351 for which we only obtained a @xmath3-band image ) greyscale representations of the @xmath3 broad - band and @xmath10 and @xmath11 colour index images , radial profiles of position angle and ellipticity as determined from the fit to the @xmath9-band image ( @xmath3 for ngc 3351 ) , and radial surface brightness ( @xmath9 ) and colour ( @xmath10 and @xmath11 ) profiles .
we tabulated the values plotted as a function of radius , but publish the tables in electronic form only .
they are also available from the authors . for 10 of our 12 sample galaxies ,
nicmos imaging with the _ hst _ is available from the _ hst _ archive .
we retrieved the re - reduced f160w ( comparable to @xmath9-band ) images from the archive .
however , we improved the quality of some of these images by doing additional data reduction to remove artifacts .
we relied on header information to place the images on an astrometrically correct grid .
the images are all taken with the nic2 camera , with a pixel size of 0.075 . in most cases ,
the image retrieved from the archive is a combination of several individual exposures . for ngc 3516 and ngc 3982 , however , one single exposure was available , and these images in fact improved most due to our additional data reduction .
regan & mulchaey ( 1999 ) published two of these images ( ngc 3516 and ngc 3982 ) .
we show the central areas of all images in fig . 1
, with the same scale and orientation . in most of the objects , a wealth of detail can be seen in the cnr .
the emission usually coincides with the location of the sf ring , and most of the individual bright knots are due to regions of current sf .
dust lanes and/or sf regions often outline spiral - like patterns , which will be discussed in more detail below , and in paper iii .
in this section , we describe some of the results of the nir imaging of our sample galaxies , as shown in detail in fig . 1 ( _ hst _ nir images ) and fig . 2 ( ground based multi - band images and profile fits ) . a more systematic study of parameters derived from these data in combination with optical imaging of the complete galaxy discs is forthcoming ( papers ii and iii ) .
our broad - band nir images are remarkably smooth , and do not show any structure in the cnr ( fig .
the colour index images , however , show a red ring - like structure , possibly outlining a single spiral arm that departs from the nucleus towards the west side of the image and continues to wrap around the nucleus until it closes in a ring .
this ring appears more continuous and broader in the northern part than in the south where it looks patchy .
the feature seen in our nir colour index map can not be an artifact of combining two images with slightly different spatial resolution , because such an artifact would show up as a complete ring , whereas the observed red feature is not at constant galactocentric radius .
pogge ( 1989a ) saw an incomplete nuclear ring in emission with a number of distinct `` hot spots '' .
the _ hst _ @xmath9-band image resolves the nuclear ring into a series of tightly wound spiral armlets , outlined in dust and stellar populations .
the ring region is of a relatively low amplitude compared to the central bulge component , in contrast to nuclear rings in other objects , e.g. ngc 3351 , as can be observed in fig . 1 .
the location of the ring corresponds to a bump in all radial profiles ( i.e. , surface brightness , colour , ellipticity and position angle ) .
the @xmath10 colour of the ring is redder by 0.1 magnitudes than the background .
no isophotal twists are seen in the ellipticity and position angle profiles or in the contour map of the @xmath3-band image .
thus we confirm that there is no evidence for a nuclear bar ( regan & elmegreen 1997 ) .
we find very well - defined mini - spiral structure in the central region of this galaxy ( fig .
reynaud & downes ( 1997 ; see also reynaud & downes 1998 , 1999 ) suggest that the molecular gas distribution in the central 5 kpc of this galaxy is concentrated along arc - like shock features .
our @xmath10 image shows that these arcs are part of a well - defined nuclear spiral structure .
furthermore , reynaud & downes suggest that a patchy molecular ring may lie inside these shock fronts , possibly connected to them .
2b reveals that the spiral structure forms a pseudoring around the nucleus .
even the broad - band image shows considerable structure in the cnr .
no isophote twists can be seen in the position angle plot .
distinct regions , presumably of enhanced sf , can be noticed along the mini - spiral in the broad - band nir images .
these regions coincide in position with the dark ( red ) lanes seen to outline the mini - spiral in the @xmath10 image .
this implies that the mini - spiral in @xmath10 is not necessarily a dust spiral ( as would have been tempting to conclude from the colour index image alone ) but may well be delineated by either emission from young stars ( e.g. red supergiants ) , or from hot dust .
we conclude that we see a mini - spiral in the nir , outlined by emission from sf and dust .
this mini - spiral is well traced also on the _ hst _ @xmath9-band image ( fig .
1 ) as a collection of distinct luminous regions outlining the armlets , which are accompanied by dust lanes . like in the case of ngc 1300 , a bump in the radial profiles is visible at the location of the ring .
the @xmath10 colour of the ring is 0.05 magnitudes redder than the background colour .
no pronounced change of slope is detected in the @xmath9-band surface brightness profile , which overall decreases more smoothly than in the other galaxies discussed here .
2c shows several peaks of sf in the cnr of this starburst galaxy , which , as expected , are resolved in much more detail in the _ hst _ nicmos image we obtained from the _ hst _ archive ( fig . 1 ) .
these peculiar ` hot spots ' in the nuclear region have been identified and described in different ways by various authors .
marcelin , boulesteix & georgelin ( 1983 ) found six ` hot spots ' forming a linear structure that crosses the central region . a later study in infrared and radio by wynn
williams & becklin ( 1985 ) showed that bursts of sf were not confined to the visible hot spots .
simons et al . ( 1988 ) noticed the presence of organized dust structure in the cnr using nir and optical imaging .
we see patches of dust in our images , but no coherent structure .
regan & elmegreen ( 1997 ) found that this galaxy has an isophotal twist , using a @xmath3-band image of the inner bar region , although the results from an earlier study by elmegreen et al .
( 1996 ) were ambiguous .
similar twists have been interpreted as nuclear bars or triaxial structures ( e.g. , shaw et al .
1995 ; friedli & martinet 1993 ; wozniak et al .
however , our images show such an abundance of structure that it is hard to imagine that the measured changes in the radial behaviour of ellipticity or pa would have any significance in this respect . for that reason
, we set the ellipticity and the pa to fixed values corresponding to the parameters of the isophotes on the outskirts of our images when we fit our radial colour profiles .
as in , e.g. , ngc 1300 , the circumnuclear ring - like region of enhanced sf shows up in the surface brightness profile as a change of slope , and in the colour profiles as a significant red bump at the same radius .
our colour index maps suggest that the nuclear hot spots may be part of a pseudoring .
we can identify more than six clumps in the broad - band images , and even more in the colour index map .
this is in good agreement with the idea that obscuration by dust is the reason for a previous underestimate of the extent of this sf region ( cf .
jackson et al .
1991 ) . as in ngc 1530
, we see significant emission from sf regions in the nir , including the @xmath3-band , which must at least in part be due to red supergiant emission .
the red @xmath10 and @xmath11 colours of these regions makes them stand out clearly in the colour index maps .
we only obtained a @xmath3-band image ( fig .
2d ) , which shows an incomplete ring of presumably sf clumps .
the _ hst _ @xmath9-band image ( fig .
1 ) shows the same ring but at significantly higher resolution , and resolved into a large number of individual luminous regions , presumably star - forming .
some of the brightest knots can be recognised in the uv ( @xmath23 ) _ hst _ image published by colina et al .
( 1997 ) , but others , e.g. the one north of the nucleus , are absent from the uv image .
dust extinction is the most likely candidate mechanism for this absence , although stellar population differences can be envisaged as culprits as well .
spiral arm structure is not obvious in the cnr , and if it is present at all it is very tightly wound .
there is only little evidence of dust organised into lanes , and the best example of such a dust lane is seen towards the south - east in the _ hst _ image . in contrast , and as a direct result of their much lower spatial resolution , shaw et al .
( 1995 ) , found evidence for a circumnuclear ring only in their @xmath10 colour map .
as in the other galaxies discussed before in this section , there is a dichotomy in the slope of the @xmath3-band surface brightness profile .
the change occurs at the radius of the sf `` ring '' , and can be explained as the transition between an inner active sf region and the quiescent disc around it .
pa and ellipticity are difficult to fit in that area , but their radial profiles show consistent behaviour in- and outside the nuclear ring .
our nir imaging of this barred galaxy reveals a lot of structure , including a double peak in the central part of the nir broad - band images ( fig .
recent adaptive optics images obtained with the _ cfht _ ( by f. combes & j.h .
knapen , private communication ) confirm this double peak , but spectroscopic follow - up observations are needed to confirm whether these peaks are in fact two nuclei .
nir imaging is available .
the colour index images show a pair of long and straight dust lanes that come into the cnr through the main bar , and intricate dust lane structure in the central region .
the nuclear double peak is very obvious in the colour index maps .
elmegreen et al . (
1997 ) showed a ring in @xmath10 with five discrete clumps of sf . the signal to noise ratio in our images is too low to reveal these sf clumps .
the previously reported isophote twist ( e.g. , pompea & rieke 1990 ) is obvious from our imaging , as well as from the pa and ellipticity profiles .
the double bar , outlined by the two separate peaks in the ellipticity profile , is most probably not exclusively due to the nuclear double peak , because the accompanying isophote twist in fact starts somewhat outside the region of influence of the double peak .
the surface brightness and colour profiles show the hump at the cnr radius that is seen in many similar galaxies presented in this paper .
the only obvious feature in our images is the very red nucleus . neither the broad - band _ hst _ @xmath9-image , nor our ground - based nir colour maps reveal any other structure ( fig . 1 , 2f ) .
regan & mulchaey ( 1999 ) used an _ hst _
wfpc2nicmos colour index map to show that a single spiral dust pattern dominates the circumnuclear morphology of this galaxy .
they described a strong red dust lane that emerges from a blue feature north of the nucleus at a radius of 3@xmath24 .
the spatial resolution of our images is not high enough to reveal such fine details .
the differences between our images and _ hst _ images are probably due to the longer spectral baseline that regan & mulchaey used ( 0.55 @xmath12 m to 1.6 @xmath12 m ) and the different spatial resolution .
quillen et al .
( 1999 ) noticed that the inner @xmath14-band isophotes are slightly elongated in a direction roughly perpendicular to the outer bar , so the galaxy may be doubly barred .
our @xmath9-band profile indeed shows an isophote twist of about 50 ( fig .
quillen et al . ( 1999 ) saw a curved dust feature at about 4 arcsec south of the nucleus in their _ hst _ wfpc2nicmos colour index map .
we see a multi - armed spiral pattern in our images ( fig .
2 g ) , in agreement with regan & mulchaey ( 1999 ) who describe the global morphology of this galaxy as multi - armed .
there is a lot of structure in our ground - based images , but a detailed comparison with the colour index maps obtained by regan & mulchaey ( 1999 ) is difficult due to the lower spatial resolution of our images . the _ hst _
nir image ( fig .
1 ) is the one used by regan & mulchaey , and shows the rather faint spiral structure in the cnr .
our colour index map shows that this galaxy has a small red nucleus , classified as sy2 .
the ellipticity of the isophotes reaches a maximum at a radius of 9 arcsec .
this could correspond to a ring , a small bar or a triaxial bulge .
the radial colour profiles ( @xmath10 and @xmath11 ) follow a characteristic shape , becoming very red close to the nucleus , with the colour most likely due to emission from dust heated by the agn radiation field .
the colours become bluer until a certain radius ( the location of the ring ) , after which they remain constant .
such a profile shape is only seen in the agns of our sample .
the difference in colour between the nucleus and the ring radius is about 0.2 magnitudes in both @xmath10 and @xmath11 .
buta & crocker ( 1993 ) classify this galaxy as having a nuclear ring , based on their h@xmath2 data .
elmegreen et al .
( 1997 ) , using nir observations , did not detect any ring . according to them , the ring consists of very young stars which do not show up in the nir .
however , we can see a well - defined ring in our @xmath10 and @xmath11 images ( fig .
there is a pair of dust lanes which connect the bar to the nuclear ring in the south and north .
the reddest colours are seen where the dust lanes merge with the nuclear ring .
colina et al .
( 1997 ) present a uv ( @xmath23 ) _ hst _ image of the cnr of ngc 4303 , which shows spiral structure outlining massive sf , continuing all the way into the unresolved core on the ne side of the nucleus .
the uv sf spiral is strongest on the side opposite to where we see the largest concentrations of dust ( darkest patches in fig .
2h ) in our nir colour index maps .
our _ hst _ @xmath9-band image shows some spiral structure in the nir but emission is dominated by the central bulge component .
there are strong isophotal twists within the central 7 arcsec , and a possible nuclear bar with a radius of 2 arcsec where the ellipticity reaches a peak of 0.2 and the position angle is constant at about 220 .
the signature of the ring can be recognised as a red peak in the @xmath10 and @xmath11 colour profiles , as well as in all other profiles .
our ellipticity and pa profiles provide some evidence for a nuclear bar with a radius of 12 arcsec .
the corresponding pa twist can be recognized in the contours of the broad - band image .
benedict et al . ( 1993 ) , using the wfpc camera on the _ hst _ , found an oval distortion with a length of 8 arcsec in the nuclear region .
wozniak et al . ( 1995 ) suggested that this galaxy could be another example of a double - barred galaxy , although they found it difficult to confirm this with their ground - based image at a considerably lower spatial resolution than our new data .
the colour maps published by shaw et al .
( 1995 ) revealed the presence of the circumnuclear ring , which is so prominent in the optical ( e.g. , morgan 1958 ) , and especially in h@xmath2 ( e.g. , pogge 1989a ) .
this ring is redder than all the other regions in the galaxy , having colours consistent with those of typical old stellar populations in ellipticals and spiral bulges .
we see a smooth and continuous ring in our @xmath10 colour index map ( fig .
the _ hst _
@xmath9-image ( fig .
1 ) shows a wealth of emitting structure in narrow tightly wound spiral armlets in the nuclear ring .
fine dust lanes are seen to accompany the spirals .
the radial profiles also show the signature of the ring , most clearly with colours that are about 0.1 magnitude redder than the background .
the prominent star - forming cnr of ngc 4321 ( m100 ) has been studied in great detail using , e.g. , optical and nir imaging , co interferometry , and modelling ( e.g. , pogge 1989 ; knapen et al .
1995a , b ; knapen 1998 ; wada et al .
1998 ; garca burillo et al . 1998
; ryder & knapen 1999 ; knapen et al .
knapen et al .
( 1995b ) concluded that m100 has a circumnuclear starburst maintained by a global bar - driven density wave .
as already shown in detail by knapen et al .
( 1995a ) , the @xmath3-image of this region is generally smooth , in contrast to the appearance in optical and h@xmath2 .
two symmetrically placed ` hot spots ' , named k1 and k2 , are obvious in fig . 2j , both in emission in the broad - band image , and as red features in the colour index images .
ryder & knapen ( 1999 ) recently used nir imaging and spectroscopy to confirm the suspicion that k1 and k2 are in fact regions of enhanced sf , and the @xmath3 emission from those regions is partly due to young stars ( knapen et al .
1995a , b ) .
our nir imaging confirms the location of dust lanes and suspected sf regions , shown by knapen et al .
( 1995a ) in their @xmath5 colour index map .
the locus of the circumnuclear ring - like structure shows up prominently in all radial profiles .
unfortunately , no _ hst _
nir images are available .
ngc 5248 is a galaxy with an hii nucleus and a lot of sf activity in the cnr .
elmegreen et al . (
1997 ) found very conspicuous central spiral arms , and several hotspots that form a ring - like spiral pattern .
buta and crocker ( 1993 ) detected a nuclear ring with a diameter of 1017 arcsec .
this activity shows up clearly in our broad - band nir images , as well as in our colour index maps ( fig .
spiral structure with star - forming arms , accompanied by dust lanes , is the dominant feature .
the western spiral arm has colours which are redder by about 0.1 mag in @xmath10 than its counterpart in the east .
our @xmath10 image has been published earlier by laine et al .
they compared it with the images obtained using adaptive optics , which show a nuclear grand - design spiral structure .
this nuclear spiral , at scales of tens of pc , is not expected to show up in a single broad - band nir image , even at _ hst _ resolution , and in fact does not show up ( fig . 1 ) .
the _ hst _ @xmath9-band image does show a wealth of structure in the cnr , again in the form of emitting regions distributed along spiral arm fragments , and accompanied by less luminous regions which may well be dusty . as in other galaxies
, we can see the signature of the ring as a peak at a radius of @xmath25 arcsec in all radial profiles .
there is no evidence for nested bars .
the dust structure in the circumnuclear ring is the most conspicuous feature in our @xmath10 colour index map , but the star - forming regions in the ring can also be seen in the broad - band images ( fig .
the bar dust lanes connect to this nuclear ring in the northeast and southwest .
several sites of sf are located along the ring , and its presence is also seen in the ellipticity and pa profiles , as well as in all other profiles ( fig .
we can also see a blue ring at about 2 arcsec , within the red nuclear ring surrounding it .
this blue ring shows up as a dip in the @xmath10 profile .
its nature is not clear , and needs further study .
the _ hst _
@xmath9-band image , used also by prez et al .
( 2000 ) , shows a picture also seen in , e.g. , ngc 3351 and ngc 4314 , namely of bright emitting knots distributed along tightly wound spiral armlets in the nuclear ring , accompanied by dust lanes .
we have presented examples of circumnuclear rings and dust lane patterns in the cnrs of a dozen barred galaxies . in most galaxies ,
rings are well defined , in others , nuclear spirals with different curvature radii are a predominant circumnuclear feature .
there are also some galaxies where both features coexist .
our sample is composed of 12 galaxies , most of which have nuclear rings .
six of them also have clear nuclear mini - spirals ( ngc 1530 , ngc 3504 , ngc 4314 , ngc 4321 , ngc 5248 and ngc 6951 ) , while in a few others there is less clear or circumstantial evidence for mini - spirals .
simulations ( e.g. athanassoula 1992 ; byrd et al .
1994 ; heller & shlosman 1994 ; knapen et al .
1995b ; piner , stone & teuben 1995 ; see also review by shlosman 1999 ) have shown that circumnuclear rings can arise as a consequence of bar - driven inflow and the existence of dynamical resonances in the bar .
shocks form in the gas in regions of orbit crowding along the leading edges of the bar , and dense gas accumulates in these regions . in the shock region
the gas loses angular momentum through torques exerted by the bar , and flows inward
. the pattern of the gas flow can be quite complex , but if two ilrs are present , inflowing gas accumulates in a ring between them ( reviewed by shlosman 1999 ) .
if there is no ilr , the gas may continue to flow inward , resulting in a nuclear starburst rather than a ring ( telesco , dressel & wolstencroft 1993 ) .
other mechanisms for the origin of circumnuclear rings that have been put forward as possible origins include shear in the differentially rotating disc ( buta & combes 1996 ) , supermassive black hole binaries ( taniguchi & wada 1996 ) , and minor mergers ( taniguchi 1999 ) .
our sample galaxies were selected to have some kind of circumnuclear structure , and except in ngc 3516 and ngc 3982 , all have clear circumnuclear star - forming ring - like regions .
sf is visible in most galaxies in the high - resolution _ hst _ nicmos images ( fig . 1 ) , and also in a number of galaxies in the ground - based images
. however , in all galaxies with circumnuclear sf , dust lane structure outlining small - scale spiral arms is present , as can be seen in the colour index maps ( fig . 2 ) .
this is true even for ngc 1300 where the sf regions are conspicuously absent from our ground - based nir data and very weak in _ hst _ nir images .
we conclude that in this class of barred galaxies with circumnuclear sf , dust lane structure in the cnr is always direct evidence of the accompanying sf , which may also show up in broad - band nir imaging .
nir data are not optimal for determining the sizes of the circumnuclear features , and we will come back to this issue in future papers , where imaging will be used to localize the sf activity in the cnr .
however , the radial profiles often allow the determination of the diameter of the `` rings '' : typical major axis diameters are 1 - 2 kpc .
the first suggestion about the link between dust lanes and shocks in the gas flow was made by prendergast ( 1962 ) .
gas can follow simple periodic orbits that do not intersect when the bar or oval distortion is not very strong . when the bar is strong , the families of periodic orbits intersect and shocks are formed ( e.g. sanders , teuben and van albada 1983 ) .
dust lanes in bars , as described by athanassoula ( 1992 ) , can be : 1 . straight and parallel to the bar major axis , or 2 .
curved and have their concave sides towards the bar major axis .
the detailed shape of the loci of offset shocks changes with the parameters of the potential ( athanassoula 1992 ) .
the simulations presented by athanassoula show that there is a very clear sequence of shapes for the shock loci . for axial ratios corresponding to fat bars or ovals , and presumably to what could theoretically be called a weak bar , the shock lanes are curved with their concave parts towards the bar major axis . for slim ( `` strong'')bars ,
the dust lanes straighten out , their curvature decreases and their shape becomes gradually more straight . for even higher axial ratios the shock loci tend towards the bar major axis .
thus , strong bars should have straight dust lanes , while ovals or weak bars should have more curved dust lanes .
athanassoula ( 1992 ) also proposed that the strength of sf near the dust lanes depends on the strength of the bar .
the dust lanes observed on the leading edges of most bars with high ellipticity ( strong ) are very smooth and show no hii regions or other signs of recent sf . in the case of curved dust lanes strings of hii regions
are accompanying the dust lanes . due to the enhanced gas mass
fraction , dust is expected to be prominent in the cnrs themselves , and this , in fact , is seen clearly in most cases in the colour index images , and also often in the broad - band images , especially the _ hst _ nicmos images .
such dust lanes are expected to be continuations of the dust lanes in the bar , and this is what is seen in our images ( see shlosman 1999 for a theoretical review ) .
nir colour index images are not very sensitive to changes in stellar populations , and can outline dust structure clearly ( see e.g. the @xmath5 map of m100 in knapen et al .
we see clear and abundant observational evidence for dust lanes on several scales , but most clearly in the cnrs .
dust lanes in the bars are not well visible in general in our nir imaging due to the lower signal to noise ratios achieved in the bar regions . in paper ii , we present optical colour index maps which outline the dust lane structure in the bars and discs of our sample galaxies more clearly , and we study the relationship between the shape of the dust lanes , the axial ratio , and the sf in the bar in more detail .
our colour maps have shown the presence of circumnuclear rings in most of the galaxies in our sample .
these rings are generally redder than other regions in the galaxy by about 0.05 0.1 magnitude in @xmath10 and @xmath11 . from the colour index maps or colour profiles alone
it is not possible to make meaningful statements about quantities of extinguishing dust implied by the redder colours .
in fact , there are indications from both optical and nir imaging and spectroscopy that the red colours in cnrs in galaxies like the ones studied here may be influenced by young stars , e.g. red supergiants ( knapen et al .
1995a , b ; knapen 1996 ; elmegreen et al . 1997
; ryder & knapen 1999 ) . in paper iii , we will compare the precise location of the @xmath10 and @xmath11 features with those of sf regions as seen in emission , and try to place quantitative limits on the origins of the red light in the cnrs .
two of the host galaxies of sy nuclei ( ngc 3516 and ngc 3982 ) , which have colours consistent with those measured by peletier et al .
( 1999 ) , also show a peculiar shape in the colour profiles , starting from a very red value and steeply decreasing until the radius of the ring , remaining constant afterwards .
peletier et al .
( 1999 ) suggested that the red @xmath11 ( or @xmath10 ) colours in the cores of many sy galaxies could be due to a significant fraction of thermal radiation from hot dust heated by the sy nucleus .
however , the two other agn hosts in our sample ( ngc 4303 and ngc 6951 ) do not show significantly red nuclei .
one of the main data products presented in this paper is a set of detailed radial profile fits to the nir images .
we present ( fig .
2a - l ) radial profiles of ellipticity , major axis position angle , nir colour , and @xmath9-band surface brightness .
we will discuss the existence and properties of ( primary and secondary ) bars in paper ii , where we combine the nir profiles presented here with @xmath26-band profiles of the complete discs , covering also the primary bar , if present .
certain aspects of the radial profiles have already been discussed in section 3 , but there is one characteristic which is worth mentioning here .
apart from ngc 3516 and ngc 3982 , the radial profiles show characteristic behaviour near the radius of the circumnuclear `` ring '' . in seven
( out of 10 ) galaxies , _ all _ profiles ( surface brightness , colour , ellipticity and position angle ) show either a change in slope , where the surface brightness profile becomes steeper , or a bump , when the colour profile shows a limited redder region caused by either dust or sf in or near the ring
see section 4.3 .
these characteristics are accompanied by some change in ellipticity and position angle . in a further two galaxies ( ngc 1530 and ngc 4314 ) only the surface brightness profile
does not follow this trend , whereas in ngc 5248 the colour profile is the deviant one .
of course in ngc 2903 and ngc 3351 we can not judge the behaviour of ellipticity / position angle and colour profiles , respectively .
we postulate that the features in the surface brightness and colour profiles are due directly to a relatively young population .
firstly , an old , bulge - like , population would not be expected to give rise to strong high - resolution features such as the ones discussed here . secondly ,
if the red bumps in the colour profiles were due to extinction by dust only , they should be accompanied by relative dips in the corresponding surface brightness profiles ( such dust would remove light in each of the @xmath8 and @xmath3-bands ) . instead
, red bumps in the colour profiles are in all cases accompanied by bumps in the surface brightness profiles : relative excesses in emission , red in colour , that are naturally explained by excess emission produced locally by relatively young stars , probably accompanied by dust .
we will come back to this issue in paper iii , where we combine the nir data with imaging .
this is the first paper in a series exploring the relationships between ring - like star - forming regions in the central few kpc regions of barred galaxies and the morphological properties of their hosts . in this paper , we present new sub - arcsec resolution nir images obtained with the _ cfht _ of the central regions of a sample of 12 barred galaxies with circumnuclear sf activity , and _ hst _ archival nir images of most of the sample objects . in paper
iii , we will combine these data with optical imaging of the complete hosts , as presented in paper ii .
our nir images reveal a wealth of structure , caused by young stars and dust . in colour index maps , structures such as sf sites and dust lanes are generally apparent , even in those galaxies where the broad - band images appear featureless .
the colour maps of most galaxies show evidence of an often tightly wound spiral structure . in some cases ,
this spiral structure is visible also in the broad - band images , and as expected due to the higher spatial resolution , more often so in the _ hst _ than in the ground - based images .
where circumnuclear structure is visible in the _ hst _ images ( in about three quarters of all cases ) , it shows large numbers of small emitting regions , presumably sf regions , often outlined in spiral arm fragments , and accompanied by dust lanes .
circumnuclear spiral structure thus appears to be common in barred spiral galaxies with circumnuclear sf . in most of our sample galaxies , and in all those where the star - forming nuclear ring is well defined , radial profiles of surface brightness , colour , ellipticity and position angle show a characteristic bump or change in slope at the radius where the circumnuclear ring is located .
_ acknowledgements _ this research has made use of the nasa / ipac extragalactic database ( ned ) which is operated by the jet propulsion laboratory , california institute of technology , under contract with the national aeronautics and space administration . | we present sub - arcsecond resolution ground - based near - infrared images of the central regions of a sample of twelve barred galaxies with circumnuclear star formation activity , which is organized in ring - like regions typically one kiloparsec in diameter .
we also present _
hubble space telescope _ near - infrared images of ten of our sample galaxies , and compare them with our ground - based data . although our sample galaxies were selected for the presence of circumnuclear star formation activity , our broad - band near - infrared images are heterogeneous , showing a substantial amount of small - scale structure in some galaxies , and practically none in others .
we argue that , where it exists , this structure is caused by young stars , which also cause the characteristic bumps or changes in slope in the radial profiles of ellipticity , major axis position angle , surface brightness and colour at the radius of the circumnuclear ring in most of our sample galaxies . in 7 out of 10 _ hst _ images , star formation in the nuclear ring is clearly visible as a large number of small emitting regions , organised into spiral arm fragments , which are accompanied by dust lanes .
nir colour index maps show much more clearly the location of dust lanes and , in certain cases , regions of star formation than single broad - band images .
circumnuclear spiral structure thus outlined appears to be common in barred spiral galaxies with circumnuclear star formation .
# 1 2_2h@xmath0 2_h_2@xmath1 # 1.#2 # 1#2 # 1.#2 # 1#2 # 1.#2 # 1#2 # 1h#2m#3s # 1^h#2^m#3^s # 1d#2m#3s # 1#2#3 # 1h#2m#3.#4s # 1^h#2^m#3#4 # 1d#2m#3.#4s # 1#2#3#4 # 1.#2 # 1^m -0.55em.0.22em#2 = = = = = = = = # 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 = = = = = = = = [ firstpage ] galaxies : spiral galaxies : starburst galaxies : evolution galaxies : structure infrared : galaxies |
the isotropy of the cosmic expansion is well - constrained for the early universe , particularly by cosmic microwave background observations , and is a basic tenet of physical cosmology .
the change from a matter - dominated to a dark energy - dominated universe in recent times , however , raises the possibility of a dark energy - driven anisotropic expansion if dark energy is itself anisotropic .
there is no obvious reason for such symmetry breaking , but observational tests of something as fundamental as the isotropy of the hubble expansion should be made for late times ( the current epoch ) .
one such test is possible via extragalactic proper motions : if the expansion is anisotropic , then quasars and galaxies will stream toward directions of faster expansion and away from directions of slower expansion .
the signature of anisotropic expansion in a homogeneous universe is thus a curl - free proper motion vector field ( to first order ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the term `` cosmic parallax '' has been used by some to indicate a general relative angular motion of objects in the universe ( e.g. , * ? ? ?
* ; * ? ? ?
* ) and used by others in a more canonical sense to indicate apparent angular motion induced by the motion of the observer ( e.g. , * ? ? ?
we favor the latter usage ; this paper therefore treats the apparent proper motion induced by anisotropic cosmic expansion @xcite , referenced to the international celestial reference frame ( icrf ) in the current epoch .
the observed proper motions are therefore relative , but are not necessarily induced by the observer s motion . in this paper , we present a simple model of anisotropic expansion and fit the model to the @xcite proper motion catalog to place a new constraint on the isotropy of the hubble expansion and thus on the isotropy of dark energy .
we assume @xmath3 km s@xmath1 mpc@xmath1 and a flat cosmology ( this treatment is independent of specific assumptions about @xmath4 and @xmath5 , provided @xmath6 ) .
as described in @xcite and @xcite , a homogeneous but anisotropic bianchi i model with metric @xmath7 has three different expansion rates , @xmath8 , @xmath9 , and @xmath10 , where the hubble parameter as observed is @xmath11 and the friedmann - robertson - walker metric is recovered for @xmath12 .
this metric has a flat geometry and no global vorticity , but the anisotropic expansion will produce a shearing velocity field , causing objects to stream toward directions of faster expansion and away from directions of slower expansion .
the shear can be characterized by the fractional deviation from the average hubble expansion today ( @xmath13 ) , @xmath14 where @xmath15 , @xmath16 , or @xmath17 , isotropy corresponds to @xmath18 ( no deviation from @xmath19 in any direction ) , and the expansion is `` conserved '' : @xmath20 ( the directionless overall @xmath19 is preserved , despite anisotropy ) . under the simplifying assumption of straight geodesics ( incorrect , but a small error as demonstrated by @xcite ) , the sky signal of an anisotropic expansion is a curl - free quadrupolar proper motion vector field that is independent of distance .
note that this is a `` real - time '' signal , meaning that the time derivatives are with respect to small coordinate time intervals today ( decades ) and therefore a constant @xmath19 is a good assumption .
the application of this model by @xcite , @xcite , and @xcite to the apparent proper motions of extragalactic objects , however , retained a metric fixed with respect to the equatorial coordinate system and did not allow the anisotropy to have an arbitrary orientation .
application of this model to observations by @xcite was not rotationally invariant ( the result would depend on the choice of coordinate system ) . in order to obtain a fully general model ,
we require rotational invariance appropriate for vector spherical harmonics ( the divergence and curl of the scalar spherical harmonics , @xmath21 ) : all orders @xmath22 for a given degree @xmath23 must be included in a model fit to data @xcite . here
we demonstrate that all quadrupolar ( @xmath24 ) curl - free vector spherical harmonic orders naturally arise from an arbitrary rotation of the anisotropic coordinate system .
using an arbitrary rotation from the equatorial reference frame to the one described by equation ( [ eqn : metric ] ) , we can allow for any anisotropy orientation .
the anisotropy also need not be triaxial ; there could simply be a single direction of high or low expansion , provided the expansion conservation condition is satisfied .
rotations are made about @xmath25 by angle @xmath26 , about the new @xmath27 by @xmath28 , and about the final @xmath29 axis by @xmath30 .
thus , the anisotropic bianchi i axes can be expressed in the equatorial coordinate system via @xmath31 , where @xmath32 is the rotation matrix about axis @xmath33 by angle @xmath34 .
the equatorial coordinates of the anisotropy axes are therefore : @xmath35 @xmath36 @xmath37 the anisotropic expansion proper motion vector field @xmath38 coefficients are listed in table [ tab : shear_theory ] ( see appendix [ appdx ] for the full equation ) .
these were obtained by taking the time derivatives of the equatorial coordinates , @xmath39 and @xmath40 , expressed in terms of the shear terms @xmath41 .
the general form of the vector field is @xmath42 \bmath{\hat{e}}_i \label{eqn : sheareqn}\end{aligned}\ ] ] where there are no imaginary coefficients for @xmath43 and the sum over real and imaginary coefficients is implied .
the hubble constant can be written as @xmath44 @xmath45as yr@xmath1 ( for @xmath46 km s@xmath1 mpc@xmath1 ) , so a shear of 10% would produce streaming motions of order 1.5 @xmath45as yr@xmath1
. comparison of this derivation of the shear field @xmath38 to the spheroidal ( curl - free or e - mode ) quadrupolar vector spherical harmonics @xcite reveals an exact one - to - one correspondence ( table [ tab : shear_theory ] ) .
the treatment here is therefore rotationally invariant ( as desired ) , and it is therefore completely equivalent to fit the spheroidal @xmath24 vector spherical harmonics to a proper motion vector field to test the isotropy of expansion .
it is not , however , correct to fit or select single or selected orders of a vector spherical harmonic model , as has been done previously .
table [ tab : shear_theory ] lists the equivalent vector spherical harmonic coefficients ( see appendix [ appdx ] for the full equation ) , following the @xcite conventions : @xmath47 note that this equation has absorbed the factors of 2 for the @xmath48 orders and the factors of @xmath49 for the imaginary terms into the coefficients @xmath50 , in contrast to the definitions used by @xcite . [
cols="^,^,^,^,^,^ " , ] the quadrupolar ( @xmath24 ) spheroidal vector spherical harmonic coefficients @xmath51 and the total power in the @xmath24 curl - free order @xmath52 follow the @xcite conventions .
parenthetical values are @xmath0 uncertainties .
the units are @xmath45as yr@xmath1 .
the z - score of a one - sided significance test is 1.1 , so the power in this mode is not significant .
figure [ fig : shear ] shows the model fit to the proper motion vector field for the 429 objects in the @xcite catalog .
positive deviation from the hubble expansion ( @xmath53 ) appears as an antipodal pair of convergent points , and negative deviation appears as an antipodal pair of divergent points .
the equatorial coordinates of the best - fit ( but not significant ) anisotropy axes are : @xmath54 and @xmath55 , @xmath56 and @xmath57 , and @xmath58 and @xmath59 .
these results are insensitive to the initial parameter assumptions .
while the anisotropy axes can be interchanged arbitrarily through rotations , the shear values and directions are stable best - fit solutions .
the solution is likewise insensitive to the sometimes large intrinsic proper motions or large proper motion uncertainties of individual objects . restricting the sample to the 358 objects with proper motions and proper motion uncertainties less than 100 @xmath45as
yr@xmath1 has a negligible impact on the model parameters or uncertainties listed in table [ tab : shear_meas ] .
the power in the spheroidal vector spherical harmonics of degree @xmath23 is @xmath60 a scalar under coordinate rotation @xcite . for the @xmath24 fit in table [ tab : shear_meas ] ,
the power is @xmath61 @xmath45as yr@xmath1 . the z - score of a one - sided significance test of the mean - error - reduced power ( * ?
* equations ( 85 ) and ( 87 ) ) , is 1.1 , so the power in this mode is not significant . and [ tab : e2quad_meas ] .
the measured anisotropy is not significant , either by parameter in or total power .
the high and low deviations from the average hubble expansion are indicated as converging ( @xmath62 ) and diverging ( @xmath63 ) loci , including error bars .
the upper right scale bar indicates the amplitude of the vectors , the dotted line shows the galactic plane , the solid circle indicates the galactic center , the open circles mark the galactic poles , and the crosses indicate the @xmath64 direction.,scaledwidth=48.0% ]
while the bianchi i metric model assumed for the data analysis is specific , the fitting function is a general @xmath24 spheroidal vector spherical harmonic ; the model plays a role in specific parameter estimation , but the vector field sky pattern is generic .
one can therefore adapt the fit parameters to different cosmological models , and one can fit additional vector spherical harmonic terms ( e.g. , * ? ? ? * ; * ? ? ?
figure [ fig : shear ] shows a remarkable ( but non - significant ) alignment of the fit anisotropy axes with the galactic plane - equator intersection .
this could be driven by a vertical spheroidal dipole component induced by vertical ( out - of - the - plane ) solar acceleration ( a second aberration drift ; * ? ? ?
. a dipole fit to the proper motion vector field ( after subtracting the galactocentric acceleration described in section [ sec : methods ] ) is not significant and does not point toward the @xmath29 axis ( the points of convergence ) . since the anisotropy is neither significant in total power nor in individual parameters , this surprising alignment seems to be coincidental and should not be over - interpreted .
the main limitation to this technique is the proper motion precision and sample size .
the overall proper motion precision of individual objects will improve with time , as will the sample size of icrf `` defining sources , '' but these will be slow , secular improvements . the next order - of - magnitude improvement will be provided by the gaia mission .
gaia is an optical astrometry mission that will measure 500,000 quasar proper motions with @xmath6580 @xmath45as astrometry for @xmath66 mag stars @xcite .
unlike radio sources , compact optical extragalactic sources do not show significant internal intrinsic proper motions , so gaia proper motion catalogs will not exhibit the uncorrelated proper motion signals that contaminate radio measurements .
we estimate that the gaia mission will constrain anisotropy below 1% .
we have demonstrated how anisotropic hubble expansion can be measured or constrained using extragalactic proper motions , and we applied this technique to the best current proper motion catalog @xcite to place a new constraint on the isotropy of the hubble expansion and thus on the isotropy of dark energy .
no significant anisotropy was detected ; the hubble expansion is isotropic to 7% ( @xmath0 ) , corresponding to streaming motions of 1 @xmath45as yr@xmath1 , in the best - constrained directions ( @xmath219% and + 17% in the least - constrained directions ) and does not significantly deviate from isotropy in any direction .
the gaia mission , which is expected to obtain proper motions for 500,000 quasars , will likely constrain the anisotropy below 1% .
the author thanks the anonymous referee for helpful comments and o. titov and s. b. lambert for making their proper motion catalog publicly available .
this research has made use of the vizier catalogue access tool , cds , strasbourg , france .
the original description of the vizier service was published in a&as 143 , 23 .
this research has made use of nasa s astrophysics data system bibliographic services .
amendola , l. , eggers b. o. , valkenburg , w. , & wong , y. y. y. 2013 , journal of cosmology and astroparticle physics , 12 , 042 darling , j. 2013 , apj , 777 , l21 de bruijne , j. , perryman , m. , lindegren , l. , jordi , c. , hg , e. , katz , d. , & cropper , m. 2005 , gaia - jdb-022 technical note ding , f. & croft , r. a. c. 2009 , mnras , 397 , 1739 fontanini , m. , west , e. j. , & trodden , m. 2009 , phys . rev .
d , 80 , 123515 mignard , f. & klioner , s. 2012 , a&a , 547 , a59 quercellini , c. , cabella , p. , amendola , l. , quartin , m. , & balbi , a. 2009 , phys .
d , 80 , 063527 reid , m. j. , menten , k. m. , brunthaler , a. , zheng , x. w. , dame , t. m. , xu , y. , wu , y. , zhang , b. , sanna , a. , sato , m. , hachisuka , k. , choi , y. k. , immer , k. , moscadelli , l. , rygl , k. l. j. , & bartkiewicz , a. 2014 , in press ( arxiv 1401.5377 ) titov , o. 2009 , in proc .
19th european vlbi for geodesy and astrometry ( evga ) working meeting , ed .
g. bourda , et al .
, 14 titov , o. & malkin , z. 2009 , a&a , 506 , 1477 titov , o. , lambert , s. b. , & gontier , a .- m .
2011 , a&a , 529 , a91 titov , o. & lambert , s. 2013 , a&a , 559 , a95 xu , m. h. , wang , g. l. , & zhao , m. 2012 , a&a , 544 , a135
the complete vector field described in section [ sec : theory ] , equation ( [ eqn : sheareqn ] ) , and table [ tab : shear_theory ] is @xmath67 \right . \nonumber\\
\left . + { 1\over2 } \cos\alpha\sin\delta \left [ \left(\sigma_{y'}+{1\over2}\,\sigma_{x'}\right ) \left(2\cos\alpha^*\cos\delta^*\sin2\psi^*+\sin\alpha^*\sin2\delta^*\cos2\psi^*\right ) + { 3\over2}\,\sigma_{x'}\,\sin\alpha^*\sin2\delta^*\right ]
\right . \nonumber\\ \left . + { 1\over4 } \sin2\alpha\cos\delta \left [ \left(\sigma_{y'}+{1\over2}\,\sigma_{x'}\right ) \left(3\cos2\alpha^*\cos2\psi^*-\cos2\alpha^*\cos2\delta^*\cos2\psi^ * -4\sin2\alpha^*\sin\delta^*\sin2\psi^*\right ) \right.\right .
\nonumber\\ \left.\left .
-{3\over2}\,\sigma_{x'}\,\cos2\alpha^*\left(1+\cos2\delta^*\right)\right ] \right.\nonumber\\ \left .
+ { 1\over4 } \cos2\alpha\cos\delta \left [ \left(\sigma_{y'}+{1\over2}\,\sigma_{x'}\right ) \left(-3\sin2\alpha^*\cos2\psi^*+\sin2\alpha^*\cos2\delta^*\cos2\psi^ * -4\cos2\alpha^*\sin\delta^*\sin2\psi^*\right ) \right.\right .
\nonumber\\ \left .
\left .
+ { 3\over2}\,\sigma_{x'}\,\sin2\alpha^*\left(1+\cos2\delta^*\right)\right]\right ) \bmath{\hat{e}}_\alpha \nonumber\\ + h_\circ \left({3\over8}\sin2\delta\left[-\left(\sigma_{y'}+{1\over2}\,\sigma_{x'}\right)\cos2\psi^*\left(1+\cos2\delta^*\right ) + { 1\over2}\,\sigma_{x'}\left(1 - 3\cos2\delta^*\right)\right ] \right .
\nonumber\\ \left.+{1\over2}\cos\alpha\cos2\delta\left[-\left(\sigma_{y'}+{1\over2}\,\sigma_{x'}\right)\left(2\sin\alpha^*\cos\delta^*\sin2\psi^*-\cos\alpha^*\sin2\delta^*\cos2\psi^*\right ) + { 3\over2}\,\sigma_{x'}\,\cos\alpha^*\sin2\delta^*\right ] \right .
\nonumber\\ \left .
+ { 1\over2}\sin\alpha\cos2\delta\left[\left(\sigma_{y'}+{1\over2}\,\sigma_{x'}\right)\left(2\cos\alpha^*\cos\delta^*\sin2\psi^*+\sin\alpha^*\sin2\delta^*\cos2\psi^*\right ) + { 3\over2}\,\sigma_{x'}\,\sin\alpha^*\sin2\delta^*\right ] \right .
\nonumber\\ \left .
+ { 1\over8}\cos2\alpha\sin2\delta\left[\left(\sigma_{y'}+{1\over2}\,\sigma_{x'}\right)\left(3\cos2\alpha^*\cos2\psi^*-\cos2\alpha^*\cos2\delta^*\cos2\psi^ * -4\sin2\alpha^*\sin\delta^*\sin2\psi^*\right ) \right.\right .
\nonumber\\ \left.\left .
-{3\over2}\,\sigma_{x'}\,\cos2\alpha^*\left(1+\cos2\delta^*\right)\right ] \right .
\nonumber\\ \left .
+ { 1\over8}\sin2\alpha\sin2\delta\left[-\left(\sigma_{y'}+{1\over2}\,\sigma_{x'}\right)\left(-3\sin2\alpha^*\cos2\psi^*+\sin2\alpha^*\cos2\delta^*\cos2\psi^ * -4\cos2\alpha^*\sin\delta\sin2\psi^*\right ) \right.\right.\nonumber\\ \left.\left .
-{3\over2}\,\sigma_{x'}\ , \sin2\alpha^*\left(1+\cos2\delta^*\right ) \right ] \right ) \bmath{\hat{e}}_\delta.\ \
\label{eqn : fullshear}\end{aligned}\ ] ] the e - mode ( curl - free ) quadrupole vector spherical harmonic is @xmath68 | the isotropy of the universal hubble expansion is a fundamental tenet of physical cosmology , but it has not been precisely tested during the current epoch , when dark energy is dominant .
anisotropic expansion will produce a shearing velocity field , causing objects to stream toward directions of faster expansion and away from directions of slower expansion .
this work tests the basic cosmological assumption of isotropic expansion and thus the isotropy of dark energy .
the simplest anisotropy will manifest as a quadrupolar curl - free proper motion vector field .
we derive this theoretical signature using a tri - axial expanding metric with a flat geometry ( bianchi i model ) , generalizing and correcting previous work .
we then employ the best current data , the titov & lambert ( 2013 ) proper motion catalog of 429 objects , to measure the isotropy of universal expansion .
we demonstrate that the hubble expansion is isotropic to 7% ( @xmath0 ) , corresponding to streaming motions of 1 microarcsecond yr@xmath1 , in the best - constrained directions ( @xmath219% and + 17% in the least - constrained directions ) and does not significantly deviate from isotropy in any direction .
the gaia mission , which is expected to obtain proper motions for 500,000 quasars , will likely constrain the anisotropy below 1% .
astrometry cosmology : observations cosmology : theory cosmology : miscellaneous dark energy proper motions |
one of new directions in the heavy quark physics is related with a description of baryons with two heavy quarks @xcite . in a large heavy quark mass limit ,
a typical size of a @xmath10-diquark configuration is much less than the characteristic radius of light hadrons @xmath11 . for light quarks ,
the heavy - heavy diquark object looks like a heavy antiquark at low virtualities , so that one can relate the @xmath12-baryon characteristics with the properties of a meson with a single heavy quark @xcite .
the form factors of the doubly heavy baryons are straightforwardly related with those of heavy light mesons @xcite , described by the isgur - wise functions , up to the accuracy of a probability for the transition of the doubly heavy diquark to another one .
the latter point of view to the @xmath10 diquark allows one also to formulate some rules , describing the process of the @xmath12 baryon production @xcite .
first , the process under consideration can be represented as the process of a hard production of the @xmath10 diquark , that further has a hadronization into the @xmath12 baryon .
the hadronic production of the diquark with a mixed flavor is calculated under the analogy with the @xmath13 heavy quarkonium production , on the basis of computations of the fourth @xmath3-order diagrams in the qcd perturbation theory .
second , the nonperturbative soft part of the matrix element describes the quark binding in the @xmath10 diquark and it is given by the diquark wave function at the origin . at present , the basis for the current estimates of the @xmath12 baryon production cross section is the consideration of the @xmath10 diquark production cross sections at large transverse momenta via the fragmentation mechanism @xcite .
the fragmentation function of the heavy quark @xmath14 is taken in the same form as the function of fragmentation into the heavy quarkonium @xmath13 @xcite with the same quantum numbers over the lorentz group .
this approach is quite excusive and valid in the high energy @xmath15-annihilation , where the fragmentational mechanism does practically dominate @xcite , but the situation is more complex for the hadronic production @xcite .
the complication is caused by that a number of the leading order diagrams , describing the hadronic production of doubly heavy diquark , is much greater than the number of diagrams in the @xmath15-interaction .
moreover , the former diagrams do not allow one completely to interpret them in terms of the fragmentational mechanism .
the fragmentation is not exactly defined in the hadronic production . as was shown @xcite ,
the diagrams of the recombinational type give a dominant contribution even at quite large transverse momenta .
so , a simple formula , corresponding to the convolution of the differential cross section of the heavy quark production with the function of the heavy quark fragmentation into the @xmath16 quarkonium or @xmath10 diquark , gives only a cross section evaluation in the order of magnitude . to obtain an exact result
is necessary to take into account all 36 diagrams of the fourth order over @xmath3 .
in what follows , we present the results of exact calculations for the @xmath17 and @xmath18 baryons yields in the framework of the computations for the heavy quark pair production in the fourth order diagrams of qcd .
the difference between the given approach and the fragmentational one is connected with that the contribution of the complete set of diagrams is calculated with no neglection of a part of them .
for the states with a various quantum numbers , the performed calculations result in the yield ratios , different from those of in the approximate approach of the fragmentation .
the calculation technique , applied in the present paper , is analogous to that of the hadronic production of @xmath19 mesons @xcite .
the only difference is due to the binding of two heavy quarks ( @xmath20 and @xmath21 ) in contrast to the binding of a heavy quark with a heavy antiquark .
we suppose that the binding energy in the diquark is much less than the masses of quarks , composing the diquark , and hence , the quarks are on the mass shells .
therefore , the quark four - momenta are related with the diquark momentum @xmath22 in the following way @xmath23 where @xmath24 is the diquark mass , @xmath25 are the quark masses . in the given approach
, the diquark production can be described by the 36 leading order feynman diagrams , corresponding to the production of four free quarks in the way of combining of two quarks into the colour antitriplet diquark with the given quantum numbers over the lorentz group .
the latter procedure is performed by means of the projection operators @xmath26 for the scalar state of diquark ( the corresponding baryon is denoted as @xmath27 ) , @xmath28 for the vector state of diquark ( the baryons are denoted as @xmath29 and @xmath30 ) .
to produce the quarks , composing the diquark , in the @xmath31-state , one has to introduce the colour wave function as @xmath32 , into the diquark production vertex , so that @xmath33 is the colour index of the first quark , @xmath34 is that of the second one , and @xmath35 is the colour index of the diquark . the diquark production amplitude @xmath36 is expressed via the amplitude @xmath37 for the free quark production in kinematics ( [ 1 ] ) as @xmath38 where @xmath39 are the spiralities of corresponding quarks , @xmath40 are its colour indices , @xmath41 is the diquark radial wave function at the origin , and the @xmath42 operators have the following explicit form ( @xmath43 ) @xmath44 for the scalar state , and @xmath45 for the vector one . in the numerical calculations , giving results , which will be discussed in the next section ,
we suppose the following values of parameters @xmath46 where the @xmath47 value has been calculated by means of a numerical solution of the schrdinger equation with the martin potential @xcite , multiplied by the 1/2 factor , caused by the colour antitriplet state of quarks instead of singlet one , and the @xmath48 value is taken from ref.@xcite for the sake of convenience of the result comparison . to calculate the production cross section of the diquarks , composed of two @xmath49-quarks , one has to account for their identity .
one can easily find , that the antisymmetrization over the identical fermions leads to the scalar diquark production amplitude , equal to zero , and it results in that the amplitude of the vector @xmath50-diquark production can be obtained by the substitution of equal masses in the amplitude of the vector @xmath51-diquark production with the account for the 1/2 factor , following from the identity of quarks and antiquarks .
we assume , that the produced diquark has the fragmentation into the baryon , practically carrying away a total diquark momentum , with the unit probability .
therefore , discussing the results , we will talk on the differential cross sections of @xmath52 and @xmath2 baryons production , since we suppose that the latters negligibly small deviate from the differential cross sections of diquarks .
the total energy dependence of the gluonic production cross sections of @xmath0 ( @xmath53 ) and @xmath1 ( @xmath54 ) baryons is shown on fig.1 and in tab.1 .
to compare , the predictions of the fragmentational mechanism for @xmath1 ( solid line ) and @xmath0 ( dashed line ) are also presented .
one can see from the figure , that the fragmentational production mechanism , assuming the validity of factorization in the cross section at @xmath55 and at large transverse momenta via the formula @xmath56 with @xmath57 , does not work at low gluon energies , where it overestimates the cross section , because of the incorrect evaluation of the phase space , and it is not valid also at large energies , where the predictions of the fragmentational mechanism are essentially less that the exact results .
so , the fragmentational values underestimate the @xmath1 and @xmath0 cross sections in 9 and 4 times , respectively , at @xmath58 gev .
when the fragmentational predictions give the ratio @xmath59 , the exact perturbative calculations result in @xmath60 even at @xmath58 gev .
the agreement with the fragmentational production at @xmath58 gev is poor even at large transverse momenta of the baryon , as one can see from the distributions over @xmath61 for the @xmath1 and @xmath0 production , shown on fig.2 in comparison with the predictions of the fragmentational mechanism .
note , that in contrast to the doubly heavy baryon production , the exact perturbative calculations of the gluonic production of @xmath62 mesons with @xmath63 gev at @xmath58 gev agree with the fragmentational predictions . for the baryon production ,
a visible deviation is observed up to the largest values of @xmath61 . the differential cross section @xmath64 of the @xmath65 production in @xmath66 interactions at @xmath67 tev
is presented on fig.3 in comparison with the fragmentational predictions . in the @xmath19 meson production
, the difference between the exact and fragmentational approaches was slightly hidden due to the convolution with the hadron structure functions . in the process of baryon production , this is not the case , and the invalidity of the fragmentational approach explicitly follows from the form of the distribution under consideration . the calculation results on the @xmath2 production point out that the deviation between the exact perturbative and fragmentational values is also essential as in the @xmath1 production .
this disagreement can be noted on fig.4 and in tab.2 , where the dependence of the @xmath2 production cross section on the energy of the gluon interactions is presented , as well as on fig.5 , where the differential cross section @xmath64 of the @xmath2 production in the gluon interactions is shown at @xmath58 gev in comparison with the fragmentational mechanism predictions . from the latter figure
, one can see that even at @xmath68 gev , the exact result slightly overestimates the fragmentation . the differential cross section @xmath64 of the @xmath2 production in @xmath66 interactions at the energy @xmath67 tev
is shown on fig.6 in comparison with the fragmentational estimate .
one can see , that at the reasonable values of @xmath61 , the fragmentational result is approximately 3 times less than the exact perturbative one . at the chosen values of parameters and with the account for the cuts over the transverse momentum and rapidity of the baryons ( @xmath8 gev and @xmath9 ) ,
the production cross section of the @xmath69-wave @xmath17-baryons and its antiparticles is evaluated as @xmath70 nb , and the total cross section of the @xmath69-wave @xmath18-baryon production with the account for antiparticles is equal to @xmath71 nb . after the expected end of run ib at tevatron with the integral luminosity @xmath72 , one has the yields of @xmath73 of @xmath17-baryons and @xmath74 of @xmath18-baryons .
thus , we have shown , that the calculations in the leading order approximation of the qcd perturbation theory for the gluonic production of the doubly heavy diquarks , having the hadronization into the baryons , lead to the essential discrepancy in the differential as well as total cross sections of the baryon production in comparison with the predictions of the mechanism of the heavy quark fragmentation into the diquark .
* m.a.moinester , tel aviv univ .
preprint taup 2238 - 95 , 1995 [ hep - ph/ 9503289 ] ; + e.bagan et al . , z.phys .
c64 , 57 ( 1994 ) ; + m.a.sanchis-lozano , nucl.phys .
b440 , 251 ( 1995 ) ; + v.v.kiselev , a.k.likhoded , m.v.shevlyagin , phys.lett .
b332 , 411 ( 1994 ) .
h.georgi , m.b.wise , phys.lett .
b243 , 279 ( 1990 ) ; + c.d.carone , phys .
b253 , 408 ( 1991 ) ; + m.j.savage , m.b.wise , phys.lett .
b248 , 177 ( 1990 ) .
m.neubert , phys.rep .
245 , 259 ( 1994 ) .
a.falk et al . , phys.rev .
d49 , 555 ( 1994 ) .
h.chang , y .- q.chen , phys.rev .
d46 , 3845 ( 1992 ) ; + e.braaten , k.cheung , t.c.yuan , phys.rev .
d48 , 4230 , 5049 ( 1993 ) .
v.v.kiselev , a.k.likhoded , m.v.shevlyagin , z.phys .
c23 , 77 ( 1994 ) .
a.v.berezhnoy , a.k.likhoded , m.v.shevlyagin , yad.fiz .
58 , 730 ( 1995 ) ; + a.v.berezhnoy , a.k.likhoded , o.p.youshchenko , preprint ihep 95 - 59 , protvino , 1995 [ hep - ph/ 9504302 ] ; + a.v.berezhnoy , a.k.likhoded , m.v.shevlyagin , phys.lett .
b342 , 351 ( 1995 ) ; + k.kolodzej , a.leike , r.rckl , preprint mpi - pht/95 - 36 , 1995 [ hep - ph/ 9505298 ] .
a.martin , phys.lett .
b92 , 338 ( 1980 ) .
.the dependence of the gluonic production cross sections for the @xmath0 and @xmath1 baryons on the total energy .
( the error on last digits stands in brackets . ) [ cols="^,^,^ " , ]
* the gluonic production cross sections of @xmath0 ( @xmath53 ) and @xmath1 ( @xmath54 ) in comparison with the predictions of the fragmentational mechanism for @xmath0 ( dashed line ) and @xmath1 ( solid line ) . * the distributions over the transverse momentum in the gluonic production of @xmath65 in comparison with the fragmentation result at the interaction energy 100 gev .
here and in what follows , the solid and dashed lines correspond to the @xmath1 and @xmath0 production , respectively , and the exact perturbative and fragmentational results are shown as histograms and smooth curves , respectively . * the differential cross section @xmath75 of the @xmath65 production in @xmath76 collisions versus the transverse momentum of the @xmath65-baryon at the hadron interaction energy 1.8 tev , in comparison with the fragmentation result . * the gluonic production cross section of @xmath2 ( @xmath54 ) in comparison with the predictions of the fragmentational mechanism ( solid line ) . * the distribution over the transverse momentum in the gluonic production of @xmath2 in comparison with the fragmentation result at the interaction energy 100 gev . * the differential cross section @xmath75 of the @xmath2 production in @xmath76 collisions versus the transverse momentum of the @xmath2-baryon at the hadron interaction energy 1.8 tev , in comparison with the fragmentation result . | in the framework of qcd perturbation theory , total and differential cross sections of the @xmath0 , @xmath1 and @xmath2 baryons production in gluon collisions are calculated in the leading order over @xmath3 for the doubly heavy ( @xmath4 ) and ( @xmath5 ) diquarks . at both small and large transverse momenta of the baryons , a use of the mechanism of the heavy quark fragmentation into the heavy diquark is shown to underestimate the cross section values in comparison with the exact numerical calculations of complete set of diagrams .
the expected in tevatron experiments yield of baryons with two heavy quarks is evaluated as @xmath6-baryons and @xmath7-baryons at @xmath8 gev and @xmath9 of the baryon momentum and rapidity cuts , with account for the antiparticle yields .
ihep 95 - 87 + hep - ph/9507242 + * hadronic production of baryons , containing two heavy quarks * + a.v.berezhnoy , v.v.kiselev , a.k.likhoded + _ institute for high energy physics , protvino 142284 , russia _ + e - mail : [email protected] |
we start by analyzing the amplitude damping channel with kraus operators @xmath20 where @xmath21 is the probability of decay @xmath22 .
this map is agnostic on the direction of the @xmath23 and @xmath24 axis of the bloch sphere : a rotation around the @xmath25 axis leaves @xmath26 unchanged and adds an inconsequential phase factor to @xmath27 .
thus we can optimize the single - probe input state among the family @xmath28 .
the optimal state has @xmath29 , and its qfi is @xmath30 . to show that entanglement - assisted strategies perform better , we compare this qfi with the one of an entangled state of probe and ancilla @xmath31
. this state might not be the optimal state for the entanglement - assisted strategy , but it outperforms the previous one .
indeed its corresponding output state @xmath32 $ ] ( where the map acts only on the probe qubit and not on the ancilla ) has a qfi of @xmath33 ( also shown by @xcite ) , if one optimizes over @xmath34 for each @xmath21 .
even the simple choice @xmath35 is advantageous for all @xmath21 as its qfi is @xmath36 .
these qfis are compared in fig .
[ fig : adcsingle]a .
one observable that achieves the qcr bound for the last qfi is @xmath37 , with @xmath38 and @xmath39 . indeed , measuring the observable @xmath40 , the error on @xmath1 is @xmath41 this expression is optimized for @xmath42 , where the cramer - rao bound of the optimal single - probe state is beaten performing bell measurements between probe and ancilla . as discussed above , we can perform the optimization @xmath43 using a feedback strategy @xcite that uses an additional phase factor @xcite in the interferometer to drive the global phase to this `` sweet spot '' .
the ancilla - assisted advantage persists also when we use entangled probes .
the case without ancillas was analyzed in @xcite where an upper bound @xmath44 to the fisher information was given , which is achievable in the high - noise regime @xmath45 . in the case of two qubit probes
, one can also perform a numerical optimization of the qfi on generic two - qubit states ( presented in fig .
[ fig : adcsingle]b , solid line ) .
interestingly , the performance of the optimal state is similar to the `` noon '' state @xmath46 , which is known to be optimal in the noiseless case .
however , all these strategies can be beaten by using ancillas ( as was noted also in @xcite , although no explicit example was presented there ) . in fig .
[ fig : adcsingle]b , dotted - dashed line , we show the performance of a the four - qubit noon state @xmath47 of two probes and two ancillas : it beats the optimal state of two probes for noise levels @xmath48 , where its qfi is @xmath49 , which is optimal for @xmath50 , where @xmath51 is an integer .
one observable that achieves the qcr for the single - qubit probe is @xmath52 , where @xmath53 and @xmath54 . note that this particular four - qubit state is not necessary optimal .
the generalized pauli channel is described by @xmath55= ( 1- { p_1 - p_2 -p_3 } ) \rho + { p_1 } \sigma_x \rho \sigma_x ^\dagger + { p_2 } \sigma_y \rho \sigma_y ^\dagger + { p_3 } \sigma_z \rho \sigma_z ^\dagger , \label{eq : pauli}\ ] ] with @xmath56 , @xmath57 , @xmath58 pauli matrices , and @xmath59 , @xmath60 and @xmath61 probabilities .
two special pauli channels are the dephasing and the depolarizing noise .
dephasing noise corresponds to @xmath62 in and it was proved that ancillas offer no advantage over the unentangled probe @xcite : for both cases the optimal qfi of a single - qubit probe is @xmath63 . depolarising noise corresponds to having @xmath64 in : it describes an isotropic loss of coherence . as was also shown in @xcite , in this case ancillas
do help : indeed for a single - qubit probe , the optimal state is @xmath65 where the qfi is @xmath66 , whereas the qfi for a probe maximally entangled with an ancilla is @xmath67 , which is always greater .
this result is very surprising since the depolarizing channel can be seen as a time - sharing ( with probability @xmath68 ) of a noiseless channel and a channel where the state of the probe is replaced by a maximally mixed state @xmath69 ( useless for estimation ) . for both of these channels
the use of an ancilla gives no advantage : the ancilla becomes important only when they are randomly time - shared .
one observable that achieves the qcr for the single - qubit probe is @xmath70 , and for the ancilla - assisted case , @xmath71 ( identical with the one for the adc ) .
even though dephasing and depolarizing channels commute with the unitary @xmath15 , in general the pauli channel does not : we will consider the case where the noise acts only after @xmath15 .
for a single qubit probe in the generic initial state @xmath72 , the qfi is @xmath73 \end{aligned}\ ] ] ( @xmath74 stands for `` no ancilla '' ) , which is maximized for @xmath29 for all @xmath1 , and ( as before ) one can optimize over @xmath1 ( or , equivalently , @xmath75 ) using a feedback strategy . to prove that the presence of an ancilla is beneficial
, we consider a maximally entangled state @xmath76 of probe and ancilla . in this case
the qfi is @xmath77 we performed a numerical search over the parameter space @xmath78 . for all possible values of @xmath79 , the expression in eq . is larger than that of eq . for all @xmath75 .
to illustrate this , the comparison between and is shown in fig . [
fig : pauli ] , for the case when @xmath80 and @xmath81 .
the probe - ancilla observable @xmath82 achieves the qcr bound relative to the qfi of , with @xmath83 , @xmath84 .
note that when @xmath85 and @xmath86 , the qfi is unity : this represents the case of orthogonal noise when the ancilla strategy can recover the full information on the phase even in the presence of noise , as was pointed out in @xcite .
we propose here an experimental scheme to implement our proposal , based on the use of a single photon , where two qubits are encoded in the path and polarisation degrees of freedom .
such a scheme , that can simulate the ancilla - assisted strategy , is shown in fig.[fig : experimentalscheme ] .
the experiment uses a single photon which is prepared in a polarization - path entangled state as follows and the noise acts on the polarization degree of freedom .
the initial polarisation state of the photon in this scheme is @xmath87 .
a polarizing beam splitter ( pbs ) then transmits @xmath88 and reflects @xmath89 .
therefore , pbs1 acts as an effective cnot gate and puts into a polarization - path entangled state @xmath90 ( red solid = a , blue dashed = b ) .
the state is then transformed according to the unitary @xmath15 and the noisy channel .
different noise models can be added , e.g. using the techniques implemented in @xcite .
if the bit is flipped , the flipped component is directed by pbs2 and pbs3 onto another path , which interfere at bs2 . the half wave plate ( hwp )
rotates v polarization to h so that they interfere at the 50:50 beam splitter ( bs ) .
the which - arm statistics after the bs are effective projective bell measurements in this basis .
that is , at bs1 , the outputs correspond to projecting onto @xmath91 , and at bs2 , @xmath92 .
this scheme is easily implementable with present - day technologies , e.g. in @xcite similar schemes were experimentally realized and controlled noise was introduced in an ancilla - assisted scenario for different purposes .
in conclusion , we have studied the role of entangled ancillas in metrology for the important classes of qubit noise models .
we have shown that , for a single probe , in the presence of amplitude damping , depolarizing noise as well as general pauli noise , an entanglement - assisted scheme provides an advantage in the efficiency of phase measurement over the unentangled case for all ranges of noise regimes .
we also derived the optimal measurement procedures which achieve the cramer - rao bound .
bennett , s.j .
wiesner , phys .
lett . * 69 * , 2881 ( 1992 ) . c. h. bennett , p. w. shor , j. a. smolin , and a. v. thapliyal , phys .
* 83 * , 3081 ( 1999 ) . c. h. bennett , p. w. shor , j. a. smolin , and a. v. thapliyal , ieee trans .
theory * 48 * , 2637 ( 2002 ) , eprint quant - ph/0106052 .
z. shadman , h. kampermann , c. macchiavello and d. bruss , new j. phys .
* 12 * , 073042 ( 2010 ) .
v. giovannetti , s. lloyd , and l. maccone , phys . rev
. lett . * 96 * , 010401 ( 2006 ) .
v. giovannetti , s. lloyd , and l. maccone , science * 306 * , 1330 ( 2004 ) . w. van dam , g. m. dariano , a. ekert , c. macchiavello and m. mosca , phys . rev . lett . * 98 * , 090501 ( 2007 ) .
v. giovannetti , s. lloyd , and l. maccone , nature photonics * 5 * , 222 ( 2011 ) .
j. p. dowling , contemporary physics * 49 * , 125 ( 2008 ) .
w. dr , m. skotiniotis , f. frwis , b. kraus , phys .
* 112 * , 080801 ( 2014 ) .
g. arrad , y. vinkler , d. aharonov , and a. retzker , phys .
* 112 * , 150801 ( 2014 ) .
e. m. kessler , i. lovchinsky , a. o. sushkov , and m. d. lukin , phys .
lett . * 112 * , 150802 ( 2014 ) .
s. a. haine , s .
s. szigeti , phys .
a * 92 * , 032317 ( 2015 ) , s a. haine , s s. szigeti , m. d. lang , c .
m. caves , phys .
rev . a * 91 * 041802 , ( 2015 ) s. l. braunstein and c. m. caves , phys .
* 72 * , 3439 ( 1994 ) .
s. l. braunstein , c. m. caves , and g. milburn , annals phys .
* 247 * , 135 ( 1996 ) .
g. y. xiang , b. l. higgins , d. w. berry , h. m. wiseman , g. j. pryde , nature photonics * 5 * , 43 ( 2011 ) .
h. t. dinani and d. w. berry , phys .
a * 90 * , 023856 ( 2014 ) .
a. shaji and c. m. caves , phys .
a * 76 * , 032111 ( 2007 ) .
s. f. huelga , c. macchiavello , t. pellizzari , a.k .
ekert , m.b .
plenio , j.i .
cirac , phys .
* 79 * , 3865 ( 1997 ) .
m. kacprowicz , r. demkowicz - dobrzanski , w. wasilewski , k. banaszek , i. a. walmsley , nature photonics * 4 * , 357 ( 2010 ) .
r. chaves , j. b. brask , m. markiewicz , j. kolodyski , a. acn , phys .
. lett . * 111 * , 120401 ( 2013 ) .
l. maccone and g. de cillis , phys .
a * 79 * , 023812 ( 2009 ) .
r. ozeri arxiv:1310.3432 ( 2013 ) .
r. demkowicz - dobrzanski , j. kolodynski , m. guta , nature communications * 3 * , 1063 ( 2012 ) . b. escher , r. de matos filho , and l. davidovich , nature physics * 7 * , 406 ( 2011 ) .
m. jarzyna , r. demkowicz - dobrzanski , new j. phys . * 17 * , 013010 ( 2015 ) .
knysh , e.h .
chen , g.a .
durkin , arxiv:1402.0495 ( 2014 ) .
s. alipour , m. mehboudi , a.t .
lett . * 112 * , 120405 ( 2014 ) .
s. l. braunstein phys .
* 69 * , 3598 ( 1992 ) .
s. alipour , a. t. rezakhani phys . rev . a * 91 * , 042104 ( 2015 ) .
r. demkowicz - dobrzanski , k. banaszek , r. schnabel phys .
a * 88 * , 041802(r ) ( 2013 ) .
d. w. berry , h. m. wiseman phys .
* 85 * , 5098 ( 2000 ) ; d. w. berry , h. m. wiseman , j. k. breslin phys .
a * 63 * , 053804 ( 2001 ) ; d. berry , arxiv quant - ph/0202136 ( 2002 ) .
a. hentschel , b.c .
sanders phys .
lett . * 104 * , 063603 ( 2010 ) ; a. hentschel , b.c .
sanders phys . rev .
lett . * 107 * , 233601 ( 2011 ) .
durkin , j.p .
dowling phys .
lett . * 99 * , 070801 ( 2007 ) .
seshadreesan , s. kim , j. p. dowling , h. lee phys .
a * 87 * , 043833 ( 2013 ) .
holevo , _ probabilistic and statistical aspects of quantum theory _ ( north holland pub .
co. , amsterdam , 1982 ) . for the general case
see also : g.m .
dariano , c. macchiavello and m.f .
sacchi , phys .
a * 248 * , 103 ( 1998 ) . c. w. helstrom , `` quantum detection and estimation theory , '' ( academic press , new york , 1976 ) | entanglement - assisted quantum communication employs pre - shared entanglement between sender and receiver as a resource .
we apply the same framework to quantum metrology , introducing shared entanglement between the preparation and the measurement stage , namely using some entangled ancillary system that does not interact with the system to be sampled .
this is known to be useless in the noiseless case , but was recently shown to be useful in the presence of noise . here
we detail how and when it can be of use .
for example , surprisingly it is useful when randomly time sharing two channels where ancillas do not help ( depolarizing ) .
we show that it is useful for all levels of noise for many noise models and propose a simple optical experiment to test these results .
entanglement - assisted communication @xcite employs pre - shared entanglement between sender and receiver in addition to the signals sent through the channel .
this doubles the capacity of a noiseless channel , as illustrated in the well known superdense coding protocol @xcite , and is even more beneficial in the presence of noise @xcite . here
we apply the same framework to quantum metrology @xcite , which studies how quantum effects may aid parameter estimation . a quantum parameter estimation is composed of three stages ( fig . [ f : schema]a ) : the preparation stage , where some probe systems are initialized ; the sampling stage , where the probes interact with the system to be sampled ( this interaction encodes the parameter on the probes ) ; the measurement stage , where the probes are measured and the outcome is processed to yield the parameter estimate .
entanglement - assisted quantum metrology ( fig .
[ f : schema]b ) refers to the scenario in which the probes are entangled with an ancilla that does not participate to the sampling stage ( similar ideas have also been studied in the context of noisy channel estimation @xcite ) . then at the measurement stage a joint measurement is performed between probes and ancilla . it was recently realized that this is useful in the presence of noise @xcite , although it was known to be useless in the noiseless case @xcite .
here we detail how entangled ancillas can be used , what are the gains one can achieve and which explicit measurement strategies can achieve such gains , analyzing the most important qubit channels . we also propose a simple experiment that can test our results .
the entanglement - assisted scenario should not be confused with the generic use of entanglement in quantum estimation , that has been extensively studied previously , e.g. @xcite , showing how entangled probes achieve better precision than unentangled ones .
it is also different from the application of quantum error correction to metrology @xcite where all the systems are involved in the interaction ( fig .
[ f : schema]c ) , although some scenarios analyzed in refs . @xcite
go beyond this conventional quantum error correction scheme .
instead , in the entanglement - assisted scenario , the ancilla does not interact and can be considered as noiseless , if we suppose ( as is often the case ) that the noise is relevant especially during the interaction stage .
this translates into a reduced noise acting on the global state and a reduced resource count : indeed in quantum metrology the resource count refers to the number of times that the probed system is sampled @xcite .
so , the entangled ancillas should be accounted for as separate resources , as is done in quantum communication @xcite .
we will detail the increase in achievable precision in the presence of an entangled ancilla with respect to the one achievable in its absence , instead of dealing with the issue of whether one can beat the standard quantum limit or achieve the heisenberg bound .
indeed , it is known that the heisenberg scaling can not be achieved asymptotically for many of these noise models @xcite , although it can often be achieved in the non - asymptotic regime @xcite .
interestingly , it has been pointed out @xcite that the heisenberg scaling _ can _ be recovered through entanglement - assisted metrology even asymptotically in the very special case of orthogonal noise , even though only a lower scaling is achieved in the absence of ancillas @xcite .
however a general analysis of entanglement - assisted metrology was lacking up to now : our results show the attainable precision enhancement for the case of the pauli channels ( including the fully depolarizing case ) and of the amplitude damping . moreover ,
dephasing and erasure noise do not allow for any enhancement in the entanglement - assisted scenario , although entanglement among the probes is helpful @xcite .
our analysis then concludes the study of entanglement - assisted estimation for all relevant qubit channels .
most previous literature studies the precision through the quantum cramer - rao ( qcr ) bound , with the promise that such a lower bound to precision is significant because it is asymptotically achievable
. however , since we want achieveability in the non - asymptotic regime , we must provide estimation strategies and prove that they achieve the qcr , at least in a feedback scenario @xcite .
indeed , both the qcr and the error in the employed strategy may depend on the unknown parameter to be estimated @xcite , but a feedback strategy converges exponentially to the `` sweet spot '' @xcite , so that it has only a logarithmic cost in terms of resources @xcite . even though the rigorous way to determine phase errors is the holevo variance @xcite , as is usually done in the literature we will use the customary variance , since the two match for sufficiently precise estimation strategies .
the qcr @xcite is a lower bound to the precision @xmath0 of the estimation of a parameter @xmath1 : under reasonable hypotheses , @xmath2 where @xmath3 is the number of times the estimation is repeated , and @xmath4 is the quantum fisher information ( qfi ) associated to the global state @xmath5 of probes and ancillas ( after the interaction @xmath6 with the probed system ) . the qfi is @xmath7 where @xmath8 , @xmath9 and @xmath10 are the eigenvalues and eigenvectors of @xmath5 .
the map @xmath6 encodes the phase parameter @xmath1 onto the probes : @xmath11 $ ] , where @xmath12 is the initial state . without loss of generality , for qubit channels
we suppose that the phase is encoded onto the computational basis by the unitary @xmath13 .
for the sake of simplicity we will consider situations where the noise maps @xmath14 acts after @xmath15 , namely @xmath16 .
if the noise map and @xmath15 do not commute , this is an important restriction of our analysis ( required to make the problem tractable , as nontrivial effects arise otherwise @xcite ) , but it is not a restriction for the erasure , amplitude damping and depolarizing noise models which commute with @xmath15 . to find the best bound , one must maximize the qfi , which in general depends both on the input state @xmath12 and on the unknown parameter @xmath1 .
the former optimization depends on the noise map , the latter can be taken care of using feedback mechanisms @xcite .
for all noise maps , we can use the convexity of qfi @xcite to choose a pure input state @xmath17 . indeed , supposing that @xmath18 , we have @xmath19)\leqslant \sum_j\lambda_jj(\lambda_\varphi[|j\>\<j|])$ ] .
( the qfi is not convex if @xmath9 depended on @xmath1 , but an extended convexity still holds @xcite . ) |
the c@xmath0h radical ( ethynyl ) is an intermediate of significant importance in the chemistry of hydrocarbon combustion , particularly in acetylene - air mixtures.@xcite it is also related to the substituted ethynyl species which are the crucial intermediates in the hydrogen abstraction - acetylene addition ( haca ) mechanism that leads to the formation of polyaromatic ring species on the way to soot in fuel rich environments .
@xcite recently , a haca - like mechanism has also been invoked to explain the growth of single - walled carbon nanotubes in low temperature catalyzed - chemical vapor deposition syntheses using acetylene and inert gas flowing over mildly heated metal - containing catalysts . in this model ,
the adsorbed c@xmath0h intermediate acts as a continuously regenerated `` catalyst '' in the process.@xcite the radical was one of the earliest molecules identified in interstellar space via its microwave spectrum .
@xcite models of interstellar chemistry@xcite invoke reaction pathways involving c@xmath0h to explain the presence of polyaromatic hydrocarbons ( pahs ) in the interstellar and star - formation regions in space .
finally , investigations of the atmosphere of titan , a moon of saturn , have shown that photochemistry in its upper atmosphere leads to significant concentrations of polyynes , assumed to be produced in similar reaction sequences to those involved in the terrestrial formation of pahs and soot , and including ethynyl - like intermediates in a chain reaction synthetic pathway .
@xcite in support of the extensive reaction network modeling in multiple environments , rate constants have been measured for reactions of ethynyl radicals with various partners over a range of temperatures .
@xcite following the astronomical observation of the ethynyl radical,@xcite the gas - phase microwave spectrum was recorded and the molecular structure determined in the 1980s.@xcite prior to this , laboratory observations had been limited to work in low temperature matrices.@xcite the infrared spectrum of the radical is complicated by the presence of a low - lying _ a _ @xmath3 excited electronic state which splits due to the renner - teller effect .
the three resulting electronic surfaces are vibronically coupled , so the vibrationally excited levels of the ground _ x _ @xmath7 state all contain some @xmath8-state character .
the curl@xcite group reported extensive work on the infrared spectrum in the 3@xmath9 m ( c - h stretching ) region and hirota@xcite recorded and analyzed the spectrum in the c - c stretching region and identified hot bands originating in the low frequency bending vibrational mode .
all observed infrared spectra showed effects resulting from the strong vibronic interactions and local @xmath10-dependent perturbations due to background levels derived from combinations of lower frequency modes .
many more spectroscopic studies were reported in subsequent years and pfelzer__et al.__@xcite survey the work up to the mid 1990 s .
detailed understanding of the vibronic interactions leading to the observed spectra has relied upon _ ab initio _ methods to label the vibronic levels observed and to guide assignments , particularly at higher internal energies .
peri _ _ et al.__@xcite made a state - of - the - art calculation of the vibronic states up to approximately 5000 @xmath2 in 1992 , and were able to identify the vibronic levels observed to that time .
later , tarroni and carter@xcite reported more accurate calculations based on variational solutions of the nuclear motion on the coupled high level _ ab initio _ potential energy surfaces .
importantly , the second of these papers reported absorption intensities to vibronically excited levels from the zero point level of the ground state thereby allowing vibronic state labels to be assigned to observed spectra with confidence in energy regions where the density of states is large .
in low temperature matrices , forney _ _ et al.__@xcite had reported infrared spectra to much shorter wavelengths than previously seen , and hsu _ _ et al.__@xcite identified many higher lying vibronic states in laser - induced fluorescence ( lif ) spectra in the @xmath11 band system .
the vibronic structure out to similar energy levels was also observed in an anion photodetachment experiment by zhou _ _
et al.__@xcite recently , sharp - williams _ et al . _
observed new gas phase spectra of c@xmath0h between 3600 and 4100 @xmath2 in a jet cooled sample,@xcite assigning the spectra by comparison to the tarroni and carter s results@xcite and identifying multiple local perturbations in the rotational structure of the upper vibronic levels .
additional bands of c@xmath0h in the 4250@xmath124550 @xmath2 region have very recently been analyzed in a fourier transform emission spectrum , and a band near 6340 @xmath2 has been characterized by diode laser spectroscopy in a concentration modulated discharge source.@xcite both the calculations@xcite and matrix spectrum@xcite show that the most intense bands in the entire infrared and near - infrared spectra of the radical lie around 7100 @xmath2 or 1.4 @xmath9 m .
the vibronic levels at this energy are strongly mixed , but derive much of their intensity from the near - vertical transition to the @xmath8 ( 0,0,2 ) state , in a region of spectrum where no gas phase spectra have yet been reported .
we expect that the spectra in this wavelength region will be advantageous for kinetic or non - intrusive remote sensing applications .
in addition to the strength of these near - infrared transitions , efficient spectroscopic sources and detectors are available at these wavelengths , and optical fibers are well suited for coupling probe radiation in and out of the sample . in this work ,
we report measurements of the c@xmath0h spectrum between approximately 6630 and 7135 @xmath2 by diode laser transient absorption spectroscopy . we have identified and rotationally assigned three band systems in this region , based on ground state combination differences @xcite and comparison to the predictions of tarroni and carter.@xcite in the course of our spectroscopic studies , we were initially puzzled to find that the decay rate of the vibronic ground state
c@xmath0h radical signals was several times slower than expected , based on published room temperature rate constants @xcite for the reaction of c@xmath0h with our chosen photolytic precursor , cf@xmath13c@xmath6h , at the sample pressures we had chosen for convenient collection of radical spectra . a time - dependent spectral simplification observed in ar : cf@xmath13c@xmath6h mixtures that we attributed to vibrational thermalization was not observed in ar : c@xmath0h@xmath0 mixtures , where a dense set of both assigned and unassigned lines all decayed at comparable rates .
the ground state radical lifetimes were also longer than anticipated in acetylene , based on published thermal rate constants and the sample composition . to investigate these effects further , we modified the detection system slightly for fixed frequency kinetic studies , and undertook some pressure and dilution studies , to see the effects of competitive reaction and relaxation on the time - dependent spectra .
the kinetic observations and discussions are reported following the spectroscopic observations and analysis below .
one important reason to characterize these strong bands of c@xmath0h in the 7000 @xmath2 region is for future use in kinetic investigations . in this preliminary study ,
the complex interplay of hot formation , vibronic relaxation and energy - dependent reactivity is evident but far from fully characterized .
useful comparisons can be made with previous spectroscopic work where conditions reported for observing spectra of different hot or cold - band transitions in c@xmath0h had been optimized .
additional comparisons can be made with previous kinetic spectroscopy studies of c@xmath0h reactions , where either complete or negligible vibronic relaxation prior to reaction is a recognized precondition for the measurement of reliable thermal rate constants .
hirota and co - workers @xcite had used 193 nm photolysis of a 2:3 mixture of c@xmath0h@xmath0:ar at 2 torr total pressure to search for c@xmath0h transitions in the 1850 @xmath2 region .
they reported many lines , but none that could be assigned to cold bands of c@xmath0h , unless they replaced the ar with h@xmath0 , d@xmath0 , co or ch@xmath14 as a buffer gas .
he and n@xmath0 were also reported to be ineffective at vibrational relaxation .
these observations are hard to understand given current understanding of the kinetics .
the acetylene pressure was high enough to make the estimated lifetime of c@xmath0h shorter than the microsecond detector response , and adding hydrogen would seemingly only decrease the lifetime by means of an additional reactive loss channel.@xcite clear spectroscopic evidence of the @xmath15 fundamental of c@xmath0h was nevertheless obtained .
hsu et al.@xcite used pump - probe delay time - dependent lif intensities to help assign a selection of bands originating in various vibrationally excited states of c@xmath0h ( @xmath16 ) . they reported successively longer lifetimes for lower energy states , but no resolved growth kinetics for any bands , following 193 nm excitation of c@xmath0h@xmath0/h@xmath0 mixtures . in a subsequent report,@xcite the substitution of sf@xmath17 for ar , he , or h@xmath0 in the c@xmath0h@xmath0 photolysis sample was found to increase the relative intensity of otherwise weak lif bands originating in low bending levels of c@xmath0h , consistent with enhanced vibrational relaxation by sf@xmath17 .
no lif transitions originating from the ground vibrational state of c@xmath0h ( @xmath16 ) have been reported , presumably due to poor franck - condon factors . for kinetic spectroscopy studies , vibronic relaxation prior to reaction
is typically tested by varying the pressure or choice of an inert buffer to verify insensitivity of the observed reaction rates.@xcite direct measurements of hot band relaxation@xcite are preferable , but not always feasible , to confirm the timescale of vibrational thermalization . using the newly assigned bands
, we report the results of our investigation into pressure - dependent kinetics of state - resolved relaxation in competition reaction with two c@xmath0h precursors : c@xmath0h@xmath0 and cf@xmath18c@xmath0h .
the vibronic relaxation appears to be slower than has often been claimed .
the transient absorption spectra were recorded using a spectrometer very similar to that described previously .
@xcite ethynyl radicals were formed by 193 nm arf excimer laser photolysis of a sample of cf@xmath19c@xmath20h ( trifluoromethyl acetylene , or 3,3,3 trifluoropropyne ) obtained from synquest laboratories , diluted in argon .
a sacher laserteknik tunable diode laser , model tec 500 , with output centered at 1450 nm was the spectroscopic source .
one portion of the laser power was sent to a wavemeter ( burleigh wa-1500 ) , and another portion was used as a reference beam for active noise reduction .
the main beam was directed through a herriott - type absorption cell with 35 passes between mirrors separated by 2 meters .
signal and reference beams were imaged onto an _ a.c._-coupled dual ingaas photoreceiver ( new focus 1617 ) . the excimer laser beam , with typical pulse energies of 30 mj in a 3 @xmath21 beam , passed through the center of the absorption cell and overlapped the probe beam in the central portion of the cell .
we estimate the effective absorption path length through the photolysis volume was about 25 meters .
mixtures of ar and cf@xmath19c@xmath20h in a 1:1 ratio were prepared in a large storage flask at pressures of approximately one atmosphere .
this gas mixture flowed slowly at 5 sccm though the cell .
total pressures used for the spectral scans were typically 1 torr .
mixtures of acetylene in ar were also used for comparison in some spectral regions , producing similar but weaker spectra with generally faster decay rates .
because of the stronger and longer - lasting signals , the systematic spectroscopy scans were carried out with the cf@xmath18c@xmath20h precursor .
some additional exploratory measurements of the relaxation and reaction kinetics were made at fixed probe frequencies on selected ground state ethynyl radical transitions , as a function of precursor and inert buffer gas ( argon ) concentrations .
these measurements used a _
d.c._-coupled balanced photoreceiver ( new focus 1817 ) , with sequential measurement and off - line subtraction of on- and off - resonant signals to remove the reproducible background associated with acoustic transients and increase the fidelity of the transient signals .
tests with varying flow rates or a reduction of the excimer repetition rate below its typical value of 10 hz showed no sign of secondary photolysis or sample depletion .
initial spectroscopic searches for the strong bands observed in the low temperature matrix spectrum @xcite and predicted by tarroni and carter@xcite were carried out by photolysis of a slowly flowing sample of cf@xmath18c@xmath20h in ar at 193 nm .
transient absorption bands presumed to be due to c@xmath20h were detected throughout the search region .
further support for the attribution of absorption lines to c@xmath20h was obtained by observing some of the same lines in spectra using c@xmath20h@xmath20 as the photolytic precursor .
spectra recorded immediately following the photolysis pulse were much more complicated than spectra at longer delay times , which we attribute to collisional cooling of an initially formed hot distribution .
the time development of signals in samples containing c@xmath6h@xmath6 was quite different to that observed with cf@xmath18c@xmath20h , as might be expected due to differences in absorption cross section , quantum yields for c@xmath6h , the initial state distributions , relaxation kinetics , and rates of reaction with the precursors .
some further kinetic investigations are presented below , but for the purposes of the spectroscopic studies , cf@xmath18c@xmath20h was found to provide larger signals and simpler spectra after partial relaxation , when compared to c@xmath22h@xmath6 samples .
therefore , all of the subsequent spectroscopic measurements were made using cf@xmath18c@xmath20h precursor in ar .
examples of a section of the observed spectra near 7125 @xmath2 at short and long time delays with the cf@xmath18c@xmath20h precursor are shown in figure [ figure1 . ] .
this spectrum was obtained by averaging 40 photolysis laser shots per wavelength step ( typically 0.005 @xmath2 ) in a 1:1 mixture of cf@xmath18c@xmath20h and argon at a total pressure of 1.0 torr .
figure [ figure1.](a ) shows a complex c@xmath6h spectrum at zero time delay following the photolysis pulse with a time gate of 1.0 @xmath23s .
after a 3.6 @xmath23s time delay , the congestion is substantially reduced , leaving a much simpler spectrum , as shown in figure [ figure1.](b ) .
two branches , each with a bandhead , can be identified and assigned to _
_ r__@xmath24 and _ _ r__@xmath6 branches of a @xmath25 band .
perturbations starting at _ _
r__@xmath6(15.5 ) and _ _ r__@xmath24(16.5 ) are also evident in figure [ figure1 . ] .
the strong band observed in the ne matrix spectrum at 7064 @xmath2 @xcite has been attributed to a combination of two bands,@xcite a @xmath25 band calculated at 7106 @xmath2 and a @xmath26 band at 7099 @xmath2 , consistent with the sub - band identifications above .
the @xmath1 lower state in both of these transitions is the ground state of the radical , and analysis of the cold spectra proceeded by searching for ground state combination differences based on the ground state constants of gottlieb et al.@xcite assuming vibronic bands of the type predicted .
@xmath1 state energy levels are described by a rotational quantum number ( @xmath10 ) and a parity label ( + or @xmath12 ) with @xmath27 ( f@xmath28 ) or @xmath29 ( f@xmath0 ) , where @xmath30 is the total angular momentum excluding spin .
the lowest level @xmath31 ( @xmath32 , f@xmath33 ) has ( + ) parity and the parity labeling alternates with @xmath30 for @xmath1 states . as a result , _
p- _ and _ r- _ branches are doubled .
the doublet splitting is illustrated in figure [ figure2 . ] ( b ) for lines in the @xmath34 transition at 7111.4 @xmath2 and is due to the difference in the spin - rotation splittings between the ground and the excited states .
the early time spectrum in figure [ figure2.](a ) shows additional unassigned , presumably hot band , lines that are gone at the time of the delayed spectrum shown in figure [ figure2.](b ) .
a more detailed comparison of the time dependence of ground state and vibronically excited states for different sample mixtures and pressures is presented in section [ sec : kinetics ] below .
the energy levels of the @xmath3 state are also labeled by rotational ( @xmath10 ) and parity quantum numbers .
two sub - states @xmath35 and @xmath36 result from spin - orbit coupling .
analysis below shows the spin - orbit constant , a , is small and negative , therefore the @xmath36 and @xmath35 energy levels map onto f@xmath28 and f@xmath0 , respectively . as a result ,
6 main branches ( p@xmath28 , p@xmath0 , q@xmath28 , q@xmath0 , r@xmath28 , and r@xmath0 ) were observed .
the four additional satellite branches , with f@xmath37f@xmath0 were not observed here .
seeking further support for some of the @xmath25 assignments , we searched the 6700 @xmath2 region for hot band transitions coming from the _ x _ ( 0,1@xmath38,0 ) ( @xmath39 symmetry ) to the same upper level states .
the fundamental bending vibrational frequency is 371.6034(4 ) @xmath2 ; rotational and fine structure parameters for this level were determined by kanamori et .
@xcite we did not find these hot band transitions above our noise level at the calculated frequencies , likely due to poor franck - condon factors .
instead , an additional @xmath26 band was identified in this region .
figure [ figure3 . ] shows a portion of this spectrum at 0 @xmath9s and 3.6 @xmath9s , with some _
_ r__@xmath24 , _ _
r__@xmath22 , _ _ p__@xmath24 and _ _ p__@xmath6 branch lines labeled .
the band origin of this @xmath26 is at 6695.678 @xmath40 .
complete data files of the two spectral regions between 6630 and 7135 @xmath2 studied here , at three different times ( 0 , 2.6 and 3.6 @xmath23s ) are included as supplementary material.*(supplemental material ) * the @xmath1 rotational hamiltonian can be written as : @xmath41 leading to the energy level expression : @xmath42 the b@xmath43 , d@xmath43 , and @xmath44 rotational and spin - rotational ground state constants were fixed to the literature values @xcite .
the energies for the @xmath3 state were modeled by including the spin - orbit interaction ( a ) , rotational constant ( b ) quartic centrifugal distortion correction ( d ) , the spin - rotation ( @xmath45 ) , and the @xmath46-doubling ( @xmath47 ) . in the hund s case ( a ) basis limit , the effective hamiltonian for the @xmath3 state is @xcite :
@xmath48 in equation [ eq3 ] , @xmath49 and @xmath50 are the shift operators of @xmath51 , the total angular momentum in the absence of nuclear spin , and of @xmath52 , the total electron spin , and @xmath53 is the azimuthal coordinate of the unpaired electron .
the energies of the @xmath3 state were determined by numerical diagonalization of the hamiltonian matrix representation of dimension 4x4 constructed in a sequentially coupled , non - parity conserving hund s case ( a ) basis set .
the energy levels can be modeled using either the hund s case ( a ) basis set , which is a good approximation for the low - rotational levels or hund s case ( b ) basic set , which is better for higher rotational levels ( n @xmath54 11 ) .
the calculated energy levels are the same in both hund s case ( a ) and case ( b ) limits .
@xcite in the @xmath26 band at 7088 @xmath2 , 80 transition wavenumbers were identified via combinational differences using the ground state parameters constants of gottlieb et al.@xcite however , only 32 transition wavenumbers were entered directly into the fit to the upper state parameters , due to multiple perturbations in the upper state energy levels .
deviations in excess of 0.015 @xmath2 , more than three times the measurement error , were considered evidence of local perturbations strong enough to exclude a measured transition from the fit .
the measured line centers are presented in table [ table1 ] , along with the assignments and differences between the calculated and observed transition wavenumbers .
similarly , 66 transition wavenumbers in the @xmath26 band at 6696 @xmath2 were identified by combination differences using the same ground state parameters , and 49 transition wavenumbers were used in the fit .
the line centers are presented in table [ table2 ] , along with the assignments and differences between the calculated and observed transition wavenumbers .
the transitions from @xmath34 bands at 6696 and 7088 @xmath2 were
least squares fit to the simple energy expressions in equation [ eq2 ] , described above , to determine the excited state constants including b@xmath55 , d@xmath55 , @xmath56 and the origin for both upper @xmath57 states .
the standard deviation of the fits were 0.004 @xmath58 and 0.009 @xmath58 for the bands at 6696 and 7088 @xmath2 respectively .
the optimized parameters and associated errors are given in table [ table4 ] .
in the @xmath25 band at 7110 @xmath2 , 69 transition wavenumbers were identified via combinational differences using the ground state parameters .
of these , 57 were included in the least squares fit to determine the excited state parameters _
_ a__@xmath55 , _ _
b__@xmath55 , _ _
d__@xmath55 , @xmath59 , @xmath60 , and the band origin .
the standard deviation of the fit is 0.010 @xmath58 .
the optimized parameters and associated errors are given in table [ table4 ] .
the transient absorption of c@xmath6h transitions for different photolytic precursors and buffer gas pressures offers a view of the competition between collisional thermalization and reactive loss .
figure [ figure4 . ] illustrates the time - dependent absorption signals observed on the _
r_(4 ) line of the 7088 @xmath58 band at a constant cf@xmath13c@xmath6h precursor partial pressure of 50 mtorr with decreasing amounts of ar buffer gas .
small corrections have been applied to normalize for changes in photolysis and probe laser intensity variations , so that the percent absorption signals are directly comparable .
the rise rate becomes progressively slower as the ar pressure decreases , and the peak amplitude decreases . at 5 and 10
torr , the decay rate is about the same , and in good agreement with the published @xcite room temperature rate constant of 3.5@xmath6110@xmath62 @xmath63 molec@xmath64 s@xmath64 for the c@xmath6h + cf@xmath13c@xmath65h reaction , and the partial pressure of cf@xmath13c@xmath6h .
the delayed maximum and slower decay at the lowest total pressure of 0.25 torr superficially resembles the expectations of slower vibrational relaxation . when the vibrational relaxation cascade is no longer rapid compared to the reactive loss
, the observed decay rate will no longer be a measure of the reaction rate , as the relaxation will continue to provide a delayed source of radicals in the probed energy level .
the absolute amplitudes tell an additional story , however . if the changing ar pressure only affected the thermalization rate , and the reaction rate depended only on the constant partial pressure of cf@xmath13c@xmath20h
, the observed time dependence would be described by an exponentially decaying total population of c@xmath6h times a monotonically increasing , time - dependent fractional population in the detected state .
a signal probing c@xmath6h ( 0,0,0 ) at low ar pressure could never exceed the signals at higher ar pressure under these conditions , contrary to our observations .
the most likely explanation appears to be that the unrelaxed population in higher energy states does not react as rapidly as the thermalized c@xmath6h radicals , so that delayed relaxation extends the lifetime of the total c@xmath6h radical pool .
a selection of r - branch lines probing rotational states from @xmath30=4 to @xmath30=15 in the 7088 @xmath2 band were measured to assess the time - dependent rotational distributions , and to test for the possibility of high rotational states affecting the kinetics . for a 1% mixture of cf@xmath18c@xmath0h in ar at 5 torr ,
the normalized time - dependent signals were superimposable throughout the rise and fall , as one would expect for rotational thermalization faster than the @xmath9s time scale .
a plot of the data is included in the * supplementary material , figure s1*. we note that the apparent temperature of the stationary rotational distribution would be 270 @xmath66 10 k if the rotational line intensities are interpreted with unperturbed line strength factors .
the apparently colder than ambient rotational temperature can be explained by an increasing rotational mixing with spectroscopically dark levels at higher @xmath10 .
the upper level of this transition is a @xmath67 state whose @xmath68state character is predominantly @xmath69.@xcite this level will be mixed with the dark @xmath70 components of the same ( 030 ) level in the renner - teller mixed @xmath8 state , via a _
k_-type resonance@xcite that includes a @xmath71dependent mixing term .
we have modeled this interaction and confirmed that a mixing coefficient defined as a fractional contribution of dark state wavefunction of @xmath72 , when combined with this @xmath73dependence , accounts the observed reduction in the intensity of the higher @xmath10 lines in the band . an analogous but contrasting set of pressure dependent measurements in argon / acetylene mixtures is illustrated in figure [ figure5 . ] . as with cf@xmath13c@xmath6h ,
increasing the ar partial pressure at fixed precursor pressure makes the signal rise more quickly to a larger maximum and a faster decay .
higher pressures are required to reach a limiting decay rate , which at 10 torr , is still about 30% slower than expected , given the 50 mtorr partial pressure of c@xmath0h@xmath0 and a room temperature rate constant of 1.3 x 10@xmath74 @xmath75 molec@xmath76 s@xmath76 .
@xmath77@xcite apparently even in the presence of ar at 10 torr the vibrational thermalization is not yet complete on the time scale of reaction with c@xmath0h@xmath0 .
the differences between cf@xmath13c@xmath6h and c@xmath0h@xmath0 can be further illustrated by comparing the growth and decay of an absorption signal originating in the vibrationless state of c@xmath0h with an unassigned hot band line observed at 7111.470 @xmath2 with either precursor .
this line was illustrated in figure [ figure2 .
] , and is typical of many , observed to decay quickly as the ground state lines grew in . figure [ figure6 . ] compares these signals for two different probed transitions on samples of the two photolytic precursors , each prepared as a 200:1 mixture of ar : precursor , measured at 10 torr total pressure .
the large signals probe the @xmath78 level of the ground state , and illustrate the smaller signals and faster reaction for c@xmath0h@xmath0 than for cf@xmath13c@xmath6h .
the smaller , faster signals , expanded in the inset figure contrast the hot band kinetics in the same two samples . without knowing the exact energy of the probed state , no quantitative conclusions can be drawn , but it can be seen that even a 200:1 ratio of ar buffer to precursor is insufficient open up a clear separation of time scales for vibrational relaxation and reactive loss of the vibrationless c@xmath0h in the case of c@xmath0h@xmath0 , since the decays of the vibrationless and hot band levels are similar , while it is adequate in the case of cf@xmath19c@xmath22h .
this may be due to an initially higher extent of internal excitation in the acetylene photodissociation , perhaps also affected by differences in the intrinsic competition between reaction and relaxation for collisions of the hot radicals with the two different precursor molecules .
the initial c@xmath0h internal energy distribution of the photofragments produced in acetylene photodissociation has been characterized by h atom rydberg time - of - flight measurements .
@xcite the structured kinetic energy spectra indicate preferential excitation in high bending states , low rotational excitation , and a significant population at peaks near the energy of the @xmath8 state .
the available energy for c@xmath0h fragments in the 193 nm photodissociation of cf@xmath18c@xmath0h is about 30% higher ( 198 kj mol@xmath76 vs. 151 kj mol @xmath76 ) , @xcite but the additional degrees of freedom in the cf@xmath18 co - fragment compared to h atoms provide statistical grounds to expect a cooler c@xmath0h internal energy distribution from cf@xmath18c@xmath0h than from c@xmath0h@xmath0 . no direct comparison has been reported , however , to our knowledge . a striking difference in the relative intensities of prompt and delayed spectra under the high concentration conditions initially used to investigate the spectroscopy was one of the observations that led us to the current kinetic measurements . in our preliminary spectroscopic measurements ,
a substantial spectral simplification was observed in the cf@xmath18c@xmath0h : ar samples during the first few microseconds , as illustrated in figures [ figure1.]-[figure3 . ] . selected short scans of c@xmath0h@xmath0:ar samples at the same 1:1 mixing ratio and total pressure of 1
torr showed the same transitions , different initial relative intensities , a more rapid decay , but no significant differences in the decay rates of the lines assignable to ground state levels and other unassigned lines that behaved like hot bands in the cf@xmath18c@xmath0h : ar sample .
we now attribute this behavior to the dominance of the reaction of c@xmath0h with c@xmath0h@xmath0 in competition with less efficient vibrational relaxation by either ar or c@xmath0h@xmath0 .
we believe most of the reactive loss seen under these conditions is occurring from vibrationally excited c@xmath0h .
previously , there have been no reported gas phase measurements of c@xmath6h in the 6700 @xmath40 and 7100 @xmath58 regions studied here .
however , the origins determined for all three bands identified and assigned are in good agreement with the calculated values from tarroni et al .
@xcite compared to the earlier matrix work,@xcite a new @xmath26 band was identified and assigned to the transition from the ground state to the x(0,8@xmath79,2 ) state again based on tarroni and carter.@xcite this state has a very small fraction of @xmath80 electronic state mixing , and was assigned to the vibrational combination level of @xmath81 state .
the standard deviation for the fit of this band is smaller than the other two observed bands in this region , possibly reflecting its less mixed character .
we estimate the overall intensity ratio between the three bands at 6696 , 7088 and 7110 @xmath2 is about 1:4:2 , respectively , which is in good agreement with the calculated intensity ratios.@xcite neglecting small nuclear hyperfine coupling effects , the observed discrete @xmath10-dependent perturbations require perturbing levels of the same @xmath10 and parity to lie nearby in energy . by examining plots ( figure [ figure7 . ] ) of observed@xmath12calculated energies against @xmath82 , for the upper @xmath1 state of the band at 7088 @xmath2 , local perturbations at @xmath83 ( @xmath12 parity ) , @xmath84 ( + parity ) , and for all @xmath10 associated with @xmath85 can be seen .
for the first two levels , only one @xmath10-component of a given @xmath30 is perturbed , implying that the perturbing level must belong to a state where the parity does not@xcite alternate with @xmath30 , _
i.e. _ probably a nearby @xmath86 state . however , the observed @xmath87 state at 7110 @xmath2 is _ not _ the perturber because @xmath83 ( @xmath88 parity ) of the @xmath1 state is pushed down by a state that has @xmath83 at higher energy , while the @xmath83 ( @xmath12 parity ) of the 7110 @xmath2 @xmath87 state has a lower energy , by approximately 0.5 @xmath2 , as shown in figure [ figure8 . ] .
the observed @xmath3 state level in question is also perturbed and pushed down in energy , as can be seen in table iii .
similarly , @xmath89 of the @xmath1 state lies well below the energy of the corresponding level in the @xmath3 state which therefore can not cause the observed @xmath1 perturbation either .
we conclude the perturbing level is a spectroscopically dark state .
supplementary data available@xcite for the published calculations list all @xmath67 and @xmath3 levels up to 10 000 @xmath2 in energy , but there is no other state which could plausibly lead to the low - j perturbations , implying it is actually a @xmath90 or higher angular momentum state . above @xmath91 ,
the 7088 @xmath2 @xmath1 state suffers a much larger perturbation as can be seen in figure [ figure7 . ] .
this perturbation affects both @xmath92 and @xmath12 parity levels equally so is qualitatively different to that seen for the two lower rotational levels .
however , at these @xmath10-values , all c@xmath0h states will be close to case(b ) , and so the arguments used to infer the character of the perturber above can not be applied .
the @xmath93 level is calculated@xcite to lie some 40 @xmath2 below the 7088 @xmath2 experimental level , and this could plausibly be the source of the observed high-@xmath10 perturbations in 7088 @xmath2 band .
the @xmath25 band at 7110 @xmath2 shows strong @xmath94 branches and much weaker _ p _ branches .
for example at the position of @xmath95(12.5 ) and @xmath96(11.5 ) , the observed spectroscopic features are barely above the noise .
this unusual intensity pattern is not unknown for vibronic transitions involving @xmath97 and @xmath1 states@xcite and occurs when @xmath981 perturbations are present .
the @xmath99 branch intensity pattern for transitions terminating in the mixed level can be very complex .
however , the @xmath981 perturbation always affects the intensity of @xmath100 = + 1 ( @xmath101-branch ) and @xmath881(@xmath102-branch ) in the opposite direction .
if the perturbation is strong , a whole branch ( e.g @xmath101-branch ) can be missing while the other ( e.g. @xmath102-branch ) gets much higher intensity as was observed in nco by dixon @xcite .
the numerous additional lines observed in the prompt spectrum of the recorded region are likely hot bands of c@xmath0h . from the generally reliable excited vibronic energies calculated by tarroni and carter@xcite and experimental rotational constants for many low - lying excited levels
, we may hope to identify some of these additional bands using combination differences , although we have so far been unsuccessful in this attempt .
clearly , the kinetic analysis of vibronic relaxation would be more compelling with secure spectral assignments . because many free radicals are formed with substantial internal energy , and undergo rapid reactions that can compete with relaxation , there is recent interest in improving theoretical treatments of non - thermal effects in combustion @xcite and atmospheric chemistry .
@xcite energetic radicals or adducts may react completely , prior to the establishment of any local thermodynamic equilibrium of internal states , under conditions when the reaction rates and even branching ratios may be rapidly changing with internal energy .
the case of c@xmath6h may be a good place to investigate the behavior at intermediate pressures of inert colliders , where the competition between relaxation and reaction can be tuned and observed , rather than deliberately suppressed .
the vibronic mixing of the _ a _ @xmath3 and _ x _ @xmath1 states in c@xmath6h can make the energy dependent reactivity more complex than in other radicals with a simpler electronic structure . the time dependence of ground state absorption signals for a photolytically generated hot radical like c@xmath0h will depend on the initial state distribution , the vibronic relaxation cascade due to energy transfer collisions with both inert and unsuccessfully reactive collision partners , as well as the reactive loss of the radicals in their evolving , nonthermal energy distribution .
the energy dependence of the reaction rate plays a well - known and central role in the conventionally measured temperature - dependent rate constants of thermalized radicals .
the energy transfer characteristics are typically less easily measured under the high energy conditions relevant to chemical activation or photodissociation .
it is , however , frequently possible to devise conditions when the subtle details of the energy transfer are unimportant , and internal energy relaxation eigenvalues are well separated from the chemically significant eigenvalues of the master equation that describes the coupled reaction and relaxation processes .
@xcite previous studies have shown dramatically increased reactivity of unrelaxed @xcite and partially relaxed @xcite c@xmath0h radicals from the photodissociation of acetylene in collisions with h@xmath20 , for example .
the thermal rate constant for this reaction near room temperature is 5.7 @xmath61 10@xmath103 @xmath104 molec@xmath76 sec@xmath76 , increasing sharply with temperature , following nonlinear arrhenius behavior.@xcite it may be important to distinguish between the temperature dependence of the thermal rate constant and the variable reactivity of internally excited states of the radical .
theoretical calculations@xcite of possible vibrational mode - specific effects in this reaction and its reverse do not find vibrational enhancement of the reaction with stretching excitation of c@xmath0h and even find an inhibition with bending excitation .
the potential energy surface@xcite used for the full - dimensional dynamics calculations@xcite does not , however , include the low energy _ a _ state , which has a barrierless path to h + c@xmath0h@xmath0 , according to the calculations of peeters et al.@xcite the low energy _
a _ @xmath3 state of c@xmath0h may contribute further complexity to an energy - dependent reactivity , as transition states and barriers to reactions may be quite different for _ a _ and _ x _ state interactions .
the vibronic mixing of _ a _ and _ x _ states is strong enough , however , that they may not be kinetically distinct species , but rather an energy - dependent mixture , with a correspondingly mixed reactivity .
@xcite even the lowest vibrationally excited states nominally belonging to the _ x _ @xmath7 state of c@xmath0h are calculated to have about 0.05 @xmath8-state character , and the zero - order vibrationless level of the @xmath8 state is quite evenly fragmented among five vibronic levels with energies between 3604 and 4593 @xmath2 , none of which have more than 40% @xmath8 state character.@xcite in contrast to the reaction with hydrogen , the thermal rate constant for the reaction of c@xmath0h with acetylene is much faster : 1.3 @xmath61 10@xmath74 @xmath75 molec@xmath76 s@xmath76 near 300 k , with little temperature dependence , as reviewed by van look and peeters.@xcite the trends in transient ground state population illustrated in figure [ figure5 . ] and even more so in figure [ figure4 . ] for cf@xmath18c@xmath0h suggest that slowing down the relaxation of the hot radicals at lower buffer pressure inhibits the reactive loss .
such behavior can be expected for a barrierless association followed by a submerged barrier to fragmentation.@xcite a comparison of the transient spectra of hot and cold band lines following c@xmath0h@xmath0 photodissociation similar to our figure [ figure6 .
] was reported by yan , et al.@xcite this work has been widely cited @xcite as evidence that in the presence of 20 torr he , the vibrational thermalization of c@xmath0h will be complete in less than a microsecond . without fully reproducing the conditions of yan al . @xcite we can not be certain , but it appears likely that in the 3.5% mixture of acetylene and he used for this measurement , the disappearance rate of vibrationally excited c@xmath0h is dominated by reaction with c@xmath0h@xmath0 rather than vibrational relaxation by he . at the lower acetylene pressures typically used in subsequent kinetic experiments , the reactive loss is suppressed and vibrational excitation can be expected to persist to longer times .
sf@xmath17 has often also been added to test gas mixtures to accelerate vibrational relaxation , although several workers have reported that it was found to have no effect on ground state kinetic spectroscopy of c@xmath0h and was therefore left out.@xcite until future experiments are able to provide a direct monitor of the vibrational relaxation , we can only echo a summary statement from laufer and fahr @xcite in their 2004 review of the reactions and kinetics of ethynyl radicals from acetylene photodissociation : the role of vibrational excitation in the subsequent chemistry is not clear . " as the computational machinery for _ ab initio _ kinetics becomes more robust and well tested , the ambition grows to improve the treatment of chemical activation problems , including multiwell branching in competition with relaxation.@xcite coupled electronic states add to the challenge .
particularly for fast radical reactions that may react to completion faster than thermalization , a collection of thermal rate constants may not be sufficient to characterize a mechanism .
an extended set of state - resolved c@xmath0h kinetic measurements could provide a challenging target for the development of next - generation master - equation - based kinetic models .
the newly identified strong near - infrared bands can facilitate these measurements .
three new near - infrared bands of c@xmath0h have been observed and analyzed in the gas phase by diode laser spectroscopy .
transitions centered at 7110 , 7088 , and 6696 @xmath2 were found to be in good agreement with the calculated energies of vibronic levels of mixed @xmath4 and @xmath5 electronic character.@xcite the transitions are among the strongest of all those in the infrared and near infrared spectrum of the radical and provide a convenient spectroscopic tool for monitoring concentrations of the ground state radical in chemical mixtures .
kinetic measurements show that the competition between reactive loss and vibrational relaxation of the initially hot radicals makes interpretation of the kinetic signals complicated , especially because of the presence of the low - lying excited electronic state which has different reactivity and whose wavefunction contaminates all vibrationally excited levels of the ground electronic state .
these attributes combine to make c@xmath0h a particularly attractive vehicle for studying the effects of the presence of multiple potential surfaces on collisional processes in small free radicals where accurate calculations are now possible .
see supplementary material for c@xmath6h spectral data in two regions between 6630 and 7135 @xmath2 obtained at 0 , 2.6 , and 3.6 @xmath105s time delay , using 1:1 mixtures of cf@xmath19c@xmath20h and ar at a total pressure of 1.0 torr .
supplemental figure s1 and its caption , showing transient absorption signals for a selection of rotational levels in a 1% mixture of cf@xmath18c@xmath0h in ar at a total pressure of 5 torr and a boltzmann temperature analysis .
work at brookhaven national laboratory was carried out under contract no .
de - sc0012704 with the u.s .
department of energy , office of science , and supported by its division of chemical sciences , geosciences and biosciences within the office of basic energy sciences .
we thank stephen klippenstein for productive discussions and exploratory calculations on the submerged barriers in the c@xmath0h + cf@xmath18c@xmath0h reaction .
73ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ]
+ 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty link:\doibase 10.1002/kin.10141 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1002/anie.201404537 [ * * , ( ) ] link:\doibase 10.1021/jp402481y [ * * , ( ) ] link:\doibase 10.1021/jp5124162 [ * * , ( ) ] link:\doibase 10.1016/j.carbon.2014.01.020 [ * * , ( ) ] link:\doibase 10.1021/jp808346h [ * * , ( ) ] link:\doibase 10.1088/0034 - 4885/78/3/036501 [ * * ( ) , 10.1088/0034 - 4885/78/3/036501 ] link:\doibase 10.1086/181646 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1073/pnas.1113827108 [ * * , ( ) ] link:\doibase 10.1006/jmsp.2001.8459 [ * * , ( ) ] in @noop _ _ , ( , ) pp . ,
in link:\doibase 10.1016/s0273 - 1177(01)00059-x [ _ _ ] , , vol . ,
( , ) pp . , link:\doibase 10.1021/cr030039x [ * * , ( ) ] link:\doibase 10.1021/j100306a044 [ * * , ( ) ] link:\doibase 10.1021/j100383a003 [ * * , ( ) ] link:\doibase 10.1021/j100168a029 [ * * , ( ) ] link:\doibase 10.1021/j100151a026 [ * * , ( ) ] link:\doibase 10.1021/j100135a009 [ * * , ( ) ] link:\doibase 10.1021/j100044a013 [ * * , ( ) ] link:\doibase 10.1021/jp960201i [ * * , ( ) ] link:\doibase 10.1021/jp9532677 [ * * , ( ) ] link:\doibase 10.1021/jp031162v [ * * , ( ) ] link:\doibase 10.1021/jp401354x [ * * , ( ) ] link:\doibase 10.1086/183706 [ * * , ( ) ] link:\doibase 10.1086/160647 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1016/0301 - 0104(75)87025-x [ * * , ( ) ] link:\doibase 10.1063/1.1724865 [ * * , ( ) ] link:\doibase 10.1063/1.443331 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1063/1.445112 [ * * , ( ) ] link:\doibase 10.1063/1.448927 [ * * , ( ) ] link:\doibase 10.1016/0022 - 2852(87)90294 - 3 [ * * , ( ) ] link:\doibase 10.1063/1.452163 [ * * , ( ) ] link:\doibase 10.1016/0022 - 2860(88)80269 - 2 [ * * , ( ) ] link:\doibase 10.1063/1.453571 [ * * , ( ) ] link:\doibase 10.1063/1.454877 [ * * , ( ) ] link:\doibase 10.1016/0022 - 2852(88)90106 - 3 [ * * , ( ) ] link:\doibase 10.1006/jmsp.1996.0058 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1039/c1cp21523j [ * * , ( ) ] link:\doibase 10.1063/1.3532088 [ * * , ( ) ] \doibase http://dx.doi.org/10.1016/j.jms.2015.01.007 [ * * , ( ) ] link:\doibase 10.1021/j100280a044 [ * * , ( ) ] link:\doibase 10.1139/v94 - 085 [ * * , ( ) ] link:\doibase 10.1021/jp1115965 [ * * , ( ) ] @noop _ _ , ed .
( , , ) @noop * * , ( ) link:\doibase 10.1063/1.470361 [ * * , ( ) ] link:\doibase 10.1063/1.475415 [ * * , ( ) ] @noop * * , ( ) @noop _ _ , ed .
( , ) link:\doibase 10.1098/rsta.1960.0003 [ * * , ( ) ] link:\doibase 10.1021/jp062693x [ * * , ( ) ] , \doibase http://dx.doi.org/10.1016/j.proci.2014.05.118 [ * * , ( ) ] link:\doibase 10.1126/science.1224106 [ * * , ( ) ] link:\doibase 10.1021/acs.jpca.5b00640 [ * * , ( ) ] link:\doibase 10.1021/acs.jpca.5b06025 [ * * , ( ) ] link:\doibase 10.1063/1.1436481 [ * * , ( ) ] link:\doibase 10.1063/1.468060 [ * * , ( ) ] link:\doibase { 10.1063/1.2206180 } [ * * , ( ) ] link:\doibase { 10.1063/1.3120607 } [ * * , ( ) ] link:\doibase 10.1021/jp407701z [ * * , ( ) ] link:\doibase 10.1021/acs.jpca.5b01127 [ * * , ( ) ]
] cdcdcdcdc n@xmath43 & r@xmath106 & obs@xmath88calc & r@xmath107 & obs@xmath88calc & p@xmath106 & obs@xmath88calc & p@xmath107 & obs@xmath88calc + 1 & 7093.035 * & -0.055 & 7093.035 * & -0.069 & & & & + 2 & 7095.640 & 0.016 & 7095.640 & -0.003 & & & & + 3 & 7098.034 & 0.003 & 7098.034 * & -0.022 & 7078.468 * & -0.056 & 7078.468 * & -0.066 + 4 & 7100.303 & -0.009 & 7100.303 * & -0.038 & 7075.245 & 0.013 & 7075.245 & -0.002 + 5 & 7102.500 * & 0.035 & 7102.500 & 0.001 & 7071.812 & -0.001 & 7071.812 * & -0.022 + 6 & 7104.475 & -0.014 & 7104.380 * & -0.148 & 7068.256 & -0.013 & 7068.256 * & -0.038 + 7 & 7106.485 * & 0.102 & 7106.419 & -0.009 & 7064.632 * & 0.035 & 7064.632 & 0.004 + 8 & 7108.153 & 0.006 & 7108.202 & 0.006 & 7060.786 & -0.012 & 7060.691 * & -0.143 + 9 & 7109.791 & 0.012 & 7109.843 & 0.010 & 7056.969 * & 0.098 & 7056.902 & -0.009 + 10 & 7111.303 * & 0.026 & 7111.354 * & 0.018 & 7052.817 & 0.003 & 7052.862 & 0.002 + 11 & 7112.634 & -0.007 & 7112.692 & -0.013 & 7048.638 & 0.011 & 7048.684 & 0.007 + 12 & 7113.888 * & 0.020 & 7113.950 & 0.013 & 7044.333 * & 0.024 & 7044.387 * & 0.023 + 13 & 7114.955 & -0.001 & 7115.023 & -0.007 & 7039.850 & -0.007 & 7039.907 & -0.010 + 14 & 7115.939 * & 0.034 & 7116.014 * & 0.030 & 7035.293 * & 0.022 & 7035.348 & 0.012 + 15 & 7116.899 * & 0.188 & 7116.966 * & 0.171 & 7030.554 & 0.005 & 7030.614 & -0.005 + 16 & 7118.899 * & 1.526 & 7118.987*@xmath108 & 1.525 & 7025.726 * & 0.037 & 7025.791 * & 0.027 + 17 & 7117.238 * & -0.650 & 7117.365 * & -0.617 & 7020.877 * & 0.187 & 7020.936 * & 0.167 + 18 & 7117.652 * & -0.602 & 7117.800 * & -0.553 & 7017.073 * & 1.524 & 7017.208 * & 1.575 + 19 & 7117.937 * & -0.531 & 7118.095 * & -0.477 & 7009.619 * & -0.645 & 7009.749 * & -0.605 + 20 & 7118.129 * & -0.400 & 7118.277 * & -0.360 & 7004.234 * & -0.599 & 7004.378 * & -0.550 + 21 & & & & & 6998.728 * & -0.526 & 6998.875 * & -0.479 + 22 & & & & & 6993.122 * & -0.402 & 6993.274 * & -0.355 + cdcdcdcdc n@xmath43 & r@xmath106 & obs@xmath88calc & r@xmath107 & obs@xmath88calc & p@xmath106 & obs@xmath88calc & p@xmath107 & obs@xmath88calc + 1 & 6701.285 & -0.008 & 6701.285 & -0.001 & & & & + 2 & 6703.991 & -0.002 & 6703.991 * & 0.009 & 6689.773 & -0.009 & 6689.773 & -0.001 + 3 & 6706.622 & 0.003 & 6706.612 & 0.007 & 6686.717 & -0.010 & 6686.717 & 0.001 + 4 & 6709.172 & 0.001 & 6709.151 & -0.002 & 6683.600 & -0.001 & 6683.589 & 0.003 + 5 & 6711.647 & 0.001 & 6711.626 & 0.001 & 6680.399 & -0.002 & 6680.383 & 0.000 + 6 & 6714.045 & 0.001 & 6714.023 & 0.004 & 6677.133 & 0.005 & 6677.108 & 0.002 + 7 & 6716.365 & 0.004 & 6716.334 & 0.001 & 6673.786 & 0.007 & 6673.757 & 0.003 + 8 & 6718.596 & 0.000 & 6718.564 & 0.000 & 6670.355 & 0.002 & 6670.329 & 0.005 + 9 & 6720.744 & -0.001 & 6720.705 & -0.004 & 6666.856 & 0.007 & 6666.818 & 0.002 + 10 & 6722.805 & -0.001 & 6722.757 & -0.009 & 6663.265 & 0.002 & 6663.227 & 0.000 + 11 & 6724.724 * & -0.050 & 6724.700 * & -0.032 & 6659.594 & 0.001 & 6659.547 & -0.006 + 12 & 6726.640 & -0.008 & 6726.603 & 0.002 & 6655.835 & -0.002 & 6655.788 & -0.005 + 13 & 6728.426 & 0.004 & 6728.378 & 0.006 & 6651.934 * & -0.057 & 6651.918 * & -0.026 + 14 & 6730.117 * & 0.024 & 6730.064 * & 0.025 & 6648.043 & -0.008 & 6647.997 & -0.004 + 15 & 6731.704 * & 0.049 & 6731.652 * & 0.054 & 6644.019 & 0.004 & 6643.966 & 0.005 + 16 & 6733.200 * & 0.095 & 6733.147 * & 0.103 & 6639.898 * & 0.021 & 6639.845 * & 0.026 + 17 & & & & & 6635.679 * & 0.045 & 6635.626 * & 0.054 + 18 & & & & & 6631.375 * & 0.094 & 6631.315 * & 0.099 + c d c d c d c d c d c d c n@xmath43 & r@xmath106 & obs@xmath88calc & r@xmath107 & obs@xmath88calc & q@xmath106 & obs@xmath88calc & q@xmath107 & obs@xmath88calc & p@xmath106 & obs@xmath88calc & p@xmath107 & obs@xmath88calc + 2 & & & & & 7106.633 & -0.003 & 7109.394 & 0.011 & & & & + 3 & & & 7118.985 * @xmath108 & -0.022 & 7106.287 * & -0.062 & 7108.538 & -0.015 & & & & + 4 & 7119.585 & -0.009 & 7120.878 & -0.005 & 7105.685 * & -0.125 & 7107.783 * & 0.087 & 7094.684 & 0.013 & & + 5 & 7121.538 * & -0.018 & 7122.627 & -0.003 & 7105.060 & 0.009 & 7106.970 * & 0.22 & 7091.194 * & -0.023 & 7092.767 * & -0.017 + 6 & 7123.318 & -0.001 & 7124.239 & 0.010 & 7104.025 * & -0.063 & 7105.673 & -0.009 & 7087.551 & 0.000 & 7088.835 & 0.000 + 7 & 7124.870 & -0.013 & 7125.672 & 0.001 & 7102.983 * & 0.040 & 7104.474 & -0.002 & 7083.682 & -0.007 & 7084.759 & 0.001 + 8 & 7126.259 & -0.001 & 7126.952 & 0.004 & 7101.729 * & 0.103 & 7103.134 & 0.013 & 7079.623 & -0.006 & 7080.539 & 0.005 + 9 & 7127.446 & -0.004 & 7128.061 & 0.005 & & & & & & & 7076.159 & 0.005 + 10 & 7128.450 & -0.006 & 7128.981 & -0.010 & & & 7099.934 & -0.006 & 7070.934 & 0.007 & 7071.617 & 0.006 + 13 & 7129.277 & -0.001 & 7130.755 & 0.006 & 7096.621 & 0.007 & 7098.101 & -0.006 & 7066.297 & -0.002 & 7066.903 & 0.003 + 11 & 7129.906 & -0.008 & 7129.748 & -0.004 & 7094.623 & 0.030 & & & 7061.482 & -0.006 & 7062.016 & -0.003 + 12 & 7130.378 & 0.012 & 7130.318 & -0.017 & & & 7093.971 & 0.028 & 7056.497 & 0.003 & & + 14 & 7130.642 & 0.011 & 7130.961 & 0.001 & 7090.024 & -0.002 & & & 7051.313 & -0.005 & 7051.720 & -0.014 + 15 & 7130.700 & -0.009 & 7131.016 & 0.016 & 7087.466 & -0.013 & 7089.087 & -0.015 & 7045.971 & 0.013 & 7046.324 & -0.014 + 16 & & & 7130.906 * & 0.052 & & & & & & & & + ccccc & x(0,0,0 ) & x(0,8@xmath109,2 ) & x(0,2@xmath109,3)a(0,3,0)@xmath110 & x(0,3@xmath38,3)a(0,0,2)@xmath38 + & @xmath111 & @xmath112 & @xmath112 & @xmath86 + a & & & & -4.29(2 ) + b & 1.4568256(2 ) & 1.42115(5 ) & 1.3941(2 ) & 1.3719(1 ) + dx10@xmath113 & 0.357(3 ) & 1.582(3 ) & 0.89(7 ) & 0.68(4 ) + @xmath45 & 0.0020883(1 ) & 0.0015(2 ) & -0.0071(5 ) & -0.0336(6 ) + p+2q & & & & -0.0925(8 ) + q & & & & -0.00447(3 ) + origin & & 6695.678(2 ) & 7087.645(7 ) & 7109.646(6 ) + & + origin & & 6696.7 & 7099.0 & 7106.7 + intensity & & 327.5 & 1384.1 & 744.2 + + + | transient diode laser absorption spectroscopy has been used to measure three strong vibronic bands in the near infrared spectrum of the c@xmath0h , ethynyl , radical not previously observed in the gas phase .
the radical was produced by ultraviolet excimer laser photolysis of either acetylene or ( 1,1,1)-trifluoropropyne in a slowly flowing sample of the precursor diluted in inert gas , and the spectral resolution was doppler - limited .
the character of the upper states was determined from the rotational and fine structure in the observed spectra and assigned by measurement of ground state rotational combination differences .
the upper states include a @xmath1 state at 6696 @xmath2 , a second @xmath1 state at 7088 @xmath2 , and a @xmath3 state at 7110 @xmath2 . by comparison with published calculations ( r. tarroni and s. carter , _ j. chem
. phys _ * 119 * , 12878 ( 2003 ) and _ mol . phys_. * 102 * , 2167 ( 2004 ) ) , the vibronic character of these levels was also assigned .
the observed states contain both @xmath4 and @xmath5 electronic character .
several local rotational level perturbations were observed in the excited states .
kinetic measurements of the time - evolution of the ground state populations following collisional relaxation and reactive loss of the radicals formed in a hot , non - thermal , population distribution were made using some of the strong rotational lines observed .
the case of c@xmath6h may be a good place to investigate the behavior at intermediate pressures of inert colliders , where the competition between relaxation and reaction can be tuned and observed to compare with master equation models , rather than deliberately suppressed to measure thermal rate constants . |
the skyrme model @xcite , despite is relative successes @xcite , can only be considered as a prototype of an effective theory of qcd . indeed
, large @xmath0 analysis @xcite suggests that bosonization of qcd would most likely involve an infinite number of mesons .
if this is the case , then taking the appropriate decoupling limits ( or large mass limit ) for higher spin mesons leads to an all - orders lagrangian in derivatives of pion fields . for
now however , fixing the form of the effective lagrangian from an exact low - energy limit of qcd seems to be out of our reach and the alternative has been to propose simple effective lagrangians @xcite and rely on a few nucleons properties to set the parameters they depend on . from that point of view ,
a rather stringent test for such models lies in whether they could accurately describe the data for the electromagnetic form factors of nucleons for moderate values of momentum transfer .
the electromagnetic form factors of the semiclassically quantized su(2 ) skyrmion were studied systematically by braaten et al @xcite but this first attempt did not take into account relativistic kinematical corrections which are important for momentum transfers @xmath1 gev@xmath2 these correction were implemented to the original skyrme model predictions by holzwarth @xcite using the prescription of ji @xcite .
actually , holzwarth also introduced a second correction to the skyrme model to account for vector meson effects either by introducing a multiplicative factor to reflect the contributions of the poles coming from these mesons , or by adding the vector mesons as dynamical degrees of freedom in the lagrangian . yet , in spite of remarks and suggestions in @xcite , no analysis regarding how higher - order lagrangians could reproduce the nucleon electromagnetic form factors or conversely , how the form factors could help construct a more accurate effective lagragian . in this work
, we calculate the electromagnetic form factors for a class of higher - order ( skyrme - like ) models introduced in @xcite . for simplicity
, we limit our numerical analysis to one and two - parameter models with a proper treatment of relavistic effects , and evaluate to what extent the experimental data of the electromagnetic form factors of proton can suggest a form of higher - order lagrangians or discriminate among possible candidates .
note that neutron data are ignored here since they are plagued with large uncertainties .
let us first introduce the lagrangian density for the skyrme model @xmath3 ^{2}\label{eq : skyrme_lagr}%\ ] ] where @xmath4 is the pion decay constant , @xmath5 is the left - handed chiral courant @xmath6 and the skyrme constant @xmath7 is a dimensionless constant .
@xmath8 is a @xmath9 field related to the pion field @xmath10 by @xmath11 .
the field configurations with finite energy must satisfy the boundary condition @xmath12 these configurations fall into topological sectors characterized by @xmath13}\right ) , \label{eq : charge}%\ ] ] a topological invariant taking integral values .
skyrme interpreted this topological invariant as the baryon number .
accordingly , the lowest - energy @xmath14 sector is identified with the nucleon . in this sector , the lowest - energy field configuration is given by the hedgehog ansatz @xmath15 \label{eq : hedgehog}%\ ] ] where @xmath16 satisfies the boundary conditions @xmath17 and @xmath18 . with the convenient change of scale , we can use @xmath19 and @xmath20 as units of length and energy respectively and rewrite the lagrangian density ( [ eq : skyrme_lagr ] ) as @xmath21 where @xmath22 $ ] . a pion mass term @xcite
is usually added to account for the chiral symmetry breaking observed in nature @xmath23 this term serves as a regulator for the magnetic radii and form factors ( see ( [ eq : ray_m ] ) ) of nucleon otherwise they would diverge @xcite . using the hedgehog ansatz ( [ eq : hedgehog ] ) , one
obtain the mass of the static skyrmion @xmath24
+ 2\beta^{2}(1-\cos f)}\right\ } } .
\end{split } \label{eq : masse_u}%\ ] ] where @xmath25 has now been rescaled and @xmath26 the stable static soliton is obtained by minimizing the mass and requires to solve the chiral equation @xmath27 + a\left [ { f^{\prime2}\frac{{c}}{{s}}% + \frac{{sc}}{{r^{2}}}-2\frac{{f^{\prime}}}{r}}\right ] -\beta^{2}% s=0.\label{eq : chirale}%\ ] ] with the boundary conditions @xmath17 and @xmath18 for @xmath28 for simplicity here , we used @xmath29 , @xmath30 , @xmath31 .
fluctuations around this static soliton soliton should be quantized .
quantization of the skyrmion is usually performed with the introduction of spin and isospin rotation matrix as a collective coordinate @xcite .
the spin / isospin rotations of skyrmion takes the form @xmath32 with @xmath33 an arbitrary time - dependent su(2 ) matrix . substituting ( [ eq : quanti ] ) in ( [ eq : skyrme_u ] )
, one gets @xmath34=-m+\frac { i(i+1)}{2\mathcal{i}}%\ ] ] where @xmath35 is defined in ( [ eq : masse_u ] ) and @xmath36 } % \ ] ] is the moment of inertia of the skyrmion . here @xmath37 is the spin or isospin of the nucleon .
the parameters @xmath38 and @xmath7 are fixed using two experimental input ( mass of the nucleons or else ) .
one is then able to reproduce the static properties of baryons within a 30 % accuracy @xcite .
these methods can also be generalized to extensions of the skyrme model as we will show in the following section .
the skyrme model is more a prototype for an low - energy pion interactions than a full effective field theory .
higher - order terms are expected to appear in addition to ( [ eq : skyrme_u ] ) and ( [ eq : masse_u ] ) but in its most general form , the lagrangian would involve a increasing number of terms at each order in pion field derivatives making the treatment practically intractable .
one of us has proposed a special class of models whose energy density , assuming the hedgehog ansatz , is at most linear in @xmath39 @xcite .
this requirement is sufficient to determine a unique term to each order in derivatives and turns out to have deeper geometrical meaning @xcite .
the static energy density coming from the lagrangian of order @xmath40 in derivatives of the field takes the form @xmath41 } \label{eq : energ_marleau}%\ ] ] where @xmath29 and @xmath42 . using the hedgehog ansatz ,
the first two terms arise from the non - linear @xmath43 and the skyrme terms @xmath44\]]@xmath45\ ] ] while the third term leads to @xmath46 as for the term proposed by jackson et al @xcite to allow for the dynamics of the @xmath47 meson in the skyrme model .
generalizing to all order , the static energy associated to this class of all - order lagrangian can be written in a simple form @xmath48 where a specific model is characterized by a choice of the parameters @xmath49 or equivalently of the function @xmath50 and @xmath51 .
yet , @xmath52 is not completely arbitrary . requiring that a unique soliton solution exists sets some constraints on @xmath52 @xcite : @xmath53 for @xmath54 .
the mass of the soliton , including the pion mass term , is @xmath55 and leads to the generalized chiral equation @xmath56 + a\chi^{\prime\prime } \left ( a\right ) \left [ { f^{\prime2}\frac{{c}}{{s}}+\frac{{sc}}{{r^{2}}% } -2\frac{{f^{\prime}}}{r}}\right ] -\beta^{2}s=0.\ ] ] in that context , the skyrme lagrangian corresponds to @xmath57 .
the moment of inertia of the soliton also take a simple form@xmath58 } .\label{eq : inertie_gene}%\ ] ] most all - orders model depend on more than two parameters . in the next section
, we analyze the behavior of the electromagnetic form factors for a few models in the hope that these could help fix the @xmath49 coefficients and obtain a better agreement with the the experimental data in general .
the electromagnetic form factors of the proton in the breit frame , for spacelike momentum transfer @xmath59 , are the fourier transforms of its electric charge and magnetic moment densities : @xmath60@xmath61 with @xmath62 and @xmath63 \label{eq : den_inertie}%\ ] ] are the baryon density and moment of inertia density respectively . here
@xmath64 is the spherical bessel function of order @xmath65 and @xmath66 is the nucleons mass
. both densities ( [ eq : den_bar ] ) and ( [ eq : den_inertie ] ) are normalized @xmath67 while the electromagnetic form factors satisfy the normalization condition @xmath68 where @xmath69 is the magnetic moment of the proton with baryonic square radius @xmath70 however , the definitions ( [ eq : fac_e ] ) and ( [ eq : fac_m ] ) only hold in the breit frame moving at velocity @xmath71 with respect to the nucleon rest frame where the chiral profile @xmath72 is computed . a correction for this lorentz boost must be applied .
ji @xcite has proposed a simple prescription to circumvent this difficulty : @xmath73@xmath74 where @xmath75 and @xmath76 are given respectively by ( [ eq : fac_e ] ) and ( [ eq : fac_m ] ) and @xmath77 is the lorentz factor @xmath78 with the nucleon mass @xmath35 .
unfortunately , the boost transformations ( [ eq : ge_rel ] ) and ( [ eq : gm_rel ] ) violate the so - called superconvergence rule @xmath79 which is expected to hold for electromagnetic form factors . indeed
the limit @xmath80 in the breit frame corresponds to @xmath81 in the rest frame and generally @xmath82 does not vanish .
a possible approach to restore superconvergence is to relax the condition that @xmath35 must take the value of the nucleon mass and instead allow @xmath35 to vary in ( [ eq : correction ] ) in order to get the best agreement with the data of @xmath83 at the highest available values of @xmath84 @xcite .
however for the models under study here , this procedure turned out to be unsatisfactory . the value of @xmath35 ensuring superconvergence caused the ratio of the electromagnetic form factors to become too suppressed in the large-@xmath85 limit . on the other hand ,
the models which we propose as candidates for the description of qcd at low energy are not expected to hold for large values of @xmath86 since the superconvergence rule seems to be too restrictive , it will not be applied here .
we will instead promote @xmath35 as parameter and adjust its value to provide a better fit of high-@xmath85 data .
the parameters of the skyrme model , @xmath7 and @xmath87 are usually set with either one of the following two methods : ( i ) adjust @xmath7 and @xmath4 to obtain the mass of the nucleons ( 939 mev ) and of the @xmath88 resonance ( 1232 mev ) or ( ii ) set @xmath4 according to its experimental value ( 186 mev ) and adjust @xmath7 to reproduce the nucleon-@xmath88 mass split ( 295 mev ) . for comparison purposes ,
we adopt the second method also used in @xcite and assume that the pion mass takes its physical value .
we consider here two simple extensions of the skyrme model that fall into the class of models described in section [ sec : all ] @xmath89 model a adds a single term ( and a single parameter @xmath90 ) of order six in derivative of the field to the skyrme model whereas model b which depends on two additionnal parameter has a rational form which hopefully could allow to reproduce poles due to vector mesons .
this latter model is a generalization of a model introduced by jackson et al @xcite with @xmath91 and @xmath92 .
fixing the model parameters @xmath93 and @xmath94 as well as the scale parameter @xmath35 is a tedious procedure which requires a few steps : first , we choose a set of parameters ( @xmath90 for model a or @xmath95 and @xmath94 for model b ) and solve the differential equation ( [ eq : gen_eom ] ) for massive pions and use the method described above ( inputs are @xmath4 and nucleon-@xmath88 mass split ) to fit for the appropriate value of @xmath96 in the second step , we compute the form factors , compare them with data and adjust @xmath35 to minimize the @xmath97 of the ratio of the electromagnetic form factors .
the whole procedure is repeated with different sets of parameters until we get the configuration @xmath98 or @xmath99 with lowest @xmath100 the results for the electromagnetic form factor are presented in fig .
[ fig : no_lambda].we get @xmath101 @xmath102 and @xmath103 gev for model a and @xmath104 @xmath105 @xmath106 and @xmath103 gev for model b. the results for the skyrme model are also shown for comparison ( @xmath107 and @xmath108 gev ) . according to fig .
[ fig : no_lambda ] , both models present improvements over the skyrme model , especially for the ratio @xmath109 of the form factors . despite its simplicity ,
model a seems to overcome model b. however both models exhibits a rapid divergence of the magnetic form factor of proton . , @xmath110 ( full lines ) ,
model b : @xmath111 , @xmath112 , @xmath113 ( dashed lines ) and skyrme model : @xmath107 ( dotted lines ) .
the data are from @xcite ( full diamonds ) and @xcite ( full squares).,scaledwidth=80.0% ] in view of these results , we must conclude that although the models examined here provide clear improvements , they can not reproduce the data for the magnetic form factor adequately .
the problem originate to their inability to mimic the pole coming from the vector mesons . computing these predictions for an arbitrary number of skyrme model extensions being prohibitive
, we resort to another approach @xcite .
the @xmath114 meson effects are incorporated by multiplying the form factors ( [ eq : fac_e ] ) and ( [ eq : fac_m ] ) by @xmath115 with @xmath116 = 770 mev . the electromagnetic mean square radii of proton then become @xmath117@xmath118 this correction introduced an additional parameter whose allowed values go from the purely pionic model ( @xmath119 ) to the mesonic dominance ( @xmath120 ) .
it should be noted that the parameter @xmath121 does not have any effect on the ratio of the electromagnetic form factors . a similar effect
could also be included to take into account the @xmath122 meson effects , but corresponding factor would involve a second @xmath121 parameter and our calculations indicates that this additionnal factor does not leads to noticeable amelioration of the behavior .
figure [ fig : modab ] presents the electromagnetic form factors as obtained from model a and b with pole effects introduced through ( [ eq : lamb ] ) .
all the models parameters as well as the parameters @xmath35 and now @xmath121 are fitted using the procedure described above to minimize the @xmath97 of the magnetic form factor .
the best accord with data are acheived for values of the parameters @xmath123 @xmath124 with @xmath125 gev and @xmath126 for model a and @xmath104 @xmath127 and @xmath106 with @xmath128 gev and @xmath129 for model b. the results for the skyrme model are found to be constistent with those of refs .
@xcite with @xmath130 gev and @xmath131 .
clearly , the inclusion of vector mesons effects produce a significant improvement @xmath132 over the results of fig .
[ fig : no_lambda ] . again both extensions of the skyrme model , model a and b ,
are better at reproducing the data with model a being slightly superior despite its relative simplicity .
furthermore , as shown in table [ tab : results ] , the addition of higher orders terms have little effect on other static properties of the proton with these choice of parameters ( note that the magnetic moment of proton predicted by the models are all lower than its experimental value but one has to remember that these properties are subject to quantum corrections @xcite ) .
the values of @xmath121 indicate a strong dominance of the @xmath114 meson as in @xcite .
our results therefore indicates that the data are best reproduced by a higher - order effective lagrangian in the low - energy limit of qcd .
unfortunately , as we have noticed , the determination of a more acurate effective lagrangian faces theoretical uncertainties which remain to be addressed .
first , the boost prescription ( [ eq : ge_rel ] ) and ( [ eq : gm_rel ] ) , not being compatible with superconvergence , prevents us from reproducing results for @xmath133 .
an adequate boost prescription holding account of superconvergence should allow to extend the analysis to higher momentum tranfer .
secondly , we have only explored the possibilities of two types of skyrme model extensions .
these could not mimick the vector meson effects satisfactorily without the artifact of equation ( [ eq : lamb ] ) . finding a form of skyrme - like lagrangian that would allow to bypass the approach in ( [ eq : lamb ] ) and avoid
the introduction of the @xmath121 parameter remains a challenge .
[ c]lccccccc & skyrme & & model a & & model b & & exp .
+ @xmath35 ( gev@xmath134 ) & @xmath135 & & @xmath136 & & @xmath137 & & @xmath138 + @xmath121 & @xmath139 & & @xmath140 & & @xmath141 & & @xmath138 + @xmath142 ( @xmath143 ) & @xmath144 & & @xmath145 & & @xmath146 & & @xmath147 + @xmath148 ( fm ) & @xmath149 & & @xmath150 & & @xmath151 & & @xmath152 + @xmath153 ( fm ) & @xmath154 & & @xmath155 & & @xmath156 & & @xmath157 + | we compare the prediction of skyrme - like effective lagrangians with data for electromagnetic form factors of proton and consider the possibility of fixing the parameters of these higher - order lagrangians .
our results indicate that one or two - parameter models can lead to better agreement with the data but more accurate determination of the effective lagragian faces theoretical uncertainties . |
the notion of duality of projective varieties , which appears in various branches of mathematics , has been a subject of study since the beginnings of algebraic geometry @xcite . given an embedded projective variety @xmath6 , its dual variety @xmath7 is the closure in the dual projective space @xmath8 of the hyperplanes intersecting the regular points of @xmath9 non transversally . a projective variety @xmath9 is _ self - dual _ if it is isomorphic to its dual @xmath7 as embedded projective varieties .
the expected codimension of the dual variety is one .
if this is not the case , @xmath9 is said to be defective .
self - dual varieties other than hypersurfaces are defective varieties with `` maximal '' defect .
let @xmath10 be an algebraically closed field of characteristic @xmath2 .
let @xmath0 be an algebraic torus over @xmath10 and @xmath5 a finite dimensional @xmath0-module . in this paper
we characterize self - dual projective toric varieties @xmath11 equivariantly embedded , in terms of the combinatorics of the associated configuration of weights @xmath12 ( cf .
theorems [ teo : bside ] and [ teo : aside ] ) and in terms of the interaction of the space of relations of these weights with the torus orbits ( cf
. theorems [ teoeqauto ] and [ teo : constaut+gde ] ) .
in particular , we show that @xmath9 is self - dual if and only if @xmath13 and the smallest linear subspaces containing @xmath14 and @xmath7 have the same dimension , see theorems [ thm : autdim ] and [ thm : autdimgral ] . given a basis of eigenvectors of @xmath5 and the configuration of weights of the torus action on @xmath5 , it is not difficult to check the equality of the dimensions of @xmath9 and its dual ( for instance , by means of the combinatorial characterization of the tropicalization given in @xcite ) .
but the complete classification of defective projective toric varieties in an equivariant embedding is open in full generality and involves a complicated combinatorial problem . for smooth toric varieties
this characterization is obtained in @xcite ; the case of @xmath15-factorial toric varieties is studied in @xcite . for non necessarily normal projective toric varieties of codimension two , a characterization is given in @xcite .
this has been extended for codimensions three and four in @xcite . for smooth projective varieties ,
a full list of self - dual varieties is known @xcite .
this list is indeed short and reduces in the case of toric varieties to hypersurfaces or segre embeddings of @xmath16 , for any @xmath17 , under the assumption that @xmath18 .
this was expected to be the whole classification under the validity of hartshorne s conjecture @xcite .
we prove that this is indeed the whole list of self - dual smooth projective toric varieties in [ th : regcase ] .
there exist some classical examples of self - dual non smooth varieties , as the quartic kummer surface .
tevelev and popov gave new families of non smooth self - dual varieties that come from actions of isotropy groups of complex symmetric spaces on the projectivized nilpotent varieties of isotropy modules ( @xcite , @xcite ) .
as a consequence of theorem [ teo : bside ] , it is easy to construct new families of self - dual projective toric varieties in terms of the gale dual configuration ( see definition [ def : gale ] ) .
a big class of self - dual toric varieties are the toric varieties associated to lawrence configurations ( see definition [ df : law ] ) , which contain the configurations associated to the segre embeddings .
lawrence constructions are well known in the domain of geometric combinatorics , where they are one of the prominent tools to visualize the geometry of higher dimensional polytopes ( see ( * ? ? ?
* chapter 6 ) ) ; the commutative algebraic properties of the associated toric ideals are studied in @xcite .
we show in section [ sec : examples ] other non lawrence concrete examples for any dimension bigger than @xmath19 and any codimension bigger than @xmath20 .
we also introduce the notion of strongly self - dual varieties ( see definition [ def : strong ] ) , which in the toric case is not only related to the geometry of the configuration of weights but also to number theoretic aspects .
this concept is useful for the study of the existence of rational multivariate hypergeometric functions @xcite . in section [ sec : prelim ]
we gather some preliminary results about embedded projective toric varieties and duality of projective varieties . in section [ sec : orbits ] we characterize self - dual projective toric varieties in terms of the geometry of the action of the torus and we give precise assumptions under which self - dual projective varieties are precisely those with maximal defect . in section [ sec : comb ] we give two ( equivalent ) combinatorial characterizations of self - duality . in section [ sec : examples ] we collect several new examples of self - dual ( non smooth ) projective toric varieties . finally , in section [ sec : strong ] we study strongly self - dual toric varieties .
_ acknowledgments : _ we are grateful to eduardo cattani for his suggestion of the statement of theorem [ teo : bside ] and to vic reiner for his help in the statement of theorem [ teo : aside ] .
in this section we collect some well known results and useful observations on projective toric varieties and duality of projective varieties .
let @xmath0 be an algebraic torus over an algebraically closed field @xmath1 of characteristic @xmath2 .
we denote by @xmath21 the lattice of characters of @xmath0 ; recall that @xmath22=\bigoplus_{\lambda\in{{\mathcal x}}(t ) } \bbbk \lambda$ ] .
any finite dimensional rational @xmath0-module @xmath5 , @xmath23 , decomposes as a direct sum of irreducible representations @xmath24 with @xmath25 , @xmath26 , for all @xmath27 .
the action of @xmath0 on @xmath5 canonically induces an action @xmath28 on the associated projective space , given by @xmath29=[t\cdot v]$ ] , where @xmath30\in { { \mathbb p}}(v)$ ] denotes the class of @xmath31 . recall that an irreducible @xmath0-variety @xmath9 is called _ toric _ if there exists @xmath32 such that the orbit @xmath33 is open in @xmath9 .
let @xmath34 ( which may contain repeated elements ) be the associated set of weights of a finite dimensional @xmath0-module @xmath5 we call @xmath12 the _ configuration of weights _ associated to the @xmath0-module @xmath5 . to any basis @xmath35 of eigenvectors
we can associate a projective toric variety by @xmath36\bigr)}\subset { { \mathbb p}}(v).\ ] ] denote by @xmath37 .
the dense orbit @xmath38\bigr)$ ] in @xmath39 coincides with the intersection @xmath40 .
observe that since @xmath41 is equal to @xmath42\bigr)$ ] , it follows that @xmath41 is maximal among the dimensions of the toric subvarieties of @xmath43 i.e. those of the form @xmath44)}$ ] for some @xmath30\in { { \mathbb p}}(v)$ ] .
based on the decomposition ( [ eq : decomp ] ) , in ( * ? ? ?
* proposition ii.5.1.5 ) it is proved that any projective toric variety in an equivariant embedding is of type @xmath39 for some @xmath0-module @xmath5 and a basis on eigenvectors @xmath45 of @xmath5 , in the following sense .
let @xmath46 be a @xmath0-module and @xmath47 a toric subvariety ; then there exists @xmath48 ( with possible repetitions ) and a @xmath0-equivariant linear injection @xmath49 , @xmath50 , such that the induced equivariant morphism @xmath51 gives an isomorphism @xmath52 .
moreover , let @xmath53 be another @xmath0-module , @xmath54 a basis of eigenvectors of @xmath53 such that @xmath55 , and consider @xmath56 , given by @xmath57 .
clearly , @xmath58 is an isomorphism of @xmath0-modules , and its induced morphism @xmath59 is an isomorphism such that @xmath60 . in view of the preceding remark ,
the following notation makes sense .
[ def : vartor2 ] the projective toric variety @xmath61 associated to the configuration of weights @xmath12 is defined as @xmath62\bigr)}\subset { { \mathbb p}}(v),\ ] ] where @xmath5 is a @xmath0-module with @xmath12 as associated configuration of weights .
we can make a series of reductions on @xmath12 and @xmath0 , as in @xcite .
first , the following easy lemma allows to reduce our problem to the case of a faithful representation .
[ lem : faith ] given a @xmath0-module @xmath5 of finite dimension and @xmath12 the associated configuration of weights , consider the torus @xmath63 , where @xmath64 denotes the @xmath65-submodule generated by @xmath12 .
the representation of @xmath0 in @xmath66 induces a faithful representation @xmath67 which has the same set theoretical orbits in @xmath5 .
we can then replace @xmath0 by the torus @xmath68 .
it is easy to show that this is equivalent to the fact that @xmath69 , which we will assume from now on without loss of generality .
next , we enlarge the torus without affecting the action on @xmath43 ; this will allow us to easily translate affine relations to linear relations on the configuration of weights .
if we let the algebraic torus @xmath70 act on @xmath5 by @xmath71 , then the actions @xmath72 and @xmath73 have the same set theoretical orbits . more in general ,
let @xmath74 and @xmath75 .
consider the @xmath0-action on @xmath5 given by @xmath76 .
the actions @xmath77 and @xmath78 coincide on @xmath43 , and the corresponding variety @xmath79 coincides with @xmath61 .
hence , we can assume that there is a splitting of @xmath80 in such a way that @xmath81 for all @xmath82 , @xmath83 and @xmath84 .
in fact , the previous reductions are comprised in the following more general setting : [ step : affine ] consider @xmath85 two tori and two finite configurations of @xmath86 weights @xmath87 , @xmath88 .
assume that there exists a @xmath15-affine transformation @xmath89 such that @xmath90 for all @xmath91
. then @xmath92 .
[ rem : variado ] ( 1 ) the dimension of the projective toric variety @xmath61 equals the dimension of the affine span of @xmath12 .
\(2 ) note that if @xmath93 is contained in an hyperplane off the origin , then @xmath94 precisely when @xmath95 and the elements in @xmath12 are a basis of @xmath21 .
\(3 ) if we denote by @xmath96 the dimension of the affine span of @xmath12 , then @xmath61 is an hypersurface if and only if @xmath97 .
in this situation , either @xmath12 coincides with the set of vertices of its convex hull @xmath98 , or @xmath99 contains only one element @xmath100 in its relative interior , and @xmath101 is the set of vertices .
we finish this paragraph by recalling some basic facts about the geometric structure of a toric variety @xmath61 .
[ lem : affineopen ] let @xmath102 be a configuration , and assume that the set of vertices of @xmath99 is equal to @xmath103 . set @xmath104\bigr)$ ] , @xmath105 .
then @xmath106 is an affine toric @xmath0-variety , and there exist @xmath0-equivariant open immersions @xmath107 , in such a way that @xmath108\bigr).\ ] ] in particular , @xmath61 is a normal variety if and only if @xmath109 for all @xmath110 .
moreover , @xmath61 is a smooth variety if for all @xmath111 , there are exactly @xmath112 edges of @xmath99 from @xmath113 , and the subset @xmath114 is basis of @xmath21 , where @xmath115 is the `` first '' point on an edge from @xmath113 .
see for example ( * ? ? ?
* appendix to chapter 3 ) .
let @xmath116 be a lattice of rank @xmath117 .
we let @xmath118 and consider the @xmath1-vector space @xmath119 .
recall that given a basis @xmath120 of @xmath121 , we can identify @xmath121 with @xmath122 and @xmath123 with @xmath124 .
[ def : regconf ] a _ lattice configuration _ @xmath125 is a finite sequence of lattice points .
we say that a configuration @xmath12 is _ regular _ if it is contained in a hyperplane off the origin .
[ exa : regularconf ] let @xmath0 be an algebraic torus , and let @xmath126 be a configuration of weights
. then the following are equivalent : \(1 ) the configuration @xmath12 is regular ; \(2 ) up to affine isomorphism , @xmath12 has the form @xmath127 for all @xmath128 ; \(3 ) there exists a splitting @xmath80 , such that under the identification @xmath129 , the weights of @xmath12 are of the form @xmath130 , @xmath131 .
see also the reductions made before lemma [ step : affine ] .
we denote by @xmath132 the lattice of affine relations among the elements of @xmath12 , that is @xmath133 belongs to @xmath134 if and only if @xmath135 and @xmath136 .
if @xmath12 is regular , then @xmath134 coincides with the lattice of linear relations among the elements of @xmath12 .
note that these ( affine or linear ) relations among the elements of @xmath12 can be identified with the affine relations among the elements of the configuration @xmath137 .
thus , given any configuration @xmath138 , we can embed it in @xmath139 via @xmath140 so that affine dependencies are translated to linear dependencies .
in fact , the map @xmath141 is an injective affine linear map .
more in general , we have the following definition .
we say that two configurations @xmath142 , are _ affinely equivalent _ if there exists an affine linear map @xmath143 ( defined over @xmath15 ) such that @xmath144 sends @xmath145 bijectively to @xmath146 ( that is , @xmath144 defines an injective map from the affine span of @xmath145 to the affine span of @xmath146 ) .
so , if @xmath145 and @xmath146 are affinely equivalent , they have the same cardinal and moreover , @xmath147 .
any property of a configuration @xmath12 shared by all its affinely equivalent configurations is called an _ affine invariant _ of @xmath12 . in this terminology ,
lemma [ step : affine ] asserts that the projective toric variety @xmath61 is an affine invariant of the configuration @xmath12 .
[ def : pyramid ] we say that @xmath148 is a _ pyramid _ ( or a _ pyramidal configuration _ ) if there exists an affine hyperplane @xmath149 such that @xmath150 , i.e. if all points in @xmath12 but one lie in @xmath149 , or equivalently , if there exist and index @xmath151 and an affine linear function @xmath152 such that @xmath153 for all @xmath154 and @xmath155 .
more precisely , we say that @xmath12 is a _
@xmath156-pyramidal configuration _ if , after reordering , there exists a splitting of the lattice as a direct sum of lattices @xmath157 , with @xmath158 a basis of @xmath159 and @xmath160 , with @xmath146 not a pyramidal configuration of @xmath161 . in particular , the @xmath2-pyramidal configurations are the _ non pyramidal configurations_. [ obsnucleo ] being a pyramid is clearly an affine invariant of a configuration .
it is straightforward to check that @xmath12 is a non pyramidal configuration if and only if there exists a relation @xmath162 with @xmath163 , i.e. if @xmath164 is not contained in a coordinate hyperplane .
[ def : join ] let @xmath165 two @xmath1-vector spaces of respective dimensions @xmath166 and @xmath167 two projective varieties .
recall that the _ join _ of @xmath9 and @xmath168 is the projective variety @xmath169:\,[x ] \in x , [ y ] \in y \bigr\ } } \subset { { \mathbb p}}(v_1 \times v_2),\ ] ] that is , the cone over the join @xmath170 is the product of the cones over @xmath9 and @xmath168 .
we set @xmath171 \in { { \mathbb p}}(v_1\times v_2),\;[y ] \in y\bigr\ } \subset { { \mathbb p}}(v_1 \times v_2).\ ] ] we define analogously @xmath172
. we will denote @xmath173 .
observe that for any @xmath174 , @xmath175 for any @xmath176 .
if @xmath9 and @xmath168 are non empty , then @xmath177
. remark that given @xmath178 , @xmath179 , @xmath180 , then @xmath181 we will denote this variety by @xmath182 . given two projective toric varieties @xmath183 and @xmath184 ,
then their join is also a toric variety .
explicitly : [ obse : parajoin ] ( 1 ) let @xmath185 be a splitting of @xmath0 as a product of tori , @xmath186 and @xmath187 .
let @xmath188 , @xmath189 for all @xmath190 , and @xmath191 , @xmath192 for all @xmath193 .
then @xmath194 is a @xmath0-module for the product action @xmath195 .
moreover , @xmath5 decomposes in simple submodules as @xmath196 .
consider the @xmath197-toric varieties @xmath198 ( @xmath199 ) and let @xmath200 .
the projective toric variety associated to @xmath12 is then the join @xmath201 .
\(2 ) in the particular case when @xmath202 is a @xmath156-pyramidal configuration with @xmath203 as in definition [ def : pyramid ] , let @xmath204 , @xmath205 , @xmath206 and @xmath5 as above .
we then have that @xmath207 .
that is , @xmath61 is the _ cone over @xmath184 with vertex @xmath208_. next , we describe the toric varieties associated to configurations with repeated weights , as cones over non degenerate ( i.e. not contained in a proper linear subspace ) projective toric varieties .
[ lema : splittingjoin ] let @xmath209 be a configuration of @xmath86 weights , with @xmath113 appearing @xmath210 times and @xmath211 if @xmath212 . if we set @xmath213 , then the smallest linear subspace that contains @xmath61 has codimension @xmath156 . in particular
, @xmath61 is a non degenerate variety if and only if the configuration @xmath12 has no repeated elements .
let @xmath214 be a basis of associated eigenvectors of @xmath5 , with @xmath215 for all @xmath216 , @xmath217 .
consider a hyperplane @xmath218 of equation @xmath219 where @xmath220 are the coordinates in the basis @xmath221 .
then @xmath222 if and only if @xmath223\subset \pi$ ] for all @xmath224 . as @xmath225=[\sum \lambda_i(t ) v_{i , j_i } ] \in \pi$ ] , this is equivalent to the equalities @xmath226 since @xmath227 are different weights , we deduce that @xmath228 for all @xmath229 .
it follows that the maximum codimension of a subspace that contains @xmath61 is @xmath230 . on the other hand , clearly @xmath231 where the subspace @xmath232 has codimension @xmath156 .
[ lem : otromas ] let @xmath233 be a configuration of @xmath86 weights , with @xmath113 appearing @xmath210 times and @xmath211 if @xmath212 .
set @xmath213 and let @xmath234 with @xmath215 for all @xmath235 , @xmath236 , @xmath217 .
let @xmath237 and consider @xmath238 .
then @xmath61 is isomorphic to @xmath239 .
let @xmath240 the linear isomorphism defined by @xmath241 the associated projective map clearly sends @xmath61 to the join @xmath239 . in proposition [ prop : jointoric ] below
we combine remark [ obse : parajoin ] and lemmas [ lema : splittingjoin ] and [ lem : otromas ] , in order to describe a projective toric variety as a cone over a non degenerate projective _ toric _ variety that is not a cone ( that is , the associated configuration is non pyramidal ) .
[ rem : xnd ] let @xmath242 be a _ non linear _
irreducible projective variety .
let @xmath243 be the minimal linear subspace containing @xmath9 , and let @xmath156 be the codimension of @xmath149 .
then @xmath244 and if @xmath245 denotes the variety @xmath9 as a subvariety of @xmath149 , then @xmath246 .
since @xmath245 is non degenerate , it follows that there exists @xmath247 such that @xmath248 , where @xmath249 .
hence , we have an identification @xmath250 in particular , @xmath251 .
observe that @xmath247 is a non degenerate subvariety .
moreover , we can assume that @xmath168 is not a cone .
in this case , we will denote @xmath252 .
if moreover @xmath9 is an equivariantly embedded toric variety , then we can choose @xmath253 as @xmath254 in the following proposition .
when @xmath9 is linear , @xmath255 , @xmath256 and @xmath168 is empty .
[ prop : jointoric ] let @xmath233 be a configuration of @xmath86 weights , with @xmath113 appearing @xmath210 times and @xmath211 if @xmath212 . set @xmath213 and assume that @xmath237 is @xmath257-pyramidal .
then there exists a splitting @xmath185 such that , after reordering of the elements in @xmath258 , it holds that @xmath259 , where @xmath260 is a basis of @xmath261 and @xmath262 is a non pyramidal configuration , as in definition [ def : pyramid ] .
moreover , we have that @xmath263 in the special case when @xmath61 is linear , @xmath264 is empty .
we set @xmath265 , with @xmath215 for all @xmath235 , @xmath236 , @xmath217 , and @xmath266 .
assume that @xmath258 is a @xmath257-pyramidal configuration , and let @xmath267 .
then , there exists a splitting @xmath185 such that , after reordering of @xmath258 , @xmath260 is a basis of @xmath261 and @xmath262 is a non pyramidal configuration .
hence , @xmath268 by lemma [ lem : otromas ] , we can assume that @xmath269 , and so @xmath270 as claimed .
we recall the classical notion of the dual variety of a projective variety .
[ def : dual ] let @xmath5 be a @xmath10-vector space of finite dimension and denote by @xmath271 its dual @xmath1-vector space .
let @xmath11 be an irreducible projective variety .
the _ dual variety _ of @xmath9 is defined as the closure of the hyperplanes intersecting the regular part @xmath272 of @xmath9 non transversally : @xmath273\in { { \mathbb p}}(v^\vee ) \mathrel { : } \exists x\in x_{reg}\,,\ f|_{t_xx}\equiv 0\bigr\}}\subset { { \mathbb p}}(v^\vee).\ ] ] as usual , @xmath274 denotes the embedded tangent space of @xmath9 at @xmath275 .
note that @xmath276 .
we set by convention , @xmath277 .
self - duality is not an intrinsic property , it depends on the projective embedding .
it can be proved that @xmath7 is an irreducible projective variety and that @xmath278 ( see for example @xcite ) . for a generic variety @xmath11 , @xmath279 .
if @xmath280 , it is said that @xmath9 has _ defect _ @xmath281 .
[ def : selfdual ] an irreducible projective variety @xmath282 is called _ self - dual _ if @xmath9 is isomorphic to @xmath7 as embedded projective varieties , that is if there exists a ( necessarily linear ) isomorphism @xmath283 such that @xmath284 . a self - dual projective variety @xmath285 of dimension @xmath286 ( i.e. , which is not a hypersurface ) has positive defect @xmath287 .
the defect of the whole projective space @xmath288 is @xmath289 .
[ rk : identesp - espdual ] recall that given a basis @xmath290 of @xmath5 , we can identify @xmath43 with @xmath291 by means of @xmath292 , where @xmath293 is the dual basis of @xmath221 . then , via the choice of a basis of @xmath5 , we can look at the dual variety inside the same projective space .
self - duality can be reformulated as follows : @xmath6 is self - dual if there exists @xmath294 such that @xmath284 .
let @xmath5 be a @xmath0-module of finite dimension @xmath86 over a @xmath96-dimensional torus @xmath0 and let @xmath12 be the associated configuration of weights . in view of the considerations of the preceding subsections , we assume from now on and without loss of generality , that @xmath48 is a regular configuration , possibly with repeated elements , such that @xmath69 .
the regularity of @xmath12 implies in particular the existence of a splitting @xmath295 as in remark [ exa : regularconf ] .
then , @xmath61 is a @xmath296-dimensional subvariety of the @xmath297-dimensional projective space @xmath43 and the lattice @xmath134 has rank @xmath298 .
the dual variety @xmath299 has the following interpretation . for @xmath300\in{{\mathbb p}}(v^\vee)$ ] , let @xmath301 $ ] , @xmath302 $ ] .
then @xmath299 is obtained as the closure of the set of those @xmath300\in{{\mathbb p}}(v^\vee)$ ] such that there exists @xmath235 with @xmath303 for all @xmath304 .
@xmath305 in @xcite a rational parameterization of the dual variety @xmath299 was obtained .
we adapt this result to our notations . as before , @xmath45 is a basis of eigenvectors , @xmath25 , and @xmath306 is a basis of @xmath164 .
we denote by @xmath307 the @xmath308-dimensional @xmath1-vector space @xmath309 and we identify @xmath43 with @xmath291 by means of the chosen basis @xmath310 of eigenvectors ( and its dual basis ) as in remark [ rk : identesp - espdual ] .
[ thm : dualvarparam ] let @xmath311 as before .
then the mapping @xmath312 defined by @xmath313,s ) \mapsto s\cdot\bigl[\sum a_iv_i\bigr]\ ] ] has image dense in @xmath299 .
that is , the morphism @xmath314\ ] ] is a rational parameterization of @xmath299 , and @xmath315 this last equality , which expresses the dual variety as the closure of the union of the torus orbits of all the classes in the vector space of relations of the configuration @xmath12 , is the starting point of our classification of self - dual projective toric varieties , which we describe in the sequel .
let @xmath0 be a torus of dimension @xmath96 and @xmath5 a rational @xmath0-module of dimension @xmath86 with associated configuration of weights @xmath316 .
we assume that @xmath69 and keep the notations of the preceding section .
given @xmath317\in { { \mathbb t}}^{n-1}$ ] , we denote by @xmath318\bigr)=\bigl[\sum p_ix_iv_i\bigr]$ ] the diagonal linear isomorphism defined by @xmath319 . in this subsection we characterize self - dual projective toric varieties associated to a configuration of weights @xmath12 which define a _
non pyramidal _ configuration , in terms of the orbits of the torus action .
note that the whole projective space @xmath43 can be seen as a toric projective variety associated to a @xmath320-pyramidal configuration and its dual variety is empty .
but we now show that for non pyramidal configurations the dimension of the dual variety @xmath299 can not be smaller than the dimension of the toric variety @xmath61 .
this result has been proved by zak @xcite for any non degenerate smooth projective variety .
[ lema : nonpiramidaldim ] if @xmath12 is a non pyramidal configuration , then @xmath321 . indeed , if @xmath12 is not a pyramidal configuration , then by remark [ obsnucleo ] we know that there exists @xmath322 such that @xmath323 for all @xmath91 .
hence , if we identify @xmath43 with @xmath324 by means of the dual basis , then @xmath325)}= m_p\bigl(\overline{{{\mathcal o}}\bigl ( \bigl[{\textstyle \sum_iv_i}\bigr]\bigr)}\bigr ) = m_p(x_a).\ ] ] since @xmath326 , we have that @xmath327 and the result follows
. we identify @xmath43 with @xmath291 by means of the chosen basis @xmath310 of eigenvectors ( and its dual basis ) as in remark [ rk : identesp - espdual ] .
the following is the main result of this subsection .
[ teoeqauto ] let @xmath328 be a non pyramidal configuration .
the following assertions are equivalent . 1 .
@xmath61 is a self - dual projective variety .
there exists @xmath329 such that @xmath330 .
there exists @xmath331 such that @xmath332 .
4 . for all @xmath333 , @xmath334 . 5 . for all @xmath335 , @xmath336 .
we prove @xmath337 and @xmath338 , the rest of the implications being trivial .
@xmath337 : by proposition [ thm : dualvarparam ] , @xmath339 for all @xmath340 . since @xmath341 , equality holds in the last equation .
@xmath338 : let @xmath329 be such that @xmath342 . if @xmath343 , then @xmath344
. then , @xmath345 and the result follows .
the equivalence between ( i ) and ( v ) in theorem [ teoeqauto ] implies that as soon as an the dual of an equivariantly embedded projective toric variety of the form @xmath61 has the same dimension of the variety , there exists a linear isomorphism between them .
[ thm : autdim ] let @xmath328 be a configuration of weights which is non pyramidal . then @xmath61 is self - dual if and only if @xmath341 .
this result is not true in general for projective toric varieties not equivariantly embedded , even for rational planar curves ( for which the dual is again a curve , but not necessarily isomorphic ) .
we now address the complete characterization of self - dual projective toric varieties associated to an arbitrary configuration of weights @xmath346 .
we keep the notations of the preceding section .
we begin by recalling a well known result about duality of projective varieties : [ lem : tevelev ] let @xmath347 be a non linear irreducible subvariety .
\(1 ) assume that @xmath9 is contained in a hyperplane @xmath348 . if @xmath349 is the dual variety of @xmath9 , when we consider @xmath9 as a subvariety of @xmath288 , then @xmath7 is the cone over @xmath349 with vertex @xmath319 corresponding to @xmath149 .
\(2 ) conversely , if @xmath7 is a cone with vertex @xmath319 , then @xmath9 is contained in the corresponding hyperplane @xmath149 .
when @xmath9 is linear , @xmath350 is empty . as an immediate application of lemma [ lem : tevelev ]
, we have the following characterization of self - dual equivariantly embedded projective toric hypersurfaces .
note that the only linear varieties which are self - dual are the subspaces of dimension @xmath351 in @xmath352 .
in particular , the only hyperplanes which are self dual are points in @xmath353 . [ coro : noname ]
let @xmath0 be an algebraic torus and @xmath328 a configuration such that @xmath61 is a non linear hypersurface .
then @xmath61 is self - dual if and only if @xmath61 is not a cone .
assume that @xmath61 is a cone .
then by lemma [ lem : tevelev ] , it follows that @xmath299 is contained in a hyperplane , hence @xmath61 is not self - dual .
if @xmath61 is not a cone , then @xmath12 is non pyramidal ( see remark [ obse : parajoin ] ) , and it follows from lemma [ lema : nonpiramidaldim ] that @xmath354 . if @xmath355 , then @xmath356 and hence @xmath357 , which is a contradiction .
it follows that @xmath358 and hence theorem [ thm : autdim ] implies that @xmath61 is self - dual .
applying lemma [ lem : tevelev ] , we can reduce the study of duality of projective varieties to the study of non degenerate projective varieties that are not a cone .
[ prop : arb ] let @xmath242 be an irreducible projective variety .
let @xmath351 be the codimension of the minimal subspace of @xmath359 containing @xmath9 .
then , with the notations of remark [ rem : xnd ] , the following assertions hold : \(1 ) if @xmath360 , with @xmath361 self - dual , then @xmath9 is self - dual .
\(2 ) if @xmath9 is self - dual , then @xmath362 , and @xmath363 , that is @xmath364 let @xmath365 . applying recursively lemma [ lem : tevelev ] ( see remark [ rem : xnd ] ) we obtain that @xmath366 in particular , @xmath367 , and the maximal subspace that contains @xmath7 has codimension @xmath368 . in order to prove
( 1 ) , assume that @xmath363 and @xmath253 is self - dual .
then @xmath369 .
since @xmath253 is self - dual , there exists an isomorphism @xmath370 such that @xmath371 .
it is clear that @xmath144 extends to an isomorphism @xmath372 such that @xmath373 . in order to prove ( 2 ) , assuming @xmath9 is self - dual and writing @xmath9 as in remark [ rem : xnd ] , it follows that @xmath363 , and hence @xmath374 . in our toric setting , proposition [ prop : arb ] can be improved , so that we obtain a geometric characterization of self - dual projective toric varieties .
[ thm : autdimgral ] let @xmath12 be an arbitrary lattice configuration .
then @xmath61 is self - dual if and only if @xmath375 and the smallest linear subspaces containing @xmath61 and @xmath299 have the same ( co)dimension . by proposition [ prop : jointoric ] , @xmath376 where @xmath377 is a non pyramidal configuration without repeated weights . by theorem
[ thm : autdim ] , @xmath378 is self dual if and only if @xmath379 . the result follows now from proposition [ prop : arb ] . combining proposition [ prop : jointoric ] and theorem [ thm : autdimgral ] we obtain the following explicit combinatorial description of self - dual toric varieties .
[ teo : constaut+gde ] let @xmath380 be a configuration of @xmath86 weights with each @xmath113 appearing @xmath210 times , @xmath381 if @xmath212 .
let @xmath382 be the associated configuration without repeated weights .
then @xmath61 is self - dual if and only if the following assertions hold .
\(i ) @xmath258 is a @xmath156-pyramidal configuration , where @xmath383 .
( ii)there exists a splitting @xmath185 such that , after reordering of the elements in @xmath258 , it holds that @xmath259 , where @xmath384 is a basis of @xmath261 and @xmath385 is a non pyramidal configuration , as in remark [ obse : parajoin ] .
definition [ def : pyramid ] .
moreover , the @xmath386toric projective variety @xmath387 , @xmath388 , is self - dual .
it follows from theorem [ teo : constaut+gde ] that if @xmath61 is a self - dual toric variety with @xmath12 is pyramidal , then there are repeated weights in @xmath12 .
note that any configuration in @xmath389 of the form @xmath390 where @xmath391 , with @xmath392 the first canonical basis vector and @xmath393 non pyramidal , has repeated weights without begin pyramidal ( but it becomes a pyramid when we avoid repetitions ) .
in this section we will characterize self - duality of projective toric varieties of type @xmath61 in combinatorial terms .
we make explicit calculations for the algebraic torus @xmath394 acting on @xmath395 , in order to give an interpretation of the conditions of theorem [ teoeqauto ] in terms of the configuration @xmath12 and in terms if its _ gale dual configuration _ , whose definition we recall below .
we refer the reader to ( * ? ? ?
* chapter 6 ) for an account of the basic combinatorial notions we use in what follows .
let @xmath397 .
we identify the lattice of characters @xmath398 with @xmath399 .
thus , any character @xmath400 is of the form @xmath401 , where @xmath402 and @xmath403 .
we take the canonical basis of @xmath395 as the basis of eigenvectors of the action of @xmath0 .
that is , if @xmath404 , @xmath0 acts on @xmath395 by @xmath405 for all @xmath406 .
then , @xmath407\bigr)}= \overline{\bigl\{[t^{\lambda_1}:\dots : t^{\lambda_n}]:\,t\in ( \bbbk^{*})^d\bigr\ } } \subset { { \mathbb p}}^{n-1}.\ ] ] by abuse of notation we also set @xmath408 the matrix with columns the weights @xmath113 . in view of the reductions made in section [ sec : prelim ] we assume without loss of generality that the first row of @xmath12 is @xmath409 and that the columns of @xmath12 span @xmath122 . the homogeneous ideal @xmath410 in @xmath411 $ ] of the associated projective toric variety @xmath61 is the binomial ideal ( @xcite ) @xmath412 thus , @xmath413 \in { { \mathbb p}}^{n-1}:\;x^{a}=x^{b } , \;\forall\;a , b \in \mathbb{n}^n\;\text{such that}\;aa = ab\bigr\},$ ] and it is easy to see that @xmath414 \in { { \mathbb p}}^{n-1}:\,x^{v^{+}}-x^{v^{-}}=0,\,\forall\,v\in \mathcal r_a\bigr\},\ ] ] where @xmath415 , @xmath416 ( and so @xmath417 ) . for @xmath418
we then have @xmath419 \in { { \mathbb p}}^{n-1}\mathrel { : } p^{v^{-}}x^{v^{+}}-p^{v^{+}}x^{v^{-}}=0 , \,\forall\,v \in \mathcal r_a \bigr\}. \end{split}\ ] ] if @xmath12 is a non pyramidal configuration , then theorem [ teoeqauto ] can be rephrased in terms of a geometric condition on the gale dual of @xmath12 .
[ def : gale ] let @xmath408 with rank @xmath96 .
let @xmath420 be a basis of @xmath164 .
we say that the matrix @xmath421 with columns the vectors @xmath422 is a _ gale dual matrix _ of @xmath12 .
let @xmath423 be the configuration of rows of @xmath424 , that is @xmath425 ( observe that we allow repeated elements ) .
the configuration @xmath426 is called a _ gale dual _ configuration of @xmath12 .
remark that @xmath427 .
\(1 ) since @xmath164 is an affine invariant of the configuration @xmath12 , it follows that two affinely equivalent configuration share their gale dual configurations .
\(2 ) the configuration @xmath12 is non pyramidal if and only @xmath428 for all @xmath91 .
\(3 ) when @xmath12 is regular , @xmath164 is the integer kernel @xmath429 of the matrix @xmath12 .
\(4 ) for any gale dual matrix of @xmath12 , the morphism @xmath430 , defined by @xmath431 gives a parameterization of @xmath432 , where we denote @xmath433 .
[ rmk : exproof ] by theorem [ teoeqauto ] , @xmath61 is self - dual if and only if there exists @xmath434 such that @xmath435 . by the remarks in subsection
[ rk : ector ] , it follows that @xmath61 is self - dual if and only if for some such @xmath436 we have that @xmath437 for all @xmath438 and @xmath439 .
assume @xmath61 is self - dual .
then , given any choice of gale dual configuration , we deduce that for all @xmath440 and @xmath441 , we have that @xmath442 for ( some , or in fact all ) @xmath434 .
since this gives an equality in the polynomial ring @xmath443 $ ] , both sides must have the same irreducible factors .
but @xmath444 and @xmath445 are associated irreducible factors if and only if @xmath446 and @xmath447 are collinear vectors .
we deduce that for any line @xmath448 in @xmath449-space @xmath450 and for all @xmath451 , @xmath452 hence , @xmath453 for all @xmath454 , or equivalently , @xmath455 .
in fact , this last condition is not only necessary but also sufficient .
we give a proof of both implications using results about the tropicalization of the dual variety @xmath61 as described in @xcite .
first we recall that given a dual gale configuration @xmath456 , and a subset @xmath457 , the _ flat _
@xmath458 of @xmath426 associated to @xmath459 is the subset of all the indices @xmath460 such that @xmath446 belongs to the subspace generated by @xmath461 .
[ teo : bside ] let @xmath462 a non pyramidal configuration and @xmath424 a gale dual for @xmath12 as in ( [ def : gale ] ) .
then @xmath463 is self - dual if and only if for any line @xmath448 through the origin in @xmath450 we have that @xmath464 .
since we are dealing with affine invariants , we can assume that @xmath12 is a regular configuration .
by theorem [ thm : autdim ] , we know that @xmath61 is self - dual if and only if @xmath112 equals @xmath465 . given a vector @xmath466 , we define a new vector @xmath467 by @xmath468 if @xmath469 and @xmath470 if @xmath471 .
if follows from ( * ? ? ? * corollary 4.5 ) that @xmath472 if and only if for any vector @xmath439 , the vector @xmath473 lies in the row span @xmath474 of the matrix @xmath12 .
but since we are assuming that @xmath475 , this is equivalent to the condition that @xmath476 . by duality , this is in turn equivalent to the fact that for any @xmath477 , the inner product @xmath478 that is to say , @xmath61 is self dual if and only if for any @xmath439 the sum @xmath479 .
but the sets @xmath480 of non zero coordinates of the vectors in the space of linear relations @xmath164 coincide with the flats of the gale configuration @xmath426 .
so , @xmath61 is self - dual if and only if for any flat @xmath481 the sum @xmath482 .
it is clear that this happens if and only if the same condition holds for all the one - dimensional flats , i.e. if for any line @xmath448 through the origin the sum @xmath483 .
the assumption that @xmath12 is a non pyramidal configuration in theorem [ teo : bside ] is crucial , as the following example shows .
let @xmath12 be a configuration such that @xmath164 has rank @xmath20 .
then @xmath164 is spanned by a single vector , whose coordinates add up to @xmath2 .
so , the condition in theorem [ teo : bside ] that the sum of the @xmath446 in this line equals @xmath2 is satisfied .
but by corollary [ coro : noname ] if @xmath12 is a pyramid , then @xmath61 is not self - dual . in this paragraph
we characterize the non pyramidal configurations @xmath484 whose gale dual configurations are as in theorem [ teo : bside ] .
we keep the assumptions that @xmath485 and that @xmath12 is non pyramidal .
we begin with some basic definitions about configurations . given @xmath486 , we call @xmath487 the _ support _ of the relation and denote @xmath488 .
we say that @xmath113 belongs to the relation if @xmath489 .
recall that any affine relation @xmath490 satisfies @xmath136 .
it is said that @xmath491 is a _ circuit _ if there is no non trivial affine dependency relation with support strictly contained in @xmath492 .
in other words , a circuit is a minimal affine dependency relation .
[ rem : circuitnumber ]
let @xmath258 be a circuit of a configuration @xmath12 , and let @xmath474 be the minimal face containing @xmath258 . if @xmath493 denotes the dimension of the affine span of @xmath474 , then @xmath258 has at most @xmath494 elements . two elements @xmath495 of a configuration @xmath449 are _ parallel _ if they generate the same straight line trough the origin .
in particular , @xmath496 and @xmath497 .
the elements @xmath495 are _ antiparallel _ if they are parallel and point into opposite directions .
two elements @xmath498 of a configuration @xmath12 are _ coparallel _ if they belong exactly to the same circuits .
\(1 ) coparallelism is an equivalence relation .
we denote by @xmath499 the coparallelism class of the element @xmath500 .
\(2 ) it is easy to see that @xmath501 and @xmath502 are coparallel if and only if they belong to the same affine dependency relations .
\(3 ) the definition of coparallelism can be extended to pyramidal configurations as follows . if @xmath100 is such that it does not belong to any dependency relation , then @xmath503 .
otherwise , @xmath499 consists , as above , of all elements of @xmath12 belonging to the same circuits as @xmath501 .
the condition that @xmath12 is not a pyramid is then equivalent to the condition that @xmath504 for all @xmath100 .
[ lem : copara ] let @xmath505 be a gale dual of @xmath12 .
then @xmath113 is coparallel to @xmath506 if and only if @xmath507 and @xmath508 are parallel elements of @xmath426 .
let @xmath424 the @xmath509-matrix with rows given by @xmath426 as in definition [ def : gale ] . as @xmath12 is not a pyramid ,
no row @xmath446 of @xmath424 is zero .
any element @xmath510 is of the form @xmath511 , for some @xmath512 .
then @xmath113 is coparallel to @xmath506 if and only if for any nonzero @xmath512 it holds that @xmath513 precisely when @xmath514 .
it is clear that this happens if and only if @xmath515 for a non zero constant @xmath516 ,
that is , if and only if @xmath517 are parallel . let @xmath518 be a configuration .
a subconfiguration @xmath519 is a _ face complement _ if @xmath520 is the intersection of the configuration @xmath12 with a face of @xmath521 , the convex hull of @xmath12 .
[ rem : deprel ] let @xmath518 be a configuration . a subconfiguration @xmath519 is a face complement if and only if there exists a dependency relation such that @xmath522 indeed , a dependency relation @xmath523 with all @xmath524 can be extended with zero coordinates to a relation @xmath525 among all @xmath446 s .
thus , @xmath257 lies in the row space of @xmath12 and so there exists @xmath526 such that @xmath527 .
it follows that the linear form associated to @xmath528 vanishes on the complement of @xmath258 , and all the points of @xmath258 lie in the same open half space delimited by the kernel of @xmath528 .
[ lem : paracop ] let @xmath518 be a configuration .
a coparallelism class @xmath519 is a face complement if and only if and only if there exist @xmath529 @xmath530 and @xmath531 are antiparallel . if @xmath258 is a coparallelism class , we know by lemma [ lem : copara ] that all @xmath532 are parallel .
it is then clear that a dependency relation @xmath257 as in remark [ rem : deprel ] exists if and only if two of the vectors @xmath533 are antiparallel .
[ def : pfc ] let @xmath534 be a configuration and @xmath535 a face complement .
we say that @xmath258 is a _ parallel face complement _ if @xmath258 and @xmath536 lie in parallel hyperplanes .
note that in this case both @xmath258 and @xmath520 are facial . in figure 1 below
there are three configurations of @xmath537 lattice points in @xmath538-dimensional space ( the @xmath537 vertices in each polytope ) .
the @xmath19 vertices marked with big dots in each of the configurations define a coparallelism class @xmath258 . in the first polytope @xmath539 ,
@xmath258 is not a face complement ; in the second polytope @xmath540 , @xmath258 is a face complement but not a parallel face complement ; in the third polytope @xmath541 , @xmath258 is a parallel face complement .
the characterization in our next theorem proves that only the toric variety corresponding to this last configuration is self - dual .
is self - dual . ]
it is straightforward to check that if @xmath542 are affinely equivalent configurations and @xmath144 is an affine linear map sending bijectively @xmath145 to @xmath146 , then @xmath144 preserves coparallelism classes , faces and parallelism relations .
indeed , all these notions can be read in a common gale dual configuration . moreover
, we can translate theorem [ teo : bside ] as follows .
[ teo : aside ] let @xmath534 be a non pyramidal configuration .
the projective toric variety @xmath61 is self - dual if and only if any coparallelism class of @xmath12 is a parallel face complement .
let @xmath426 be a gale dual of @xmath12 as in definition [ def : gale ] .
by lemma [ lem : copara ] , coparallelism classes @xmath543 in @xmath12 are in correspondence with parallel vectors @xmath532 in the dual space ( i.e. lines containing vectors of @xmath426 ) . but now , @xmath258 is a parallel face complement if and only if there exists @xmath544 such that @xmath545 for all @xmath546 and @xmath547 for all @xmath548 .
reciprocally , the sum of the vectors @xmath549 implies the existence of such an @xmath528 as in the proof of lemma [ lem : paracop ] .
the result now follows from theorem [ teo : bside ] .
we have the following easy lemma .
[ lem : ccpfc ] assume that @xmath12 is a non pyramidal self - dual lattice configuration . for any @xmath550 ,
the coparallelism class @xmath551 has at least two elements and it is a face of @xmath99 .
it follows from definition [ def : pfc ] that there exists a linear function @xmath58 taking value @xmath2 on @xmath552 and value @xmath20 on @xmath551 .
then , @xmath551 is the facial subset of @xmath12 supported by the hyperplane @xmath553 .
if @xmath554 , then by theorem [ teo : aside ] @xmath555 is a face , and hence @xmath12 would be a pyramid .
it follows that so @xmath556 , for any @xmath550 .
we give in lemma [ lem : easy ]
ii ) an example of a self - dual lattice configuration @xmath12 which has an interior point . however , this can not happen if @xmath61 is not a hypersurface , as the following proposition shows . [
prop : onlyvertex ] let @xmath328 be a configuration without repetitions such that @xmath61 is self - dual , with @xmath557 .
then the interior of the convex hull @xmath99 does not contain elements of @xmath12 and any facial subset of @xmath12 contains at most one lattice point in its interior .
since @xmath558 has no repeated elements , it follows from theorem [ teo : constaut+gde ] that @xmath12 is non pyramidal .
then , as @xmath61 is not a hypersurface , if follows from remark [ rem : variado ] that @xmath559 , where @xmath96 is the dimension of the affine span of @xmath12 .
assume that there exists @xmath550 belonging to the relative interior of an @xmath560-dimensional face @xmath474 of @xmath99 .
then @xmath561 is a convex combination of the vertices of @xmath474 , and thus @xmath562 .
but by lemma [ lem : ccpfc ] , @xmath551 is a facial subset of @xmath12 , and thus a face of @xmath474 , which intersects the relative interior of @xmath474 .
then , @xmath563 .
then , @xmath564 , with @xmath565 vertices of @xmath474 .
we claim that @xmath566 are affinely independent and thus , @xmath551 is a circuit . indeed ,
for any @xmath567 , @xmath568 , and so there can not exist an affine dependence relation with support strictly contained in @xmath569 .
in particular , @xmath570 , @xmath571 are the vertices of @xmath474 and @xmath561 is the only interior lattice point in @xmath572 . therefore , if the relative interior of @xmath99 contains one element @xmath573 , it follows that @xmath12 is a circuit , and hence @xmath97 , see remark [ rem : circuitnumber ] .
that is , @xmath61 is an hypersurface .
[ ex : intpointface ] consider the configuration @xmath12 given by the columns of the matrix @xmath574 the associated toric variety has dimension @xmath538 in @xmath575 , so it is not a hypersurface .
no point of @xmath576 lies in the interior , but there are two faces of @xmath12 ( namely , the segments with vertices @xmath577 and @xmath578 , respectively ) which do have a point of @xmath12 in their relative interior .
note that @xmath579 is not smooth .
it is a complete intersection but the four fixed points @xmath580 are not regular , as can be checked by the drop in rank of the jacobian matrix .
this could be seen directly in the geometry of the configuration .
the convex hull of @xmath12 is a simple polytope ( in fact , it is a simplex ) of dimension @xmath538 lying in the hyperplane @xmath581 , but fixing the origin at any of the four vertices , the first lattice points in the @xmath538 rays from that vertex do not form a basis of the lattice @xmath582 . note that there is a splitting of the @xmath583-torus @xmath584 as a product of tori of dimension @xmath19 corresponding respectively to the first three and last three weights in @xmath12 .
we end this paragraph by showing another interesting combinatorial property of configurations associated to self - dual toric varieties .
[ prop : passagetoface ] let @xmath102 be a non pyramidal configuration such that @xmath61 is self dual and let @xmath585 be an arbitrary non empty subset of @xmath12 . then , either @xmath585 is a pyramidal configuration or @xmath586 is self - dual and , moreover , there exists a face @xmath474 of @xmath99 such that @xmath587 .
assume that @xmath588 is non pyramidal , and consider @xmath589 .
it is clear that @xmath590 .
hence , if @xmath591 is a basis of @xmath592 , then there exists a @xmath593-basis of @xmath594 of the form @xmath595 .
let @xmath596 be a @xmath597-basis of @xmath164 , and @xmath598 its associated gale dual configuration .
then there exists an invertible @xmath599-matrix @xmath121 such that @xmath600 since @xmath12 is self - dual , it follows from theorem [ teo : bside ] that the rows @xmath446 are such that the sum of vectors @xmath446 in the same line through the origin is zero .
hence , the matrix @xmath601 satisfies the same property . as @xmath585 is non pyramidal ,
no row of @xmath602 is zero .
therefore , @xmath603 , and hence @xmath602 , also satisfy the property that the sum of all its row vectors in a line trough the origin is equal to zero .
hence , @xmath586 is self - dual .
moreover , the sum of the row vectors of @xmath264 is zero , and it follows from remark [ rem : deprel ] that @xmath585 is facial .
in this section we use our previous results in order to obtain new families of projective toric varieties that are self - dual . in particular , we obtain many new examples of non smooth self - dual projective varieties .
we also identify all the smooth self - dual projective varieties of the form @xmath61 .
we retrieve in this ( toric ) case ein s result , without needing to rely on hartshorne s conjecture .
[ df : law ] we say that a configuration @xmath12 of @xmath604 lattice points is _ lawrence _ if it is affinely equivalent to a configuration whose associated matrix has the form @xmath605 where @xmath606 denotes the @xmath607 identity matrix .
equivalently , @xmath12 is a lawrence configuration if it is affinely equivalent to a cayley sum of @xmath86 subsets , each one containing the vector @xmath2 and one of the column vectors of @xmath121 .
lawrence configurations are a special case of _ cayley configurations _
( see @xcite ) .
the lawrence configuration associated to the matrix is the cayley configuration of the two - point configurations consisting of the origin and one column vector of @xmath121 . in the smooth case , cayley configuration of strictly equivalent polytopes
correspond to toric fibrations ( see @xcite ) .
it is straightforward to verify that if @xmath12 is lawrence , then 1 .
@xmath12 is pyramidal if and only if @xmath121 is pyramidal .
we immediately deduce from theorem [ teo : bside ] the following result .
[ lauself - duales ] if @xmath12 is a non pyramidal lawrence matrix then @xmath61 is self - dual .
[ ex : segre ] the well known fact that the segre embedding of @xmath16 in @xmath609 is self - dual follows directly from corollary [ lauself - duales ] , the image of the segre morphism @xmath610,\ ] ] where @xmath611 , y = [ y_0:y_1:\dots : y_m]$ ] , is a projective toric variety with associated matrix @xmath612 & & & 1 & 0\ar@{.}[rrr ] & & & 0\\ 0\ar@{.}[rrr ] & & & 0 & 1\ar@{.}[rrr ] & & & 1 \\ 1\ar@{.}[dddrrr ] & 0\ar@{.}[rr]\ar@{.}[ddrr ] & & 0\ar@{.}[dd ] & 1\ar@{.}[dddrrr ] & 0\ar@{.}[rr]\ar@{.}[ddrr ] & & 0\ar@{.}[dd]\\ 0\ar@{.}[dd]\ar@{.}[ddrr ] & & & & 0\ar@{.}[dd]\ar@{.}[ddrr ] & & & \\ & & & 0 & & & & 0\\ 0\ar@{.}[rr ] & & 0 & 1 & 0\ar@{.}[rr ] & & 0 & 1 } } \right)\ ] ] the sum of the first two rows equals the sum of the last @xmath613 rows .
it is easy to see that @xmath12 is affinely equivalent to the configuration @xmath614 with associated matrix @xmath615 the matrix @xmath614 is a non pyramidal lawrence matrix , hence @xmath616 is self - dual .
we end this paragraph by proving that segre embeddings of @xmath617 , @xmath618 are the unique smooth self - dual projective toric varieties that are not a hypersurface .
we begin with an easy lemma which classifies all smooth hypersurfaces of the form @xmath61 [ lem : easy ] let @xmath12 be a lattice configuration such that @xmath61 is a smooth hypersurface .
then , @xmath12 is of one of the following forms : * @xmath12 consists of a two equal points , and so @xmath619 . * @xmath12 consists of three collinear points with one of them the mid point of the others , and so @xmath620 .
* @xmath12 consists of four points @xmath621 with @xmath622 , and so @xmath623 is the segre embedding of @xmath624 in @xmath625 . when @xmath61 is a hypersurface , an equation for @xmath61 is given by @xmath626 , where the transpose of the row vector @xmath627 is a choice of gale dual of @xmath12 .
the cases i ) , ii ) and iii ) in the statement correspond to the row vectors @xmath628 and @xmath629 , respectively ( or any permutation of the coordinates ) , and it is straightforward to check that @xmath61 is smooth .
it is easy to verify that in any other case , there exists a point @xmath630 where @xmath631 and all its partial derivatives vanish at @xmath632 .
we saw in example [ ex : intpointface ] that a non - pyramidal self - dual lattice configuration @xmath12 with @xmath633 can have a point in the interior of a proper face .
moreover , more complicated situations can happen : [ ex : missingpts ] consider the following dimension @xmath538 configuration @xmath634 , @xmath635 .
then , @xmath636 and @xmath61 is self - dual because the following is a choice of gale dual @xmath637 : @xmath638 all the points in @xmath12 are vertices of the polytope @xmath639 , but @xmath640 .
indeed , there is a lattice point in the middle of each of the segments @xmath641 $ ] , @xmath642 $ ] , which are faces of @xmath643 .
it is clear that @xmath61 is not smooth ( for instance looking at the first lattice points in all the edges emanating from @xmath644 ) ( nor embedded by a complete linear system ) .
however , the following result shows that when @xmath61 is smooth and self - dual , the situation is nicer .
[ lem : descartando ] let @xmath12 be a lattice configuration without repeated points such that @xmath61 is self - dual and _
smooth_. then , unless @xmath61 is the quadratic rational normal curve in ii ) of lemma [ lem : easy ] , no facial subset of @xmath12 contains a point of @xmath12 in its relative interior .
assume @xmath645 has no repeated points and there exists @xmath550 and a proper face @xmath474 of @xmath99 containing @xmath561 in its relative interior .
then , @xmath646 is not a pyramid , and it follows from proposition [ prop : passagetoface ] that @xmath647 is self - dual . since @xmath648 is also smooth , proposition [ prop : onlyvertex ] implies that @xmath647 is a hypersurface .
we deduce from lemma [ lem : easy ] that @xmath572 has dimension one and consists ( up to reordering ) of @xmath538 points @xmath649 with @xmath650 .
we can choose a gale dual @xmath449 of @xmath12 of the form : @xmath651 with @xmath652 the @xmath653 column vector with transpose @xmath654 .
we see that the coparallelism class of each @xmath113 is contained in @xmath572 and no class can consist of a single element because @xmath12 is not a pyramid
. therefore , @xmath655 ; that is , any two of the first @xmath538 rows of @xmath449 are linearly dependent
. we can thus find another choice of gale dual @xmath601 of @xmath12 of the form : @xmath656 then , there is a splitting of the torus and @xmath61 can not be smooth , with arguments similar to those in example [ ex : intpointface ] , because @xmath12 has no repeated points and so there is no linear equation in the ideal @xmath410 .
we now characterize the segre embeddings @xmath16 in @xmath609 from example [ ex : segre ] in terms of the gale dual configuration .
[ lem : bsegre ] a toric variety @xmath657 is the segre embedding of @xmath16 if and only if any gale dual @xmath658 of @xmath12 has the following form : @xmath659 and , up to reordering , the rows of @xmath660 of @xmath449 satisfy @xmath661 and @xmath662 , for all @xmath663 .
it is clear that any gale dual to the matrix @xmath614 in is of this form . and
it is also straightforward to check that any matrix @xmath449 as in the statement is a gale dual of this @xmath614 .
we can now prove the complete characterization of smooth self - dual varieties @xmath61 .
[ th : regcase ] the only self - dual smooth non linear projective toric varieties equivariantly embedded are the toric hypersurfaces described in ii ) and iii ) of lemma [ lem : easy ] and the segre embeddings @xmath16 in @xmath609 for @xmath664 .
we proceed by induction in the codimension of @xmath12 . by lemma [ lem :
easy ] , the result is true when @xmath61 is a hypersurface .
assume then that @xmath633 .
now , by lemma [ lem : descartando ] , we know that all the points in @xmath12 are vertices of @xmath99 .
let @xmath258 be a coparallelism class and let @xmath665 .
then , @xmath586 is smooth and it is non pyramidal .
indeed , we can choose a gale dual @xmath449 of @xmath12 of the form : @xmath666 where @xmath667 correspond to the elements of @xmath258 .
if @xmath585 is a pyramid , is is easy to show that at least one row of @xmath668 must be zero , and it follows that the corresponding point of the configuration belongs also to @xmath258 , and thus is a contradiction .
hence , it follows from proposition [ prop : passagetoface ] that @xmath586 is self - dual with @xmath669 and no interior point . therefore , by induction , @xmath586 is the segre embedding of @xmath670 in @xmath671 for @xmath672 ( including the hypersurface case @xmath624 ) .
in particular , @xmath673 is even .
assume @xmath674 .
let @xmath675 be a choice of gale dual of @xmath585 as in lemma [ lem : bsegre ] , with rows @xmath676 with @xmath677 a basis of @xmath678 and @xmath679 .
add another integer affine relation with coprime entries as the first column , to form a matrix @xmath601 whose columns are a @xmath15-basis of relations of @xmath12 of the form : @xmath680 now , each coparallelism class of any @xmath681 ( with respect to @xmath585 ) has two elements when @xmath682 , and so it can not be `` broken '' when considering coparallelism classes in @xmath12 , since it is not a pyramid . then , via column operations we can assume that @xmath683 is of the form @xmath684 . in case @xmath685 , then @xmath686 and the unique coparallelism class could be broken , but at most in two pieces with two elements each , and again we have the same formulation for @xmath683 . in both cases ,
if @xmath687 , then we have a splitting , which implies that either there is a repeated point ( if @xmath688 ) or @xmath61 is not smooth
. then @xmath689 .
consider the subconfiguration @xmath690 of @xmath12 obtained by forgetting the two columns corresponding to the rows @xmath691 and @xmath692 of @xmath602 .
since the vectors @xmath446 with complementary indices add up to zero , it follows that @xmath690 is facial and again , @xmath693 is smooth .
we deduce that @xmath694 and @xmath695 , which implies that @xmath61 is the segre embedding of @xmath696 in @xmath697 .
we begin by recalling a result of landman ( @xcite ) for smooth varieties : if @xmath285 is a non linear smooth projective variety such that @xmath698 with defect @xmath699 , then @xmath700 .
we have the following easy corollary .
[ cor : lardman ] let @xmath701 with maximal rank @xmath96 associated to a regular configuration of weights and let @xmath702 be the projective toric variety associated to @xmath12 . if @xmath61 is non linear , smooth and self - dual then @xmath86 is even .
the dimension of the projective @xmath61 is @xmath117 because @xmath12 is regular .
then if @xmath61 is self - dual we have that the defect of @xmath61 equals @xmath703 . hence , if @xmath61 is moreover non linear and smooth we deduce from landman s result that @xmath704 or equivalently , @xmath705 , and so @xmath86 must be even .
we use the previous corollary together with theorem [ teo : bside ] to construct families of non regular self - dual varieties .
[ ej : familia ] consider the families of matrices @xmath706 for @xmath707 , @xmath708 defined by : @xmath709 clearly , @xmath710 is a choice of a gale dual matrix of @xmath711 . observe that as @xmath712 , the configuration @xmath711 is not a pyramid and @xmath713 .
moreover , it is easy to show that if @xmath714 , then @xmath715 and @xmath716 are not isomorphic as embedded varieties because they have different degree .
the degree of @xmath717 is the normalized volume of the convex hull of the points in the configuration @xmath711 ( @xcite ) and it can be computed easily in terms of the gale dual configuration .
since the conditions of theorem [ teo : bside ] hold , it follows that @xmath715 is self - dual for all @xmath718 .
moreover , @xmath719 is odd and so we deduce from corollary [ cor : lardman ] that @xmath715 is a singular variety .
the difference between its dimension and its defect is @xmath720 .
we can generalize example [ ej : familia ] in order to construct families of non degenerate projective toric self - dual varieties of arbitrary dimension greater than or equal to @xmath538 and of arbitrary codimension greater than or equal to @xmath19 .
_ families of self - dual varieties of any dimension @xmath721_. let any @xmath722 and @xmath723 non zero integer numbers satisfying @xmath724 .
consider the planar lattice configuration @xmath725 let @xmath12 be any lattice configuration with gale dual @xmath726 .
then , @xmath12 is not a pyramid and the associated projective toric variety @xmath727 is self - dual by theorem [ teo : bside ] , with dimension @xmath728 . when @xmath729 , the dimension of @xmath715 is @xmath538 .
the case @xmath730 corresponds to the segre embedding of @xmath731 in @xmath575 .
already for @xmath732 , the configuration @xmath711 does not contain all the lattice points in its convex hull .
if we add those `` remaining '' points to the configuration , the associated toric variety is no longer self - dual . _ families of self - dual varieties of any codimension @xmath733_. using the same ideas of the previous example
, we can construct pairs @xmath734 with @xmath12 a non pyramidal configuration and @xmath449 its gale dual satisfying the hypothesis of theorem [ teo : bside ] , so that @xmath61 is self - dual , with arbitrary codimension @xmath17 . for any @xmath722 set @xmath735 . as usual
, @xmath736 denotes the canonical basis in @xmath737 . for any choice of non zero integers @xmath738 with @xmath724 consider the following lattice configuration in @xmath737 : @xmath739 for any lattice configuration @xmath740 with this gale dual
, @xmath711 is not a pyramid and its associated self - dual toric variety @xmath741 has dimension @xmath742 and codimension @xmath613 .
we are interested now in characterizing a particular interesting case of self - dual projective toric varieties .
[ def : strong ] we say that a projective variety @xmath743 is _ strongly self - dual _ if @xmath9 coincides with @xmath744 under the canonical identification between @xmath288 and its dual projective space as in remark [ rk : identesp - espdual ] .
assume that this condition holds and @xmath12 is non pyramidal .
it follows from theorem [ teoeqauto ] that for any @xmath748 , @xmath749 .
but since @xmath750\bigr)$ ] , we deduce that @xmath751\bigr)}=x_a$ ] , that is @xmath752 .
assume that @xmath61 is strongly self - dual .
then ( a ) holds by theorem [ teo : bside ] . by proposition [ prop : strong ] , we know that @xmath755 , and this last variety is cut out by the @xmath308 binomials @xmath756
conversely , condition ( a ) implies the equalities @xmath760 of the polynomials in @xmath560 on both sides up to constant , as in remark [ rmk : exproof ] .
then , condition ( b ) ensures that this constant is @xmath20 .
therefore , @xmath755 , and so @xmath61 is strongly self - dual by proposition [ prop : strong ] .
[ teo : caraclawself - dual ] let @xmath12 be a non pyramidal lawrence configuration consisting of @xmath604 points in @xmath763 , as in @xmath764 .
then @xmath61 is strongly self - dual if and only if there exists a subset @xmath765 of rows of the lower matrix @xmath766 such that @xmath767 is an odd number for all @xmath768 . by corollary [ lauself - duales ] ,
@xmath61 is self - dual for any non pyramidal lawrence configuration @xmath12 .
thus , @xmath61 is strongly self - dual if and only if conditions ( b ) in theorem [ teo : stself - dual ] are satisfied .
if @xmath769 is a gale dual configuration for @xmath121 , then @xmath770 defines a gale dual configuration for @xmath12 . conditions ( b ) are then equivalent in this case to the equalities @xmath771 this is in turn equivalent to the condition that for all @xmath772 , the sum @xmath773 .
but this is equivalent to the fact that the vector @xmath774 lies in the row span of @xmath121 when we reduce all its entries modulo @xmath19 .
denoting classes in @xmath775 with an over - line , this condition means that there exist @xmath776 such that @xmath777 the segre embeddings in example [ ex : segre ] have associated lawrence matrices as in ( [ eq : segre ] ) , where @xmath121 is a matrix with a single row of with all entries equal to @xmath20 .
they clearly satisfy the hypotheses of theorem [ teo : caraclawself - dual ] .
then , for any @xmath779 , the segre embedding of @xmath16 is a strongly - self dual projective toric variety .
popov , _ self - dual algebraic varieties and nilpotents orbits_. proceedings of the international colloquium on algebra , arithmetic and geometry , mumbai , 2000 , tata inst .
research , narosa publ .
house , 2002 , 509533 .
popov ; e.a .
tevelev , _ self - dual projective algebraic varieties associated with symmetric spaces_. in : popov , vladimir l. ( ed . ) , algebraic groups and algebraic varieties , encyclopaedia of mathematical sciences , vol .
132 , invariant theory and algebraic transformation groups vol .
iii , springer verlag , 131167 ( 2004 ) . | let @xmath0 be a torus over an algebraically closed field @xmath1 of characteristic @xmath2 , and consider a projective @xmath0-module @xmath3 .
we determine when a projective toric subvariety @xmath4 is self - dual , in terms of the configuration of weights of @xmath5 . |
in recent years , the triplet superconductivity has become one of the modern subjects for researchers in the field of superconductivity @xcite .
particularly , the nonunitary spin triplet state in which cooper pairs may carry a finite averaged intrinsic spin momentum has attracted much attention in the last decade @xcite . a triplet state in the momentum space @xmath1
can be described by the order parameter @xmath2 in a 2@xmath32 matrix form in which @xmath4 are 2@xmath32 pauli matrices @xmath5 .
the three dimensional complex vector @xmath6 ( gap vector ) describes the triplet pairing state . in the nonunitary state ,
the product @xmath7 is not a multiple of the unit matrix .
thus in a non - unitary state the time reversal symmetry is necessarily broken spontaneously and a spontaneous moment @xmath8 appears at each point @xmath1 of the momentum space . in this case
the macroscopically averaged moment @xmath9 integrated on the fermi surface does not vanish .
the value @xmath10 is related to the net spin average by @xmath11 $ ] .
it is clear that the total spin average over the fermi surface can be nonzero . as an application ,
the nonunitary bipolar state of @xmath0wave pairing symmetry has been considered for the @xmath12phase of superconductivity in the @xmath13 compound which has been created at low temperatures @xmath14 and small values of the magnetic field @xmath15 @xcite . in the present paper , the ballistic josephson weak link via an interface between two superconducting bulks with different orientations of the crystallographic axes
is investigated .
this type of weak link structure can be used for the demonstration of the pairing symmetry in the superconducting phase @xcite .
consequently , we generalize the formalism of paper @xcite for the weak link between triplet superconducting bulks with a nonunitary order parameter . in the paper @xcite , the josephson effect in the point contact between unitary @xmath0wave triplet superconductors has been studied .
also , the effect of misorientation on the charge transport has been investigated and a spontaneous current tangential to the interface between the @xmath0wave superconductors , has been observed . in this paper the nonunitary bipolar @xmath0wave model of the order parameter
is considered .
it is shown that the current - phase diagrams are totally different from the current - phase diagrams of the junction between the unitary triplet ( axial and planar ) @xmath0wave superconductors @xcite .
roughly speaking , these different characters can be used to distinguish between nonunitary bipolar @xmath0wave superconductivity and the other types of superconductivity . in the weak - link structure between the nonunitary @xmath0wave superconductors , the spontaneous current parallel to the interface
has been observed as a fingerprint for unconventional superconductivity and spontaneous time reversal symmetry breaking .
the effect of misorientation on the spontaneous and josephson currents is investigated .
it is possible to find the value of the phase difference in which the josephson current is zero but the spontaneous current , which is produced by the interface and is tangential to the interface , is present . in some configurations and at the zero phase difference , the josephson current is not generally zero but has a finite value .
this finite value corresponds to a spontaneous phase difference which is related to the misorientation between the gap vectors@xmath16 .
the arrangement of the rest of this paper is as follows . in sec.[section2 ] we describe the configuration that we have investigated . for a non - self - consistent model of the order parameter ,
the quasiclassical eilenberger equations @xcite are solved and suitable green functions have been obtained analytically . in sec.[section3 ]
the formulas obtained for the green functions have been used for the calculation of the current densities at the interface .
an analysis of numerical results will be presented in sec.[section4 ] together with some conclusions in sec.[section5 ] .
we consider a model of a flat interface @xmath17 between two misorientated nonunitary @xmath0wave superconducting half - spaces ( fig.1 ) as a ballistic josephson junction . in the quasiclassical ballistic approach , in order to calculate the current , we use `` transport - like '' quasiclassical eilenberger equations @xcite for the energy integrated green functions @xmath18 @xmath19 = 0 , \label{eilenberger}\ ] ] and the normalization condition @xmath20 , where @xmath21 are discrete matsubara energies @xmath22 @xmath14 is the temperature , @xmath23 is the fermi velocity and @xmath24 in which @xmath25 are pauli matrices . .
] the matsubara propagator @xmath26 can be written in the form : @xmath27 where the matrix structure of the off - diagonal self energy @xmath28 in the nambu space is @xmath29 the nonunitary states , for which @xmath30 are investigated .
fundamentally , the gap vector ( order parameter ) @xmath16 has to be determined numerically from the self - consistency equation @xcite , while in the present paper , we use a non - self - consistent model for the gap vector which is much more suitable for analytical calculations @xcite .
solutions to eq .
( [ eilenberger ] ) must satisfy the conditions for the green functions and the gap vector @xmath16 in the bulks of the superconductors far from the interface as follow : @xmath31 \cdot \hat{\sigma}i\hat{\sigma}_{2 } \\
i\hat{\sigma}_{2}\left [ i\mathbf{d}_{n}^{\ast } + \mathbf{d}_{n}^{\ast } \times \mathbf{a}_{n}\right ] \cdot \hat{\sigma } & -\varepsilon \hat{\sigma}_{2}(1+\mathbf{a}_{n}\cdot \hat{\sigma})\hat{\sigma}_{2 } \end{array } \right ) \label{bulk solution}\ ] ] where @xmath32 and @xmath33}{\varepsilon _ { m}^{2}+\mathbf{d}_{n}\cdot \mathbf{d } _ { n}^{\ast } + \sqrt{(\varepsilon _ { m}^{2}+\mathbf{d}_{n}\cdot \mathbf{d } _ { n}^{\ast } ) ^{2}+(\mathbf{d}_{n}\times \mathbf{d}_{n}^{\ast } ) ^{2}}}}\ ] ] @xmath34 where @xmath35 is the external phase difference between the order parameters of the bulks and @xmath36 label the left and right half spaces respectively .
it is clear that poles of the green function in the energy space are in @xmath37 consequently , @xmath38 and @xmath39 in which @xmath40 is the energy value of the poles .
( [ eilenberger ] ) has to be supplemented by the continuity conditions at the interface between superconductors . for all quasiparticle trajectories ,
the green functions satisfy the boundary conditions both in the right and left bulks as well as at the interface .
the system of equations ( [ eilenberger ] ) and the self - consistency equation for the gap vector @xmath16 @xcite can be solved only numerically . for the junction between nonunitary bipolar @xmath0wave bulks , @xmath41 , geometry ( i ) and different misorientations .
currents are given in units of @xmath42 for unconventional superconductors such solution requires the information about the interaction between the electrons in the cooper pairs and the nature of unconventional superconductivity in novel compounds which in most cases are unknown .
also , it has been shown that the absolute value of a self - consistent order parameter is suppressed near the interface and at the distances of the order of the coherence length , while its dependence on the direction in the momentum space almost remains unaltered @xcite .
this suppression of the order parameter changes the amplitude value of the current , but does not influence the current - phase dependence drastically .
for example , it has been verified in refs.@xcite for the junction between unconventional @xmath43-wave superconductors , in ref.@xcite for the case of unitary `` @xmath44-wave '' superconductors and in ref.@xcite for pinholes in @xmath45 that there is good qualitative agreement between self - consistent and non - self - consistent results for not very large angles of misorientation .
it has also been observed that the results of the non - self - consistent model in @xcite are similar to experiment @xcite .
consequently , despite the fact that this solution can not be applied directly to a quantitative analysis of a real experiment , only a qualitative comparison of calculated and experimental current - phase relations is possible . in our calculations , a simple model of the constant order parameter up to the interface
is considered and the pair breaking and scattering on the interface are ignored .
we believe that under these strong assumptions our results describe the real situation qualitatively . in the framework of such a model
, the analytical expressions for the current can be obtained for a certain form of the order parameter .
the solution of eq . ( [ eilenberger ] ) allows us to calculate the current densities .
the expression for the current is : @xmath46 where @xmath47 stands for averaging over the directions of an electron momentum on the fermi surface @xmath48 and @xmath49 is the electron density of states at the fermi level of energy . for the junction between nonunitary bipolar @xmath0wave bulks , @xmath41 , geometry ( ii ) and different misorientations .
] we assume that the order parameter is constant in space and in each half - space it equals its value ( [ bulk order parameter ] ) far from the interface in the left or right bulks .
for such a model , the current - phase dependence of a josephson junction can be calculated analytically .
it enables us to analyze the main features of current - phase dependence for any model of the nonunitary order parameter .
the eilenberger equations ( [ eilenberger ] ) for green functions @xmath26 , which are supplemented by the condition of continuity of solutions across the interface , @xmath17 , and the boundary conditions at the bulks , are solved for a non - self - consistent model of the order parameter analytically . in the ballistic case
the system of equations for functions @xmath50 and @xmath51 can be decomposed into independent blocks of equations .
the set of equations which enables us to find the green function @xmath52 is : @xmath53 where @xmath54 the eqs .
( [ a])-([d ] ) can be solved by integrating over the ballistic trajectories of electrons in the right and left half - spaces .
the general solution satisfying the boundary conditions ( [ bulk solution ] ) at infinity is @xmath55 @xmath56 @xmath57 @xmath58 where @xmath59 is the time of flight along the trajectory , @xmath60 and @xmath61 by matching the solutions ( [ e]-[h ] ) at the interface @xmath62 , we find constants @xmath63 and @xmath64 indices @xmath36 label the left and right half - spaces respectively .
the function @xmath65 which is a diagonal term of the green matrix and determines the current density at the interface , @xmath17 , is as follows : @xmath66^{2 } } \label{charge - term}\ ] ] where @xmath67 we consider a rotation @xmath68 only in the right superconductor ( see fig.1 ) , i.e. , @xmath69 @xmath70 is the unit vector in the momentum space .
component of the current tangential to the interface versus the phase difference @xmath71 for the junction between nonunitary bipolar @xmath0wave superconducting bulks , @xmath41 , geometry ( i ) and the different misorientations . ] the crystallographic @xmath72-axis in the left half - space is selected parallel to the partition between the superconductors ( along the @xmath73-axis in fig.1 ) .
to illustrate the results obtained by computing the formula ( [ charge - term ] ) , we plot the current - phase diagrams for two different geometries . these geometries correspond to the different orientations of the crystals in the right and left sides of the interface ( fig.1):(i ) the basal @xmath74-plane in the right side has been rotated around the @xmath72-axis by @xmath75 ; @xmath76 .
( ii ) the @xmath72-axis in the right side has been rotated around the @xmath77-axis by @xmath75 ; @xmath78.further calculations require a certain model of the gap vector ( order parameter ) @xmath16 .
in the present paper , the nonunitary @xmath0wave gap vector in the @xmath12phase ( low temperature @xmath14 and low field @xmath15 ) of superconductivity in @xmath13 compound has been considered .
this nonunitary bipolar state which explains the weak spin - orbit coupling in @xmath13 is @xcite : @xmath79 the coordinate axes @xmath80 are selected along the crystallographic axes @xmath81 in the left side of fig.[fig1 ] .
the function @xmath82@xmath83 describes the dependence of the gap vector on the temperature @xmath14 ( our numerical calculations are done at the low value of temperature @xmath84 ) . using this model of the order parameter ( [ model ] ) and solution to the eilenberger equations ( [ charge - term ] ) , we have calculated the current density at the interface numerically .
these numerical results are listed below:1 ) the nonunitary property of green s matrix diagonal term consists of two parts . the explicit part which is in the @xmath85 mathematical expression in eq.([charge - term ] ) and the implicit part in the @xmath86 and @xmath87 terms .
these @xmath86 and @xmath87 terms are different from their unitary counterparts . in the mathematical expression for @xmath86 the nonunitary mathematical terms @xmath88 are presented .
the explicit part will be present only in the presence of misorientation between gap vectors,@xmath89 , but the implicit part will be present always .
so , in the absence of misorientation @xmath90 , although the implicit part of nonunitary exists the explicit part is absent . for the junction between nonunitary bipolar @xmath0wave superconducting bulks , @xmath41 , geometry ( ii ) and the different misorientations ( @xmath91component ) . ]
this means that in the absence of misorientation , current - phase diagrams for planar unitary and nonunitary bipolar systems are the same but the maximum values are slightly different.2 ) a component of current parallel to the interface @xmath92 for geometry ( i ) is zero similar to the unitary case @xcite while the other parallel component @xmath93 has a finite value ( see fig.[fig4 ] ) .
this latter case is a difference between unitary and nonunitary cases . because in the junction between unitary @xmath0wave superconducting bulks all parallel components of the current ( @xmath93 and @xmath92 ) for geometry ( i ) are absent @xcite.3 ) in figs.[fig2],[fig3 ]
, the josephson current @xmath94 is plotted for a certain nonunitary model of @xmath0wave and different geometries .
figs.[fig2],[fig3 ] are plotted for the geometry ( i ) and geometry ( ii ) respectively .
they are completely unusual and totally different from their unitary counterparts which have been obtained in @xcite.4 ) in fig.[fig2 ] for geometry ( i ) , it is observed that by increasing the misorientation , some small oscillations appear in the current - phase diagrams as a result of the non - unitary property of the order parameter , .
also , the josphson current at the zero external phase difference @xmath95 is not zero but has a finite value .
the josephson current will be zero at the some finite values of the phase difference.5 ) in fig.[fig3 ] for geometry ( ii ) , it is observed that by increasing the misorientation the new zeros in current - phase diagrams appear and the maximum value of the current will be changed non - monotonically . for the junction between nonunitary bipolar @xmath0wave superconducting bulks , @xmath41 , geometry ( ii ) and the different misorientations ( @xmath96component ) .
] in spite of the fig[fig2 ] for geometry ( i ) , the josephson currents at the phase differences @xmath95 , @xmath97 and @xmath98 are zero exactly.6 ) the current - phase diagram for geometry ( i ) and @xmath99component ( fig.[fig4 ] ) is totally unusual . by increasing the misorientation , the maximum value of the current increases .
the components of current parallel to the interface for geometry ( ii ) are plotted in fig.[fig5 ] and fig[fig6 ] .
all the terms at the phase differences @xmath100 , @xmath101 and @xmath98 are zero .
the maximum value of the current - pase diagrams is not a monotonic function of the misorientation .
thus , we have theoretically studied the supercurrents in the ballistic josephson junction in the model of an ideal transparent interface between two misoriented @xmath13 crystals with nonunitary bipolar @xmath0wave superconducting bulks which are subject to a phase difference @xmath35 .
our analysis has shown that misorientation between the gap vectors creates a current parallel to the interface and different misorientations between gap vectors influence the spontaneous parallel and normal josephson currents .
these have been shown for the currents in the point contact between two bulks of unitary axial and planar @xmath44-wave superconductor in @xcite separately .
also , it is shown that the misorientation of the superconductors leads to a spontaneous phase difference that corresponds to the zero josephson current and to the minimum of the weak link energy in the presence of the finite spontaneous current .
this phase difference depends on the misorientation angle .
the tangential spontaneous current is not generally equal to zero in the absence of the josephson current .
the difference between unitary planar and nonunitary bipolar states can be used to distinguish between them .
this experiment can be used to test the pairing symmetry and recognize the different phases of @xmath13 .
99 m. sigrist and k. ueda , _ rev .
phys . _ * 63 * , 239 ( 1991 ) .
y. maeno , h. hashimoto , k. yoshida , s. nashizaki , t. fujita , j.g .
bednorz , and f. lichenberg , _ nature _ , * 372 * , 532 ( 1994 ) .
mackenzie and y. maeno , _ rev .
phys . _ * 75 * , 657 ( 2003 ) .
h. tou , y. kitaoka , k. ishida , k. asayama , n. kimura , y. onuki , e. yamamoto , y. haga , and k. maezawa , _ phys .
lett . _ * 80 * , 3129 ( 1998 ) .
k. machida , t. nishira , and t. ohmi , _ j. phys
_ , * 68 , * 3364 ( 1999 ) . t. ohmi and k. machida , _ phys . rev .
lett . _ * 71 * , 625 ( 1993 ) .
n. stefanakis , _ phys .
b _ * 65 * , 064533 ( 2002 ) .
r. mahmoodi , s.n .
shevchenko and yu.a .
kolesnichenko , _ fiz .
_ * 28 * , 262(2002 ) [ _ sov . j. low temp . phys .
_ * 28 * , 184 ( 2002 ) ] .
i.o . kulik and a.n .
omelyanchouk , _ fiz .
temp . _ * 4 * , 296 ( 1978 ) [ _ sov . j. low temp .
_ , * 4 * , 142 ( 1978 ) ] .
amin , m.coury , s.n .
rashkeev , a.n .
omelyanchouk , and a.m. zagoskin , _ physica b , _ * 318 * , 162 ( 2002 ) .
g. eilenberger , _ z. phys . , _
* 214 * , 195 ( 1968 ) .
s. barash , a.m. bobkov , and m. fogelstrm , _ phys .
b _ * 64 * , 214503 ( 2001 )
viljas , _ cond - mat_/0004246 .
s .- k . yip ,
lett . _ * 83 * , 3864 ( 1999 ) .
s. backhaus , s. pereverzev , r.w .
simmonds , a. loshak , j.c .
davis , and r.e .
packard , _ nature _ * 392 * , 687 ( 1998 ) . | a stationary josephson effect in a weak - link between misorientated nonunitary triplet superconductors is investigated theoretically .
the non - self - consistent quasiclassical eilenberger equation for this system has been solved analytically . as an application of this analytical calculation ,
the current - phase diagrams are plotted for the junction between two nonunitary bipolar @xmath0wave superconducting banks .
a spontaneous current parallel to the interface between superconductors has been observed .
also , the effect of misorientation between crystals on the josephson and spontaneous currents is studied .
such experimental investigations of the current - phase diagrams can be used to test the pairing symmetry in the above - mentioned superconductors . |
khz quasi - periodic oscillations ( khz qpos ) observed in neutron - star low - mass x - ray binaries ( ns lmxbs ) will give an important clue to understand the innermost structure of relativistic disks of ns lmxbs , if their origin is clarified .
so far , many models of khz qpos are proposed . among them
trapped disk oscillations are one of promising candidates , and trapping of p- , g- , and c- mode oscillations in relativistic disks has been extensively studied ( see wagoner 1999 , kato 2001 , and kato et al .
2008 for review , and see fu & lai 2009 and lai & tiang 2009 for related topics ) .
in addition to the above trapped oscillations , kato ( 2010 , 2011a , b , 2012a , b ) suggested that trapped , two - armed ( @xmath0 ) , vertical p - mode oscillations in geometrically thin ( magnetized ) disks are one of possible origins of khz qpos observed in neutron - star x - ray sources . in this model two trapped oscillations , both of which are fundamental modes ( @xmath1 ) in the vertical direction but one is the fundamental ( @xmath5 ) and the other the first overtone ( @xmath6 ) in the radial direction , are regarded to be the twin khz qpos .
frequencies of these two oscillations change with correlation by time changes of disk temperature , disk vertical thickness and toroidal magnetic fields in disks .
this correlated change of the frequencies is considered to be the cause of the observed correlated change of the twin khz qpos ( kato 2011b , 2012b ) . in calculating the frequencies of
trapped , two - armed p - mode oscillations , we have assumed that the disks are isothermal in the vertical direction and have toroidal magnetic fields .
furthermore , for simplicity , the toroidal magnetic fields are assumed to be distributed in the vertical direction so that the alfv ' en speed , @xmath7 , is constant in the vertical direction ( kato 2011a , 2012a ) . even in this simplified case ,
semi - analytical calculations of frequencies of trapped vertical p - mode oscillation are still complicated .
hence , in previous studies ( kato 2011a , 2012a ) we have assumed that the oscillations are nearly vertical and the motions in the radial direction associated with the oscillations can be treated as small perturbations over the vertical motions .
the approximation is allowed in qualitative arguments , but not accurate enough to do quantitative comparison of our model with observations .
hence , in this paper , returning to a simpler case , i.e. , to the case of non - magnetized disks , we examine the trapping of two - armed vertical p - mode oscillations , since in this simpler case the frequencies of trapped oscillations can be calculated analytically without introducing the approximation that the oscillations are nearly vertical .
the results show that even in this case of non - magnetized disks , the observed correlation of twin khz qpos and the correlation between khz qpos and horizontal branch oscillation can be both described by moderate time variations of disk temperature , if the disks are strongly truncated in the vertical thickness .
distinct from previous papers by kato ( 2011a , b , 2012a , b ) , no magnetic field is assumed in this paper . except for this ,
our disk models are the same as those adopted in previous papers .
that is , the unperturbed axisymmetric disks are vertically isothermal .
thus , the vertical hydrostatic equilibrium gives that in the vertical direction the density @xmath8 is distributed as @xmath9 where the scale height @xmath10 is related to the isothermal accoustic speed @xmath11 and the vertical epicyclic frequency @xmath12 by @xmath13 here , @xmath14 is the radial coordinate of the cylindrical coordinates ( @xmath14,@xmath15,@xmath16 ) whose @xmath16-axis is perpendicular to the unperturbed disk plane and the origin is at the disk center . it is noted that the newtonian formulation is adopted in this paper , except when the radial distributions of @xmath17 , @xmath12 , and @xmath18 are considered , where @xmath18 is the radial epicyclic frequency and @xmath17 is the angular velocity of disk rotation and taken to be the keplerian angular velocity , @xmath19 , since geometrically thin disks are considered .
we assume that the isothermal disks are terminated at a certain height and surrounded by hot , low - density corona .
the height of termination , @xmath20 , is taken to be a free parameter .
hereafter , we introduce a dimensionless parameter , @xmath21 , defined by @xmath22 , which is a function of @xmath14 in general , but taken to be a constant independent of @xmath14 , since the trapped region of oscillations is not wide , except for the case of small @xmath21 , as is shown below ( see figures 2 and 3 ) .
the disk temperature is also a parameter describing disks . as a reference
, we adopt the temperature in the standard disks where the gas pressure dominates over the radiation pressure and the opacity comes from the free - free processes . in this case , if the conventional viscosity parameter @xmath23 is 0.1 and the mass accretion rate normalized by the eddington critical accretion rate is 0.3 , the square of the isothermal accoustic speed , @xmath24 , is given by ( e.g. , kato et al .
2008 ) @xmath25 where @xmath26 is the mass of the central source and @xmath27 is the schwarzschild radius defined by @xmath28 . in this paper
we adopt @xmath29 $ ] as a dimensionless parameter describing the temperature distribution in disks .
@xmath30 is , in general , a function of @xmath14 , but taken to be constant independent of @xmath14 , since the trapped region of oscillations is generally not wide , as mentioned before . in summary
, we specify our disk models by two parameters : @xmath31 and two additional parameters @xmath32 and @xmath33 , where @xmath33 is the spin parameter of the central source ( @xmath34 ) .
wave equations describing isothermal oscillations in vertically isothermal and geometrically thin disks have been well studied ( e.g. , kato 2001 for review ) .
it is convenient to introduce the variable @xmath35 defined by @xmath36 , where @xmath37 is the pressure variation over the unperturbed pressure @xmath38 and @xmath39 is the unperturbed density .
then , under the approximation that the radial wavelength of the perturbations is smaller than the scale length of the radial variations of the equilibrium properties of the disk , we have a wave equation in terms of @xmath35 as ( e.g. , kato 2001 ) @xmath40 + \frac{(\omega - m\omega)^2}{c_{\rm s}^2}h_1=0 , \label{3.1}\ ] ] where the perturbations are taken to be proportional to exp@xmath41 $ ] , @xmath42 and @xmath43 are being , respectively , the frequency and the azimuthal wavenumber of perturbations . in this paper
we consider only the case of @xmath0 , but @xmath43 is retained hereafter without specifying @xmath0 so that the origin of @xmath43 can be traced back .
let us introduce a dimensionless variable @xmath44 defined by @xmath45 .
then , equation ( [ 3.1 ] ) is written as @xmath46 \nonumber \\ & & + \frac{(\omega - m\omega)^2}{\omega_\bot^2}h_1=0 . \label{3.2}\end{aligned}\ ] ] wave equations similar to or more general than equation ( [ 3.2 ] ) have been solved by the wkb method , approximately separating @xmath47 as @xmath48 ( e.g. , silbergleit et al 2001 , ortega - rodrigues et al .
2008 , kato 2012a ) . in the present case
, @xmath35 is simply separated as @xmath49 , and we have a set of two equations ( e.g. , okazaki et al .
1987 ) : @xmath50 and @xmath51 + \biggr[1-\frac{k\omega_\bot^2}{(\omega - m\omega)^2}\biggr]f=0 , \label{3.4}\ ] ] where @xmath52 is the separation constant and can be obtained by solving equation ( [ 3.3 ] ) with relevant boundary conditions , which is given in the next subsection . in the case where the isothermal disks extend infinity in the vertical direction (
i.e. , @xmath53 ) , the boundary condition of finite wave - energy density at @xmath54 requires that @xmath52 is zero or a positive integer , say @xmath55 , and @xmath56 is a hermite polynomial , @xmath57 ( okazaki et al.1987 )
. the integer @xmath55 thus represents the number of nodes of @xmath35 in the vertical direction . in the present problem
, we are interested in vertical p - mode oscillations and thus @xmath55 starts from @xmath1 , not from @xmath58 ( @xmath58 corresponds to the horizontal p - mode oscillations and the oscillations of @xmath58 is outside of our interest in this paper . ) hereafter , we use @xmath59 to denote node number of @xmath35 in the vertical direction . in the case of vertically terminated disks ( @xmath60 ) , we impose boundary conditions that the lagrangian pressure variation vanishes at @xmath61 , i.e. , @xmath62 at @xmath61 , since in the coronae the perturbations propagate away rapidly because of high temperature . if we adopt the approximation that the oscillations are nearly vertical there , this boundary condition can be reduced to ( kato 2012a ) @xmath63 where @xmath64 is the @xmath16-component of velocity associated with the oscillations .
since the @xmath16-component of equation of motion is related to @xmath35 by @xmath65 , this boundary condition can be expressed in terms of @xmath66 as @xmath67 the eigenvalue @xmath52 obtained by solving equation ( [ 3.3 ] ) with this boundary condition is expressed hereafter by @xmath68 , since it depends upon @xmath21 and the number of node , @xmath55 .
the tilde is attached to @xmath52 in order to avoid confusion with @xmath69 in kato ( 2012a ) .
the results are summarized in table 1 for some values of @xmath55 and @xmath21 .
the eigenvalue @xmath70 is related to @xmath71 obtained by the boundary condition @xmath72 by @xmath73 . ]
@xmath69 and @xmath74 , respectively , when the equation is solved by the boundary condition @xmath72 .
then , @xmath69 and @xmath74 are related to @xmath70 and @xmath66 given in this paper by @xmath75 and @xmath76 , respectively .
this can be shown by substituting @xmath74 given above into equation ( * ) and using the fact that @xmath66 is a solution of equation ( * ) .
] .eigenvalue @xmath70 of vertical oscillations [ cols="^,^,^,^,^,^,^,^",options="header " , ] we solve equation ( [ 3.4 ] ) by the wkb method with @xmath70 given above . by introducing @xmath77 defined by , as silbergleit et al .
( 2001 ) did , @xmath78 ^ 2-\kappa^2(r'))\biggr\}dr ' , \label{3.5}\ ] ] where @xmath79 is the radius where inner boundary condition is imposed , we can reduce equation ( [ 3.4 ] ) in the form @xmath80 where @xmath81}. \label{3.7}\ ] ] as @xmath79 we take the radius of the marginally stable circular orbit and assume that @xmath35 vanishes there , i.e , @xmath82 .
, is consistent with the boundary condition adopted by kato ( 2012a ) , which is @xmath83 at @xmath84 .
this is because the @xmath16-component of equation of motion is @xmath65 and since @xmath85 when @xmath35 is separated as @xmath86 and @xmath82 . ]
the other boundary condition is that outside of the capture radius , @xmath87 , which is defined by @xmath88 , the oscillations are spatially damped .
then , as the trapping condition , we have ( see silbergleit et al .
2001 ; kato 2012a ) @xmath89^{1/2}}{c_{\rm s } } \biggr[1-\frac{{\tilde k}_{n,{\rm s}}\omega_\bot^2}{(\omega - m\omega)^2}\biggr]^{1/2}dr = \pi(n_{\rm r}+3/4 ) , \label{3.8}\ ] ] where @xmath90 is the node number of @xmath91 in the radial region of @xmath92 .
this equation gives the frequencies of trapped oscillations ( see also the appendix ) , as functions of @xmath55 , @xmath93 , @xmath30 , and @xmath21 for given @xmath26 and @xmath33 .
numerical results obtained by solving equation ( [ 3.8 ] ) are shown here mainly in the case of @xmath94 with some values of @xmath33 , @xmath30 , and @xmath21 . among many oscillation modes of @xmath95 and @xmath90 , our main interests here are on oscillations modes of @xmath96 , ( 1,1 ) , and ( 2,0 ) , since we think that the former two will correspond to the set of twin khz qpos , and the third one will to the horizontal branch oscillation ( hbo ) .
( 80mm,80mm)figure-1.eps figure 1 shows the @xmath30-dependence of frequencies of the trapped oscillations .
four oscillation modes of @xmath96 , ( 1,1 ) , ( 2,0 ) , and ( 2,1 ) are shown for two cases of @xmath53 and @xmath97 with @xmath98 .
for comparison , the results obtained by the approximation method that the horizontal motions associated with the oscillations are small perturbations over the vertical ones ( kato 2010 , 2011a ) are shown by thin curves .
it is noted that in the case of @xmath1 , the frequencies are independent of @xmath21 , since @xmath70 is 1.0 and free from @xmath21 .
frequencies of oscillations decrease in the order of @xmath96 , ( 1,1 ) , ( 2,0 ) , and ( 2,1 ) , as long as @xmath99 , and decrease as @xmath30 increases .
it is also noted that the frequencies derived by using the approximation of nearly vertical oscillations are systematically higher than those derived without the approximation , especially in the case of @xmath1 .
( 80mm,80mm)figure-2.eps figure 2 is the propagation diagram for two oscillations of @xmath96 and ( 1,1 ) in the case of @xmath100 .
disk parameters adopted in figure 2 are the same as those in figure 1 . for comparison ,
the oscillations derived by the approximation of nearly vertical oscillations are shown by thin curves . to understand the propagation diagram of figure 2 ,
let us examine propagation and evanescent regions of an oscillation with frequency @xmath101 .
since the propagation regions are specified by @xmath102 [ see equation ( [ 3.6 ] ) ] , we know that the regions are @xmath103 and @xmath104 in the case where @xmath105 .
the former is the propagation region of vertical p - mode oscillations and separated into two regions of @xmath106 and @xmath107 , and the latter is the propagation region of g - mode oscillations and @xmath108 . in the case of vertical p - mode oscillations with @xmath1 , we have @xmath109 .
hence , for oscillations of @xmath0 , @xmath110 is a propagation region , which is the region of our interest in this paper . the radius where @xmath111 is the capture radius , @xmath87 . outside of @xmath87
, there is an evanescent region which extends till the radius of @xmath112 ( inner lindblad radius ) .
the region specified by @xmath113 is again a propagation region , but the vertical p - mode oscillations which are nearly trapped in the region of @xmath110 will penetrate only weakly to this propagation region . it is noted that inside the region of @xmath113 there is the corotation point , @xmath114 , where the oscillations will be damped . in the present oscillations , however , their damping at the corotation resonance will be negligible , since wave leakage to the point will be negligible by spatial damping in the evanescent region , as mentioned above .
the trapping condition ( [ 3.8 ] ) determines the frequency and the capture radius which are shown in figure 2 by thick horizontal lines .
the thin horizontal lines show the case where the oscillations are approximated to be nearly vertical ones .
( 80mm,80mm)figure-3.eps figure 3 shows the propagation diagram of oscillations of @xmath115 and @xmath2 in the case of @xmath100 .
three cases of @xmath21 ( i.e. , @xmath53 , @xmath116 , and 1.3 ) are shown with other parameters being the same as those in figures 1 and 2 .
a purpose of showing this figure is to demonstrate that in the case of @xmath115 the curve specifying the capture radius , i.e. , the radius of @xmath117 depends on @xmath21 since @xmath70 depends on @xmath21 .
it is notices that when @xmath21 is close to 1 , the trapped region is rather wide , since @xmath70 is large and close to 4.0 .
( 80mm,80mm)figure-4.eps figure 4 shows the frequency - spin relation for four oscillation modes of @xmath96 , ( 1,1 ) , ( 2,0 ) , and ( 2,1 ) in the case of @xmath100 with @xmath94 .
two cases of @xmath53 and @xmath97 are shown .
as @xmath33 increases the frequencies of trapped oscillations increase , since the oscillations are more localized in the inner region as @xmath33 increases .
we take the standpoint that two oscillations of @xmath118 ( 1,0 ) and ( 1,1 ) with @xmath0 are , respectively , the upper and lower khz qpos . in addition
, we suppose that the ( 2,0 ) oscillation corresponds to the horizontal branch oscillation ( hbo ) .
these possibilities are examined here by comparing correlation curves derived by the model with observations .
we examine how the set of ( 1,0 ) and ( 1,1 ) oscillations move on a frequency - frequency diagram when @xmath30 is changed from 1.0 to 10.0 .
first , two cases of @xmath98 and 0.2 are shown in figure 5 with @xmath94 .
these correlation curves are superposed on the diagram showing the observed frequency - frequency plots of the twin khz qpos .
it should be noted here that the correlation curves are free from @xmath21 , since the frequencies of the @xmath1 oscillations are independent of @xmath21 .
an increase of @xmath33 shifts the correlation curve in the upper - right direction .
let us next consider effects of mass of the central source on the correlation curve .
figure 6 is for @xmath119 and three cases of @xmath98 , 0.2 , and 0.4 are shown .
this figure shows that an increase of mass shifts the correlation curve in the lower - left direction on the diagram , compared with the corresponding cases of @xmath94 .
( 80mm,80mm)figure-5.eps ( 80mm,80mm)figure-6.eps from figures 5 and 6 we see that the observed correlation between the twin khz qpos of typical z - sources can be well described by @xmath120 if the mass of the central source is slightly larger than @xmath121 .
a larger @xmath33 , say @xmath122 , is required if the mass is larger , say @xmath123 . in previous papers ( kato 2011a , b , 2012a , b ) , we have calculated frequencies of trapped vertical p - mode oscillations and correlations among them in the case where the disks have toroidal magnetic fields . in these papers
, however , we adopted the approximation that the oscillations are nearly vertical in the sense that the horizontal motions associated with the oscillations can be treated as small perturbations on the vertical motions .
this approximation is qualitatively acceptable , but not always relevant to quantitative arguments . in order to see this more in detail ,
we show here the results derived by using the approximation in the particular case of no magnetic fields , since in this case we have derived the frequencies without the approximation .
the correlation curves of the ( 1,0 ) and ( 1,1 ) oscillations derived by using the approximation are shown in figure 7 , when @xmath30 is changed from 1.0 to 10.0 in two cases of @xmath98 and 0.2 with @xmath124 .
this figure should be compared with figure 5 obtained without the approximation .
the comparison shows that in the approximate case the correlation curve shifts in the upper - right direction almost along the curve , compared with the case without the approximation . in other words , if an observed frequency correlation is described by the results obtained by using the approximation , we have larger values of @xmath30 compared with those derived without the approximation .
this can be understood from figure 1 , where the frequencies calculated by the approximation are shown to be systematically higher than those calculated without using the approximation .
comparison of figure 7 with figure 5 also shows that the approximation gives systematically a larger @xmath33 , if other parameter values are fixed to be the same in both cases .
( 80mm,80mm)figure-7.eps the frequency correlation between two oscillations of @xmath125 and @xmath126 is examined in order to to see whether the calculated correlation curve can describe the observational one between the upper khz qpo and hbo found by psaltis et al .
two cases of @xmath127 and @xmath128 are shown in figure 8 by changing @xmath30 from 1.0 to 20.0 .
the mass of the central source is @xmath94 and the spin parameter @xmath98 .
the abscissa is the frequency of the ( 1,0 ) oscillation , and the ordinate is that of the ( 2,0 ) oscillation .
the curves are superposed on the frequency - frequency diagram plotting observational data points ( figure 2.9 in a review paper by van der klis 2004 ) .
the blue curve is for @xmath127 and the red one is for @xmath128 the left - end of these curves is for @xmath129 and the right - end is for @xmath130 . for comparison ,
the correlation curve between the ( 1,0 ) and ( 2,1 ) oscillations is also shown by green curve in the case of @xmath128 ( in this case the ordinate is the frequency of ( 2,1 ) oscillation . )
( 100mm,100mm)figure-8.eps we are not interested here in the hbo harmonics and the sub hbo , since their amplitude are small and will be subsidiary phenomena . in the case of @xmath127
the correlation curve calculated is roughly parallel to the observed khz qpo - hbo correlation curve , but runs much above the observed one on the diagram . as @xmath21 decreases the curve moves downward on the diagram .
the figure shows that in the case of @xmath128 the correlation curve is close to the observed sequence of the correlation .
we should notice , however , that in the present paper we have calculated the frequency correlation of ( 1,0 ) and ( 2,0 ) oscillations only in a particular case of no magnetic fields .
if disks have strong magnetic fields , @xmath21 required to describe observations will not be so small as @xmath128 and also the maximum value of @xmath30 required will become smaller than that shown in figure 8 ( see next section ) .
trapping of two - armed ( @xmath0 ) vertical p - mode oscillations in vertically isothermal disks are examined .
the disks are assumed to be terminated at a certain height by hot , low - density corona .
the height of the termination is taken to be a parameter .
the above issue has been already examined in the previous papers ( kato 2011a , b , 2012a , b ) in the case where the disks have toroidal magnetic fields , introducing the approximation that the oscillations are nearly vertical in the sense that the horizontal motions associated with the oscillations are small perturbations over the vertical ones . in the case where disks have no magnetic fields ,
however , the trapping can be examined without the above approximation .
hence , returning to this simpler case of no magnetic fields , we re - examine in this paper the trapping of the vertical p - mode oscillations of @xmath0 .
the purposes of this paper are thus i ) to examine the case of no magnetic fields more accuratly without the approximation , and ii ) to show that even in the case of no magnetic fields the calculated correlation curve can describe the observed one , if disk parameters such as @xmath131 and @xmath132 are taken moderately .
figure 1 shows that in the case of oscillations of @xmath1 , the frequencies calculated without the above - mentioned approximation are systematically lower compared with those calculated with the approximation . due to this systematical difference , in the case where the approximation is not used , the correlation curve between the ( 1,0 ) and ( 1,1 ) oscillations shifts on the frequency frequency diagram in the left - downward direction almost along the curve derived with the approximation ( compare figures 5 and 7 ) .
we suppose that the same situation occurs even in magnetized disks .
that is , if we could calculate the correlation curves without the approximation in the case of magnetized disks , they will shift in the left - downward direction along the curve derived in paper i. this means that the same correlation range on the diagram can be derived by smaller values of @xmath30 and @xmath133 , if frequencies coud be calculated without the approximation .
in other words , such strong magnetic fields required in paper i may be unnecessary , if we could derive and solve wave equations in magnetized disks without the approximation .
this is in favor of our model , since the magnetic fields required in paper i seem to be somewhat stronger than those usually supposed in accretion disks .
next , we examine whether the ( 2,0 ) oscillation can be regarded as the horizontal branch oscillation ( hbo ) .
figure 8 shows that in non - magnetized disks the frequency correlation between the ( 1,0 ) and ( 2,0 ) oscillations is close to the observed correlation between the upper khz qpo and hbo , if @xmath21 is taken to be as small as @xmath128 .
this suggests that the disks of horizontal branch phase of z - sources are strongly diminished in their vertical thickness . concerning the disks of the horizontal branch , we considered in paper i two possibilities : the sequence of horizontal branch is i ) a sequence of change of toroidal magnetic fields ( figure 9 of paper i ) or ii ) a sequence of change of disk temperature with moderate or strong toroidal magnetic fields ( figure 10 of paper i ) .
the first possibility seems to be less promising , since the correlation curve calculated in paper i has a sharper gradient on the frequency - frequency diagram , compared with the curve of observations ( see figure 9 of paper i ) .
this sharp gradient comes from the fact that as the toroidal magnetic fields become strong , the frequency of the ( 2,0 ) oscillation decreases strongly when @xmath115 .
\omega_\bot^2 $ ] [ see equation ( 47 ) of kato ( 2012a ) or equation ( 27 ) of kato ( 2011a ) ] .
this means that the frequency @xmath42 given by @xmath134^{1/2 } \omega_\bot$ ] decreases , as @xmath135 increases from 1 to 2 by increase of @xmath136 when @xmath115 ( @xmath137 when @xmath115 ) . ]
hence , the correlation curve calculated by changing the strength of toroidal magnetic fields has a sharp gradient on the frequency - frequency diagram .
this situation will not be much changed , even if we could calculate the frequency of trapped ( 2,0 ) oscillations in magnetized disks without using the approximation of nearly vertical oscillation .
compared with this , the correlation curve obtained by changing the disk temperature can better describe the observational correlation curve . in the limit of no magnetic fields
the calculated curve is roughly parallel to the observed correlation curve on the frequency frequency diagram and moves downward on the diagram as @xmath21 decreases ( see figure 8) .
if toroidal magnetic fields are present ( @xmath136 being taken to be constant during temperature change ) , the correlation curve calculated will move downward furthermore compared with the case of no magnetic field , since an increase of @xmath136 will have a similar effect to a decrease of @xmath21 [ see equation @xmath134^{1/2 } \omega_\bot$ ] ] .
increases @xmath69 , while an increase of @xmath136 increases @xmath135 .
hence , both have similar effects on frequency @xmath42 . ]
this suggests that the observed correlation curve will be described by @xmath21 larger than @xmath128 , when there are toroidal magnetic fields of moderate strength with @xmath138 const .
the correlation curve shown in figure 10 of paper i will be also consistent with this argument , since the correlation curve in the figure will shift in the right - downward direction if we could calculate the correlation curve without using the approximation of nearly vertical oscillations .
the above argument suggests that the horizontal branch of z - sources is a sequence of temperature change in disks whose vertical thickness is strongly diminished .
the direction of temperature increase is from right to left on the horizontal branch , since the direction of decrease of observed qpo frequencies is from right to left on the branch .
the temperature change will be a result of a change of mass accretion rate .
thus , this disk model of the z - branch is similar to and support the disk model proposed by church et al .
( 2006 ) and bolucinska - church et al .
( 2010 ) from spectral analyses , although the origins of qpos are different .
one of possible causes why the disks of horizontal branch are vertically thin will be strong radiation from the central source resulting from high mass accretion and disk evaporation toward the corona , as church et al .
( 2006 ) and bolucinska - church et al .
( 2010 ) argue . if moderately strong toroidal magnetic fields are present in disks
, it will be in favor of our model of qpos , although only the limiting case of no magnetic fields is examined in this paper .
it is noted here that disks with strong magnetic fields ( the plasma @xmath30 being less than unity and as small as 0.1 ) are expected at a stage of transition from optically thin flows ( adafs ) to optically thick disks by increase of mass accretion rate ( machida et al .
2006 , oda et al .
2007 , 2009 , 2010 ) , although the horizontal branch of z - sources is different from such stage .
it should be emphasized that in our model of khz qpos the correlation curve between the ( 1,0 ) and ( 1,1 ) oscillations moves almost along a common curve on the frequency - frequency diagram for changes of disk parameters , as far as the mass and spin of the central source are fixed . on the other hand , the correlation curve between the ( 1,0 ) and ( 2,0 ) oscillations are rather sensitive to the disk parameters such as @xmath21 .
the former is in favor of our model , since the common curve mentioned above can well describe the observed correlation curve for reasonable mass and spin parameter of the central source .
however , the latter has two sides for and against our model .
if our model is correct , it is a good tool to determine the disk structure of horizontal branch of z - sources .
it is , however , unclear why disk thickness ( and magnetic fields ) in real disks is really adjusted to values relevant to describe observations . in our model the radius of the inner edge of disks
is assumed to be unchanged during the disk evolution along the sequence .
it is unclear at the present stage of our model whether time change of the inner edge of disks is required to describe the observed correlation curve by our model .
the observed amplitude of hbos is much larger than that of khz qpos .
whether this can be described by our model is a subject to be discussed further .
one of possibilities will be the difference of width of the trapped regions . in our model
, the trapped region of the ( 1,0 ) oscillation is only in the innermost region , but that of the ( 2,0 ) oscillation is wide when @xmath21 is small ( see figure 3 ) .
this difference will be one of causes of amplitude difference between khz qpos and hbos .
it is noted that the trapped region of the ( 1,1 ) oscillation is also wide ( see figure 2 ) , but in this case the oscillation has one wavelength in the trapped region .
hence , by a cancellation by phase mixing the observed amplitude will not become large .
it is also noted that a wide trapped region implies that our basic approximation that @xmath139 and @xmath30 are constant in the trapped region will not always be relevant .
radial variations of @xmath139 and @xmath30 should be taken into account to do more detailed comparison with observations . in a previous paper ( kato 2012a )
, we have examined the effects of toroidal magnetic fields on frequencies of trapped oscillations in truncated disks . in the study
we have introduced an approximation that the vertical p - mode oscillations consist of nearly vertical motions and thus the horizontal motions associated with them can be taken to be small perturbations over the vertical ones . under this approximation ,
in addition to the wkb approximation , we obtained an equation describing the frequencies of trapped oscillations [ see equations ( 39 ) , ( 41 ) , and ( 44 ) in kato ( 2012a ) ] @xmath140^{1/2}}{\omega_\bot h } \frac{\epsilon^{1/2}}{a_{n,{\rm s}}^{1/2}}dr = \pi \biggr(n_{\rm r}+\frac{3}{4}\biggr ) , \label{a1}\ ] ] where @xmath141 is a dimensionless quantity given by equation ( 37 ) in kato ( 2012a ) .
furthermore , @xmath142 is a dimensionless quantity defined by [ see equations ( 12 ) and ( 17 ) in kato ( 2012a ) ] @xmath143 and is assumed to be @xmath144 ( kato 2012a ) . since in the case of no magnetic fields ( @xmath145 )
, @xmath141 is reduced to @xmath146 equation ( [ a1 ] ) can be written as @xmath147^{1/2}}{c_{\rm s } } \biggr[\frac{(\omega - m\omega)^2-{\tilde k}_{n,{\rm s}}\omega_\bot^2 } { { \tilde k}_{n,{\rm s}}\omega_\bot^2}\biggr]^{1/2}dr = \pi\biggr(n_{\rm r}+\frac{3}{4}\biggr ) .
\label{a4}\ ] ] this is the equation corresponding to equation ( [ 3.8 ] ) in this paper , which is @xmath147^{1/2}}{c_{\rm s } } \biggr[\frac{(\omega - m\omega)^2-{\tilde k}_{n,{\rm s}}\omega_\bot^2 } { ( \omega - m\omega)^2}\biggr]^{1/2}dr = \pi\biggr(n_{\rm r}+\frac{3}{4}\biggr )
. \label{a5}\ ] ] this means that the perturbation method adopted by kato ( 2012a ) is valid only when in the propagation region of oscillations we have @xmath148 or @xmath149 this is nothing but the condition of @xmath144 , as is expected . in the case of two armed oscillations , @xmath0 , with @xmath1
, we have @xmath109 .
hence , this condition is realized when @xmath42 is comparable with @xmath150 . in oscillations with @xmath115 , @xmath151 in disks with @xmath53 and @xmath152 in the disks terminated in the vertical thickness ( see table 1 ) .
hence , this approximation is valid when @xmath42 is moderately smaller than @xmath150 as is shown in figure 1 . | trapping of two - armed ( @xmath0 ) vertical p - mode oscillations in relativistic disks are examined .
the disks are assumed to be isothermal in the vertical direction , but are truncated at a certain height by the presence of corona .
the same issues have been examined in a previous paper ( kato 2012a ) . in this paper , unlike the previous paper , however , we do not use the approximation that the oscillations are nearly vertical , but limit to a simpler case of no magnetic field . as in the previous paper ,
the results suggest that the two basic oscillation modes [ both are the fundamental ( @xmath1 ) in the vertical direction but in the horizontal direction one is the fundamental ( @xmath2 ) and the other the first overtone ( @xmath3 ) ] correspond to the twin khz qpos .
second , the oscillation mode which is the first overtone @xmath4 in the vertical direction and the fundamental in the horizontal direction ( @xmath2 ) will correspond to the horizontal branch oscillation ( hbo ) of z - sources .
the latter suggests that the horizontal branch of z - sources is a sequence of temperature change in disks whose vertical thickness is strongly terminated
. the temperature increases leftward along the sequence from the apex between normal and horizotal branches . |
a large amount of research has been devoted to the task of defining and identifying communities in social and information networks , _
i.e. _ , in graphs in which the nodes represent underlying social entities and the edges represent interactions between pairs of nodes .
most recent papers on the subject of community detection in large networks begin by noting that it is a matter of common experience that communities exist in such networks .
these papers then note that , although there is no agreed - upon definition for a community , a community should be thought of as a set of nodes that has more and/or better connections between its members than between its members and the remainder of the network .
these papers then apply a range of algorithmic techniques and intuitions to extract subsets of nodes and then interpret these subsets as meaningful communities corresponding to some underlying `` true '' real - world communities . in this paper , we explore from a novel perspective several questions related to identifying meaningful communities in large sparse networks , and we come to several striking conclusions that have implications for community detection and graph partitioning in such networks .
we emphasize that , in contrast to most of the previous work on this subject , we look at very large networks of up to millions of nodes , and we observe very different phenomena than is seen in small commonly - analyzed networks . at the risk of oversimplifying the large and often intricate body of work on community detection in complex networks ,
the following five - part story describes the general methodology : 1 .
data are modeled by an `` interaction graph . '' in particular , part of the world gets mapped to a graph in which nodes represent entities and edges represent some type of interaction between pairs of those entities .
for example , in a social network , nodes may represent individual people and edges may represent friendships , interactions or communication between pairs of those people .
the hypothesis is made that the world contains groups of entities that interact more strongly amongst themselves than with the outside world , and hence the interaction graph should contain sets of nodes , _
i.e. _ , communities , that have more and/or better - connected `` internal edges '' connecting members of the set than `` cut edges '' connecting the set to the rest of the world .
3 . a objective function or metric is chosen to formalize this idea of groups with more intra - group than inter - group connectivity . 4 . an algorithm is then selected to find sets of nodes that exactly or approximately optimize this or some other related metric .
sets of nodes that the algorithm finds are then called `` clusters , '' `` communities , '' `` groups , '' `` classes , '' or `` modules '' . 5 .
the clusters or communities or modules are evaluated in some way .
for example , one may map the sets of nodes back to the real world to see whether they appear to make intuitive sense as a plausible `` real '' community .
alternatively , one may attempt to acquire some form of `` ground truth , '' in which case the set of nodes output by the algorithm may be compared with it . with respect to points ( 1)(4 ) , we follow the usual path . in particular , we adopt points ( 1 ) and ( 2 ) , and we then explore the consequence of making such a choice , _
i.e. _ , of making such an hypothesis and modeling assumption . for point
( 3 ) , we choose a natural and widely - adopted notion of community goodness ( community quality score ) called _ conductance _ , which is also known as the normalized cut metric @xcite .
informally , the conductance of a set of nodes ( defined and discussed in more detail in section [ sxn : related : lowcondalgs ] ) is the ratio of the number of `` cut '' edges between that set and its complement divided by the number of `` internal '' edges inside that set .
thus , to be a good community , a set of nodes should have small conductance , _
i.e. _ , it should have many internal edges and few edges pointing to the rest of the network .
conductance is widely used to capture the intuition of a good community ; it is a fundamental combinatorial quantity ; and it has a very natural interpretation in terms of random walks on the interaction graph .
moreover , since there exist a rich suite of both theoretical and practical algorithms @xcite , we can for point ( 4 ) compare and contrast several methods to approximately optimize it .
to illustrate conductance , note that of the three @xmath2-node sets @xmath3 , @xmath4 , and @xmath5 illustrated in the graph in figure [ fig : conduct2 ] , @xmath4 has the best ( the lowest ) conductance and is thus the most community - like .
-nodes sets that have been marked , @xmath4 has the best ( _ i.e. _ , the lowest ) conductance , as it has the lowest ratio between the number of edges cut and the number of edges inside .
so , set @xmath4 is the best @xmath2-node community or the most community - like set of @xmath2 nodes in this particular network.,scaledwidth=50.0%,height=226 ] however , it is in point ( 5 ) that we deviate from previous work . instead of focusing on individual groups of nodes and trying to interpret them as `` real '' communities , we investigate statistical properties of a large number of communities over a wide range of size scales in over @xmath0 large sparse real - world social and information networks .
we take a step back and ask questions such as : how well do real graphs split into communities ? what is a good way to measure and characterize presence or absence of community structure in networks ?
what are typical community sizes and typical community qualities ? to address these and related questions
, we introduce the concept of a _ network community profile ( ncp ) plot _ that we define and describe in more detail in section [ sxn : ncpp : def ] .
intuitively , the network community profile plot measures the score of `` best '' community as a function of community size in a network .
formally , we define it as the conductance value of the minimum conductance set of cardinality @xmath6 in the network , as a function of @xmath6 .
as defined , the ncp plot will be np - hard to compute exactly , so operationally we will use several natural approximation algorithms for solving the minimum conductance cut problem in order to compute different approximations to it . by comparing and contrasting these plots for a large number of networks , and by computing other related structural properties ,
we obtain results that suggest a significantly more refined picture of the community structure in large real - world networks than has been appreciated previously .
we have gone to a great deal of effort to be confident that we are computing quantities fundamental to the networks we are considering , rather than artifacts of the approximation algorithms we employ . in particular : * we use several classes of graph partitioning algorithms to probe the networks for sets of nodes that could plausibly be interpreted as communities .
these algorithms , including flow - based methods , spectral methods , and hierarchical methods , have complementary strengths and weaknesses that are well understood both in theory and in practice .
for example , flow - based methods are known to have difficulties with expanders @xcite , and flow - based post - processing of other methods are known in practice to yield cuts with extremely good conductance values @xcite . on the other hand ,
spectral methods are known to have difficulties when they confuse long paths with deep cuts @xcite , a consequence of which is that they may be viewed as computing a `` regularized '' approximation to the network community profile plot .
( see section [ algo - notes - section ] for a more detailed discussion of these and related issues . ) * we compute spectral - based lower bounds and also semidefinite - programming - based lower bounds for the conductance of our network datasets .
* we compute a wide range of other structural properties of the networks , _ e.g. _ , sizes , degree distributions , maximum and average diameters of the purported communities , internal versus external conductance values of the purported communities , etc .
* we recompute statistics on versions of the networks that have been modified in well - understood ways , _ e.g. _ , by removing small barely - connected sets of nodes or by randomizing the edges .
* we compare our results across not only over @xmath0 large social and information networks , but also numerous commonly - studied small social networks , expanders , and low - dimensional manifold - like objects , and we compare our results on each network with what is known from the field from which the network is drawn .
to our knowledge , this makes ours the most extensive such analysis of the community structure in large real - world social and information networks .
* we compare results with analytical and/or simulational results on a wide range of commonly and not - so - commonly used network generation models @xcite . * main empirical findings : * taken as a whole ,
the results we present in this paper suggest a rather detailed and somewhat counterintuitive picture of the community structure in large social and information networks .
several qualitative properties of community structure , as revealed by the network community profile plot , are nearly universal : * up to a size scale , which empirically is roughly @xmath0 nodes , there not only exist cuts with relatively good conductance , _ i.e. _ , good communities , but also the slope of the network community profile plot is generally sloping downward .
this latter point suggests that smaller communities can be combined into meaningful larger communities , a phenomenon that we empirically observe in many cases .
* at the size scale of roughly @xmath0 nodes , we often observe the global minimum of the network community profile plot ; these are the `` best '' communities , according to the conductance measure , in the entire graph .
these are , however , rather interestingly connected to the rest of the network ; for example , in most cases , we observe empirically that they are a small set of nodes barely connected to the remainder of the network by just a _ single _ edge .
* above the size scale of roughly @xmath0 nodes , the network community profile plot gradually increases , and thus there is a nearly inverse relationship between community size and community quality . as a function of increasing size ,
the best possible communities become more and more `` blended into '' the remainder of the network .
intuitively , communities blend in with one another and gradually disappear as they grow larger .
in particular , in many cases , larger communities can be broken into smaller and smaller pieces , often recursively , each of which is more community - like than the original supposed community . *
even up to the largest size scales , we observe significantly more structure than would be seen , for example , in an expander - like random graph on the same degree sequence .
a schematic picture of a typical network community profile plot is illustrated in figure [ fig : intro_ncpp ] . in
red ( labeled as `` original network '' ) , we plot community size vs. community quality score for the sets of nodes extracted from the original network . in black
( rewired network ) , we plot the scores of communities extracted from a random network conditioned on the same degree distribution as the original network .
this illustrates not only tight communities at very small scales , but also that at larger and larger size scales ( the precise cutoff point for which is difficult to specify precisely ) the best possible communities gradually `` blend in '' more and more with the rest of the network and thus gradually become less and less community - like .
eventually , even the existence of large well - defined communities is quite questionable if one models the world with an interaction graph , as in point ( 1 ) above , and if one also defines good communities as densely linked clusters that are weakly - connected to the outside , as in hypothesis ( 2 ) above . finally , in blue ( bag of whiskers )
, we also plot the scores of communities that are composed of disconnected pieces ( found according to a procedure we describe in section [ sxn : obs_struct ] ) .
this blue curve shows , perhaps somewhat surprisingly , that one can often obtain better community quality scores by combining unrelated disconnected pieces . to understand the properties of generative models sufficient to reproduce the phenomena we have observed , we have examined in detail the structure of our social and information networks .
although nearly every network is an exception to any simple rule , we have observed that an `` octopus '' or `` jellyfish '' model @xcite provides a rough first approximation to structure of many of the networks we have examined .
that is , most networks may be viewed as having a `` core , '' with no obvious underlying geometry and which contains a constant fraction of the nodes , and then there is a periphery consisting of a large number of relatively small `` whiskers '' that are only tenuously connected to the core .
figure [ fig : intro_graph ] presents a caricature of this network structure . of course
, our network datasets are far from random in numerous ways_e.g .
_ , they have higher edge density in the core ; the small barely - connected whisker - like pieces are generally larger , denser , and more common than in corresponding random graphs ; they have higher local clustering coefficients ; and this local clustering information gets propagated globally into larger clusters or communities in a subtle and location - specific manner .
more interestingly , as shown in figure [ fig : phicore ] in section [ sxn : obs_struct : remove_whiskers ] , the core itself consists of a nested core - periphery structure .
* main modeling results : * the behavior that we observe is not reproduced , at even a qualitative level , by any of the commonly - used network generation models we have examined , including but not limited to preferential attachment models , copying models , small - world models , and hierarchical network models . moreover , this behavior is qualitatively different than what is observed in networks with an underlying mesh - like or manifold - like geometry ( which may not be surprising , but is significant insofar as these structures are often used as a scaffolding upon which to build other models ) , in networks that are good expanders ( which may be surprising , since it is often observed that large social networks are expander - like ) , and in small social networks such as those used as testbeds for community detection algorithms ( which may have implications for the applicability of these methods to detect large community - like structures in these networks ) . for the commonly - used network generation models , as well as for expander - like , low - dimensional , and small social networks , the network community profile plots are generally downward sloping or relatively flat .
although it is well understood at a qualitative level that nodes that are `` far apart '' or `` less alike '' ( in some sense ) should be less likely to be connected in a generative model , understanding this point quantitatively so as to reproduce the empirically - observed relationship between small - scale and large - scale community structure turns out to be rather subtle .
we can make the following observations : * very sparse random graph models with no underlying geometry have relatively deep cuts at small size scales , the best cuts at large size scales are very shallow , and there is a relatively abrupt transition in between .
( this is shown pictorially in figure [ fig : intro_ncpp ] for a randomly rewired version of the original network . )
this is a consequence of the extreme sparsity of the data : sufficiently dense random graphs do not have these small deep cuts ; and the relatively deep cuts in sparse graphs are due to small tree - like pieces that are connected by a single edge to a core which is an extremely good expander . *
a forest fire generative model @xcite , in which edges are added in a manner that imitates a fire - spreading process , reproduces not only the deep cuts at small size scales and the absence of deep cuts at large size scales but other properties as well : the small barely connected pieces are significantly larger and denser than random ; and for appropriate parameter settings the network community profile plot increases relatively gradually as the size of the communities increases . * the details of the `` forest fire '' burning mechanism are crucial for reproducing how local clustering information gets propagated to larger size scales in the network , and those details shed light on the failures of commonly - used network generation models . in the forest fire model ,
a new node selects a `` seed '' node and links to it . then with some probability it `` burns '' or adds an edge to the each of the seed s neighbors , and so on , recursively .
although there is a `` preferential attachment '' and also a `` copying '' flavor to this mechanism , two factors are particularly important : first is the local ( in a graph sense , as there is no underlying geometry in the model ) manner in which the edges are added ; and second is that the number of edges that a new node can add can vary widely , depending on the local structure around the seed node .
depending on the neighborhood structure around the seed , small fires will keep the community well - separated from the network , but occasional large fires will connect the community to the rest of the network and make it blend into the network core .
thus , intuitively , the structure of the whiskers ( components connected to the rest of the graph via a single edge ) are responsible for the downward part of the network community profile plot , while the core of the network and the manner in which the whiskers root themselves to the core helps to determine the upward part of the network community profile plot . due to local clustering effects , whiskers in real networks
are larger and give deeper cuts than whiskers in corresponding randomized graphs , fluctuations in the core are larger and deeper than in corresponding randomized graphs , and thus the network community profile plot increases more gradually and levels off to a conductance value well below the value for a corresponding rewired network . * main methodological contributions :
* to obtain these and other conclusions , we have employed approximation algorithms for graph partitioning to investigate structural properties of our network datasets . briefly
, we have done the following : * we have used metis+mqi , which consists of using the popular graph partitioning package metis @xcite followed by a flow - based mqi post - processing @xcite . with this procedure , we obtain sets of nodes that have very good conductance scores . at very small size scales
, these sets of nodes could plausibly be interpreted as good communities , but at larger size scales , we often obtain tenuously - connected ( and in some cases unions of disconnected ) pieces , which perhaps do not correspond to intuitive communities . *
thus , we have also used the local spectral method of anderson , chung , and lang @xcite to obtain sets of nodes with good conductance value that that are `` compact '' or more `` regularized '' than those pieces returned by metis+mqi .
since spectral methods confuse long paths with deep cuts @xcite , empirically we obtain sets of nodes that have worse conductance scores than sets returned by metis+mqi , but which are `` tighter '' and more `` community - like . '' for example , at small size scales the sets of nodes returned by the local spectral algorithm agrees with the output of metis+mqi , but at larger scales this algorithm returns sets of nodes with substantially smaller diameter and average diameter , which seem plausibly more community - like . we have also used what we call the bag - of - whiskers heuristic to identify small barely connected sets of nodes that exert a surprisingly large influence on the network community profile plot . both metis+mqi and the local spectral algorithm scale well and thus either may be used to obtain sets of nodes from very large graphs . for many of the small to medium - sized networks
, we have checked our results by applying one or more other spectral , flow - based , or heuristic algorithms , although these do not scale as well to very large graphs . finally , for some of our smaller network datasets
, we have computed spectral - based and semidefinite - programming - based lower bounds , and the results are consistent with the conclusions we have drawn
. * broader implications : * our observation that , independently of the network size , compact communities exist only up to a size scale of around @xmath0 nodes agrees well with the `` dunbar number '' @xcite , which predicts that roughly @xmath7 individuals is the upper limit on the size of a well - functioning human community .
moreover , we should emphasize that our results do not disagree with the literature at small sizes scales .
one reason for the difference in our findings is that previous studies mainly focused on small networks , which are simply not large enough for the clusters to gradually blend into one another as one looks at larger size scales . in order to make our observations
, one needs to look at large number ( due to the complex noise properties of real graphs ) of large networks .
it is only when dunbar s limit is exceeded by several orders of magnitude that it is relatively easy to observe large communities blurring together and eventually vanishing .
a second reason for the difference is that previous work did not measure and examine the _ network community profile _ of cluster size vs. cluster quality .
finally , we should note that our explanation also aligns well with the _ common bond _ vs. _ common identity _ theory of group attachment @xcite from social psychology , where it has been noted that bond communities tend to be smaller and more cohesive @xcite , as they are based on interpersonal ties , while identity communities are focused around common theme or interest .
we discuss these implications and connections further in section [ sxn : discussion ] .
the rest of the paper is organized as follows . in section
[ sxn : related ] we describe some useful background , including a brief description of the network datasets we have analyzed .
then , in section [ sxn : ncpp ] we present our main results on the properties of the network community profile plot for our network datasets .
we place an emphasis on how the phenomena we observe in large social and information networks are qualitatively different than what one would expect based on intuition from and experience with expander - like graphs , low - dimensional networks , and commonly - studied small social networks .
then , in sections [ sxn : obs_struct ] and [ algo - notes - section ] , we summarize the results of additional empirical evaluations
. in particular , in section [ sxn : obs_struct ] , we describe some of the observations we have made in an effort to understand what structural properties of these large networks are responsible for the phenomena we observe ; and in section [ algo - notes - section ] , we describe some of the results of probing the networks with different approximation algorithms in an effort to be confident that the phenomena we observed really are properties of the networks we study , rather than artifactual properties of the algorithms we chose to use to study those networks . we follow this in section [ sxn : models ] with a discussion of complex network generation models .
we observe that the commonly - used network generation models fail to reproduce the counterintuitive phenomena we observe .
we also notice that very sparse random networks reproduce certain aspects of the phenomena , and that a generative model based on an iterative `` forest fire '' burning mechanism reproduces very well the qualitative properties of the phenomena we observe .
finally , in section [ sxn : discussion ] we provide a discussion of our results in a broader context , and in section [ sxn : conclusion ] we present a brief conclusion .
in this section , we will provide background on our data and methods . we start in section [ sxn : related : networkdata ] with a description of the network datasets we will analyze .
then , in section [ sxn : related : clusters ] , we review related community detection and graph clustering ideas . finally , in section [ sxn : related : lowcondalgs ] , we provide a brief description of approximation algorithms that we will use .
there exist a large number of reviews on topics related to those discussed in this paper .
for example , see the reviews on community identification @xcite , data clustering @xcite , graph and spectral clustering @xcite , graph and heavy - tailed data analysis @xcite , surveys on various aspects of complex networks @xcite , the monographs on spectral graph theory and complex networks @xcite , and the book on social network analysis @xcite . see section [ sxn : discussion ] for a more detailed discussion of the relationship of our work with some of this prior work .
we have examined a large number of real - world complex networks .
see tables [ tab : data_statsdesc_1 ] , [ tab : data_statsdesc_2 ] , and [ tab : data_statsdesc_3 ] for a summary . for convenience
, we have organized the networks into the following categories : social networks ; information / citation networks ; collaboration networks ; web graphs ; internet networks ; bipartite affiliation networks ; biological networks ; low - dimensional networks ; imdb networks ; and amazon networks .
we have also examined numerous small social networks that have been used as a testbed for community detection algorithms ( _ e.g. _ , zachary s karate club @xcite , interactions between dolphins @xcite , interactions between monks @xcite , newman s network science network @xcite , etc . ) , numerous simple network models in which by design there is an underlying geometry ( _ e.g. _ , power grid and road networks @xcite , simple meshes , low - dimensional manifolds including graphs corresponding to the well - studied `` swiss roll '' data set @xcite , a geometric preferential attachment model @xcite , etc . ) , several networks that are very good expanders , and many simulated networks generated by commonly - used network generation models(_e.g .
_ , preferential attachment models @xcite , copying models @xcite , hierarchical models @xcite , etc . ) .
network & @xmath8 & @xmath9 & @xmath10 & @xmath11 & @xmath12 & @xmath13 & @xmath14 & @xmath15 & @xmath16 & description + + delicious & 147,567 & 301,921 & 0.40 & 0.65 & 4.09 & 48.44 & 0.30 & 24 & 6.28 & del.icio.us collaborative tagging social network + epinions & 75,877 & 405,739 & 0.48 & 0.90 & 10.69 & 183.88 & 0.26 & 15 & 4.27 & who - trusts - whom network from epinions.com @xcite + flickr & 404,733 & 2,110,078 & 0.33 & 0.86 & 10.43 & 442.75 & 0.40 & 18 & 5.42 & flickr photo sharing social network @xcite + linkedin & 6,946,668 & 30,507,070 & 0.47 & 0.88 & 8.78 & 351.66 & 0.23 & 23 & 5.43 & social network of professional contacts + livejournal01 & 3,766,521 & 30,629,297 & 0.78 & 0.97 & 16.26 & 111.24 & 0.36 & 23 & 5.55 & friendship network of a blogging community @xcite + livejournal11 & 4,145,160 & 34,469,135 & 0.77 & 0.97 & 16.63 & 122.44 & 0.36 & 23 & 5.61 & friendship network of a blogging community @xcite + livejournal12 & 4,843,953 & 42,845,684 & 0.76 & 0.97 & 17.69 & 170.66 & 0.35 & 20 & 5.53 & friendship network of a blogging community @xcite + messenger & 1,878,736 & 4,079,161 & 0.53 & 0.78 & 4.34 & 15.40 & 0.09 & 26 & 7.42 & instant messenger social network + email - all & 234,352 & 383,111 & 0.18 & 0.50 & 3.27 & 576.87 & 0.50 & 14 & 4.07 & research organization email network ( all addresses ) @xcite + email - inout & 37,803 & 114,199 & 0.47 & 0.82 & 6.04 & 165.73 & 0.58 & 8 & 3.74 & ( all addresses but email has to be sent both ways ) @xcite + email - inside & 986 & 16,064 & 0.90 & 0.99 & 32.58 & 74.66 & 0.45 & 7 & 2.60 & ( only emails inside the research organization ) @xcite + email - enron & 33,696 & 180,811 & 0.61 & 0.90 & 10.73 & 142.36 & 0.71 & 13 & 3.99 & enron email dataset @xcite + answers & 488,484 & 1,240,189 & 0.45 & 0.78 & 5.08 & 251.78 & 0.11 & 22 & 5.72 & yahoo answers social network + answers-1 & 26,971 & 91,812 & 0.56 & 0.87 & 6.81 & 59.17 & 0.08 & 16 & 4.49 & cluster 1 from yahoo answers + answers-2 & 25,431 & 65,551 & 0.48 & 0.80 & 5.16 & 56.57 & 0.10 & 15 & 4.76 & cluster 2 from yahoo answers + answers-3 & 45,122 & 165,648 & 0.53 & 0.87 & 7.34 & 417.83 & 0.21 & 15 & 3.94 & cluster 3 from yahoo answers + answers-4 & 93,971 & 266,199 & 0.49 & 0.82 & 5.67 & 94.48 & 0.08 & 16 & 4.91 & cluster 4 from yahoo answers + answers-5 & 5,313 & 11,528 & 0.41 & 0.73 & 4.34 & 29.55 & 0.12 & 14 & 4.75 & cluster 5 from yahoo answers + answers-6 & 290,351 & 613,237 & 0.40 & 0.71 & 4.22 & 57.16 & 0.09 & 22 & 5.92 & cluster 6 from yahoo answers + + cit - patents & 3,764,105 & 16,511,682 & 0.82 & 0.96 & 8.77 & 21.34 & 0.09 & 26 & 8.15 & citation network of all us patents @xcite + cit - hep - ph & 34,401 & 420,784 & 0.96 & 1.00 & 24.46 & 63.50 & 0.30 & 14 & 4.33 & citations between physics ( arxiv hep - th ) papers @xcite + cit - hep - th & 27,400 & 352,021 & 0.94 & 0.99 & 25.69 & 106.40 & 0.33 & 15 & 4.20 & citations between physics ( arxiv hep - ph ) papers @xcite + blog - nat05 - 6 m & 29,150 & 182,212 & 0.74 & 0.96 & 12.50 & 342.51 & 0.24 & 10 & 3.40 & blog citation network ( 6 months of data ) @xcite + blog - nat06all & 32,384 & 315,713 & 0.87 & 0.99 & 19.50 & 153.08 & 0.20 & 18 & 3.94 & blog citation network ( 1 year of data ) @xcite + post - nat05 - 6 m & 238,305 & 297,338 & 0.21 & 0.34 & 2.50 & 39.51 & 0.13 & 45 & 10.34 & blog post citation network ( 6 months ) @xcite + post - nat06all & 437,305 & 565,072 & 0.22 & 0.38 & 2.58 & 35.54 & 0.11 & 54 & 10.48 & blog post citation network ( 1 year ) @xcite + + ata - imdb & 883,963 & 27,473,042 & 0.87 & 0.99 & 62.16 & 517.40 & 0.79 & 15 & 3.48 & imdb actor collaboration network from dec 2007 + ca - astro - ph & 17,903 & 196,972 & 0.89 & 0.98 & 22.00 & 65.70 & 0.67 & 14 & 4.21 & co - authorship in astro - ph of arxiv.org @xcite + ca - cond - mat & 21,363 & 91,286 & 0.81 & 0.93 & 8.55 & 22.47 & 0.70 & 15 & 5.36 & co - authorship in cond - mat category @xcite + ca - gr - qc & 4,158 & 13,422 & 0.64 & 0.78 & 6.46 & 17.98 & 0.66 & 17 & 6.10 & co - authorship in gr - qc category @xcite + ca - hep - ph & 11,204 & 117,619 & 0.81 & 0.97 & 21.00 & 130.88 & 0.69 & 13 & 4.71 & co - authorship in hep - ph category @xcite + ca - hep - th & 8,638 & 24,806 & 0.68 & 0.85 & 5.74 & 12.99 & 0.58 & 18 & 5.96 & co - authorship in hep - th category @xcite + ca - dblp & 317,080 & 1,049,866 & 0.67 & 0.84 & 6.62 & 21.75 & 0.73 & 23 & 6.75 & dblp co - authorship network @xcite + l|r|r|r|r|r|r|r|r|r|l network & @xmath8 & @xmath9 & @xmath10 & @xmath11 & @xmath12 & @xmath13 & @xmath14 & @xmath15 & @xmath16 & description + + web - berkstan & 319,717 & 1,542,940 & 0.57 & 0.88 & 9.65 & 1,067.55 & 0.32 & 35 & 5.66 & web graph of stanford and uc berkeley @xcite + web - google & 855,802 & 4,291,352 & 0.75 & 0.92 & 10.03 & 170.35 & 0.62 & 24 & 6.27 & web graph google released in 2002 @xcite + web - notredame & 325,729 & 1,090,108 & 0.41 & 0.76 & 6.69 & 280.68 & 0.47 & 46 & 7.22 & web graph of university of notre dame @xcite + web - trec & 1,458,316 & 6,225,033 & 0.59 & 0.78 & 8.54 & 682.89 & 0.68 & 112 & 8.58 & web graph of trec wt10 g web corpus @xcite + + as - routeviews & 6,474 & 12,572 & 0.62 & 0.80 & 3.88 & 164.81 & 0.40 & 9 & 3.72 & as from oregon exchange bgp route view @xcite + as - caida & 26,389 & 52,861 & 0.61 & 0.81 & 4.01 & 281.93 & 0.33 & 17 & 3.86 & caida as relationships dataset + as - skitter & 1,719,037 & 12,814,089 & 0.99 & 1.00 & 14.91 & 9,934.01 & 0.17 & 5 & 3.44 & as from traceroutes run daily in 2005 by skitter + as - newman & 22,963 & 48,436 & 0.65 & 0.83 & 4.22 & 261.46 & 0.35 & 11 & 3.83 & as graph from newman @xcite + as - oregon & 13,579 & 37,448 & 0.72 & 0.90 & 5.52 & 235.97 & 0.46 & 9 & 3.58 & autonomous systems @xcite + gnutella-25 & 22,663 & 54,693 & 0.59 & 0.83 & 4.83 & 10.75 & 0.01 & 11 & 5.57 & gnutella network on march 25 2000 @xcite + gnutella-30 & 36,646 & 88,303 & 0.55 & 0.81 & 4.82 & 11.46 & 0.01 & 11 & 5.75 & gnutella p2p network on march 30 2000 @xcite + gnutella-31 & 62,561 & 147,878 & 0.54 & 0.81 & 4.73 & 11.60 & 0.01 & 11 & 5.94 & gnutella network on march 31 2000 @xcite + edonkey & 5,792,297 & 147,829,887 & 0.93 & 1.00 & 51.04 & 6,139.99 & 0.08 & 5 & 3.66 & p2p edonkey graph for a period of 47 hours in 2004 + + iptraffic & 2,250,498 & 21,643,497 & 1.00 & 1.00 & 19.23 & 94,889.05 & 0.00 & 5 & 2.53 & ip traffic graph a single router for 24 hours + atp - astro - ph & 54,498 & 131,123 & 0.70 & 0.87 & 4.81 & 16.67 & 0.00 & 28 & 7.78 & authors - to - papers network of astro - ph @xcite + atp - cond - mat & 57,552 & 104,179 & 0.65 & 0.79 & 3.62 & 10.54 & 0.00 & 31 & 9.96 & authors - to - papers network of cond - mat @xcite + atp - gr - qc & 14,832 & 22,266 & 0.47 & 0.60 & 3.00 & 9.72 & 0.00 & 35 & 11.08 & authors - to - papers network of gr - qc @xcite + atp - hep - ph & 47,832 & 86,434 & 0.60 & 0.76 & 3.61 & 16.80 & 0.00 & 27 & 8.55 & authors - to - papers network of hep - ph @xcite + atp - hep - th & 39,986 & 64,154 & 0.53 & 0.68 & 3.21 & 13.07 & 0.00 & 36 & 10.74 & authors - to - papers network of hep - th @xcite + atp - dblp & 615,678 & 944,456 & 0.49 & 0.64 & 3.07 & 13.61 & 0.00 & 48 & 12.69 & dblp authors - to - papers bipartite network + spending & 1,831,540 & 2,918,920 & 0.34 & 0.58 & 3.19 & 1,536.35 & 0.00 & 26 & 5.62 & users - to - keywords they bid + hw7 & 653,260 & 2,278,448 & 0.99 & 0.99 & 6.98 & 346.85 & 0.00 & 24 & 6.26 & downsampled advertiser - query bid graph + netflix & 497,959 & 100,480,507 & 1.00 & 1.00 & 403.57 & 28,432.89 & 0.00 & 5 & 2.31 & users - to - movies they rated . from netflix prize @xcite + queryterms & 13,805,808 & 17,498,668 & 0.28 & 0.41 & 2.53 & 14.92 & 0.00 & 86 & 19.81 & users - to - queries they submit to a search engine + clickstream & 199,308 & 951,649 & 0.39 & 0.87 & 9.55 & 430.74 & 0.00 & 7 & 3.83 & users - to - urls they visited @xcite + + bio - proteins & 4,626 & 14,801 & 0.72 & 0.91 & 6.40 & 24.25 & 0.12 & 12 & 4.24 & yeast protein interaction network @xcite + bio - yeast & 1,458 & 1,948 & 0.37 & 0.51 & 2.67 & 7.13 & 0.14 & 19 & 6.89 & yeast protein interaction network data @xcite + bio - yeastp0.001 & 353 & 1,517 & 0.73 & 0.93 & 8.59 & 20.18 & 0.57 & 11 & 4.33 & yeast protein - protein interaction map @xcite + bio - yeastp0.01 & 1,266 & 8,511 & 0.79 & 0.97 & 13.45 & 47.73 & 0.44 & 12 & 3.87 & yeast protein - protein interaction map @xcite + l|r|r|r|r|r|r|r|r|r|l network & @xmath8 & @xmath9 & @xmath10 & @xmath11 & @xmath12 & @xmath13 & @xmath14 & @xmath15 & @xmath16 & description + + road - ca & 1,957,027 & 2,760,388 & 0.80 & 0.85 & 2.82 & 3.17 & 0.06 & 865 & 310.97 & california road network + road - usa & 126,146 & 161,950 & 0.97 & 0.98 & 2.57 & 2.81 & 0.03 & 617 & 218.55 & usa road network ( only main roads ) + road - pa & 1,087,562 & 1,541,514 & 0.79 & 0.85 & 2.83 & 3.20 & 0.06 & 794 & 306.89 & pennsylvania road network + road - tx & 1,351,137 & 1,879,201 & 0.78 & 0.84 & 2.78 & 3.15 & 0.06 & 1,064 & 418.73 & texas road network + powergrid & 4,941 & 6,594 & 0.62 & 0.69 & 2.67 & 3.87 & 0.11 & 46 & 19.07 & power grid of western states power grid @xcite + mani - faces7k & 696 & 6,979 & 0.98 & 0.99 & 20.05 & 37.99 & 0.56 & 16 & 5.52 & faces ( 64x64 grayscale images ) ( connect 7k closest pairs ) + mani - faces4k & 663 & 3,465 & 0.90 & 0.97 & 10.45 & 20.20 & 0.56 & 29 & 8.96 & faces ( connect 4k closest pairs ) + mani - faces2k & 551 & 1,981 & 0.84 & 0.94 & 7.19 & 12.77 & 0.54 & 32 & 11.07 & faces ( connect 2k closest pairs ) + mani - facesk10 & 698 & 6,935 & 1.00 & 1.00 & 19.87 & 25.32 & 0.51 & 6 & 3.25 & faces ( connect every to 10 nearest neighbors ) + mani - facesk3 & 698 & 2,091 & 1.00 & 1.00 & 5.99 & 7.98 & 0.45 & 9 & 4.89 & faces ( connect every to 5 nearest neighbors ) + mani - facesk5 & 698 & 3,480 & 1.00 & 1.00 & 9.97 & 12.91 & 0.48 & 7 & 4.03 & faces ( connect every to 3 nearest neighbors ) + mani - swiss200k & 20,000 & 200,000 & 1.00 & 1.00 & 20.00 & 21.08 & 0.59 & 103 & 37.21 & swiss - roll ( connect 200k nearest pairs of nodes ) + mani - swiss100k & 19,990 & 99,979 & 1.00 & 1.00 & 10.00 & 11.02 & 0.59 & 162 & 58.32 & swiss - roll ( connect 100k nearest pairs of nodes ) + mani - swiss60k & 19,042 & 57,747 & 0.93 & 0.96 & 6.07 & 7.03 & 0.59 & 243 & 89.15 & swiss - roll ( connect 60k nearest pairs of nodes ) + mani - swissk10 & 20,000 & 199,955 & 1.00 & 1.00 & 20.00 & 25.38 & 0.56 & 10 & 5.47 & swiss - roll ( every node connects to 10 nearest neighbors ) + mani - swissk5 & 20,000 & 99,990 & 1.00 & 1.00 & 10.00 & 12.89 & 0.54 & 13 & 8.34 & swiss - roll ( every node connects to 5 nearest neighbors ) + mani - swissk3 & 20,000 & 59,997 & 1.00 & 1.00 & 6.00 & 7.88 & 0.50 & 17 & 6.89 & swiss - roll ( every node connects to 3 nearest neighbors ) + + atm - imdb & 2,076,978 & 5,847,693 & 0.49 & 0.82 & 5.63 & 65.41 & 0.00 & 32 & 6.82 & actors - to - movies graph from imdb ( imdb.com ) + imdb - top30 & 198,430 & 566,756 & 0.99 & 1.00 & 5.71 & 18.19 & 0.00 & 26 & 8.32 & actors - to - movies graph heavily preprocessed + imdb - raw07 & 601,481 & 1,320,616 & 0.54 & 0.79 & 4.39 & 20.94 & 0.00 & 32 & 8.55 & country clusters were extracted from this graph + imdb - france & 35,827 & 74,201 & 0.51 & 0.76 & 4.14 & 14.62 & 0.00 & 20 & 6.57 & cluster of french movies + imdb - germany & 21,258 & 42,197 & 0.56 & 0.78 & 3.97 & 13.69 & 0.00 & 34 & 7.47 & german movies ( to actors that played in them ) + imdb - india & 12,999 & 25,836 & 0.57 & 0.78 & 3.98 & 31.55 & 0.00 & 19 & 6.00 & indian movies + imdb - italy & 19,189 & 37,534 & 0.55 & 0.77 & 3.91 & 11.66 & 0.00 & 30 & 6.91 & italian movies + imdb - japan & 15,042 & 34,131 & 0.60 & 0.82 & 4.54 & 16.98 & 0.00 & 19 & 6.81 & japanese movies + imdb - mexico & 13,783 & 36,986 & 0.64 & 0.86 & 5.37 & 24.15 & 0.00 & 19 & 5.43 & mexican movies + imdb - spain & 15,494 & 31,313 & 0.51 & 0.76 & 4.04 & 14.22 & 0.00 & 28 & 6.44 & spanish movies + imdb - uk & 42,133 & 82,915 & 0.52 & 0.76 & 3.94 & 15.14 & 0.00 & 23 & 7.04 & uk movies + imdb - usa & 241,360 & 530,494 & 0.51 & 0.78 & 4.40 & 25.25 & 0.00 & 30 & 7.63 & usa movies + imdb - wgermany & 12,120 & 24,117 & 0.56 & 0.78 & 3.98 & 11.73 & 0.00 & 22 & 6.26 & west german movies + + amazon0302 & 262,111 & 899,792 & 0.95 & 0.97 & 6.87 & 11.14 & 0.43 & 38 & 8.85 & amazon products from 2003 03 02 @xcite + amazon0312 & 400,727 & 2,349,869 & 0.94 & 0.99 & 11.73 & 30.33 & 0.42 & 20 & 6.46 & amazon products from 2003 03 12 @xcite + amazon0505 & 410,236 & 2,439,437 & 0.94 & 0.99 & 11.89 & 30.93 & 0.43 & 22 & 6.48 & amazon products from 2003 05 05 @xcite + amazon0601 & 403,364 & 2,443,311 & 0.96 & 0.99 & 12.11 & 30.55 & 0.43 & 25 & 6.42 & amazon products from 2003 06 01 @xcite + amazonall & 473,315 & 3,505,519 & 0.94 & 0.99 & 14.81 & 52.70 & 0.41 & 19 & 5.66 & amazon products ( all 4 graphs merged ) @xcite + amazonallprod & 524,371 & 1,491,793 & 0.80 & 0.91 & 5.69 & 11.75 & 0.35 & 42 & 11.18 & products ( all products , source+target ) @xcite + amazonsrcprod & 334,863 & 925,872 & 0.84 & 0.91 & 5.53 & 11.53 & 0.43 & 47 & 12.11 & products ( only source products ) @xcite + * social networks : * the class of social networks in table [ tab : data_statsdesc_1 ] is particularly diverse and interesting .
it includes several large on - line social networks : a network of professional contacts from linkedin ( linkedin ) ; a friendship network of a livejournal blogging community ( livejournal01 ) ; and a who - trusts - whom network of epinions ( epinions ) .
it also includes an email network from enron ( email - enron ) and from a large european research organization . for the latter we generated three networks
: email - inside uses only the communication inside organization ; email - inout also adds external email addresses where email has been sent both way ; and email - all adds all communication inside the organization and to the outside world . also included in the class of social networks are networks that are not the central focus of the websites from which they come , but which instead serve as a tool for people to share information more easily .
for example , we have : the networks of a social bookmarking site delicious ( delicious ) ; a flickr photo sharing website ( flickr ) ; and a network from yahoo ! answers question answering website ( answers ) . in all these networks , a node refers to an individual and an edge is used to indicate that means that one person has some sort of interaction with another person , _
e.g. _ , one person subscribes to their neighbor s bookmarks or photos , or answers their questions .
* information and citation networks : * the class of information / citation networks contains several different citation networks .
it contains two citation networks of physics papers on arxiv.org , ( cit - hep - th and cit - hep - ph ) , and a network of citations of us patents ( cit - patents ) .
( these paper - to - paper citation networks are to be distinguished from scientific collaboration networks and author - to - paper bipartite networks , as described below . )
it also contains two types of blog citation networks . in the so - called post networks , nodes are
posts and edges represent hyperlinks between blog posts ( post - nat05 - 6 m and post - nat06all ) . on the other hand ,
the so - called blog network is the blog - level - aggregation of the same data , _
i.e. _ , there is a link between two blogs if there is a post in first that links the post in a second blog ( blog - nat05 - 6 m and blog - nat06all ) .
* collaboration networks : * the class of collaboration networks contain academic collaboration ( _ i.e. _ , co - authorship ) networks between physicists from various categories in arxiv.org ( ca - astro - ph , etc . ) and between authors in computer science ( ca - dblp ) .
it also contains a network of collaborations between pairs of actors in imdb ( ata - imdb ) , _ i.e. _ , there is an edge connecting a pair of actors if they appeared in the same movie .
( again , this should be distinguished from actor - to - movie bipartite networks , as described below . ) * web graphs : * the class of web graph networks includes four different web - graphs in which nodes represent web - pages and edges represent hyperlinks between those pages .
networks were obtained from google ( web - google ) , the university of notre dame ( web - notredame ) , trec ( web - trec ) , and stanford university ( web - berkstan ) .
the class of internet networks consists of various autonomous systems networks obtained at different sources , as well as a gnutella and edonkey peer - to - peer file sharing networks .
* bipartite networks : * the class of bipartite networks is particularly diverse and includes : authors - to - papers graphs from both computer science ( atp - dblp ) and physics ( atp - astro - ph , etc . ) ; a network representing users and the urls they visited ( clickstream ) ; a network representing users and the movies they rated ( netflix ) ; and a users - to - queries network representing query terms that users typed into a search engine ( queryterms ) .
( we also have analyzed several bipartite actors - to - movies networks extracted from the imdb database , which we have listed separately below . )
* biological networks : * the class of biological networks include protein - protein interaction networks of yeast obtained from various sources .
* low dimensional grid - like networks : * the class of low - dimensional networks consists of graphs constructed from road ( road - ca , etc . ) or power grid ( powergrid ) connections and as such might be expected to `` live '' on a two - dimensional surface in a way that all of the other networks do not .
we also added a `` swiss roll '' network , a @xmath17-dimensional manifold embedded in @xmath18-dimensions , and a `` faces '' dataset where each point is an @xmath19 by @xmath19 gray - scale image of a face ( embedded in @xmath20 dimensional space ) and where we connected the faces that were most similar ( using the euclidean distance ) .
* imdb , yahoo ! answers and amazon networks : * finally , we have networks from imdb , amazon , and yahoo !
answers , and for each of these we have separately analyzed subnetworks .
the imdb networks consist of actor - to - movie links , and we include the full network as well as subnetworks associated with individual countries based on the country of production . for the amazon networks , recall that amazon sells a variety of products , and for each item @xmath3 one may compile the list the up to ten other items most frequently purchased by buyers of @xmath3 .
this information can be presented as a directed network in which vertices represent items and there is a edge from item @xmath3 to another item @xmath4 if @xmath4 was frequently purchased by buyers of @xmath3 .
we consider the network as undirected .
we use five networks from a study of clauset _ et al . _
@xcite , and two networks from the viral marketing study from leskovec _ et al .
_ @xcite . finally , for the yahoo !
answers networks , we observe several deep cuts at large size scales , and so in addition the full network , we analyze the top six most well - connected subnetworks . in addition to providing a brief description of the network , tables [ tab : data_statsdesc_1 ] , [ tab : data_statsdesc_2 ] and [ tab : data_statsdesc_3 ] show the number of nodes and edges in each network , as well as other statistics which will be described in section [ sxn : obs_struct : stats ] .
( in all cases , we consider the network as undirected , and we extract and analyze the largest connected component . ) the sizes of these networks range from about @xmath21 nodes up to nearly @xmath22 million nodes , and from about @xmath23 edges up to more than @xmath0 million edges .
all of the networks are quite sparse their densities range from an average degree of about @xmath24 for the blog post network , up to an average degree of about @xmath25 in the network of movie ratings from netflix , and most of the other networks , including the purely social networks , have average degree around @xmath26 ( median average degree of @xmath27 ) . in many cases
, we examined several versions of a given network .
for example , we considered the entire imdb actor - to - movie network , as well as sub - pieces of it corresponding to different language and country groups .
detailed statistics for all these networks are presented in tables [ tab : data_statsdesc_1 ] , [ tab : data_statsdesc_2 ] and [ tab : data_statsdesc_3 ] and are described in section [ sxn : obs_struct ] . in total , we have examined over @xmath0 large real - world social and information networks , making this , to our knowledge , the largest and most comprehensive study of such networks .
hierarchical clustering is a common approach to community identification in the social sciences @xcite , but it has also found application more generally @xcite . in this procedure , one first defines a distance metric between pairs of nodes and then produces a tree ( in either a bottom - up or a top - down manner ) describing how nodes group into communities and how these group further into super - communities . a quite different approach that has received a great deal of attention ( and that will be central to our analysis )
is based on ideas from _ graph partitioning _ @xcite . in this case , the network is a modeled as simple undirected graph , where nodes and edges have no attributes , and a partition of the graph is determined by optimizing a merit function .
the graph partitioning problem is find some number @xmath6 groups of nodes , generally with roughly equal size , such that the number of edges between the groups , perhaps normalized in some way , is minimized .
let @xmath28 denote a graph , then the _ conductance _
@xmath29 of a set of nodes @xmath30 , ( where @xmath31 is assumed to contain no more than half of all the nodes ) , is defined as follows .
let @xmath32 be the sum of degrees of nodes in @xmath31 , and let @xmath33 be the number of edges with one endpoint in @xmath31 and one endpoint in @xmath34 , where @xmath34 denotes the complement of @xmath31 .
then , the conductance of @xmath31 is @xmath35 , or equivalently @xmath36 , where @xmath37 is the number of edges with both endpoints is @xmath31 .
more formally : given a graph @xmath38 with adjacency matrix @xmath3 the _ conductance of a set _ of nodes @xmath31 is defined as : @xmath39 where @xmath40 , or equivalently @xmath41 , where @xmath42 is a degree of node @xmath43 in @xmath38 .
moreover , in this case , the _ conductance of the graph _
@xmath38 is : @xmath44 [ def : conductance ] thus , the conductance of a set provides a measure for the quality of the cut @xmath45 , or relatedly the goodness of a community @xmath31 .
indeed , it is often noted that communities should be thought of as sets of nodes with more and/or better intra - connections than inter - connections ; see figure [ fig : conductance ] for an illustration .
when interested in detecting communities and evaluating their quality , we prefer sets with small conductances , _
i.e. _ , sets that are densely linked inside and sparsely linked to the outside .
although numerous measures have been proposed for how community - like is a set of nodes , it is commonly noted_e.g .
_ , see shi and malik @xcite and kannan , vempala , and vetta @xcite that conductance captures the `` gestalt '' notion of clustering @xcite , and as such it has been widely - used for graph clustering and community detection @xcite .
there are many other density - based measures that have been used to partition a graph into a set of communities @xcite .
one that deserves particular mention is modularity @xcite . for a given partition of a network into a set of communities
, modularity measures the number of within - community edges , relative to a null model that is usually taken to be a random graph with the same degree distribution .
thus , modularity was originally introduced and it typically used to measure the strength or quality of a particular partition of a network .
we , however , are interested in a quite different question than those that motivated the introduction of modularity . rather than seeking to partition a graph into the `` best '' possible partition of communities
, we would like to know how good is a particular element of that partition , _
i.e. _ , how community - like are the best possible communities that modularity or any other merit function can hope to find , in particular as a function of the size of that partition .
in addition to capturing very well our intuitive notion of what it means for a set of nodes to be a good community , the use of conductance as an objective function has an added benefit : there exists an extensive theoretical and practical literature on methods for approximately optimizing it .
( finding cuts with exactly minimal conductance is np - hard . ) in particular , the theory literature contains several algorithms with provable approximation performance guarantees .
first , there is the spectral method , which uses an eigenvector of the graph s laplacian matrix to find a cut whose conductance is no bigger than @xmath29 if the graph actually contains a cut with conductance @xmath46 @xcite .
the spectral method also produces lower bounds which can show that the solution for a given graph is closer to optimal than promised by the worst - case guarantee .
second , there is an algorithm that uses multi - commodity flow to find a cut whose conductance is within an @xmath47 factor of optimal @xcite .
spectral and multi - commodity flow based methods are complementary in that the worst - case @xmath48 approximation factor is obtained for flow - based methods on expander graphs @xcite , a class of graphs which does not cause problems for spectral methods , whereas spectral methods can confuse long path with deep cuts @xcite , a difference that does not cause problems for flow - based methods .
third , and very recently , there exists an algorithm that uses semidefinite programming to find a solution that is within @xmath49 of optimal @xcite .
this paper sparked a flurry of theoretical research on a family of closely related algorithms including @xcite , all of which can be informally described as combinations of spectral and flow - based techniques which exploit their complementary strengths . however , none of those algorithms are currently practical enough to use in our study . of the above three theoretical algorithms , the spectral method is by far the most practical .
also very common are recursive bisection heuristics : recursively divide the graph into two groups , and then further subdivide the new groups until the desired number of clusters groups is achieved .
this may be combined with local improvement methods like the kernighan - lin and fiduccia - mattheyses procedures @xcite , which are fast and can climb out of some local minima .
the latter was combined with a multi - resolution framework to create metis @xcite , a very fast program intended to split mesh - like graphs into equal sized pieces .
the authors of metis later created cluto @xcite , which is better tuned for clustering - type tasks .
finally we mention graclus @xcite , which uses multi - resolution techniques and kernel @xmath6-means to optimize a metric that is closely related to conductance .
while the preceding were all approximate algorithms for finding the lowest conductance cut in a whole graph , we now mention mqi @xcite , an _ exact _ algorithm for the slightly different problem of finding the lowest conductance cut in _ half _ of a graph .
this algorithm can be combined with a good method for initially splitting the graph into two pieces ( such as metis or the spectral method ) to obtain a surprisingly strong heuristic method for finding low conductance cuts in the whole graph @xcite .
the exactness of the second optimization step frequently results in cuts with extremely low conductance scores , as will be visible in many of our plots .
mqi can be implemented by solving single parametric max flow problems , or sequences of ordinary max flow problems .
parametric max flow ( with mqi described as one of the applications ) was introduced by @xcite , and recent empirical work is described in @xcite , but currently there is no publicly available code that scales to the sizes we need .
ordinary max flow is a very thoroughly studied problem . currently , the best theoretical time bounds are @xcite , the most practical algorithm is @xcite , while the best implementation is hi_pr by @xcite .
since metis+mqi using the hi_pr code is very fast and scalable , while the method empirically seems to usually find the lowest or nearly lowest conductance cuts in a wide variety of graphs , we have used it extensively in this study
. we will also extensively use local spectral algorithm of andersen , chung , and lang @xcite to find node sets of low conductance , _
i.e. _ , good communities , around a seed node .
this algorithm is also very fast , and it can be successfully applied to very large graphs to obtain more `` well - rounded '' , `` compact , '' or `` evenly - connected '' communities than those returned by meits+mqi .
the latter observation ( described in more detail in section [ algo - notes - section ] ) is since local spectral methods also confuse long paths ( which tend to occur in our very sparse network datasets ) with deep cuts .
this algorithm takes as input two parameters the seed node and a parameter @xmath50 that intuitively controls the locality of the computation and it outputs a set of nodes .
local spectral methods were introduced by spielman and teng @xcite , and they have roughly the same kind of quadratic approximation guarantees as the global spectral method , but they have computational cost is proportional to the size of the obtained piece @xcite .
in this section , we discuss the _ network community profile plot _ ( ncp plot ) , which measures the quality of network communities at different size scales .
we start in section [ sxn : ncpp : def ] by introducing it .
then , in section [ sxn : ncpp : low_small ] , we present the ncp plot for several examples of networks which inform peoples intuition and for which the ncp plot behaves in a characteristic manner
. then , in sections [ sxn : ncpp : large_sparse ] and [ sxn : ncpp : large_sparse_more ] we present the ncp plot for a wide range of large real world social and information networks .
we will see that in such networks the ncp plot behaves in a qualitatively different manner . in order to more finely resolve community structure in large networks
, we introduce the _ network community profile plot _ ( ncp plot ) . intuitively , the ncp plot measures the quality of the best possible community in a large network , as a function of the community size . formally
, we may define it as the conductance value of the best conductance set of cardinality @xmath6 in the entire network , as a function of @xmath6 .
given a graph @xmath38 with adjacency matrix @xmath3 , the _ network community profile plot ( ncp plot ) _
plots @xmath51 as a function of @xmath6 , where @xmath52 where @xmath53 denotes the cardinality of the set @xmath31 , and where the conductance @xmath54 of @xmath31 is given by equation ( [ eqn : conductance_set ] ) .
since this quantity is intractable to compute , we will employ well - studied approximation algorithms for the minimum conductance cut problem to approximate it .
in particular , operationally we will use several natural heuristics based on approximation algorithms to do graph partitioning in order to compute different approximations to the ncp plot .
although other procedures will be described in section [ algo - notes - section ] , we will primarily employ two procedures .
first , metis+mqi , _ i.e. _ , the graph partitioning package metis @xcite followed by the flow - based post - processing procedure mqi @xcite ; this procedure returns sets that have very good conductance values .
second , the local spectral algorithm of andersen , chung , and lang @xcite ; this procedure returns sets that are somewhat more `` compact '' or `` smoothed '' or `` regularized , '' but that often have somewhat worse conductance values .
just as the conductance of a set of nodes provides a quality measure of that set as a community , the shape of the ncp plot provides insight into the community structure of a graph as a whole .
for example , the magnitude of the conductance tells us how well clusters of different sizes are separated from the rest of the network . one might hope to obtain some sort of `` smoothed '' measure of the notion of the best community of size @xmath6 ( _ e.g. _ , by considering an average of the conductance value over all sets of a given size or by considering a smoothed extremal statistic such as a @xmath55-th percentile ) rather than conductance of the best set of that size .
we have not defined such a measure since there is no obvious way to average over all subsets of size @xmath6 and obtain a meaningful approximation to the minimum . on the other hand
, our approximation algorithm methodology implicitly incorporates such an effect .
although metis+mqi finds sets of nodes with extremely good conductance value , empirically we observe that they often have little or no internal structure they can even be disconnected . on the other hand ,
since spectral methods in general tend to confuse long paths with deep cuts @xcite , the local spectral algorithm finds sets that are `` tighter '' and more `` well - rounded '' and thus in many ways more community - like .
( see sections [ sxn : related : lowcondalgs ] and [ algo - notes - section ] for details on these algorithmic issues and interpretations . )
the ncp plot behaves in a characteristic manner for graphs that are `` well - embeddable '' into an underlying low - dimensional geometric structure . to illustrate this ,
consider figure [ fig : ncpp_lowdim ] . in figure
[ fig : ncpp_lowdim : toy ] , we show the results for a @xmath56-dimensional chain , a @xmath17-dimensional grid , and a @xmath18-dimensional cube . in each case , the ncp plot is steadily downward sloping as a function of the number of nodes in the smaller cluster .
moreover , the curves are straight lines with a slope equal to @xmath57 , where @xmath58 is the dimensionality of the underlying grids .
in particular , as the underlying dimension increases then the slope of the ncp plot gets less steep .
thus , we observe : if the network under consideration corresponds to a @xmath58-dimensional grid , then the ncp plot shows that @xmath59 this is simply a manifestation of the isoperimetric ( _ i.e. _ , surface area to volume ) phenomenon : for a grid , the `` best '' cut is obtained by cutting out a set of adjacent nodes , in which case the surface area ( number of edges cut ) increases as @xmath60 ) , while the volume ( number of vertices / edges inside the cluster ) increases as @xmath61 .
this qualitative phenomenon of a steadily downward sloping ncp plot is quite robust for networks that `` live '' in a low - dimensional structure , _
e.g. _ , on a manifold or the surface of the earth .
for example , figure [ fig : ncpp_lowdim : power ] shows the ncp plot for a power grid network of western states power grid @xcite , and figure [ fig : ncpp_lowdim : road ] shows the ncp plot for a road network of california .
these two networks have very different sizes the power grid network has @xmath62 nodes and @xmath63 edges , and the road network has @xmath64 nodes and @xmath65 edges and they arise in very different application domains . in both cases , however , we see predominantly downward sloping ncp plot , very much similar to the profile of a simple @xmath17-dimensional grid . indeed ,
the `` best - fit '' line for power grid gives the slope of @xmath66 , which by ( [ eqn : slope_dim ] ) suggests that @xmath67 , which is not far from the `` true '' dimensionality of @xmath17 .
moreover , empirically we observe that minima in the ncp plot correspond to community - like sets , which are occasionally nested .
this corresponds to hierarchical community organization .
for example , the nodes giving the dip at @xmath68 are included in the nodes giving the dip at @xmath69 , while dips at @xmath70 and @xmath71 are both included in the dip at @xmath72 . in a similar manner ,
figure [ fig : phimodels : manifold ] shows the profile plot for a graph generated from a `` swiss roll '' dataset which is commonly examined in the manifold and machine learning literature @xcite . in this case
, we still observe a downward sloping ncp plot that corresponds to internal dimensionally of the manifold ( 2 in this case ) .
finally , figures [ fig : phimodels : expander1 ] and [ fig : phimodels : expander2 ] show ncp plots for two graphs that are very good expanders .
the first is a @xmath73 graph with @xmath74 nodes and a number of edges such that the average degree is @xmath75 , @xmath27 , and @xmath76 .
the second is a constant degree expander : to make one with degree @xmath58 , we take the union of @xmath58 disjoint but otherwise random complete matchings , and we have plotted the results for @xmath77 . in both of these cases ,
the ncp plot is roughly flat , which we also observed in figure [ fig : ncpp_lowdim : toy ] for a clique , which is to be expected since the minimum conductance cut in the entire graph can not be too small for a good expander @xcite .
somewhat surprisingly ( especially when compared with large networks in section [ sxn : ncpp : large_sparse ] ) , a steadily decreasing downward ncp plot is seen for small social networks that have been extensively studied in validating community detection algorithms .
several examples are shown in figures [ fig : ncpp_small ] .
for these networks , the interpretation is similar to that for the low - dimensional networks : the downward slope indicates that as potential communities get larger and larger , there are relatively more intra - edges than inter - edges ; and empirically we observe that local minima in the ncp plot correspond to sets of nodes that are plausible communities .
consider , _
e.g. _ , zachary s karate club @xcite network ( zacharykarate ) , an extensively - analyzed social network @xcite .
the network has @xmath78 nodes , each of which represents a member of a karate club , and @xmath79 edges , each of which represent a friendship tie between two members .
figure [ fig : ncpp_small : karate_graph ] depicts the karate club network , and figure [ fig : ncpp_small : karate_plot ] shows its ncp plot .
there are two local minima in the plot : the first dip at @xmath80 corresponds to the cut @xmath3 , and the second dip at @xmath81 corresponds to cut @xmath4 .
note that cut @xmath4 , which separates the graph roughly in half , has better conductance value than cut @xmath3 .
this corresponds with the intuition about the ncp plot derived from studying low - dimensional graphs .
note also that the karate network corresponds well with the intuitive notion of a community , where nodes of the community are densely linked among themselves and there are few edges between nodes of different communities . [ cols="^,^ " , ] thus , intuitively , one can think of small well - separated communities those below the @xmath82 size scale that consist of subsets of the small trees starting to grow , and as they pass the @xmath82 size scale and become bigger and bigger , they blend in more and more with the central part of the network , which ( since it exhibits certain expander - like properties ) does not have particularly well - defined communities .
note ( more generally ) that if there are @xmath82 nodes in the small tree at the top of the graph , then the dip in the ncp plot in figure [ fig : models_clique2 ] is of depth @xmath83 .
in particular , if @xmath84 then the depth of this cut is @xmath85 . intuitively , the ncp plot increases since the `` cost '' per edge for every additional edge inside a cluster increases with the size of the cluster .
for example , in cut @xmath3 in figure [ fig : models_clique1 ] , the `` price '' for having @xmath18 internal edges is to cut @xmath27 edges , _
i.e. _ , @xmath17 edges cut per edge inside . to expand the cluster by just a single edge
, one has to move one level down in the tree ( toward the cut @xmath4 ) where now the price for a single edge is @xmath75 edges , and so on .
the question arises now as to whether we can find a simple generative model that can explain both the existence of small well - separated whisker - like clusters and also an expander - like core whose best clusters get gradually worse as the purported communities increase in size .
intuitively , a satisfactory network generation model must successfully take into account the following two mechanisms : * the model should produce a relatively large number of relatively small but still large when compared to random graphs
well connected and distinct whisker - like communities .
( this should reproduce the downward part of the community profile plot and the minimum at small size scales . ) * the model should produce a large expander - like core , which may be thought of as consisting of intermingled communities , perhaps growing out from the whisker - like communities , the boundaries of which get less and less well - defined as the communities get larger and larger and as they gradually blend in with rest of the network .
( this should reproduce the gradual upward sloping part of the community profile plot . ) the so - called _ forest fire model _
@xcite captures exactly these two competing phenomena .
the forest fire model is a model of graph generation ( that generates directed graphs an effect we will ignore ) in which new edges are added via a recursive `` burning '' mechanism in an epidemic - like fashion . since the details of the recursive burning process are critical for the model s success , we explain it in some detail . to describe the forest fire model of @xcite , let us fix two parameters , a _
forward burning probability _ @xmath86 and a _ backward burning probability _ @xmath87 .
we start the entire process with a single node , and at each time step @xmath88 , we consider a new node @xmath32 that joins the graph @xmath89 constructed thus far .
the node @xmath32 forms out - links to nodes in @xmath89 as follows : * node @xmath32 first choose a node @xmath90 , which we will refer to as a `` seed '' node or an `` ambassador '' node , uniformly at random and forms a link to @xmath90 .
* node @xmath32 selects @xmath91 out - links and @xmath92 in - links of @xmath90 that have not yet been visited .
( @xmath91 and @xmath92 are two geometrically distributed random numbers with means @xmath93 and @xmath94 , respectively .
if not enough in - links or out - links are available , then @xmath32 selects as many as possible . )
let @xmath95 denote the nodes at the other ends of these selected links .
* node @xmath32 forms out - links to @xmath95 , and then applied step ( ii ) recursively to each of the @xmath95 , except that nodes can not be visited a second time during the process .
thus , burning of links in the forest fire model begins at node @xmath90 , spreads to @xmath95 , and proceeds recursively until the process dies out .
one can view such a process intuitively as corresponding to a model in which a person comes to the party and first meets an ambassador who then introduces him or her around .
if the person creates a small number of friendships these will likely be from the ambassadors `` community , '' but if the person happens to create many friendships then these will likely go outside the ambassador s circle of friends .
this way , the ambassador s community might gradually get intermingled with the rest of the network .
joins the network and selects a seed node @xmath90 .
middle : @xmath32 then attaches itself by recursively linking to @xmath90 s neighbors , @xmath90 s neighbor - neighbors , and so on , according to the `` forest fire '' burning mechanism described in the text .
right : a new node @xmath96 joins the network , selects seed @xmath97 , and recursively adds links using the same `` forest fire '' burning mechanism .
notice that if @xmath96 causes a large `` fire , '' it links to a large number of existing nodes .
in this way , as potential communities around node @xmath90 grow , the ncp plot is initially decreasing , but then larger communities around @xmath90 gradually blend - in with the rest of the network , which leads the ncp plot to increase.,title="fig:",scaledwidth=15.0% ] joins the network and selects a seed node @xmath90 .
middle : @xmath32 then attaches itself by recursively linking to @xmath90 s neighbors , @xmath90 s neighbor - neighbors , and so on , according to the `` forest fire '' burning mechanism described in the text .
right : a new node @xmath96 joins the network , selects seed @xmath97 , and recursively adds links using the same `` forest fire '' burning mechanism . notice that if @xmath96 causes a large `` fire , '' it links to a large number of existing nodes . in this way , as potential communities around node @xmath90 grow , the ncp plot is initially decreasing , but then larger communities around @xmath90 gradually blend - in with the rest of the network , which leads the ncp plot to increase.,title="fig:",scaledwidth=15.0% ] joins the network and selects a seed node @xmath90 .
middle : @xmath32 then attaches itself by recursively linking to @xmath90 s neighbors , @xmath90 s neighbor - neighbors , and so on , according to the `` forest fire '' burning mechanism described in the text .
right : a new node @xmath96 joins the network , selects seed @xmath97 , and recursively adds links using the same `` forest fire '' burning mechanism . notice that if @xmath96 causes a large `` fire , '' it links to a large number of existing nodes . in this way , as potential communities around node @xmath90 grow , the ncp plot is initially decreasing , but then larger communities around @xmath90 gradually blend - in with the rest of the network , which leads the ncp plot to increase.,title="fig:",scaledwidth=17.0% ] [ fig : forestfire ] two properties of this model are particularly significant .
first , although many nodes might form one or a small number of links , certain nodes can produce large conflagrations , burning many edges and thus forming a large number of out - links before the process ends .
such nodes will help generate a skewed out - degree distribution , and they will also serve as `` bridges '' that connect formerly disparate parts of the network
. this will help make the ncp plot gradually increase .
second , there is a locality structure in that as each new node @xmath32 arrives over time , it is assigned a `` center of gravity '' in some part of the network , _
i.e. _ , at the ambassador node @xmath90 , and the manner in which new links are added depends sensitively on the local graph structure around node @xmath90 . not only does the probability of linking to other nodes decrease rapidly with distance to the current ambassador , but because of the recursive process regions with a higher density of links tend to attract new links .
figure [ fig : forestfire ] illustrates this .
initially , there is a small community around node @xmath90 .
then , node @xmath32 joins and using the forest fire mechanism locally attaches to nodes in the neighborhood of seed node @xmath90 .
the growth of the community around @xmath90 corresponds to downward part of the ncp plot .
however , if a node @xmath96 then joins and causes a large fire , this has the effect of larger and larger communities around @xmath90 blending into and merging with the rest of the network .
cc community profile plots for the forest fire model at various parameter settings .
the backward burning probability is @xmath98 , and we increase ( left to right , top to bottom ) the forward burning probability @xmath99 .
note that the largest and smallest values for @xmath86 lead to less realistic community profile plots , as discussed in the text .
, title="fig:",scaledwidth=45.0% ] & community profile plots for the forest fire model at various parameter settings .
the backward burning probability is @xmath98 , and we increase ( left to right , top to bottom ) the forward burning probability @xmath99 .
note that the largest and smallest values for @xmath86 lead to less realistic community profile plots , as discussed in the text .
, title="fig:",scaledwidth=45.0% ] + community profile plots for the forest fire model at various parameter settings .
the backward burning probability is @xmath98 , and we increase ( left to right , top to bottom ) the forward burning probability @xmath99 .
note that the largest and smallest values for @xmath86 lead to less realistic community profile plots , as discussed in the text .
, title="fig:",scaledwidth=45.0% ] & community profile plots for the forest fire model at various parameter settings .
the backward burning probability is @xmath98 , and we increase ( left to right , top to bottom ) the forward burning probability @xmath99 .
note that the largest and smallest values for @xmath86 lead to less realistic community profile plots , as discussed in the text .
, title="fig:",scaledwidth=45.0% ] + community profile plots for the forest fire model at various parameter settings .
the backward burning probability is @xmath98 , and we increase ( left to right , top to bottom ) the forward burning probability @xmath99 .
note that the largest and smallest values for @xmath86 lead to less realistic community profile plots , as discussed in the text .
, title="fig:",scaledwidth=45.0% ] & community profile plots for the forest fire model at various parameter settings .
the backward burning probability is @xmath98 , and we increase ( left to right , top to bottom ) the forward burning probability @xmath99 .
note that the largest and smallest values for @xmath86 lead to less realistic community profile plots , as discussed in the text . ,
title="fig:",scaledwidth=45.0% ] + not surprisingly , however , the forest fire model is sensitive to the choice of the burning probabilities @xmath86 and @xmath100 . we have experimented with a wide range of network sizes and values for these parameters , and in figure [ fig : phiforestfire ] , we show the community profile plots of several @xmath101 node forest fire networks generated with @xmath98 and several different values of @xmath86 .
the first thing to note is that since we are varying @xmath86 the six plots in figure [ fig : phiforestfire ] , we are viewing networks with very different densities . next , notice that if , _ e.g. _ , @xmath102 or @xmath103 then we observe a very natural behavior : the conductance nicely decreases , reaches the minimum somewhere between @xmath26 and @xmath0 nodes , and then slowly but not too smoothly increases .
not surprisingly , it is in this parameter region where the forest fire model has been shown to exhibit realistic time evolving graph properties such as densification and shrinking diameters @xcite
. next , also notice that if @xmath86 is too low or too high , then we obtain qualitatively different results .
for example , if @xmath104 , then the community profile plot gradually decreases for nearly the entire plot . for this choice of parameters ,
the forest fire does not spread well since the forward burning probability is too small , the network is extremely sparse and is tree - like with just a few extra edges , and so we get large well separated `` communities '' that get better as they get larger . on the other hand , when burning probability is too high , _
e.g. _ , @xmath105 , then the ncp plot has a minimum and then rises extremely rapidly . for this choice of parameters ,
if a node which initially attached to a whisker successfully burns into the core , then it quickly establishes many successful connections to other nodes in the core .
thus , the network has relatively large whiskers that failed to establish such a connection and a very expander - like core , with no intermediate region , and the increase in the community profile plot is quite abrupt .
we have examined numerous other properties of the graphs generated by the forest fire model and have found them to be broadly consistent with the social and information networks we have examined .
one property , however , that is of particular interest is what the whiskers look like .
figure [ fig : whiskff ] shows an example of several whiskers generated by the forest fire model if we choose @xmath106 and @xmath107 .
they are larger and more well - structured than the tree - like whiskers from the random graph model of section [ sxn : models : sparse_gw ] .
also notice that they all look plausibly community - like with a core of the nodes densely linked among themselves and the bridge edge then connects the whisker to the rest of the network . and @xmath98 .
the green square node belongs to the network core , and by cutting the edge connecting it with red circular node we separate the community of circles from the rest of the network ( depicted as a green square).,title="fig:",width=83 ] and @xmath98 .
the green square node belongs to the network core , and by cutting the edge connecting it with red circular node we separate the community of circles from the rest of the network ( depicted as a green square).,title="fig:",width=83 ] and @xmath98 .
the green square node belongs to the network core , and by cutting the edge connecting it with red circular node we separate the community of circles from the rest of the network ( depicted as a green square).,title="fig:",width=75 ] and @xmath98 .
the green square node belongs to the network core , and by cutting the edge connecting it with red circular node we separate the community of circles from the rest of the network ( depicted as a green square).,title="fig:",width=75 ] and @xmath98 .
the green square node belongs to the network core , and by cutting the edge connecting it with red circular node we separate the community of circles from the rest of the network ( depicted as a green square).,title="fig:",width=86 ] and @xmath98 .
the green square node belongs to the network core , and by cutting the edge connecting it with red circular node we separate the community of circles from the rest of the network ( depicted as a green square).,title="fig:",width=56 ] and @xmath98 .
the green square node belongs to the network core , and by cutting the edge connecting it with red circular node we separate the community of circles from the rest of the network ( depicted as a green square).,title="fig:",width=56 ] we conclude by noting that there has also been interest in developing hierarchical graph generation models , _
i.e. _ , models in which a hierarchy is given and the linkage probability between pairs of nodes decreases as a function of their distance in the hierarchy @xcite . the motivation for this comes largely from the intuition that nodes in social networks and are joined in to small relatively tight groups that are then further join into larger groups , and so on . as figures [ fig : phimodels : bh ] and [ fig : phimodels : cga ] make clear , however , such models do not immediately lead to community structure similar to what we have observed and which has been reproduced by the forest fire model . on the other hand , although there are significant differences between hierarchical models and the forest fire model , @xcite notes that there are similarities .
in particular , in the forest fire model a new node @xmath32 is assigned an ambassador @xmath90 as an entry point to the network . this is analogous to a child having a parent in the hierarchy which helps to determine how that node links to the remainder of the network .
similarly , many hierarchical models have a connection probability that decreases exponentially in the hierarchical tree distance . in the forest fire model ,
the probability that a node @xmath32 will burn along a particular path to another node @xmath108 is exponentially small in the path length , although the analogy is not perfect since there may exist many possible paths .
in this section , we discuss several aspects of our main results in a broader context . in particular , in section [ sxn : discussion : ground_truth ] , we compare to several data sets
in which there is some notion of `` ground truth '' community and we also describe several broader non - technical implications of our results .
then , in section [ sxn : discussion : community ] , we describe recent work on community detection and identification . finally , in section [ sxn : discussion : technical ]
, we discuss several technical and algorithmic issues and questions raised by our work . in this subsection
, we examine the relationship between network communities of the sort we have been discussing so far and some notion of `` ground truth . ''
when considering a real network , one hopes that the output of a community finding algorithms will be `` real '' communities that exist in some meaningful sense in the real world . for example , in the karate club network in figure [ fig : ncpp_small : karate_graph ] , the cut found by the algorithm corresponds in some sense to a true community , in that it splits the nodes almost precisely as they split into two newly formed karate clubs . in this section ,
we take a different approach : we take networks in which there are explicitly defined communities , and we examine how well these communities are separated from the rest of the network . in particular , we examine a minimum conductance profile of several network datasets , where we can associate with each node one or more community labels which are exogenously specified .
note that we are overloading the term `` community '' here , as in this context the term might mean one of two things : first , it can refer to groups of nodes with good conductance properties ; and second , it can refer to groups of nodes that belong to the same self - defined or exogenously - specified group .
we consider the following five datasets : * livejournal12 @xcite : livejournal is an on - line blogging site where users can create friendship links to other users .
in addition , users can create groups which other users can then join . in livejournal , there are @xmath109 such groups , and a node belongs to @xmath110 groups on the average .
thus , in addition to the information in the interaction graph , we have labels specifying those groups with which a user is associated , and thus we may view each such group as determining a `` ground truth '' community . * ca - dblp @xcite : we considered a co - authorship network in which nodes are authors and there is an edge if authors co - authored at least one paper . here , publication venues ( _ e.g. _ , journals and conferences ) can play the role of `` ground truth '' communities .
that is , an author is a member of a particular group or community if he or she published at a particular conference or in a particular journal . in our dblp network , there are @xmath111 such groups , with a node belonging to @xmath112 on the average . * amazonallprod @xcite : this is a network of products that are commonly purchased together at amazon.com .
( intuitively one might expect that , _
e.g. _ , gardening books are frequently purchased together , so the network structure might reflect a well - connected cluster of gardening books . ) here , each item belongs to one or more hierarchically organized categories ( book , movie genres , product types , etc . ) , and products from the same category define a group which we will view as a `` ground truth '' community .
items can belong to @xmath113 different groups , and each item belongs to @xmath114 groups on the average . * atm - imdb : this network is a bipartite actors - to - movies network composed from imdb data , and an actor @xmath3 is connected to movie @xmath4 if @xmath3 appeared in @xmath4 . for each movie
we also know the language and the country where it was produced . countries and languages may be taken as `` ground truth '' communities or groups , where every movie belongs to exactly one group and actors belongs to all groups to which movies that they appeared in belong . in our dataset , we have @xmath115 language groups and @xmath116 country groups . * email - inside and email - inout @xcite : this is an email communication network from a large european research organization conducting research in natural sciences : physics , chemistry , biology and computer science
. each of @xmath117 members of the organization belongs to exactly one of @xmath118 departments , and we use the department memberships to define `` ground truth '' communities .
although none of these notions of `` ground truth '' is perfect , many community finding algorithms use precisely this form of anecdotal evaluation : a network is taken , network communities are found , and then the correspondence of network communities to `` ground truth '' communities is evaluated .
note , in contrast , we are evaluating how `` ground truth '' communities behave at different size scales with respect to our methodology , rather than examining how the groups we find relate to `` ground truth '' communities .
furthermore , note that the notions of `` ground truth '' are not all the same we might expect that people publish papers across several different venues in a very different way than actors appear in movies from different countries .
more detailed statistics for each of these networks may be found in tables [ tab : data_statsdesc_1 ] , [ tab : data_statsdesc_2 ] and [ tab : data_statsdesc_3 ] . to examine the quality of `` ground truth '' communities in the these network datasets , we take all groups and measure the conductance of the cut that separates that group from the rest of the network .
thus , we generated ncp plots in the following way .
for every `` ground truth '' community , we measure the conductance of the cut separating it from the rest of the graph , from which we obtain a scatter plot of community size versus conductance . then , we take the lower - envelope of this plot , _
i.e. _ , for every integer @xmath6 we find the conductance value of the community of size @xmath6 that has the lowest conductance .
figure [ fig : explicitcmty ] shows the results for these network datasets ; the figure also shows the ncp plot obtained from using the local spectral algorithm on both the original network ( plotted in red ) and on the rewired network ( plotted in black ) .
several observations can be made : * the conductance of `` ground truth '' communities follows that for the network communities up to until size @xmath26-@xmath0 nodes , _
i.e. _ , larger communities get successively more community - like . as `` ground truth '' communities get larger , their conductance values tend to get worse and worse , in agreement with network communities discovered with graph partitioning approximation algorithms .
thus , the qualitative trend we observed in nearly every large sparse real - world network ( of the best communities blending in with the rest of the network as they grow in size ) is seen to hold for small `` ground truth '' communities .
* one might expect that the ncp plot for the `` ground truth '' communities ( the green curves ) will be somewhere between the ncp plot of the original network ( red curves ) and that for the rewired network ( black curves ) , and this is seen to be the case in general .
the ncp plot for network communities goes much deeper and rises more gradually than for `` ground truth '' communities .
this is also consistent with our general observation that only small communities tend to be dense and well separated , and to separate large groups one has to cut disproportionately many edges .
* for the two social networks we studied ( livejournal12 and ca - dblp ) , larger `` ground truth '' communities have conductance scores that get quite `` random '' , _
i.e. _ , they are as well separated as they would be in a randomly rewired network ( green and black curves overlap ) .
this is likely associated with the relatively weak and overlapping notion of `` ground truth ''
we associated with those two network datasets .
on the other hand , for amazonallprod and atm - imdb networks , the general trend still remains but large `` ground truth '' communities have conductance scores that lie well below the rewired network curve .
our email network illustrates a somewhat different point . the ncp plot for email - inside
should be compared with that for email - inout , which is displayed in figure [ fig : phidatasets1 ] .
the email - inside email network is rather small , and so it has a decreasing community profile plot , in agreement with the results for small social networks . since communication is mainly focused between the members of the same department , both network and `` ground truth '' communities
are well expressed .
next , compare the ncp plot of email - inside with that of email - inout ( figure [ fig : phidatasets1 ] ) .
we see that the ncp plot of email - inside slopes downwards ( as we consider only the communication inside the organization ) , but as soon as we consider the communication inside the organization and to the outside world ( email - inout , or alternatively , see email - enron ) then we see a completely different and more familiar picture the ncp plot drops and then slowly increases .
this suggests that the organizational structure , ( _ e.g. _ , departments ) manifest themselves in the internal communication network , but as soon as we put the organization into the broader context ( _ i.e. _ , how it communicates to the rest of the world ) then the internal department structure seems to disappear .
in contrast to numerous studies of community structure , we find that there is a natural size scale to communities .
communities are relatively small , with sizes only up to about @xmath0 nodes .
we also find that above size of about @xmath0 , the `` quality '' of communities gets worse and worse and communities more and more `` blend into '' the network .
eventually , even the existence of communities ( at least when viewed as sets with stronger internal than external connectivity ) is rather questionable .
we show that large social and information networks can be decomposed into a large number of small communities and a large dense and intermingled network `` core''we empirically establish that the `` core '' contains on average @xmath119 of the nodes and @xmath120 of all edges .
but , as demonstrated by figure [ fig : phicore ] , the `` core '' itself has a nontrivial structure in particular , it has a core - whisker structure that is analogous to the original complete network . * the dunbar number : * our observation on the limit of community size agrees with dunbar @xcite who predicted that roughly @xmath7 is the upper limit on the size of a well - functioning human community .
moreover , allen @xcite gives evidence that on - line communities have around @xmath121 members , and on - line discussion forums start to break down at about @xmath122 active contributors .
church congregations , military companies , divisions of corporations , all are close to the sum of @xmath7 @xcite .
we are thus led to ask : why , above this size , is community quality inversely proportional to its size ? and why are ncp plots of small and large networks so different ?
previous studies mainly focused on small networks ( _ e.g. _ , see @xcite ) , which are simply not large enough for the clusters to gradually blend into one another as one looks at larger size scales .
our results do not disagree with literature at small sizes .
but it seems that in order to make our observations one needs to look at large networks .
it is only when dunbar s limit is passed that we find large communities blurring and eventually vanishing .
a second reason is that previous work did not measure and examine the _ network community profile _ of cluster size vs. cluster quality .
* common bond vs. common identity communities : * dunbar s explanation aligns well with the common bond vs. common identity theory of group attachment @xcite from social psychology .
common identity theory makes predictions about people s attachment to the group as a whole , while common bond theory predicts people s attachment to individual group members .
the distinction between the two refers to people s different reasons for being in a group . because they like the group as a whole we get identity - based attachment , or because they like individuals in the group we get bond - based attachment .
anecdotally , bond - based groups are based on social interaction with others , personal knowledge of them , and interpersonal attraction to them . on the other hand ,
identity - based groups are based on common identity of its members , _
e.g. _ , liking to play a particular on - line game , contributing to wikipedia , etc .
it has been noted that bond communities tend to be smaller and more cohesive @xcite , as they are based on interpersonal ties , while identity communities are focused around common theme or interest .
see @xcite for a very good review of the topic . translating this to our context , the bond vs. identity communities mean that small , cohesive and well - separated communities are probably based on common bonds , while bigger groups may be based on common identity , and it is hard to expect such big communities to be well - separated or well - expressed in a network sense .
this further means the transition between common bond ( _ i.e. _ , maintaining close personal ties ) and common identity ( _ i.e. _ , sharing a common interest or theme ) occurs at around one hundred nodes .
it seems that at this size the cost of maintaining bond ties becomes too large and the group either dies or transitions into a common identity community .
it would be very interesting as a future research topic to explore differences in community network structure as the community grows and transitions from common bond to common identity community .
* edge semantics : * another explanation could be that in small , carefully collected networks , the semantics of edges is very precise while in large networks we know much less about each particular edge , _
e.g. _ , especially when online people have very different criteria for calling someone a friend .
traditionally social scientists through questionnaires `` normalized '' the links by making sure each link has the same semantics / strength . * evidence in previous work : * there has also been some evidence that hints towards the findings we make here . for example , clauset _ et al . _
@xcite analyzed community structure of a graph related to the amazonallprod , and they found that around @xmath123 of the nodes belonged to the largest `` miscellaneous '' community .
this agrees with the typical size of the network core , and one could conclude that the largest community they found likely corresponds to the intermingled core of the network , and most of the rest of the communities are whisker - like .
in addition , recently there have been several works hinting that the network communities subject is more complex than it seems at the first sight . for example
, it has been found that even random graphs can have good modularity scores @xcite . intuitively , random graphs have no community structure , but there can still exist sets of nodes with good community scores , at least as measured by modularity ( due to random fluctuations about the mean ) .
moreover , very recently a study of robustness of community structure showed that the canonical example of presence of community structure in networks @xcite may have no significant community structure @xcite . * more general thoughts : * our work also raises an important question of what is a natural community size and whether larger communities ( in a network sense ) even exist .
it seems that when community size surpasses some threshold , the community becomes so diverse that it stops existing as a traditionally understood `` network community . '' instead , it blends in with the network , and intuitions based on connectivity and cuts seem to fail to identify it .
approaches that consider both the network structure and node attribute data might help to detect communities in these cases . also , conductance seems like a very reasonable measure that satisfies intuition about community quality , but we have seen that if one only worries about conductance , then bags of whiskers and other internally disconnected sets have the best scores .
this raises interesting questions about cluster compactness , regularization , and smoothness : what is a good definition of compactness , what is the best way to regularize these noisy networks , and how should this be connected to the notion of community separability ?
a common assumption is that each node belongs to exactly one community .
our approach does not make such an assumption . instead ,
for each given size , we independently find best set of nodes , and `` communities '' of different sizes often overlap .
as long there is a boundary between communities ( even if boundaries overlap ) , cut- and edge - density- based techniques ( like modularity and conductance ) may have the opportunity to find those communities .
however , it is the absence of clear community boundaries that makes the ncp plot go upwards .
a great deal of work has been devoted to finding communities in large networks , and much of this has been devoted to formalizing the intuition that a community is a set of nodes that has more and/or better intra - linkages between its members than inter - linkages with the remainder of the network . very relevant to our work
is that of kannan , vempala , and vetta @xcite , who analyze spectral algorithms and describe a community concept in terms of a bicriterion depending on the conductance of the communities and the relative weight of inter - community edges .
flake , tarjan , and tsioutsiouliklis @xcite introduce a similar bicriterion that is based on network flow ideas , and flake _ et al . _
@xcite defined a community as a set of nodes that has more intra - edges than inter - edges .
similar edge - counting ideas were used by radicchi _
et al . _
@xcite to define and apply the notions of a strong community and a weak community . within the `` complex networks '' community , girvan and newman @xcite proposed an algorithm that used `` centrality '' indices to find community boundaries . following this ,
newman and girvan @xcite introduced _
modularity _ as _ a posteriori _ measure of the strength of community structure .
modularity measures inter- ( and not intra- ) connectivity , but it does so with reference to a randomized null model .
modularity has been very influential in the recent community detection literature @xcite , and one can use spectral techniques to approximate it @xcite .
on the other hand , guimer , sales - pardo , and amaral @xcite and fortunato and barthlemy @xcite showed that random graphs have high - modularity subsets and that there exists a size scale below which communities can not be identified . in part as a response to this , some recent work has had a more statistical flavor @xcite . in light of our results , this work seems promising , both due to potential `` overfitting '' issues arising from the extreme sparsity of the networks , and also due to the empirically - promising regularization properties exhibited by local spectral methods .
we have made extensive use of the local spectral algorithm of andersen , chung , and lang @xcite .
similar results were originally proven by spielman and teng @xcite , who analyzed local random walks on a graph ; see chung @xcite for an exposition of the relationship between these methods .
andersen and lang @xcite showed that these techniques can find ( in a scalable manner ) medium - sized communities in very large social graphs in which there exist reasonably well - defined communities . in light of our results ,
such methods seem promising more generally . other recent work that has focused on developing local and/or near - linear time heuristics for community detection
include @xcite .
in addition to this work we have cited , there exists work which views communities from a very different perspective .
for example , kumar _
et al . _
@xcite view communities as a dense bipartite subgraph of the web ; gibson , kleinberg , and raghavan @xcite view communities as consisting of a core of central authoritative pages linked together by hub pages ; hopcroft _ et al . _
@xcite are interested in the temporal evolution of communities that are robust when the input data to clustering algorithms that identify them are moderately perturbed ; and palla _ et al . _
@xcite view communities as a chain of adjacent cliques and focus on the extent to which they are nested and overlap
. the implications of our results for this body of work remain to be explored . in this subsection
, we describe the relationship between our work and recent work with similar flavor in graph partitioning , algorithms , and graph theory .
recent work has focused on the expansion properties of power law graphs and the real - world networks they model .
for example , mihail , papadimitriou , and saberi @xcite , as well as gkantsidis , mihail , and saberi @xcite , studied internet routing at the level of autonomous systems ( as ) , and showed that the preferential attachment model and a random graph model with power law degree distributions each have good expansion properties if the minimum degree is greater than @xmath17 or @xmath18 , respectively .
this is consistent with the empirical results , but as we have seen the as graphs are quite unusual , when compared with nearly every other social and information network we have studied . on the other hand , estrada has made the observation that although certain communication , information , and biological networks have good expansion properties , social networks do not @xcite .
this is interpreted as evidence that such social networks have good small highly - cohesive groups , a property which is not attributed to the biological networks that were considered . from the perspective of our analysis , these results are interesting since it is likely that these small highly - cohesive groups correspond to sets near the global minimum of the network community profile plot .
reproducing deep cuts was also a motivation for the development of the geometric preferential attachment models of flaxman , frieze , and vera @xcite .
note , however , that the deep cuts they obtain arise from the underlying geometry of the model and thus are nearly bisections .
consider also recent results on the structural and spectral properties of very sparse random graphs . recall that the @xmath124 random graph model @xcite consists of those graphs on @xmath125 nodes , in which there is an edge between every pair vertices with a probability @xmath126 , independently .
recall also that if @xmath127 , then a typical graph in @xmath124 has a giant component , _
i.e. _ , connected subgraph consisting of a constant fraction of the nodes , but the graph is not fully connected @xcite .
( if @xmath128 , the a typical graph is disconnected and there does not exist a giant component , while if @xmath129 , then a typical graph is fully connected . ) as noted , _
e.g. _ , by feige and ofek @xcite , this latter regime is particularly difficult to analyze since with fairly high probability there exist vertices with degrees that are much larger than their expected degree . as reviewed in section [ sxn : models : sparse_gw ] , however , this regime is not unlike that in a power law random graph in which the power law exponent @xmath130 @xcite .
chakrabarti _ et al . _
@xcite defined the `` min - cut '' plot which has similarities with our ncp plot .
they used a different approach in which a network was recursively bisected and then the quality of the obtained clusters was plotted against as a function of size ; and the `` min - cut '' plots were only used as yet - another statistic to test when assessing how realistic are synthetically generated graphs .
note , however , that the `` min - cut '' plots have qualitatively similar behavior to our ncp plots , _
i.e. _ , they initially decrease , reach a minimum , and then increase . of particular interest to us
are recent results on the mixing time of random walks in this @xmath127 regime of the @xmath124 ( and the related @xmath73 ) random graph model .
benjamini , kozma , and wormald @xcite and fountoulakis and reed @xcite have established rapid mixing results by proving structural results about these very sparse graphs . in particular
, they proved that these graphs may be viewed as a `` core '' expander subgraph , whose deletion leaves a large number of `` decorations , '' _
i.e. _ , small components such that a bounded number are attached to any vertex in the core .
the particular constructions in their proofs is complicated , but they have a similar flavor to the core - and - whiskers structure we have empirically observed .
similar results were observed by fernholz and ramachandran @xcite , whose analysis separately considered the @xmath17-core of these graphs and then the residual pieces .
they show that a typical longest shortest path between two vertices @xmath131 and @xmath32 consists of a path of length @xmath48 from @xmath131 to the @xmath17-core , then a path of length @xmath48 across the @xmath17-core , and finally a path of length @xmath48 from the @xmath17-core to @xmath32 . again , this is reminiscent of the core - and - whiskers properties we have observed . in all these cases , the structure is very different than traditional expanders @xcite , which we also empirically observe .
eigenvalues of power law graphs have also been studied by mihail and papadimitriou @xcite , chung , lu , vu @xcite , and flaxman , frieze , and fenner @xcite .
we investigated statistical properties of community - like sets of nodes in large real - world social and information networks .
we discovered that community structure in these networks is very different than what we expected from the experience with small networks and from what commonly - used models would suggest . in particular , we defined a _ network community profile plot ( ncp plot ) _ , and we observed that good network communities exist only up to a size scale of @xmath1 nodes .
this agrees well with the observations of dunbar . for size scales above @xmath1 nodes ,
the ncp plot slopes upwards as the conductance score of the best possible set of nodes gets gradually worse and worse as those sets increase in size .
thus , if the world is modeled by a sparse `` interaction graph '' and if a density - based notion such as conductance is an appropriate measure of community quality , then the `` best '' possible `` communities '' in nearly every real - world network we examined gradually gets less and less community - like and instead gradually `` blends in '' with the rest of the network , as the purported communities steadily grow in size .
although this suggests that large networks have a _ core - periphery _ or _ jellyfish _ type of structure , where small `` whiskers '' connect themselves into a large dense intermingled network `` core , '' we also observed that the `` core '' itself has an analogous core - periphery structure .
none of the commonly - used network generation models , including preferential - attachment , copying , and hierarchical models , generates networks that even qualitatively reproduce this community structure property .
we found , however , that a model in which edges are added recursively , via an iterative `` forest fire '' burning mechanism , produces remarkably good results .
our work opens several new questions about the structure of large social and information networks in general , and it has implications for the use of graph partitioning algorithms on real - world networks and for detecting communities in them .
we thank reid andersen , christos faloutsos and jon kleinberg for discussions , lars backstrom for data , and arpita ghosh for assistance with the proof of theorem [ thm : maingw ] .
r. andersen , f.r.k .
chung , and k. lang .
local graph partitioning using pagerank vectors . in _
focs 06 : proceedings of the 47th annual ieee symposium on foundations of computer science _ , pages 475486 , 2006 .
s. arora , e. hazan , and s. kale .
@xmath132 approximation to sparsest cut in @xmath133 time . in _
focs 04 : proceedings of the 45th annual symposium on foundations of computer science _ , pages 238247 , 2004 .
m. babenko , j. derryberry , a. goldberg , r. tarjan , and y. zhou .
experimental evaluation of parametric max - flow algorithms . in _
wea 07 : proceedings of the 6th workshop on experimental algorithms _ , pages 256269 , 2007 .
l. backstrom , d. huttenlocher , j. kleinberg , and x. lan .
group formation in large social networks : membership , growth , and evolution . in _
kdd 06 : proceedings of the 12th acm sigkdd international conference on knowledge discovery and data mining _ , pages 4454 , 2006 .
a. broder , r. kumar , f. maghoul , p. raghavan , s. rajagopalan , r. stata , a. tomkins , and j. wiener .
graph structure in the web . in _
www 00 : proceedings of the 9th international conference on world wide web _ , pages 309320 , 2000 .
cherkassky and a.v .
goldberg . on implementing push - relabel method for the maximum flow problem . in _
ipco 95 : proceedings of the 4th international ipco conference on integer programming and combinatorial optimization _ , pages 157171 , 1995 .
a. fabrikant , e. koutsoupias , and c.h .
heuristically optimized trade - offs : a new paradigm for power laws in the internet . in _
icalp 02 : proceedings of the 29th international colloquium on automata , languages and programming _ , pages 110122 , 2002 .
m. faloutsos , p. faloutsos , and c. faloutsos . on power - law relationships of the internet topology . in _
sigcomm 99 : proceedings of the conference on applications , technologies , architectures , and protocols for computer communication _ , pages 251262 , 1999 .
flake , s. lawrence , and c.l .
efficient identification of web communities . in _
kdd 00 : proceedings of the 6th acm sigkdd international conference on knowledge discovery and data mining _ ,
pages 150160 , 2000 .
flaxman , a.m. frieze , and j. vera . a geometric preferential attachment model of networks .
in _ waw 04 : proceedings of the 3rd workshop on algorithms and models for the web - graph _ , pages 4455 , 2004 .
flaxman , a.m. frieze , and j. vera . a geometric preferential attachment model of networks ii . in _ waw 07 : proceedings of the 5th workshop on algorithms and models for the web - graph _ ,
pages 4155 , 2007 . c. gkantsidis , m. mihail , and a. saberi .
conductance and congestion in power law graphs . in _ proceedings of the 2003 acm sigmetrics international conference on measurement and modeling of computer systems _ , pages 148159 , 2003 .
j. hopcroft , o. khan , b. kulis , and b. selman .
natural communities in large linked networks . in _
kdd 03 : proceedings of the 9th acm sigkdd international conference on knowledge discovery and data mining _ , pages 541546 , 2003 .
r. kumar , j. novak , and a. tomkins .
structure and evolution of online social networks . in _
kdd 06 : proceedings of the 12th acm sigkdd international conference on knowledge discovery and data mining _ , pages 611617 , 2006 .
r. kumar , p. raghavan , s. rajagopalan , d. sivakumar , a. tomkins , and e. upfal .
stochastic models for the web graph . in _
focs 00 : proceedings of the 41st annual symposium on foundations of computer science _ , pages 5765 , 2000 .
k. lang and s. rao . a flow - based method for improving the expansion or conductance of graph cuts . in _
ipco 04 : proceedings of the 10th international ipco conference on integer programming and combinatorial optimization _ , pages 325337 , 2004 .
t. leighton and s. rao .
an approximate max - flow min - cut theorem for uniform multicommodity flow problems with applications to approximation algorithms . in _
focs 88 : proceedings of the 28th annual symposium on foundations of computer science _ ,
pages 422431 , 1988 .
j. leskovec , d. chakrabarti , j. kleinberg , and c. faloutsos .
realistic , mathematically tractable graph generation and evolution , using kronecker multiplication . in _
ecml / pkdd 05 : proceedings of the european international conference on principles and practice of knowledge discovery in databases _ , pages 133145 , 2005 .
j. leskovec , j. kleinberg , and c. faloutsos .
graphs over time : densification laws , shrinking diameters and possible explanations . in _
kdd 05 : proceeding of the 11th acm sigkdd international conference on knowledge discovery in data mining _ , pages 177187 , 2005 .
j. leskovec , k.j .
lang , a. dasgupta , and m.w .
statistical properties of community structure in large social and information networks . in _ www 08 : proceedings of the 17th international conference on world wide web _ , pages 695704 , 2008 .
d. lusseau , k. schneider , o.j .
boisseau , p. haase , e. slooten , and s.m .
the bottlenose dolphin community of doubtful sound features a large proportion of long - lasting associations .
, 54:396405 , 2003 .
spielman and s .- h . teng .
spectral partitioning works : planar graphs and finite element meshes . in _ focs 96 : proceedings of the 37th annual ieee symposium on foundations of computer science _ , pages 96107 , 1996 .
spielman and s .- h .
nearly - linear time algorithms for graph partitioning , graph sparsification , and solving linear systems . in _
stoc 04 : proceedings of the 36th annual acm symposium on theory of computing _ ,
pages 8190 , 2004 . | a large body of work has been devoted to defining and identifying clusters or communities in social and information networks , _
i.e. _ , in graphs in which the nodes represent underlying social entities and the edges represent some sort of interaction between pairs of nodes .
most such research begins with the premise that a community or a cluster should be thought of as a set of nodes that has more and/or better connections between its members than to the remainder of the network . in this paper
, we explore from a novel perspective several questions related to identifying meaningful communities in large social and information networks , and we come to several striking conclusions . rather than defining a procedure to extract sets of nodes from a graph and then attempt to interpret these sets as a `` real '' communities , we employ approximation algorithms for the graph partitioning problem to characterize as a function of size the statistical and structural properties of partitions of graphs that could plausibly be interpreted as communities .
in particular , we define the _ network community profile plot _ , which characterizes the `` best '' possible community according to the conductance measure over a wide range of size scales .
we study over @xmath0 large real - world networks , ranging from traditional and on - line social networks , to technological and information networks and web graphs , and ranging in size from thousands up to tens of millions of nodes .
our results suggest a significantly more refined picture of community structure in large networks than has been appreciated previously .
our observations agree with previous work on small networks , but we show that large networks have a very different structure .
in particular , we observe tight communities that are barely connected to the rest of the network at very small size scales ( up to @xmath1 nodes ) ; and communities of size scale beyond @xmath1 nodes gradually `` blend into '' the expander - like core of the network and thus become less `` community - like , '' with a roughly inverse relationship between community size and optimal community quality .
this observation agrees well with the so - called dunbar number which gives a limit to the size of a well - functioning community . however , this behavior is not explained , even at a qualitative level , by any of the commonly - used network generation models . moreover , it is exactly the opposite of what one would expect based on intuition from expander graphs , low - dimensional or manifold - like graphs , and from small social networks that have served as testbeds of community detection algorithms .
the relatively gradual increase of the network community profile plot as a function of increasing community size depends in a subtle manner on the way in which local clustering information is propagated from smaller to larger size scales in the network .
we have found that a generative graph model , in which new edges are added via an iterative `` forest fire '' burning process , is able to produce graphs exhibiting a network community profile plot similar to what we observe in our network datasets . ' '' '' ' '' '' |
dispersion of particles in stochastic and turbulent flows is a key fundamental problem @xcite with applications in a huge number of disciplines going from atmospheric and ocean sciences @xcite , to environmental sciences @xcite , chemical engineering and astrophysics @xcite . at high reynolds numbers ,
molecular diffusion makes a negligible contribution to spatial transport , and so turbulence dominates not only the transport of momentum , but also that of temperature , humidity , salinity and of any chemical species or concentration field .
mixing can be approached from an eulerian point of view , studying the spatial and temporal evolution of a concentration field @xcite , and also by using a lagrangian approach in terms of the relative dispersion of pairs of particles @xcite . notwithstanding the enormous literature on the topic , a stochastic model for particle trajectories in turbulent flows
whose basic assumptions are fully justified is yet to come @xcite .
the modeling of pair dispersion for tracers was pioneered in @xcite , where a _
locality assumption _ was introduced , see also @xcite for a recent historical review . in a modern language ,
richardson s approach is built up in analogy with diffusion , replacing molecular fluctuations with turbulent fluctuations , acting differently at different scales .
hence , in a turbulent flow , diffusivity is enhanced because the instantaneous separation rate depends on the local turbulent conditions encountered by pairs along their path : @xmath4 where @xmath5 is the amplitude of the separation vector between the two particles , @xmath6 , and @xmath7 is a scalar eddy - diffusivity . for the eddy - diffusivity approach to be valid , separations @xmath8 have to be chosen larger than the initial ones , @xmath9 , and smaller than the integral scale of the flow , @xmath10 .
moreover , time lags have to be large enough , so that the memory of the initial separation is lost .
+ in the light of kolmogorov 1941 theory , ( see @xcite , the scalar eddy - diffusivity can be modeled as follows : @xmath11 where @xmath12 is the eulerian longitudinal velocity difference along the direction of particle separation @xmath13 , @xmath14 , @xmath15 is a dimensionless constant , and @xmath16 is the rate of kinetic energy dissipation in the flow . in the above equation , @xmath17 is the correlation time of the lagrangian velocity differences at scale @xmath13 , @xmath18 ^ 2\rangle } \int_0^t \langle \delta_r { \bf u}({\bf r}(t),t ) \cdot \delta_r { \bf u}({\bf r}(s),s ) \rangle\ , ds\ , .
\label{eq : tau}\ ] ] in the inertial range of scales , by dimensional considerations , we can write @xmath19 , and @xmath20 , from which the celebrated richardson s @xmath21 law of eqn .
( [ eq:43 ] ) follows . as a consequence , eqn .
( [ eq:1 ] ) predicts a super - diffusive growth for the particle separation : @xmath22 and the dependence on the initial conditions is quickly forgotten .
in fact , when released into a fluid flow , tracer pairs separate ballistically at short time lags , _ ' a la _ batchelor @xcite , keeping memory of their initial longitudinal velocity difference , @xmath23 , up to time lags of the order of @xmath24 .
only later on , richardson s super - diffusive regime follows .
+ richardson s approach is exact if we assume that tracers disperse in a @xmath25correlated in time velocity field .
in such a case , the probability density function ( pdf ) of observing two tracers at separation @xmath8 at time @xmath26 , @xmath27 , satisfies a fokker - planck diffusive equation with a space dependent diffusivity coefficient , @xmath7 @xcite : @xmath28\,,\ ] ] where @xmath7 is a function of the velocity correlation evaluated at the current separation , only .
+ the richardson equation ( [ eq : rich ] ) with initial condition @xmath29 can be solved , see e.g. , @xcite , and the solution has an asymptotic , large time form ( independent of the initial condition @xmath9 and @xmath30 ) of the kind : @xmath31 \,,\ ] ] where @xmath32 is a normalization constant .
the richardson pdf is perfectly self - similar , so that all positive moments behave according to the dimensional law , @xmath33 .
+ there are many reasons for which the richardson distribution can not exactly describe the behaviour of tracer pairs in real flows .
the most important ones are : ( i ) the nature of the temporal correlations in the fluid flow @xcite ; ( ii ) the non - gaussian fluctuations of turbulent velocities @xcite ; ( iii ) the small - scale effects induced by the dissipation sub - range , and ( iv ) the large - scale effects induced by the flow correlation length .
these last two are of course connected to finite reynolds - number effects . + it is worth noticing that formally any diffusion coefficient of the form @xmath34 , with @xmath35 , is compatible with the @xmath36 law , however different results
would then be obtained for the functional form of @xmath37 @xcite . + ( red ) and heavy particles with @xmath38 ( blue ) , simultaneously emitted from a source of size @xmath39 .
trajectories are recorded from the emission time , up to the time @xmath40 after the emission.,width=377 ] since richardson s seminal work , pair dispersion has been addressed in a large number of experimental and numerical studies , in the @xmath41 inverse energy cascade @xcite and in the direct enstrophy cascade ( see e.g. , @xcite ) , as well as in the @xmath42 direct energy cascade @xcite , in convective turbulent flows @xcite and in synthetic flows @xcite . in @xcite , kinematic and direct numerical simulations
have also been used to compare forward and backward relative dispersion in three dimensional turbulent flows .
comprehensive reviews on the topic can be found in @xcite,@xcite and @xcite . + despite the huge amount of theoretical , numerical and experimental works devoted to this issue , it is fair to say that at the moment there is neither a clear consensus in favour of the richardson s approach , nor a clear disproof . the main practical reason is due to the fact that the predictions -if correct- , are applicable to tracer pairs whose evolution has been for all times in the inertial range of scales : @xmath43\ , , \label{eq : ic}\ ] ] where @xmath44 is the viscous scale of the turbulent flow .
in other words , we should record tracer dispersion at space and time scales _ unaffected _ by viscous or integral scale effects .
this is of course a strong requirement which is particularly difficult to match in any experimental or numerical test because of the natural limitations in the accessible reynolds number @xmath45 , i.e. , in the scale separation range @xmath46 .
moreover the viscous scale itself @xmath44 and the stretching rate at this scale are strongly fluctuating quantities in turbulent flows @xcite , causing further difficulties when pair statistics must be limited to a pure inertial range behaviour .
it is worth noticing that a possible way out is to resort to exit - time statistics @xcite , which will be discussed in section [ sec : exit ] .
+ to avoid viscous effects on the pair dispersion statistics , it is also common to study pairs whose initial separation is well inside the inertial range , @xmath47 , paying the price to be dominated for long times by the initial condition and therefore mostly accessing the batchelor regime @xcite .
alternatively , numerical simulations of particles evolving in stochastically generated velocity fields are a useful tool to describe ( possibly non gaussian and non self - similar ) inertial range pair dispersion @xcite .
note however that kinematic simulations might lead to a mean - square separation of the particle pairs with a power law different from the richardson s law @xcite .
+ for the reasons ( i)-(ii ) listed above , it is well possible that even in a infinite reynolds number limit , the richardson s prediction may turn out to be wrong
. effects of time correlations have been discussed by many authors @xcite , in connection to the problem of the formally admissible _ infinite propagation speed _ present in any diffusive approach _
la _ fokker - planck @xcite , and also in relation to the possible non - markovian nature of the lagrangian position and velocity process @xcite .
+ summarising , it is extremely important to clarify with high accuracy if the deviations from richardson s theory observed in laboratory experiments and numerical simulations , at finite reynolds numbers , are due to sub - leading effects associated to the lack of scale - separation or not . in the latter case , it means that they would survive even in the @xmath48 limit .
this is the aim of the research presented in this paper .
+ we use direct numerical simulations of isotropic and homogeneous three dimensional turbulence at @xmath0 , seeded with an unprecedented number of particles ( emitted from localised sources in different locations inside the flow ) , in order to increase the total number of pairs starting with an initially small separation and to minimise local anisotropy and non - homogeneity . + we present results for both tracers and point - like heavy particles , without feedback on the flow as in the original problem attacked by richardson . when particles have inertia , new scenarios arise @xcite , because of the non homogeneous spatial distribution @xcite and the very intermittent nature of relative velocity increments characterised by the presence of quasi - singularities @xcite .
+ not surprisingly , and in the absence of a theory , empirical observations are in this case even less stringent , also because of the need to specify the initial distributions of both particle positions and velocities .
two types of experiments can be done with inertial particles .
the first consists of studying relative dispersion as a function of the distribution of initial separations only . in practice
, inertial particles are allowed to reach their stationary spatial and velocity distributions inside a bounded volume ( stationary distribution on a fractal dynamical attractor in phase - space ) , after which their dispersion properties are measured , conditioning on the initial distance @xcite .
the second consists in directly injecting inertial particles in the flow , with prescribed initial velocity and separation distributions .
the first protocol is more relevant to study relative dispersion properties in connection with spatial clustering , particularly effective at small stokes numbers ( e.g. , the spatial preferential concentration and trapping in coherent structures in the flow ) , and in connection with caustics , strongly modifying the relative velocities at large stokes numbers @xcite .
the second protocol is more relevant in geophysical and industrial applications , where transient behaviours are crucial as in the case of volcanic eruptions , leakages of contaminants , or pollutant emissions .
+ in this paper , we are interested in the latter case , for which we designed the simplest procedure of having inertial particles emitted in the same positions and with the same velocities of the tracers ( see figure [ fig_two_stokes ] ) .
this choice turns out to be optimal to better understand the statistics of tracers also , as it will become clearer in the sequel .
the main results of the paper can be briefly anticipated .
first , we quantify the importance of viscous - scale fluctuations for tracer separations , showing that they easily affect the separation evolution for time scales much larger than what predicted by the dimensional batchelor time , @xmath49 .
this effect is so huge that , in current experiments , it spoils any possibility to assess tracer scaling properties in the _ expected _ inertial range of time scales . to overcome this problem
, we suggest that heavy pairs can be used as dispersing particles that are able to dynamically filter out viscous - time fluctuations . by measuring pair dispersion of heavy pairs at different degree of inertia , we are able to observe a much clearer convergence towards an inertial - range regime for the relative separation .
we interpret this result as an indication of the existence of infinite reynolds - number corrections to the richardson s prediction , as expected on the basis of a eulerian - lagrangian multifractal approach @xcite , yet never observed in real data . the multifractal approach ( mf ) is also used to determine a set of specific correlation moments involving powers of the pair separation distance and relative velocity of tracer pairs , which are expected to be statistically invariant along lagrangian trajectories in the inertial range @xcite .
+ the paper is organised as follows . in section
[ sec : numerics ] , we present the details of our numerical study . in section [ sec : direct ] , we introduce the relative dispersion statistics for both tracers and heavy particles . in sub - section
[ sec : direct - sub ] , we discuss the effects due to finite reynolds numbers and those induced by the fluctuations of the viscous scale . in sub - section
[ sec : pdf ] , we present the numerical data for the probability density functions . in connection with the issue of intermittent corrections , in sub - section [ sec : direct - multi ] , we briefly review the multifractal prediction for the pair dispersion of tracers developed by @xcite , and we test it against our data . in section [ sec : exit ] , we discuss the results concerning the exit - time statistics , probing pairs which separate very slowly .
the fluid is described by the navier - stokes equations for the velocity field @xmath50 @xmath51 the statistically homogeneous and isotropic external forcing @xmath52 injects energy in the first low - wavenumber shells , by keeping constant in time their spectral content @xcite .
the kinematic viscosity @xmath53 is chosen such that the kolmogorov length scale is @xmath54 , where @xmath55 is the grid spacing , so that a good resolution of the small - scale velocity dynamics is obtained .
.parameters of the numerical simulations : taylor - scale based reynolds number @xmath45 , grid resolution @xmath56 , kolmogorov length scale @xmath44 in simulation units ( su ) , grid spacing @xmath57 ( su ) , mean energy dissipation @xmath16 ( su ) , kinematic viscosity @xmath53 ( su ) , kolmogorov time - scale @xmath58 ( su ) , large - scale eddy turnover time @xmath59 ( in units of @xmath60 ) , root - mean - square velocity @xmath61 ( su ) , @xmath62 number of trajectories of inertial particles emitted for each stokes number @xmath63 from each local source and for each puff , @xmath64 number of sources in the flow , @xmath65 total number of particle pairs emitted in all simulations per stokes number ( @xmath66 runs with @xmath67 local sources , each emitting @xmath68 puffs ) , @xmath69 maximal temporal length of particle trajectories ( in units of @xmath60 ) . [ cols="^,^,^,^,^,^,^,^,^,^,^,^,^ " , ] . [
tab : param ] the numerical domain is cubic , with periodic boundary conditions in the three space directions ; a fully dealiased pseudo - spectral algorithm with second - order adam - bashforth time - stepping is used .
we performed a series of direct numerical simulations with resolution of @xmath70 grid points and reynolds number at the taylor scale @xmath71 .
the flow is seeded with bunches of tracers and heavy particles , emitted in different fluid locations to reduce the large scale correlations , and local inhomogeneous / anisotropic effects .
each bunch is emitted within a small region of space , of kolmogorov scale size , in puffs of @xmath72 particles each , for tracers and heavy particles . in a single run , there are @xmath67 of such point sources , releasing about @xmath68 puffs with a frequency comparable with the inverse of the kolmogorov time .
we collected statistics over @xmath66 different runs . as a result , we follow a total amount of @xmath73 pairs .
+ the heavy particles are assumed to be of size much smaller than the kolmogorov scale of the flow and with a negligible reynolds number relative to the particle size . in this limit
, their equations of motion take the simple form : @xmath74\,,\ ] ] where the dots denote time derivatives .
the particle position and velocity vectors are @xmath75 and @xmath76 , respectively ; @xmath77 is the eulerian fluid velocity evaluated at the particle position .
the particle response time is @xmath78 .
the flow kolmogorov time scale , appearing in the definition of the stokes number , @xmath79 , is @xmath80 .
particle - particle interactions and the feedback of the particles onto the flow are here neglected . in this work ,
we show results for the following set of stokes numbers : @xmath81 and @xmath82 .
additional details of the runs can be found in table [ tab : param ] .
in figure [ fig_two_stokes ] , we compare the time evolution up to a time of the order of the large scale eddy turn over time , @xmath59 , of a bunch of tracers and a bunch of heavy particles with @xmath83 , both emitted in a region of strong shear .
+ , in log - lin scale , of the mean - square separation of heavy particle pairs , normalised with the curve for tracer pairs.,width=377 ] at a time lag roughly equal to @xmath84 after the emission , it can be observed an abrupt transition in the particle dispersion of tracers ( red in the online version ) , occurring when most of the pairs reaches a relative distance of the order of @xmath85 .
later on , we again notice the presence of many pairs with mutual separations much larger than the mean one .
the trajectories of the heavy particles ( blue in the online version ) show a different evolution . after the emission , they tend to remain at a mutual distance of the order of @xmath44 for a very long time , thus dispersing much less .
because they respond to fluid fluctuations with a time lag of the order of their stokes time , they tend to keep their initial velocity unchanged before relaxing onto the underlying fluid velocities . as a result ,
inertial particles behave as if their batchelor time was much longer than that of tracers ( in our dns the latter is small , @xmath86 , because the source is strongly localised ) .
our main observation here is that the larger the stokes number , the longer is the filtering time that heavy particles apply to the local stretching properties of the carrying fluid : since inertia is moderate in the present experiment , this fact will allow us to have a very effective method to reduce effects from viscous scale fluctuations in the particles statistics and to better disentangle inertial range properties in the pair dispersion evolution . +
this is more quantitatively understood in figure [ fig_r2_compared ] , where the second order moment of the relative separation for tracers , @xmath87 , and heavy particles , @xmath83 , are plotted .
heavy particles tend to separate less since they are unaffected by turbulent fluctuations up to time and spatial scales large enough for their inertia to become sub - dominant with respect to the underlying turbulent fluctuations . on a dimensional ground
, such a scale can be easily estimated to be of the order of @xmath88 when @xmath89 @xcite .
let us notice that in this respect the choice of the initially prescribed velocity distribution plays a key role , since inertial particles are emitted with a velocity equal to that of the underlying fluid .
the quantity which is better suited to quantify this observation is the ratio between relative separations : @xmath90 in the present experiment we observe @xmath91 at any scale and time , as shown in the lower inset of figure [ fig_r2_compared ] . by prescribing an initial distribution equal to the stationary pdf of heavy particle velocity increments , the trend would have been the opposite , with inertial particles at short times separating much faster than tracers , i.e. @xmath92 up to a separation @xmath93 @xcite . in the main body of figure [ fig_r2_compared ]
, we notice that for both tracers and heavy particles , and for time lags large enough , the separation curves tend toward a @xmath94 richardson - like behaviour , but without showing any clear scaling , as also measured by the compensated plot in the upper inset of figure [ fig_r2_compared ] . .
right panel : simultaneous realisation of a tracer bunch , emitted from a different source , showing a much smaller dispersion .
both emissions are recorded up to the time @xmath95.,title="fig:",width=241 ] .
right panel : simultaneous realisation of a tracer bunch , emitted from a different source , showing a much smaller dispersion .
both emissions are recorded up to the time @xmath95.,title="fig:",width=241 ] to better appreciate the importance of viscous effects on pair dispersion , we show in figure [ fig_2_bunches ] , the time evolution of two different bunches of tracers emitted in different positions in the flow .
the initial size of the puffs is of the order of the viscous scale .
the bunch on the left is emitted in a region where the stretching rate has a typical value of the order of its root mean square , @xmath96 , while the bunch on the right is emitted in a region where the local stretching rate is unusually small . as a result
this second bunch separates with a much longer delay with respect to the average behaviour .
+ in figure [ fig_2_bunches_r2 ] , the mean squared separations measured for pairs belonging to each of these two bunches are compared with the pair separation averaged over the full statistics .
the bunch emitted in a region where the stretching rate assumes the typical value exits the viscous region in a time lag of the order of @xmath58 , and soon approaches the inertial range behaviour compatible with @xmath97 .
pairs belonging to the other bunch keep a small separation @xmath98 for a time lag up to @xmath99 , a time comparable to the integral time - scale @xmath59 , and never recover the @xmath94 scaling behaviour along the whole duration of our simulation .
the examples shown in figure [ fig_2_bunches_r2 ] are meant to represent the huge variations that affect pair separation statistics for time lags of the order of @xmath58 .
these variations are the hardest obstacle to assess pure _ inertial _ range properties in any experimental or numerical set - up , in addition to the challenge of following particle trajectories for a time lag long enough and in a volume large enough .
+ our idea here is to use heavy particles as _ smart _ passive , but dynamical , objects to filter out such huge viscous effects , without affecting the long - time and large - scale physics .
indeed , heavy particles are less affected by fluctuations of the local viscous scale , since they respond to the fluid with their stokes time ( see figure [ fig_two_stokes ] ) . moreover ,
heavy pairs experience a less fluctuating local stretching rate , as also measured by the distribution of the finite - time lyapunov exponents as a function of the stokes number @xcite .
finally , we recall that because of the injection choice here adopted , caustics in the heavy particles velocity distribution manifest themselves only at a later stage . .
data from the left panel are represented by ( @xmath100 ) ; while data from the right panel are represented by ( @xmath101 ) .
the continuous curve is the mean square separation averaged over the whole statistical database.,width=377 ] as for the heavy pairs dispersion , the different degrees of fluctuations are better quantified in figure [ fig_flat ] , where we show the ratio of the third and fourth order moments of the separation distribution along the particle trajectories , normalised to the second order one @xmath102 for convenience , we refer to @xmath103 and @xmath104 as generalised skewness and flatness , respectively . at the transition between the viscous and the inertial range of scales ,
i.e. , for @xmath105 , tracers and small stokes particles possess a generalised flatness @xmath106 .
this is the signature of an extremely intermittent distribution ( for a @xmath107squared distribution with three degrees of freedom , we would have @xmath108 , while for the richardson distribution @xmath109 ) .
moreover , it is evident that the bump displayed by the generalised flatness around @xmath110 is influenced by the behaviour at shorter times and influences the behaviour at much larger times .
in other words , it is the quantitative counterpart of the big variations shown in the two examples in figure [ fig_2_bunches_r2 ] .
+ the filtering effect of the particle stokes time is the reason for which heavy pairs initially possess a smaller flatness .
this is particularly evident for the @xmath83 case , which exhibits a smoother transition toward a scaling behaviour for @xmath111 , supporting the idea the inertia helps to remove viscous fluctuations from the physics of the inertial range . at time
lags @xmath112 , the generalised coefficients @xmath103 and @xmath104 for heavy particles with @xmath83 are larger than those of the tracer pairs .
this is probably the result of the relaxation onto the tracer power - law behaviour , which happens only at a later stage .
finally , it is clear from figure [ fig_flat ] that the data set at the moderate inertia of @xmath83 is less affected by the strong viscous bump at @xmath113 and that therefore promises to be the best candidate to test inertial range statistical properties .
and @xmath114 , respectively ; also plotted are the expected values for a @xmath115-squared distribution with three degrees of freedom , which are @xmath116 for the flatness and @xmath117 for the skewness.,width=453 ] the probability density functions @xmath37 are plotted in figure [ fig_pdf_naked ] , for different values of inertia @xmath118 and @xmath2 , and at different time lags after the emission from the source . in order to highlight the dynamics of those particles filling the left tail ( i.e. separating less than the average ) , for figure [ fig_pdf_naked ] we have selected pairs with initial separation @xmath119\eta$ ] @xcite .
for pairs with different inertia @xmath120 highlighting the left tails behaviour .
plots refer to times @xmath121 , @xmath122 and @xmath123 after the emission .
the power - law scaling @xmath124 , with @xmath125 , is plotted .
the power - law scaling @xmath126 is also reported : note that for @xmath127 the correlation dimension @xmath128 is @xmath129 , while for @xmath130 it is @xmath131 @xcite .
the two power laws @xmath126 for @xmath127 and @xmath130 are indistinguishable in the scale of the plot , hence we reported the slope for @xmath130 only .
( right panels ) : lin - log plot of the same separation pdfs , at the same time lags . for these pdfs pairs are selected with initial separation @xmath132 \eta$ ] . , title="fig:",width=238 ] for pairs with different inertia @xmath120 highlighting the left tails behaviour .
plots refer to times @xmath121 , @xmath122 and @xmath123 after the emission .
the power - law scaling @xmath124 , with @xmath125 , is plotted .
the power - law scaling @xmath126 is also reported : note that for @xmath127 the correlation dimension @xmath128 is @xmath129 , while for @xmath130 it is @xmath131 @xcite .
the two power laws @xmath126 for @xmath127 and @xmath130 are indistinguishable in the scale of the plot , hence we reported the slope for @xmath130 only .
( right panels ) : lin - log plot of the same separation pdfs , at the same time lags . for these pdfs pairs are selected with initial separation @xmath132 \eta$ ] . , title="fig:",width=238 ] for pairs with different inertia @xmath120 highlighting the left tails behaviour .
plots refer to times @xmath121 , @xmath122 and @xmath123 after the emission .
the power - law scaling @xmath124 , with @xmath125 , is plotted .
the power - law scaling @xmath126 is also reported : note that for @xmath127 the correlation dimension @xmath128 is @xmath129 , while for @xmath130 it is @xmath131 @xcite .
the two power laws @xmath126 for @xmath127 and @xmath130 are indistinguishable in the scale of the plot , hence we reported the slope for @xmath130 only .
( right panels ) : lin - log plot of the same separation pdfs , at the same time lags . for these pdfs pairs are selected with initial separation @xmath132 \eta$ ] .
, title="fig:",width=238 ] for pairs with different inertia @xmath120 highlighting the left tails behaviour .
plots refer to times @xmath121 , @xmath122 and @xmath123 after the emission .
the power - law scaling @xmath124 , with @xmath125 , is plotted .
the power - law scaling @xmath126 is also reported : note that for @xmath127 the correlation dimension @xmath128 is @xmath129 , while for @xmath130 it is @xmath131 @xcite .
the two power laws @xmath126 for @xmath127 and @xmath130 are indistinguishable in the scale of the plot , hence we reported the slope for @xmath130 only .
( right panels ) : lin - log plot of the same separation pdfs , at the same time lags . for these pdfs pairs are selected with initial separation @xmath132 \eta$ ] .
, title="fig:",width=238 ] for pairs with different inertia @xmath120 highlighting the left tails behaviour .
plots refer to times @xmath121 , @xmath122 and @xmath123 after the emission .
the power - law scaling @xmath124 , with @xmath125 , is plotted .
the power - law scaling @xmath126 is also reported : note that for @xmath127 the correlation dimension @xmath128 is @xmath129 , while for @xmath130 it is @xmath131 @xcite .
the two power laws @xmath126 for @xmath127 and @xmath130 are indistinguishable in the scale of the plot , hence we reported the slope for @xmath130 only .
( right panels ) : lin - log plot of the same separation pdfs , at the same time lags . for these pdfs pairs are selected with initial separation @xmath132 \eta$ ] . , title="fig:",width=238 ] for pairs with different inertia @xmath120 highlighting the left tails behaviour .
plots refer to times @xmath121 , @xmath122 and @xmath123 after the emission .
the power - law scaling @xmath124 , with @xmath125 , is plotted .
the power - law scaling @xmath126 is also reported : note that for @xmath127 the correlation dimension @xmath128 is @xmath129 , while for @xmath130 it is @xmath131 @xcite .
the two power laws @xmath126 for @xmath127 and @xmath130 are indistinguishable in the scale of the plot , hence we reported the slope for @xmath130 only .
( right panels ) : lin - log plot of the same separation pdfs , at the same time lags . for these pdfs pairs are selected with initial separation @xmath132 \eta$ ] .
, title="fig:",width=238 ] consider first the behaviour at large separations .
one clearly sees the effect anticipated earlier .
heavy particles tends to separate less .
the effect becomes less and less visible with time , because inertia is forgotten on a time scale roughly proportional to the stokes time .
pairs with @xmath83 , however , accumulate a delay in separation that is never recovered , even at large times @xmath133 .
+ for the left tails , associated to pairs that do not separate , the trend as a function of stokes is the opposite .
we recall that from classical arguments @xcite , we expect that the fraction of tracer pairs at a distance @xmath8 would behave as a power law @xmath124 , where @xmath134 is the space dimension , see eq .
( [ eq : pdfrich ] ) .
similarly , for heavy pairs , we expect to observe the scaling @xmath135 , where @xmath136 measures the spatial correlation dimension . at short
time lags , @xmath137 , the effect of inertia quickly appears and we measure a higher probability to observe pairs at very small separations : this happens because heavy pairs are less affected than tracers by intermittent events of anomalously slow separations , and hence rapidly populate the left tail of the distribution . as time goes on , @xmath138 , we observe that pairs with moderate inertia , namely @xmath127 and @xmath139 , clearly show the tendency to clusterise on a fractal set @xcite characterised by the spatial correlation dimension @xmath140 @xcite , where @xmath141 is the spatial dimension of the flow . , at time lags @xmath142 ( left panel ) and @xmath143 ( right panel ) , for different stokes numbers.,title="fig:",width=245 ] , at time lags @xmath142 ( left panel ) and @xmath143 ( right panel ) , for different stokes numbers.,title="fig:",width=245 ] differently , we clearly observe that the left tail of the tracer pdf superposes well with that of the largest stokes , @xmath83 : this is because the correlation dimension for this high stokes number is @xmath144 . at a later time lag , @xmath145
, it is very hard to detect a power - law scaling in the left tail , even with the large database of the present experiment : by this time lag , most of the pairs have reached larger separations . since pair dispersion takes place at finite reynolds numbers , it is clear that asymptotic power - law behaviours , see eq .
( [ eq : pdfrich ] ) , can be observed in a limited range of space and time scales , only .
to summarise , the observed power - law scalings at small separations reproduce the classical expectations for tracers and heavy pairs reported in @xcite , and based on the richardson s model . to detect intermittency effects ,
that we expect to be present due to tracer pairs that separate much less than the average , different observables are needed . this will be the object of the exit - time analysis in the last section .
+ things become more interesting when the pair separation distribution are plotted in dimensionless units . in figure
[ fig_pdf_normalized ] , we show the pdfs measured over the whole statistical database , as a function of the pair distance and also in terms of the normalised relative separation , @xmath146 it is important to notice that the differences as a function of the stokes number previously observed are fully reabsorbed once dimensionless quantities are used .
this supports a strong universality for large separations as a function of @xmath63 , at least up to the values here studied .
should the pdf data follow the richardson s prediction , we would see a perfect rescaling on a stretched exponential curve , for all times and all separations .
it is evident that this is not the case .
moreover , we stress that the most important departures from the richardson s prediction develop on the far right tails , i.e. for intense fluctuations due to pairs that separate much more than the average .
this fact was already observed using the same dataset by @xcite .
previous numerical and experimental studies were limited to events with a probability larger than @xmath147 ( see e.g. , @xcite , where departures from the richardson s prediction could not be unambiguously detected . + it must be noticed that the renormalisation in terms of the separation @xmath148 brings some extra difficulties in the interpretation of data .
indeed , since the mean squared separation , @xmath149 , is increasing with time , at large times it might well happen that the far right tails of the pdfs are completely dominated by large - scale effects , @xmath150 .
the opposite happens for the events close to the peak , which can be affected by viscous contributions , @xmath151 , for small times . in both cases ,
finite reynolds - number effects come into play . as a result
, these rescalings do not allow straightforward conclusions , neither to confirm neither to exclude the alleged departures from the distribution predicted by richardson in the infinite reynolds limit . + , selecting the pair distances to be in the range @xmath152 \eta$ ] .
left panel is for tracers , while the right panel is for heavy pairs with @xmath83 .
symbols are drawn for a subset of points only for clarity.,title="fig:",width=245 ] , selecting the pair distances to be in the range @xmath152 \eta$ ] .
left panel is for tracers , while the right panel is for heavy pairs with @xmath83 .
symbols are drawn for a subset of points only for clarity.,title="fig:",width=245 ] in figure [ fig_pdf_inertial ] we show the same data of figure [ fig_pdf_normalized ] , but conditioning the relative separation to belong to the inertial range of scales , @xmath153 .
a more coherent picture now emerges , since we note that ( i ) curves belonging to different time lags develop clearly non overlapping tails ; ( ii ) for times large enough a universal , stokes independent , regime seems to develop ; ( iii ) the rapid fall - off of the left tails of the pdfs for extreme separations disappears .
these observations suggest the possibility to identify inertial range statistical properties that show a reynolds - independent departure from the richardson s prediction and that can not be attributed to viscous or large - scale effects .
the presence of inertial - range deviations from the pure self - similar behaviour of pair separation statistics predicted by the richardson s approach should not be surprising , although never observed . in the @xmath42 direct energy cascade regime ,
anomalous scaling is measured both in the statistics of eulerian longitudinal and transverse velocity increments @xcite , and in the statistics of lagrangian velocity increments along single particle trajectories @xcite .
a simple argument predicts the presence of intermittent corrections in the high - order moments of the relative particles separation , @xmath154 @xcite .
the starting point is the exact relation for the moment of order @xmath155 of the pair separation , @xmath156 where @xmath12 is the velocity increment measured along the tracer pair trajectories , separated by a distance @xmath8 . here , for simplicity , we have neglected the tensorial structure and the time dependency of @xmath8 and @xmath12 is understood .
let us suppose that the above correlation can be estimated with fully eulerian quantities , then the multifractal approach @xcite could be employed to obtain : @xmath157 where @xmath158 is the probability to observe an eulerian velocity fluctuation at the scale @xmath8 with a local scaling exponent @xmath159 , @xmath160 , as a function of the spectrum of fractal dimension , @xmath161 .
in order to relate the above eulerian estimate with the lagrangian one in ( [ eq : rp ] ) , one can adopt the dimensional _ bridge _ relation supposing that lagrangian velocity fluctuations at separation @xmath5 are connected to eulerian spatial fluctuations at scale @xmath8 with @xmath162 .
this amounts to say that one can use the same fractal dimensions @xmath161 for the lagrangian and eulerian velocity statistics .
a quantitative support for this hypothesis has been made by a validation for one particle quantities against numerical and experimental results in @xcite .
the same argument applied to two - particle quantities allows to rewrite ( [ eq : multi ] ) as a function of the time lag , @xmath26 , along the particle separations , @xmath163 after time integration , a saddle point approximation can be used in the limit @xmath164 , when the smallest exponent dominates the integral @xcite : @xmath165 this is the multifractal theory for tracer pair separation statistics in the presence of an eulerian velocity field @xcite , whose multiscale statistics is described by the set of fractal dimension @xmath161 . to our knowledge ,
such a prediction has never been tested , either on experimental or on numerical data , except for the validation against stochastic velocity fields reported in @xcite .
note that the multifractal spectrum is in general non - linear in the order @xmath155 ; since by the @xmath166-law , the exponent of the third order longitudinal eulerian structure function must be unity , we must also have that @xmath167 @xcite . in figure [ fig_multi ]
, we compare the multifractal spectrum , @xmath168 , obtained by using two possible forms for @xmath161 ( extracted , respectively , from the longitudinal eulerian structure functions , @xmath169 , and from the transverse ones , @xmath170 ) in statistically homogeneous and isotropic @xmath42 turbulence .
such small uncertainty is to be considered as our prediction error on the shape of the function @xmath161 , which can not be deduced by first principles .
a more detailed discussion about this point can be found in @xcite . + ( @xmath101 ) , and from the scaling exponents of eulerian transversal structure functions , @xmath170 ( @xmath100 ) .
the continuous line is the dimensional richardson scaling , @xmath171.,width=340 ] to measure the scaling behaviours , we use the procedure known as extended self similarity ( ess ) @xcite , which amounts to study the relative scaling of moments with respect to a reference one , whose exponent is constrained by an exact relation . in figure [ fig_fit ] , the @xmath172th order moment of particle separation is compensated with the second order one : @xmath173 both for tracer pairs and for heavy pairs at @xmath83 . in the inertial range where the prediction of equation ( [ eq : multi_espo ] ) is expected to be valid
, we should observe a plateau .
it is evident that the multifractal prediction works better than the dimensional one , suggesting that the approach goes in the right direction .
however , because the plateau is very narrow it is necessary to wait for data at higher reynolds before making any firm conclusion .
the claim is that the observed departure from the multifractal prediction in the tracers statistics is due to contamination induced by viscous effects , which we have seen to be very strong for the @xmath87 case .
the situation becomes more interesting for heavy particles at @xmath83 for which viscous effects have a smaller impact . here
the multifractal scaling gives a larger plateau for the @xmath174 moment , and shows the beginning of a plateau for the @xmath175 moment .
we thus have an indication of an inertial - range intermittent effect for particle pair separation statistics in homogeneous and isotropic @xmath42 turbulence .
+ this is the main result of this paper . + and @xmath175 to the second order one , as a function of time . for each moment
, the upper curve is obtained by compensating with the multifractal exponent , @xmath168 , obtained from the eulerian transverse structure functions , while the middle curve is obtained by compensating with the multifractal exponent from the eulerian longitudinal structure functions ( see previous figure and discussion in the text ) .
the lower curve is obtained by compensating with the richardson s dimensional scaling @xmath176 .
( left ) moments of tracer separation ; ( right ) moments of separation of heavy pairs with @xmath83 .
the error bars are given by the root mean square of equation ( [ eq : ess ] ) computed on two equal sub - ensembles of the whole statistics.,width=491 ] .
[ fig_fit ] on the basis of the multifractal formalism , it is straightforward to conclude that there exist correlation functions that should be _ statistically preserved _ along pair trajectories , for scales well within the inertial range @xcite .
indeed , by again applying the saddle - point estimate to the inertial range scaling of the mixed separation - velocity moments , according to multifractal model we get that @xmath177 if the scaling exponents of the particle separation compensates that of the eulerian velocity moment : @xmath178 . in figure
[ fig_grisha ] , we measure the mixed separation - velocity correlations of order @xmath174 and @xmath175 in ess , that is with reference to the moment of order two .
the scaling obtained by using the @xmath179 exponents is the one that works better for inertial range time lags , i.e. @xmath180 . of order @xmath175 and @xmath174 , with respect to a reference order , @xmath181 , vs time .
the moments are compensated with the dimensional prediction of the theory of richardson , and with the longitudinal intermittent scaling exponents.,width=510 ]
another interesting question concerns the statistical properties of _ weak _ separation events , i.e. the left tail of the probability density function . in order to assess the importance of these events
, one can not resort to negative moments of the separation statistics , because these are ill defined .
an alternative approach which overcomes this difficulty is to use _ inverse _ statistics , see @xcite , i.e. , exit - times .
exit - times also allow for a clearer separation of scales as stressed by @xcite .
the idea consists in fixing a set of thresholds , @xmath182 with @xmath183 and @xmath184 , and in calculating the probability density function of the time @xmath185 needed for the pair separation to change from @xmath148 to @xmath186 .
formally this corresponds to calculating the first passage time .
the advantage of this approach is that all pairs are sampled when they belong to eddies of similar size , between @xmath148 and @xmath186 .
this limits the effect of pairs that at a given time lag , since they have separated very fast or very slow , might be at very different separation scales .
in other words , in the exit - time statistics the _ contamination _ from viscous and large - scale cut - offs should be less important .
+ for particle pairs with initial condition @xmath187 , a perfectly reflecting boundary condition at @xmath188 and an absorbing boundary condition at @xmath189 , the pdf of exit time , @xmath190 , is given by : @xmath191 using the richardson s distribution of eq .
( [ eq : pdfrich ] ) for @xmath192 , we get @xmath193 an asymptotic form of exit - time pdf behaves according to the following expression , @xmath194\,,\ ] ] where @xmath195 is a dimensionless constant , for details see @xcite , and @xmath196 is the mean exit - time . note that eq .
( [ eq : pdfexit2 ] ) contains only dimensionless parameters and it is thus a universal result .
we note that while positive moments @xmath197 preferably sample pairs that separate slowly , negative moments , @xmath198^p \rangle$ ] , are dominated by pairs that separate fast , @xcite . from eq .
( [ eq : pdfexit2 ] ) , a prediction can be obtained for the mean exit - time @xcite , @xmath199 from which it follows that according to the richardson self - similar behaviour , we expect @xmath200 .
the continuous straight line is the richardson s prediction.,width=472 ] in figure [ fig:10 ] , we plot the exit - time pdfs for the tracer pairs , calculated for @xmath201 . first , let us notice that the super - exponential decay observed for the large distance case is probably due to a systematic bias induced by the fact that we have a finite length in the trajectories and therefore very long exit times can not be measured .
second , we observe that the curves do not overlap meaning that the pdfs are not self - similar . concerning the statistical accuracy , this result is a clear improvement of the one reported in @xcite : now deviations from the richardson s prediction are evident because of the huge statistics achieved in the present numerical experiment .
even though the asymptotic distribution is still given by an exponential decay , as it should be expected from rare events following a poissonian process , the whole pdf shape can not be superposed using only the mean exit - time as a normalising factor . + concerning the positive moments of the exit times , we use relative scaling properties to test a breaking of the self - similar properties .
in figure [ fig : t3t4p ] we verify that we have the statistical convergence needed to measure moments of order @xmath202 and @xmath203 . in figure
[ fig : ess_exit ] , we show the ratio of @xmath204 moments of order @xmath205 with respect to @xmath206 , for @xmath207 . for separations within the inertial range @xmath208
the breaking of self - similarity is evident .
it is difficult to conclude if these are true reynolds independent corrections , and if they are affected by the finite length of the particle trajectories .
more data will be needed to get a more quantitative understanding of this effect .
furthermore , we recall that there is not a multifractal prediction for the behaviour of the positive exit - time moments of relative dispersion , since these would be associated to the negative moments of the eulerian velocity increments @xcite . with @xmath209 ( left ) and @xmath174 ( right ) for @xmath207 , testing the statistical convergence.,width=491 ] and @xmath174 to @xmath206 , as a function of thresholds for @xmath207 . for different values of @xmath210
we observe the same behaviours.,width=340 ]
in this work we have numerically studied the relative dispersion statistics of tracer and heavy particles emitted from point - like sources in a homogeneous and isotropic turbulent flow , at resolution @xmath70 . when particles are injected at separations of the order of the viscous scale , the fluctuations of the local stretching rate have a huge impact , influencing the mean square separation , up to time lags of the order of the integral time - scale .
such a dramatic effect has hindered the possibility of studying inertial range quantities in lagrangian experiments or numerical simulations up to now . in this study , the large statistical database ,
the use of conditional statistics and the information obtained by comparing tracers and inertial particles evolution enabled us to highlight with a great precision , and for the first time , deviations in the pair separation distribution from the self - similar behaviour predicted by richardson . such deviations are manifest in the statistical behaviour of the right tail of the separation pdf , and are due to tracer pairs that separate much faster than the average .
we use similar measurements for heavy particles at moderate inertia , which filter out fluctuations of the viscous scales , to give a higher confidence that the corrections observed for tracers are pure inertial range effect .
moreover , by conditioning the relative separation to belong to the inertial range of scales , a universal behaviour develops , which is stokes independent and is clearly different from the richardson s prediction .
the behaviour of the conditioned pdfs support the idea that finite reynolds - number effects , even if present , are sub - dominant . furthermore , the numerical results indicate that tracer dispersion is intermittent , with deviations from a self - similar scaling visible already in the low order moments behaviour .
the observed intermittent corrections to the richardson s prediction are qualitatively consistent with a multifractal prediction for the scaling behaviour of relative separation moments of tracer pairs , although in a narrow scaling region .
numerical and experimental results at higher reynolds number are requested to further support these findings . by measuring the exit - time statistics ,
we provide also an evidence of the non self - similar character of slow pair dispersion events .
the statistics of the shapes of the puffs along the temporal evolution is also a crucial point that worths to be studied , involving more information about the multi - particle separation connected to geometry and shapes @xcite we thank jrmie bec and gregory eyink for useful discussions .
we acknowledge financial support from eu - cost action mp0806 `` particles in turbulence '' .
this work is part of the research program of the foundation for fundamental research on matter ( fom ) , which is part of the netherlands organisation for scientific research ( nwo ) .
numerical simulations were performed within the hpc iscra class a projects `` point '' and `` flag '' at cineca ( italy ) .
acknowledge partial funding from the european research council under the european community s seventh framework programme , erc grant agreement n. 339032 . | results from direct numerical simulations of particle relative dispersion in three dimensional homogeneous and isotropic turbulence at reynolds number @xmath0 are presented .
we study point - like passive tracers and heavy particles , at stokes number @xmath1 and @xmath2 .
particles are emitted from localised sources , in bunches of thousands , periodically in time , allowing to reach an unprecedented statistical accuracy , with a total number of events for two - point observables of the order of @xmath3 .
+ the right tail of the probability density function for tracers develops a clear deviation from richardson s self - similar prediction , pointing to the intermittent nature of the dispersion process . in our numerical experiment ,
such deviations are manifest once the probability to measure an event becomes of the order of -or rarer than- one part over one million , hence the crucial importance of a large dataset .
+ the role of finite - reynolds effects and the related fluctuations when pair separations cross the boundary between viscous and inertial range scales are discussed . an asymptotic prediction based on the multifractal theory for inertial range intermittency and valid for large reynolds numbers
is found to agree with the data better than the richardson theory .
the agreement is improved when considering heavy particles , whose inertia filters out viscous scale fluctuations . by using the exit - time statistics
we also show that events associated to pairs experiencing unusually slow inertial range separations have a non self - similar probability distribution function .
intermittency , multiphase and particle - laden flows , turbulent mixing |
searches for gravitational microlensing events by monitoring stars located in the large magellanic cloud ( lmc ) and the galactic bulge are being carried out and @xmath1 events have been detected ( alcock et al . 1997a , 1997b ; ansari et al .
1996 ; udalski et al . 1997
; alard & guibert 1997 ) .
all of the information from current microlensing experiments is obtained from the einstein timescale , @xmath2 .
one can obtain information about individual lenses because the einstein timescale is related to the physical parameters of the lens by @xmath3 where @xmath4 is the physical size of the einstein ring , @xmath5 is the mass of the lens , @xmath6 is the lens - source transverse speed , @xmath7 , @xmath8 , and @xmath9 are the separations between the observer , lens , and source star . in addition , one can obtain information about galactic structure and the halo macho fraction from the measured optical depth , which is directly proportional to the sum of the timescales of individual events . however , the einstein timescale alone can not fully constrain the physical parameters of individual lenses .
this is because the einstein timescale depends on a combination of the physical parameters of the lens , i.e. , @xmath5 , @xmath6 , and @xmath7 . as a result
, the nature of machos is still very uncertain despite the large number of detected events . therefore it is very important to devise a general method that can resolve the degeneracy in the lens parameters . in general ,
the problem of breaking the lens parameter degeneracy has been approached in two different ways .
the first approach is through measuring the lens parallax ( gould 1994b ) .
if the parallax of an event is measured , one can obtain the projected speed @xmath10 .
the second approach is through measuring the lens proper motion , @xmath11 , where @xmath12 is the angular einstein ring radius .
when both the values of @xmath13 and @xmath14 are determined , one can completely break the degeneracy and obtain the physical parameters of individual lenses by @xmath15^{-1}. } \eqno(2)\ ] ] many methods to measure the lens parallax and the proper motion have been proposed .
the parallaxes of lensing events can be measured in a general fashion from the simultaneous observations of the event from the ground and space ( gould 1994b , 1995 ; han & gould 1995 ; boutreux & gould 1997 ; gaudi & gould 1997 ) . one way to measure the lens proper motion is by utilizing the finite source size effect of lensing events , which can be measured photometrically ( gould 1994a ; nemiroff & wickramasinghe 1994 ; witt & mao 1994 ; witt 1995 ; peng 1997 ) , spectroscopically ( maoz & gould 1994 ; loeb & sasselov 1995 ) , and astrometrically ( mao & witt 1998 ) .
proper motions can also be measured by using lunar occultation ( han , narayanan , & gould 1996 ) and by analyzing the light curves of binary source events ( han & gould 1997 ) . however , unlike the measurement of lens parallax , measurement of @xmath14 using the above methods is only possible for a few rare cases , such as very close lens - source encounter events .
a much more general way of measuring the proper motion is by conducting an accurate astrometric measurement of the separation between the two source images during the event ( miyamoto & yoshii 1995 ) .
this suggestion , however , is implausible because the expected separation for typical values of lens mass and distance is too small for direct observation of individual images .
however , several planned interferometric observation missions , e.g. , the _ space interferometry mission _ ( allen , shao , & peterson 1997 , hereafter sim ) , can be used to measure the astrometric displacements in the light centroid caused by gravitational microlensing in events detected photometrically from the ground ( walker 1995 ; h-1pt g , novikov & polnarev 1995 ; paczyski 1998 ; boden , shao , & van buren 1998 ) . during the event ,
the apparent position of the lensed source star traces out an ellipse ( ` astrometric ellipse ' ) , whose ellipticity depends on the lens - source impact parameter .
the proper motion can be measured because the size of the astrometric ellipse is directly proportional to the einstein ring radius ( see 2 ) . however , except for a brief discussion of bright lenses by boden et al .
( 1997 ) , all previous studies are based on the assumption of a dark lens . in this paper
, we analyze the influence of bright lenses on the astrometric measurements of the centroid displacement and investigate this effect on the determination of @xmath0 .
we find that if an event is caused by a bright lens , the centroid displacement is distorted by the flux of the lens and the resulting astrometric ellipse becomes rounder and smaller with increasing lens brightness , causing an incorrect determination of the angular einstein ring radius .
a lens - blended event can not be distinguished from a dark lens event from the trajectory of the measured centroid displacements alone because both events have elliptical trajectories : the degeneracy between dark and bright lens events .
we also find that even with information from the analysis of the light curve of the event it will still be difficult to resolve the degeneracy caused by the bright lens because the light curve is also affected by lens blending .
when a source star is microlensed , it is split into two images . for a source with a separation @xmath16 from a lens ,
the images are seen at @xmath17 \theta_{\rm e } , \eqno(3)\ ] ] with individual amplifications of @xmath18 . \eqno(4)\ ] ] the lens - source separation in units of the angular einstein ring radius is related to the lensing parameters by @xmath19^{1/2};\qquad { \cal t}= { t - t_0\over t_{\rm e } } , \eqno(5)\ ] ] where @xmath20 is the impact parameter and @xmath21 is the time of maximum amplification . by symmetry , the images with respect to the lens lie at the same azimuth as the source , one on the same and the other on the opposite side of the source . under the assumption that the lens is dark ,
the centroid of the image pair is defined by @xmath22\theta_{\rm e}. \eqno(6)\ ] ] then the location of the image centroid relative to the source star is determined by @xmath23 where the @xmath24 and @xmath25 axes represent the directions which are parallel and normal to the lens - source transverse motion .
if we let @xmath26 and @xmath27 , @xmath28 , the coordinates are related by @xmath29 where @xmath30 and @xmath31 .
therefore , the trajectory of the apparent source star image centroid traces out an ellipse during the event , the ` astrometric ellipse ' .
once the astrometric ellipse is measured , one can determine the impact parameter of the event because the axis ratio of the ellipse is related to the impact parameter .
what makes the measurement of astrometric shift useful lies in the fact that the size of the astrometric ellipse is directly proportional to the angular einstein ring radius .
therefore , one can uniquely determine the proper motion by fitting the observed astrometric shifts to the theoretical trajectories in equations ( 7 ) combined with the timescale which is determined from the photometric observation of the event . in figure 1
, we present the astrometric ellipses for various values of @xmath20 ( upper panel ) and the changes in the semimajor axis ( middle panel ) and in the axis ratio ( lower panel ) of the ellipse as a function of the impact parameter .
if a lensing event is caused by a bright lens , on the other hand , the centroid shift is distorted by the flux of the lens : ` astrometric lens blending . ' the centroid displacement , which is measured with respect to a new reference position for the unlensed lens - source centroid by subtracting the relative motion of the center with respect to nearby stars , for the lens - blended event becomes @xmath32 the distortion factor caused by lens blending is related to the flux ratio between the lens and the source star before gravitational lensing amplification by @xmath33 \over \left [ 1+(\ell_l/\ell_s)\right]\left [ 1+(\ell_l/\ell_s ) u(u^2 + 4)^{1/2}/(u^2 + 2)\right ] \right\}. \eqno(10)\ ] ] the new trajectory is also an ellipse , but with different shape and size from that of an event caused by a dark lens . in the upper panel of figure 2 ,
we present the astrometric ellipses for various values of @xmath34 for an example bright lens event with @xmath35 , where @xmath36 is the apparent magnitude difference between the lens and source . by comparing the shapes and sizes of these ellipses for different values of @xmath34 ,
one finds two trends : first , as the lens / source flux ratio increases , the astrometric ellipse becomes rounder .
the axis ratios are @xmath37 , 0.446 , 0.532 , and 0.624 for @xmath38 , 1 , 0 , and @xmath39 , respectively , compared to the value of @xmath40 for the dark lens case .
second , the size of the ellipse ( measured by the size of the semimajor axis ) decreases with increasing lens brightness .
the semimajor axes of the ellipses in units of @xmath0 when @xmath38 , 1 , 0 , and @xmath39 are @xmath41 , 0.284 , 0.230 , and 0.153 , respectively , compared to the value of @xmath42 for the dark lens case .
in the lower panel of figure 3 , we present the changes in the axis ratio ( @xmath43 : deformation factor ) and the semimajor axis ( @xmath44 : contraction factor ) of the astrometric ellipse as a function of @xmath34 .
note that unlike the semimajor axis , the semiminor axis , @xmath45 , does not monotonically decrease with increasing lens brightness .
it increases when @xmath34 is large , but then decreases as the lens brightness increases further .
this transient increase in @xmath46 arises because the ellipse becomes both rounder and smaller .
a lens - blended event can not be distinguished from a dark lens event just from the trajectory of the measured centroid displacements because both events have the same elliptical trajectories : the degeneracy between dark and bright lens events .
if the centroid shifts of a bright lens event is fit to a model for a dark lens event not knowing the event is affected by the lens brightness , the size of the angular einstein ring radius will be underestimated .
since the size of the angular einstein ring radius is directly proportional to the size of the astrometric ellipse , the ratio of the observed to the true size of the angular einstein ring radius is determined by @xmath47 where the term @xmath48 represents the contraction factor of the semimajor axis and the term @xmath49 is included because @xmath0 is determined not from @xmath50 but from @xmath51 , which is the impact parameter determined from the shape of the observed astrometric ellipse . in table 1
, we list the values of best - fitting @xmath51 and @xmath52 which are determined from the dark lens fit to the observed astrometric centroid displacements of an example event with @xmath53 when it is blended with various lens brightnesses . from the table ,
one finds that increasing the lens brightness causes @xmath51 to increase and the determined values of @xmath54 to decrease compare to their true values .
an interesting fact to note is that although one obtains incorrect values of @xmath0 if lens - blending is not considered , the determined proper motions of events are relatively less affected by lens blending .
this is because not only the astrometrically determined value of @xmath0 , but also the photometrically determined @xmath2 is underestimated due to lens blending .
when the light curve of an event is blended with a bright lens , the determined timescale is reduced by a factor @xmath55^{1/2};\qquad a_{\rm min}=0.34(1+\ell_l/\ell_s)+1 \eqno(12)\ ] ] ( nemiroff 1997 ; han 1998 ) . in table 1 , we list the ratios of the photometrically determined timescale and resulting proper motion to their true values for events blended by lenses with various brightnesses .
one finds that despite substantial changes in the values of both @xmath0 and @xmath2 , the determined proper motions do not deviate seriously from the true values .
in the previous section , we showed that if an event is caused by a bright lens , the centroid displacement is distorted by the flux of the lens and the resulting astrometric ellipse becomes rounder and smaller with increasing lens brightnesses , causing an incorrect determination of @xmath0 .
however , since the astrometric measurements of the centroid displacements will be conducted for events detected photometrically , one might think that it would be possible to correct the astrometric lens - blending effect with additional information from the analysis of the light curves .
if the correct value of the impact parameter is obtained from the light curve , it will be possible to determine the lens / source flux ratio from the comparison of @xmath51 and @xmath50 .
once @xmath56 is determined , it will be explicit to obtain the correct value of @xmath0 .
however , we find that it will still be difficult to correct the lens - blending effect even with photometric information because light curve itself is also affected by lens blending : photometric lens blending . to demonstrate this we simulate lens - blended events toward both the galactic bulge and lmc under realistic observational conditions .
we determine the uncertainty of the photometrically recovered @xmath20 by conducting both bright and dark lens fit to the simulated event light curve .
the result of the fit is obtained by computing @xmath57 , i.e. , @xmath58 ^ 2 , \eqno(13)\ ] ] where @xmath59 is the number of measurements , @xmath60 is the photometric precision , and @xmath61 is the number of parameters , which is 5 ( @xmath62 , @xmath21 , @xmath2 , @xmath20 , and @xmath56 ) . the light curve with lens blending is represented by @xmath63 where @xmath64 is the absolute amplification of the event .
the events are assumed to be observed with a frequency of @xmath65 during @xmath66 .
for the values of the einstein timescales , we adopt @xmath67 and 40 days for the galactic bulge and lmc events , respectively , based on the most frequent values of actual events toward each direction ( alcock et al . 1997a , 1997b ) .
the photometric precision is assumed to be @xmath68 for the galactic bulge events , while we assume @xmath69 for the lmc events as the source stars are faint .
the uncertainties for both fits are determined by the @xmath70 level , which is equivalent to @xmath71 . in table 2
, we list the uncertainties in the photometrically determined impact parameters for bright lens events by both dark and bright lens fits . from the table , one finds the best - fitting impact parameter from the astrometric dark lens fit , @xmath51 listed in table 1 , lies within the uncertainty range of @xmath20 obtained from the photometric bright lens fit . although the best - fitting impact parameter from the astrometric dark lens fit is just beyond the uncertainty range of @xmath20 determined from the photometric dark lens fit , it will be very difficult to distinguish the two values considering the uncertainties accompanied to @xmath51 .
this implies that lens - blending effect can not be detected from the simple comparison of the impact parameters which are determined both astrometrically and photometrically .
in addition , one finds that the best - fitting @xmath20 from the photometric dark lens fit lies within the uncertainty range of @xmath20 obtained from the photometric lens blending fit , implying that lens blending can not also be easily detected photometrically .
alard , c , & guibert , j. 1997 , , 326 , 1 alcock , c. , et al .
1997a , , 479 , 119 alcock , c. , et al . 1997b , , 486 , 697 allen , r. , shao , m. , & peterson , d. 1997 proc .
spie , 2871 , 504 ansari , r. , et al .
1996 , , 314 , 94 boden , a. f. , shao , m. , & van buren , d. 1998 , , submitted boutreux , t. , & gould , a. 1996 , , 462 , 705 gaudi , b. s. , & gould , a. 1997 , , 477 , 152 gould , a. 1994a , , 421 , l71 gould , a. 1994b , , 421 , l75 gould , a. 1995 , , 447 , 491 han , c. 1998 , , 500 , 569 han , c. , & gould , a. 1995 , , 447 , 5e han , c. , & gould , a. 1997 , , 480 , 196 han , c. , narayanan , v. k. , & gould , a. 1996 , , 461 , 587 h-1pt g , e. , novikov , i. d. , & polnarev , a. g. 1995 , , 294 , 287 loeb , a. , & sasselov , d. 1995 , , 449 , l33 mao , s. , & witt , h. j. 1998 , preprint ( astro - ph/9804045 ) maoz , d. , & gould , a. 1994 , , 425 , l67 miyamoto , m. , & yoshii , y. 1995 , , 110 , 1427 nemiroff , r. j. 1997 , , 486 , 693 nemiroff , r. j. , & wickramasinghe , w. a. d. t. 1994 , , 424 , l21 paczyski , b. 1998 , , 404 , l23 peng , e. w. 1997 , , 475 , 43 udalski , a. , et al . 1997 , acta .
astron . , 47 , 169 walker , m. a. 1995 , , 453 , 37 witt , h. j. 1995 , , 449 , 42 witt , h. j. , & mao , s. 1994 , , 430 , 505 note. the uncertainties of the photometrically determined values of the impact parameter for events blended by lenses with various brightnesses .
the uncertainties are determined by conducting both bright and dark lens fits to the simulated event light curves constructed under reasonable observational conditions .
upper panel : the trajectories of the source image centroid displacements of an event caused by a dark lens for various values of @xmath20 . here
, @xmath24 and @xmath25 axes represent the directions which are parallel and normal to the lens - source transverse motion , respectively .
middle and lower panels : the changes of semimajor axis ( middle panel ) and the axis ratio ( lower panel ) of the astrometric ellipse as a function of impact parameter .
upper panel : the trajectories of the source image centroid displacements of events caused by bright lenses .
we present the trajectories for various values of magnitude difference between the lens and the source star , @xmath34 , for an example event with @xmath35 .
lower panel : the changes in the axis ratio ( @xmath43 : deformation factor ) and in the semimajor axis ( @xmath44 : contraction factor ) of the astrometric ellipse of bright lens event as a function of @xmath34 . | in current microlensing experiments , the information about the physical parameters of individual lenses are obtained from the einstein timescales .
however , the nature of machos is still very uncertain despite the large number of detected events .
this uncertainty is mainly due to the degeneracy of the lens parameters in the measured einstein timescales .
the degeneracy can be lifted in a general fashion if the angular einstein ring radius @xmath0 , and thus the macho proper motion , can be measured by conducting accurate astrometric measurements of centroid displacement in the source star image . in this paper , we analyze the influence of bright lenses on the astrometric measurements of the centroid displacement and investigate this effect on the determination of @xmath0 .
we find that if an event is caused by a bright lens , the centroid displacement is distorted by the flux of the lens and resulting astrometric ellipse becomes rounder and smaller with increasing lens brightness , causing an incorrect determination of the angular einstein ring radius .
a lens - blended event can not be distinguished from a dark lens event from the trajectory of the measured centroid displacements alone because both events have elliptical trajectories : the degeneracy between dark and bright lens events .
we also find that even with information from the analysis of the light curve of the event it will still be difficult to resolve the degeneracy caused by the bright lens because the light curve is also affected by lens blending . 6.5 in ccc 1 & 1 & 1 + 1 & 1 & 1 + 4 & 3 & 8 submitted to _ the astrophysical journal _ : july 16 , 1998 preprint : cnu - a&ss-06/98 |
thermodynamics of quantum - optical systems has intrigued scientists ever since masers and lasers were realized experimentally .
scovil and schulz - dubois @xcite analyzed a three - level maser in the framework of a heat engine .
based on a boltzmann distribution of atomic populations , they gave an intuitive definition of the engine s efficiency , and showed it to be less than an or equal to the carnot efficiency .
using the concept of negative temperature @xcite , and motivated by ramsey s @xcite work on spin temperature , scovil and schulz - dubois @xcite extended their analysis of three level systems to cases where the reservoirs temperature is negative , and introduced the concept of negative efficiencies .
alicki studied a generic open quantum system coupled to heat reservoirs , and under the influence of varying external conditions ( such as a time dependent field ) @xcite .
alicki partitioned the energy of a quantum system into heat and work using the time dependencies of the density and hamiltonian operators . based on alicki s definitions for heat and work , kosloff analyzed two coupled oscillators interacting with hot and cold thermal reservoirs in the framework of a heat engine , and showed that the engine s efficiency complies with the second law of thermodynamics @xcite .
in later work , geva and kosloff studied a three - level amplifier coupled to two heat reservoirs @xcite @xcite . in their model
the external field influences the dissipative terms , and the second law of thermodynamics is generally satisfied .
this paper is to some extent a continuation of the studies discussed in the previous paragraph .
in contrast with previous work , in our approach the matter and the radiation field are treated as a bipartite system that is fully quantized , as opposed to a forced unipartite system .
this treatment of the working medium ( the material system ) and the work source ( the radiation field ) on an equal footing requires some new thermodynamic developments , that we adapt from @xcite .
the general methodology is applied to an extended dissipative jaynes - cummings model ( ed - jcm ) , which consists of a three - level material system coupled to two thermal heat baths and a quantized cavity mode .
we show that this system provides a simple model of light amplification , which can then be analyzed using formulations of the first and second law of thermodynamics for bipartite systems .
the heat flux and power calculated with this model lead to an engine efficiency that is in quantitative agreement with the efficiency formula intuitively defined by scovil and schulz - dubois .
a semiclassical counterpart of the ed - jcm equations is then presented and solved completely at steady state , giving the efficiency formula of scovil and schultz - dubois analytically from fundamental thermodynamic fluxes .
this paper is arranged in the following manner .
section ii is a brief introduction to the thermodynamics of bipartite systems . in section iii
we define the ed - jcm master equation . in section
iv we present numerical results for the ed - jcm model , showing that it acts as a simple model for a quantum amplifier . in section
v we discuss the entropic behavior of the full system and its individual components , its behavior at steady state and the role of entanglement . in section vi we give a thermodynamical analysis of the ed - jcm .
we formulate the first law of thermodynamics in two different ways .
we then show that the second law of thermodynamics is satisfied in two formulations ( clausius , carnot ) , and that the efficiency of the ed - jcm heat engine agrees with that defined intuitively by scovil and schulz - dubois . in section vii we compare the steady state heat flux and power of the fully quantum model with a semiclassical version of the ed - jcm , and derive the engine efficiency formula of scovil and schulz - dubois _ analytically _ from fundamental thermodynamic fluxes .
section viii concludes .
a bipartite system is described by a density matrix of a @xmath0 hilbert space .
the partial density matrix of one part is obtained by tracing over the other : @xmath1 the entropy of a quantum system is given by the von neumann entropy @xcite : @xmath2 the evolution of a bipartite system is given by the following master equation : @xmath3+\mathcal{l}_{d}[\mbox{\boldmath$\rho_{ab}$}],\label{masterab}\ ] ] where @xmath4=-\frac{i}{\hbar}[\mbox{\boldmath$h$ } , \mbox{\boldmath$\rho_{ab}$}]$ ] is the hamiltonian part of the lindblad super operator , and @xmath5 $ ] is the dissipative part of the lindblad super operator . the bipartite time independent hamiltonian is given by : @xmath6 where @xmath7 and @xmath8 are the hamiltonians of subsystems @xmath9 and @xmath10 , and @xmath11 is the coupling term between them . here and throughout the article , we use bold letters to signify operators that have a tensor product structure .
heat flux and power of the individual parts of the system are defined by @xcite : @xmath12\mbox{\boldmath$h_{a(b)}$}\}\\\nonumber p_{a(b)}&\equiv&-\frac{i}{\hbar}\textrm{tr}\{\mbox{\boldmath$\rho_{ab}$}[\mbox{\boldmath$h_{a(b)}$ } , \mbox{\boldmath$v_{ab}$}]\}.\label{qdotpab}\end{aligned}\ ] ] the energy flux of the full system is due only to the dissipative part of the lindblad super operator : @xmath13\mbox{\boldmath$h$}\}.\label{eab}\ ] ]
consider a three - level system interacting resonantly with one quantized cavity mode and two thermal photonic reservoirs as depicted in fig .
[ edjcm ] .
three level system interacting with two heat reservoirs ( hot and cold ) and a quantized cavity mode . ]
the system is governed by the following master equation in the interaction picture : @xmath14+\mathcal{l}_{dc}[\mbox{\boldmath$\rho_{mf}$}]+\mathcal{l}_{dh}[\mbox{\boldmath$\rho_{mf}$}].\label{mqe}\ ] ] the letters in the subscripts have the following significance : @xmath15=matter , @xmath16=field , @xmath17=dissipative , @xmath18=hamiltonian , @xmath19=cold , @xmath20=hot .
the hamiltonian part of the liouvillian is given by : @xmath21=-\frac{i}{\hbar}[\mbox{\boldmath$v_{mf}$},\mbox{\boldmath$\rho_{mf}$ } ] , \label{eq : ixn_rep_h}\ ] ] where @xmath22 is a resonant jcm type interaction hamiltonian , @xmath23 being the matter - field coupling constant . @xmath24 $ ] and @xmath25 $ ] are the dissipative cold and hot lindblad super operators , respectively : @xmath26&=&\gamma_{02}\{(n_{02}\!+\!1)([\mbox{\boldmath$\sigma_{02}\rho_{mf}$},\mbox{\boldmath$\sigma_{02}^{\dag}$}]\!+\![\mbox{\boldmath$\sigma_{02}$},\!\mbox{\boldmath$\rho_{mf}\sigma_{02}^{\dag}$}])\!+\ ! n_{02}([\mbox{\boldmath$\sigma_{02}^{\dag}\rho_{mf}$},\mbox{\boldmath$\sigma_{02}$}]\!+\![\mbox{\boldmath$\sigma_{02}^{\dag}$},\mbox{\boldmath$\rho_{mf}\sigma_{02}$}])\}\nonumber\\ \!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\!\mathcal{l}_{dh}[\mbox{\boldmath$\rho_{mf}$}]&=&\gamma_{01}\{(n_{01}\!+\!1)([\mbox{\boldmath$\sigma_{01}\rho_{mf}$},\mbox{\boldmath$\sigma_{01}^{\dag}$}]\!+\![\mbox{\boldmath$\sigma_{01}$},\mbox{\boldmath$\rho_{mf}\sigma_{01}^{\dag}$}])\!+\ ! n_{01}([\mbox{\boldmath$\sigma_{01}^{\dag}\rho_{mf}$},\mbox{\boldmath$\sigma_{01}$}]\!+\![\mbox{\boldmath$\sigma_{01}^{\dag}$},\mbox{\boldmath$\rho_{mf}\sigma_{01}$}])\},\label{ldch}\end{aligned}\ ] ] where @xmath27 and @xmath28 are the weiskopf - wigner decay constant associated with the cold and hot reservoirs , respectively , and @xmath29 and @xmath30 are the number of thermal photons in the cold and hot reservoirs , respectively . note that direct dissipation occurs only through matter - reservoir coupling ( the cold photonic reservoir couples levels @xmath31 and @xmath32 , the hot photonic reservoir couples levels @xmath31 and @xmath33 ) , and is typically used to represent atomic decay in quantum optics .
the matter creation and annihilation operators are in tensor product form @xmath34 , and their matrix form is given by : @xmath35 the reservoirs temperature is given by : @xmath36 where @xmath37 is the central frequency of the cold ( hot ) reservoir .
the ed - jcm master equation ( equation [ mqe ] ) can be obtained by summing the hamiltonian contribution and the two dissipative contributions .
alternatively , it can be derived for a three - level system with a break in symmetry using the weak coupling ( to the reservoirs ) , markovian , and weiskopf - wigner approximations in a similar fashion to the simple jcm with master equation with atomic damping which is derived in appendix i. the hamiltonian ( energy operator ) of the full matter - field system is given by : @xmath38 where @xmath39 and @xmath40 are the matter and field hamiltonians , respectively , and @xmath41 is given by : @xmath42 under matter - field resonance ( @xmath43 ) the hamiltonian in the interaction picture is unchanged and is not time dependent ( @xmath44 ) since @xmath45=0 $ ] ( when there is no resonance one can still transform to an interaction picture in which the hamiltonian is unchanged @xcite ) . however , as indicated previously ( eq . [ eq : ixn_rep_h ] ) , in the interaction picture , the hamiltonian part of the evolution of the density matrix is only via the interaction term @xmath46 .
before we move on to discuss the ed - jcm as a quantum amplifier , we wish to discuss the main differences between the ed - jcm and the quantum theory of the laser due to scully and lamb ( sl ) @xcite @xcite .
firstly , in the sl model the material system ( the atom ) has either four levels @xcite or five levels , whereas in the ed - jcm the matter has three levels .
secondly , in the sl model the transitions between the two upper lasing levels and the two lower levels is achieved through a phenomenological decay , whereas in the ed - jcm population may also be pumped from the ground state to the two upper lasing levels through the full dissipative lindblad super operator .
thirdly , in the sl model the atom is assumed to be injected into the cavity in the upper lasing level and interact with the cavity for a time @xmath47 , whereas in the ed - jcm the matter is in continuous contact with the quantized cavity mode , and amplification is achieved for a wide range of initial states .
finally , in the sl model the field is allowed to decay using the weiskopf - wigner formalism , whereas in the ed - jcm discussed in this paper the field does not decay . in principle , cavity losses can be introduced to the ed - jcm .
however , we do not consider field damping in this paper , which allows us to compare the thermodynamical fluxes in the quantum ed - jcm with their analog in a semiclassical ed - jcm ( section vii ) and a similar model by geva and kosloff in which field damping is not included @xcite @xcite . the differences between our model and that of the sl model will be seen below to play a crucial role in our ability to give a thermodynamic foundation of amplification .
the ed - jcm master equation , eq .
[ mqe ] , was solved using the standard runge - kutta method ( fourth - order @xcite ) for various choices of parameters .
the accuracy of the solution was checked by decreasing the step size .
furthermore , in order to test whether the numerical solution captures all time scales ( especially the rapid oscillations ) , the algorithm was tested on the simple jcm @xcite which can be solved analytically @xcite @xcite . in all plots presented here @xmath48 , and quantities
are given in atomic units .
the condition @xmath49 corresponds physically to a situation where the coupling between the matter and the selected quantized cavity mode is much stronger than the matter - reservoir coupling .
the energy flux of the full matter - field system and the individual subsystems is given by : @xmath50\mbox{\boldmath$h$}\}\nonumber\\ \dot{e}_{m}&\equiv&\textrm{tr}\{\dot{\rho}_{m}h_{m}\}=-\frac{i}{\hbar}\textrm{tr}\{\mbox{\boldmath$\rho_{mf}$}[\mbox{\boldmath$h_{m}$ } , \mbox{\boldmath$v_{mf}$}]\}+\textrm{tr}\{\mathcal{l}_{d}[\mbox{\boldmath$\rho_{mf}$}]\mbox{\boldmath$h_{m}$}\}=p_{m}+\dot{q}_{m}\nonumber\\ \dot{e}_{f}&\equiv&\textrm{tr}\{\dot{\rho}_{f}h_{f}\}=-\frac{i}{\hbar}\textrm{tr}\{\mbox{\boldmath$\rho_{mf}$}[\mbox{\boldmath$h_{f}$ } , \mbox{\boldmath$v_{mf}$}]\}=p_{f},\label{firstlawall}\end{aligned}\ ] ] where @xmath51 , @xmath52 , and @xmath53 are the energy fluxes of the full matter - field system , the matter , and the field , respectively .
@xmath54 and @xmath55 are obtained from @xmath56 by a partial trace over the field or a partial trace over the matter , respectively . @xmath57 and @xmath58 are the hamiltonians of the matter and field subsystems , respectively , without the tensor product with the identity .
note that the energy fluxes of the individual subsystems in eq .
[ firstlawall ] are defined via @xmath54 and @xmath55 together with the subsystem hamiltonians @xmath59 . in the next two subsections we discuss the transient and steady state energetic behavior of the ed - jcm . since there is direct dissipation only through matter - reservoir coupling ,
there is no heat flux associated with the field ( this is physically expected , and was shown analytically elsewhere @xcite ) .
the energy of the full matter - field system and of the individual subsystems is plotted in fig .
[ edjcmenergy ] for an initial state where the matter is in state @xmath33 and the selected cavity mode has no photons ( @xmath60 ) .
and the selected cavity mode has no photons ( @xmath60 ) .
note that at long times there is a steady state increase in the field s energy.,width=302 ] at short times , @xmath61 , the matter and field energies oscillate at a frequency of @xmath62 @xcite . here
@xmath63 is the effective decay constant .
moreover , at short times the well known collapse and revival phenomena @xcite @xcite is observed for a sufficiently excited initial coherent state as depicted in fig .
[ edjcmenergyzoom ] . ) .
the revival time ( top ) is @xmath64 , where @xmath15 is a positive integer .
the rabi oscillation time ( bottom ) is @xmath65.,width=302 ] in order to monitor the field s coherence we calculate the quantum optical husimi - kano @xmath66 function which is defined by @xcite : @xmath67 where @xmath68 is a ( generally complex ) coherent state . in fig .
[ edjcmq ] we plot the @xmath66 function at four different times ( @xmath69 , @xmath70 , @xmath71 , @xmath72 ) for the initial state @xmath73 , @xmath74 .
function of the selected cavity mode , @xmath75 .
a ) : at @xmath69 the @xmath66 function is a narrow 2d gaussian .
b ) and c ) : at @xmath76 the @xmath66 function is spread in phase space inhomogeneously .
d ) : at @xmath77 the @xmath66 function has expanded ( due to amplification of the selected mode ) into a radially symmetric annulus ( all the initial phase information is lost).,width=377 ] the @xmath66 function at @xmath69 is that of a coherent state with the phase centered around the real axis ( fig . [ edjcmq]a ) . at transient times , the @xmath66 function spreads in phase space , but it is not homogeneous as seen in fig . [ edjcmq]b and fig .
[ edjcmq]c . at much longer times , @xmath78 , the matter energy decreases to a steady state value , while the field energy increases with a steady state power of @xmath79 ( the numerical value of a linear fit to the last @xmath80 points , @xmath81 ) as seen in fig .
[ edjcmenergy ] .
another indication for an increase in the field s energy is seen in the steady state increase in the full system energy .
thus , the field and the matter - field system as a whole never reach a steady state for the type of evolution discussed in this paper . the relation between the full system energy and the energy of the individual component subsystems
will be discussed in the next sections . a steady state increase in the field
s energy is clearly an amplification of the selected cavity mode .
this behavior contrasts with the simple jcm in which the atom and field oscillate forever ( the atom oscillates between the excited and ground states while the field oscillates between the @xmath31 and @xmath33 fock states ) .
amplification of the selected cavity mode will occur with any other coherent state , including the @xmath31 fock state .
the fact that the field s energy increases monotonically is not unreasonable , since the harmonic oscillator is infinite , and since we do not consider direct dissipation of the cavity mode ( which could be modeled by transmissive mirrors if desired ) .
the collapse and revival phenomenon at longer times is completely damped due to the dissipative contribution to the liouvillian as seen in fig .
[ edjcmenergyzoom ] . at these long times
all phase ( internal coherence ) information is lost : the @xmath66 function is radially symmetric and is dispersed on a bigger area ( bottom of fig . [ edjcmq]c ) . from this time
onwards the shape of the @xmath66 function remains unchanged , and it expands fully symmetrically .
the decay of the initial field coherence is also reflected in the decay of the off - diagonal field density matrix elements . at @xmath82 ,
the off - diagonal matrix elements are @xmath83 times smaller than their initial value , and are practically zero .
all the remaining density matrix elements are diagonal with a poissonian - like photon distribution whose average number of photons increases with time . the full density matrix can be divided into a @xmath84 block matrix , each block associated with one element of the matter density matrix . at long times ,
@xmath85 , the matter - field inter - coherence is maintained by the non - vanishing matrix elements : @xmath86 of the full density matrix .
these elements correspond to matter - field coupling , maintained via the structure of the jcm hamiltonian .
other density matrix elements at these long times are @xmath87 orders of magnitude smaller than the dominant non vanishing elements discussed above .
we consider now the entropy in the ed - jcm . in the next two subsections we discuss the transient and steady state entropic behavior . in subsection c
we discuss the relation between the entropies of the individual subsystems and entanglement , both at transient and steady state times . in fig .
[ edjcmentropy ] we plot the entropy of the full matter - field system and the individual subsystems for the initial state @xmath88 , @xmath89 . .
a ) : entropy of the full matter - field system ( solid line ) and the individual subsystem ( field dotted line , matter dash - dot line ) for the full evolution .
b ) : matter entropy for times @xmath90 .
c ) : entropy of the full matter - field system for times @xmath85.,width=377 ] the entropy plots at the top of fig .
[ edjcmentropy ] show that there is a rapid rise ( @xmath91 ) in the entropies of the matter ( dash - dot line ) , the field ( dotted line ) , and the full matter - field system ( solid line ) .
this overall rise in entropy is discussed in subsection c ) and is attributed to the dissipative nature of the problem . at times
@xmath92 , the entropy of the individual subsystems ( matter or field ) is oscillatory , as seen by the matter entropy plot in the middle of fig .
[ edjcmentropy ] .
this behavior is typical of the simple jcm @xcite @xcite . at this stage @xmath93
; in subsection c ) we attribute the excess entropy to entanglement .
[ edjcmentropy]a suggests that at @xmath94 the matter - field system has reached a steady state .
indeed , at times @xmath94 the energy of the matter remains constant ( dash - dot line in fig .
[ edjcmenergy ] ) .
however , as was indicated in the previous section , the field energy plot ( dotted line in fig .
[ edjcmenergy ] ) and the matter - field energy plot ( solid line in fig . [ edjcmenergy ] ) both show a constant rise for @xmath78 .
furthermore , a closer inspection of the matter - field entropy ( fig . [ edjcmentropy]b ) reveals a constant slight rise in entropy at times @xmath78 ( a similar rise in the field entropy is also observed ) .
moreover , the field density matrix eigenvalues change in the second and third significant figures over a @xmath95 time scale .
these findings give further proof of the fact that the field and matter - field system as a whole never reach a steady state .
we will now analyze the nature of the entropies associated with the subsystems .
the entropy of the individual parts of a bipartite system is closely tied to the issue of entanglement @xcite @xcite .
an important measure for entanglement is the conditional entropy , defined for the matter - field system by : @xmath96 where @xmath97 is the conditional entropy of the matter , and @xmath98 is the conditional entropy of the field .
in contrast with the conditional entropy in classical bipartite systems , the conditional entropy in quantum bipartite systems can assume negative values . in this case , the correlation between the two parts of the system is of a purely quantum nature , and the system is therefore entangled .
a fine example for entanglement in the context of our work is the simple jcm .
consider an initial state given by : @xmath99 , where the atom ( indicated by subscript @xmath100 ) is in the excited state and the cavity mode is empty . under the jcm hamiltonian ( pure hamiltonian dynamics ) , the full atomic - field entropy is constant @xmath101 .
however , during most of the evolution time the conditional entropies of both the atom and field ( which are equal ) are negative . in our case , at short times , @xmath102 we find that the matter s conditional entropy is negative .
thus , the excess entropy @xmath93 at these times is attributed to entanglement .
a more powerful test for entanglement , introduced originally by peres @xcite , is the negativity of the partially transposed density matrix .
the partially transposed density matrix is defined by : @xmath103 a sufficient condition for entanglement is the negativity of @xmath104 .
however , since this test applies only to finite dimensional density matrices , one should take care not to mistake truly negative eigenvalues with negative eigenvalues that are an artifact of truncation of an infinite hilbert space @xcite .
indeed , at times smaller than the typical decay time ( @xmath105 ) , we find that the matter - field partially transposed density matrix @xmath106 is negative ( negative eigenvalues with a substantial absolute value are found up to a time @xmath107 , and hence the matter - field system is entangled . at @xmath94
the matter and field conditional entropies are positive .
moreover , the conditional entropies are almost equal to the partial entropies : @xmath108 , and @xmath106 is positive ( as was indicated before ) .
all these findings lead us to conclude that in all likelihood at long times the matter - field system is only weakly classically correlated .
we summarize this section by stating that at short times ( @xmath109 ) , when the partial entropies are oscillatory ( see bottom of fig .
[ edjcmentropy ] ) , the matter - field system is entangled , as verified by the negative conditional entropies and the negative partially transposed full density matrix . however , as dissipation sets in , the matter - field system becomes less and less entangled . at @xmath94 ,
when the partial entropies are not oscillating any more , the matter - field system in all likelihood is not entangled ( as verified by the positive conditional entropies and the positive partially transposed full density matrix ) , and the overall rise in entropy is attributed to the dissipative lindblad super operator .
the first law of thermodynamics is essentially given in equation [ firstlawall ] . however , some fine details need more clarification .
the first law of thermodynamics for the full matter - field system in differential form is given by : @xmath110h\}=\dot{q}_{m}+\dot{q}_{f}+\dot{q}_{v}=\dot{q}_{m}+\dot{q}_{v},\ ] ] where @xmath111\mbox{\boldmath$h_{f}$}\}=0 $ ] as was shown elsewhere @xcite , and @xmath112\mbox{\boldmath$v_{mf}$}\}$ ] .
@xmath113\mbox{\boldmath$h_{m}$}\}=\textrm{tr}\{\mathcal{l}_{dc}[\mbox{\boldmath$\rho_{mf}$}]\mbox{\boldmath$h_{m}$}\}+\textrm{tr}\{\mathcal{l}_{dh}[\mbox{\boldmath$\rho_{mf}$}]\mbox{\boldmath$h_{m}$}\}=\dot{q}_{mc}+\dot{q}_{mh}$ ] is the heat flux associated with the matter and it is composed of heat fluxes from / to the cold and hot heat reservoirs .
note that to an observer looking on the matter - field system as a whole , the full system is only dissipating heat . another way to formulate the first law of thermodynamics
is based on the energy flux of individual subsystems .
the first law of thermodynamics for the matter and field separately ( in differential form ) is given by : @xmath114\}+\textrm{tr}\{\mathcal{l}_{d}[\mbox{\boldmath$\rho_{mf}$}]\mbox{\boldmath$h_{m}$}\}=p_{m}+\dot{q}_{m}\label{emdot}\\ \dot{e}_{f}&\equiv&\textrm{tr}\{\dot{\rho}_{f}h_{f}\}=-\frac{i}{\hbar}\textrm{tr}\{\mbox{\boldmath$\rho_{mf}$}[\mbox{\boldmath$h_{f}$ } , \mbox{\boldmath$v_{mf}$}]\}=p_{f},\label{efdot}\end{aligned}\ ] ] where @xmath115\}$ ] , and @xmath116\}$ ] are the power terms .
since we are considering the case of perfect matter - field resonance , @xmath117 ( @xmath118=-[\mbox{\boldmath$h_{f}$},\mbox{\boldmath$v_{mf}$}]$ ] ) , hence : @xmath119 it may be shown that @xmath120 vanishes if the off - diagonal matrix elements of @xmath56 are purely imaginary .
note that to an observer looking on the matter alone work flux ( power ) and heat fluxes are identified according to eq .
[ emdot ] , in agreement with the traditional thermodynamic partitioning of energy into work and heat .
the field , which is the work source , either receives or emits energy to the working medium ( the matter ) in the form of power . in this paper
we are interested in optical amplification . under such conditions , at steady state the energy flux balance is such that @xmath121 and the three - level system operates thermodynamically as a heat engine . the second law of thermodynamics is obtained via the entropy production function of the full bipartite matter - field system , which is defined by @xcite , @xcite : @xmath122 where @xmath123 is the entropy production associated with the bipartite matter - field density matrix ( via differentiation of the von neumann entropy ) , and @xmath124 is the entropy production associated with the reservoirs ( via the heat flux from / to the reservoirs ) given by : @xmath125 where @xmath126 , and @xmath127(\mbox{\boldmath$h_{m}$}+\mbox{\boldmath$v_{mf}$})\}=\dot{q}_{mc(h)}+\dot{q}_{vc(h)}$ ] .
spohn showed that for a completely positive map ( such as the lindblad super operator ) @xcite : @xmath128 equation [ entprod1 ] represents the differential form of the second law of thermodynamics in clausius s formulation , since the sum of the entropy changes of the system and reservoirs is guaranteed to be positive .
we now define a new entropy production function : @xmath129 where @xmath130 is the entropy production associated with the matter density matrix ( via differentiation of the matter von neumann entropy ) , and @xmath131 is the entropy production associated with the reservoirs , @xmath132 taking into account the contribution only from the matter heat flux @xmath133 the physical idea behind @xmath134 is that it is built only from matter thermodynamic fluxes : the intrinsic entropy flux @xmath130 and the entropy flux @xmath131 arising just from matter heat fluxes .
for many initial matter states @xmath135 at all times , and hence @xmath136 . moreover ,
when the matter reaches a steady state we always find numerically ( irrespective of the initial matter state ) that @xmath135 .
this is the case also in the semiclassical ed - jcm discussed in section vii , where it can be shown analytically that @xmath137 .
therefore , at steady state , @xmath134 is physically similar to the entropy production function in the semiclassical case , @xmath138 , where the field is not quantized .
in contrast with @xmath138 and @xmath139 , @xmath134 is not guaranteed to be positive at all times ( especially at times @xmath140 , due to the highly oscillatory nature of the partial entropy at short times ) . however , when the matter reaches a steady state ( @xmath141 ) , the increase in the field s entropy is marginal ( as was indicated before ) , and the main source of entropy production is the heat flux from / to the heat reservoirs ( @xmath142 ) .
thus , when the matter reaches a steady state @xmath143 and since the matter operates in a heat engine mode ( @xmath144 and @xmath145 ) , we obtain carnot s efficiency formula : @xmath146 where we have used eq .
[ emdot ] with eq .
[ dotqm ] , and eq . [ sigmss ] with eq .
[ jm ] . for example ,
the efficiency of the heat engine for the choice of parameters discussed in the previous plots and for various initial field strengths ( ranging from an empty cavity up to @xmath147 photons ) is @xmath148 , which is less than the carnot efficiency which is @xmath149 .
scovil and schulz - dubois gave an intuitive , but non - thermodynamics definition of the efficiency of the three - level system operating as a maser @xcite : @xmath150 where @xmath151 is the signal ( maser ) frequency , and @xmath152 is the pump frequency ( central frequency of the hot reservoir , @xmath153 ) . by substituting our initial choice of parameters ( @xmath154 , and @xmath155 )
we see that our numerical result agrees precisely with the efficiency estimated by scovil and schulz - dubois .
we find numerically that the efficiency calculated from eq .
[ carnot ] agrees precisely with the efficiency calculated using eq .
[ scovil_djt ] .
although we do not have analytic proof of this equivalence for the case of quantized light , we show below that this equivalence can be derived analytically when the light is treated classically .
a work source is the physical entity on which work is done , or which performs work on a system ( working medium ) .
conventional wisdom in classical thermodynamics states that a work source s entropy is constant during the operation of a heat engine @xcite . whether the classical engine operates cyclically , as in the usual carnot cycle , or synchronously , the working medium returns to its initial state .
the working assumption in thermodynamics is that entropy may be produced at the boundary of the working medium and the heat reservoirs , but not at the boundary with the work source .
the work source in the quantum amplifier discussed in this paper is the selected cavity mode which is amplified . at steady state ,
the density matrix of the matter becomes constant and thus its entropy is unchanged from this time onwards .
since energy is flowing from the hot reservoir to the cold reservoir and work is produced in the form of amplification of the cavity mode , this corresponds to the engine operating in synchronous mode with the cavity mode as the work source .
inspection of fig .
[ edjcmentropy ] shows that the entropy of the light is not constant .
even after the matter reaches a steady state , the entropy of the light continues to grow linearly in time .
the semiclassical ed - jcm master equation is similar to the quantum ed - jcm master equation given by equation [ mqe ] . however , since the selected quantized cavity mode is replaced by a time dependent field , major differences arise .
the field is considered as an external degree of freedom , and hence it has no entropy content .
we propagate a @xmath84 density matrix representing the matter only , and all operators are represented by @xmath84 matrices ( as opposed to @xmath156 in the fully quantized case ) .
finally , the hamiltonian part of the liouvillian assumes a different form , where the creation and annihilation field operators are replaced by clockwise and anti - clockwise oscillating exponents . despite the last difference
, we note that in perfect matter - field resonance the hamiltonian part of the liouvillian in the interaction picture is time independent .
the semiclassical hamiltonian is given by : @xmath157 where @xmath158 is the matter hamiltonian as given in equation [ jcmh ] ( without the tensor product with @xmath159 ) , and @xmath160 is the interaction hamiltonian with a classical single coherent mode in the rwa .
@xmath161 is the semiclassical matter - field coupling constant ( which can be obtained via the semiclassical coupling matrix element @xcite ) given by ( atomic units ) : @xmath162 where @xmath163 is the dipole operator , @xmath164 is the field polarization , and @xmath165 is the field amplitude which can be estimated by calculating the average value of the quantum field operator for a coherent state @xcite : @xmath166 where @xmath167 is the mode frequency ( not necessarily in resonance with the atomic transition ) , @xmath168 is the cavity volume , and @xmath169 is the field strength .
the quantum matter - field coupling constant is given by ( atomic units ) @xcite : @xmath170 combining equations [ lsc ] , [ e0 ] , and [ lq ] we obtain that : @xmath171 the dissipative part of the liouvillian is identical to equation [ ldch ] ( without the tensor product with @xmath159 ) .
substitution of equation [ edjcmsc ] and equations [ ldch ] ( not in tensor product form ) into equation [ mqe ] yields a set of equations for the matter density matrix elements . in the interaction picture
( with @xmath172 given by @xmath57 ) and assuming perfect matter - field resonance , these equations take the form : @xmath173 heat flux ( @xmath174 ) and power ( @xmath175 ) for unipartite systems with external ( time dependent ) forcing were originally defined by alicki @xcite : @xmath176h\}\label{qal}\\ p&=&\textrm{tr}\left\{\rho\frac{\partial h}{\partial t}\right\}\label{pal}.\end{aligned}\ ] ] before we derive the steady state power and heat flux we wish to discuss the main differences between the semiclassical ed - jcm and the semiclassical theory of the laser due to lamb @xcite .
firstly , in lamb s model the material system ( the atom ) has two levels , whereas in the semiclassical ed - jcm the matter has three levels .
secondly , in lamb s model pumping and decay of the two lasing levels are phenomenological ( where the pumping function affects the field and thus the interaction term in the hamiltonian ) , whereas in the semiclassical ed - jcm pumping and dumping of matter population from the ground state to the two upper lasing levels is achieved through the full dissipative lindblad superoperator .
thirdly , in lamb s model the field is allowed to decay phenomenologically , whereas in the semiclassical ed - jcm discussed in this paper the field does not decay .
finally , in lamb s model , maxwell s equations for the classical field are solved self- consistently with a quantum perturbative solution of the atomic density matrix , whereas in the semiclassical ed - jcm discussed here the field is not accounted for directly . as was mentioned previously , in the semiclassical model of geva and
kosloff the field is not accounted for directly as well .
therefore , cavity damping is not incorporated , and negative steady state power in the atom signifies an increase in the field s energy .
the steady state solutions for @xmath177 and @xmath178 is @xmath179 , since @xmath180 , @xmath181 , and @xmath182 ( after applying the steady state condition @xmath183 ) .
combining the equations for @xmath184 and @xmath185 at steady state ( @xmath186 ) yields a central equation : @xmath187 where @xmath188 is the phase of the @xmath189 density matrix element .
there are now three possible physical solutions .
a. @xmath190 .
this yields : @xmath191 where @xmath192 . note that this corresponds to a very specific choice of parameters .
b. @xmath193 .
this yields a situation where there is no inversion of the atomic levels : @xmath194 moreover , it leads to a positive atomic steady state power which corresponds to attenuation of the electromagnetic field .
this is outside the scope of the current paper , and will be explored in more detail elsewhere @xcite . c. @xmath195
this yields a situation where there is an inversion of the atomic levels : @xmath196 the steady state solutions for the @xmath197 , @xmath198 , and @xmath199 density matrix elements is obtained through the solution of the following set of equations : @xmath200 the solution of equation [ atomicssmat ] can be written as : @xmath201 where @xmath202 are given in the appendix ii .
we are now in position to compare the steady state heat fluxes and power of the fully quantum model with the analytical solutions of the semiclassical model . applying alicki s definitions ( eq . [ qal ] and eq .
[ pal ] ) to the semiclassical ed - jcm at steady state yields : @xmath203 where @xmath204 is a positive constant given in the appendix .
we note that at steady state @xmath205v\}=0 $ ] , and thus @xmath206h_{m}\}$ ] . under the condition @xmath49 ,
the reservoir heat fluxes and power for the fully quantum ed - jcm are found numerically to be independent of @xmath169 for the range @xmath207 ( which corresponds to an initial coherent state ranging from no photons at all to @xmath147 photons in the cavity ) .
there are @xmath208 deviations for the higher field strength range ( where the initial number of photons in the cavity is close to @xmath147 ) due to a slightly rougher truncation of the fock space .
the analytical semiclassical hot reservoir heat flux and power in the range @xmath209 are practically independent of @xmath169 , and agree almost perfectly with the numerical steady state fluxes in the fully quantum model .
however , as @xmath169 decreases below @xmath210 the semiclassical reservoir heat fluxes and power change dramatically .
this is of course expected , as @xmath211 , and thus when the field s amplitude decreases below @xmath210 , @xmath161 is no longer much bigger than @xmath212 . under the condition @xmath49 , we find essentially perfect agreement between the numerical steady state fluxes of the fully quantum ed - jcm and the analytical steady state fluxes of the semiclassical ed - jcm .
therefore we can state that as far as thermodynamical fluxes are considered , the semiclassical ed - jcm captures the true physical picture .
one important exception is that in the semiclassical treatment , if there is no initial field present at all ( @xmath213 ) , amplification can not take place . for completeness
, we note that the steady state amplification described above is only one of several thermodynamic modes of operation of the light - matter system .
consider the expression for the steady state power : @xmath214 where @xmath215 .
a mathematically feasible solution for @xmath216 is obtained only when @xmath217 ( @xmath218 is a positive constant ) .
substituting @xmath216 into equations [ deltapi2 ] and [ powerpi2 ] reveals that atomic inversion and amplification go together hand in hand .
inversion in the two excited state levels implies negative power ( corresponding to an amplification of the electromagnetic field ) and vice versa .
however , a full solution of case b reveals that @xmath219 . in this case , a mathematically feasible solution for @xmath216 is obtained only when @xmath220 .
therefore , @xmath216 is a symmetric function of @xmath221 .
the absolute value of the atomic coherence at steady state is plotted in fig .
[ sccoherence ] for three parameter ranges .
the different thermodynamic modes of operation will be described in more detail in a forthcoming publication @xcite . .
b ) @xmath222 .
c ) @xmath223.,width=377 ] it can be seen that substantial atomic coherence is observed only when @xmath224 . what about the engine s efficiency ? in the previous section we mentioned that the ( numerical ) efficiency of the quantum amplifier always matches the ratio obtained from scovil and schulz - dubois s intuitive definition .
before we calculate the engine s efficiency , we wish to obtain carnot s formulation of the second law in differential form .
we begin with spohn s entropy production function : @xmath225 where @xmath226 is the three - level system entropy change , and @xmath227 is the entropy flux from / to the hot ( cold ) reservoir . at steady state @xmath228
we now wish to rewrite @xmath229 in terms of @xmath230 and @xmath231 .
the quantity @xmath232 , which measures the energy flux including the atomic - field interaction energy is given by : @xmath233 at steady state : @xmath234}{f}.\label{edot3lscss}\ ] ] the quantity @xmath235 is zero only at perfect atomic - field resonance .
however , the quantity @xmath236 , which measures the energy flux without the atomic - field interaction energy , and was introduced originally in @xcite , is zero at steady state , as @xmath158 does not depend on time .
expanding @xmath52 yields : @xmath237 where @xmath238h_{m}\}$ ] and @xmath239\}$ ] are the alternative definitions for heat flux and power introduced in @xcite . at steady state : ( 1 ) @xmath240 and hence @xmath241 , and ( 2 ) since @xmath205v\}=0 $ ] , @xmath242 .
therefore we can replace @xmath229 in equation [ spohn3lsc ] with @xmath243 where @xmath244 and obtain : @xmath245 which is carnot s efficiency formula in differential form .
we note that equation [ carnot3lsc ] is always true regardless of a resonance condition .
moreover , we wish to emphasize that rewriting @xmath229 in terms of @xmath230 and @xmath246 for non - resonant cases is possible only through the alternative approach to energy flux in unipartite systems discussed in @xcite .
substitution of @xmath246 ( which is identical with @xmath231 at perfect resonance ) and @xmath230 into the engine s efficiency formula at steady state ( @xmath247 ) yields : @xmath248 which is identical with the maser s efficiency defined intuitively by scovil and schulz - dubois .
our model offers a statistical description for the reservoirs , and it allows us to derive thermodynamic fluxes , which in turn yield scovil and schulz - dubois s efficiency formula .
we note that geva and kosloff @xcite also considered a semiclassical model for a three - level amplifier .
the main difference between their model and the semiclassical ed - jcm is that in geva and kosloff s model the time dependence of the classical field affects the dissipative superoperator . as a result ,
the steady state efficiency in the model by geva and kosloff depends on the power of the field , and hence it is generally not the same as in scovil and schulz - dubois s intuitive definition .
in their early work @xcite scovil and schulz - dubois asserted that the ground state population ( @xmath249 ) is bigger than the populations in the two excited states ( @xmath197 and @xmath198 ) .
this is indeed verified in appendix ii .
they also asserted that the inversion ratio between the two excited levels is given by : @xmath250 this assertion appears to be well motivated physically , since it would seem that at steady state the population ratios between the two excited levels and the pumping level should be related by boltzmann factors .
however , it turns out that this is not correct .
while it is true that the matter reaches a steady state , as seen in both the quantum and semiclassical models , the matter - field system as a whole does not reach a steady state , as was seen by solving the fully quantum model in this paper .
moreover , there is no a priori requirement of what steady state populations will be attained .
we will now demonstrate that the ratio between the two excited levels asserted by scovil and schulz - dubois is not correct . substituting the expressions for the reservoirs temperatures given in equation [ restemp ] into equation [ scovilwrong ] yields : @xmath251 substituting @xmath252 yields @xmath253 . in appendix
ii we give analytical expressions for all the density matrix elements at steady state , from which a closed formula for @xmath254 may be obtained : @xmath255 where @xmath256 are positive constants given in appendix ii . substituting @xmath257 in the analytical expression for @xmath254 yields ( similarly to the quantum model ) only a marginal inversion ratio between the two excited levels , @xmath258 for field strengths @xmath259 , respectively , where equation [ scovilwrongsub ] yields @xmath253 .
we have analyzed a fully quantum model in which a three - level material system is coupled to a single quantized cavity mode and two thermal photonic reservoirs in a framework of a heat engine .
this gives what is arguably the the simplest possible quantum model for light amplification . at the same time
, it permits a full thermodynamic analysis . unlike previous work @xcite , the field is not considered as an external time dependent force acting on the matter , but it is an integral part of the quantum system , allowing us to treat both light and matter on equal footing .
we solved the ed - jcm master equation numerically , and showed that indeed amplification of the selected cavity mode occurs even in this simple model .
however , initial field coherence is lost , as seen by the radially symmetric @xmath66 function for @xmath260 .
moreover , we find that the quantized field mode has an entropy content that changes dramatically at short times , and increases very slowly for @xmath260 . the matter - field system as a whole never reaches a steady state : at @xmath260 the energy in the field continues to increase linearly in time , which can be analyzed thermodynamically in terms of power generation from energy in the hot reservoir .
the three - level matter system , obtained by performing the partial trace of the full system over the field , does reach a steady state as seen by constant steady state energy and entropy .
another aspect of the quantum treatment that can not be dealt with at all within the framework of the semiclassical ed - jcm is entanglement .
we showed that at short times @xmath140 the matter - field system is entangled , as seen by the negative conditional entropies and the negativity of the partially transposed density matrix .
however , at longer times @xmath261 we believe that the matter - field system is classically correlated but not entangled , as the conditional entropies ( which are almost equal to the partial entropies ) and the partially transposed density matrix are both positive . based on our previous work on bipartite systems governed by a time independent master equation
@xcite we were able to derive the fundamental laws of thermodynamics .
the first law is obtained both for the full matter - field system and for the individual ( partially traced ) subsystems , using thermodynamical fluxes of heat flux and power .
the second law of thermodynamics in differential form is guaranteed to exist for the full matter - field system through spohn s @xcite entropy production function .
we define a new entropy production function @xmath262 based on matter thermodynamical fluxes . through @xmath262
we show that at steady state , when the main entropy production is due to heat fluxes from / to the heat reservoirs , carnot s efficiency formula is obtained in differential form .
a strong motivation for this work comes from an early paper by scovil and schulz - dubois @xcite in which they analyze a three - level maser as a heat engine . in their work , they intuitively defined the engine s efficiency as the ratio between the maser frequency and the pumping frequency .
however , they do not connect this efficiency with explicit expressions for work and heat , as expected from a thermodynamical analysis of a heat engine . in our quantized field treatment , the efficiency formula of scovil and schulz - dubois
was found to be in complete agreement with numerical calculations based on thermodynamical power and heat fluxes .
we have also analyzed a semiclassical version of the ed - jcm .
we obtained closed analytical expressions for power and heat flux at steady state that are in virtually perfect agreement with those obtained numerically for the fully quantum ed - jcm .
one may conclude from this that as far as steady state thermodynamical fluxes are concerned , the semiclassical model is sufficient .
furthermore , from our analytical results for power and heat flux we were able to recover scovil and schulz - dubois s efficiency formula analytically .
one of the assertions in the work of scovil and schulz - dubois is that the ratio of populations in the two excited levels is given by a product of boltzmann factors .
we showed analytically that this last assertion does not hold in general . in future work , we intend to explore further the other thermodynamic scenarios implied by the present model , both semiclassically and quantum mechanically .
of particular interest is the reversal of the present mode of operation of the engine so that it operates as a refrigerator for light .
this work was supported by the german - israeli foundation for scientific research and development .
99 h. e. d. scovil , e. o. schulz - dubois , phys . rev . lett . * 2 * , 262 ( 1959 ) . e. m. purcell , r. v. pound , phys . rev . * 81 * , 279 ( 1951 ) .
n. f. ramsey , phys . rev . * 103 * , 102 ( 1956 ) .
j. e. geusic , e. o. schulz - dubois , h. e. d. scovil , phys . rev . * 156 * , 343 ( 1967 ) .
r. alicki , j. phys .
a * 12 * , l103 ( 1979 ) .
r. kosloff , j. chem . phys . * 80 * , 1625 ( 1984 ) . e. geva , r. kosloff , phys .
e , * 49 * , 3903 ( 1994 ) .
e. geva , r. kosloff , j. chem . phys . * 104 * , 7681 ( 1996 ) .
e. boukobza , d. j. tannor , to be published .
j. von neumann , _ mathematische grundlagen der quantenmechanik _ ,
berlin : springer ( 1932 ) .
m. o. scully , m. s. zubairy , _ quantum optics _ , cambridge university press ( 1997 ) .
s. stenholm , phys .
* 6 * , 1 ( 1973 ) .
m. o. scully , w. e. lamb jr . ,
* 16 * , 853 ( 1966 ) .
m. o. scully , w. e. lamb jr .
, phys . rev . *
159 * , 208 ( 1967 ) .
j. h. mathews _ numerical methods for mathematics , science , and engineering _ , prentice - hall , inc .
( 1992 ) .
e. t. jaynes , f. w. cummings , proc .
ieee * 51 * , 89 ( 1963 ) . s. j. d. phoenix , p. l. knight , annals of physics * 186 * , 381 ( 1988 ) .
e. boukobza , d. j. tannor , phys .
a * 71 * , 063821 ( 2005 ) .
j. h. eberly , n. b. narozhny , j. j. sanchez - mondragon , phys .
* 44 * , 1323 ( 1980 ) .
j. gea - banacloche , phys .
lett . * 65 * , 3385 ( 1990 ) .
w. p. schleich , _ quantum optics in phase space _ , wilew - vch verlag berlin gmbh ( 2001 ) .
n. j. cerf , c. adami , phys .
lett . * 79 * , 5194 ( 1997 ) .
a. peres , phys .
77 * , 1413 ( 1996 ) .
h. spohn , j. math .
* 19 * , 1227 ( 1978 ) .
f. reif , _ fundamentals of statistical and thermal physics _ , mcgraw - hill ( 1965 ) .
r. loudon , _ the quantum theory of light _ , oxford university press ( 1983 ) .
w. e. lamb jr . , phys . rev . *
134 * , a1429 ( 1964 ) .
e. boukobza , d. j. tannor , in preparation .
a. j. van wonderen , phys .
a * 56 * , 3116 ( 1997 ) .
f. farhadmotamed , a. j. van wonderen , k. lendi , j. phys .
a * 31 * , 3395 ( 1998 ) .
the master equation for the resonant jaynes - cummings model ( jcm ) with atomic damping in the interaction representation is given by : @xmath263+\mathcal{l}_{d}[\mbox{\boldmath$\rho_{af}^{i}$}],\label{jcmdafi1}\ ] ] where @xmath264 is the combined atom - field density matrix in the interaction picture , and @xmath265 $ ] and @xmath266 $ ] are given by : @xmath267&=&-\frac{i}{\hbar}[\mbox{\boldmath$v^{i}$},\mbox{\boldmath$\rho^{i}_{af}$}]=-i[\lambda(\mbox{\boldmath$\sigma^{-}a^{\dag}$}+\mbox{\boldmath$\sigma^{+}a$}),\mbox{\boldmath$\rho^{i}_{af}$}]\nonumber\\ \mathcal{l}_{d}[\mbox{\boldmath$\rho_{af}^{i}$}]&= & \!\!\gamma(n_{th}\!+\!1)([\mbox{\boldmath$\sigma^{-}$},\mbox{\boldmath$\rho^{i}_{af}\sigma^{+}$}]\!\!+\!\![\mbox{\boldmath$\sigma^{-}\rho^{i}_{af}$},\mbox{\boldmath$\sigma^{+}$}])\!\!+ \!\!\gamma n_{th}\!([\mbox{\boldmath$\sigma^{+}$},\mbox{\boldmath$\rho^{i}_{af}\sigma^{-}$}]\!+\![\mbox{\boldmath$\sigma^{+}\rho^{i}_{af}$},\mbox{\boldmath$\sigma^{-}$}]),\label{jcmdafi2}\end{aligned}\ ] ] where @xmath268 and @xmath269 are atomic ( field ) creation and annihilation operators ( @xmath270 ; @xmath271 being the pauli @xmath272 matrix ) .
@xmath23 , @xmath212 , @xmath273 are the atomic - field coupling constant , weiskopf - wigner decay constant , and the number of thermal photons , respectively .
note that since equation [ jcmdafi1 ] is a master equation of a bipartite system , all the operators in equation [ jcmdafi2 ] are implicitly tensor products .
for example , @xmath274 is shorthand notation for @xmath275 .
the damped jcm master equation is usually obtained by adding the hamiltonian part and the dissipative part .
we note that van wonderen gave an analytical solution for the atomic density matrix in the damped jcm @xcite , and later studied the entropic behavior of the atom @xcite . in this appendix , we derive the full jcm master equation by applying the weak - coupling , markovian and weiskopf - wigner approximations , and using a set of unitary transformations .
the derivation of the dissipative part follows closely the derivation given by scully and zubairy .
we start with the full system ( atom - field)-bath hamiltonian in the schrdinger picture : @xmath276 where @xmath277 , @xmath278 , @xmath279 are given by : @xmath280 we denote by @xmath281 the atom - field system , by @xmath282 the bath which is composed of an infinite number of oscillators where the operators of each oscillator are denoted by subscript @xmath283 , and by @xmath284 the atomic-@xmath283th mode coupling constant .
the hat notation indicates that all operators are implicitly tensor products with the appropriate identity operators .
for example @xmath285 , @xmath286 , and @xmath287 .
the above hamiltonian is written under the rotating wave approximation ( rwa ) meaning that only energy conserving terms are considered .
note that only the atom is coupled directly to the bath modes .
the evolution of the full system - bath is purely hamiltonian : @xmath288.\ ] ] we now move to the system - bath interaction picture ( denoted by superscript @xmath289 ) : @xmath290,\label{jcmdsbi}\end{aligned}\ ] ] where : @xmath291 } , \end{aligned}\ ] ] where @xmath292 . in the derivation of equation [ jcmdsbi ] we made use of the identity
@xmath293=0 $ ] .
a perturbation expansion to second order in @xmath279 yields : @xmath294 -\frac{1}{\hbar^{2}}\int_{0}^{t}{dt'[\mbox{\boldmath$\hat{v}_{sb}^{\bar{i}}(t)$},[\mbox{\boldmath$\hat{v}_{sb}^{\bar{i}}(t')$},\mbox{\boldmath$\rho_{sb}^{\bar{i}}(t')$}]]}.\ ] ] consider the weak system - bath coupling limit , that is @xmath295 , where @xmath296 is any correlation between the system and bath which fulfills @xmath297 ( this holds for @xmath298 both in the schrdinger and interaction pictures ) . in this case
the atom - field system evolves according to : @xmath299\ } -\frac{1}{\hbar^{2}}\textrm{tr}_{b}\left\{\int_{0}^{t}{dt'[\mbox{\boldmath$\hat{v}_{sb}^{\bar{i}}(t)$},[\mbox{\boldmath$\hat{v}_{sb}^{\bar{i}}(t')$},\mbox{\boldmath$\rho_{s}^{\bar{i}}$}\otimes\mbox{\boldmath$\rho_{b}(0)$}]]}\right\}. \label{rhosnon}\ ] ] note that in equation [ rhosnon ] , @xmath300 and @xmath301 .
the explicit form of equation [ rhosnon ] is given by : @xmath302e^{i\omega_{k}t}}\nonumber\\ & -&\int_{0}^{t}dt'\sum_{k , k'}\lambda_{k}\lambda_{k ' } \left\{[\mbox{\boldmath$\hat{\bar{\sigma}}^{-}(t)\hat{\bar{\sigma}}^{+}(t')\rho_{s}^{\bar{i}}(t')$}-\mbox{\boldmath$\hat{\bar{\sigma}}^{+}(t')\rho_{s}^{\bar{i}}(t')\bar{\sigma}_{-}(t)$ } ] e^{i\omega_{k}t - i\omega_{k'}t'}\langle \mbox{\boldmath$a_{k}^{\dag}a_{k'}$}\rangle\right.\nonumber\\ & + & [ \mbox{\boldmath$\hat{\bar{\sigma}}^{+}(t)\hat{\bar{\sigma}}^{-}(t')\rho_{s}^{\bar{i}}(t')$}-\mbox{\boldmath$\hat{\bar{\sigma}}^{-}(t')\rho_{s}^{\bar{i}}(t')\bar{\sigma}^{+}(t)$ } ] e^{-i\omega_{k}t+i\omega_{k'}t'}\langle \mbox{\boldmath$a_{k}a_{k'}^{\dag}$}\rangle\nonumber\\ & + & \left.[\mbox{\boldmath$\hat{\bar{\sigma}}^{-}(t)\hat{\bar{\sigma}}^{-}(t')\rho_{s}^{\bar{i}}(t')$ } -\mbox{\boldmath$\hat{\bar{\sigma}}^{-}(t')\rho_{s}^{\bar{i}}(t')\hat{\bar{\sigma}}^{-}(t)$ } -\mbox{\boldmath$\hat{\bar{\sigma}}^{-}(t)\rho_{s}^{\bar{i}}(t')\hat{\bar{\sigma}}^{-}(t')$}\right.\nonumber\\ & + & \left.\mbox{\boldmath$\rho_{s}^{\bar{i}}(t')\hat{\bar{\sigma}}^{-}(t')\hat{\bar{\sigma}}^{-}(t)$}]e^{i\omega_{k}t - i\omega_{k'}t'}\langle \mbox{\boldmath$a_{k}^{\dag}a_{k'}^{\dag}$}\rangle\right\}+\textrm{h.c.},\label{rhosexp0}\end{aligned}\ ] ] where h.c .
refers to all terms on the rhs .
the bath density matrix is now assumed to be composed of a product of oscillatory modes each being in a thermal state , that is : @xmath303 where @xmath304 is the average number of thermal photons in the @xmath283th mode . with this assumption equation [ rhosexp0 ]
reduces to : @xmath305 e^{i\omega_{k}(t - t')}\bar{n}_{k}\right.\\ & + & \left.[\mbox{\boldmath$\bar{\sigma}^{+}(t)\bar{\sigma}^{-}(t')\rho_{s}^{\bar{i}}(t')$}-\mbox{\boldmath$\bar{\sigma}^{-}(t')\rho_{s}^{\bar{i}}(t')\bar{\sigma}^{+}(t)$ } ] e^{-i\omega_{1}(t - t')}(\bar{n}_{k}+1)\right\}+\textrm{h.c.}\label{rhosexp1}\end{aligned}\ ] ] the sum over @xmath283 is now replaced by an integral : @xmath306 where @xmath307 is the quantization volume and @xmath308 is the @xmath283th mode oscillation frequency . substituting @xmath309 ( @xmath310 is the transition dipole matrix element , and @xmath311 is the angle between @xmath310 and the electric field polarization vector ) , and integrating in the weiskopf - wigner approximation ( extending the lower limit of the integral over @xmath308 from @xmath312 to @xmath313 , and replacing @xmath314 by @xmath167 ) simplifies equation [ rhosexp1 ] : @xmath315=-\gamma n_{th } [ \mbox{\boldmath$\bar{\sigma}^{-}(t)\bar{\sigma}^{+}(t)\rho_{s}^{\bar{i}}(t)$}-\mbox{\boldmath$\bar{\sigma}^{+}(t)\rho_{s}^{\bar{i}}(t)\bar{\sigma}^{-}(t)$}]\nonumber\\ & -&\gamma(n_{th}+1)[\mbox{\boldmath$\bar{\sigma}^{+}(t)\bar{\sigma}^{-}(t)\rho_{s}^{\bar{i}}(t)$}-\mbox{\boldmath$\bar{\sigma}^{-}(t)\rho_{s}^{\bar{i}}(t)\bar{\sigma}^{+}(t)$ } ] + \textrm{h.c.},\label{rhosexp2}\end{aligned}\ ] ] where @xmath316 @xmath317 is the average number of thermal photons , and @xmath318 is the decay rate .
we now move to the system schrdinger picture : @xmath319+\mbox{\boldmath$e^{-\frac{i}{\hbar}h_{s}t}\mathcal{\bar{l}}_{d}[\rho_{s}^{\bar{i}}]e^{\frac{i}{\hbar}h_{s}t}$}.\label{jcmdss}\end{aligned}\ ] ] using the definitions for @xmath320 and @xmath321 ( after tracing out the bath ) it is easily shown that @xmath322\mbox{\boldmath$e^{\frac{i}{\hbar}h_{s}t}$}=\mathcal{l}_{d}[\mbox{\boldmath$\rho_{s}$}]$ ] .
finally , the master equation for the system in the schrdinger picture is given by : @xmath323+\mathcal{l}_{d}[\mbox{\boldmath$\rho_{s}$}]\nonumber\\ \mathcal{l}_{h}[\mbox{\boldmath$\rho_{s}$}]&=&-\frac{i}{\hbar}[\mbox{\boldmath$h_{s}$},\mbox{\boldmath$\rho_{s}$}]\nonumber\\ \mathcal{l}_{d}[\mbox{\boldmath$\rho_{s}$}]&= & \gamma(n_{th}+1)([\mbox{\boldmath$\sigma^{-}$},\mbox{\boldmath$\rho_{s}\sigma_{+}$}]+[\mbox{\boldmath$\sigma^{-}\rho_{s}$},\mbox{\boldmath$\sigma^{+}$}])+ \gamma n_{th}([\mbox{\boldmath$\sigma^{+}$},\mbox{\boldmath$\rho_{s}\sigma^{-}$}]+[\mbox{\boldmath$\sigma^{+}\rho_{s},\sigma_{-}$}]),\label{rhosexps}\end{aligned}\ ] ] where we deliberately omitted the superscript @xmath324 labeling the schrdinger picture .
to summarize , we went through the following path : @xmath325 the first transition takes us from the system - bath schrdinger picture to the system - bath interaction picture through a unitary transformation .
tracing over the bath under the weak coupling , markovian , and weiskopf - wigner approximations leads us to the system dissipative interaction picture .
finally , by applying a unitary transformation we move to the system schrdinger picture .
to complete the analysis we now move to the standard interaction picture which includes both the hamiltonian and the dissipative parts : @xmath326+\mbox{\boldmath$e^{\frac{i}{\hbar}(h_{a}+h_{f})t}$}\mathcal{l}_{d}[\mbox{\boldmath$\rho_{s}$}]\mbox{\boldmath$e^{\frac{i}{\hbar}(h_{a}+h_{f})t}$},\label{jcmdsi}\end{aligned}\ ] ] where it can be shown that : @xmath327\mbox{\boldmath$e^{\frac{i}{\hbar}(h_{a}+h_{f})t}$}=\mathcal{l}_{d}[\mbox{\boldmath$\rho_{s}^{i}$}]$ ] .
equation [ jcmdsi ] is now identical with equation [ jcmdafi2 ] , with subscript @xmath281 replacing subscript @xmath328 .
the density matrix for the semiclassical ed - jcm operating as an amplifier ( @xmath329 ) is given by : @xmath330 where @xmath202 are given by : @xmath331 @xmath332 are all positive constants , and since @xmath333 .
thus the population in the zeroth ( pumping ) level is always greater then the population in either level @xmath31 or @xmath33 . since @xmath334 , a mathematical feasible expression
is obtained only if @xmath217 . | thermodynamics of a three - level maser was studied in the pioneering work of scovil and schulz - dubois [ phys .
rev . lett . * 2 * , 262 ( 1959 ) ] . in this work
we consider the same three - level model , but treat both the matter and light quantum mechanically .
specifically , we analyze an extended ( three - level ) dissipative jaynes - cummings model ( ed - jcm ) within the framework of a quantum heat engine , using novel formulas for heat flux and power in bipartite systems introduced in our previous work [ e. boukobza and d. j. tannor , pra ( in press ) ] .
amplification of the selected cavity mode occurs even in this simple model , as seen by a positive steady state power .
however , initial field coherence is lost , as seen by the decaying off - diagonal field density matrix elements , and by the husimi - kano q function .
we show that after an initial transient time the field s entropy rises linearly during the operation of the engine , which we attribute to the dissipative nature of the evolution and not to matter - field entanglement .
we show that the second law of thermodynamics is satisfied in two formulations ( clausius , carnot ) and that the efficiency of the ed - jcm heat engine agrees with that defined intuitively by scovil and schulz - dubois .
finally , we compare the steady state heat flux and power of the fully quantum model with the semiclassical counterpart of the ed - jcm , and derive the engine efficiency formula of scovil and schulz - dubois _ analytically _ from fundamental thermodynamic fluxes . |
graphical analysis ( ga ) has been routinely used for quantification of positron emission tomography ( pet ) radioligand measurements . the first ga method for measuring primarily tracer uptakes for irreversible kinetics
was introduced by patlak , @xcite , and extended for measuring tracer distribution ( accumulation ) in reversible systems by logan , @xcite .
these techniques have been utilized both with input data acquired from plasma measurements and using the time activity curve from a reference brain region .
they have been used for calculation of tracer uptake rates , absolute distribution volumes ( dv ) and dv ratios ( dvr ) , or , equivalently , for absolute and relative binding potentials ( bp ) .
they are widely used because of their inherent simplicity and general applicability regardless of the specific compartmental model .
the well - known bias , particularly for reversible kinetics , in parameters estimated by ga is commonly attributed to noise in the data , @xcite , and therefore techniques to reduce the bias have concentrated on limiting the impact of the noise .
these include ( i ) rearrangement of the underlying system of linear equations so as to reduce the impact of noise yielding the so - called _ multi - linear _ method ( ma1 ) , @xcite , and a second multi - linear approach ( ma2 ) , @xcite , ( ii ) preprocessing using the method of generalized linear least squares ( glls ) , @xcite , yielding a hybrid glls - ga method , @xcite , ( iii ) use of the method of perpendicular least squares , @xcite , also known as total least squares ( tls ) , @xcite , ( iv ) likelihood estimation , @xcite , ( v ) tikhonov regularization @xcite , ( vi ) principal component analysis , @xcite , and ( vii ) reformulating the method of logan so as to reduce the noise in the denominator , @xcite . here , we turn our attention to another important source of the bias : the model error which is implicit in ga approaches .
the bias associated with ga approaches has , we believe , three possible sources .
the bias arising due to random noise is most often discussed , but errors may also be attributed to the use of numerical quadrature and an approximation of the underlying compartmental model .
it is demonstrated in section [ sec : theory ] that not only is bias an intrinsic property of the linear model for limited scan durations , which is exaggerated by noise , but also that it may be dominated by the effects of the model error .
indeed , numerical simulations , presented in section [ sec : validatetheory ] , demonstrate that large bias can result even in the noise - free case .
conditions for over- or under - estimation of the dv due to model error and the extent of bias of the logan plot are quantified analytically .
these lead to the design of a bias correction method , section [ sec : method ] , which still maintains the elegant simplicity of ga approaches .
this bias reduction is achieved by the introduction of a simple nonlinear term in the model . while this approach adds some moderate computational expense , simulations reported in section [ sec : resultssc0 ] for the fibrillar amyloid @xmath0 radioligand [ 11c ] benzothiazole - aniline ( pittsburgh compound - b [ pib ] ) , @xcite , illustrate that it greatly reduces bias .
relevant observations are discussed in section [ sec : diss ] and conclusions presented in section [ sec : conc ] .
for the measurement of dv , existing linear quantification methods for reversible radiotracers with a known input function , i.e. the unmetabolized tracer concentration in plasma , are based on the following linear approximation of the true kinetics developed by logan , @xcite : @xmath3 here @xmath4 is the measured _ tissue time activity curve _ (
ttac ) , @xmath5 is the _ input function _ , dv represents the _ distribution volume _ and quantity @xmath6 is a constant . with known @xmath4 and @xmath5
we can solve for dv and @xmath6 by the method of linear least squares .
this model , which we denote by ma0 to distinguish it from ma1 and ma2 introduced in @xcite , approximately describes tracer behavior at equilibrium . dividing through by @xmath4 , showing that the dv is the linear slope and @xmath7 the intercept , yields the original logan graphical analysis model , denoted here by logan - ga , @xmath8 in which the dv and intercept @xmath7 are obtained by using linear least squares ( ls ) for the sampled version of ( [ eq : logan ] ) .
although it is well - known that this model often leads to under - estimation of the dv it is still widely used in pet studies . an alternative formulation based on ( [ eq : ma0 ] )
is the so - called ma1 , @xmath9 for which the dv can again be obtained using ls @xcite .
recently another formulation , obtained by division in ( [ eq : ma0 ] ) by @xmath5 instead of @xmath4 , has been developed by zhou _
et al _ , @xcite .
but , as noted by varga _
et al _ in @xcite the noise appears in both the independent and dependent variables in ( [ eq : logan ] ) and thus tls may be a more appropriate model than ls for obtaining the dv . whereas it has been concluded through numerical experiments for tracer [ @xmath10f]fcway and [ @xmath11c]mdl 100,907 , @xcite , that ma1 ( [ eq : multi - lin1 ] ) performs better than other linear methods , including logan - ga ( [ eq : logan ] ) , tls and ma2 @xcite , none of these techniques explicitly deals with the inherent error due to the assumption of model ma0 ( [ eq : ma0 ] ) .
the focus here is thus examination of the model error specifically for logan - ga and ma1 , from which a new method for reduction of model error is designed . the general three - tissue compartmental model for the reversible radioligand binding kinetics of a given brain region or a voxel can be illustrated as follows , @xcite : ( 9.200000,3.400000)(0.000000,-3.400000 ) ( 0.0000,-1.2000)(2.0000,1.2000)[c ] ( 2.0000,-1.0000)(1.6000,0.8000)[c ] ( 3.6000,-1.2000)(2.0000,1.2000)[c ] ( 3.6000,-0.7750)(-1,0)1.6000 ( 2.0000,-0.4250)(1,0)1.6000 ( 5.6000,-1.0000)(1.6000,0.8000)[c ] ( 7.2000,-1.2000)(2.0000,1.2000)[c ] ( 7.2000,-0.7750)(-1,0)1.6000 ( 5.6000,-0.4250)(1,0)1.6000 ( 3.6000,-2.2000)(2.0000,1.0000)[c ] ( 3.6000,-3.4000)(2.0000,1.2000)[c ] ( 4.7750,-2.2000)(0,1)1.0000 ( 4.4250,-1.2000)(0,-1)1.0000 here @xmath5 ( kbq / ml ) is the input function , i.e. the unmetabolized radiotracer concentration in plasma , and @xmath12 , @xmath13 and @xmath14 ( kbq / g ) are free radioactivity , nonspecific bound and specific bound tracer concentrations , resp . , and @xmath15 ( ml / min / g ) and @xmath16 ( 1/min ) , @xmath17 , are rate constants . the dv is related to the rate constants as follows @xcite , @xmath18 the numerical implementation for estimating the unknown rate constants of the differential system illustrated in figure [ fig:3 t ] is difficult because three exponentials are involved in the solution of this system , @xcite .
specifically , without the inclusion of additional prior knowledge , the rate constants may be unidentifiable , @xcite .
fortunately , for most tracers it can safely be assumed that @xmath19 and @xmath20 reach equilibrium rapidly for specific binding regions .
then it is appropriate to use a two - tissue four - parameter ( 2t-4k ) model by binning @xmath13 and @xmath12 to one compartment @xmath21 .
this is equivalent to taking @xmath22 , and hence @xmath23 . on the other hand , for regions without specific binding activity , we know @xmath24 which is equivalent to taking @xmath25 , and it is again appropriate for most radioligands to bin @xmath13 and @xmath12 . the one - tissue compartmental model is then appropriate for regions without specific binding activity .
for some tracers , however , for example the modeling of pib in the cerebellar reference region , the best data fitting is obtained by using the 2t-4k model without binning @xmath13 and @xmath12 , @xcite . assuming the latter , the dv is given by @xmath26 , and @xmath27 , for regions with and without specific binding activity , resp . ignoring the notational differences between the two models , for regions with and without specific binding activity , they are both described by the same abstract mathematical 2t-4k model equations . here , without loss of generality , we present the 2t-4k model equations for specific binding regions , @xmath28 to obtain the equations appropriate for regions without specific binding activity , @xmath14 is replaced by @xmath13 and @xmath29 and @xmath30 are interpreted as the association and dissociation parameters of regions without specific binding activity . to simplify the explanation @xmath14 , @xmath29 and @xmath30
are used throughout for both regions with and without specific binding activity , with the assumption that @xmath14 , @xmath29 and @xmath30 should automatically be replaced by @xmath13 , @xmath31 and @xmath32 respectively , when relevant . the solution of the linear differential system ( [ eq:2t4kc1])-([eq:2t4kcs ] )
is given by @xmath33 where @xmath34 represents the convolution operation , @xmath35 the overall concentration of radioactivity is @xmath36 integrating ( [ eq:2t4kc1])-([eq:2t4kcs ] ) and rearranging yields @xmath37 this is model ( [ eq : ma0 ] ) when @xmath14 is linearly proportional to @xmath4 for a time window within the scan duration of @xmath38 minutes .
the accuracy of linear methods based on ( [ eq : ma0 ] ) is thus dependent on the validity of the assumption that @xmath14 and @xmath39 are approximately linearly proportional to @xmath4 over a time window within @xmath40 $ ] .
logan observed that @xmath39 and @xmath14 are roughly proportional to @xmath4 , after some time point @xmath41 , @xcite .
if the assumption of linear proportionality breaks down for the given window , @xmath42 $ ] , bias in the estimated uptake rate or dv will be introduced , as shown later in section [ sec : resultssc0 ] , due to the intrinsic model error of a ga method . indeed , in section
[ subsec : equilibrium ] we show that , for the pib radioligand on some regions with small @xmath30 , there is no window within a @xmath2 minutes scan duration where @xmath14 and @xmath4 are linearly proportional .
this is despite the apparent good linearity , visually , of the logan plot of @xmath43 against @xmath44 . waiting for equilibrium , which may take several hours , is impractical in terms of patient comfort , cost and measurement of radioactivities .
the limitation of the constant approximation can be analysed theoretically .
because @xmath45 and @xmath5 is very small for large time the convolution @xmath46 is relatively small .
we can safely assume that the ratio of @xmath47 to @xmath48 is roughly @xmath49 for @xmath50 .
then @xmath14 , see equation ( [ sol : cs ] ) , is approximately proportional to @xmath51 for @xmath50 . in our tests with pib , the neglected component @xmath52 is less than @xmath53 for @xmath54 min .. on the other hand , this is not the case for @xmath39 , see equation ( [ sol : c_1 ] ) , because @xmath55 and @xmath56 need not be of the same scale .
for example , if @xmath57 we know @xmath58 from ( [ symb : a1 ] ) , thus @xmath59 .
specifically , @xmath60 may not be small in relation to @xmath61 .
thus , it is not appropriate , as is assumed for the logan - ga ( [ eq : logan ] ) and other linear methods derived from ma0 , to approximate @xmath62 as constant for @xmath63 $ ] .
one may argue that if @xmath64 is close to @xmath65 the term @xmath66 in @xmath4 could be ignored .
then the ratio of @xmath4 to @xmath14 would be close to constant after @xmath41 , and the resulting estimates of the dv using logan - ga ( [ eq : logan ] ) and ma1 ( [ eq : multi - lin1 ] ) would be reasonable . while it is easy to verify that @xmath64 is positive and bounded above by one , this fraction need not be close to its upper bound .
indeed , for realistic test data , see table [ tab : ki ] , @xmath67 .
the simulations presented in tables [ tab : overest ] and [ tab : sc0 ] validate that a small value of this fraction may cause a problem in the estimation of the dv using the linear logan - ga and ma1 methods .
it is immediate using @xmath68 , and positivity of both @xmath69 and @xmath70 , that @xmath71 is bounded above and below , @xmath72 and @xmath73 determines the variation in @xmath71 .
if @xmath30 is small the bound is not tight and the dv estimated by logan - ga , or a linear method derived from ma0 , may not be accurate , see for example the regions of interest ( rois ) * 1 * , * 3 * and * 6 * in the test examples reported in table [ tab : ki ] .
we reiterate that , by the discussion above , the variation for roi * 6 * , within which no specific binding activity exist , is determined by @xmath74 .
this relationship between the size of @xmath30 and the bias in the logan - ga estimate of the dv is illustrated in figure [ fig : bias_k4 ] of section [ subsec : equilibrium ] for the test data of table [ tab : ki ] .
the complete mathematical result for the model error of logan - ga and ma0 is presented in the _
appendix_. similar results , omitted here to save space , can be obtained for ma1 .
the main conclusion is that both logan - ga and ma0 can lead to an over - estimation of the dv .
this contrasts the standard view of these methods .
we summarize in the following theorem , for which the main idea is to show that replacing ( [ b_range ] ) which occurs on the right hand side of ( [ eq : sfunc ] ) by a constant intercept @xmath6 introduces an error in the least squares solution for the dv which can be specifically quantified .
[ thm : logan_bias ] suppose logan - ga , or respectively ma0 , are used for noise - free data acquired for @xmath75 frames with frame time @xmath76 and @xmath77 .
then , with @xmath71 as defined in ( [ b_range ] ) , for each method the same conclusions are reached : * the dv is over - estimated ( under - estimated ) if @xmath78,$ ] is a non - constant decreasing ( increasing ) function , and * the dv is exact if @xmath78,$ ] is a constant function ; let @xmath79 be the true value of the dv , and define the variation of a function over @xmath80 $ ] by @xmath81}{\mathbf { x}}(t)-\displaystyle\min_{t \in [ t_l , t_n ] } { \mathbf { x}}(t)|.\ ] ] then the bias in @xmath82 calculated by logan - ga is bounded by @xmath83 where @xmath84 .
this theorem is an immediate result of lemma [ lemma : ma1_ana ] and corollary [ cor : logan - ga ] in the _ appendix _ for the vectors obtained from the sampling of the functions @xmath85 @xmath86 at discrete time points @xmath87 the relevant vectors are defined by @xmath88 , @xmath89 , @xmath90 , where the division corresponds to componentwise division .
it is easy to check that all these vectors are positive vectors , @xmath91 , @xmath92 , @xmath93 and @xmath94 are non - constant increasing vectors and @xmath95 is decreasing .
thus all conditions for lemma [ lemma : ma1_ana ] and corollary [ cor : logan - ga ] are satisfied .
note that in the denominator of ( [ logandv_bound ] ) the simplification @xmath96 is used . in the latter discussion we may use the variation ( increasing or decreasing ) of @xmath70 instead of that of @xmath71 because @xmath97 it is not surprising that the properties of logan - ga and ma0 are similar . indeed ,
ma0 is none other than weighted logan - ga with weights @xmath98 , which changes the noise structure in the variables .
in contrast to the conventional under - estimation observations , it is suprising that the dv may be over - estimated .
however , the over - estimation is indeed observed in the tests presented in section [ sec : overest ] and [ sec : resultssc0 ] .
inequality ( [ logandv_bound ] ) indicates that logan - type linear methods will work well for data for which @xmath99 is flat .
unfortunately , @xmath99 may become flat only for a late time interval .
thus our interest , in section [ sec : method ] , is to better estimate the dv using a reasonable ( practical ) time window , which may include the window over which @xmath70 is still increasing .
our initial focus is on the modification of logan - type methods .
then , in section [ sec : validatetheory ] we present numerical simulations using noise - free data which illustrate the difficulties with logan - ga and ma1 , and support the results of theorem [ thm : logan_bias ] .
in the previous discussion we have seen the theoretical limitations of the logan - ga and ma1 methods . here
we present a new model and associated algorithm which assists with reducing the bias in the estimation of the dv .
observe that , @xmath100 , implies that @xmath101 , where @xmath102 can be ignored for @xmath50 .
therefore , for @xmath50 ( [ eq : root ] ) can be approximated by a new model as follows @xmath103 where @xmath104 and @xmath105 .
this suggests new algorithms should be developed for estimation of parameters dv , @xmath106 , @xmath107 and @xmath108 . here ,
a new approach , based on the basis function method ( bfm ) in @xcite , in which @xmath108 is discretized , is given by the following algorithm .
[ alg : biascor_guo]given @xmath109 and @xmath98 for @xmath110 and @xmath77 , the dv is estimated by performing the following steps . 1 .
calculate @xmath111 and intercept @xmath7 , using logan - ga .
2 . set @xmath112 and @xmath113 if @xmath114 otherwise @xmath115 .
3 . form discretization @xmath116 , @xmath117 for @xmath108 , with equal spacing logarithmically between @xmath118 and @xmath119 .
for each @xmath120 solve the linear ls problem , i.e. cast it as a multiple linear regression problem with @xmath121 as the dependent variable .
@xmath122 with data at @xmath123 , @xmath124 , to give values @xmath125 , @xmath126 and @xmath127 . : 5 .
determine @xmath128 for which residual is minimum over all @xmath120 .
set @xmath111 , @xmath106 and @xmath107 to be @xmath129 , @xmath130 and @xmath131 , resp .
* remarks : * + 1 .
[ remark : alpha1 ] the interval for @xmath108 is determined as follows : first the lower bound @xmath132 for @xmath108 is suitable for most tracers , but could be reduced appropriately .
this lower bound is not the same as that on @xmath133 used in bfm , in which @xmath133 is required to be greater than the decay constant of the isotope , @xcite .
second by point ( [ item : est_b ] ) of corollary [ cor : logan - ga ] in the _ appendix a _ , @xmath6 should be positive and near the average value of @xmath71 , where , by ( [ ineq : bounds_s ] ) , @xmath134 .
on the other hand , @xmath135 if @xmath136 is small relative to @xmath137 .
thus , @xmath108 is linked with @xmath6 through @xmath71 .
this is used to give the estimate of the upper bound on @xmath108 . practically , it is possible that the logan - ga may yield an intercept @xmath138 , then we set @xmath115 .
2 . numerically , because @xmath139 is much larger than both @xmath4 and @xmath14 for @xmath50 , the estimate of dv is much more robust to noise in the formulation , including both model and random noise effects , than are the estimates of @xmath106 and @xmath107 . therefore , while @xmath106 and @xmath107 may not be good estimates of @xmath140 and @xmath141 , resp . for noisy data ,
the estimate of dv will still be acceptable .
consequently , it is possible that logan - ga and ma0 will produce reasonable estimates for dv , even when the model error is non negligible .
3 . the algorithm can be accelerated by employing a coarse - to - fine multigrid strategy .
the coarser level grid provides bounds for the fine level grid .
the grid resolution can be gradually refined until the required accuracy is satisfied .
we present a series of simulations which first validate the theoretical analysis of section [ sec : theory ] for noise - free data , and then numerical experiments which contrast the performance of algorithm [ alg : biascor_guo ] with logan - ga , ma1 and nonlinear kinetic analysis ( ka ) algorithms for noisy data .
we assume the radioligand binding system is well modeled by the 2t-4k compartmental model and focus the analysis on the bias in the estimated dv which can be attributed to the simplification of the 2t-4k model . for the simulation
we use representative kinetic parameters for brain studies with the pib tracer .
these kinetic parameters , detailed in table [ tab : ki ] , are adopted from published clinical data , @xcite .
the simulated regions include the posterior cingulate ( pcg ) , cerebellum ( cere ) and a combination of cortical regions ( cort ) .
the kinetic parameters of each roi are also associated with the subject medical condition , namely normal controls ( nc ) and alzheimer s disease ( ad ) diagnosed subjects .
the kinetic parameters for the first seven rois are from @xcite while the last four are from @xcite .
rate constants for rois * 5 * to * 11 * are directly adopted from the published literature , while those for rois * 1 * to * 4 * are rebuilt from information provided in @xcite .
the values for rois * 1 * to * 4 * and * 8 * to * 11 * represent average values for each group , while those for rois * 5 * and * 6 * are derived from one ad subject and those for roi * 7 * from another ad subject . .rate constants for eleven rois , including pcg , cere , and cort , for ad and nc adopted from @xcite . for rois * 6 * , * 7 * , * 10 * and * 11 * no specific binding activity is assumed , i.e. @xmath25 , @xmath142 ; while for rois * 1 * to * 5 * , * 8 * and * 9 * we assume that the free and nonspecific compartments rapidly reach equilibrium , i.e. @xmath22 , @xmath143 . coefficients @xmath144 and @xmath145 are defined in ( [ symb : a1 ] ) .
the values for rois * 1 * to * 4 * and * 8 * to * 11 * represent average values for each group , while those for rois * 5 * and * 6 * are derived from one ad subject and those for roi * 7 * from another ad subject . [ tab : ki ] [ cols="^,^,^,^,^,^,^,^,^,^",options="header " , ] in contrasting the results with respect to only the bias in the calculation of the dv it is clear that algorithm [ alg : biascor_guo ] leads to significantly more robust solutions than logan - ga1 and ma1 for noise - free data . on the other hand
, the ka approach can lead to very good solutions , comparable and perhaps marginally better than algorithm [ alg : biascor_guo ] . for roi * 6 * , for which the ka solution is significantly better , we recall that the solution depends on the initial values of the parameters . changing the initial @xmath32 to @xmath146 , the resulting bias in the dv of roi * 6 * calculated by ka is increased to @xmath147 . on the other hand , algorithm [ alg : biascor_guo ] is not dependent on specifying initial values , and is thus more computationally robust . while the results with noise - free data support the use of algorithm [ alg : biascor_guo ] , it is more critical to assess its performance for noise - contaminated simulations . the experimental evaluation for noisy data is based on the noise - free input @xmath148 and noise - free output @xmath4 , one output ttac for each of the eleven parameter sets given in table [ tab : ki ] .
noise contamination of the input function and these ttacs is obtained as follows . for a given noise - free decay - corrected concentration ttac , @xmath4 , gaussian ( @xmath149 ) noise at each time point @xmath123 is modeled using the approach in @xcite . the standard deviation in the noise at each time point @xmath123 , depends on the frame time interval @xmath150 in seconds , the tracer decay constant @xmath151 ( @xmath152 for @xmath153 ) and a scale factor @xmath154 @xmath155 the resulting coefficients of variation @xmath156 ( ratio @xmath157 to @xmath98 ) , for scale factors @xmath65 and @xmath158 , are illustrated in figure [ fig : cv ] .
the noise in the input function can be attributed to two sources , system and random noise .
although the random @xmath159-ray emission follows a poisson distribution , we use the limiting result that a large mean poisson distribution is approximately gaussian to model this randomness as gaussian .
thus both sources are modeled as gaussian but with different variance . consider first the following model for determining the randomness of the @xmath160ray emissions .
suppose a @xmath161 ml blood sample is placed in a @xmath159-ray well counter which has efficiency @xmath162 and the measured counts over @xmath163 seconds are @xmath164
. then the measured decay corrected concentration ( kbq / ml ) is @xmath165 where @xmath166 is a normalization factor to convert the counts to `` kilo '' counts .
then , assuming that the mean of @xmath109 ( or its true value ) is @xmath167 as given in ( [ inpf.eq ] ) , the standard deviation in the measurement of @xmath109 due to random effects is @xmath168 .
the coefficient of variation , c@xmath169 , which results from this random noise is shown in figure [ fig : cv ] .
it is assumed in the experiments that each blood sample has volume @xmath170 , the count duration is @xmath171 seconds and the well counter efficiency is @xmath172 .
then , denoting the coefficient of variation due to system noise by c@xmath173 , the noise - contaminated input is given by @xmath174 where @xmath175 is selected from a standard normal distribution ( g@xmath176 ) , and in the simulations we use c@xmath177 , see figure [ fig : inpf ] .
two hundred random noise realizations are generated for each input - ttac pair , and for each noise level ( @xmath178 , @xmath158 ) .
the distribution volume is calculated for each experimental pair using logan - ga , ma1 , ka and algorithm [ alg : biascor_guo ] . in each case
two scan durations are considered , @xmath179 and @xmath2 minutes respectively , and @xmath180 minutes .
unlike the noise - free case , the numerical quadrature for @xmath181 uses only the samples at scan points @xmath109 .
we present histograms for the percentage relative error of the bias @xmath182 in order to provide a comprehensive contrast of the methods . figure
[ fig : dv - histall ] shows the histograms for all eleven rois , with the range of the error for each method indicated in the legend .
the figures ( a)-(b ) are for scan windows of @xmath2 minutes , for noise scale factors @xmath178 and @xmath183 while ( c)-(d ) are for scan windows of @xmath1 minutes .
figure [ fig : dv - hist3 ] provides equivalent information for a representative cortical region roi * 3*. it is clear that the distributions of the relative errors for ka and ma1 are far from normal ; ka has a significant positive tail while logan - ga has strong negative bias .
ma1 has unacceptably long tails except for the case of low noise with long scan duration , i.e. @xmath178 with @xmath2 minutes scan duration . on the other hand ,
the histogram for algorithm [ alg : biascor_guo ] is close to a gaussian random distribution ; the mean is near zero and the distribution is approximately symmetric .
moreover , algorithm [ alg : biascor_guo ] performs well , and is only outperformed marginally by ma1 for the lower noise and longer time window case . on the other hand , there are some situations , particularly for ma1 , in which the relative error is less than @xmath184 ; in other words , the calculated dvs are negative .
such _ unsuccessful _ results occur only for the higher noise level ( @xmath183 ) . while there was only one such occurrence for the logan - ga ( @xmath1 min . with roi * 9 * ) , there were @xmath185 such occurrences for ma1 , @xmath186 for the shorter time interval of @xmath1 minutes ( rois * 1 * , * 3 * , * 4 * , * 5 * , * 6 * , * 8 * and * 9 * ) and @xmath187 for the longer interval of @xmath2 minutes , ( rois * 1 * and * 6 * ) .
the reason for the negative dv for ma1 is discussed in section [ subsec : negdv ] .
from the results for the higher noise @xmath183 we conclude that algorithm [ alg : biascor_guo ] using the shorter @xmath1 minutes scan duration outperforms the other algorithms , even in comparison to their results for the longer scan duration .
obviously algorithm [ alg : biascor_guo ] is more expensive computationally than logan - ga and ma1 . in the simulations , the average cpu time , in seconds , per ttac was @xmath188 , @xmath189 , @xmath190 and @xmath191 , for logan - ga , ma1 , ka and algorithm [ alg : biascor_guo ] , respectively . the high cost of the ka results from the requirement to use a nonlinear algorithm .
because the ka requires a good initial estimate for the parameters the cost is variable for each ttac ; it is dependent on whether the supplied initial value is a good initial estimate .
indeed the ka results take from @xmath192 to @xmath193 seconds , while the costs using the other methods are virtually ttac independent .
the graphical analysis methods of logan - type rely on the assumption that the ratio @xmath14 to @xmath4 is approximately constant within a chosen window @xmath42 $ ] . this ratio is plotted against time for the simulated data for rois * 1 * to * 11 * in figure [ fig : cs2ct90 ] .
it is clear that the ratios for rois * 1 * , * 3 * and * 6 * have not reached equilibrium even by @xmath2 minutes .
these are the three data sets with the largest bias reported in section [ sec : overest ] and with smallest @xmath30 ( resp .
@xmath32 ) .
it is certain that equilibrium is eventually reached .
these curves first increase to a peak at about @xmath194 minutes for rois * 1 * and * 3 * and at about @xmath195 minutes for roi * 6 * and then decrease before reaching approximately constant values ( figure [ fig : cs2ct720 ] ) . on the other hand , increasing the scan duration to more than two hours is not practical .
moreover , as illustrated in figure [ fig : ga - plot ] , using the linearity of @xmath196 versus @xmath197 to verify whether equilibrium has been reached may be misleading .
for example , it would appear that all eleven data sets have achieved equilibrium after roughly @xmath198 minutes .
the arrow in figure [ fig : ga - plot ] points to the marker corresponding to the data calculated at the middle point of the frame from @xmath198 to @xmath185 minutes .
( y - axis ) against @xmath199 ( x - axis ) for all test rois except rois * 3 * , * 5 * and * 11 * for the first @xmath2 minutes .
the last eight points correspond to the time interval @xmath198 to @xmath2 minutes .
the curves for rois * 3 * , * 5 * and * 11 * are similar to those for rois * 1 * , * 4 * and * 10 * resp .. the arrow points to the first frame falling in this interval for roi * 6*. ] we illustrate the relation between the bias in the estimate of dv calculated by logan - ga and @xmath30 in figure [ fig : bias_k4 ] .
as discussed in section [ subsec : errana ] , a small value of @xmath30 may cause a large variation in @xmath71 .
this graph verifies that the magnitude of the bias decreases as @xmath30 increases , further verifying that large bias in dv may arise purely due to modeling assumptions in the absence of noise in the data .
for the eleven rois , assuming noise - free data , a scan duration of @xmath2 minutes and @xmath180 minutes . the specific data pairs ( @xmath30 , bias )
are , for rois * 1 * to 11 , respectively , @xmath200 , @xmath201 , @xmath202 , @xmath203 , @xmath204 , @xmath205 , @xmath206 , @xmath207 , @xmath208 , @xmath209 , and @xmath210 . ]
both logan - ga and ma1 , ( [ eq : logan ] ) and ( [ eq : multi - lin1 ] ) resp .
, require the calculation of integrals @xmath211 and @xmath181 .
assume the noise - free measurements @xmath98 are derived from the integral over the @xmath212th frame duration .
thus we can easily recover its integral without introducing error while quadrature error for calculation of @xmath181 due to using a limited number of plasma samples is unavoidable .
the accuracy of the numerical quadrature impacts the accuracy of the parameter estimates .
note that we classify the noise effects as another source of bias in dv .
we recalculate the dv for the experiments reported in section [ sec : overest ] , but now using numerical quadrature for calculation of @xmath181 with data sampled one time point per time frame .
the bias for each roi of the estimated dv using @xmath2 minutes scan data with @xmath180 minutes is @xmath213 , @xmath214 , @xmath215 , @xmath216 , @xmath217 , @xmath218 , @xmath219 , @xmath220 , @xmath221 , @xmath222 and @xmath223 when calculated using logan - ga , and @xmath224 , @xmath225 , @xmath226 , @xmath227 , @xmath228 , @xmath229 , @xmath230 , @xmath231 , @xmath232 , @xmath233 and @xmath234 calculated using ma1 .
it is interesting to note that the dv calculated for roi * 7 * is no longer an over - estimate .
this does not contradict the result of theorem [ thm : logan_bias ] , which predicts that the dv for roi * 7 * will be over - estimated due to model error , provided that the other aspects of the calculation are accurate . now using a less accurate quadrature
the negative bias due to quadrature error canceled the positive bias due to the model error .
indeed , for all eleven test cases the impact of the less accurate quadrature is to shift the bias down , i.e. it is more negative as compared to the equivalent more accurate calculations shown in table [ tab : overest ] . in the eleven simulated rois , large under - estimation of the dv calculated by logan - ga and ma1
is observed for rois * 1 * ( nc cort ) , * 3 * ( nc pcg ) and * 6 * ( ad cere ) . a lower value of the dv in the cortical regions of ncs and in the cerebellum for ad subjects will result in under - estimation of the dvr for ncs and over - estimation of the dvr for ad subjects when the cerebellum is used as the reference region for the dvr calculation .
thus , the difference between ad and nc can be artificially enhanced , and viewed as a positive outcome associated with the bias of logan - ga and ma1 .
this conclusion , however , can not be generalized .
it is unknown whether it is always the case that ad / nc have small / large @xmath32 in cerebellar regions and relatively large / small @xmath30 in cortical regions .
confirmation of these assertions would suggest , based on the discussion in sections [ subsec : errana ] and [ subsec : equilibrium ] , that the dvr is over - estimated for ad subjects and under - estimated for healthy subjects ( also see figure [ fig : bias_k4 ] ) .
in addition , more subtle differences , such as the ones between mild cognitive impairment ( mci ) and nc , or among nc with differential genetic risk for ad , may make the effects of bias much less predictable .
consequently , we evaluate the quantification methods based on their bias because the goal of these methods is to estimate the dv as accurately as possible .
as noted in section [ subsec : noisy ] , ma1 generates some results with negative dvs .
such results are reported as _
unsuccessful _ in ichise s original paper @xcite
. careful study of these results shows that the negative dvs arise when @xmath235 has the wrong sign . for most radioligand binding studies
@xmath236 is a small positive number because @xmath237 , which is usually larger than @xmath238 , see remark ( [ remark : alpha1 ] ) of algorithm [ alg : biascor_guo ] .
thus a small error in the estimate of @xmath235 due to large noise in the data may change its sign .
this in turn impacts the sign of the estimate of the dv .
in this article , we quantified the model error in estimating distribution volume using graphical analysis methods .
we described the conditions under which the dv is either over- or under - estimated , and quantified the bias caused by model error .
we validated our findings through simulations with noise - free data . to reduce the impact of model error
, we added a simple nonlinear term to the fundamental linear model ma0 , and presented a new algorithm for its solution .
simulations with noisy data demonstrate that the new algorithm is cost - effective and robust even for shorter scan durations . for pib - pet studies ,
the new method using shorter scan data ( @xmath179 minutes ) outperforms , or is at least as good as , logan - ga , ma1 and ka methods using longer scan data ( @xmath2 minutes ) .
the proposed approach can be easily extended for dvr estimation .
this is a focus of our future work .
this work was supported by grants from the state of arizona ( to drs .
guo , reiman , chen and renaut ) , the nih ( r01 ag031581 , r01 mh057899 and p30 ag19610 to dr .
reiman ) and the nsf ( dms 0652833 to dr . renaut and dms 0513214 to drs .
renaut and guo ) .
the authors thank researchers from the university of pittsburgh for their published findings , including information about pib input function and rate constants .
here we present the theoretical result from which theorem [ thm : logan_bias ] is obtained .
we use the notation that @xmath239 , @xmath145 , @xmath240 , @xmath241 and @xmath242 , @xmath243 , @xmath240 , @xmath244 , are vectors with entries @xmath245 and @xmath246 , resp .
the notation @xmath247 and @xmath248 denotes component wise division and multiplication , namely entries @xmath249 and @xmath250 , @xmath251 is @xmath252 and @xmath253 is the euclidean norm .
we call @xmath254 decreasing ( * increasing * ) if @xmath255 ( @xmath256 ) , and non - constant decreasing ( * non - constant increasing * ) if it is decreasing ( * increasing * ) and at least one of the @xmath257 ( @xmath258 ) signs is strict , @xmath259 ( @xmath260 ) . if all of the @xmath257 ( @xmath258 ) signs are strict , we call @xmath254 strictly decreasing ( * strictly increasing * )
. a vector @xmath254 is constant if @xmath261 for some constant @xmath262 and for all @xmath212 .
( chebyshev s sum inequality @xcite ) given real numbers + @xmath255 and @xmath263 then @xmath264 similarly , if @xmath255 and @xmath265 then @xmath266 in the above chebyshev s sum inequalities the numbers are not required to be positive and the equality is true if and only if one of the two vectors , @xmath254 or @xmath267 , is a constant vector .
if @xmath254 and @xmath267 are positive vectors , the chebyshev s sum inequalities can be expressed as @xmath268 and @xmath269 .
[ lemma : baseineq ] if @xmath91 , @xmath95 and @xmath270 are positive real vectors , of which @xmath91 is a increasing vector and @xmath95 is a decreasing vector , then 1 .
[ lemma2_1 ] @xmath271 if @xmath272 is a non - constant increasing vector .
the inequality is strict if @xmath91 is _ strictly _ increasing .
[ lemma2_2]@xmath273 if @xmath274 is a non - constant decreasing vector .
the inequality is strict if @xmath91 is _ strictly _ increasing .
[ lemma2_3]@xmath275 if @xmath274 is a constant vector , 4 .
[ lemma2_4 ] @xmath276 if @xmath277 is a non - constant decreasing vector .
the inequality is strict if @xmath91 is _ strictly _ increasing .
[ lemma2_5]@xmath278 if @xmath277 is a non - constant increasing vector .
the inequality is strict if @xmath91 is _ strictly _ increasing .
[ lemma2_6]@xmath279 if @xmath277 is a constant vector .
we only prove the first case .
the proof for the other items follows similarly .
we use mathematical induction . for the lowest dimension @xmath280 ,
@xmath281 the last reduction follows from the monotonicity of @xmath91 , @xmath95 , which implies @xmath282 , and the non - constant increasing assumption of @xmath274 , which guarantees @xmath283 . when @xmath91 is strictly increasing @xmath284 .
under this condition @xmath285 for @xmath280 . assuming the inequality @xmath271 is true for dimension @xmath286 , i.e. @xmath287 then for @xmath288 @xmath289 the last reduction is based on the monotonicity of @xmath290 and @xmath274 .
when @xmath91 is strictly increasing @xmath291 for all @xmath292 the inequality will be strict because at least one of the terms @xmath293 is positive based on the monotonicity condition .
the result thus follows by induction for all integers @xmath294 .
the following corollary now follows immediately by observing that @xmath272 increases when @xmath270 increases and @xmath295 decreases when @xmath270 decreases .
[ cor : ineq_monos ] if @xmath91 , @xmath95 and @xmath270 are positive real vectors , of which @xmath91 is a strictly increasing vector and @xmath95 is a decreasing vector , then 1 .
@xmath296 if @xmath270 is a decreasing vector .
2 . @xmath297 if @xmath270 is an increasing vector .
[ lemma : ma1_ana ] if @xmath91 , @xmath95 , @xmath93 and @xmath270 are positive real vectors , of which @xmath91 is strictly increasing , @xmath95 is decreasing , and @xmath91 , @xmath93 , @xmath270 and @xmath298 satisfy @xmath299 ; and @xmath300=\mathrm{argmin}\|{{\mathbf p}}x - b{{\mathbf q}}-{{\mathbf r}}\|_2 ^ 2 $ ] ; then 1 .
the estimated solution @xmath301 and exact solution @xmath298 are related by * @xmath302 if @xmath274 is a non - constant decreasing vector , * @xmath303 if @xmath274 is a non - constant increasing vector , * @xmath304 if @xmath274 is a constant vector ; 2 .
the following inequality is true without any monotonicity assumptions : @xmath305 3 .
[ lemma_bsign ] the sign of the intercept @xmath306 is determined as follows : * @xmath307 if @xmath277 is a non - constant decreasing vector , * @xmath308 if @xmath277 is a non - constant increasing vector , * @xmath309 if @xmath277 is a constant vector ; 4 .
given @xmath310 , the ls solution of @xmath311 for @xmath6 is @xmath312 ; 5 . given @xmath312 , the ls solution of @xmath311 for @xmath313 and the true solution @xmath298 have the same relationship as stated in the first conclusion of this theorem .
it is easy to verify that the ls solution of @xmath314 is @xmath315 the proof then follows as outlined below : 1 .
replace @xmath93 in the expression for @xmath301 with @xmath316 .
then @xmath317 and the results immediately follow from lemma [ lemma : baseineq ] ( [ lemma2_1])-([lemma2_3 ] ) and the fact @xmath318 when @xmath91 is not linear proportional to @xmath95 .
because @xmath319 and similarly @xmath320 we have @xmath321 using the fact @xmath322 and ( [ eq : hatx ] ) , we conclude the inequality is true .
again we replace @xmath93 with @xmath316 , then the expression for @xmath306 becomes @xmath323 the results immediately follow from lemma [ lemma : baseineq ] ( [ lemma2_4])-([lemma2_6 ] ) and the fact @xmath318 when @xmath91 and @xmath95 do not have the same direction .
4 . this result is easily verified .
5 . given @xmath312 ,
the ls solution of @xmath311 for @xmath313 is @xmath324 the results now follow from lemma [ lemma : baseineq ] .
we now transform the exact equation to @xmath325 and rewrite the results using vectors @xmath89 , @xmath90 and @xmath88 .
correspondingly , we find the ls solution of @xmath326 for @xmath327 , @xmath328 .
[ cor : logan - ga ] if @xmath92 , @xmath94 and @xmath329 are positive , of which @xmath92 is strictly increasing , @xmath92 , @xmath94 , @xmath329 and @xmath298 satisfy @xmath330 ; and @xmath300=\mathrm{argmin}\|\bar{{{\mathbf p } } } x - b{{\mathbf e}}-\bar{{{\mathbf r}}}\|_2 ^ 2 $ ] , then 1 .
[ item : estx ] the estimated solution @xmath301 and the exact solution @xmath298 are related by * @xmath302 if @xmath329 is a non - constant decreasing vector , * @xmath303 if @xmath329 is a non - constant increasing vector , * @xmath304 if @xmath329 is a constant vector ; + moreover , the following inequality is true without any monotonicity assumptions .
@xmath331 2 .
[ item : est_b ] the sign of the intercept @xmath306 is determined as follows : * @xmath307 if @xmath332 is a non - constant decreasing vector , * @xmath308 if @xmath332 is a non - constant increasing vector , * @xmath309 if @xmath332 is a constant vector . + in addition , * @xmath333 if @xmath329 is a non - constant decreasing vector , * @xmath334 if @xmath329 is a non - constant increasing vector , * @xmath335 if @xmath329 is a constant vector ; 3 .
given @xmath310 , the ls solution of @xmath336 for @xmath6 is @xmath337 ; 4 . given @xmath337 , the ls solution of @xmath336 for @xmath313 and the true solution
@xmath298 are related as stated in the first conclusion of this theorem .
most results are a direct corollary of lemma [ lemma : ma1_ana ] by setting @xmath338 .
we only prove the new results ( [ item : estx ] ) and ( [ item : est_b ] ) . 1 .
we just need to prove the bounds for @xmath339 . setting @xmath338
in ( [ eq : hatx ] ) we have @xmath340 because @xmath341 @xmath342 and @xmath343 we obtain @xmath344 2 . setting @xmath338 in ( [ eq : hatb ] ) we have @xmath345 the results on the sign follow from lemma [ lemma : ma1_ana ] ( [ lemma_bsign ] ) . for the remaining three inequalities , we only prove the case for which @xmath329 is decreasing .
proofs of the other two are similar . setting @xmath338 in ( [ eq : hatb ] ) we have @xmath346
in remark 2 , we claimed that `` the estimate of dv is much more robust to noise in the formulation than are the estimates of @xmath106 and @xmath107 because @xmath139 is much larger than both @xmath4 and @xmath14 for @xmath50 '' .
here we present a theoretical explanation , which is helpful for algorithm design in quantification . instead of considering a general linear equation , which is out of the range of this paper , we assume a system of equations @xmath347 with only two independent variables @xmath348^t$ ] .
the two columns of the system matrix @xmath106 are denoted by @xmath349 and @xmath350 , i.e. @xmath351 $ ] .
[ thm : var ] suppose the linear system @xmath352 , for @xmath351 $ ] , has the exact solution @xmath353 $ ] , the uncorrelated noise vector @xmath354 obeys a multi - variable gaussian distribution with zero means and common variance @xmath355 and that @xmath356 .
then least squares solution @xmath357^t$ ] has the following statistical properties 1 .
@xmath358 and @xmath359 , and 2 .
@xmath360 .
we assume matrix @xmath106 has the following singular value decomposition @xmath361=u s v^t = u \left ( \begin{array}{cc } s_1 & 0\\ 0 & s_2\\ \vdots & \dots\\ 0&0 \end{array } \right ) \left ( \begin{array}{cc } \cos\theta & \sin\theta\\ -\sin\theta&\cos\theta \end{array } \right ) , \ ] ] in which @xmath362 .
then @xmath363 where @xmath364 because @xmath365 is an unitary matrix and @xmath356 we immediately derive the the following inequality from equation ( [ eq : svd ] ) : @xmath366 this inequality is equivalent to @xmath367 which implies @xmath368 , i.e. @xmath369 and @xmath370 , and @xmath371 .
if we denote the two rows of matrix @xmath372 by @xmath373 and @xmath374 than @xmath375 because @xmath368 and @xmath376 we conclude @xmath377 if we let @xmath378 and @xmath379 be the two rows of matrix @xmath380 then @xmath381 and @xmath382 because @xmath365 is unitary .
thus @xmath383 let @xmath384 it is clear @xmath385 and @xmath386 because the means of @xmath354 are zero , and @xmath387 and @xmath388 resp .. therefore @xmath389 .
because @xmath390 we conclude @xmath391 and @xmath392 this result is illustrated by the following simple example : @xmath393 the first column is much larger than the second column .
if we add @xmath394 noise to the right hand side , i.e. @xmath395 and @xmath396 , and perform simulation with @xmath166 realizations the distribution of the resulted @xmath397 and @xmath398 are illustrated in figure [ fig : diffsens ] .
these results are consistent with the conclusions in theorem [ thm : var ] .
[ fig : diffsens ]
integrating ( [ eq:2t4kc1 ] ) and ( [ eq:2t4kcs ] ) from @xmath49 to @xmath399 we obtain @xmath400 taking the sum of equations ( [ eq : int2t4kc1 ] ) and ( [ eq : int2t4kcs ] ) yields : @xmath401 and canceling @xmath402 from ( [ eq : int2t4kcs1 ] ) using ( [ eq : ctsum ] ) gives : @xmath403 this can be transformed to ( [ eq : root ] ) immediately by using @xmath404 .
j. logan , j. s. fowler , n. d. volkow , a. p. wolf , s. l. dewey , d. j. schlyer , r. r. macgregor , r. hitzemann , b. bendriem , s. j. gatley , graphical analysis of reversible radioligand binding from time - activity measurements applied to [ -11-methyl]-(-)-cocaine studies in human subjects , j. cereb .
blood flow metab . 10
( 1990 ) 740747 .
r. buchert , f. wilke , j. van den hoff , j. mester , improved statistical power of the multilinear reference tissue approach to the quantification of neuroreceptor ligand binding by regularization , j. cereb .
blood flow metab .
23 ( 5 ) ( 2003 ) 612620 .
y. zhou , w. ye , j. r. brai , a. h. crabb , j. hilton , d. f. wong , a consistent and efficient graphical analysis method to improve the quantification of reversible tracer binding in radioligand receptor dynamic studies , neuroimage , 44 ( 3 ) ( 2009 ) 661670 .
c. a. mathis , y. wang , d. p. holt , g. f. huang , m. l. debnath , w. e. klunk , synthesis and evaluation of 11-labeled 6-substituted 2-aryl benzothiazoles as amyloid imaging agents , j. med .
46 ( 2003 ) 27402755 .
j. j. frost , k. h. douglass , h. s. mayberg , r. f. dannals , j. m. links , a. a. wilson , h. t. ravert , w. c. crozier , h. n. j. wagner , multicompartmental analysis of [ 11]-carfentanil binding to opiate receptors in humans measured by positron emission tomography , j. cereb .
blood flow metab . 9
( 1989 ) 398409 .
j. c. price , w. e. klunk , b. j. lopresti , x. lu , j. a. hoge , s. k. ziolko , d. p. holt , c. c. meltzer , s. t. dekosky , c. a. mathis , kinetic modeling of amyloid binding in humans using imaging and ittsburgh compound- , j. cereb .
blood flow metab .. 25 ( 11 ) ( 2005 ) 15281547 .
s. g. mueller , m. w. weiner , l. j. thal , r. c. petersen , c. r. jack , w. jagust , j. q. trojanowski , a. w. toga , l. beckett , ways toward an early diagnosis in lzheimer s disease : the lzheimer s disease neuroimaging initiative ( ) , alzheimer s dement . 1 ( 1 ) ( 2005 ) 5566 . | graphical analysis methods are widely used in positron emission tomography quantification because of their simplicity and model independence .
but they may , particularly for reversible kinetics , lead to bias in the estimated parameters .
the source of the bias is commonly attributed to noise in the data .
assuming a two - tissue compartmental model , we investigate the bias that originates from model error .
this bias is an intrinsic property of the simplified linear models used for limited scan durations , and it is exaggerated by random noise and numerical quadrature error .
conditions are derived under which logan s graphical method either over- or under - estimates the distribution volume in the noise - free case .
the bias caused by model error is quantified analytically .
the presented analysis shows that the bias of graphical methods is inversely proportional to the dissociation rate .
furthermore , visual examination of the linearity of the logan plot is not sufficient for guaranteeing that equilibrium has been reached .
a new model which retains the elegant properties of graphical analysis methods is presented , along with a numerical algorithm for its solution .
we perform simulations with the fibrillar amyloid @xmath0 radioligand [ 11c ] benzothiazole - aniline using published data from the university of pittsburgh and rotterdam groups .
the results show that the proposed method significantly reduces the bias due to model error .
moreover , the results for data acquired over a @xmath1 minutes scan duration are at least as good as those obtained using existing methods for data acquired over a @xmath2 minutes scan duration . bias ; graphical analysis ; logan plot ; pet quantification ; pib ; alzheimer s disease ; distribution volume . |
at sufficiently low particle densities and small layer separations , double - layer quantum hall ( 2lqh ) systems at total filling factor unity ( @xmath3 ) can be described as broken - symmetry states@xcite that possess interlayer phase coherence , even in the absence of interlayer tunneling.@xcite the 2lqh system can be mapped to an equivalent spin-1/2 system by equating `` up '' ( `` down '' ) pseudospins with electrons in the upper ( lower ) layer.@xcite ( the electrons are assumed to have their real spins polarized . )
the exchange energy between the electrons produces a pseudospin stiffness @xmath4 that seeks to keep the pseudospins aligned locally . at finite layer separation @xmath5 , the direct ( hartree )
energy produces a local capacitive charging energy that is minimized when the two layers have equal electron density .
thus the expectation value of the @xmath6 component of the pseudospin vanishes and the pseudospin system has an `` easy - plane '' anisotropy that gives the itinerant ferromagnet an @xmath7 symmetry@xcite ( in the absence of interlayer tunneling ) .
the expectation value of a pseudospin at location @xmath8 can therefore be specified by its angle @xmath9 in the @xmath10 plane . in the absence of interlayer tunneling ,
the 2lqh system picks out a particular global value of @xmath11 in the ground state , spontaneously breaking the @xmath12 symmetry of the @xmath7 ferromagnet .
this gives rise to a linearly dispersing goldstone mode at long wavelengths.@xcite recent measurements of the zero - bias tunneling conductance in 2lqh systems show a huge resonant enhancement when interlayer coherence is present.@xcite this enhancement is related to the goldstone mode of the broken @xmath12 symmetry , and it has been proposed that the dispersion of the goldstone mode can be observed in tunneling conductance measurements by applying a weak parallel magnetic field.@xcite the @xmath7 pseudomagnet also possesses vortex excitations called `` merons '' ; unlike those in an ordinary ferromagnet , these vortices are electrically charged , and the lowest - energy charged excitations of the system consist of vorticity - neutral meron pairs.@xcite there is also a finite - temperature kosterlitz - thouless ( kt ) phase transition due to the @xmath7 symmetry and the finite pseudospin stiffness of the ferromagnet . in the presence of interlayer tunneling ,
the @xmath12 invariance associated with conservation of the charge difference between the two layers is lost .
the finite interlayer tunneling @xmath13 of the electrons acts as an effective zeeman pseudofield of magnitude @xmath14 along the pseudospin @xmath15 axis and aligns the pseudospins , so that @xmath16 in the ground state .
the goldstone mode disappears , and the collective mode becomes gapped . in the presence of interlayer tunneling ,
merons of opposite vorticity are bound together by a `` string '' that produces a linear attraction between the merons , eliminating the finite - temperature kt transition.@xcite murphy _ et al . _
have investigated the effect of an in - plane magnetic field @xmath1 on 2lqh systems.@xcite by varying @xmath1 and studying the energy gap obtained from activation energy measurements of the longitudinal resistivity , they find evidence for a phase transition between two competing qh ground states at a critical value @xmath17 .
these two ground states are understood in the pseudospin language as being due to a competition between the tunneling energy @xmath13 and the exchange energy @xmath4.@xcite read has presented an appealing analysis of charged ( meron pair ) excitations in this system , focusing on the value of the charge gap near the commensurate - incommensurate transition.@xcite the presence of @xmath1 periodically shifts the phase of the tunneling matrix elements , resulting in an effective zeeman field for the pseudospins that rotates along the planar direction perpendicular to @xmath18 , with a wavelength @xmath19 , which is the distance required to contain one flux quantum @xmath20 between the layers .
the net result is that for gradual variations of the pseudospin texture ( gradual on the scale of the magnetic length @xmath21 , where @xmath22 is the strength of the magnetic field normal to the plane ) , the energy of the @xmath7 pseudospin system has the pokrovsky - talapov ( pt ) form,@xcite @xmath23 , \ ] ] up to a constant , where @xmath24 defines the parallel magnetic - field wave vector , @xmath25 is the tunneling energy ( where @xmath26 is the tunneling energy when @xmath27),@xcite and @xmath28 is the pseudospin stiffness within the hartree - fock approximation ( hfa ) . here
@xmath29 is the layer imbalance , and @xmath30 is the interlayer exchange stiffness when the layers are balanced : @xmath31 , or @xmath32 .
the layer separation is @xmath5 , and @xmath33 is the filling factor of layer @xmath34 . for small @xmath35 ( small @xmath1 )
, the ground - state energy is minimized by having the pseudospins point in the direction of the local ( rotating ) pseudospin zeeman field , so that @xmath36 .
this is the commensurate ground state , and it minimizes the pseudospin zeeman ( tunneling ) energy .
however , it does so at the expense of the exchange energy , since neighboring pseudospins are no longer parallel .
in the limit of large @xmath35 , the cost in exchange energy for the pseudospins to align with the rapidly rotating zeeman field is prohibitive , and the pseudospins become ( nearly ) uniformly polarized ( constant @xmath11 ) , just as if @xmath37 .
the state with uniformly polarized pseudospin is the large-@xmath35 limit of the incommensurate state .
it turns out that for all finite @xmath0 , the translational symmetry of the pseudospin polarization is broken , and a soliton - lattice ( sl ) state is obtained in the incommensurate phase .
the sl phase of the pt model can also undergo a separate finite - temperature dislocation - mediated kt transition that restores the translational symmetry.@xcite this work focuses on calculating the ground - state properties of the sl state , for all @xmath0 .
interestingly , it is not necessary to determine @xmath9 in order to calculate the total energy of the system.@xcite from the total energy , we calculate thermodynamic quantities such as the sl stiffnesses , extending the results of ref . for the stiffnesses and the kt temperature to all @xmath1 .
we also calculate the sl contribution to the magnetization and susceptibility , and discuss some possibilities of measuring these quantities experimentally .
the plan of this paper is as follows . in sec .
[ sec : singsols ] , we discuss the single - soliton solutions that follow from the equation of motion obtained from the pt energy .
for @xmath38 the solitons proliferate ; in sec .
[ sec : solat ] , the interaction between soliton lines is discussed , and the separation between solitons as a function of @xmath39 is derived . in sec .
[ sec : stiff ] , the compressional and shear elastic constants are analyzed , and an estimate is made of the kosterlitz - thouless temperature for melting the soliton lattice , as a function of @xmath35 .
the interlayer phase coherent 2lqh state has a diamagnetic response to an applied in - plane magnetic field;@xcite section [ sec : mag ] gives a calculation of the in - plane magnetization due to the 2lqh state , as a function of @xmath35 .
we conclude with a summary of our results .
when @xmath38 , it is convenient to define @xmath40 , so that eq .
( [ eq : ept ] ) becomes @xmath41 .\ ] ] minimizing @xmath42 with respect to variations in @xmath43 gives the sine - gordon equation : @xmath44 where @xmath45 .
we shall give numerical values for our results for a hypothetical `` typical '' gaas ( effective mass @xmath46 and relative dielectric constant @xmath47 ) 2lqh sample , which for the sake of definiteness we assign the following sample parameters : total particle areal density @xmath48 , layer ( midwell to midwell ) separation @xmath49 nm , and tunneling energy @xmath50 mev .
such a sample would have @xmath51 nm , @xmath52 , @xmath53 mev for @xmath54 , and @xmath55 mev . in the hfa ,
@xmath56 mev and @xmath57 nm . the commensurate state minimizes the tunneling energy by having @xmath58 , so that the phase angle @xmath59 follows the tumbling zeeman pseudofield .
the energy per area of this state is @xmath60 . in the limit of large @xmath35 ,
the incommensurate state with constant @xmath11 has a lower energy per area , equal to zero .
these energies are plotted as the solid and dashed lines in fig .
[ hannafig1 ] .
we therefore estimate that there is a phase transition near the point where the commensurate - state and the constant-@xmath11 incommensurate - state energies are equal , at @xmath61 .
it turns out , however , that the incommensurate state lowers its energy by breaking translation invariance , so that at finite @xmath13 and @xmath62 , the value of @xmath43 depends on position .
equation ( [ eq : sg1 ] ) possesses soliton solutions . to see this ,
let us seek solutions of the form @xmath63 , \ ] ] where @xmath64 could be any unit vector in the @xmath10 plane .
then eq . ( [ eq : sg1 ] ) becomes @xmath65 note that this equation is equivalent to the equation of motion for a pendulum of length @xmath66 , , if we replace , ( time ) , , and ( tunneling amplitude ) .
this analogy is very useful in finding the soliton solutions for the pt model .
in particular , the pendulum can make full circles in a given direction .
this corresponds to the sl state in the 2lqh system . in analogy with the pendulum problem
, we may define a conserved quantity analogous to the total ( kinetic plus potential ) energy of a pendulum : @xmath67 differentiating eq .
( [ eq : q2 ] ) with respect to @xmath68 and using eq .
( [ eq : sg2 ] ) shows that @xmath69 , so that @xmath70 is a constant of the motion . defining @xmath71 then leads to the equation @xmath72 it is straightforward to solve eq .
( [ eq : beta ] ) when @xmath73 by writing @xmath74 , so that @xmath75 , giving @xmath76 , where @xmath77\ ] ] represents a single soliton in the @xmath64-direction , centered at @xmath78 , with width @xmath79 .
this is shown in fig .
[ hannafig2 ] .
note that @xmath80 changes by @xmath81 as @xmath82 goes from @xmath83 to @xmath84 .
this corresponds to the motion of a pendulum that just barely completes a full revolution , and whose period goes to infinity .
the energy per length of a single soliton may be computed by substituting eq .
( [ eq : sol ] ) into eq .
( [ eq : tp2 ] ) and subtracting the commensurate - state ( @xmath85 ) energy , to obtain @xmath86 where @xmath87 is the sample length in the planar direction perpendicular to @xmath64 . the lowest ( soliton ) and highest ( anti - soliton ) energy solutions occur for @xmath88 .
since solitons in the @xmath89 directions are equivalent to anti - solitons in the @xmath90 , we shall speak only about solitons with orientations defined by @xmath64 .
the lowest - energy soliton state has @xmath91 , and its energy per length is @xmath92 which goes to zero when @xmath93 , where @xmath94 the value of the critical wave vector @xmath95 will depend on the layer imbalance @xmath29 .
equations ( [ eq : tnu ] ) and ( [ eq : rhonu ] ) give @xmath96 in the hfa , where @xmath97 is the value of @xmath95 when the layers are balanced ( @xmath32 ) . equation ( [ eq : qcmz ] ) implies that the value of @xmath95 where the ci transition occurs could be fine - tuned by adjusting the layer imbalance @xmath98 i.e. , by adjusting the gate voltages on the 2lqh sample .
such a procedure might be very useful in investigating the ci transition . when @xmath99 and @xmath100 , the lowest - energy charged excitations are finite - length soliton lines with charged meron ends
i.e. , charged vortices bound by a soliton `` string '' whose tension is given by eq .
( [ eq : essbar ] ) .
as @xmath35 increases , the soliton - line `` string tension '' gets weaker , so that the coulomb repulsion of the merons stretches out the string and lowers the energy of the charged excitation.@xcite at @xmath101 the soliton - line `` string tension '' goes to zero , and it costs zero energy to make infinitely long soliton lines . since the creation energy per length for a soliton decreases linearly with @xmath35 ( with @xmath1 ) for @xmath102 , it becomes energetically favorable to form solitons .
the number of solitons created is determined by the competition between the ( negative ) creation energy per soliton versus the repulsive ( positive energy ) interactions between solitons .
note that @xmath103 ( the value of @xmath35 at which the commensurate - state and constant-@xmath11 incommensurate state energies are equal ) , so that for @xmath38 it is energetically favorable to create solitons . because the solitons are weakly repulsive , the result is a soliton - lattice state that we describe below and illustrate in fig .
[ hannafig3 ] .
an analogous effect occurs in long josephson junctions , where application of a magnetic field parallel to two superconducting planes in close proximity produces @xmath81 twists in the josephson phase and generates a sl state.@xcite to summarize , for @xmath104 , we obtain the commensurate phase in which @xmath105 . for @xmath106 , we introduce a single soliton , corresponding to the marginal case of a pendulum that makes exactly one full revolution and has an infinite period of oscillation . for @xmath38
, we obtain a soliton lattice , corresponding to a pendulum making complete revolutions in one direction .
it has been argued that the commensurate to incommensurate soliton - lattice ( ci ) transition at @xmath106 can be identified with the phase transition between 2lqh ground states seen by murphy and co - workers;@xcite we therefore make the identification @xmath107 .
we shall now use eq .
( [ eq : beta ] ) to determine the sl spacing @xmath108 .
we do this by noting that over one period of the soliton lattice , @xmath43 changes by @xmath81 , so that @xmath109 changes by @xmath110 .
we therefore express @xmath108 as @xmath111 where we have used eq .
( [ eq : beta ] ) and have defined @xmath112 , and where @xmath113 is the complete elliptic integral of the first kind.@xcite we define the sl wave vector @xmath114 , so that eq .
( [ eq : ls ] ) may be written in terms of @xmath115 as @xmath116 note that @xmath117 corresponds to the ci transition , where @xmath118 and @xmath119 , whereas @xmath120 corresponds to the constant-@xmath11 incommensurate state , where @xmath121 . from eqs .
( [ eq : qsoqc ] ) and ( [ eq : kfunc ] ) , it follows that @xmath122 in order to determine the value @xmath123 of the sl wave vector that minimizes the total energy , we express the energy per area from eq .
( [ eq : tp2 ] ) as an integral over @xmath109 [ cf .
( [ eq : ls ] ) ] and obtain @xmath124 \\ \nonumber & \rightarrow & \left \lbrace \begin{array}{llll } \frac{1}{2}\rho_{\rm s}q^2 - t/2\pi\ell^2 , & q < q_{\rm c } \\ \rho_{\rm s } \left [ \frac{1}{2}q^2 - { \bf q } \cdot { \bf q}_{\rm s } \right . & ~ \\ \left . \qquad + q_{\rm c } q_{\rm s } \left ( 1 + 4 e^{-\frac{2\pi}{q_{\rm s}\xi } } \right ) - 1/\xi^2 \right ] , & \eta \rightarrow 1 \\ \rho_{\rm s }
\left [ \frac{1}{2 } \left({\bf q}-{\bf q}_{\rm s}\right)^2 - 1/\left(2q_{\rm s}\xi^2\right)^2 \right ] , & \eta \rightarrow 0 \end{array } \right .\end{aligned}\ ] ] where @xmath125 is the complete elliptic integral of the second kind,@xcite and @xmath126 is the sample length in the planar direction parallel to @xmath64 .
agreement between eqs .
( [ eq : esol ] ) and ( [ eq : elxly ] ) in the thermodynamic limit is obtained by equating @xmath127 , so that @xmath128 at @xmath106 .
all the terms in the @xmath129 limit ( near the ci transition , where the solitons are well separated ) of eq .
( [ eq : elxly ] ) have simple physical interpretations .
the first and last terms , @xmath130 , constitute the ( commensurate - phase ) energy per area of the pt model in the absence of solitons .
the second and third terms , are just the creation energy per area for the solitons in terms of interacting soliton lines [ see eq .
( [ eq : essbar ] ) ] .
the fourth term , @xmath131 , is the exponentially weak repulsive interaction energy per area between the solitons . because @xmath132 is the density of soliton lines
, the fourth term says that near the ci transiton , the interaction energy per length between two parallel , straight , and infinitely long solitons lines separated by a distance @xmath108 is @xmath133 hence for @xmath38 soliton lines proliferate rapidly until the repulsion between the solitons become significant .
the notion of an effective repulsive interaction between sine - gordon solitons was discussed by perring and skyrme,@xcite who obtained the exponentially weak repulsion found above .
the arguments of ref .
imply that when the solitons are close together ( @xmath134 ) at large @xmath39 , the repulsive potential energy per length between soliton lines is @xmath135 .
this latter repulsion is due to boundary condition that @xmath43 must change by @xmath81 over the small distance @xmath108 , which implies a large gradient energy . the value of @xmath136 which minimizes the energy per area [ eq .
( [ eq : elxly ] ) ] is found by setting to zero @xmath137 where we have used the identity@xcite @xmath138 it is not difficult to show that the second derivative of the energy per area with respect to @xmath139 is always positive ; thus the extremum calculated above is a minimum .
it follows from eq .
( [ eq : eom ] ) that the energy is minimized for @xmath140 , and for @xmath141 such that @xmath142 where @xmath143 .
we define the ci closeness parameter @xmath144 which vanishes at the ci transition ; from eqs .
( [ eq : efunc ] ) and ( [ eq : qoqc ] ) , it follows that for @xmath145 , @xmath146 , \ ] ] so that asymptotically , @xmath147 equations ( [ eq : qsoqc ] ) and ( [ eq : qoqc ] ) together allow us to determine the equilibrium sl wave vector @xmath148 that minimizes the energy , in terms of the parallel - field wave vector @xmath149.@xcite we have plotted this in fig .
[ hannafig4 ] . from eqs .
( [ eq : qsoqc ] ) , ( [ eq : qoqc ] ) , and ( [ eq : etabar ] ) , it follows that @xmath150 where the @xmath151 limit is true asymptotically.@xcite we note however that , as found by pokrovsky and talapov@xcite and discussed in ref . , the meandering of soliton lines at finite temperature renormalizes the dependence of the soliton - line density on the parallel magnetic field , so that @xmath152 sufficiently close to the ci transition . the minimized value @xmath153 of the energy per area at @xmath154 is found by using eqs .
( [ eq : elxly ] ) and ( [ eq : qoqc ] ) in eq .
( [ eq : efunc ] ) to obtain @xmath155 the sl state energy per area is represented by the dots in fig .
[ hannafig1 ] .
although it is not needed for calculating the stiffnesses or susceptibility of the sl , the sl solution for @xmath156 is given by@xcite @xmath157 = { \rm sn}({\bf\hat{e}}_1\cdot ( { \bf r } - { \bf r}_0)/\eta\xi,\eta ) , \ ] ] where sn denotes the sine - amplitude jacobian elliptic function with parameter @xmath158.@xcite near the ci transition , when @xmath159 , @xmath43 is very close to being a periodic superposition of single - soliton solutions , spaced apart by @xmath160 , so that @xmath161 , where @xmath80 denotes the single - soliton solution , eq . (
[ eq : sol ] ) . in the sl state , especially away from the ci transition , it is sometimes useful to work with @xmath162 because it is periodic in the sl spacing , so that @xmath163 , where @xmath164 . in the limit @xmath165 , @xmath166 and @xmath167 , so that we may regard @xmath168 as a small quantity .
expressing the sine - gordon equation [ eq .
( [ eq : sg2 ] ) ] in terms of @xmath169 and working to lowest order in @xmath169 gives @xmath170 $ ] , so that @xmath171 .\ ] ]
the elastic constants of the soliton lattice are given by the stiffness tensor @xmath172 .
the change in the energy due to varying the spacing between the soliton lines is characterized by the compressional stiffness @xmath173 .
the shear stiffness @xmath174 is associated with the change in energy due to shearing the upper and lower ends of the soliton lines in opposite directions , and is equivalent to a combined rotation and compression .
we use the calculated stiffnesses to describe the @xmath1-dependence of the kt temperature@xcite for the dislocation - mediated kt melting transition@xcite of the soliton - lattice .
we calculate the stiffness tensor by two methods .
first , we calculate the stiffness @xmath172 that is obtained by differentiating @xmath175 in eq .
( [ eq : elxly ] ) with respect to the @xmath176 components of @xmath136 for fixed @xmath149 at the extremal , where eq .
( [ eq : eom ] ) is zero .
then we calculate the stiffness tensor @xmath177 obtained by calculating the effects of fluctuations of the angle variable @xmath156 away from its ground - state value , eq .
( [ eq : thetasoln ] ) .
we begin by calculating the stiffness tensor @xmath172 from the dependence of the ground - state energy [ eq .
( [ eq : elxly ] ) ] on the soliton - lattice wave vector @xmath139 .
the expressions we obtain for @xmath172 by this method have been carried out for all values of @xmath178 , and agree with the results obtained in ref .
, in the limit @xmath119 .
because the stiffnesses involve the second derivative of @xmath175 with respect to the components of @xmath136 at fixed @xmath149 , the terms in eq .
( [ eq : elxly ] ) that depend explicitly on @xmath149 ( including the term @xmath179 that gives the orientational dependence of the energy per area ) do not contribute to @xmath172 .
thus @xmath180_{\bf q } \nonumber \\ & = & \rho_{\rm s } \lim_{{\bf q}_{\rm s}\rightarrow \bar{\bf q}_{\rm s } } \left [ \left ( \delta_{ij } - \frac{\bar{q}_{{\rm s}i}\bar{q}_{{\rm s}j } } { \bar{q}_{\rm s}^2 } \right ) \frac{1}{\bar{q}_{\rm s } } + \frac{\bar{q}_{{\rm s}i}\bar{q}_{{\rm s}j}}{\bar{q}_{\rm s}^2 } \frac{\partial}{\partial q_{\rm s } } \right ] \\
\nonumber & & \qquad \qquad \qquad \times \left [ q_{\rm c } \frac{e(\eta)}{\eta } \right ] _ { \bf q } , \end{aligned}\ ] ] where we have used eqs . (
[ eq : eom ] ) and ( [ eq : qoqc ] ) .
since @xmath181 , it follows that @xmath182 . using the results of sec .
[ sec : solat ] , and the identity@xcite @xmath183 from which it follows that @xmath184 we find that the compressional elastic constant @xmath173 is equal to @xmath185_{\bf q } = \frac{\partial q}{\partial\bar{q}_{\rm s } } \\ \nonumber & & = \frac{16(1-\bar{\eta}^2)}{(\bar{q}_{\rm s}\xi)^3 ( q\xi ) \bar{\eta}^4 } \rightarrow \left \lbrace \begin{array}{cc } ( 2/\pi^2 ) \epsilon \ln^2 ( 1/\epsilon ) , & q / q_{\rm c } \rightarrow 1 \\ 1 - \frac{3}{32 } \left ( \frac{\pi}{4 } \frac{q_{\rm c}}{q } \right ) ^4 , & q / q_{\rm c } \rightarrow \infty \end{array } \right .\end{aligned}\ ] ] in the limit @xmath186 when the soliton lines are far apart , @xmath173 is very small [ of order @xmath187 , see eq .
( [ eq : qsofq ] ) ] .
the energy cost of compression very close to the ci transition is due to the exponentially weak intersoliton interaction energy .
the energy per area due to the string tension of the soliton lines ( the term @xmath188 ) does not contribute to @xmath173 , although it does contribute to @xmath174 , as we explain below . as explained in ref .
, soliton lines meander at finite temperature and are no longer straight ; collisions between meandering soliton lines produce an effective entropic repulsion between the solitons that dominates the exponential repulsion at any nonzero temperature , for @xmath189 sufficiently large .
this effect renormalizes the compressional stiffness @xmath173 upwards so that it becomes proportional to @xmath190 . in the limit
@xmath191 , the tunneling term in the pt energy ( [ eq : tp2 ] ) fluctuates on very short length scale and averages to zero , so that eq .
( [ eq : tp2 ] ) becomes the isotropic xy model ; thus one expects @xmath173 to approach the pseudospin stiffness @xmath4 in that limit . from eqs .
( [ eq : qoqc ] ) and ( [ eq : kij ] ) , it follows that the shear elastic constant @xmath174 is given by @xmath192 as expected , the shear stiffness @xmath174 approaches the pseudospin stiffness @xmath4 in the limit @xmath191 .
but in the limit @xmath159 , @xmath174 diverges as @xmath193 .
the origin of this effect is that the shear motion described by @xmath174 increases the total length of the soliton lines . by definition
, @xmath194 describes a shear displacement in which @xmath195 changes , while @xmath196 remains fixed : i.e. , the solitons lines tilt away from their equilibrium `` vertical '' ( @xmath197 ) direction by a small angle , while keeping their `` horizontal separation '' @xmath198 constant .
this shear motion reduces the mean soliton separation ( the separation along the direction perpendicular to the tilted soliton lines ) to , so that . the @xmath174-shear corresponds to a global rotation plus a compression of the sl .
packing the solitons lines more closely together in a fixed sample area produces more total soliton line length , which costs more soliton - line creation energy . because the term @xmath199 that contains the orientational dependence of the energy per area [ eq .
( [ eq : elxly ] ) ] is linearly proportional to @xmath200 , it can not contribute to the stiffnesses ( which are proportional to second derivatives of the energy per area with respect to the components of @xmath201 ) at all .
it might be supposed that the rotation of the soliton lines that occurs upon shearing should cost energy , but this is not so , because a shear is a combination of rotation and compression , rather than a pure rotation .
the creation energy per area of the sl near the ci transition ( @xmath202 ) is @xmath203 , and consists of two terms .
the first term ( @xmath204 ) is just the total line length of the solitons times the line tension , divided by the total area .
the second term ( @xmath205 ) depends explictly on the angle @xmath206 between @xmath149 and @xmath136 , and determines the orientation of the sl because it is minimized by choosing @xmath136 along @xmath35 ( i.e. , @xmath207 ) ; thus , a different choice of sl orientation ( i.e. , a pure rotation of the soliton lines ) would cost more energy . interestingly , the second term in the creation energy is unchanged by a shear , because the energy cost of rotating the soliton lines is exactly offset by the increase in total soliton line length : @xmath208\cos\phi = -q \bar{q}_{\rm s}$ ] , independent of @xmath206 .
the only change in the creation energy comes from the first term , which depends only on the density of soliton lines : . sufficiently close to the ci transition ( i.e. , when @xmath209 ) , the exponentially small interactions may be neglected in comparison to the creation energy .
for @xmath159 and small shear ( @xmath210 ) , @xmath211 - t/2\pi\ell^2 , \end{aligned}\ ] ] so that @xmath212 as @xmath159 , in agreement with the results of ref . .
the fact that bilayer phase - coherent 2lqh states can exhibit a finite - temperature kosterlitz - thouless ( kt ) transition in the absence of interlayer tunneling ( @xmath37 ) has been discussed in earlier work.@xcite although finite @xmath13 removes the possibility of a kt transition in the commensurate phase of 2lqh systems by altering the nature of the long - range interaction between vortices ( `` merons '' in this case),@xcite the sl phase does support a finite - temperature kt transition due to dislocation - mediated melting of the sl.@xcite as discussed in ref . , the kt temperature may be estimated as @xmath213 where we have used the zero - temperature values for @xmath172 that we calculated previously to make a rough estimate the kt temperature . as mentioned earlier , finite - temperature fluctuation effects can strongly renormalize @xmath173,@xcite and may also effect @xmath174 .
our results agree with those of ref . in the limit @xmath119 .
we have plotted the compressional ( @xmath173 ) and shear ( @xmath174 ) stiffnesses in fig .
[ hannafig5 ] , together with the kt temperature .
the kt transition would be most easily measured from the temperature dependence of the linear response to oppositely directed currents in each layer .
this would require double - layer electron devices with layers that could be contacted separately .
unfortunately , the leakage currents produced when the interlayer tunneling is not vanishingly small would make it difficult , perhaps impossible , to set up oppositely directed currents in each layer .
however , because the sl dislocations are electrically charged , it might be possible that the kt transition could be signalled by an increase in the usual longitudinal resistivity @xmath214 , measured in devices with the current flowing in the same direction in both layers .
the increase in @xmath214 would originate from the proliferation of unbound charged dislocations above the transition temperature .
we now calculate an alternate stiffness tensor @xmath177 by examining the effect of deviations of the angle variable @xmath156 from its ground - state value .
we write @xmath215 , where @xmath216 is the ground - state solution that minimizes the pt energy ( [ eq : tp2 ] ) and therefore satisfies eq .
( [ eq : sg1 ] ) , and @xmath217 is the deviation of @xmath43 from its ground - state value .
we do not include dynamics here , because our focus is on ground - state , rather than excited - state , properties .
the pt energy for @xmath43 is @xmath218 \equiv e_{\rm pt}[\tilde{\theta}_0 ] + \delta h$ ] , where @xmath219 , given by eq . (
[ eq : tp2 ] ) , is the pt energy from which the ground - state @xmath216 is determined via eq .
( [ eq : sg1 ] ) , and @xmath220 \\ \nonumber & = & \frac{1}{2 } \int \frac{d^2r}{2\pi\ell^2 } \delta\tilde{\theta } \left [ t\cos\tilde{\theta}_0 - 2\pi\ell^2 \rho_{\rm s } { \bf \nabla}^2 \right ] \delta\tilde{\theta } , \end{aligned}\ ] ] where we have kept terms up to quadratic order in @xmath221 .
there are no terms linear in @xmath221 because @xmath216 is determined by minimizing @xmath219 with respect to variations in @xmath43 .
the total energy is minimized by choosing @xmath221 from among the eigenvalues of the bracketed schrdinger - like operator in eq .
( [ eq : delh ] ) , so that @xmath222 \delta\tilde{\theta}_{\bf q } = { \cal e}_{\bf q } \delta\tilde{\theta}_{\bf q}\ ] ] if we take @xmath223 so that @xmath224 , then @xmath216 depends only on @xmath15 , @xmath225 is translationally invariant in the @xmath226 direction , and we may write @xmath227 . the term @xmath228 in eq .
( [ eq : scheq ] ) is periodic in the @xmath15 direction , with a period of @xmath229 . as shown in ref .
, when @xmath216 has the form ( [ eq : thetasoln ] ) , eq .
( [ eq : scheq ] ) becomes lam s equation , after a simple rescaling of @xmath15 .
lam s equation has three simple solutions , two of which have low - energy limits .
the first type of solution has zero energy and corresponds to a uniform translation of the vortex lines , @xmath230 , where @xmath231 is the @xmath15 component of @xmath232 in eq .
( [ eq : thetasoln ] ) . of greatest interest to us are the type of solutions which have been called `` vortex oscillations '' in the context of long josephson junctions in parallel magnetic fields.@xcite from ref . , it follows that in the long - wavelength limit , @xmath233 where @xmath173 is equal to the compressional stiffness in eq .
( [ eq : k11 ] ) , and @xmath234 is the crystal momentum along the @xmath15 direction .
although @xmath235 , we find that @xmath236 .
the reason for the discrepancy between @xmath237 and @xmath174 is not obvious .
it may be that using @xmath238 for the transverse stiffness is valid only at very long wavelengths @xmath239,@xcite where @xmath240 is read s estimate of the minimum @xmath226-distance for sl dislocations to interact logarithmically.@xcite it is also possible that @xmath237 , which is based on a calculation that assumes @xmath221 is everywhere small , is not able to describe uniform shear , which can move solitons lines far from their equilbrium positions on the scale of the soliton line thickness @xmath79 .
the relationship between @xmath237 and @xmath174 requires further clarification , including a stronger argument for preferring @xmath174 over @xmath237 in estimating the kt temperature .
the interlayer phase coherent 2lqh state exhibits an in - plane magnetization @xmath241 in the presence of an in - plane magnetic field @xmath242 .
the in - plane magnetization can be calculating by differentiating the minimized ground - state energy per volume with respect to the parallel magnetic field : @xmath243 where @xmath153 is given by eq .
( [ eq : emin ] ) . in order to carry out the differentiation in eq .
( [ eq : magpar ] ) , we first differentiate eq .
( [ eq : qoqc ] ) with respect to @xmath35 and make use of eq .
( [ eq : qsoqc ] ) to obtain @xmath244 we note here that the @xmath245 term in eq .
( [ eq : detadq ] ) arises from differentiating @xmath246 in eq .
( [ eq : qoqc ] ) with respect to @xmath35 ; the @xmath35-dependence of @xmath246 is due to the dependence of the tunneling matrix element on the parallel magnetic field , @xmath247 , which is a single - particle effect.@xcite the tunneling part of the equilibrium energy per area will also give a contribution to the in - plane magnetization proportional to @xmath248 , again due to dependence of @xmath13 on @xmath35 .
it is convenient to separate @xmath241 into two parts : , where @xmath249 is calculated at fixed @xmath13 ( @xmath13 independent of @xmath35 ) , and @xmath250 arises from the @xmath35-dependence of @xmath13 . using eqs .
( [ eq : magpar ] ) and ( [ eq : detadq ] ) it is straightforward to show that the sl contribution to the parallel magnetization is @xmath251 where @xmath252 sets the scale of the sl magnetization , and the numerical estimate of @xmath253 is given for the `` typical '' gaas sample described in sec .
[ sec : singsols ] .
for such a sample , @xmath254 sets the scale of the sl magnetic moment per unit area .
thus magnetometers with sensitivities in the range of @xmath255 am@xmath256 to @xmath257 am@xmath256 would require sample areas in the range of 1 mm@xmath256 and 1 cm@xmath256 to measure @xmath259 .
the magnitude of the sl magnetization behaves like @xmath260 the sl magnetization may also be calculated directly from the pseusdospin supercurrent density,@xcite @xmath261 , and the definition of the magnetic moment .
the electrical current @xmath262 in layers 1 and 2 is @xmath263 the magnetization produced by the above current is therefore@xcite @xmath264 in agreement with eq .
( [ eq : magsl ] ) . the 2lqh interlayer phase coherent state
may be regarded as a pseudospin - channel superconductor.@xcite the magnetization @xmath241 is due to pseudospin supercurrents , corresponding to electrical currents traveling in opposite directions in each layer , which partially screen @xmath242 . for @xmath99 , @xmath242 is maximally excluded from the region between the planes ; but when @xmath265 , additional @xmath1 penetrates the region between the plates in the form of solitons that each carry a single flux quantum @xmath20 ( corresponding to a phase change @xmath266 ) , leading to a precipitous decline in the magnetization .
the direction of @xmath241 is opposite to @xmath242 , in accord with lenz s law .
an exactly analogous effect occurs for magnetic fields applied parallel to narrow insulating regions ( josephson junctions ) between superconductors .
the contribution to the magnetization due to the @xmath35-dependence of the tunneling matrix element @xmath13 is @xmath267 .\end{aligned}\ ] ] in the commensurate phase , @xmath268 , and the magnitude of the tunneling contribution to the parallel magnetization behaves like @xmath269 the sl magnetization is plotted in fig .
[ hannafig6 ] .
it is useful to compare the total sl magnetization @xmath253 to the scale of the landau diamagnetism in a @xmath3 2lqh system : @xmath270 where @xmath271 is the effective bohr magneton , and we have made use of the parameters for the `` typical '' gaas sample described in sec .
[ sec : singsols ] .
this shows that the sl magnetization is expected to be roughly an order of magnitude smaller than the landau diamagnetism .
it is interesting to note that the weak signals associated with orbital diamagnetism in two - dimensional electron systems have been measured at high magnetic fields by torsional magnetometry,@xcite squid magnetometry,@xcite and micromechanical cantilever magnetometry.@xcite the torque on the 2lqh sample in the presence of both a perpendicular and parallel magnetic field has both a landau - diamagnetic component @xmath272 and a sl component , @xmath273 .
the smallness of the parallel moment @xmath274 is offset by the large perpendicular magnetic field @xmath22 in the expression for the torque @xmath275 .
the sl magnetization @xmath249 might also be measured using high - field magnetometry techniques mentioned above .
it follows from eq .
( [ eq : magpar ] ) that when @xmath274 is plotted against @xmath1 , the area under under the resulting curve from @xmath276 to @xmath277 is @xmath278 \\ \nonumber & = & \left(\frac{t \langle m_x \rangle}{2\pi\ell^2d}\right)_{b_\parallel=0 } , \end{aligned}\ ] ] where @xmath279 is the ground - state expectation value of the @xmath15 component of the pseudospin order parameter , which has a value of @xmath280 in the hfa , so that @xmath281 ( in the hfa ) when the layers are balanced ( @xmath32 )
. thus the area under the @xmath274 versus @xmath1 curve may be regarded either as a measurement of @xmath13 ( if quantum fluctuations in the ground state are neglected ) , or as a measure of order - parameter suppression ( of @xmath282 ) due to quantum fluctuations@xcite ( if @xmath13 can be measured separately ) .
equations ( [ eq : magsl ] ) and ( [ eq : qsofq ] ) show that @xmath283 in the commensurate phase ( @xmath284 ) , where @xmath285 sets the scale of the sl contribution to the magnetic susceptibility .
the sl contribution to the parallel - field magnetic susceptibility is defined for fixed @xmath13 ( independent of @xmath35 ) : @xmath286 , & q / q_{\rm c } \rightarrow 1^+ \\ -\frac{3}{2 } \left ( \frac{\pi}{4 } \frac{q_{\rm c}}{q }
\right ) ^4 , & q / q_{\rm c } \rightarrow \infty \\ \end{array } \right .\end{aligned}\ ] ] where we have used eqs . (
[ eq : qoqc ] ) and ( [ eq : k11 ] ) , @xmath287 and the numerical estimate of @xmath288 is given for the `` typical '' gaas sample described in sec .
[ sec : singsols ] . here
@xmath289 is the fine - structure constant .
we have plotted the susceptibility in fig .
[ hannafig7 ] .
note that near the ci transition , the susceptibility diverges like @xmath290 , with logarithmic corrections .
it might be possible to measure the sl magnetization or even the sl magnetic susceptibility by varying the gate voltages of the sample , in order to adjust @xmath95 and tune close to the ci transition . measuring @xmath291 might be possible using ac modulation of the gate voltages in order to ac modulate the layer imbalance @xmath98 and therefore the critical wave vector @xmath95 . by such a method ,
the ratio @xmath292 could be ac modulated just above and below unity , allowing @xmath291 to be determined at or near the ci transition . as an example , we compute here that part of @xmath293 which is proportional to @xmath291 : @xmath294 where we have used eq ( [ eq : qcmz ] ) in the last line . equation ( [ eq : dmdm ] ) shows that @xmath293 has a contribution proportional to @xmath291 , which diverges like @xmath295 near the ci transition .
however , eq . ( [ eq : dmdm ] ) vanishes for balanced layers ( @xmath32 ) ; thus , layer imbalance is required to measure @xmath293 .
it turns out that when the layers are not balanced ( @xmath296 ) , the solitons have a nonzero electric dipole moment per length;@xcite this changes the intersoliton repulsion from being exponentially weak to a power law .
the net result is that the stiffness @xmath173 near the ci transition is strengthened from being essentially linear in @xmath297 to being proportional to @xmath298 , so that the divergence in @xmath291 becomes an inverse square - root singularity . in real samples
, we expect this singularity to be smoothed out by finite temperature and disorder . in practice ,
@xmath98 is varied by adjusting gate voltages . for a small change @xmath299 in a gate voltage away from balance ( @xmath32 ) , the change @xmath300 in the layer imbalance is linearly proportional to @xmath299 . for the `` typical '' gaas sample described in sec .
[ sec : singsols ] , a rough estimate indicates that @xmath301 if we take @xmath302 m to be the distance between the gate and the double layers .
there is an additional complication in changing the layer imbalance by adjusting one of the gate voltages .
unless both the back and front gates are adjusted together in a coordinated way , the change in gate voltage @xmath299 will also change the total filling factor @xmath303 by an amount @xmath304 , where @xmath305 if @xmath302 m.
in the presence of a sufficiently strong parallel magnetic field , a @xmath54 2lqh device undergoes a transition to a soliton - lattice ( sl ) state . we have investigated the ground - state properties of the sl state for all values of the parallel magnetic field , with an eye towards possible experimentally measurable effects .
we found that the sl contribution to the orbital magnetization rises in the commensurate phase ( @xmath99 ) with @xmath35 , and quickly drops to zero in the incommensurate phase ( @xmath62 ) .
an estimate of the size of the sl magnetization shows that it could be detected by sensitive magnetometry techniques.@xcite the sl magnetic susceptibility shows a singularity at @xmath101 , and it was proposed that this signature of the ci transition might be detected by varying the gate voltages so as tune close to the ci transition .
the longitudinal and transverse sl stiffnesses were computed and used to estimate the temperature of the kt transition , which might be indicated experimentally by an increase in @xmath306 at the transition .
a more sensitive signal of the kt transition would be obtained by measuring the transresistivity ( coulomb drag ) as a function of temperature , although this would require separately contactable layers .
however , the leakage currents produced by any sizeable interlayer tunneling might make it impractical to set up oppositely directed currents in each layer . in this paper
, we have largely neglected the existence of disorder .
as pointed out by fisher@xcite and discussed by read@xcite , randomness in the tunneling @xmath13 due , for example , to small variations in the barrier thickness , pins the sl domain walls randomly and destroys the long - range order , no matter how weak the randomness .
this puts limits on how closely @xmath35 can approach @xmath246 i.e. , how small @xmath307 can be .
we have not discussed here the dynamics of the solitons
i.e. , the motion of individual solitons@xcite or the collective motion of the sl@xcite which also produce experimental signatures of the incommensurate sl state .
we have focused instead on ground - state properties .
one of us has recently found that when the layer densities are made unequal by adjusting the gate voltages of the device ( keeping the total filling factor equal to one ) , the layer densities in the sl state become `` rippled '' , resulting in a dipole density wave.@xcite preliminary calculations show that the sudden onset of such a rippled sl state may give large contributions to the differential capacitance of 2lqh systems , especially for the interlayer capacitance ( `` eisenstein ratio'').@xcite further work on the properties of the rippled state is in progress , but it may be that sensitive measurements of differential capacitance in unbalanced ( @xmath308 ) 2lqh systems could detect the ci transition .
c.b.h . would like to thank b. i. halperin , n. read , and t. c. lubensky for helpful discussions , and the institute for theoretical physics ( itp ) at the university of california , santa barbara , where part of this work was carried out , for providing support through the itp scholars program .
this work was supported by an award from research corporation and by the national science foundation under grants dmr-9972332 , dmr-9714055 , and dmr-0087133 . for a review of double - layer quantum hall systems ,
see the experimental chapter by j. p. eisenstein and the theoretical chapter by s. m. girvin and a. h. macdonald in _ novel quantum liquids in low - dimensional semiconductor structures _ , edited by s. das sarma and a. pinczuk ( wiley , new york , 1995 ) .
j. p. eisenstein , h. l. strmer , v. narayanamurti , a. y. cho , a. c. gossard , and c. w. tu , phys .
lett . * 55 * , 875 ( 1985 ) ; j. p. eisenstein , h. l. strmer , v .
narayanamurti , a. y. cho , and a. c. gossard , surf .
sci . * 170 * , 271 ( 1986 ) ; a. potts , r. a. shepherd , w. g. herrenden - harker , m. elliot , c. l. jones , a. usher , g. a. c. jones , d. a. ritchie , e. h. linfield , m. grimshaw , j. phys .
cond . mat .
* 8 * , 5189 ( 1996 ) ; s. a. j. wiegers , m. specht , l. p. levy , m. y. simmons , d. a. ritchie , a. cavanna , b. etienne , g. martinez , and p. wyder , physica b * 251 * , 115 ( 1998 ) .
h. l. strmer , t. haavasoja , v. narayanamurti , a. c. gossard , and w. wiegmann , j. vac .
b * 1 * , 423 ( 1983 ) ; ines meinel , dirk grundler , silke bargstdt - franke , and christian heyn , appl .
. lett . * 70 * , 3305 ( 1997 ) ; i. meinel , t. hengstmann , d. grundler , d. heitmann , m. bichler , and w. wegscheider , phys .
* 82 * , 819 ( 1999 ) . | double - layer quantum hall systems possess interlayer phase coherence at sufficiently small layer separations , even without interlayer tunneling .
when interlayer tunneling is present , application of a sufficiently strong in - plane magnetic field @xmath0 drives a commensurate - incommensurate ( ci ) transition to an incommensurate soliton - lattice ( sl ) state .
we calculate the hartree - fock ground - state energy of the sl state for all values of @xmath1 within a gradient approximation , and use it to obtain the anisotropic sl stiffness , the kosterlitz - thouless melting temperature for the sl , and the sl magnetization .
the in - plane differential magnetic susceptibility diverges as @xmath2 when the ci transition is approached from the sl state . |
approximately 10% of the frii radio galaxies in the 3crr catalog are classified as ` x - shaped ' or winged because of their unusual radio morphologies @xcite .
these radio galaxies have two pairs of misaligned radio lobes of approximately equal linear extent .
they also tend to have low radio luminosities , typically lying near the fri / frii division .
there are two general hypotheses about the nature and origin of these structures . on the one hand
, it has been argued that the x - shaped morphology is essentially a hydrodynamic phenomenon which is the result of supersonic or buoyant flow of radio plasma in an asymmetric gas distribution . @xcite and @xcite argued that the x - shaped radio morphology is a result of the result of strong backflow of material behind the terminal hot spots of radio galaxy jets and subsequent buoyant evolution of the wings. @xcite hypothesized that the x - shaped morphology is the direct result of the supersonic expansion / inflation of the lobe into an elliptical atmosphere , and that all radio galaxies in such environments should exhibit this phenomenon . on the other hand
, it has been argued that this structure results from precession or reorientation of the jet axis , possibly as a result of a galaxy interaction and/or merger event of two supermassive black holes @xcite . in this scenario , the wings
represent the remnant lobe of an earlier epoch of nuclear activity , and only the primary jet / lobe is currently being powered by the nuclear outflow .
this latter scenario has generated particular interest recently because it implies that these galaxies may be strong sources of gravitational radiation @xcite .
the sensitivity of chandra and xmm - newton has been a crucial factor in making progress in the study of the gaseous environments of the radio galaxies and better understanding issues related to the dynamics of the radio structures .
in addition to emission from the thermal coronae , _ chandra _ and xmm - newton observations of radio galaxies have detected dozens of examples of non - thermal x - ray emission from their jets and hotspots .
the standard paradigm suggests that the x - ray emission observed from these features has different origins depending on the power of the radio galaxy . the x - ray emission from the knots of jets of fr
i radio galaxies are generally attributed to synchrotron emission from a population of ultra - relativistic electrons @xcite .
this conclusion is supported by several pieces of evidence including their broad - band spectra .
well studied examples include centaurus a @xcite , m87 @xcite , and 3c 66b @xcite . on the other hand , inverse - compton scattering of seed photons by relativistic electrons
is generally believed to be responsible for the x - ray emission from jets @xcite and hotspots @xcite of the more powerful frii sources .
well known hotspot examples include cygnus a @xcite and 3c 295 @xcite . to date ,
only two examples of x - ray emission from fr ii hotspots have been attributed to synchrotron emission , 3c 390.3 @xcite and pictor a @xcite , although @xcite argue that the x - ray emission from a large number of low power hotspots is at least partly synchrotron in origin .
3c403 is the nearest x - shaped radio galaxy in the 3crr catalog @xcite , and is an ideal target to study the x - shaped phenomenon . in this paper we present results from a _
observation of this radio galaxy .
we use archival vla and hst observations to make a multi - frequency study of the various components .
we have resolved x - ray emission from four distinct components : the bright active nucleus , the hot ism , several compact radio components , and diffuse x - ray emission coincident with the radio lobes .
this is the first high - resolution x - ray observation of an x - shaped radio galaxy and the primary goal of this observation was to study the ambient ism confining the radio emitting components .
in addition , the non - thermal x - ray emission that we detect from several of the compact radio components allows us to draw strong conclusions regarding their emission mechanisms .
this paper is organized as follows .
section two contains a brief description of the data , an observation log , and a summary of our analysis . in section three ,
we discuss the emission from the hot ism and nucleus , and we describe how we have separated these components spatially and spectrally . in section four , we discuss the non - thermal x - ray emission associated with the radio hotspots and lobes . the nature of x - shaped radio galaxies and the implications of our results are discussed in section five .
the final section contains a brief summary and conclusion . throughout this paper
, we assume the wmap cosmology ( @xmath7=71 , @xmath8=0.27 , @xmath9=0.73 ) , so that the measured redshift of the host galaxy ( @xmath10 ) corresponds to a luminosity distance of 260.6 mpc , and 1@xmath2 = 1.127 kpc .
the radio galaxy 3c403 was observed for 49470 s with the _ chandra_/acis - s instrument ( obsid 2968 ) in faint mode on december 12 , 2002 .
the nucleus of the galaxy was located at the best focus .
the lightcurve of events in the 5.0 - 10.0 kev bandpass on the entire s3 chip , excluding the nucleus and any other point sources visible by eye , was created using 259 s bins and examined for periods of flaring background .
intervals where the rate was more than 3@xmath11 above the mean rate were removed . the mean rate with periods of flaring removed is consistent with the nominal quiescent acis - s background listed in the proposer s observation guide .
a total of 4600 s of data was excluded , leaving 44870 s of good time .
bad pixels , hot columns , and columns along node boundaries were also removed .
this galaxy is located at relatively low galactic latitude ( @xmath12 , @xmath13 ) , so the nominal blank sky background files generated from deep , high galactic latitude observations are not appropriate .
fortunately the radio galaxy covers only a small region of the s3 chip , so that local background for spectral analysis was estimated from a large region on the s3 chip north of the source .
absorption by foreground gas in our galaxy ( @xmath14=1.54@xmath410@xmath15 @xmath6 ) was included in all spectral fits .
we have combined data from our _ chandra_/acis - s observation with archival vla observations ( 1.48 and 8.4 ghz ) and an hst observation to make a multi - frequency study of various components .
detailed discussions of the 1.48 and 8.4 ghz vla data have been presented in @xcite and @xcite , respectively .
we will use the nomenclature of the @xcite throughout this paper when referring to specific knots or radio features .
an 8.4 ghz radio map of this radio galaxy is shown in figure [ radioimg ] .
the primary lobes can be seen to the e and w of the central nucleus .
the lower surface brightness ` wings ' that give this object its x - shaped morphology extend to the n and s. several hotspots can be seen in the eastern lobe , and perhaps a hint of a jet between the nucleus and knot f6 .
the x - ray and radio positions of the active nucleus agree to @xmath10.5@xmath2 .
this radio galaxy was also observed for 280 s with the hst / wfpc2 instrument in 1995 as part of the hst 3cr snapshot survey using the f702w filter ( pivot wavelength of 6919 ) @xcite .
we obtained the reprocessed data from the hst archive and used the iraf synphot package to apply photometric calibrations .
the fluxes were reddening - corrected using the dust maps of @xcite . an adaptively smoothed , exposure corrected , background subtracted x - ray image in the 0.5 - 2.0 kev bandpass with 8.4 ghz radio contours overlaid is shown in figure [ xradovl ] .
x - ray emission is detected from at least four distinct components : the active nucleus , the hot gas of the host galaxy , several of the radio hot spots , and diffuse emission from the lobes and the wings .
a raw x - ray image in the 0.5 - 2.0 kev bandpass of the central region ( 20@xmath1614@xmath2 ) , including the hot ism and the active nucleus , is shown in figure [ central ] .
the point source at the center is the agn , and the extensions to the ne and sw are the hot ism .
the morphology of the ism is highly elliptic , and co - aligned with , but more eccentric than , the optical isophotes of the stellar light distribution @xcite .
we wish to separate the emission of the hot ism from that of the active nucleus in order to study each component separately . unfortunately , these components are not well separated spatially .
the major and minor axes of the ism shown in figure [ central ] are approximately 7.5@xmath2 and 3.0@xmath2 in length , respectively .
the 90% encircled energy radius for a point source on - axis at 1.49 kev is @xmath10.9@xmath2 .
we have used images in four x - ray bands to guide our spectral analysis in order to separate the emission from the two components .
an examination of images in the 0.5 - 0.7 , 0.7 - 1.0 , 1.0 - 1.5 , and 1.5 - 2.0 kev bands showed that above 1 kev , most of the emission is unresolved and is probably due to the active nucleus .
below 0.7 kev , the contribution from the central point source is small , and most or all of the emission is from the ism .
we extracted the spectrum , shown in figure [ spectrum ] , from a 7.5@xmath2 radius circular region centered on the nucleus .
the data have been binned to a minimum of 10 counts per bin , and the background was determined from a large region on the s3 chip approximately 4@xmath17 to the n of the nucleus , well removed from any detectable emission from the galaxy .
the spectrum consists of two peaks : one centered at @xmath10.7 kev with a high energy tail , and a second that peaks at around 5 kev .
the count rate in the 0.25 - 10.0 kev band was 4.7@xmath410@xmath18 cts s@xmath19 .
the majority of the counts in this spectrum are from the unresolved active nucleus .
this suggests that this spectrum moderately suffers from the effects of pile - up .
based on figure 6.19 of the _ chandra _ proposer s observatory guide ( pog ) , we estimate a pile - up fraction of @xmath15% .
we note that we fitted the data using coarser binning ( 25 and 50 counts per bin ) to ensure the applicability of gaussian statistics and find statistically identical parameter values and uncertainties .
we initially modeled the spectrum with a three - component model including a heavily absorbed ( @xmath2010@xmath5 @xmath6 ) power law , an fe line with variable centroid and width , and an apec model .
the effect of the pile - up was modeled by using the xspec function ` pileup ' @xcite , and galactic absorption ( @xmath14=1.54@xmath410@xmath15 @xmath6 ) was included in all fits . for simplicity ,
the elemental abundance of the apec model was fixed at 1.0 times solar , typical of the central regions of early - type galaxies . if allowed to vary freely , the elemental abundance was poorly constrained . at the relatively low best - fit temperature ( @xmath10.3 kev ) , the emission is line dominated , and the elemental abundance can be traded against the normalization .
the radiative cooling function , @xmath21 , scales nearly linearly with abundance at these temperatures and abundances .
none of our conclusions below are sensitive to this choice , although a larger elemental abundance would reduce our estimates of the gas density below ( @xmath22 ) . in this scenario ,
the power law represents emission from the agn , and the apec model the emission from the hot ism .
we could not obtain an acceptable fit with this three component model as there were large residuals between 1 and 2 kev .
inspection of images in the 1 - 2 kev band suggests that most of the emission in this band is is due to the unresolved agn .
we therefore added a second , less absorbed power law to the model and were able to obtain an acceptable fit .
an examination of the images in several bands ( described above ) suggests that this second power law is heavily absorbed .
we fixed the absorption and spectral index of this second power law to 4@xmath410@xmath15 @xmath6 and 2.0 respectively to minimize the number of free parameters ; only the normalization of this component was allowed to vary freely .
the choice of a spectral index that is steeper than that of the primary absorbed power law and the choice of the value of @xmath14 constrain this component to contribute little above and below the 1 - 2 kev band .
the best - fit model has been overlaid onto figure [ spectrum ] .
the data and model around the fe k line is more clearly shown in figure [ fespectrum ] .
the best - fit values of the parameters and 90% confidence intervals are summarized in table [ specparms ] .
the morphology of the x - ray emission of the central region of the hot ism is highly elliptical and co - aligned with the optical isophotes of the host galaxy .
we have overplotted the contours from an adaptively smoothed x - ray image in the 0.5 - 2.0 kev band onto an hst / wfpc2 image of the host galaxy as shown in figure [ xoptovl ] .
the position angle of the diffuse emission is not aligned with the primary e / w radio components , and is almost perpendicular to the n and s radio wings .
the spatial relationship between the lobes , the wings , and the ism , at least near the nucleus , is qualitatively what we would expect in the buoyancy / backflow model .
based on the spectral analysis above , the unabsorbed x - ray luminosity of the ism within 8.5 kpc of the nucleus is 2.9@xmath410@xmath23 ergs s@xmath19 in the 0.25 - 10.0 kev bandpass .
this x - ray luminosity is at the high end of the @xmath24 versus @xmath25 relationship derived from observations of large samples of elliptical galaxies in @xcite and @xcite .
the temperature of the gas in this central region is @xmath10.3 kev , rather low for a massive elliptical galaxy , perhaps indicative of a cooling flow .
we fitted an elliptical @xmath26-model plus constant background to an x - ray image of the central 5@xmath2 of the galaxy in the 0.3 - 1.0 kev bandpass using the sherpa software package in order to characterize quantitatively the morphology of this emission .
as demonstrated above , the contribution from the nucleus in this band is small .
the free parameters of this model were @xmath26 , the scale parameter @xmath27 , the normalization , the ellipticity , @xmath28 , the position angle of the major axis of the ellipse , @xmath29 , and the background , @xmath30 .
the telescope psf was computed using the ciao program _
mkpsf _ to remove the instrumental response .
both @xmath28 and @xmath29 were well constrained , with @[email protected] and @xmath29=44@xmath322@xmath33 ( measured east of north ) .
the uncertainties are at 90% confidence . based on the spectral analysis described above , the 0.3 - 1.0 kev band contains little emission from the active nucleus . the parameters @xmath26 and @xmath27 were poorly constrained , but @[email protected] and @xmath27@xmath353@xmath2 ( @xmath353.5 kpc ) .
using nominal values of @xmath26=0.55 and @xmath27=0.6@xmath2 , the central proton density , @xmath36 , is @xmath11.14 @xmath37 for the best - fit spectrum described above .
the values of @xmath27 and @xmath36 are somewhat smaller and larger , respectively , than typically observed in elliptical galaxies , although it is not uncommon to find small scale , dense cores in the ism at the centers of radio galaxies @xcite .
the cooling time of this material is very short , @xmath110@xmath38 years , because of the high density and relatively low temperature .
the cooling time is insensitive to the adopted abundance , since a metal abundance higher than we have assumed ( 1.0 times the solar value ) would decrease the central density but increase the cooling time ( the cooling time scales roughly as z@xmath39 ) . on the kpc scale ,
the x - ray morphology , and presumably the density and pressure profiles , are highly elliptical ( @xmath28=0.57 ) .
in contrast , the ellipticity of the stellar light profile is only 0.25 at 6.9@xmath2 ( the effective radius of the de vaucouleurs profile for the host of 3c403 ) from the nucleus @xcite .
we have detected x - ray emission to only half this distance along the semi - major axis , but an examination of the central 3@xmath2 of the hst images clearly shows that the stellar distribution has a roughly constant ellipticity ( @xmath10.25 ) down to about 1@xmath2 , where the distribution becomes circular . the distribution of the gas is therefore more elliptical than that of stars . in the central 10 - 15 kpc of a elliptical galaxies , the stellar mass typically dominates the dark matter so the gravitational potential should be determined by the distribution of the former . if the gas is in hydrostatic equilibrium with the gravitating matter , we expect that the ellipticity of the gas and that of the gravitating matter should be similar , but in fact this is not the case in 3c403 .
we suggest two possible explanations for this .
the first , and most likely in our view , is that the gas is not in hydrostatic equilibrium .
the cooling time of this gas is less than 10@xmath38 years , roughly the same age as the radio lobes .
it is plausible that the gas is cooling and falling in toward the center .
the inflation of the radio lobe may have disrupted the central region as well . if the gas is moving , it may well have a different morphology than the gravitating matter . as an alternative , it is possible that the emission which we have been attributing to the ism is , in fact , due to a partially resolved kiloparsec scale jet .
we consider this unlikely for two reasons .
first , there is no evidence for any radio counterpart to this emission .
the x - ray to radio spectrum would have to be unusually flat .
second , the axis of this hypothetical kpc scale jet is not aligned with the larger scale , tens of kpc radio features , which implies that the jet has been sharply bent . on larger ( tens of kpc ) scales ,
the data are not of sufficient quality for us to determine whether the gas distribution remains elliptical at or beyond the boundaries of the wings and lobes .
we can , however , address the simpler question of whether the radio wings are greatly overpressurized relative to the ambient medium .
there is statistically significant emission above background out to about 50@xmath2 from the nucleus , although the profile of the emission is poorly constrained .
this could represent emission associated with a group of galaxies for which the optical host of 3c403 is the dominant member .
the surface brightness profile of the gas from the core to @xmath150@xmath2 can not be described by a single @xmath26-profile , however .
given the small core radius and high density of the emission in the central region , it is not reasonable simply to extrapolate this to larger radii to estimate the gas pressure for a comparison with the pressure of the lobes and the wings .
for simplicity , and because it would be difficult to maintain a gas distribution to large radii that is more elliptical than the stellar mass distribution , we assume that the profile is spherically symmetric and parameterize the data with a circular @xmath26-profile . we note that using an elliptical profile ( e.g. with the same ellipticity that we found for the smaller scale emission ) would not significantly change our conclusions below . a plot of the surface brightness in circular annuli versus distance from the nucleus is shown in figure [ sbprof ] .
x - ray knots , hotspots , and background point sources have been excluded .
we measure a flux of 1.32@xmath410@xmath40 ergs @xmath6 s@xmath19 ( unabsorbed ) in the 0.1 - 10.0 kev band in an annular region 50@xmath2 in radius from the nucleus , excluding the central 10@xmath2 .
the x - ray luminosity of this emission is @xmath11.1@xmath4 10@xmath23 ergs s@xmath19 in the same band .
we have assumed a gas temperature of 0.7 kev , typical for groups , and an elemental abundance of 0.4 times solar .
for an assumed spherical @xmath26-model profile with @xmath26=0.5 and @xmath27=30@xmath2 , the central proton density is @xmath13.3@xmath410@xmath41 @xmath37 .
the gas pressure at the distance of the center of the wings is @xmath12@xmath410@xmath42 dyn @xmath6 .
the equipartition pressure of the wings is @xmath15@xmath410@xmath43 dyn @xmath6 ( see section 4.2 ) .
if all of the x - ray emission from the wings is due to ic scattering of cmb photons , the pressure is @xmath11@xmath410@xmath42 dyn @xmath6 ( see below ) .
thus , the internal minimum pressure of the radio wings is the same order of magnitude as the pressure of the ambient medium .
our estimate of the gas pressure has large uncertainties , and we have ignored the effects of projection which would place the lobes further out into the halo and into lower density gas . given the uncertainties , it is reasonable to conclude that the lobes are in approximate pressure equilibrium with the ism .
there is no evidence to suggest that the wings are greatly overpressurized relative to the ambient medium .
the spectrum of the nucleus is well described by a two power - law model plus an fe line .
the primary power - law is heavily absorbed ( @xmath34@xmath410@xmath5 @xmath6 ) .
we attribute the primary power - law component to emission from material near the inner part of the accretion close to the central supermassive black hole .
there is no evidence for a significant amount of dust in hst observations of the host galaxy @xcite , so it is likely that this large absorption column is due to cool material in close proximity to the central black hole as well , perhaps a molecular torus .
this is consistent with previous optical emission line studies that have shown that 3c403 is a narrow emission line galaxy ( nelg ) @xcite .
this suggests that we are observing the disk / torus nearly edge on , and that the jets are advancing at an angle close to the plane of the sky .
the broad line region , if it exists , must be buried behind the torus and unobservable .
neither the photon index ( [email protected] ) nor the unabsorbed x - ray luminosity ( 1.07@xmath410@xmath44 ergs s@xmath19 in the 0.25 - 10 kev bandpass ) are unusual for the nuclei of radio galaxies .
the parameters of the second power - law component are poorly constrained , and as described above , we have fixed the power - law index ( @xmath45=2 ) and absorption ( @xmath14=4@xmath410@xmath15 @xmath6 ) of this component for consistency with the imaging data .
we suggest that this second power - law component is x - ray emission from an unresolved jet that is partially obscured , as observed with greater sensitivity in centaurus a @xcite .
the radio luminosity of the core is 6.5@xmath410@xmath15 w hz@xmath19 sr@xmath19 at 5 ghz .
the x - ray luminosity of the second power - law component is 7.1@xmath410@xmath46 w hz@xmath19 sr@xmath19 at 1 kev .
the x - ray to radio flux ratio is consistent with that found in a sample of low - power radio galaxies @xcite , and supports the hypothesis that this second power - law component is related to spatially unresolved emission from a jet .
it is likely that the excess column density is due to absorption within a few pc of the nucleus .
we speculate that this unresolved jet lies above the accretion disk , and our line of sight to the jet passes through a region of lower column density higher up in the torus .
there is no evidence of gas or dust on kpc scales , perhaps from a recent merger , that could be responsible for this large column .
it is also possible that the hot ism is responsible for this large column .
the column density of the hot gas within a distance @xmath27 ( the central core radius ) is @xmath47 1.0@xmath410@xmath15 @xmath6 , remarkably close to the 4@xmath410@xmath15 @xmath6 value chosen based on the multi - band images .
a careful calculation would require detailed knowledge of the ionization state of the gas , the distribution of gas along the line of sight , and the elemental abundance , all of which are poorly constrained .
alternatively , it is possible that this second power - law component is due either to electron scattering from an ionized corona or a partial covering fraction absorber , both of which are often invoked to explain the x - ray spectrum of seyfert galaxies @xcite
. the observed flux ratio of the first and second power - law components is consistent with that observed in seyfert galaxies .
the jet hypothesis is , however , more natural in our view as 3c 403 is an frii radio galaxy .
this hypothesis could be evaluated by a vlbi observation to search for a pc - scale jet near the core .
a strong fe line with ew=244@xmath3120 ev ( 90% confidence ) has also been detected .
the centroid of the line is [email protected] kev , and the width ( r.m.s . ) of the line is 80@xmath3150 ev .
the centroid of this line is consistent with fluorescent emission from cold material .
the large ew of this line is consistent with a model where the line emission originates from an absorbing region that surrounds the central supermassive black hole @xcite .
the measured width of the fe line corresponds to a velocity ( fwhm ) of 900 km s@xmath19 .
if the line broadening is the result of fluorescence from cold gas in keplerian motion around a 10@xmath48 m@xmath49 black hole , it must lie @xmath15 pc from the black hole .
it is reasonable to attribute the large column and fluorescent fe line to a molecular torus that surrounds the central supermassive black hole .
as shown in figure [ elobeovl ] , small - scale x - ray emission is associated with multiple radio components in the eastern lobe : these include the radio components f7/f8 , f6 , f5 , f1 and a distinct radio component upstream of f1 which we denote f1b .
there is no detected x - ray emission associated with the hotspot components f2 or f3 in the e lobe or with the larger hotspots p1 and p2 in the western lobe .
figure [ wings ] shows that there is also large - scale diffuse emission associated with the western lobe and with both the extended wings .
the position of the active nucleus in the x - ray data using the nominal aspect solution agrees with the radio position to better than 0.5@xmath2 .
we have not attempted to improve the aspect solution beyond this .
we begin by considering the compact components in the e lobe shown in figure [ elobeovl ] . for the x - ray emission associated with the radio hotspots f1 and f6 ,
there are enough counts to fit a power - law spectrum .
we find best - fitting photon indices of @xmath50 and @xmath51 for f1 and f6 respectively , and normalizations corresponding to flux densities at 1 kev of @xmath52 and @xmath53 njy .
for the other three compact components there are not enough counts to extract a spectrum , so we estimated the flux densities based on the count rates assuming a photon index of 2 .
( the flux density is insensitive to our assumed photon index : for a photon index of 1.5 the flux densities would decrease by @xmath54% . )
flux densities for all the eastern lobe components are tabulated in table [ compact ] .
we use the 8.4 ghz maps @xcite to measure radio flux densities corresponding to the x - ray detections . for compact radio components it
is possible to extract radio flux densities either by integrating over the whole x - ray extraction region ( with local background subtraction ) or by fitting a model , such as a homogeneous sphere ( as was done by @xcite ) or a gaussian . in table
[ compact ] we tabulate radio flux densities extracted using both methods , where they were both possible , which is true of the compact components f1 and f6 ; this illustrates the range of possible radio flux densities that can be said to correspond to the x - ray emission .
we also tabulate the ratio of x - ray to radio flux . the compact components f1 and f6 are detected with hst ; the hst image with x - ray contours overlaid
is shown in figure [ xknotovl ] .
the radio - optical - x - ray spectra of these two knots are shown in figure [ knotspec ] .
we have fitted synchrotron models to these data , and both of these components are well fitted by a single standard continuous injection model @xcite . as pointed out by @xcite ,
inverse - compton models require extreme conditions to reproduce the observed x - ray flux densities .
the knots must either be far from equipartition or they must be moving relativistically at an angle close to the line of sight .
a synchrotron model seems to be the only plausible explanation for the x - ray emission from f1 and f6 . for the other components , the existing short hst observation does not give us useful constraints (
these features are resolved at 8.4 ghz with an @xmath55 beam , and therefore spread over many hst wfc pixels ) and we do not have x - ray spectral measurements .
however , the radio to x - ray ratios for the other detected components ( table [ compact ] ) are very similar to those measured for f1 and f6 , so it would seem plausible that they have the same emission mechanism . an interesting feature of this interpretation is that some of the x - ray detected components of 3c403 are morphologically part of the jet rather than being hotspots : this is true of f1b , f5/4 and ( the clearest case ) f7/8 , and it is also debatable whether f6 should be considered a jet knot rather than a primary hotspot .
x - ray emission from hotspots is of course seen in many other sources , and has been attributed to synchrotron emission in some cases ( e.g. 3c390.3 , @xcite ) , although 3c403 is one of the clearest cases of synchrotron rather than inverse - compton emission @xcite .
however , 3c403 is one of the best examples of synchrotron x - ray emission from the jet of a powerful _ narrow - line _ radio galaxy . in unified models of frii radio galaxies ,
the jets of narrow - line radio galaxies are at large angles to the line of sight @xcite ; this is true even for lower - luminosity objects , whose unification partners must be broad - line radio galaxies rather than quasars @xcite . in both cases ,
the numbers of the different classes of objects observed require that the narrow - line objects be at angles greater than @xmath56 to the line of sight .
such large angles effectively rule out the possibility of applying the popular beamed inverse - compton model @xcite to the jet in 3c403 .
we have calculated the expected flux density in this model for the jet component f7/8 for a range of angles and bulk lorentz factors , and find that the observed flux density can only be reproduced with large bulk lorentz factors ( @xmath57 ) and small angles to the line of sight ( @xmath58 ) . while the large bulk lorentz factors can not be ruled out , such small angles to the line of sight are inconsistent with everything that is known about 3c403 .
a synchrotron model appears to be the only viable one here .
it is unclear what fraction of frii jets are dominated in the x - ray band by synchrotron emission , but they may be common .
the inverse - compton process may dominate only in those sources whose jets are highly beamed .
finally , we wish to draw attention to two other interesting features of the compact x - ray emission .
the first is the non - detection of the hotspot component f2 .
the upper limit on the x - ray emission from this hotspot is an order of magnitude below the level at which it would have been detected if its radio to x - ray ratio had been the same as the detected components. in a synchrotron interpretation , this requires a difference in the particle acceleration properties of f1 and f2 .
it is possible that f2 is a relic hotspot that is now detached from the jet flow and that has no ongoing particle acceleration .
the second interesting feature is the clear extension of the brightest x - ray feature , f6 , back towards the nucleus .
figure [ f6knot ] contains an x - ray image of the knot f6 in the 0.5 - 2.0 kev band with 8.4 ghz radio contours overlaid .
the peaks of the radio and x - ray emission are well aligned , and both knots are clearly extended .
the x - ray extension is to the w of the main peak , and the radio extension to the sw .
there is no clear radio counterpart to the x - ray extension and if the extension is considered to be radio - related , it has an x - ray to radio flux density ratio more than a factor 10 above most of the other features in the jets ( based on an upper limit to the possible associated radio emission ) .
this sort of x - ray to radio ratio is not impossible for a synchrotron feature knots in the centaurus a jet show similar properties but its location just upstream of a bright hotspot or jet knot is certainly intriguing , and reminiscent of the ` offsets ' seen in the jets of fri radio galaxies @xcite
. extended x - ray emission from the lobes of frii radio sources is normally attributed to inverse - compton emission ( e.g. @xcite ) .
typically the x - ray flux densities measured from the lobes are close to , or slightly above , the value expected for inverse - compton scattering of the microwave background radiation by lobes with magnetic field strengths close to the equipartition value .
the detected extended emission from the w lobe and wings of 3c403 is too faint and diffuse for us to fit a spectrum , so that we can not test the inverse - compton model s prediction of a flat ( photon index 1.5 ) x - ray spectrum .
however , we can estimate the 1-kev flux densities of the various components from the observed count rate . here
we assume a photon index of 1.5 in converting between counts and flux density .
the results are tabulated in table [ extended ] .
the x - ray counts are background subtracted , and the uncertainties are statistical ( poisson ) .
we measured corresponding radio fluxes from 1.4 ghz and 8.4 ghz radio maps ( the low - resolution map of @xcite and a map made from the 1.4-ghz b - configuration data of @xcite ) .
these are also tabulated in table [ extended ] .
modeling the lobes and hotspots as being in the plane of the sky , we then calculated the expected inverse - compton flux density at equipartition . to do this
we assumed no relativistic protons , a low - energy electron energy cutoff corresponding to @xmath59 , and a low - energy electron energy index of 2 ( corresponding to @xmath60 . a high - energy break and cutoff
were included to fit to the steep spectral index between 1.4 and 8.4 ghz , and the radio spectra were normalized to the appropriate fraction of the 178-mhz flux density of 28.3 jy ( see @xcite ) this is justified given the very uniform spectral index found by @xcite .
the inverse - compton calculation was then carried out using the code of @xcite .
we have assumed cylindrical geometry for the wings , and spherical geometry for the w lobe .
the results are tabulated in table [ extended ] .
the ratios of observed to predicted x - ray flux are around 4 for the two wings , which is not particularly unusual for radio galaxies @xcite and suggests field strengths within a factor of @xmath61 of the equipartition value in the wings .
the discrepancy with the equipartition prediction would be further reduced if there were any projection of the wings .
if we assume no projection , and take the x - ray flux density as giving a magnetic field strength measurement , then the electron energy density dominates over the magnetic field energy density in these lobes by around a factor 20 , and the non - thermal pressure in the wings is about @xmath62 dyn @xmath6 ( with large uncertainties because of the large formal error on the count measurement ) . the western lobe , on the other hand , has a ratio of observed to predicted flux that is much larger than is typical , and the distribution of x - ray counts in this region is also not what would be expected from a simple inverse - compton model , in which it should more or less match the low - frequency radio emission ( fig .
[ wings ] ) .
the x - ray emission is not peaked at the radio peak in the n end of the lobe , rather the x - ray emission uniformly fills the lobe .
therefore , this emission can not be explained by inverse - compton scattering of cmb photons by the electrons responsible for the radio emission .
because of the differences in radio and x - ray morphology , this emission is unlikely to be a shock heated shell of gas surrounding the lobe that would be created if the lobe were expanding into the ism supersonically @xcite .
it is possible that the x - ray emission from the w lobe is due to inverse - compton scattering of beamed ir / optical photons from the active nucleus by a lower energy , unobservable , population of relativistic electrons .
such a model has been used to explain diffuse x - ray emission from the western lobe of the quasar 3c 207 @xcite .
as summarized in the introduction , there are two general scenarios commonly invoked to explain the x - shaped morphology of radio galaxies such as 3c403 .
one line of argument suggests that the x - shaped morphology is a hydrodynamic phenomenon , and the other is that x - shaped radio galaxies are formed as the result of a sudden change in the jet axis of agn , perhaps as the result of a galaxy merger and/or the merger of two supermassive black holes . within the hydrodynamic scenario , there are two general models in the literature : the x - shaped structure is the result of strong ( perhaps highly supersonic ) backflow and subsequent buoyant evolution @xcite , or that both the lobes and wings are highly supersonic ( relative to the ambient ism / icm ) and the observed structures are the natural result of over - pressurized expansion / inflation of the lobes into an asymmetric medium @xcite .
we prefer the backflow / buoyancy model because , in our view , the observed structures of x - shaped radio galaxies are the natural consequences of strong backflows in an asymmetric medium , and there is no indication that 3c403 s lobes are highly overpressured . we know that strong backflows exist in radio galaxies , and that there are many cases where the morphology of the ambient medium in clusters , groups , and isolated ellipticals is complex .
the dynamics of the backflow of material from the jet head are determined by several factors including the jet velocity , the density contrast between the jet and the ambient medium , and the temperature and density profile of the medium into which the jet is propagating @xcite . in particular , lighter jets ( i.e. jets with a larger values of @xmath63/@xmath64 ) have more powerful backflows than heavier jets .
backflow velocities ranging from several thousand km s@xmath19 to 0.1@xmath65 have been estimated for a sample of double lobed 3cr radio galaxies @xcite .
thus backflow velocities can be orders of magnitude higher than the advance speed of the jet head . once the backflow material reaches a strong pressure gradient , it will rapidly decelerate , bend ( depending on the details of the gradient ) , and subsequently it will be driven primarily by buoyancy forces ( i.e. subsonically ) .
the density and pressure morphologies of the hot gas in clusters , groups , and elliptical galaxies is complex .
if _ chandra _
observations of clusters of galaxies are any guide , there are no such things as spherically symmetric gas halos sitting idly in their gravitating dark matter potentials ( markevitch , 2004 , private communication ) . in our view , there is no compelling reason to invoke more complex phenomena such as merging supermassive black holes to explain x - shaped radio galaxies .
the arguments against the hydrodynamic models are based primarily on three considerations : the radio morphology of these objects , evidence for recent mergers , and radio spectral differences of the wings and lobes .
none of these are strong arguments in our view and we suggest that it is quite plausible ( and even natural ) for backflow / buoyancy to create the structures that are observed .
the most important of the three are those based on the radio morphology . in some sources ,
the wings ( in projection ) are longer than the lobes ( as is the case in 3c403 ) @xcite , and in others the alignment of the wings and lobes relative to the nucleus is sometimes offset ( e.g. ngc 326 ) @xcite .
qualitative arguments are often made that the observed morphologies are unlikely to be generally formed by hydrodynamic flows .
the implicit assumption in these arguments is that the gas atmosphere that the jet is propagating into is roughly spherically symmetric .
the observed radio structures will depend critically on the distribution of the gas into which the jet is propagating .
similarly , the general paradigm suggests that the lobes of fr iis are advancing supersonically into the ambient medium , and it has been argued that it is not plausible that the wings and lobes are approximately the same length if the lobes are evolving supersonically and the wings buoyantly @xcite .
this claim is true _ if _ the lobes are advancing supersonically into the ambient medium , but there is no direct evidence to support this claim . in all cases where the internal and external pressures of fr ii lobes have been measured to date ,
the lobes appear to be close to pressure equilibrium with their surroundings , as we observe in 3c403 @xcite , implying relatively slow expansion .
the complex radio morphology of 3c403 is clearly different from the canonical lobe / hotspot morphology of classic fr ii radio galaxies such as cyg a , and its radio luminosity lies on the fri / frii dividing line .
both of these suggest that 3c403 is among the least powerful sources in this class .
the inner radio lobes of m87 , a radio galaxy on the fri / ii luminosity boundary , have detached from the shock propagating through the icm of the virgo cluster and are thus evolving buoyantly / subsonically @xcite . additionally , some have argued that there is evidence for recent mergers in the host galaxies of some of these sources .
this is certainly not a general conclusion as 3c403 , one of the best studied x - shaped radio galaxies , resides in a relatively poor environment and shows little evidence for a recent ( within the last @xmath110@xmath66 years ) merger .
hst observations of the host galaxy of 3c 403 indicate the presence of weak dust lanes , suggestive of a much older merger , certainly older than the age of the radio source , estimated to be @xmath116 myr based on radio spectral analysis @xcite . finally , multispectral radio observations of several winged sources ( e.g. ngc 326 @xcite and b2 0828 + 32 @xcite ) suggest that the plasma in the wings is considerably older than that in the lobes . variations in radio spectral index across the wings and lobes would be a natural result if the jet axis were slowly precessing
. however , this observational result is also consistent with the hydrodynamic backflow models in our view for two reasons .
first , the plasma in the wings is naturally older than that in the lobes since it traveled back from the jet head to its present location in the wings .
second , the evolutionary history of the material in the wings is clearly different from that of the material currently in the lobes in the backflow model .
the spectral differences could well be telling us something about the hydrodynamics and the nature of shock acceleration of the backflow , and be unrelated to radiative energy loss of the most energetic particles .
it is impossible to draw definitive conclusions from spectral differences between the wings and lobes .
the quality of our data presented in this paper is not sufficient for us to put strong constraints on the dynamics of the lobes / wings interaction with the ism and make a definitive , general statement regarding the nature of x - shaped radio galaxies .
we have detected non - axisymmetric structures in the central region of the host galaxy .
this result is consistent with the various hydrodynamic scenarios that have been invoked to explain this phenomenon .
deeper observations of this object and others with _ chandra _ and/or xmm / newton are required to better understand the interaction between the gaseous corona and the radio lobes in order to resolve this issue .
the primary results of this work include the following : * we have detected x - ray emission from the hot ism , the active nucleus , several radio knots , and diffuse emission from the lobes and wings of the nearby x - shaped radio galaxy 3c403 . * within @xmath15 kpc of the nucleus ,
the morphology of the hot gas is highly ( @xmath670.57 ) elliptic .
this is consistent with the hypothesis that the x - shaped radio morphology is the result of the propagation of jets in non - axisymmetric atmospheres .
the distribution of the stars in the central 5 kpc is considerably less elliptical ( @xmath670.2 ) .
our data are not of sufficient quality for us to make a good measurement of the density / pressure profile of the hot gas on larger scales , but within the uncertainties , the lobes and wings are in approximate pressure equilibrium with the hot corona .
* we have detected x - ray and optical emission from radio knots and hotspots . based on their radio / optical / x - ray ,
we attribute the emission in all cases to synchrotron emission , not ic scattering from the cmb .
all lines of evidence suggest that the jet and radio lobes lie at a large angle to the line of sight , so the effects of relativistic beaming are negligible .
this is one of the best examples to date of synchrotron x - ray emission from the compact radio features of a narrow - line fr ii radio galaxy . * the x - ray spectrum of the nucleus
is well modeled by two power laws plus a broad , fluorescent fe line from cold material .
the first power law is heavily absorbed and is buried deeply in the torus .
the second power law is less heavily absorbed , and we attribute this to x - ray emission from an unresolved jet . the ratio of x - ray to radio flux is consistent with that measured for the nuclei of other radio galaxies .
* diffuse x - ray emission is detected from the lobes and wings .
this emission can generally be explained as ic scattering from the cmb if we assume moderate departure from equipartition .
this work was supported by nasa contracts nas8 - 38248 , nas8 - 39073 , the royal society , the chandra x - ray center , and the smithsonian institution .
we would also like to thank the anonymous referee for comments and suggestions that improved this paper .
leahy , j. p. , and parma , p. 1991
, in proceedings of the 7th iap meeting , extragalactic radio sources - from beams to jets , ed .
j. roland , h. sol , and g. pelletier ( cambridge , cambridge university press ) , 307 .
.summary of spectral parameters for nucleus and hot ism .
all uncertainties are at 90% confidence .
the count rate is in the 0.25 - 10.0 kev band .
the fluxes and luminosities listed for each component are unabsorbed in the 0.25 - 10.0 kev band .
the equivalent width of the fe line is relative to the first power - law component . [
cols="^,^,^",options="header " , ] | we present results from a 49.4 ks _ chandra_/acis - s observation of the nearby ( @xmath0=0.059 ) x - shaped frii radio galaxy 3c 403 .
this is the first _
chandra _
observation of an x - shaped radio galaxy , and one of the goals of this pioneering study is to determine the relationship between the x - ray emitting gas and the x - shaped radio morphology .
we find that the x - ray isophotes of the hot gas within @xmath13.5@xmath2 of the central galaxy are highly elliptical ( [email protected] ) and co - aligned with the elliptical optical isophotes .
this supports the hypothesis that x - shaped radio sources are created by propagation of jets through asymmetric density distributions . within large uncertainties
, there is no evidence that the lobes or wings are overpressurized relative to the ism , supporting the scenario in which the wings are the result of strong backflow of material from the jet head and subsequent buoyant evolution .
we have detected x - ray emission from several of the radio knots to the e of the active nucleus , and diffuse emission from the radio lobe to the w. the x - ray emission from the eastern knots can not be explained by an inverse compton model unless they are far from equipartition . using archival hst data , optical emission is detected from two knots , and the radio / optical / x - ray spectra are well fitted by simple synchrotron models .
this is one of the strongest examples to date of x - ray synchrotron emission from multiple knots in the jet of an fr ii radio galaxy .
x - ray emission is also detected from the radio wings at a flux consistent with inverse compton scattering of cmb photons from relativistic electrons if the wings are near equipartition .
the nuclear spectrum is well described by a multi - component model that includes a heavily absorbed power law ( @xmath34@xmath410@xmath5 @xmath6 ) and a bright ( ew@xmath1244 ev ) , broadened fe line .
a second , less absorbed , power - law component , likely to represent unresolved emission from a pc - scale jet , is also required . |
algorithmic information theory ( ait , for short ) is a framework for applying information - theoretic and probabilistic ideas to recursive function theory .
one of the primary concepts of ait is the _ program - size complexity _ ( or _ kolmogorov complexity _ ) @xmath6 of a finite binary string @xmath7 , which is defined as the length of the shortest binary program for an optimal computer to output @xmath7 . here
an optimal computer is a universal decoding algorithm . by the definition , @xmath6 is thought to represent the amount of randomness contained in a finite binary string @xmath7 , which can not be captured in a computational manner . in particular , the notion of program - size complexity plays a crucial role in characterizing the _ randomness _ of an infinite binary string , or equivalently , a real . in @xcite we introduced and developed a statistical mechanical interpretation of ait .
we there introduced the notion of _ thermodynamic quantities at temperature @xmath4 _ , such as partition function @xmath0 , free energy @xmath1 , energy @xmath2 , and statistical mechanical entropy @xmath3 , into ait .
these quantities are real functions of a real argument @xmath8 .
we then proved that if the temperature @xmath4 is a computable real with @xmath9 then , for each of these thermodynamic quantities , the partial randomness of its value equals to @xmath4 , where the notion of _ partial randomness _ is a stronger representation of the compression rate by means of program - size complexity .
thus , the temperature @xmath4 plays a role as the partial randomness of all the thermodynamic quantities in the statistical mechanical interpretation of ait . in @xcite
we further showed that the temperature @xmath4 plays a role as the partial randomness of the temperature @xmath4 itself , which is a thermodynamic quantity of itself .
namely , we proved _ the fixed point theorem on partial randomness _ , which states that , for every @xmath5 , if the value of partition function @xmath0 at temperature @xmath4 is a computable real , then the partial randomness of @xmath4 equals to @xmath4 , and therefore the compression rate of @xmath4 equals to @xmath4 , i.e. , @xmath10 , where @xmath11 is the first @xmath12 bits of the base - two expansion of @xmath4 . in our second work @xcite on this interpretation
, we showed that a fixed point theorem of the same form as for @xmath0 holds also for each of free energy @xmath1 , energy @xmath2 , and statistical mechanical entropy @xmath3 .
moreover , based on the statistical mechanical relation @xmath13 , we showed that the computability of @xmath1 gives completely different fixed points from the computability of @xmath0 . in this paper
, we develop the statistical mechanical interpretation of ait further and pursue its formal correspondence to normal statistical mechanics . as a result
, we unlock the properties of the sufficient conditions further .
the thermodynamic quantities in ait are defined based on the halting set of an optimal computer . in this paper
, we show in theorem [ main ] below that there are infinitely many optimal computers which give completely different sufficient conditions in each of the thermodynamic quantities in ait .
we do this by introducing the notion of composition of computers into ait , which corresponds to the notion of composition of systems in normal statistical mechanics .
we start with some notation about numbers and strings which will be used in this paper .
@xmath14 is the set of natural numbers , and @xmath15 is the set of positive integers .
@xmath16 is the set of rationals , and @xmath17 is the set of reals . let @xmath18 with @xmath19 .
we say that @xmath20 is _ increasing _
( resp . , _ decreasing _ ) if @xmath21 ( resp .
, @xmath22 ) for all @xmath23 with @xmath24 .
normally , @xmath25 denotes any function @xmath26 such that @xmath27 .
@xmath28 is the set of finite binary strings , where @xmath29 denotes the _
empty string_. for any @xmath30 , @xmath31 is the _ length _ of @xmath7 .
a subset @xmath32 of @xmath33 is called _ prefix - free _ if no string in @xmath32 is a prefix of another string in @xmath32 . for any partial function @xmath20
, the domain of definition of @xmath20 is denoted by @xmath34 .
let @xmath35 be an arbitrary real .
we denote by @xmath36 the first @xmath12 bits of the base - two expansion of @xmath37 with infinitely many zeros , where @xmath38 is the greatest integer less than or equal to @xmath35 .
for example , in the case of @xmath39 , @xmath40 .
we say that a real @xmath35 is _ computable _ if there exists a total recursive function @xmath41 such that @xmath42 for all @xmath43 .
see e.g. weihrauch @xcite for the detail of the treatment of the computability of reals . in the following we concisely review some definitions and results of algorithmic information theory @xcite .
computer _ is a partial recursive function @xmath44 such that @xmath45 is a nonempty prefix - free set . for each computer @xmath46 and each @xmath30 ,
@xmath47 is defined by @xmath48 ( may be @xmath49 ) .
a computer @xmath50 is said to be _ optimal _ if for each computer @xmath46 there exists @xmath51 , which depends on @xmath46 , with the following property ; for every @xmath52 there exists @xmath53 for which @xmath54 and @xmath55 .
it is easy to see that there exists an optimal computer .
we choose a particular optimal computer @xmath50 as the standard one for use , and define @xmath6 as @xmath56 , which is referred to as the _ program - size complexity _ of @xmath7 or the _ kolmogorov complexity _ of @xmath7 .
it follows that for every computer @xmath46 there exists @xmath51 such that , for every @xmath57 , @xmath58 . for any @xmath59
, we say that @xmath35 is _ weakly chaitin random _ if there exists @xmath60 such that @xmath61 for all @xmath43 @xcite .
on the other hand , for any @xmath59 , we say that @xmath35 is _ chaitin random _ if @xmath62 @xcite . obviously , for every @xmath59 , if @xmath35 is chaitin random , then @xmath35 is weakly chaitin random .
we can show that the converse also hold .
thus , for every @xmath59 , @xmath35 is weakly chaitin random if and only if @xmath35 is chaitin random ( see chaitin @xcite for the proof and historical detail ) . in the works @xcite ,
we generalized the notion of the randomness of a real so that _ the degree of the randomness _ , which is often referred to as _ the partial randomness _
recently @xcite , can be characterized by a real @xmath4 with @xmath63 as follows .
let @xmath64 with @xmath65 .
for any @xmath59 , we say that @xmath35 is _ weakly chaitin @xmath4-random _ if there exists @xmath60 such that @xmath66 for all @xmath43 .
let @xmath64 with @xmath65 .
for any @xmath59 , we say that @xmath35 is _
@xmath4-compressible _ if @xmath67 , which is equivalent to @xmath68 . in the case of @xmath69 , the weak chaitin
@xmath4-randomness results in the weak chaitin randomness . for every @xmath70 $ ] and
every @xmath59 , if @xmath35 is weakly chaitin @xmath4-random and @xmath4-compressible , then @xmath71 the left - hand side of is referred to as the _ compression rate _ of a real @xmath35 in general .
note , however , that does not necessarily imply that @xmath35 is weakly chaitin @xmath4-random .
thus , the notion of partial randomness is a stronger representation of the notion of compression rate .
let @xmath64 with @xmath65 .
for any @xmath72 , we say that @xmath35 is _
chaitin @xmath4-random _
if @xmath73 . in the case of @xmath69 , the chaitin @xmath4-randomness results in the chaitin randomness
obviously , for every @xmath70 $ ] and every @xmath59 , if @xmath35 is chaitin @xmath4-random , then @xmath35 is weakly chaitin @xmath4-random .
however , in 2005 reimann and stephan @xcite showed that , in the case of @xmath74 , the converse does not necessarily hold .
this contrasts with the equivalence between the weak chaitin randomness and the chaitin randomness , each of which corresponds to the case of @xmath69 .
in this section , we review some results of the statistical mechanical interpretation of ait , developed by our former works @xcite . we first introduce the notion of thermodynamic quantities into ait in the following manner . in statistical mechanics ,
the partition function @xmath75 , free energy @xmath76 , energy @xmath77 , and entropy @xmath78 at temperature @xmath4 are given as follows : @xmath79 where @xmath80 is a complete set of energy eigenstates of a quantum system and @xmath81 is the energy of an energy eigenstate @xmath82 .
the constant @xmath83 is called the boltzmann constant , and the @xmath84 denotes the natural logarithm .
let @xmath46 be an arbitrary computer .
we introduce the notion of thermodynamic quantities into ait by performing replacements [ cs06 ] below for the thermodynamic quantities in statistical mechanics .
[ cs06 ] 1 .
replace the complete set @xmath80 of energy eigenstates @xmath82 by the set @xmath45 of all programs @xmath85 for @xmath46 .
2 . replace the energy @xmath81 of an energy eigenstate @xmath82 by the length @xmath86 of a program @xmath85 .
3 . set the boltzmann constant @xmath83 to @xmath87 .
thus , motivated by the formulae and taking into account replacements [ cs06 ] , we introduce the notion of thermodynamic quantities into ait as follows . [ tdqait ] let @xmath46 be any computer , and let @xmath4 be any real with @xmath8 .
first consider the case where @xmath45 is an infinite set . in this case
, we choose a particular enumeration @xmath88 of the countably infinite set @xmath45 .
can be chosen quite arbitrarily , and the results of this paper are independent of the choice of @xmath89 .
this is because the sum @xmath90 and @xmath91 in definition [ tdqait ] are positive term series and converge as @xmath92 for every @xmath5 . ]
the _ partition function _
@xmath93 at temperature @xmath4 is defined as @xmath94 where @xmath95 2 . the _ free energy _
@xmath96 at temperature @xmath4 is defined as @xmath97 where @xmath98 3 .
the _ energy _
@xmath99 at temperature @xmath4 is defined as @xmath100 where @xmath101 4 .
the _ statistical mechanical entropy _
@xmath102 at temperature @xmath4 is defined as @xmath103 where @xmath104 in the case where @xmath45 is a nonempty finite set , the quantities @xmath93 , @xmath96 , @xmath99 , and @xmath102 are just defined as , , , and , respectively , where @xmath105 is an enumeration of the finite set @xmath45 .
note that , for every optimal computer @xmath106 , @xmath107 is precisely a chaitin @xmath108 number introduced by chaitin @xcite .
theorems [ cprpffe ] and [ cpreesh ] below hold for these thermodynamic quantities in ait . [ cprpffe ]
let @xmath106 be an optimal computer , and let @xmath64 . 1 .
if @xmath109 and @xmath4 is computable , then each of @xmath110 and @xmath111 converges and is weakly chaitin @xmath4-random and @xmath4-compressible .
2 . if @xmath112 , then @xmath110 and @xmath111 diverge to @xmath49 and @xmath113 , respectively .
[ cpreesh ] let @xmath106 be an optimal computer , and let @xmath64 . 1 .
if @xmath9 and @xmath4 is computable , then each of @xmath114 and @xmath115 converges and is chaitin @xmath4-random and @xmath4-compressible .
2 . if @xmath116 , then both @xmath114 and @xmath115 diverge to @xmath49 .
the above two theorems show that if @xmath4 is a computable real with @xmath5 then the temperature @xmath4 equals to the partial randomness ( and therefore the compression rate ) of the values of all the thermodynamic quantities in definition [ tdqait ] for an optimal computer .
these theorems also show that the values of all the thermodynamic quantities diverge when the temperature @xmath4 exceeds @xmath117 .
this phenomenon might be regarded as some sort of phase transition in statistical mechanics .
note here that the weak chaitin @xmath4-randomness in theorem [ cprpffe ] is replaced by the chaitin @xmath4-randomness in theorem [ cpreesh ] in exchange for the divergence at @xmath69 . in statistical mechanics or thermodynamics , among all thermodynamic quantities
one of the most typical thermodynamic quantities is temperature itself .
theorem [ fptpr ] below shows that the partial randomness of the temperature @xmath4 can equal to the temperature @xmath4 itself in the statistical mechanical interpretation of ait .
we denote by @xmath118 the set of all real @xmath5 such that @xmath4 is weakly chaitin @xmath4-random and @xmath4-compressible , and denote by @xmath119 the set of all real @xmath5 such that @xmath4 is chaitin @xmath4-random and @xmath4-compressible . obviously , @xmath120 .
each element @xmath4 of @xmath118 is a _
fixed point on partial randomness _
, i.e. , satisfies the property that the partial randomness of @xmath4 equals to @xmath4 itself , and therefore satisfies that @xmath10 .
let @xmath106 be a computer .
we define the sets @xmath121 by @xmath122 in the same manner , we define the sets @xmath123 , @xmath124 , and @xmath125 based on the computability of @xmath111 , @xmath114 , and @xmath115 , respectively .
we can then show the following .
[ fptpr ] let @xmath106 be an optimal computer
. then @xmath126 and @xmath127 .
theorem [ fptpr ] is just a fixed point theorem on partial randomness , where the computability of each of the values @xmath110 , @xmath111 , @xmath114 , and @xmath115 gives a sufficient condition for a real @xmath5 to be a fixed point on partial randomness .
thus , by theorem [ fptpr ] , the above observation that the temperature @xmath4 equals to the partial randomness of the values of the thermodynamic quantities in the statistical mechanical interpretation of ait is further confirmed .
in this paper , we investigate the properties of the sufficient conditions for @xmath4 to be a fixed point on partial randomness in theorem [ fptpr ] . using the monotonicity and continuity of the functions @xmath110 and @xmath111 on temperature @xmath4 and using the statistical mechanical relation @xmath128 , which holds from definition [ tdqait ]
, we can show the following theorem for the sufficient conditions in theorem [ fptpr ] .
[ sczf ] let @xmath106 be an optimal computer . then each of the sets @xmath121 and @xmath123 is dense in @xmath129 while @xmath130 .
thus , for every optimal computer @xmath106 , the computability of @xmath111 gives completely different fixed points from the computability of @xmath110 .
this implies also that @xmath131 and @xmath132 .
the aim of this paper is to investigate the structure of @xmath118 and @xmath119 in greater detail .
namely , we show in theorem [ main ] below that there are infinitely many optimal computers which give completely different sufficient conditions in each of the thermodynamic quantities in ait .
we say that an infinite sequence @xmath133 of computers is _ recursive _ if there exists a partial recursive function @xmath134 such that for each @xmath43 the following two hold : ( i ) @xmath135 if and only if @xmath136 , and ( ii ) @xmath137 for every @xmath135
. then the main result of this paper is given as follows .
[ main ] there exists a recursive infinite sequence @xmath133 of optimal computers which satisfies the following conditions : 1 .
@xmath138 for all @xmath139 with @xmath140 .
2 . @xmath141 and @xmath142 .
3 . @xmath143 and @xmath144 . in the subsequent sections we prove the above theorems by introducing the notion of _ composition _ of computers into ait , which corresponds to the notion of composition of systems in normal statistical mechanics .
+ let @xmath145 be computers .
the composition @xmath146 of @xmath147 , @xmath148 , @xmath149 , and @xmath150 is defined as the computer @xmath151 such that ( i ) @xmath152 , and ( ii ) @xmath153 for every @xmath154 , @xmath155 , @xmath149 , and @xmath156 . [ ocmpo ]
let @xmath145 be computers .
if @xmath147 is optimal then @xmath146 is also optimal .
we first choose particular strings @xmath157 , @xmath158 , @xmath149 , @xmath159 with @xmath160 , @xmath161 , @xmath149 , and @xmath162 .
let @xmath46 be an arbitrary computer .
then , by the definition of the optimality of @xmath147 , there exists @xmath51 with the following property ; for every @xmath52 there exists @xmath53 for which @xmath163 and @xmath55 .
it follows from the definition of the composition @xmath146 that for every @xmath52 there exists @xmath53 for which @xmath164 and @xmath165 .
thus @xmath146 is an optimal computer . in the same manner as in normal statistical mechanics
, we can prove theorem [ quc ] below for the thermodynamic quantities in ait .
in particular , the equations , , and correspond to the fact that free energy , energy , and entropy are extensive parameters in thermodynamics , respectively .
[ quc ] let @xmath145 be computers .
then the following hold for every @xmath5 .
@xmath166 for any computer @xmath46 and any @xmath43 , the computer @xmath167 is denoted by @xmath168 .
in order to prove the main result , theorem [ main ] , we also introduce the notion of _ physically reasonable computer_. for any computer @xmath46 , we say that @xmath46 is physically reasonable if there exist @xmath169 such that @xmath170
. then we can prove theorem [ inc - dec ] below in a similar manner to the proof of theorem 7 of @xcite .
[ inc - dec ] let @xmath46 be a physically reasonable computer . then each of the mapping @xmath171 , the mapping @xmath172 , and the mapping @xmath173 is an increasing real function . on the other hand , the mapping @xmath174 is a decreasing real function .
in order to prove the main result , it is also convenient to use the notion of _ computable measure machine _ , which was introduced by downey and griffiths @xcite in 2004 for the purpose of characterizing the notion of schnorr randomness of a real in terms of program - size complexity . a computer @xmath46 is called a computable measure machine if @xmath175 ( i.e. @xmath176 ) is computable . for the thermodynamic quantities in ait , we can prove theorem [ comp - prcmm ] below using theorem [ inc - dec ] .
[ comp - prcmm ] let @xmath46 be a physically reasonable , computable measure machine .
then , for every @xmath5 , the following conditions are equivalent : 1 .
@xmath4 is computable .
2 . one of @xmath93 , @xmath96 , @xmath99 , and @xmath102 is computable .
all of @xmath93 , @xmath96 , @xmath99 , and @xmath102 are computable .
[ ex - prc ] the following two computers are examples of physically reasonable , computable measure machines . * ( i ) two level system : * let @xmath177 be a particular computer for which @xmath178 .
then we see that , for every @xmath8 , @xmath179 * ( ii ) one dimensional harmonic oscillator : * let @xmath180 be a particular computer for which @xmath181 .
then we see that , for every @xmath8 , @xmath182 since @xmath183 and @xmath184 , we see that @xmath177 and @xmath180 are physically reasonable , computable measure machines .
note that theorems [ inc - dec ] and [ comp - prcmm ] certainly hold for each of the particular physically reasonable computers @xmath177 and @xmath180 .
based on theorems [ quc ] and [ comp - prcmm ] , the main result is proved as follows .
we first choose any optimal computer @xmath106 and any physically reasonable , computable measure machine @xmath46 .
then , for each @xmath43 , we denote the computer @xmath185 by @xmath186 . by theorem [ ocmpo ]
, we see that @xmath186 is optimal for every @xmath43 .
furthermore , it is easy to see that the infinite sequence @xmath133 of computers is recursive . on the other hand , it follows from theorem [ quc ] that , for every @xmath43 and every @xmath5 , @xmath187 let @xmath188 and @xmath12 be arbitrary two positive integers with @xmath189 .
then it follows from the equations that @xmath190 for every @xmath5 . in what follows , using we show that @xmath191 . in a similar manner , using , , and we can show that @xmath192 as well . now , let us assume contrarily that @xmath193
. then there exists @xmath194 such that both @xmath195 and @xmath196 are computable .
it follows from that @xmath197 thus , @xmath198 is also computable .
since @xmath46 is a physically reasonable , computable measure machine , it follows from theorem [ comp - prcmm ] that @xmath199 is also computable .
therefore , since @xmath200 is optimal , it follows from theorem [ cprpffe ] ( i ) that @xmath195 is weakly chaitin @xmath199-random . however , this contradicts the fact that @xmath195 is computable .
thus we have @xmath191 .
this completes the proof of theorem [ main ] ( i ) .
theorem [ main ] ( ii ) and theorem [ main ] ( iii ) follow immediately from theorem [ fptpr ] and the fact that @xmath201 is optimal for all @xmath202 .
this work was supported by kakenhi , grant - in - aid for scientific research ( c ) ( 20540134 ) , by scope of the ministry of internal affairs and communications of japan , and by crest of the japan science and technology agency .
k. tadaki , algorithmic information theory and fractal sets .
proceedings of 1999 workshop on information - based induction sciences ( ibis99 ) , pp
. 105110 , august 26 - 27 , 1999 , syuzenji , shizuoka , japan . in japanese .
k. tadaki , a statistical mechanical interpretation of algorithmic information theory .
local proceedings of computability in europe 2008 ( cie 2008 ) , pp . 425434 , june 15 - 20 , 2008 , university of athens , greece .
extended and electronic version available : http://arxiv.org/abs/0801.4194v1 k. tadaki , fixed point theorems on partial randomness .
proceedings of the symposium on logical foundations of computer science 2009 ( lfcs09 ) , lecture notes in computer science , springer - verlag , vol . 5407 , pp .
422440 , 2009 . | the statistical mechanical interpretation of algorithmic information theory ( ait , for short ) was introduced and developed by our former works [ k. tadaki , local proceedings of cie 2008 , pp . 425434 , 2008 ] and [ k. tadaki , proceedings of lfcs09 , springer s lncs , vol .
5407 , pp .
422440 , 2009 ] , where we introduced the notion of thermodynamic quantities , such as partition function @xmath0 , free energy @xmath1 , energy @xmath2 , and statistical mechanical entropy @xmath3 , into ait .
we then discovered that , in the interpretation , the temperature @xmath4 equals to the partial randomness of the values of all these thermodynamic quantities , where the notion of partial randomness is a stronger representation of the compression rate by means of program - size complexity .
furthermore , we showed that this situation holds for the temperature itself as a thermodynamic quantity , namely , for each of all the thermodynamic quantities above , the computability of its value at temperature @xmath4 gives a sufficient condition for @xmath5 to be a fixed point on partial randomness . in this paper , we develop the statistical mechanical interpretation of ait further and pursue its formal correspondence to normal statistical mechanics .
the thermodynamic quantities in ait are defined based on the halting set of an optimal computer , which is a universal decoding algorithm used to define the notion of program - size complexity .
we show that there are infinitely many optimal computers which give completely different sufficient conditions in each of the thermodynamic quantities in ait .
we do this by introducing the notion of composition of computers into ait , which corresponds to the notion of composition of systems in normal statistical mechanics . |
among the available methods to investigate the properties of strongly interacting many body quantum systems , quantum monte carlo ( qmc ) ones hold a relevant position since some of the qmc methods can provide `` exact '' expectation values . as for all the monte
carlo methods , qmc results are affected by statistical uncertainties , but the property that makes some of them `` exact '' is the possibility of reducing within this unavoidable errors all the systematic errors introduced by the involved approximations . this is true for boson system at zero and at finite temperature , but the studies of fermion systems , excited states or real time dynamics all suffer from sign problems which still have precluded the development of such kind of `` exact '' methods .
a first great subdivision of `` exact '' qmc methods is between finite and zero temperature methods . among @xmath1 k methods we find green function monte carlo ( gfmc),@xcite diffusion monte carlo ( dmc),@xcite path integral ground state monte carlo ( pigs)@xcite and reptation monte carlo ( rmc).@xcite at finite temperature , the major role is played by path
integral monte carlo ( pimc).@xcite the main difference between finite and zero temperature methods is that ground state methods rely , to different extents , on a trial wave function for the importance sampling , while in pimc no such wave function is involved at all .
pimc needs only information on the interaction among particles as input , and the main difficulty to overcome , beyond the propagator accuracy , is the sampling ergodicity .
recently a great step forward in this direction has been realized with the advent of the worm algorithm.@xcite the first gfmc computation did not use importance sampling,@xcite but this computation was for a small system of 32 particles .
all the other computation at zero temperature have used a model wave function that was given as input . with respect to pigs method ,
if , in principle , convergence can be achieved for any initial wave function that has a finite overlap with the ground state , the question is if convergence can be achieved in a real computation .
some progress have been achieved recently both for a finite@xcite and for a bulk boson system.@xcite for example , a jastrow wave function , that describes the bulk phase of a quantum liquid , was employed as initial wave function in the study of small parahydrogen clusters.@xcite in the case of a two dimensional crystal of @xmath0he , convergence to the same result was found in the computation of the one body density matrix starting from two initial wave functions with `` opposite '' properties,@xcite one with bose
einstein condensate and the other without .
stimulated by these results , we have undergone a systematic check for a realistic hamiltonian for @xmath0he and we have found that pigs methods converge to the exact result regardless of the chosen initial wave function .
even better , pigs converges even if the wave function contains wrong correlations or no correlations at all . as a test bed
we have considered the bulk liquid and solid phases of a strongly interacting boson system such as @xmath0he . by projecting very different wave functions
, we find converging results for diagonal properties like the energy , the radial distribution function , the static structure factor and also for off - diagonal properties like the one - body density matrix . for both the liquid and the solid phases we consider a shadow wave function ( swf),@xcite which is the best available variational wave function@xcite for @xmath0he systems , and the constant wave function ( cwf ) , which is the ground state wave function of the ideal bose gas and does not contain any correlation . the latter corresponds to have no importance sampling since the initial wave function gives the same weight to any configuration of the particles .
furthermore , for the liquid phase , we have considered also a localized wave function which describes an einstein solid .
our question is whether pigs is able to recover the exact ground state starting from really different systems , i.e. an ideal gas or an einstein crystal for the quantum liquid and an ideal gas for the quantum solid .
we find that for all the considered initial wave functions pigs converges to the exact ground state .
only the convergence rate depends on the initial wave function , and turns out to be slower for the worse ones .
the paper is organized as follows : sec .
[ sec : meth ] deals with the description of the pigs method , of the test bed systems and of the initial wave functions .
results are presented and discussed in sec .
[ sec : results ] .
[ sec : conc ] contains our conclusions .
it is well known that , given a hamiltonian @xmath2 for a quantum system of @xmath3 particles , the ground state wave function @xmath4 in the position representation can be obtained as the @xmath5 limit of an imaginary time ( @xmath6 ) evolution of an initial ( trial ) wave function @xmath7 , provided that @xmath8 : @xmath9 regardless of a normalization constant , which is not involved in the monte carlo sampling and thus will be dropped in the following , an accurate approximation of the ground state wave function is given by @xmath10 this originates from the action of the imaginary time projector @xmath11 , which exponentially removes from @xmath7 any overlap with the excited states during the imaginary time evolution .
the evolved wave function @xmath12 and eq .
provide the basis for all the zero temperature qmc methods .
a first problem rises : the imaginary time propagator @xmath13 can be accurately written only for small values of @xmath14 , but a large @xmath14 limit is necessary to ensure the convergence to @xmath4 . one possible strategy to overcome this problem is to reach the large @xmath14 limit by means of a recursive procedure , as for example in gfmc@xcite and dmc.@xcite both these methods reach an extremely accurate approximation of the ground state wave function multiplied by @xmath7 by iterating equation by means of random walks .
with these methods the trial wave function has a strategical role since it is involved at each iteration step , where it is used as importance sampling function . on the contrary , the pigs method@xcite interprets the imaginary time propagator as a density matrix operator corresponding to an inverse temperature @xmath15 , and
then , by exploiting the factorization property @xmath16 the large @xmath14 propagator @xmath13 is written as a convolution of small imaginary time propagators @xmath17 as in standard path integral formalism.@xcite thus the density matrix operator is broken up into @xmath18 small pieces with a time step @xmath19 and the approximated ground state wave function reads @xmath20 an appealing feature , peculiar of this method is that , in @xmath12 , the ansatz on @xmath7 acts only at the starting point being the full imaginary time path governed by @xmath13 , which depends only on @xmath2 .
once fixed @xmath21 , the elementary evolution step in imaginary time is obtained acting with @xmath17 on a quantum state ; this action is usually called projection step due to the resulting increased overlap with the ground state . the pigs method reaches convergence to the ground state when adding further imaginary time projections to @xmath12 does not provide any appreciable change of the results , i.e. the expectation values computed with @xmath22 is compatible with those computed with @xmath12 within the statistical errors for any integer @xmath23 . since the convergence on the ground state for a finite system is exponentially fast,@xcite the number @xmath18 of required imaginary time projections is usually limited .
nevertheless @xmath18 depends on the specific expectation value one is computing , thus a separate analysis of the convergence should be carried on for every computed quantity .
the convergence rate is determined essentially by the quality of the initial wave function .
the total imaginary time @xmath14 required to clean the initial wave function up from excited state contributions is smaller for good ( i.e. with large overlap with the ground state ) initial wave functions .
also the accuracy of the approximation used for the small time @xmath13 has a role in determining the computational cost of the projection procedure : in fact , with better approximations of the small imaginary time @xmath13 , one needs a smaller number of projection steps to reach convergence since larger time step @xmath21 are allowed .
moreover the imaginary time projection procedure should be able , in principle , to incorporate in the final wave function all the correct correlations , even if these were absent in @xmath7 .
thus it should be possible , in principle , to obtain the correct ground state wave function , if @xmath14 is large enough , even if @xmath7 is extremely poor , and the results do not suffer from any variational bias . the intent of the present study is to show that this is indeed the case with values of @xmath14 manageable with present computational resources . like for finite temperature pimc,@xcite also within pigs
the quantum system is mapped into a system of classical polymers , but in this case the polymers start and end on the initial wave function @xmath7,@xcite resulting in open linear polymers instead of the close ring ones of pimc.@xcite each open polymer represents the full imaginary time path of a quantum particle that is sampled by means of the metropolis algorithm .
thus , the entire imaginary time evolution of the system is sampled at each monte carlo step , contrary to what happens for dmc , for example .
expectation values of an operator @xmath24 reads @xmath25 and , if @xmath14 is large enough to allow convergence to the ground state , @xmath26 is the ground state expectation value , without needing any extrapolation .
however an analysis of the convergence as a function of @xmath14 is needed , as explained before .
the ground state expectation value of the energy can be obtained in several ways : the most largely employed one is the mixed estimate @xmath27 notice that in pigs this mixed estimate is exact . in fact , since the hamiltonian operator @xmath2 commutes with the imaginary time evolution operator @xmath13 , it is possible to obtain unbiased expectation value of the hamiltonian inserting @xmath2 at one of the ends of the path . if @xmath7 is an accurate wave function , this estimator is preferable with respect to the direct one given by eq . with @xmath28 , because it has typically a lower variance .
this is no more true , however , when the initial wave function is particularly poor . in this case
the fluctuations in the expectation value of the hamiltonian become so sizable that is unusable .
the direct estimator is seldom used because one has to compute the first and the second derivative of @xmath17 ; but this is not a serious problem if the imaginary time propagator is available in an analytical formulation , like the one we employed here .
when @xmath7 is a good wave function , as for the swf case , we have verified that mixed and direct estimators provide compatible results , and we report here the mixed one for the above mentioned reason . on the contrary , for poor wave functions , like cwf , we had to consider the direct estimator .
as far as @xmath0he systems are concerned , because of the bose statistic obeyed by the atoms , one has , in principle , to account for permutations in the propagator @xmath13.@xcite permutation moves are not strictly requested whenever the initial wave function has the correct bose symmetry .
in fact , the polymer configuration resulting after a permutation can be in principle reached with a combination of standard sampling moves , since the polymers are open.@xcite this is not the case if @xmath7 is not bose symmetric like the gaussian wave function : permutation cycles among particles must be introduced in the sampling in order to get the exact ground state . for all @xmath7
we have implemented permutation sampling following ref . .
in fact , implementing the sampling of permutations even for a bose symmetric @xmath7 turns out to greatly improve the ergodicity of the sampling .
in addition , we have used swap moves@xcite because they increase the sampling efficiency when computing off
diagonal properties.@xcite in order to test the pigs method convergence properties we have considered two bulk phases of a many body strongly interacting boson system : liquid and solid @xmath0he . dealing with low temperature properties , @xmath0he atoms are described as structureless zero spin bosons , interacting through a realistic two body potential , that we assume to be the hfdhe2 aziz potential @xcite . for the liquid phase ,
we have considered a cubic box with periodic boundary conditions , containing @xmath29 atoms at the equilibrium density @xmath30@xmath31 . for the solid phase we have considered a cubic box with periodic boundary conditions designed to house a fcc crystal of @xmath32 atoms at the density @xmath33@xmath31 . in both cases
we add standard tail corrections to the potential energy to account for the finite size of the system by assuming the medium homogeneous ( i.e. @xmath34 ) beyond @xmath35 , where @xmath36 is the size of the box .
obviously , this is not an accurate assumption specially for the solid phase in such a small box , but our main purpose here is to show that pigs method is able to reach the same results independently on the considered initial wave function .
moreover we have studied the fcc lattice , which is stabilized by the cell geometry and by the periodic boundary conditions , whereas at zero temperature , the hcp lattice is the stable lattice .
computations of ground state properties of bulk @xmath0he with accurate tail corrections can be found in the current literature.@xcite the standard initial wave functions commonly used within pigs method@xcite are the variational jastrow wave function ( jwf ) for the liquid and the jastrow - nosanow ( j - nwf ) for the solid .
a jwf represents the simplest possible choice of wave function for strongly interacting bosons@xcite and it contains only two body correlations . using a mcmillan pseudopotential,@xcite the unnormalized jwf reads as @xmath37 the physical meaning of this jwf is that , due to the sharp repulsive part of the interaction potential @xmath38 in the hamiltonian @xmath2 , @xmath0he atoms prefer to avoid each other .
in the j - nwf the jwf is multiplied by a term like the one in eq . below , that localizes the particles in a crystalline order . in this work , however , in order to explore the convergence properties of the pigs method , we have considered two wave functions of `` opposite '' quality : the best available one , that is the shadow wave function , and the poorest imaginable one , i.e. the constant wave function .
jwf will be considered only when computing the one body density matrix in the liquid phase .
the constant wave function is the ground state wave function of the ideal bose gas , @xmath39 it carries no correlation at all .
we choose this wave function because , allowing an unrestricted sampling of the full configurational space , it results in no importance sampling . then the whole imaginary time projection procedure is driven only by the short time evolution operator @xmath13 , without any input , and then any bias , from the initial state .
thus at the starting point the system is made up by free particles ; if after a long enough imaginary time projection , pigs turns out to be able to reach a strong correlated quantum liquid and quantum crystal by itself we can safely believe that no variational bias affects pigs results . on the other hand , we choose as @xmath7 a swf given by variational computation in order to have as reference results the ones coming from the projection of an initial wave function that is more accurate as possible , i.e. from a wave function whose overlap with the exact ground state is known to be large . in the swf , additional correlations
besides the standard two body terms are introduced via auxiliary variables which are integrated out.@xcite this is done so efficiently that the crystalline phase emerges as a spontaneously broken symmetry process , induced by the inter particles correlations as the density is increased , without the need of any a priori knowledge of the equilibrium positions and without losing the translationally invariant form of the wave function .
thus swf is able to describe both the liquid and the solid phase with the same functional form and it is explicitly bose symmetric .
the standard swf functional form reads @xmath40 where @xmath41 is the set of auxiliary shadow variables , @xmath42 is the standard jastrow two body correlation term , @xmath43 is a kernel coupling each shadow to the corresponding real variable , and @xmath44 is another jastrow term describing the inter shadow correlations . as usual,@xcite
we take @xmath43 gaussian and in @xmath45 we use the rescaled and dilated he he potential @xmath38 as pseudopotential .
the variational parameters we use were chosen in order to minimize the expectation value of the hamiltonian @xmath2 and are reported in ref . .
nowadays the swf represents the best available variational wave function for @xmath0he systems.@xcite recently , we have estimated@xcite that , when describing a two dimensional solid , swf overlap per particle with the true ground state is of about 99.8% , which ensures a fast convergence rate when projected within the pigs method .
the properties of the swf are so peculiar that the pigs method that has a swf as @xmath7 deserves an its own name and is dubbed spigs : shadow path integral ground state method.@xcite in the picture of linear polymers , the presence of the shadow variables adds two extra variational links , one at each end of the polymer . in order to test how robust pigs is , we consider also a `` wrong '' wave function : for the liquid phase we consider a gaussian wave function , where each particle is harmonically localized around fixed positions @xmath46 @xmath47 i.e. @xmath7 it the wave function of an einstein harmonic solid .
the parameter @xmath48
@xmath49 is arbitrary and it is was chosen to ensure a strong localization of the particles around the positions @xmath46 that were taken over a regular cubic lattice within the simulation box .
this wave function is evidently not translationally invariant and not bose symmetric .
furthermore it does not contain any correlation between the particles , and all the information that it carries is that of a crystalline system , i.e. gwf is an extremely poor wave function for the liquid phase .
this wrong initial wave function will provide a stringent test on the convergence properties of the pigs methods . as far as the one body density matrix computation in the liquid phase is concerned , the values of the parameters @xmath50 and @xmath51 in the jwf have been chosen equal to the ones of the corresponding jastrow term in the swf .
one of the fundamental elements of path integral projection monte carlo methods is the imaginary time propagator @xmath13 , whose accuracy turns out to be crucial to the convergence to the exact results .
the functional form of @xmath13 for a generic @xmath14 is unfortunately not known with exception of few particular cases , such as , for example , the free particle and the harmonic oscillator , but accurate approximations of @xmath13 are obtainable in the small @xmath14 regime.@xcite in this work , we have chosen the pair
suzuki approximation@xcite for the imaginary time propagator , which is a pair approximation of the fourth order suzuki chin density matrix.@xcite the suzuki chin approximation is based on the following factorization of the density matrix : @xmath52 where @xmath53 is the kinetic operator and @xmath54 and @xmath55 are given by @xmath56 and @xmath57 respectively , with @xmath58 the potential operator , @xmath59 an arbitrary constant in the range @xmath60 $ ] , @xmath61 and @xmath62 .
the resulting imaginary time propagator is accurate to order @xmath63 , and has been successfully applied to liquid @xmath0he in two and three dimensions.@xcite this approximation offers also the advantage that adjusting the parameter @xmath59 it is possible to optimize the convergence , and a standard choice for a quantum system is @xmath64.@xcite a strategy to obtain a simpler , but equally accurate , approximation consists in applying a pair product assumption.@xcite for sufficiently short time steps , in fact , the many body propagator ( in imaginary time ) is well approximated by the product of two
body propagators.@xcite in this approximation , the small time propagator reads @xmath65 where @xmath66 is given as @xmath67 & m\quad { \rm odd}\\ \\ \frac{\delta\tau}{3}\left[2v_c(r_{m})+v_e(r_{m+1})\right ] & m\quad { \rm even . }
\end{array}\right.\ ] ] the potentials @xmath68 and @xmath69 are defined as @xmath70 where @xmath71 is the potential experienced by two @xmath0he atoms at a distance @xmath72 .
the advantage is that there is no need to calculate @xmath73 .
as for the full suzuki
chin approximation,@xcite also for the pair suzuki the operators corresponding to physical observables must be inserted only on odd time slices in the imaginary time path .
energy per @xmath0he atom @xmath74 vs. imaginary time step @xmath21 .
the total projection time is @xmath75 k@xmath76 .
the calculations were carried out by projecting a swf and a cwf for a system of 64 particles at the equilibrium density @xmath77 @xmath31 .
dashed lines are quartic fits to the data .
error bars , when not shown , are smaller than the used symbols.,width=302 ] in order to fix the optimal small imaginary time step value , we have performed pigs simulations with different initial wave functions . by considering decreasing @xmath21 values with a fixed total projection time , @xmath14 , we have taken the energy per particle @xmath74 as observable of reference . as an example , our results for swf and cwf in the liquid phase are plotted in fig .
[ f : dtau ] .
we choose as optimal value @xmath78 k@xmath76 ; in fact , further reductions do not change the energy in a detectable way , i.e. within the statistical uncertainty . in fig .
[ f : dtau ] swf and cwf do not converge to the same value simply because the considered total projection time @xmath14 in this test is not enough to ensure convergence of @xmath74 to the ground state energy for cwf ( see fig .
[ f : enel ] ) .
similarly , in the solid phase we take @xmath79 k@xmath76 .
once set the optimal @xmath21 value , we have computed the diagonal properties of the system for increasing total projection time @xmath14 until we reached convergence to a value that corresponds to the exact ground state result both for the liquid and for the solid phase . in the liquid phase
we have computed also the one body density matrix . for the liquid phase we have projected a swf and a cwf .
the energy per particle as a function of the total projection time @xmath14 for both the wave functions is plotted in fig .
[ f : enel ] .
energy per particle @xmath80 as a function of the total projection time @xmath14 obtained from pigs simulations for liquid @xmath0he at the equilibrium density @xmath77 @xmath31 by projecting a swf ( filled circles ) and a cwf ( open circles ) and a gwf ( open diamonds ) .
@xmath81 result ( filled circle ) corresponds to the swf variational estimate of @xmath80 , the @xmath81 for the gwf is @xmath82 k and for cwf @xmath80 is essentially infinite .
error bars are smaller than the used symbols .
dotted line indicates the convergence value @xmath83 k.,width=302 ] we find that the energy converges , independently from the considered initial wave function , to the same value @xmath83 k. this value , in spite of the small size of the considered system , is close to the experimental@xcite result @xmath84 k. swf converges very quickly , in fact @xmath85 k@xmath76 is already enough to ensure convergence .
cwf instead , requires a three times larger imaginary time , i.e. @xmath86 k@xmath76 .
this was somehow expected , since swf is presently the best available variational wave function for @xmath0he.@xcite nevertheless , the quick convergence of also cwf is a really remarkable result .
in fact , this means that pigs efficiently includes the exact interparticle correlations through the imaginary time projections , without any need of importance sampling .
then , the choice of a good wave function , within the pigs method , becomes a matter of convenience rather than of principle , since better initial wave functions only allow for a smaller total projection time @xmath14 , and thus less cpu consuming simulations . radial distribution function @xmath87 for bulk liquid @xmath0he computed in a cubic box with @xmath29 at the density @xmath77 @xmath31 with the pigs method . a ) @xmath87 obtained by projecting a swf for @xmath88 , @xmath89 and @xmath90 k@xmath76 .
the @xmath88 result corresponds to the variational swf estimate of @xmath87 .
b ) @xmath87 obtained by projecting a swf for @xmath91 k@xmath76 and a cwf for @xmath92 k@xmath76 . in the inset
a zoom of the first maximum region .
c ) @xmath93 at different @xmath14 values , where @xmath94 is the @xmath87 computed by projecting a swf for an imaginary time equal to @xmath14 , and @xmath95 is the same but by projecting a cwf .
note the smaller scale on the vertical axis , width=309 ] static structure factor @xmath96 for bulk liquid @xmath0he computed in a cubic box with @xmath29 at the density @xmath77 @xmath31 with the pigs method .
a ) @xmath96 obtained by projecting a swf and a cwf for @xmath85 k@xmath76 .
b ) @xmath96 obtained by projecting a swf and a cwf for @xmath97 k@xmath76 .
c ) @xmath98 at different @xmath14 values , where @xmath99 is the @xmath96 computed by projecting a swf for an imaginary time equal to @xmath14 , and @xmath100 is the same but by projecting a cwf .
note the smaller scale on the vertical axis.,width=309 ] this convergence is confirmed also by the radial distribution function @xmath87 and the static structure factor @xmath96 . for such quantities ,
the convergence rate is found to be similar to the energy one . in fig .
[ f : grl ] we report the radial distribution function @xmath87 obtained by projecting both a swf and a cwf at different imaginary time values . for @xmath101 k@xmath76 ,
swf results at different @xmath14 are indistinguishable within the statistical uncertainty ( see fig . [
f : grl]a ) . in fact , with swf the exact result is reached within very few projection steps and then it is no more affected by further projections . as already pointed out , also cwf displays a fast convergence , as shown in fig . [
f : grl]c , where @xmath93 is shown . for increasing @xmath14
, @xmath102 evolves toward a flat function , meaning that the systems described starting from the two different wave functions , i.e the strongly correlated quantum liquid of swf and the ideal gas of cwf , are evolving into the same quantum liquid , which is the best reachable representation of the exact ground state of the simulated system .
the same conclusion is inferred from the evolution of the static structure factor @xmath96 , which is plotted in fig .
[ f : skl ] . in order to put a more stringent check on the pigs method ability to converge to the exact ground state without any variational bias , we have considered also a wrong initial wave function by projecting a gwf .
thus at the starting point of the imaginary time path there is now a strongly localized einstein crystal .
we find , even in this case , that the energy converges to the same value as before ( see fig . [
f : enel ] ) .
thus pigs is able not only to drop from the initial wave function the wrong information of localization , but also to generate at the same time the correct correlations among the particles .
gwf needs @xmath103
k@xmath76 to converge , which is ten times larger than the swf value .
radial distribution function @xmath87 for bulk liquid @xmath0he computed in a cubic box with @xmath29 at the density @xmath77 @xmath31 with the pigs method .
a ) @xmath87 obtained by projecting a swf for @xmath97 k@xmath76 and a gwf for @xmath104 k@xmath76 . in the inset a zoom of the first maximum region .
b ) @xmath105 at different @xmath14 values , where @xmath94 is the @xmath87 computed by projecting a swf for an imaginary time equal to @xmath14 , and @xmath106 is the same but by projecting a gwf .
note the smaller scale on the vertical axis.,width=309 ] static structure factor @xmath96 for bulk liquid @xmath0he computed in a cubic box with @xmath29 at the density @xmath77 @xmath31 with the pigs method .
a ) @xmath96 obtained by projecting a swf and a gwf for @xmath85 k@xmath76 .
it is evident in the gwf result the presence of the bragg peak .
note the logarithmic scale .
b ) @xmath96 obtained by projecting a swf for @xmath97 k@xmath76 and a gwf for @xmath107 k@xmath76 .
the bragg peak is no more present in the gwf result .
c ) @xmath108 at different @xmath14 values , where @xmath99 is the @xmath96 computed by projecting a swf for an imaginary time equal to @xmath14 , and @xmath109 is the same but by projecting a gwf .
note the change of the vertical scale .
error bars are smaller than the used symbols.,width=309 ] again this convergence is confirmed also by the radial distribution function @xmath87 and the static structure factor @xmath96 . in fig .
[ f : grl2 ] we report the radial distribution function @xmath87 obtained by projecting a gwf at different imaginary time values compared with the ones coming from the projection of swf . it is evident that small imaginary time is not enough to leave out the wrong information in the gwf . for lower @xmath14 values
, there are still reminiscences of the starting harmonic solid , which are progressively lost as the projection time increases .
this is made clearer in fig .
[ f : grl2]b where we plot the difference @xmath110 , at fixed imaginary time @xmath14 , between the @xmath87 computed by projecting the swf and the one obtained by projecting the gwf .
a similar behavior is observed in the evolution static structure factor @xmath96 , plotted in fig . [
f : skl2 ] .
for the gwf , the bragg peak shown at small @xmath14 values ( fig .
[ f : skl2]a ) , which is typical of the solid phase , becomes lower and lower as the projection time is increased ( fig .
[ f : skl2]b ) , until convergence is reached ( see fig .
[ f : skl2]c ) . from the plot of the energy per particle vs. the total imaginary time @xmath14 it is possible to estimate the overlap per particle of the initial wave function on the exact ground state.@xcite by using the results in fig .
[ f : enel ] we find that the overlap of swf is about 99% , while the gwf one is about 10% .
that swf has an high overlap with the ground state is not a surprise ; it was qualitatively expected since swf is presently the best available wave function for @xmath0he.@xcite however a 99% overlap is really remarkable and provides a further argument on the goodness of swf . on the other hand , a poor overlap of gwf
was somehow expected , since the parameter @xmath111 was chosen to strongly localize the atoms of the bulk liquid around fictitious equilibrium positions on a regular lattice .
one body density matrix @xmath112 obtained from pigs simulations for liquid @xmath0he at the equilibrium density @xmath77 @xmath31 by projecting a swf , a jwf and a gwf for an imaginary time @xmath113 , @xmath114 and @xmath115 k@xmath76 respectively .
the dotted line indicates the condensate value @xmath116 obtained from an independent pigs simulation.@xcite , width=302 ] besides the diagonal ones , also off diagonal properties , such as the one body density matrix , are accessible within pigs simulations .
the one - body density matrix @xmath117 represents the probability amplitude of destroying a particle in @xmath118 and creating one in @xmath119 .
its fourier transformation represents the momentum distribution . in first quantization @xmath112
is given by the overlap between the normalized many - body ground state wave functions @xmath120 and @xmath121 , where the configuration @xmath122 differs from @xmath123 only by the position of one of the @xmath3 atoms in the system .
if @xmath120 is translationally invariant , @xmath112 only depends on the difference @xmath124 , thus @xmath125 the bose - einstein condensate fraction @xmath126 is equal to the large distance limit of @xmath127 . in fact , if @xmath112 has a nonzero plateau at large distance , the so called off - diagonal long - range order ( odlro ) , its ft contains a dirac delta function , which indicates a macroscopic occupation of a single momentum state , i.e. bose einstein condensation .
the exact @xmath112 can be obtained in pigs simulation by substituting @xmath4 in with @xmath12 with @xmath14 large enough .
this corresponds to the simulation of a system of @xmath128 linear polymers plus a polymer which is cut into two halfs , called half
polymers , one departing from @xmath118 and the other from @xmath119 .
thus @xmath112 is obtained by collecting the relative distances among the cut ends of the two half
polymers during the monte carlo sampling .
the present computation of @xmath112 has been obtained by implementing a zero temperature version of the worm algorithm.@xcite we have worked with a fixed number of particles and not in the grand canonical ensemble , similarly to what has been done at finite temperature in ref . .
in practice this corresponds to a usual pigs calculation of @xmath112 where `` open '' and `` close '' moves have been implemented@xcite in order to visit diagonal and off - diagonal sectors within the same simulation .
the advantage of doing this does not come from the efficiency of the worm algorithm to explore off - diagonal configurations , because similar efficiency is obtained with pigs when `` swap '' moves are implemented.@xcite the benefit in using a worm - like algorithm here instead comes from the automatic normalization of @xmath112 which is a peculiarity of this method.@xcite in fig .
[ f : odlro ] we report @xmath112 obtained in pigs simulations of bulk liquid @xmath0he at @xmath77 @xmath31 by projecting either a swf , a jwf and a gwf .
all the simulations give the same result , shown in fig .
[ f : odlro ] which turns out to be compatible with the recent estimate obtained with pigs given in ref . of @xmath129 .
energy per particle @xmath80 as a function of the total projection time @xmath14 obtained from pigs simulations of an fcc @xmath0he crystal at the density @xmath130 @xmath31 by projecting a swf ( filled circles ) and a cwf ( open circles ) .
dashed line indicates the convergence value @xmath131 k.,width=302 ] radial distribution function @xmath87 for bulk solid @xmath0he computed in a cubic box with @xmath32 at the density @xmath130 @xmath31 with the pigs method .
a ) @xmath87 obtained by projecting a swf for @xmath85 and @xmath115 k@xmath76 .
b ) @xmath87 obtained by projecting a swf and a cwf for @xmath132 k@xmath76 . in the inset a zoom of the first maximum region .
c ) @xmath93 at different @xmath14 values , where @xmath94 is the @xmath87 computed by projecting a swf for an imaginary time equal to @xmath14 , and @xmath95 is the same but by projecting a cwf.,width=309 ] static structure factor @xmath96 for bulk solid @xmath0he computed in a cubic box with @xmath32 at the density @xmath130 @xmath31 with the pigs method .
a ) @xmath96 obtained by projecting a swf and a cwf for @xmath85 k@xmath76 .
b ) @xmath96 obtained by projecting a swf and a cwf for @xmath132 k@xmath76 .
the black dots are under the red ones .
c ) @xmath98 at different @xmath14 values , where @xmath99 is the @xmath96 computed by projecting a swf for an imaginary time equal to @xmath14 , and @xmath100 is the same but by projecting a cwf .
error bars are smaller than the used symbols .
notice the logarithmic scale in panels a ) and b).,width=309 ] we have performed the computation at density @xmath130 @xmath31 , where @xmath0he is in the solid phase , by projecting a swf and a cwf .
our results for the energy per particle are plotted in fig .
[ f : enes ] as a function of @xmath14 . in both cases
we find convergence to the value @xmath131
k. even in this phase the convergence of swf is faster , being @xmath85 k@xmath76 enough to reach convergence . in the case of cwf convergence is reached only for a much larger imaginary time @xmath132 k@xmath76 . also in this case convergence is obtained for the radial distribution function and for the static structure factor , reported in fig .
[ f : grs ] and fig .
[ f : sks ] respectively . from fig.[f
: grs]a it is evident that swf has reached the true ground state with few projection steps , since the results for @xmath87 at @xmath85 k@xmath76 and @xmath132 k@xmath76 are indistinguishable .
the evolution toward the correct ground state of the projected cwf is instead detectable .
the presence of the crystalline structure is mainly evident in the static structure factor , where a bragg peak grows with increasing @xmath14 ( see fig . [
f : sks]a , b ) .
the emerging of the correct solid structure by projecting a really poor wave function such as the cwf is made evident by the trend toward a flat function of the differences @xmath110 and @xmath133 plotted in fig.[f : grs]c and fig.[f : sks]c respectively .
in this work we have studied with the path integral ground state method diagonal and off - diagonal properties of a strongly interacting quantum bose system like the bulk liquid and solid phases of @xmath0he .
we have obtained convergence to the ground state values of quantities like the total energy , the radial distribution function , the static structure factor and the one - body density matrix projecting radically different wave functions : equivalent expectation values in the liquid phase have been obtained using as initial wave function a shadow wave function , a gaussian wave function with strongly localized particles of an einstein solid without interparticle correlations and also a constant wave function where all configurations of the particles are equally probable .
similarly in the solid phase equivalent expectation values have been obtained by considering a shadow wave function , which describes a solid , and a constant wave function which describes an ideal bose gas .
the present analysis demonstrates the absence of any variational bias in pigs ; a method that can be thus considered as unbiased as the finite temperature pimc .
this remarkable property comes from the accurate imaginary time propagators , exactly the same used with pimc , that do not depend on the initial trial state .
it remains true that the use of a good variational initial wave function greatly improves the rate of convergence to the exact results .
moreover , very poor wave functions have also the drawback of requiring the direct estimator for the hamiltonian ( eq .
( [ aspett ] ) with @xmath134 ) implying the necessity of an analytical formulation for the small imaginary time propagator @xmath13 or an accurate knowledge of its derivatives .
we have addressed here only the case of a realistic interaction potential among helium atoms .
however one can reasonably expect that this conclusion holds even for very different kinds of interaction , once an accurate approximation for the imaginary time propagator is known ( for example hard - spheres@xcite or hydrogen plasma@xcite ) .
as far as pathological potentials like the attractive coulomb one are concerned , pigs would suffer the same limitations of pimc if inaccurate approximations of the propagator were used.@xcite
authors acknowledge s. pilati for useful discussions . this work was supported by the infm parallel computing initiative , by the supercomputing facilities of cilea and by the mathematics department `` f. enriques '' of the universit degli studi di milano .
kalos , _ phys
. rev . _ a * 2 * , 250 ( 1970 ) .
kalos , d. levesque and l. verlet , _ phys .
rev . _ a * 9 * , 2178 ( 1974 ) .
whitlock , d.m .
ceperley , g.v .
chester and m.h .
kalos , _ phys .
_ b * 19 * , 5598 ( 1979 ) .
reynolds , d.m .
ceperley , b.j .
alder and w.a .
lester , _ j. chem .
* 77 * , 5593 ( 1982 ) .
a. sarsa , k.e .
schmidt and w.r .
magro , _ j. chem
. phys . _ * 113 * , 1366 ( 2000 ) .
s. baroni and s. moroni , _ phys .
lett . _ * 82 * , 4745 ( 1999 ) .
ceperley , _ rev .
phys . _ * 67 * , 279 ( 1995 ) .
m. boninsegni , n.v . prokofev and b.v .
svistunov , _ phys .
lett . _ * 96 * , 070601 ( 2006 ) ; _ phys .
_ e * 74 * , 036701 ( 2006 ) .
cuervo and p .- n .
roy , _ j. chem .
phys . _ * 125 * , 124314 ( 2006 ) .
e. vitali , m. rossi , f. tramonto , d.e .
galli and l. reatto , _ phys .
_ b * 77 * , 180505(r ) ( 2008 ) .
vitiello , k. runge and m.h .
kalos , _ phys
lett . _ * 60 * , 1970 ( 1988 ) .
s. moroni , d.e .
galli , s. fantoni and l. reatto , _ phys .
_ b * 58 * , 909 ( 1998 ) .
cuervo , p .-
roy and m. boninsegni , _ j. chem
. phys . _ * 122 * , 114504 ( 2005 ) .
d.e . galli and l. reatto , _ mol .
phys . _ * 101 * , 1697 ( 2003 ) ; _ j. low temp
galli and l. reatto , _ j. low temp .
phys . _ * 136 * , 343 ( 2004 ) . m. boninsegni , _ j. low temp . phys .
_ * 141 * , 27 ( 2005 ) .
aziz , v.p.s .
nain , j.s .
carley , w.l .
taylor and g.t .
mcconville , _
phys . _ * 70 * , 4330 ( 1979 ) .
see for example j. boronat and j. casulleras , _ phys .
_ b * 49 * , 8920 ( 1994 ) .
see for instance , e. feenberg , _ theory of quantum fluids _
( academic , new york , 1969 ) .
mcmillan , _ phys .
rev . _ * 138 * , a442 ( 1965 ) .
t. macfarland , s.a .
vitiello , l. reatto and g.v .
chester , _ phys .
_ b * 50 * , 13577 ( 1994 ) .
k. sakkos , j. casulleras and j. boronat , arxiv:0903.2763 .
s. pilati , _ phd thesis _ trento university , italy ( 2008 ) .
p.t . roach , j.b .
ketterson and c.w .
woo , _ phys .
rev . _ a * 2 * , 543 ( 1970 ) . c. mora and x. waintal , _ phys .
_ * 99 * , 030403 ( 2007 ) .
s. moroni and m. boninsegni , _ j. low temp .
* 136 * , 129 ( 2004 ) .
j. cao and b.j .
berne , _ j. chem
. phys . _ * 97 * , 2382 ( 1992 ) ; l.m .
ses and r. ledesma , _ j. chem .
phys . _ * 102 * , 3776 ( 1995 ) ; p. grter , d.m . ceperley and f. lalo ,
_ phys . rev .
_ * 79 * , 3549 ( 1997 ) . c. pierleoni , d.m .
ceperley , b. bernu and w.r .
magro , _ phys .
* 73 * , 2145 ( 1994 ) .
mser and b.j .
* 107 * , 571 ( 1997 ) . | generally `` exact '' quantum monte carlo computations for the ground state of many bosons make use of importance sampling .
the importance sampling is based , either on a guiding function or on an initial variational wave function . here
we investigate the need of importance sampling in the case of path integral ground state ( pigs ) monte carlo .
pigs is based on a discrete imaginary time evolution of an initial wave function with a non zero overlap with the ground state , that gives rise to a discrete path which is sampled via a metropolis like algorithm . in principle
the exact ground state is reached in the limit of an infinite imaginary time evolution , but actual computations are based on finite time evolutions and the question is whether such computations give unbiased exact results .
we have studied bulk liquid and solid @xmath0he with pigs by considering as initial wave function a constant , i.e. the ground state of an ideal bose gas .
this implies that the evolution toward the ground state is driven only by the imaginary time propagator , i.e. there is no importance sampling .
for both the phases we obtain results converging to those obtained by considering the best available variational wave function ( the shadow wave function ) as initial wave function .
moreover we obtain the same results even by considering wave functions with the wrong correlations , for instance a wave function of a strongly localized einstein crystal for the liquid phase .
this convergence is true not only for diagonal properties such as the energy , the radial distribution function and the static structure factor , but also for off diagonal ones , such as the one body density matrix .
this robustness of pigs can be traced back to the fact that the chosen initial wave function acts only at the beginning of the path without affecting the imaginary time propagator . from this analysis
we conclude that zero temperature pigs calculations can be as unbiased as those of finite temperature path integral monte carlo . on the other hand , a judicious choice of the initial wave function
greatly improves the rate of convergence to the exact results . |
one of the problems of algebraic number theory is to get deep knowledge of the galois groups of various galois extensions of number fields , especially maximal extensions of number fields with restricted ramification .
such galois groups can be regarded as tale fundamental groups of spectra of algebraic integer rings punctured at some closed points , and they play essential roles for understanding the arithmetic of number fields , analogous to the role geometric fundamental groups of manifolds do in geometry . in other words , the tale fundamental group @xmath1 of the ring of integers @xmath2 of a number field @xmath3
is isomorphic to @xmath4 where @xmath5 is the maximal extension of @xmath3 which is unramified over all finite places .
this is one of the motivations for the study of unramified extensions of number fields and their galois groups . in general
, one can know the abelianizations of the tale fundamental group @xmath6 by examining the ideal class group of @xmath2 .
results @xcite tell us that @xmath7 , where @xmath3 s are imaginary quadratic fields with absolute discriminant @xmath8 and @xmath9 is the top of the class field tower of @xmath3 .
hence we can find examples of abelian or solvable tale fundamental groups .
however , one can get little information on the structure of @xmath6 itself via class field theory , especially in the case where @xmath6 is a nonabelian simple group .
the first author has previously @xcite given an example of @xmath3 with @xmath10 under the assumption of the generalized riemann hypothesis ( grh ) .
let @xmath3 be a quadratic number field @xmath11 and @xmath12 be a splitting field of @xmath13 a polynomial that has complex roots .
the field @xmath12 is an @xmath14-extension of @xmath15 and @xmath16 is the only prime ramified in this field , with ramification index two @xcite .
moreover , @xmath12 is the unique such field @xcite . by abhyankar s lemma ,
the compositum @xmath17 is an extension of @xmath3 which is unramified over all finite places , and @xmath17 is the unique @xmath0 extension of @xmath15 ramified only at 653 with ramification index 2 . under the assumption of grh , the first author proved @xcite that the class number of @xmath17 is 1 and futhermore that @xmath18 .
we have the following natural question : is it possible to show that @xmath19 without assuming grh , thereby proving unconditionally the existence of a non - solvable tale fundamental group ? to do this ,
the first step is to unconditionally prove that @xmath17 has class number one .
if a number field has a large degree and discriminant , the computation of the class number becomes quite difficult .
the minkowski bound is too large to be useful , and the root discriminant of @xmath17 is too large to be treated by odlyzko s unconditional discriminant bounds .
however , the second author showed @xcite that by finding nontrivial lower bounds for sums over prime ideals of the hilbert class field , upper bounds can be established for class numbers of fields of larger discriminant . in this paper
, we unconditionally prove that @xmath17 has class number 1 .
in fact , @xmath17 has degree 120 , and it is the largest degree number field proven unconditionally to have class number 1 .
previously , among such fields , the one with largest degree , the real cyclotomic field of conductor 151 , has degree 75 @xcite .
[ mainresult ] the unique @xmath0 extension of @xmath15 ramified only at 653 with ramification index 2 has class number 1 .
the first author proved @xcite that if the class number of @xmath17 is less than 16 , then the class number of @xmath17 is 1 . using discriminant lower bounds ,
it was further proved , _ under the assumption of grh _ , that the class number of @xmath17 is indeed less than 16 ( and therefore is 1 ) , and also that the maximal unramified extension of @xmath3 is @xmath17 .
( main theorem of @xcite)[kimmaintheorem ] let @xmath3 be a real quadratic field @xmath21 with narrow class number @xmath22 , where @xmath23 is a prime congruent @xmath22 mod @xmath24 and @xmath25 .
suppose that there exists a totally imaginary @xmath14-extension @xmath12 over @xmath15 and @xmath23 is the only prime ramified in this field with ramification index @xmath26 . if @xmath27 , then the class number of @xmath17 is one and @xmath28 ( @xmath29 is defined as an infimum of @xmath30 over all number fields @xmath31 satisfying @xmath32 and @xmath33 ( resp .
@xmath34 ) where @xmath35 ( resp .
@xmath36 ) is the number of real ( resp .
complex ) places of a number field @xmath31 . ) in @xmath37 , the number @xmath38 represents @xmath39 in particular , the proof of theorem [ kimmaintheorem ] has the following consequence : [ cor ] with the notations and conditions being the same as above , if the class number of @xmath17 is smaller than @xmath40 , then the class number of @xmath17 is exactly one .
here we briefly sketch the main elements of the proof , the details of which are given in @xcite .
let @xmath41 denote the hilbert class field of @xmath17 , and let @xmath42 denote the hilbert @xmath43-class field for a prime @xmath43 .
every automorphism group of an abelian group of order less than 16 does not contain a subgroup isomorphic to @xmath14 , so @xmath44 is a central extension of @xmath14 by @xmath45 .
if @xmath43 is odd , @xmath44 has a nontrivial abelian quotient ( * ? ? ?
* cor 2.2 ) , contradicting that the narrow class number of @xmath3 is 1 .
thus the @xmath43-class group of @xmath17 is trivial for odd @xmath43 .
similarly , @xmath46 has a nontrivial abelian quotient if @xmath47 has order greater than 2 ( * ? ? ?
* cor 2.3 ) .
also , if @xmath47 is trivial , then we are done .
otherwise , we have @xmath46 isomorphic to @xmath0 or @xmath48 ( * ? ? ?
* cor 2.3 ) .
however , @xmath46 can not be isomorphic to @xmath0 because @xmath3 has narrow class number 1 .
finally , it has been shown @xcite that @xmath49 also leads to a contradiction . to prove unconditionally that @xmath17 has class number 1
, it remains to be shown that the class number is less than 16 .
the root discriminant @xmath50 of @xmath17 is approximately 25.5539 .
if we assume the generalized riemann hypothesis , we can use discriminant lower bounds @xcite to show that any totally complex field with degree 480 or larger must have root discriminant larger than 26.48 .
but the root discriminant of @xmath17 is equal to the root discriminant of its hilbert class field , so under grh the class number of @xmath17 must be less than @xmath24
. however , without the assumption of grh , this method fails for totally complex fields with root discriminant above @xmath51 . to make further progress
, we must find another approach that can handle large root discriminants .
such a method was introduced in @xcite for totally real fields . by finding sufficiently many integral elements with small prime norm , an upper bound for class numbers
could be established even for fields of large discriminant .
we prove a similar result of totally complex fields .
[ upperbound ] let @xmath3 be a totally complex galois number field of degree @xmath52 , and let @xmath53 for some positive constant @xmath54 .
suppose @xmath55 is a subset of the prime integers which are unramified in @xmath3 and factor into principal prime ideals of @xmath3 of degree @xmath56 .
let @xmath57 where @xmath58 is euler s constant .
if @xmath59 then we have an upper bound for the class number @xmath60 of @xmath3 , @xmath61 the proof is a modification of the argument in @xcite .
we apply poitou s version @xcite of weil s explicit formula " for the dedekind zeta function of hilbert class field @xmath62 of @xmath3 : @xmath63 where @xmath58 is euler s constant and @xmath64 since @xmath3 is totally complex .
the first sum is over the nontrivial zeros of the dedekind zeta function of @xmath62 , the second sum is over the prime ideals of @xmath62 , and @xmath65 is defined by @xmath66 by our choice of @xmath31 , the real part of @xmath67 is nonnegative everywhere in the critical strip .
indeed , on the boundary of the critical strip , the real part @xmath68 is positive , and @xmath69 as @xmath70 , so by the maximum modulus principle for harmonic functions , @xmath71 can not be negative anywhere in the critical strip .
since the root discriminant @xmath72 of @xmath3 equals the root discriminant of @xmath62 , we have @xmath73 and also @xmath74 we therefore we get the expression @xmath75 we rearrange this to get the identity @xmath76}\ ] ] where @xmath77 to get an upper bound for the class number @xmath60 , we need to bound from below the sum over the zeros and the sum over the primes .
the sum @xmath78 over the critical zeros is nonnegative since the real part of @xmath67 is nonnegative on the critical strip .
we note that principal ideals in @xmath3 totally split in the hilbert class field of @xmath3 . to find a nontrivial lower bound for the sum over prime ideals of the hilbert class field
, we consider the contribution of the @xmath79 prime ideals @xmath80 of degree @xmath56 that lie over some unramified rational prime @xmath23 : @xmath81 summing this contribution over an arbitrary set @xmath55 of unramified primes gives a lower bound for the sum over the prime ideals , proving the theorem .
in order to apply theorem [ upperbound ] , we must find sufficiently many integral elements of small prime power norm . to do this ,
we first must compute a basis of the ring of integers @xmath82 of @xmath17 . in general
, it is difficult to compute an integral basis for a number field with such large degree and takes an unfeasibly long time using the commonly implemented algorithms .
fortunately , jordi gurdia , jess montes and enric nart studied and recently implemented an algorithm that allows for fast computation of an integral basis .
their algorithm has excellent heuristic running times and low memory requirements .
the detailed algorithms are described in @xcite , @xcite and @xcite . even though their algorithms depend on a conjecture , their program verifies that the returned basis indeed generates a maximal order and warns the user if the test fails .
montes package " is available at their homepage @xcite . given a number field @xmath83 defined by an irreducible monic polynomial @xmath84 with integer coefficients , and a set @xmath55 of prime divisors of the discriminant of @xmath85 , the montes package computes an integral basis of the ring of integers @xmath86 . we must find a polynomial @xmath84 defining our field @xmath17 .
magma @xcite can use the polynomial ( [ poly1 ] ) to find the defining polynomial of @xmath17 , but the coefficients and the discriminant of such polynomial are extremely high . instead , to use the montes package efficiently ,
we must find a polynomial @xmath84 with relatively small coefficients . from the work of basmaji and kiming @xcite ,
it is known that there is a unique @xmath14-extension of @xmath15 which is ramified only at @xmath16 with ramification index @xmath26 , and so there is a unique @xmath0-extension of @xmath15 ramified only at @xmath16 with ramification index @xmath26 .
we can also check that the splitting field of @xmath87 ( see @xcite ) is an @xmath0-extension of @xmath15 , and ramified only at @xmath16 with ramification two .
therefore the splitting field of ( [ poly2 ] ) is isomorphic to @xmath17 , and we find that the defining polynomial @xmath84 of the splitting field has reasonably - sized coefficients : + @xmath88 + now let @xmath89 be a zero of the above polynomial , i.e. @xmath90 .
we can get an integral basis @xmath91 for @xmath17 by using the montes package and magma software , where each @xmath92 $ ] .
the basis is included as an ancillary file with the arxiv submission of this paper .
let @xmath93 be the 60 conjugate pairs of embeddings of @xmath94 into the complex numbers .
we have the usual embedding @xmath95 of @xmath17 into @xmath96 , @xmath97 under this map , ring of integers @xmath82 is embedded as a lattice in @xmath96 . with some abuse of notation ,
we respectively define the _ multiplicative norms _ @xmath98 of @xmath99 and @xmath100 of @xmath101 to be @xmath102 @xmath103 so that @xmath104 . in order to find an upper bound for the class number of @xmath17
, we must find sufficiently many integral elements with small multiplicative norm .
for this purpose , we desire a nice " integral basis for @xmath82 in the sense that embedded basis elements have small euclidean lengths in @xmath105 . for the basis @xmath106 found by the montes package in the previous subsection , the euclidean lengths range from @xmath107 to a rather large @xmath108 ,
so we do not yet have a nice " basis .
we thereby encounter the classical problem of lattice basis reduction .
we can apply the lenstra - lenstra - lovsz ( lll ) algorithm @xcite , but it only gets us so far : a single application of the lll algorithm in maple to our initial basis @xmath109 yields a new basis in @xmath96 with euclidean norms ranging from @xmath110 to approximately @xmath111 , a substantial improvement yet not good enough for our goal of finding integral elements of small multiplicative norm .
indeed , a vector in @xmath96 with euclidean norm of 64.51 can have a multiplicative norm larger than @xmath112 .
there exist several other methods of lattice basis reduction that can produce shorter vectors than the lll algorithm at the expense of longer running time , but here we take a somewhat naive approach that is easy to implement in maple , using its native implementation of the lll algorithm .
suppose that , given a list of basis vectors @xmath113 in @xmath114 , we want apply the lll algorithm to produce a basis of relatively short vectors . typically , for lattices of high dimension and vectors of widely varying euclidean length , the results can be improved if the @xmath115 are first ordered in ascending length prior to applying the lll algorithm .
this motivates the following approach : 1 .
sort the vectors so that the @xmath115 are in ascending length .
2 . apply the lll algorithm to the sorted list of vectors to produce a new list of basis vectors @xmath115 .
repeat steps 1 and 2 until the desired shortness of basis vectors is achieved , or until repeating steps 1 and 2 no longer produces bases with shorter vectors .
perhaps surprisingly , this simple method works quite well with the integral basis @xmath106 of @xmath82 . as mentioned above , we first applied lll to @xmath109 to get a new basis .
then we applied steps 1 and 2 repeatedly ( in fact , 279 times ) until we obtained another basis @xmath116 that has sufficiently short @xmath117 .
the @xmath117 have euclidean norms that range from @xmath118 to @xmath119 .
the inverse of change of basis matrix to get from @xmath109 to @xmath120 can be found as an ancillary file with the arxiv submission of this paper .
our next step is to find integral elements of our number field @xmath17 with small multiplicative norm .
then we can apply the theorem from @xcite to establish an upper bound for the class number . to find such elements of small norm
, we will search over sparse " vectors , i.e. vectors where almost all the coefficients are zero with respect to the basis @xmath121 .
table [ tableprimenorm ] lists the elements of small prime norm that are found by searching over sparse " vectors .
these prime integers generate principal ideals which totally split in @xmath17 into 120 principal prime ideals , each of which is generated by the given element or one of its galois conjugates .
therefore we can include the primes given in table [ tableprimenorm ] as degree 1 primes in the set @xmath55 used by theorem [ upperbound ] .
.generators of some small degree 1 primes in @xmath82 [ cols="^,^,^",options="header " , ]
we are now in a position to prove our main result by applying theorem [ upperbound ] and corollary [ cor ] . by searching for elements of small norm , and taking quotients where necessary
, we find a number of primes that can be included in the set @xmath55 of unramified primes that factor into principal prime ideals .
in particular , the following 36 primes totally split into degree 1 principal primes in @xmath17 : @xmath122 @xmath123 @xmath124 also , the following 11 primes factor into degree 2 principal primes in @xmath17 : @xmath125 finally , the following three primes factor into degree 3 principal primes in @xmath17 : @xmath126 if we include these primes in our set @xmath55 and set @xmath127 , we find that @xmath128 we can numerical calculate the integral and find that @xmath129 since the root discriminant of @xmath17 is @xmath130 , we have : @xmath131 therefore , by theorem [ upperbound ] , we get an upper bound for the class number of @xmath17 : @xmath132 since the class number is an integer , we deteremine that @xmath133 applying corollary [ cor ] , we conclude that the class number of @xmath17 is 1 .
we have shown conditionally that the class number of @xmath17 is one . thus we know that @xmath17 does not have any solvable unramified extensions . to prove unconditionally that the tale fundamental group of @xmath134 is @xmath14 , the remaining task is to show that @xmath17 does not have a nonabelian simple unramified extension . how might we rule out such nonabelian simple unramified extensions ?
there are at least two possible approaches : one is to gain sufficient knowledge of the prime ideals of the unramified extension in order to bound the degree of the extension , similar to our approach for bounding the degree of hilbert class field .
but this is difficult : we do not have access to a class field theory for nonabelian extensions as we have in the abelian case .
a possible alternative approach may be to use the theory of galois representations to attempt to rule out possible nonabelian unramified extensions of @xmath17 .
the authors would like to thank takeshi tsuji and hiroki takahashi , the organizers of the 2014 rims workshop on algebraic number theory where the authors had the opportunity be introduced .
j. basmaji , i. kiming , _ a table of @xmath14 fields _ , on artin s conjecture for odd 2-dimensional representations , 3746 , 122141 , lecture notes in math .
1585 , springer , berlin ( 1994 ) .
mr1322317 ( 96e:11141 ) .
j. gurdia , j. montes , e. nart , _ higher newton polygons in the computation of discriminants and prime ideal decomposition in number fields _ , j. thor .
nombres bordeaux 23 ( 2011 ) , no .
3 , 667696 . mr2861080 .
j. gurdia , j. montes , e. nart , _ newton polygons of higher order in algebraic number theory _ , trans .
( 2012 ) , no .
1 , 361416 . mr2833586 ( 2012k:11185 ) .
j. martinet , _ petits discriminants des corps de nombres _ , journes arithmtiques 1980 , 151193 , london math .
lecture note ser .
56 , cambridge univ . press , cambridge ( 1982 ) .
mr0697261 ( 84g:12009 ) .
serre , _ modular forms of weight one and galois representations _ , algebraic number fields : l - functions and galois properties ( proc .
durham , durham , 1975 ) , 193268 , academic press , london ( 1977 ) .
mr0450201 ( 56 # 8497 ) .
k. yamamura , _ maximal unramified extensions of imaginary quadratic number fields of small conductors _ , j. thor .
nombres bordeaux 9 ( 1997 ) , no .
2 , 405448 . mr1617407 ( 99e:11139 ) . | if a number field has a large degree and discriminant , the computation of the class number becomes quite difficult , especially without the assumption of grh . in this article
, we will unconditionally show that a certain nonabelian number field of degree 120 has class number one .
this field is the unique @xmath0 extension of the rationals that is ramified only at 653 with ramification index 2 .
it is the largest degree number field unconditionally proven to have class number 1 .
the proof uses the algorithm of gurdia , montes and nart to calculate an integral basis and then finds integral elements of small prime power norm to establish an upper bound for the class number ; further algebraic arguments prove the class number is 1 .
it is possible to apply these techniques to other nonabelian number fields as well . |
luminous infrared galaxies ( lirgs ) show strong infrared emission , with infrared ( 81000 @xmath4 m ) luminosities of l@xmath7 @xmath3 10@xmath8l@xmath6 , created by energy sources hidden behind dust . they are mostly found in gas - rich galaxy mergers @xcite .
molecular gas in galaxy mergers is largely influenced by merger - induced physical processes , and obtaining observational constraints on the spatial distribution and dynamics of molecular gas in merging lirgs is important to our understanding of gas - rich galaxy merger processes .
( sub)millimeter interferometric observations of rotational j - transition lines of molecular gas are a powerful tool for this purpose .
high - sensitivity , high - spatial - resolution ( @xmath9a few arcsec ) interferometric observations of merging lirgs using bright co molecular lines have been widely performed @xcite .
these co observations at low - j transitions ( j = 10 and 21 ) have effectively traced the low - density ( 10@xmath10 @xmath11 ) molecular gas properties in great detail due to the low dipole moment of co ( @xmath4 @xmath1 0.1 debye ) . however , in merging lirgs , the fraction of high density ( @xmath310@xmath12 @xmath11 ) molecular gas is much higher than in normal quiescent star - forming galaxies @xcite , and it is within such high - density gas that stars are actually born .
thus , it is vital to obtain observational constraints of the properties of high - density molecular gas in merging lirgs if we are to understand the essential physical processes in gas - rich galaxy mergers
. observations of molecular gas with high dipole moments , such as hcn , hco@xmath0 , and hnc ( @xmath4 @xmath3 3 debye ) , can effectively probe high - density molecular gas .
however , these molecular lines are generally much fainter than the bright co lines , and the spatial information of dense gas is still limited to nearby , very bright merging lirgs only @xcite . from these previously performed observations
, it has been shown that the spatial distribution of high - density molecular gas is significantly different from that of low - density molecular gas , in that high - density gas is more concentrated in the nuclear regions of galaxies @xcite . in merging lirgs ,
it is in the nuclear regions where very violent processes , including starbursts (= active star - formation ) and active mass accretion onto a supermassive black hole ( smbh ) occur .
additionally , feedback to the surrounding interstellar medium and entire galaxy , if present , originates from these regions @xcite .
spatially resolved interferometric observations using dense gas tracers are of particular importance in investigating the interesting nuclear regions of merging lirgs .
starburst and active galactic nucleus ( agn ) activity powered by a mass - accreting smbh can have different effects / feedback on the surrounding dense molecular gas at the merging lirg s nuclei .
it is proposed that starburst and agn activity could be distinguishable based on the line flux ratios of dense molecular gas tracers @xcite because ( 1 ) an agn has an energy source with a much higher emission surface brightness than a starburst , and thus it can heat the surrounding dust and gas to higher temperature ; and ( 2 ) an agn emits stronger x - rays than a starburst does . both of these factors could alter the chemical compositions of molecular gas in agns compared with starbursts @xcite ; therefore , agns and starbursts could exhibit different molecular line flux ratios .
high - spatial - resolution interferometric observations of multiple dense gas tracers can thus be used to scrutinize the processes deep inside the obscuring dust and gas in merging lirg nuclei . with the advent of alma ,
such observations are now feasible .
ngc 1614 ( z = 0.016 ; l@xmath7 = 10@xmath13l@xmath6 ) is a well - studied , nearby lirg .
merging signatures are clearly seen in the optical and near - infrared ( 12.5 @xmath4 m ) images as long , prominent tails around a single nucleus @xcite .
it is classified as a starburst galaxy through optical spectroscopy @xcite .
the infrared 2.530 @xmath4 m spectrum of ngc 1614 is typical of a starburst - dominated galaxy with no agn signature @xcite .
the luminosities of the starburst - generated 3.3 @xmath4 m polycyclic aromatic hydrocarbon ( pah ) emission feature and the br@xmath14 ( 4.05 @xmath4 m ) hydrogen emission , measured through slitless spectroscopy and calculated relative to the total infrared ( 81000 @xmath4 m ) luminosity , are both as high as in starburst - dominated galaxies @xcite , suggesting that the observed luminosity of ngc 1614 is totally accounted for by the detected starbursts , with no need for a significant agn contribution .
the infrared @xmath15-band ( 2.2 @xmath4 m ) spectrum shows a strong stellar - origin 2.3 @xmath4 m co absorption feature @xcite and so supports the starburst - dominated scenario .
high - spatial - resolution infrared 820 @xmath4 m imaging observations reveal spatially extended , but compact ( @xmath12@xmath2 ) , starburst - heated dust continuum emission @xcite , and the measured emission surface brightness is within the range explained by star formation , again requiring no significant contribution from an energetically important agn @xcite .
the pa@xmath14 ( 1.88 @xmath4 m ) emission from star - forming hii - regions is dominated by nuclear @xmath13@xmath2 area @xcite , which suggests that most of starburst activity in ngc 1614 is concentrated in the nuclear regions within @xmath9a few arcsec .
although alma is a very powerful tool for unveiling the molecular gas distribution in detail , it is not sensitive to spatially extended structure beyond the maximum scale , which is @xmath16@xmath2 at @xmath1350 ghz in alma cycle 0 .
the spatially resolved , but intrinsically compact ( @xmath12@xmath2 ) nuclear emission morphology and the known starburst - dominated nature make ngc 1614 an ideal target to investigate the spatial distribution and dynamics of dense molecular gas in merging lirg nuclei , as well as to obtain a template of line flux ratios of dense gas tracers in starburst - dominated galaxies , during alma cycle 0 .
we thus performed alma band 7 ( 275373 ghz ) observations of ngc 1614 to study the emission properties of hcn j = 43 , hco@xmath0 j = 43 , and hnc j = 43 lines .
the basic information of ngc 1614 is summarized in table 1 . throughout this paper , we adopt h@xmath16 @xmath17 71 km s@xmath18 mpc@xmath18 , @xmath19 = 0.27 , and @xmath20 = 0.73 @xcite , where 1@xmath2 corresponds to @xmath1320 pc at the distance of ngc 1614 ( z = 0.016 ) .
all observations were made during alma cycle 0 within the program 2011.0.00020.s ( pi = m. imanishi ) .
observation details are described in table 2 .
we adopted the widest 1.875 ghz band mode , and the total channel number was 3840 . for ngc 1614 ( z = 0.016 ) ,
hcn j = 43 ( @xmath21 = 354.505 ghz ) , and hco@xmath0 j = 43 ( @xmath21 = 356.734 ghz ) lines are simultaneously observable in alma band 7 .
four frequency setups can cover four different frequencies at the same time .
two were used to observe hcn j = 43 ( central frequency was set as @xmath22= 348.922 ghz ) and hco@xmath0 j = 43 lines ( @xmath22 = 350.920 ghz ) , and the remaining two were used to measure the continuum flux level ( @xmath22 = 337.106 ghz and 338.681 ghz ) .
the net on - source exposure time for the hcn / hco@xmath0 j = 43 line observation of ngc 1614 was @xmath126 min .
the frequency of the hnc j = 43 line ( @xmath21 = 362.630 ghz ) is separated from the hcn j = 43 and hco@xmath0 j = 43 lines , and we required independent observations .
the hnc line was covered with one spectral window ( @xmath22 = 356.920 ghz ) , and an additional second spectral window was used to probe the continuum emission ( @xmath22 = 345.079 ghz ) .
the net on - source exposure time for the hnc j = 43 line observation of ngc 1614 was @xmath125 min .
we started data analysis from calibrated data provided by the joint alma observatory .
we first checked the visibility plots to see if the signatures of the emission lines were recognizable .
the presence of hcn , hco@xmath0 , and hnc j = 43 lines were evident in the visibility plot , whereas signatures of other molecular lines were not clearly seen .
we then selected channels that were free from strong line emission to estimate the continuum flux level .
we subtracted this continuum level and performed the task `` clean '' for molecular emission lines .
the `` clean '' procedure was applied also to the continuum data .
we employed 40-channel spectral binning ( @xmath117 km s@xmath18 ) and 0.3@xmath2 pixel@xmath18 spatial binning in this clean procedure .
continuum emission properties are shown in figure 1 and table 3 .
figure 2 displays the integrated intensity ( moment 0 ) maps of molecular lines and spectra within the beam size at the continuum peak positions .
continuum emission is found to be well subtracted in the spectra ( figure 2 right panels ) , and so the moment 0 maps in figure 2 ( left ) should reflect the properties of individual molecular gas emission lines .
the peak flux , rms noise level , and synthesized beams in individual maps are summarized in table 4 ( row denoted `` all '' ) .
the peak positions of hcn j = 43 and hco@xmath0 j = 43 agree with that of continuum `` a '' , whereas that of hnc j = 43 is two pixels ( 0.6@xmath2 ) displaced to the south from the continuum `` b '' peak .
cs j = 76 line at @xmath21 = 342.883 ghz is covered in a spectral window during the observations of hcn / hco@xmath0 j = 43 line , but it is not clearly detected in the spectrum ( within the beam size ) at the nucleus , which is defined from the continuum `` a '' emission peak ( figure 2 , right ) . because the signals of the cs j = 76 emission are not clearly seen , to create the moment 0 map of cs j = 76 , we refer to the velocity profile of the brightest hco@xmath0 j = 43 emission line at the nucleus and sum signals with v@xmath23 @xmath24 c ( @xmath25-@xmath26)/@xmath26 = 46004925 [ km s@xmath18 ] .
no clear cs j = 76 emission line is seen ( @xmath93@xmath28 ) in the moment 0 map ( figure 2 , left and table 4 ) .
this undetected cs j = 76 line will not be used in the following discussion .
the moment 0 maps in figure 2 display spatially extended structures compared with the beam size , particularly for hcn and hco@xmath0 j = 43 lines .
we created spectra of the hcn , hco@xmath0 , and hnc j = 43 lines , integrated over all regions with significant signal detection ( @xmath13@xmath2 @xmath29 3@xmath2 ) , in figure 3 .
all show double - peaked emission profiles , as seen in the @xmath30co j = 21 , @xmath30co j = 32 , and @xmath31co j = 21 lines @xcite .
we applied double gaussian fits .
the derived parameters are summarized in table 5 ( rows denoted with `` all '' ) . for hco@xmath0 j
= 43 emission , a triple gaussian fit was also applied because three emission peaks were seen .
the molecular line luminosities , integrated over regions of significant signal detection ( corresponding to `` all '' in table 5 ) are summarized in table 6 , where we adopted equations ( 2 ) and ( 3 ) of @xcite .
previously obtained high - spatial - resolution images of starburst indicators , such as the pa@xmath14 emission line ( 1.88 @xmath4 m ) , the pah emission feature ( 3.3 @xmath4 m ) , the infrared 8.7 @xmath4 m continuum , and the radio 58.4 ghz ( 3.66 cm ) continuum , have revealed the presence of ring - shaped circumnuclear starbursts with radii of 0.51.0@xmath32 @xcite .
a similar ring pattern is discernible in the high - spatial - resolution ( 0.5@xmath32 @xmath29 0.4@xmath32 ) co j = 21 molecular line map @xcite .
our alma map has a resolution of 1.5@xmath32 @xmath29 1.3@xmath32 . in this map , particularly in the brightest hco@xmath0 j = 43 line , where the highest s / n ratios are achieved , the emission is more elongated toward the north - south direction than the east - west direction , suggesting that a larger amount of high - density molecular gas is distributed at the northern and southern part of the nucleus .
figure 4 displays the intensity weighted mean velocity ( moment 1 ) and intensity weighted velocity dispersion ( moment 2 ) maps of hcn and hco@xmath0 j = 43 lines .
the dense molecular gas in the northern ( southern ) part of the nucleus is clearly redshifted ( blueshifted ) relative to the nuclear velocity of ngc 1614 ( v@xmath23 = 4800 [ km s@xmath18 ] for z = 0.016 ) .
the contours of emission of the red component with v@xmath23 @xmath3 4800 [ km s@xmath18 ] and the blue component with v@xmath23 @xmath9 4800 [ km s@xmath18 ] are displayed in figure 5 .
the emission properties of the red and blue components are summarized in table 4 ( rows denoted with `` red '' and `` blue '' ) .
the emission peak positions of the red and blue components are shifted by two pixels ( 0.6@xmath2 ) to the north and south direction , respectively , from the nucleus . for hnc
j = 43 emission in figure 5 , the blue component is much brighter than the red component , which could explain the slight peak offset in the hnc j = 43 moment 0 map ( integrated over all velocity components with significant signal detection ) with respect to the nuclear position ( table 3 ) .
spectra within the beam size , at the nuclear position and the peak position of the red and blue components , are shown in figure 6 .
as expected , spectra at the red ( blue ) component peaks are redshifted ( blueshifted ) , relative to those at the nuclear position for all the hcn , hco@xmath0 , and hnc j = 43 lines .
gaussian fits were applied to the spectra .
the resulting parameters are shown in table 5 ( rows denoted with `` peak '' , `` red '' , and `` blue '' ) .
the line widths of these dense molecular gas tracers are generally larger in the blue component than the red component , which has also been seen in ionized gas @xcite .
this indicates that the southern part of the nucleus is more turbulent than the northern part .
figure 7 presents the position - velocity diagram along the north - south direction , passing through the nucleus .
the channel map of the brightest hco@xmath0 j = 43 emission line is shown in figure 8 .
the dense molecular gas in the northern region has greater velocity ( more redshifted ) than the southern molecular gas .
the red and blue components of the dense molecular gas show emission peaks at @xmath10.6@xmath2 north and south of the nucleus , respectively , roughly corresponding to the northern and southern parts of the previously identified circumnuclear starburst ring with a radius of 0.51@xmath2 @xcite .
@xcite investigated the spatial distribution of the 3.3 @xmath4 m pah emission feature , and found particularly strong pah emission at the northern and southern parts of the ring .
the 3.3 @xmath4 m pah emission is a good indicator of starburst activity @xcite .
in contrast , the pa@xmath14 ( 1.88 @xmath4 m ) emission , originating in hii regions , is strongest at the eastern and western sides of the circumnuclear starburst ring @xcite . how are these two kinds of starbursts related ?
one scenario involves an age difference .
the eastern and western starburst regions are probed with the tracers of hii regions where plenty of ionizing ( @xmath25 @xmath9 912 @xmath33 ) uv photons , usually dominated by short - lived massive o - stars , are needed . in contrast
, the 3.3 @xmath4 m pah emission mostly comes from photo - dissociation regions between hii regions and molecular gas , where pahs are excited by non - ionizing ( @xmath25 @xmath3 912 @xmath33 ) stellar uv photons @xcite . for the pah - exciting non - ionizing uv continuum
, the contribution from less massive stars than o stars is higher than that for ionizing uv photons inside hii regions .
thus , the pah emission feature is sensitive not only to very young o - stars dominant starbursts but also to slightly aged starbursts where massive o - stars have mostly died , but less massive stars ( e.g. , b - stars ) still survive and emit a sufficient quantity of non - ionizing pah - exciting uv photons .
@xcite showed that the 8.7 @xmath4 m infrared dust continuum , relative to the hii region tracer pa@xmath14 line ( 1.88 @xmath4 m ) , is enhanced at the northern and southern starburst ring . because non - ionizing uv continuum emission can contribute significantly to the infrared continuum emission but can not do so for the pa@xmath14 emission , the observed spatial variation in the 3.3 @xmath4 m pah to pa@xmath14 and 8.7 @xmath4 m to pa@xmath14 flux ratio is explainable under the scenario that the typical starburst age is older at the northern and southern ring than at the eastern and western ring @xcite . if the circumnuclear starburst ring is formed by nuclear starbursts that are progressing outward @xcite , the age difference in starbursts at different positions of the ring needs to be explained .
dust extinction is another possibility because pa@xmath14 emission at 1.88 @xmath4 m can be more highly flux - attenuated than the 3.3 @xmath4 m pah emission and the 8.7 @xmath4 m dust continuum .
radio free - free continuum emission from hii regions is less susceptible to foreground dust extinction and could help to determine whether pa@xmath14 emission is significantly affected by dust extinction .
high - spatial - resolution radio continuum maps at 5 ghz ( 6 cm ) and 8.4 ghz ( 3.6 cm ) are available @xcite and @xcite ascribe the radio 5 ghz and 8.4 ghz emission in ngc 1614 to free - free emission from hii - regions in young starbursts .
because the effects of free - free absorption inside hii regions are smaller at 8.4 ghz than at 5 ghz , the 8.4 ghz radio emission map is taken as the better probe of the true spatial distribution of hii regions . in the radio 8.4 ghz map ,
strong emission is detected in the eastern and western starburst ring @xcite , as seen in pa@xmath14 line map @xcite , confirming that luminous hii regions are present at those locations .
however , at 8.4 ghz , bright emission is seen also at the northern starburst ring , where pa@xmath14 emission is not strong @xcite .
similarly , when compared with the eastern starburst ring , the southern starburst ring is more conspicuous at 8.4 ghz than in pa@xmath14 @xcite .
the comparison of the radio 8.4 ghz ( 3.6 cm ) and 1.88 @xmath4 m pa@xmath14 emission indicates that hii - regions at the northern and southern parts of the starburst ring , unveiled by the radio 8.4 ghz emission , are not sufficiently distinguished by pa@xmath14 emission .
these northern and southern regions of the starburst ring are the locations where dense molecular gas is distributed , according to our alma data . given that dust coexists with dense molecular gas ,
dust extinction is a natural explanation for the small pa@xmath14 to 8.4 ghz flux ratio at the northern and southern starburst rings .
weak dust extinction is reported for starburst regions in ngc 1614 compared with other general starburst galaxies @xcite , based on near - infrared observations at @xmath25 @xmath9 2 @xmath4 m .
this could be due partly to the fact that observations at @xmath25 @xmath9 2 @xmath4 m , including pa@xmath14 emission , selectively trace emission from less dusty starburst regions at the eastern and western ring and do not properly probe the dusty starbursts at the northern and southern parts of the ring due to flux attenuation by dust extinction .
our alma data show that the high - density molecular gas in the northern part of the nucleus is redshifted and that gas in the southern part is blueshifted with respect to the nuclear velocity of this galaxy .
this suggests the rotation of dense molecular gas along the east - west axis @xcite .
a similar velocity pattern was found previously in the ionized gas maps @xcite and the lower density molecular gas probed with co j = 32 and j = 21 @xcite .
the observed velocity dispersion is highest in the nuclear region with @xmath180 km s@xmath18 ( figure 4 ) , although it may be affected by the beam smearing of a rotating motion at the center .
this value is similar to those measured through near - infrared spectroscopy @xcite and with ( sub)millimeter co j = 32 and co j = 21 emission lines @xcite .
the rotational motion found in our moment 1 maps ( figure 4 ) can be used to derive the dynamical mass inside the rotating dense molecular gas disk .
we used the hco@xmath0 j = 43 line because it is brighter than hcn j = 43 , and so the achieved s / n ratios are higher . in figure 4 , the rotational velocity is v @xmath1 100 km s@xmath18 at 1.52@xmath2 ( or r = 480640 pc at z = 0.016 ) from the nucleus .
the derived dynamical mass within 1.52@xmath2 radius is m@xmath34 = rv@xmath35/g / sin(i)@xmath35 = 1.52.5 @xmath29 10@xmath5m@xmath6 , where the inclination angle i = 51@xmath36 is adopted @xcite . this mass is comparable to the previously estimated values @xcite .
figure 9 is a plot of hcn - to - hco@xmath0 and hcn - to - hnc flux ratios at the j = 43 transition , derived from the spectra at the nucleus , red component peak position , blue component peak , and all regions with significant signal detection . in all data ,
hcn j = 43 flux is smaller than that of hco@xmath0 j = 43 , but higher than hnc j = 43 flux . using the j = 10 transitions of hcn , hco@xmath0 , and hnc , the possibility of distinguishing the hidden energy sources of merging lirg s dusty nuclei
is suggested @xcite . in general , hcn - to - hco@xmath0 flux ratios are small ( @xmath91 ) , and hcn - to - hnc flux ratios are large ( @xmath31 ) in starburst - dominated galaxies , whereas hcn - to - hco@xmath0 flux ratios can be high ( @xmath31 ) in agns .
agns could enhance hcn flux relative to hco@xmath0 , due possibly to hcn abundance enhancement by agn radiation @xcite and/or infrared radiative pumping of hcn @xcite .
the low hcn - to - hco@xmath0 flux ratios obtained in the starburst - dominated galaxy ngc 1614 at j = 43 are similar to other starburst galaxies at j = 10 . to obtain a physical interpretation of the observed j = 43 flux ratio , we need to know the excitation conditions .
the hcn j = 10 flux of ngc 1614 was found to be 7.2 [ jy km s@xmath18 ] by @xcite based on single dish telescope observations . under thermal excitation , the hcn j = 43 flux is expected to be 16 times higher than j = 10 , and so @xmath1115 [ jy km s@xmath18 ] .
our alma observations provide an observed hcn j = 43 flux from all signal - detected regions of 2.8 [ jy km s@xmath18 ] . the smaller flux of our alma data could partly be caused by missing flux , as our alma observations are insensitive to spatially extended emission with @xmath36@xmath2 .
@xcite and @xcite estimated that in ngc 1614 , the nuclear ( @xmath9several arcsec ) co j = 10 and j = 32 emission can account for @xmath330% and @xmath345% of the total flux measured with single - dish telescopes . because the hcn j = 43 line traces higher density molecular gas ( @xmath310@xmath37 @xmath11 ) than do co j = 10 and j = 32 lines and because higher - density gas is more concentrated in the nuclear region , our alma hcn j = 43 data should recover @xmath345% of the total flux . even assuming a missing flux of a factor of @xmath12 , the hcn j = 43 flux of ngc 1614 is 5.6 [ jy km s@xmath18 ] , only @xmath95% of the expected flux ( 115 jy km s@xmath18 ) for thermal excitation .
thus , hcn j = 43 line is significantly sub - thermally excited in ngc 1614 , as observed in nearby galaxies at @xmath3100 pc scale @xcite .
since the critical density of hcn j = 43 ( n@xmath38 @xmath1 2 @xmath29 10@xmath39 @xmath11 ) is higher than that of hco@xmath0 j = 43 ( n@xmath38 @xmath1 4 @xmath29 10@xmath37 @xmath11 ) @xcite , hco@xmath0 j = 43 is more easily excited than hcn j = 43 in starbursts . in an agn ,
the emission surface brightness is higher than starburst activity , so the surrounding dust and gas are heated to a higher temperature , which may help to excite hcn j = 43 more than starburst activity .
a high hcn - to - hco@xmath0 j = 43 flux ratio could be a good diagnostic of an agn , simply because of the high excitation of hcn j = 43 in an agn , even without an hcn - abundance enhancement @xcite .
hcn - to - hco@xmath0 flux ratios of known agn - important galaxies are being measured in our alma program ( imanishi et al .
, in preparation ) , and they tend to show higher hcn - to - hco@xmath0 j = 43 flux ratios than ngc 1614 , the template starburst galaxy ( see also imanishi et al . 2010a ; sakamoto et al . 2010 ; iono et al .
since the j = 43 lines of hcn and hco@xmath0 are at higher frequencies ( shorter wavelengths ) than the lower j transition lines , the empirical energy diagnostic method , if established at j = 43 , is applicable to more distant merging lirgs using alma .
this advantage is strengthened if hcn excitation is generally thermal up to the j = 43 transition in agn - important galaxies because hcn flux increases proportional to the square of frequency , partly compensating for the increase in earth s atmospheric background emission at higher - frequency alma bands .
however , if the excitation at hcn j = 43 is sub - thermal , then hcn j = 32 or j = 21 lines may be better tracers of agn in terms of actually obtainable s / n ratios with alma . additional molecular line transition data at j = 32 and 21 for starburst - dominated galaxies and
agn - important galaxies are needed ( 1 ) to distinguish whether high hcn - to - hco@xmath0 flux ratios at j = 43 in agn - important galaxies are due to hcn abundance enhancement and/or more hcn j = 43 excitation than starbursts and ( 2 ) to identify the j transition lines that are practically the best diagnostic for separating agns from starburst - dominated galaxies .
we performed hcn , hco@xmath0 , and hnc j = 43 line observations of the well - studied starburst galaxy ngc 1614 to trace the properties of the high - density molecular gas .
our results are summarized as follows : 1 .
hcn , hco@xmath0 , and hnc j = 43 emission are clearly detected in the nuclear regions of ngc 1614 , but cs j = 76 emission and hcn j=43 line at a vibrationally - excited level ( v@xmath40=1 , l=1f ) are not .
hcn , hco@xmath0 , and hnc j = 43 emission at the northern and southern parts of the nucleus are redshifted and blueshifted , respectively , with respect to the nuclear velocity of this galaxy . when the emission is separated into the red and blue components , the red and blue components are strongest at 0.6@xmath2 north and south of the nucleus for all of the hcn , hco@xmath0 , and hnc j = 43 lines .
3 . at the peak location of the red and blue components of these dense molecular gas tracers ,
the presence of active dusty starbursts is suggested , based on the infrared 3.3 @xmath4 m pah emission , infrared 8.7 @xmath4 m dust continuum emission , and radio 8.4 ghz free - free emission , supporting the scenario that starbursts occur in dense molecular gas .
the dynamical mass derived from the red and blue dense molecular gas components , assuming rotational motion , is 1.52.5 @xmath29 10@xmath5m@xmath6 within @xmath12@xmath2 in radius .
this is similar to estimates previously obtained using other methods .
the hcn - to - hco@xmath0 flux ratios are smaller than unity , and the hcn - to - hnc flux ratios are higher than unity for j = 43 in ngc 1614 , which is a similar trend to previous observations of j = 10 for starburst - dominated galaxies .
we thank e. mullar and h. nagai for their useful advice on alma data analysis .
is supported by grants - in - aid for scientific research ( no .
22012006 ) .
this paper makes use of the following alma data : ads / jao.alma#2011.0.00020.s .
alma is a partnership of eso ( representing its member states ) , nsf ( usa ) , and nins ( japan ) , together with nrc ( canada ) and nsc and asiaa ( taiwan ) , in cooperation with the republic of chile .
the joint alma observatory is operated by eso , aui / nrao , and naoj .
aalto , s. , radford , s. j. , e. , scoville , n. z. , & sargent , a. i. 1997 , apj , 475 , l107 aalto , s. , wilner , d. , spaans , m. , wiedner , m. c. , sakamoto , k. , black , j. h. , & caldas , m. 2009 , a&a , 493 , 481 alonso - herrero , a. , engelbracht , c. w. , rieke , m. j. , rieke , g. h. , & quillen , a. c. 2001 , apj , 546 , 952 bernard - salas , j. , et al .
2009 , apjs , 184 , 230 brandl , b. r. , et al .
2006 , apj , 653 , 1129 bryant , p. , & scoville , n. z. 1999 , aj , 117 , 2632 casoli , f. , willaime , m. -c . , viallefond , f. , & gerin , m. 1999 , a&a , 346 , 663 costagliola , f. , et al .
2011 , a&a , 528 , 30 de robertis , m. m. , & shaw , r. a. 1988 , apj , 329 , 629 diaz - santos , t. , alonso - herrero , a. , colina , l. , packham , c. , radomski , j. t. , & telesco , c. m. 2008 , apj , 685 , 211 di matteo , t. , springel , v. , & hernquist , l. 2005 , nature , 433 , 605 downes , d. , & eckart , a. 2007 , a&a , 468 , l57 downes , d. , & solomon , p. m. 1998 , apj , 507 , 615 evans , a. s. , mazzarella , j. m. , surace , j. a. , & sanders , d. b. 2002 , apj , 580 , 749 gao , y. , & solomon , p. m. 2004 , apjs , 152 , 63 haan , s. , et al .
2011 , aj , 141 , 100 harada , n. , herbst , e. , & wakelam , v. 2010 , apj , 721 , 1570 hopkins , p. f. , hernquist , l. , cox , t. j. , di matteo , t. , martini , p. , robertson , b. , & springel , v. 2005 , apj , 630 , 705 hopkins , p. f. , hernquist , l. , cox , t. j. , di matteo , t. , robertson , b. , & springel , v. 2006 , apjs , 163 , 1 imanishi , m. , & dudley , c. c. 2000 , apj , 545 , 701 imanishi , m. , dudley , c. c. , & maloney , p. r. 2006a , apj , 637 , 114 imanishi , m. , imase , k. , oi , n. , & ichikawa , k. 2011 , aj , 141 , 156 imanishi , m. , nakagawa , t. , ohyama , y. , shirahata , m. , wada , t. , onaka , t. , & oi , n. 2008 , pasj , 60 , s489 imanishi , m. , & nakanishi , k. 2006 , pasj , 58 , 813 imanishi , m. , nakanishi , k. , kuno , n. , & kohno , k. 2004 , aj , 128 , 2037 imanishi , m. , nakanishi , k. , & kohno , k. 2006b , aj , 131 , 2888 imanishi , m. , nakanishi , k. , tamura , y. , oi , n. , & kohno , k. 2007 , aj , 134 , 2366 imanishi , m. , nakanishi , k. , tamura , y. , & peng , c. -h .
2009 , aj , 137 , 3581 imanishi , m. , nakanishi , k. , yamada , m. , tamura , y. , & kohno , k. 2010a , pasj , 62 , 201 imanishi , m. , nakagawa , t. , shirahata , m. , ohyama , y. , & onaka , t. 2010b , apj , 721 , 1233 iono , d. , et al .
2007 , apj , 659 , 283 iono , d. , et al .
2013 , pasj , in press ( arxiv:1305.4535 ) iono , d. , ho , p. t. p. , yun , m. s. , matsushita , s. , peck , a. b. , & sakamoto , k. 2004 , apj , 616 , l63 izumi , t. , et al .
2013 , pasj , in press ( arxiv:1306:0507 ) kewley , l. j. , heisler , c. a. , dopita , m. a. , & lumsden , s. 2001 , apjs , 132 , 37 knudsen , k. k. , walter , f. , weiss , a. , bolatto , a. , riechers , d. a. , & menten , k. 2007 , apj , 666 , 156 kohno , k. 2005 , in aip conf .
783 , the evolution of starbursts , ed .
s. httemeister , e. manthey , d. bomans , & k. weis ( new york : aip ) , 203 ( astro - ph/0508420 ) komatsu , e. , et al .
2009 , apjs , 180 , 330 konig , s. , aalto , s. , muller , s. , beswick , r. j. , & gallagher iii , j. s. 2013 , a&a , 553 , a72 krips , m. , neri , r. , garcia - burillo , s. , martin , s. , combes , f. , gracia - carpio , j. , & eckart , a. 2008 , apj , 677 , 262 lintott , c. , & viti , s. 2006 , apj , 646 , l37 meijerink , r. , & spaans , m. 2005 , a&a , 436 , 397 meijerink , r. , spaans , m. , & israel , f. p. 2007 ,
a&a , 461 , 793 miles , j. w. , houck , j. r. , hayward , t. l. , & ashby , m. l. n. 1996 , apj , 465 , 191 moorwood , a. f. m. 1986 , a&a , 166 , 4 nakanishi , k. , okumura , s. k. , kohno , k. , kawabe , r. , & nakagawa , t. 2005 , pasj , 57 , 575 neff , s. g. , hutchings , j. b. , stanford , s. a. , & unger , s. w. 1990 , aj , 99 , 1088 olsson , e. , aalto , s. , thomasson , m. , & beswick , r. 2010 , a&a , 513 , a11 perez - beaupuits , j. p. , aalto , s. , & gerebro , h. 2007 , a&a , 476 , 177 ridgway , s. e. , wynn - williams , c. g. , & becklin , e. e. 1994 , apj , 428 , 609 rothberg , b. , & joseph , r. d. 2004 , aj , 128 , 2098 sakamoto , k. , et al .
2009 , apj , 700 , l104 sakamoto , k. , aalto , s. , costagliola , f. , martin , s. , ohyama , y. , wiedner , m. c. , & wilner , d. j. 2013 , apj , 764 , 42 sakamoto , k. , aalto , s. , evans , a. s. , wiedner , m. , & wilner , d. 2010 , apj , 725 , l228 sanders , d. b. , mazzarella , j. m. , kim , d. -c . , surace , j. a. , & soifer , b. t. 2003 , apj , 126 , 1607 sanders , d. b. , & mirabel , i. f. 1996 , ara&a , 34 , 749 scoville , n. z. , sanders , d. b. , sargent , a. i. , soifer , b. t. , & tinney , c. g. 1989 , apj , 345 , l25 sellgren , k. 1981 , apj , 245 , 138 shier , l. m. , rieke , m. j. , & rieke , g. h. 1994 , apj 433 , l9 soifer , b. t. et al .
2001 , aj , 122 , 1213 solomon , p. m. , downes , d. , & radford , s. j. e. 1992 , apj , 387 , l55 solomon , p. m. , & vanden bout , p. a. 2005 , ara&a , 43 , 677 springel , v. , di matteo , t. , & hernquist , l. 2005 , mnras , 361 , 776 trung , d. v. , lo , k. y. , kim , d. -c .
, gao , y. , & gruendl , r. a. 2001 , apj , 556 , 141 vaisanen , p. , rajpaul , v. , zijlstra , a. a. , reunanen , j. , & kotilainen , j. 2012 , mnras , 420 , 2209 veilleux , s. , kim , d. -c . , sanders , d. b. , mazzarella , j. m. , & soifer , b. t. 1995 , apjs , 98 , 171 wilson , c. d. , et al .
2008 , apjs , 178 , 189 yuan , t. -t . , kewley , l. j. , & sanders , d. b. 2010 , apj , 709 , 884 ccrccl a & 344 & 14.6 ( 45@xmath28 ) & ( 04 34 00.01 , @xmath4108 34 44.9 ) & 0.32 & 1.5 @xmath29 1.3 ( 81@xmath36 ) + b & 351 & 10.9 ( 29@xmath28 ) & ( 04 34 00.03 , @xmath4108 34 44.9 ) & 0.37 & 1.3 @xmath29 1.3 ( @xmath4153@xmath36 ) + lcccl hcn j=43 & all & 1.7 ( 12@xmath28 ) & 0.14 & 1.5 @xmath29 1.3 ( 76@xmath36 ) + & red & 1.3 ( 14@xmath28 ) & 0.088 & 1.5 @xmath29 1.3 ( 76@xmath36 ) + & blue & 0.99 ( 9.4@xmath28 ) & 0.11 & 1.5 @xmath29 1.3 ( 76@xmath36 ) + hco@xmath0 j=43 & all & 7.1 ( 33@xmath28 ) & 0.22 & 1.5 @xmath29 1.3 ( 76@xmath36 ) + & red & 5.2 ( 42@xmath28 ) & 0.12 & 1.5 @xmath29 1.3 ( 76@xmath36 ) + & blue & 5.7 ( 36@xmath28 ) & 0.16 & 1.5 @xmath29 1.3 ( 76@xmath36 ) + hnc j=43 & all & 1.1 ( 7.0@xmath28 ) & 0.16 & 1.3 @xmath29 1.3 ( @xmath4153@xmath36 ) + & red & 0.57 ( 6.9@xmath28 ) & 0.084 & 1.3 @xmath29 1.3 ( @xmath4153@xmath36 ) + & blue & 1.1 ( 8.2@xmath28 ) & 0.13 & 1.3 @xmath29 1.3 ( @xmath4154@xmath36 ) + cs j=76 & all & @xmath90.45(@xmath93@xmath28 ) & 0.15 & 1.5 @xmath29 1.4 ( 81@xmath36 ) + ll|cccc hcn j=43 & all & 4733@xmath4225 + 4857@xmath4214 & 11@xmath422 + 12@xmath424 & 171@xmath4252 + 73@xmath4230 & [email protected] + & peak & 4822@xmath4210 & 11@xmath421 & 131@xmath4224 & [email protected] + & red & 4853@xmath422 & 18@xmath422 & 68@xmath427 & [email protected] + & blue & 4720@xmath4217 & [email protected] & 214@xmath4242 & [email protected] + hco@xmath0 j=43 & all & 4718@xmath429 + 4858@xmath423 & 50@xmath423 + 70@xmath426 & 200@xmath4220 + 72@xmath428 & [email protected] + & & 4650@xmath424 + 4755@xmath425 + 4857@xmath422 & 51@xmath426 + 49@xmath422 + 81@xmath428 & 70@xmath428 + 105@xmath4231 + 72@xmath425 & [email protected] + & peak & 4803@xmath424 & 41@xmath422 & 167@xmath429 & [email protected] + & red & 4854@xmath421 & 72@xmath423 & 76@xmath422 & [email protected] + & blue & 4715@xmath424 & 37@xmath422 & 180@xmath429 & [email protected] + hnc j=43 & all & 4718@xmath4211 + 4859@xmath4216 & 10@xmath422 + [email protected] & 125@xmath4227 + 50@xmath4245 & [email protected] + & peak & 4800 ( fix ) & [email protected] & 183@xmath4275 & [email protected] + & red & 4856@xmath423 & 13@xmath422 & 42@xmath427 & [email protected] + & blue & 4714@xmath429 & [email protected] & 116@xmath4223 & [email protected] + lcr hcn j=43 & ( [email protected])@xmath2910@xmath43 & ( [email protected])@xmath2910@xmath37 + hco@xmath0 j=43 & ( [email protected])@xmath2910@xmath12 & ( [email protected])@xmath2910@xmath39 + hnc j=43 & ( [email protected])@xmath2910@xmath43 & ( [email protected])@xmath2910@xmath37 + j = 43 and hnc j = 43 lines , respectively .
contours are 5@xmath28 , 15@xmath28 , 25@xmath28 , 35@xmath28 , 45@xmath28 for continuum `` a '' and 5@xmath28 , 15@xmath28 , 25@xmath28 for continuum `` b '' .
the 1 @xmath28 level is shown in table 3 and slightly differs between continua `` a '' and `` b '' . the continuum `` a '' data have higher detection significance than the continuum `` b '' data , and the peak position of the continuum `` a '' emission is ( ra , dec ) = ( 04 34 00.01 , @xmath4108 34 44.9 ) in j2000 .
we define this coordinate as the nucleus of ngc 1614 in this paper .
the synthesized beams are shown as filled white circles at the bottom left of the individual figures .
, title="fig : " ] j = 43 and hnc j = 43 lines , respectively .
contours are 5@xmath28 , 15@xmath28 , 25@xmath28 , 35@xmath28 , 45@xmath28 for continuum `` a '' and 5@xmath28 , 15@xmath28 , 25@xmath28 for continuum `` b '' .
the 1 @xmath28 level is shown in table 3 and slightly differs between continua `` a '' and `` b '' . the continuum `` a '' data have higher detection significance than the continuum `` b '' data , and the peak position of the continuum `` a '' emission is ( ra , dec ) = ( 04 34 00.01 , @xmath4108 34 44.9 ) in j2000 .
we define this coordinate as the nucleus of ngc 1614 in this paper .
the synthesized beams are shown as filled white circles at the bottom left of the individual figures .
, title="fig : " ] j = 43 , hnc j = 43 , and cs j = 76 lines of ngc 1614 .
north is top , and east is to the left .
signals in channels where line emission is recognizable are integrated to produce the moment 0 maps .
contours of the moment 0 maps are 3@xmath28 , 5@xmath28 , 7@xmath28 , 9@xmath28 , and 11@xmath28 for hcn , 5@xmath28 , 10@xmath28 , 15@xmath28 , 20@xmath28 , 25@xmath28 , and 30@xmath28 for hco@xmath0 , and 3@xmath28 , 4@xmath28 , 5@xmath28 , and 6@xmath28 for hnc . for cs j
= 76 , no emission feature with @xmath443@xmath28 is seen .
the 1 @xmath28 level is summarized in table 4 .
( _ right _ ) : the molecular line spectra within the beam size at the continuum peak position are shown .
hcn j = 43 , hco@xmath0 j = 43 , and cs j = 76 line spectra are extracted at the continuum `` a '' peak position ( table 3 ) , and hnc j = 43 spectrum is extracted at the continuum `` b '' peak position ( table 3 ) .
the down arrows indicate the expected observed frequency of hcn j = 43 , hco@xmath0 j = 43 , hnc j = 43 , and cs j = 76 lines at a redshift of z = 0.016 . in the hco@xmath0 j=43 line spectrum , the observed frequency of the vibrationally excited hcn line ( v@xmath40 = 1
, j = 43 , l = 1f ; @xmath21 = 356.256 ghz ) @xcite is shown , but its detection is not clear . assuming the same line profile as hco@xmath0 j=43 , the flux of hcn j=43 at v@xmath40=1 ( l=1f ) is estimated to be @xmath90.45 [ jy beam@xmath18 km s@xmath18 ] ( @xmath93@xmath28 ) , which is @xmath928% of that of hcn j=43 at v=0 . , title="fig : " ] j = 43 , hnc j = 43 , and cs j = 76 lines of ngc 1614 .
north is top , and east is to the left .
signals in channels where line emission is recognizable are integrated to produce the moment 0 maps .
contours of the moment 0 maps are 3@xmath28 , 5@xmath28 , 7@xmath28 , 9@xmath28 , and 11@xmath28 for hcn , 5@xmath28 , 10@xmath28 , 15@xmath28 , 20@xmath28 , 25@xmath28 , and 30@xmath28 for hco@xmath0 , and 3@xmath28 , 4@xmath28 , 5@xmath28 , and 6@xmath28 for hnc . for cs j
= 76 , no emission feature with @xmath443@xmath28 is seen .
the 1 @xmath28 level is summarized in table 4 .
( _ right _ ) : the molecular line spectra within the beam size at the continuum peak position are shown .
hcn j = 43 , hco@xmath0 j = 43 , and cs j = 76 line spectra are extracted at the continuum `` a '' peak position ( table 3 ) , and hnc j = 43 spectrum is extracted at the continuum `` b '' peak position ( table 3 ) .
the down arrows indicate the expected observed frequency of hcn j = 43 , hco@xmath0 j = 43 , hnc j = 43 , and cs j = 76 lines at a redshift of z = 0.016 . in the hco@xmath0 j=43 line spectrum , the observed frequency of the vibrationally excited hcn line ( v@xmath40 = 1 , j = 43 , l = 1f ;
@xmath21 = 356.256 ghz ) @xcite is shown , but its detection is not clear . assuming the same line profile as hco@xmath0 j=43 , the flux of hcn j=43 at v@xmath40=1 ( l=1f ) is estimated to be @xmath90.45 [ jy beam@xmath18 km s@xmath18 ] ( @xmath93@xmath28 ) , which is @xmath928% of that of hcn j=43 at v=0 . ,
title="fig : " ] + j = 43 emission lines from ngc 1614 .
the velocity is in v@xmath23 @xmath24 c ( @xmath25-@xmath26)/@xmath26 .
( _ upper left _ ) : moment 1 map of hcn j = 43 emission .
contours are 4750 and 4850 [ km s@xmath18 ] .
( _ upper right _ ) : moment 1 map of hco@xmath0 j = 43 emission .
contours are 4650 , 4750 , and 4850 [ km s@xmath18 ] .
( _ lower left _ ) : moment 2 map of hcn j = 43 emission .
contours are 20 , 40 , 60 , and 80 [ km s@xmath18 ] .
( _ lower right _ ) : moment 2 map of hco@xmath0 j = 43 emission .
contours are 20 , 40 , and 60 [ km s@xmath18 ] . ,
title="fig : " ] j = 43 emission lines from ngc 1614 .
the velocity is in v@xmath23 @xmath24 c ( @xmath25-@xmath26)/@xmath26 .
( _ upper left _ ) : moment 1 map of hcn j = 43 emission .
contours are 4750 and 4850 [ km s@xmath18 ] .
( _ upper right _ ) : moment 1 map of hco@xmath0 j = 43 emission .
contours are 4650 , 4750 , and 4850 [ km s@xmath18 ] .
( _ lower left _ ) : moment 2 map of hcn j = 43 emission .
contours are 20 , 40 , 60 , and 80 [ km s@xmath18 ] .
( _ lower right _ ) : moment 2 map of hco@xmath0 j = 43 emission .
contours are 20 , 40 , and 60 [ km s@xmath18 ] . , title="fig : " ] + j = 43 emission lines from ngc 1614 .
the velocity is in v@xmath23 @xmath24 c ( @xmath25-@xmath26)/@xmath26 .
( _ upper left _ ) : moment 1 map of hcn j = 43 emission .
contours are 4750 and 4850 [ km s@xmath18 ] .
( _ upper right _ ) : moment 1 map of hco@xmath0 j = 43 emission .
contours are 4650 , 4750 , and 4850 [ km s@xmath18 ] .
( _ lower left _ ) : moment 2 map of hcn j = 43 emission .
contours are 20 , 40 , 60 , and 80 [ km s@xmath18 ] .
( _ lower right _ ) : moment 2 map of hco@xmath0 j = 43 emission .
contours are 20 , 40 , and 60 [ km s@xmath18 ] . ,
title="fig : " ] j = 43 emission lines from ngc 1614 .
the velocity is in v@xmath23 @xmath24 c ( @xmath25-@xmath26)/@xmath26 .
( _ upper left _ ) : moment 1 map of hcn j = 43 emission .
contours are 4750 and 4850 [ km s@xmath18 ] .
( _ upper right _ ) : moment 1 map of hco@xmath0 j = 43 emission .
contours are 4650 , 4750 , and 4850 [ km s@xmath18 ] .
( _ lower left _ ) : moment 2 map of hcn j = 43 emission .
contours are 20 , 40 , 60 , and 80 [ km s@xmath18 ] .
( _ lower right _ ) : moment 2 map of hco@xmath0 j = 43 emission .
contours are 20 , 40 , and 60 [ km s@xmath18 ] . , title="fig : " ] + , and hnc j = 43 emission lines . by adopting a nuclear velocity for ngc 1614 ( z = 0.016 ) of v@xmath23 = 4800 [ km s@xmath18 ] , emission with v@xmath23 @xmath3 4800 [ km s@xmath18 ] and @xmath9 4800 [ km s@xmath18 ]
is integrated for the red and blue components , respectively .
the coordinates of the emission peaks of the red and blue components are ( 04 34 00.01 , @xmath4108 34 44.3 ) and ( 04 34 00.01 , @xmath4108 34 45.5 ) in j2000 , respectively , for all the hcn , hco@xmath0 , and hnc j = 43 lines . for hcn ,
the contour starts at 0.4 [ jy km s@xmath18 ] and increases with 0.2 [ jy km s@xmath18 ] . for hco@xmath0 ,
the contour starts at 0.5 [ jy km s@xmath18 ] and increases with 0.5 [ jy km s@xmath18 ] .
for hnc , the contour starts at 0.4 [ jy km s@xmath18 ] and increases with 0.1 [ jy km s@xmath18 ] .
, title="fig : " ] , and hnc j = 43 emission lines . by adopting a nuclear velocity for ngc 1614 ( z = 0.016 ) of v@xmath23 = 4800 [ km s@xmath18 ] , emission with v@xmath23 @xmath3 4800 [ km s@xmath18 ] and
@xmath9 4800 [ km s@xmath18 ] is integrated for the red and blue components , respectively .
the coordinates of the emission peaks of the red and blue components are ( 04 34 00.01 , @xmath4108 34 44.3 ) and ( 04 34 00.01 , @xmath4108 34 45.5 ) in j2000 , respectively , for all the hcn , hco@xmath0 , and hnc j = 43 lines . for hcn ,
the contour starts at 0.4 [ jy km s@xmath18 ] and increases with 0.2 [ jy km s@xmath18 ] . for hco@xmath0 ,
the contour starts at 0.5 [ jy km s@xmath18 ] and increases with 0.5 [ jy km s@xmath18 ] . for hnc ,
the contour starts at 0.4 [ jy km s@xmath18 ] and increases with 0.1 [ jy km s@xmath18 ] . ,
title="fig : " ] , and hnc j = 43 emission lines . by adopting a nuclear velocity for ngc 1614 ( z = 0.016 ) of v@xmath23 = 4800 [ km s@xmath18 ] , emission with v@xmath23 @xmath3 4800 [ km s@xmath18 ] and @xmath9 4800 [ km s@xmath18 ]
is integrated for the red and blue components , respectively .
the coordinates of the emission peaks of the red and blue components are ( 04 34 00.01 , @xmath4108 34 44.3 ) and ( 04 34 00.01 , @xmath4108 34 45.5 ) in j2000 , respectively , for all the hcn , hco@xmath0 , and hnc j = 43 lines . for hcn ,
the contour starts at 0.4 [ jy km s@xmath18 ] and increases with 0.2 [ jy km s@xmath18 ] . for hco@xmath0 ,
the contour starts at 0.5 [ jy km s@xmath18 ] and increases with 0.5 [ jy km s@xmath18 ] .
for hnc , the contour starts at 0.4 [ jy km s@xmath18 ] and increases with 0.1 [ jy km s@xmath18 ] .
, title="fig : " ] + | we present the results of hcn / hco@xmath0/hnc j = 43 transition line observations of the nearby starburst galaxy ngc 1614 , obtained with alma cycle 0 . we find that high density molecular gas , traced with these lines , shows a velocity structure such that the northern ( southern ) side of the nucleus is redshifted ( blueshifted ) with respect to the nuclear velocity of this galaxy .
the redshifted and blueshifted emission peaks are offset by @xmath10.6@xmath2 at the northern and southern sides of the nucleus , respectively . at these offset positions , observations at infrared @xmath33 @xmath4 m
indicate the presence of active dusty starbursts , supporting the picture that high - density molecular gas is the site of active starbursts .
the enclosed dynamical mass within the central @xmath12@xmath2 in radius , derived from the dynamics of the high - density molecular gas , is @xmath110@xmath5 m@xmath6 , which is similar to previous estimates .
finally , the hcn emission is weaker than hco@xmath0 but stronger than hnc for j = 43 for all starburst regions of ngc 1614 , as seen for j = 10 transition lines in starburst - dominated galaxies . |
the discovery of quasicrystalline solid - state structures , characterized by the presence of long - range order but no translational symmetry @xcite , which can be viewed as intermediates between crystalline and amorphous structures @xcite with icosahedral @xcite , dodecagonal @xcite , decagonal @xcite , and octagonal@xcite orientational symmetry , as well as the classical wave propagation in quasiperiodic structures @xcite , renewed the interest in studying quasiperiodic systems , in particular their transport properties @xcite .
quantities commonly studied to characterize the electron dynamics in this context are the return probability @xmath1 and the mean - square displacement @xmath2 , defined as @xmath12 and @xmath13 where @xmath14 is a wave packet initially located at a site @xmath15 : @xmath16 one of the characteristic properties of quasiperiodic systems are power - law asymptotic behaviors of these two quantities in the limit @xmath17 , when @xmath18 and @xmath19 the exponents satisfy @xmath20 , where the two equalities correspond to the limiting cases of , respectively , the absence of diffusion and the ballistic motion .
inbetween these two cases anomalous diffusion takes place ( the classical diffusion is a special case @xmath21 ) , leading to the anomalous transport _ via _
einstein s relation for the zero - frequency conductivity : @xmath22 , where @xmath23 is the diffusion constant , and eq .
( [ eq : beta ] ) leads to the generalized drude formula @xmath24 , with @xmath25 being a characteristic time beyond which the propagation becomes diffusive due to the scattering @xcite .
furthermore , the shape of the diffusion front of the wavepacket has asymptotically the form of a stretched exponential @xmath26 with @xmath27 in the limit @xmath17 @xcite .
the exponent @xmath28 characterizes the spectral properties and equals the correlation dimension of the spectral measure ( the local density of states ) @xcite .
furthermore , the wavepacket dynamics exhibits multiscaling , where different powers of displacement in eq .
( [ eq : w(t ) ] ) scale with different exponents @xmath29 [ ] .
pure - point , singular continuous and absolute continuous spectra imply , respectively , @xmath30 , @xmath31 and @xmath6 , but the reverse is not necessarily true in general : @xmath32 characterize the continuous spectra , while @xmath30 usually means a pure - point spectrum and the absence of diffusion ( @xmath33 ) , but with nontrivial exceptions @xcite .
one - dimensional quasiperiodic systems commonly have singular continuous spectra @xcite , and may furthermore exhibit transitions from pure - point to absolute continuous behavior ( being singular continuous at the transition point ) as a parameter in the hamiltonian is varied , for instance in harper s model of an electron in a magnetic field @xcite , or the kicked rotator @xcite .
eigenstates of quasiperiodic systems exhibit multifractality , when the electron wavefunction is neither localized nor extended but exhibits self - similarity @xcite ( definitions and further discussion are given in sec .
[ sec : pr ] ) , similar to the anderson model of localization for disordered conductors near the localization - delocalization transition @xcite .
the hamiltonian @xmath34 studied in this work ( defined in sec .
[ sec : definitions ] ) describes a system of @xmath35 atoms of a @xmath11-dimensional product of @xmath36th approximants of the silver - mean ( octonacci ) off - diagonal model . in a previous work @xcite , the direct sum of one - dimensional quasiperiodic systems
was studied , and it was shown that already in this case @xmath1 displays non - trivial dimensionally - dependent behavior . on the other hand , it is easy to prove analytically exactly that in quasiperiodic systems defined as a direct sum of one - dimensional quasiperiodic systems , @xmath2 does not depend on the dimensionality of such tilings . in the product systems studied here
this is not the case anymore and therefore they can be viewed as a natural next choice for the construction of interesting higher - dimensional quasiperiodic systems in the study of @xmath2 .
the rest of the paper is organized as follows : in sec .
[ sec : definitions ] we define the systems studied , whose eigenproblem is discussed in sec .
[ sec : eigenproblem ] ; in secs .
[ sec : c(t ) ] and [ sec : w(t ) ] results of the calculation of @xmath1 and @xmath2 are presented and discussed , respectively ; in sec .
[ sec : pr ] participation ratios of eigenstates are analyzed , and finally , conclusions are presented in sec . [ sec : conclusions ] .
the silver - mean chain is defined over an alphabet @xmath37 by the inflation rule @xmath38 iterated @xmath36 times starting with letter @xmath39 , with the corresponding hamiltonian @xmath40 where the sum is restricted to nearest neighbors and the hopping integrals @xmath41 take values @xmath42 and @xmath43 for , respectively , letters @xmath44 and @xmath39 ( `` large '' and `` small '' hoppings ) of the letter sequence of the approximant ( with open boundary conditions ) . among various number - theoretical properties of the approximants ,
we mention that the @xmath36th approximant s length @xmath45 satisfies @xmath46 , the so - called silver mean .
the @xmath11-dimensional quasiperiodic tilings we study are derived from the direct product of silver - mean chains @xcite with the hamiltonian @xmath47 this has a particularly simple geometrical interpretation , since it describes a particle on a @xmath11-dimensional cube , with coordinates @xmath48 , hopping only along the main diagonals of the cube : @xmath49 . since the nearest neighbors of the cube can not be connected by any number of such hops from @xmath50 , the hamiltonian ( [ eq : hamiltonian ] ) decomposes into @xmath51 parts defined on corresponding interpenetrating disconnected quasiperiodic tilings .
it can be shown that all of these hamiltonian parts are equivalent , and that the tilings they define are symmetry related , due to the reflection symmetry of the silver - mean chain about the middle bond of the chain . ) for @xmath52 .
the distances @xmath44 and @xmath39 of the silver - mean chain are taken as @xmath42 and @xmath53
. then the bond lengths of the labyrinth tiling are @xmath54 and @xmath55 .
the bipartite structure is indicated by the open and closed circles for the sites , taking into account that hopping occurs only among the sites of different kind . the second , equivalent , labyrinth tiling is shown in the figure by dots.,width=288 ] such tilings in @xmath56
are known as labyrinth tilings , and figure [ fig : labyrinth ] shows one example , an @xmath57 tiling .
the tiling has bipartite structure .
furthermore , a two - dimensional projection of the silver - mean tiling in @xmath58 , for instance , looks exactly as the tiling in figure [ fig : labyrinth ] , because the whole tiling has a layered structure : all symbols of the same kind in the figure belong to one layer .
the distance between successive layers are , of course , again determined by the silver - mean sequence .
the number of sites @xmath59 in various tilings is given in the table [ table : nm ] .
.the number of sites @xmath59 of the @xmath11-dimensional tilings of various approximants @xmath36 of the silver - mean ( octonacci ) quasicrystal . [ cols=">,>,>,>",options="header " , ]
the eigenproblem of @xmath60 is @xmath61 to which the eigenproblem of @xmath50 reduces : @xmath62 regarding the subtiling eigenproblem , its eigenstates are obtained simply by projecting the eigenstates of @xmath50 onto the subtiling ( with normalization constant @xmath63 ) due to the equivalence of the individual eigenproblems ( [ eq:1d ] ) .
thus , if @xmath64 denotes the projector onto the subspace of one subtiling , the corresponding subtiling hamiltonian is @xmath65 and its eigenstates are @xmath66 with the same eigenenergies as in eq .
( [ eq : h_m^deigenproblem ] ) .
the reduction of the problem to quasiperiodic subtilings of the product tiling therefore only reduces the degeneracy of individual eigenstates of @xmath67 , while the set of distinct eigenenergies is exactly the same . furthermore , since all the hamiltonians studied here have only off - diagonal matrix elements , their eigenenergies necessarily come in opposite - energy pairs , and thus we can restrict our discussion to the spectral properties for states with @xmath68 . figure [ fig : e(v ) ] shows the dependence of the eigenenergies on the parameter @xmath3 .
a characteristic feature of the spectra is the existence of gaps of various sizes . in @xmath69 ,
the gap - labeling theorem gives an enumeration of all the possible gaps in the spectrum @xcite .
for @xmath70 these gaps seem to close for a sufficiently large @xmath3 , and fig .
[ fig : e(v ) ] suggests that gaps persist only for @xmath71 in @xmath7 , respectively .
[ fig : e(v ) ] for @xmath70 the gaps close for smaller values of @xmath3 due to the level crossings that are a consequence of the factorizability of the eigenenergies in eq .
( [ eq : h_m^deigenproblem ] ) . due to this
we expect that the spectrum will acquire finite lebesgue measure and change from from fractal to continuous for intermediate values of @xmath3 , similar to the results of ref .
a more difficult question , that we address in the next section , is whether the spectrum is singular or absolute continuous or a mixture of the two in different regions of the spectrum for various values of @xmath3 . to this end
we first note that @xmath72 states corresponding to @xmath73 for @xmath74 are a consequence of the open boundary conditions used in this work , corresponding to the electron localized , for small @xmath3 , at the ends of the chain ( in @xmath69 and corners of the cube for @xmath70 ) , and spreading across the system as @xmath75 . in this respect
the systems studied here are different from those of ref .
[ ] , where two @xmath76 states have been induced by adding two sites at each end of the chain with @xmath77 .
however , for the results presented here this difference in boundary conditions does not play any important role .
in this section we study the scaling of the return probability @xmath1 by evaluating numerically eq .
( [ eq : c(t ) ] ) for the eigenstates of a silver - mean tiling in @xmath0 constructed , as described in the previous section , from eigenstates of the chain , obtained from the numerical diagonalization of the @xmath60 , and taking into account that the time evolution is given by @xmath78 .
we present the results in fig .
[ fig : c(t ) ] . for small values of @xmath3
there are pronounced steps , and this limit we analyze elsewhere @xcite . for each of the studied cases in the figure , there are three characteristic time - scales of the particle dynamics : ( i ) short times , characterized by a high probability to find the particle at the initial site , which corresponds to the regime when the wavepacket has only begun spreading ; ( ii ) intermediate times , when @xmath1 does seem to behave according to the power law ( [ eq : delta ] ) ; and ( iii ) the long - time limit when there is a crossover into a constant @xmath1 , due to the finite spatial extent of the studied systems .
the quantitative determination of @xmath28 is rather difficult , due to the large subdominant contributions to the asymptotic power - law behavior @xcite .
these appear through the dependence of @xmath28 on the interval of @xmath79 values taken for the least mean - squares linear fit in the log - log plot of @xmath1 .
lacking rigorous results , one attempt to characterize these corrections to eq .
( [ eq : delta ] ) would be to consider a class of power - law subdominant corrections by assuming @xmath80 it was noted in ref .
[ ] that , in the case of the fibonacci model , @xmath1 can be accurately fitted to eq .
( [ eq : subdominant ] ) under the assumption that @xmath81 is exactly equal to @xmath42 , and the authors gave some plausible but non - rigorous arguments for this assumption . on the other hand , numerical verification of eq .
( [ eq : subdominant ] ) by a least - squares nonlinear fit requires that @xmath1 is calculated for many decades of @xmath82 .
in addition , verification of @xmath83 by such a procedure would be particularly difficult near the transition where the fit would be expected to yield already @xmath84 , _
i.e. _ to decompose the fitted function into two nearly identical terms .
finally , the parameter @xmath3 would be expected to play some role also in the subdominant terms , for instance through the dependence of the value of @xmath81 on @xmath3 .
an additional difficulty was pointed out in ref .
[ ] , where the scaling exponent of the second moment of the spectral measure was considered ( which is known @xcite to be equal to @xmath28 ) for several quasiperiodic systems , including fibonacci chains , and an apparently irregular behavior of the moment was found in all of the non - fibonacci models .
this led the authors to use a fitting procedure that does not involve fitting of a straight line and even to consider the possibility that the exponent @xmath28 might not exist in these non - fibonacci systems .
in this work we fit the power law ( [ eq : delta ] ) to the calculated values of @xmath85 ( for many values of @xmath86 equidistant in the logarithmic scale ) by fitting of the expression @xmath87 in all intervals containing @xmath86 values spanning at least one decade in @xmath79 ( for easier comparison , this corresponds to a change in @xmath82 of at least @xmath88 in the plots ) , and selecting among the obtained fits of @xmath28 those that have @xmath89 values closest to @xmath90 .
should several distinct values of @xmath28 occur that all have small @xmath89 , it seems reasonable that those values corresponding to later times are closer to the correct asymptotic value of @xmath28 , just as it is reasonable to expect that for larger @xmath79 values eq .
( [ eq : subdominant ] ) becomes close to the power law ( [ eq : delta ] ) .
our fitting procedure does not assume anything in particular about the subdominant terms of the powerlaw asymptotic behavior , but rather estimates @xmath28 by requiring the absence of certain types of subdominant terms , as quantified by @xmath89 .
we show such obtained fits in fig .
[ fig : c(t)fitted ] for the case @xmath91 .
this is the most difficult system among the three cases presented in fig .
[ fig : c(t ) ] because the duration of the intermediate time regime ( ii ) , where the power law asymptotic behavior should be most likely expected , depends only on the linear size of the system , determined by @xmath36 , and not on the dimensionality @xmath11 .
the obtained values of @xmath89 in the fits presented in fig .
[ fig : c(t)fitted ] range from @xmath92 to @xmath93 .
( points ) together with fitted lines close to the power law from the procedure described in the text.,width=336 ] a characteristic feature of the fits is that for @xmath94 , the time intervals where the power law is obeyed best ( _ i.e. _ where @xmath89 is smallest ) are shorter than the intervals for @xmath95 , even though for large @xmath3 the curves for @xmath96 do not show oscillatory behavior and appear to be more similar to the straight lines .
there is , however , a small curvature in the dependence of @xmath96 on @xmath82 for the crossover times inbetween regimes ( i ) and ( ii ) as well as ( ii ) and ( iii ) .
this is hard to see in the plots , since it involves large relative changes in small curvatures , but easy to distinguish numerically in our fitting procedure .
the results for @xmath28 obtained from the above described fitting procedure applied to the results from fig .
[ fig : c(t ) ] are presented in fig .
[ fig : delta ] . the plot for @xmath56 shows @xmath97 also for @xmath98 , which is systematically larger than @xmath97 of the @xmath99 system for @xmath100 and smaller for @xmath101 , from which we speculate on the possible form of the limiting curve of the infinite system as indicated in the figure for @xmath7 .
this extrapolation of @xmath97 in @xmath7 is in agreement with semi - quantitative considerations of this limit in ref .
for @xmath69 the results suggest that @xmath5 for all @xmath102 also in the infinite system .
the exponents thus suggest that there is a singular continuous spectrum for all @xmath3 values in @xmath69 and a possible transition from singular continuous to absolute continuous spectra at @xmath103 in @xmath56 ( in agreement with a previous estimate @xcite ) , and at @xmath104@xmath105 in @xmath58
. a difficulty in obtaining exact values of @xmath106 could be due to the possibility that , for @xmath107 , the spectrum is a mixture of singular continuous and absolute continuous parts , when absolute continuous parts appear in only small intervals of the full spectrum that grow larger with increasing @xmath3 , and when the @xmath28 values obtained for @xmath3 away from @xmath106 are closer to the correct asymptotic values .
a previous study @xcite of the propagation of the projection of the wave packet ( [ eq : deltawave ] ) onto various segments of the spectrum of the silver - mean model in @xmath56 found that @xmath1 of such restricted wave packets still obeyed the power law , with exponents differing only slightly from @xmath28 obtained when the full spectrum is used .
we were able to extend this result by proving the following : if the spectrum is divided into a finite number of @xmath108 segments such that @xmath109 , where @xmath110 is the projector onto the subspace corresponding to the segment @xmath111 , and if @xmath112 in each of the @xmath108 segments , where @xmath113 is defined as in eq .
( [ eq : c(t ) ] ) with the wavepacket ( [ eq : deltawave ] ) replaced by its projection @xmath114 , and if eq .
( [ eq : delta ] ) also holds , then @xmath115 .
the main advantage of calculating the width @xmath2 of a wave packet is that it is more directly related to the transport properties than @xmath1 , as discussed in the introduction , and that its asymptotic power - law behavior ( [ eq : beta ] ) seems to be less influenced by subdominant contributions , and therefore easier to determine from numerical studies of finite - size systems , as we discuss below .
one of the main difficulties in numerical studies of the anomalous diffusion is that in order to obtain @xmath2 accurately , many eigenvectors are needed . calculating its behavior at large times also requires large system sizes .
this puts significant constraints on the system sizes that can be investigated numerically , and several approximations have been developed to circumvent these difficulties .
one of them is to expand the evolution operator in terms of small - time increments @xcite , or to study the time evolution of the position operator expanded in terms of chebyshev polynomials combined with energy filtering _ via _ a gaussian operator centered at a given energy and of the width of , _
e.g. _ , @xmath116 of the total bandwidth @xcite . in the latter work ,
generalized quasiperiodic rauzi tilings have been studied and the authors were interested how the topological connectivity of the tiling influences transport properties of the quasicrystal in @xmath7 .
the simplifying feature of the model is that the hamiltonian is sparse with no matrix elements different from @xmath42 , which significantly speeds up any algorithm where multiplication with the hamiltonian is a time - consuming part of the calculation , like the algorithm of [ ] as well as , for instance , the lanczos algorithm .
another approach was to study the evolution of a wave packet @xmath117 constructed from a subset of eigenstates from an interval @xmath111 of the full spectrum@xcite , @xmath118 using such wave packets , the largest system studied in this work is @xmath119 for @xmath58 with 3456000 sites .
figure [ fig : w(t ) ] presents results of the calculations of the spreading width @xmath2 of a wave packet initially localized on a single site using all the eigenstates of the spectrum . for small @xmath3 values
pronounced steps appear , related to the similar steps in fig .
[ fig : c(t ) ] and the structure of very narrow bands of the spectra ( cp . fig .
[ fig : e(v ) ] ) , that we discuss elsewhere @xcite .
the three regimes of wave - packet dynamics discussed in the previous section here correspond to ( i ) the nearly ballistic propagation for early times @xmath120 , ( ii ) anomalous diffusion , in intermediate intervals of @xmath79 , characterized by eq .
( [ eq : w(t ) ] ) and ( iii ) a stationary regime of approximately constant @xmath2 , due to the finite extent of the studied systems , for large @xmath79 .
the maximum of @xmath121 which would be possible for a system of linear size @xmath44 can be estimated as @xmath122 , corresponding to a state completely localized at the corners of the sample ; for the system sizes in fig .
[ fig : w(t ) ] , maximal values of @xmath123 are approx .
@xmath124 in @xmath0 , respectively . on the other hand
, a completely delocalized state would lead to @xmath125 , respectively ; thus fig .
[ fig : w(t ) ] shows that for large values of @xmath3 the wave packet becomes smeared out over the entire sample in the regime ( iii ) .
the crossover from regime ( ii ) to ( iii ) becomes longer for larger @xmath11 , which is due to the more complicated reflection of the wave packet off the boundaries in higher - dimensional tilings .
the values of the exponent @xmath29 , which were obtained using the same fitting procedure as in the previous section , are presented in fig .
[ fig : beta ] .
they seem to be independent of the dimensionality of the system .
since the calculation of @xmath2 is computationally slower than that of @xmath1 , system sizes considered here are smaller than in the previous section , and in particular , for @xmath126 , the power law was observed for only about a half a decade of time for @xmath127 , as opposed to two decades for @xmath128 , and we could not obtain a reliable exponent using our fitting procedure for @xmath129 . in the fitting procedures used in the previous section , for comparison ,
the calculated exponents were obeyed for at least a decade ( and up to 6 decades in some cases for smaller @xmath3 values ) . governing anomalous diffusion of the wave packet , as extracted from the data in fig .
[ fig : w(t ) ] .
the error bars are of the size of the symbols .
, width=288 ] in order to assess the role of topology of the quasiperiodic tiling in @xmath58 , we make a comparison of the wave - packet dynamics studied here with those of ref .
[ ] . since in the latter work energy - filtering was used as described above , we study the spreading of the wave packet ( [ eq : deltawave ] ) composed only of eigenstates with energy @xmath130 in the interval @xmath131 , where @xmath132 is chosen such that the interval contains about @xmath133 of the total number of states .
figure [ fig : d(t)part ] presents calculated values for several energies @xmath134 between the band edge and the band center .
apparently when the wave packet is composed only of band - edge states , the diffusion is anomalous with the exponent very close to the value @xmath135 obtained when the full spectrum is considered , as indicated in the figure . for other energies
, there seems to be an intermediate time regime when anomalous diffusion also takes place , with the same or smaller exponents as compared to the one found at the band edge .
furthermore , a third kind of behavior also occurs , most notably at the band center , where the wave packet seems to spread slower than any power law .
this is in qualitative agreement with the findings of [ ] in the sense that there wave packets made out of states from the band edge also spread faster but , in contradistinction with the finding here , ballistic motion was found at the band edge and anomalous diffusion at the band center .
it is unclear to us at present what kind of eigenstates are responsible for the slow diffusion seen at the band center .
qualitatively , marginally critical states @xcite might be perhaps related to this kind of wavepacket dynamics .
such constructed wavepackets will inevitably have some `` ripples '' that depend on how well eigenstates from the chosen interval @xmath111 approximate @xmath136 , which in turn depends on the choice of the starting site @xmath137 .
this can be characterized by @xmath138 , and states in fig .
[ fig : d(t)part ] have @xmath139 and @xmath140 at the band edge and the band center , respectively . in order to show some evidence that the obtained dynamics in the two cases does not depend on the choice of the initial site
, we present @xmath2 in fig .
[ fig : d(t)partb ] when the initial site is chosen two sites along the main diagonal of the system away from the center , in which case @xmath141 and @xmath142 at the band edge and the band center , respectively .
the dependence of the values of @xmath143 on the choice of the initial site is reflected in the range of values that @xmath2 takes , but does not seem to affect the two different types of dynamics . although the contribution @xmath143 of the band edge states to the wave packet in fig .
[ fig : d(t)partb ] is much smaller than to that in fig .
[ fig : d(t)part ] , the spreading of the entire wave packet is again characterized by the slope @xmath135 in agreement with the value shown in fig .
[ fig : beta ] .
the results of the previous section strongly suggest that @xmath10 is independent of @xmath11 for the class of systems @xmath50 studied here . in this section
we attempt to link that with the eigenstate properties by the analysis of the participation number .
the ( generalized ) inverse participation numbers , @xmath144 , are defined as @xmath145 the participation number @xmath146 for instance , characterizes on how many sites a given state @xmath147 is significantly different from zero : for a particle completely localized at a single site @xmath148 while for a bloch wave @xmath149 .
the participation ratio , @xmath150 , gives the fraction of the total number of sites where @xmath147 is significantly different from 0 . in the case of the eigenstates of @xmath151 , we have @xmath152 or , in other words , @xmath153 of the product state is equal to the product of the inverse participation numbers of the one - dimensional states . on the other hand ,
the inequality @xmath154 gives us insight into the nature of eigenstates of @xmath50 : they are always less or equally extended than the most extended state of the product , and more or equally extended than the least extended state of the product . in particular , they can be anisotropically extended , which happens whenever states from the product have quite different participation numbers . for each such state of @xmath151
, there is another one with the same energy and anisotropy in a different direction due to the product structure of the hamiltonian .
a previous numerical study @xcite indicates that the scaling exponents of @xmath155 of the ground state and the band center state are only slightly different . here
, we investigate the scaling of the participation ratio @xmath156 averaged over the whole spectrum as a function of @xmath45 and @xmath3 in @xmath69 , and the results are presented in fig .
[ fig : pr ] . on the number of sites @xmath45 of the silver - mean chain for @xmath157 in a doubly - logarithmic plot for @xmath158 .
the inset shows the values of the exponent @xmath159 obtained from the fit of @xmath160.,width=288 ] the results indicate that the average participation ratio scales with a power law in the number of sites of the chain , @xmath161 and the inset of fig .
[ fig : pr ] shows the values of @xmath159 obtained from the least mean - squares linear fit to the data for the largest system sizes in the figure .
the results suggest that @xmath162 , which is not surprising since in this limit the quasiperiodic tiling becomes a periodic chain ( we note that in our case @xmath163 because we have real instead of complex eigenstates so that the bloch waves have nodal structure ) . on the other end , however , @xmath164 , even though for @xmath74 the chain breaks up into clusters of atoms , and therefore @xmath165 .
this particular limit we address elsewhere @xcite .
for @xmath70 we have also calculated @xmath159 numerically for several @xmath3 values and it remains the same within the error bars , which we attribute to the multiplicative nature of the participation numbers as reflected in eqs .
( [ eq : productzq ] ) and ( [ eq : inequality ] ) .
this can be related to the independence of @xmath10 with respect to the dimensionality determined in the previous section .
we studied spectral properties , dynamics of wave packets and scaling properties of eigenstates in 1- , 2- and 3-dimensional systems obtained as direct products of the silver - mean chains , by investigating the scaling exponents @xmath166 and @xmath167 describing the asymptotic properties of , respectively , the return probability , the spreading of the wave packet , and the average participation ratio of eigenstates .
the obtained values of @xmath28 are compatible with the spectral measure being singular continuous in one dimension and undergoing a transition from singular to absolute continuous in @xmath7 with large subdominant contributions
. the latter would appear ( if such a transition indeed occurs ) as a systematic shift of the value of @xmath97 when the system size increases .
the comparison of the results for @xmath97 for @xmath98 , and @xmath168 in the two - dimensional systems ( cp .
[ fig : delta ] ) shows such a shift of the order of up to @xmath169 near the speculated transition point @xmath103 and thus corroborates such conclusion . a more quantitative characterization of these corrections , such as , for instance , whether they are of the form ( [ eq : subdominant ] ) , is beyond the scope of this work .
the obtained values for @xmath29 are independent of the dimensionality , which we have linked to the properties of the inverse participation numbers . even though these results are certainly related to the product structure of the quasiperiodic systems studied here , it is quite possible that they describe some features of generic higher - dimensional quasiperiodic systems .
in particular , we found an exact relation among the exponent @xmath28 and corresponding exponents @xmath170 of the dynamics of the projection onto subintervals of the spectrum of a wave packet initially localized at a single site , and compared the dynamics of wave packets in the silver - mean quasiperiodic tiling constructed from about @xmath133 of the total number of states near various energies in the band with a study of generalized quasiperiodic rauzi tilings in ref .
. there seems to be a similarity between the two in so far that wave packets of states near the band edge are spreading much faster than those made out of the states near the band center .
however , while in the latter work dynamics of these wave packets is , respectively , ballistic and anomalously diffusive , we find that the dynamics is , respectively , anomalously diffusive and , for the wave packets made out of band - center states , slower than any anomalous diffusion . c. tang and m. kohmoto , phys .
b * 34 * , 2041 ( 1986 ) ; m. kohmoto , b. sutherland , and c. tang , phys . rev .
b * 35 * , 1020 ( 1987 ) ; st , a. , j. stat . phys . *
56 * , 525 ( 1989 ) ; j. bellissard , b. iochum , e. scoppola , d. testard , commun . math . phys . * 125 * , 527 ( 1989 ) ; d. damanik and d. lenz , commun . math . phys . * 207 * , 687 ( 1999 ) . | spectral properties and anomalous diffusion in the silver - mean ( octonacci ) quasicrystals in @xmath0 are investigated using numerical simulations of the return probability @xmath1 and the width of the wave packet @xmath2 for various values of the hopping strength @xmath3 . in all dimensions
we find @xmath4 , with results suggesting a crossover from @xmath5 to @xmath6 when @xmath3 is varied in @xmath7 , which is compatible with the change of the spectral measure from singular continuous to absolute continuous ; and we find @xmath8 with @xmath9 corresponding to anomalous diffusion .
results strongly suggest that @xmath10 is independent of @xmath11 .
the scaling of the inverse participation ratio suggests that states remain delocalized even for very small hopping amplitude @xmath3 .
a study of the dynamics of initially localized wavepackets in large three - dimensional quasiperiodic structures furthermore reveals that wavepackets composed of eigenstates from an interval around the band edge diffuse faster than those composed of eigenstates from an interval of the band - center states : while the former diffuse anomalously , the latter appear to diffuse slower than any power law . |
light echoes from transient events provide a probe of both the circumstellar and interstellar structures in the environment ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , of the size distribution and chemical composition of the scattering dust ( e.g. , * ? ? ?
* ) , and of the detailed history of the outburst giving rise to the echo ( e.g. , * ? ? ?
echoes also could provide a geometric distance to the event , independent of the distance ladder , through analysis of polarized light @xcite .
a number of resolved light echoes have been discovered historically , including those of novae @xcite , unusual outbursting stars @xcite , erupting luminous blue variables @xcite , and ancient galactic and local group supernovae ( sne ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
see @xcite for a review .
although light echoes from more recent extragalactic sne could well be a common occurrence , observations which spatially resolve echoes are relatively rare , with only 11 occurrences known in total , as summarized in table [ tabhist ] .
inferences have been made for the existence of possible light echoes around a number of other sne , mostly from reemission of sn light by dust in the infrared @xcite .
the main obstacle to resolving the echo is that the sn host galaxies must be relatively nearby , although , even then , the structures can only be revealed by the superior angular resolution of the _ hubble space telescope _ ( _ hst _ ) . the light echo of a sn results from the luminous ultraviolet ( uv)/optical pulse scattered by dust in the sn environment .
we essentially see a reflection of the event itself @xcite .
in particular , if the light echo is detected in the uv , we have a record of the bright , rapid flash of x - ray / uv emission emerging as the sn shock breaks through the massive envelope around the progenitor star @xcite , an event that is elusive on account of its promptness and brevity ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
much of the circumstellar matter and small dust grains nearest to the sn will be destroyed by the pulse , although more distant , larger grains will survive .
the observed light echo at a particular instant in time is the intersection of a dust sheet or filament with the ( virtual ) ellipsoid surface of constant arrival time , which results in nearly all cases as a circle or arc on the sky .
we have discovered a resolved light echo around the normal sn ii - plateau ( ii - p ) 2012aw in the sb(r)b spiral host galaxy messier 95 ( m95 ; ngc 3351 ) , during the course of the legacy extragalactic uv survey ( legus ; go-13364 , pi : d. calzetti ) .
legus is a cycle 21 _ hst _ treasury program which imaged 50 nearby ( @xmath3 mpc ) galaxies in multiple bands with the wide field camera 3 ( wfc3 ) uvis channel and the advanced camera for surveys ( acs ) wide field channel .
see @xcite for a detailed introduction to the survey .
legus particularly concentrated on obtaining images of these well - studied galaxies in the wfc3/uvis f275w and f336w bands .
the sn has been extensively studied in the uv , optical , and near - infrared by @xcite , @xcite , and @xcite .
@xcite analyzed radio observations of the sn .
the sn progenitor was identified in archival _ hst _ wide - field and planetary camera 2 ( wfpc2 ) f555w and f814w and ground - based @xmath4 and @xmath5 images as a likely red supergiant ( rsg ) star by both @xcite and @xcite .
both @xcite and @xcite found that the rsg was cool ( @xmath6 ; although , @xcite more narrowly limited @xmath7 to @xmath8 k ) , must have had a dusty circumstellar environment , and was therefore quite luminous ( @xmath9 ) and of relatively high initial mass ( @xmath10@xmath11 ) , based on the stellar evolutionary models from @xcite and @xcite , respectively .
@xcite reanalyzed the data from both papers , employing a different approach for the treatment of the circumstellar extinction , and argued that the progenitor had to have been less luminous ( @xmath12 ) and less massive ( @xmath13 ) .
van dyk et al .
( in preparation ) have revisited the progenitor photometry and conclude that the star likely had @xmath14 and @xmath15 .
@xcite modeled late - time spectra from the nebular phase and found that the emission lines from likely nucleosynthetic products indicate a progenitor initial mass in the range 14@xmath16 . in this paper
we describe the observations and analysis of the light echo and the ramifications for the properties of the sn and its progenitor .
we assume a reddening- and metallicity - corrected distance modulus to the host galaxy of @xmath17 mag ( measured using cepheids ; * ? ? ?
* ) , which is a distance of @xmath18 mpc .
the inclination of m95 is relatively low , at @xmath19 , with position angle @xmath20 @xcite , so internal line - of - sight effects should be minimal .
the sn occurred @xmath21 w and @xmath22 s of the host galaxy nucleus , in a region of the galaxy with no obvious signs of recent massive star formation . @xcite
adopted 2012 march 16.1 ( jd 2456002.6 @xmath23 0.8 ) as the time of the sn explosion .
@xcite adopted march 16.0 ( jd 2456002.5 @xmath23 0.8 ) .
these are essentially the same for the purposes of our analysis below .
ut dates are used throughout .
ccccc 1987a & ii - p & lmc & 0.05 & 1,3,15 + 1991 t & ia & ngc 4527 & 15 & 9 + 1993j & iib & m81 & 3.4 & 7,10 + 1995e & ia & ngc 2441 & 49.6 & 8 + 1998bu & ia & m96 & 9.9 & 2 + 2003gd & ii - p & m74 & 10.0 & 11,12 + 2006x & ia & m100 & 15.2 & 14 + 2007af & ia & ngc 5584 & 23 & 5 + 2008bk & ii - p & ngc 7793 & 3.4 & 13 + 2012aw & ii - p & m95 & 10.0 & 16 + 2014j & ia & m82 & 3.5 & 4 +
m95 was initially observed by _ hst _ for legus with wfc3/uvis on 2014 march 1 , however , guide star tracking failed .
the galaxy was reobserved on april 23 in bands f275w ( total exposure time 2361 s ) , f336w ( 1062 s ) , f438w ( 908 s ) , f555w ( 1062 s ) , and f814w ( 908 s ) .
the pointing for the observations was designed intentionally to include the site of sn 2012aw in the total field - of - view .
the imaging data were initially run through the standard stsci pipeline procedures for calibration .
however , as described in @xcite , further processing of the ` flt ' images included a pixel - based correction for the charge transfer efficiency ( cte ) losses .
the corrected ` flc ' images were then combined using the drizzlepac software @xcite to produce a distortion - corrected mosaic in each band . the midpoint of the wfc3 observations was at 2014 april 23 5:28:06 , or jd 2456770.7 .
this is at a sn age of @xmath24 d , or 2.10 yr .
a three - color - composite image of the light echo is shown in figure [ figechocolor ] .
the echo is conspicuously blue , which is not surprising , since we expect dust to scatter preferentially the blue light from the sn .
individual images of the echo are shown for all five bands in figure [ figecho ] .
the appearance of the echo is more of a circular patch , or blob , and less obviously a single ring .
if the echo were a thin ring , though , its profile would be convolved with the wfc3/uvis psf and would not be sharply defined .
it is also entirely possible that we are actually seeing multiple rings , such as was observed for the interstellar dust - scattered echo around sn 1987a ( * ? ? ?
* see his figure 1 ) , just not as well resolved .
we know that the echo is a recent , post - sn apparition , since no evidence of it exists in the pre - explosion _ hst _ f555w and f814w images ( see @xcite , their figure 1 , and @xcite , their figure 2 )
. using 20 fiducial stars in common between the pre - sn f814w image and the legus wfc3 f814w image in the immediate sn environment , we can precisely associate the position of the sn progenitor star with the bright spot seen within the echo in figure [ figecho](e ) , with a rms uncertainty of @xmath25 ( in right ascension ) and @xmath26 ( in declination ) . there is little question that this is the sn itself , still prominently seen in f814w , relative to the fainter echo brightness .
the sn is also apparent in f438w and f555w , but less so in f275w and f336w .
a surface brightness enhancement to the southeast along the echo tends to dominate the echo s appearance from f555w blueward .
this implies that the scattering dust is not uniformly distributed , which is consistent with the filamentary nature of interstellar dust and its subsequent effect on the structure of the observed light echoes ( e.g. , * ? ? ?
* ) .
we performed aperture photometry within iraf of the entire light echo on the drizzled image mosaics in each band , through a range of apertures with increasing radii , in order to estimate the uncertainty in the measurement .
a sky annulus was used to estimate the background , which was relatively low in each band .
the vegamag zero points for an infinite aperture for the wfc3 bandpasses were adopted .
we list our photometry for the echo in table 2 .
the flux of the echo , @xmath27 , which is also listed in the table for each band , was established using these wfc3 zeropoints .
we emphasize that the photometry of the echo in table 2 actually includes both the echo _ and _ the sn , particularly at the redder bands .
to fully analyze the properties of the echo we needed to remove the sn flux from the total .
this was the most complicated step in our analysis and involved significant uncertainty .
we started with the three f814w exposures , since the sn was brightest relative to the echo in that band . even in f814w , though , the emission surrounding the sn is not necessarily spatially flat , so defining a sky background around the point source is nearly impossible .
we searched for the location and flux of the sn point source with the understanding that the sn should be the sharpest feature in the image and the nebulous background should be relatively smooth . to this end , we extracted a 19@xmath2819-pixel raster about the target in each flc exposure .
we then explored an array of positions and fluxes for the point source . for each trial position and flux , we subtracted the corresponding point - spread function ( psf ) from the raw exposures using a library psf tailored for the particular filter at the particular location in the detector .
these library psfs were constructed from a well - dithered set of data taken of the globular cluster omega centauri soon after wfc3 was installed on - board _ hst _ ; they have proven to be good to better than @xmath292% .
we then examined the residuals and identified the optimal location and flux for the sn to be the one that left the post - subtraction residuals as smooth as possible . to evaluate the smoothness of each residual image
, we compared that image with a smoothed version of itself .
the smoothed version is obtained by convolving the image with a 5@xmath285 quadratic smoothing kernel , which is equivalent to fitting for a six - parameter , two - dimensional quadratic centered on each pixel . in essence , this operation reduces the number of degrees - of - freedom by a factor of 25/6 .
the image that was closest to its smoothed version was deemed to be the smoothest .
the optimal location and flux of the point source was deemed to be the one that left the image as close as possible to the smooth version of itself . using the above strategy
, we determined a best - fit location and flux for the sn in each of the f814w flc exposures .
when the positions were transformed into the reference frame , they agreed to within a rms uncertainty of 0.09 pixel in @xmath30 and 0.03 pixel in @xmath31 . using this best - fit position
, we then found the flux of the point source in each f814w exposure that made the resulting image as smooth as possible and arrived at a flux of 650@xmath2375 e- for an exposure time of 359 s. the psfs were normalized to correspond to the flux within a 10-pixel aperture , so that is the effective aperture for these measured fluxes .
we applied a similar approach to determine a best - fit flux for the sn in each of the other filters ( 850@xmath23150 e- in f555w , 400@xmath2345 e- in f438w , 50@xmath2375 e- in f336w , and 100@xmath2350 e- in f275w ) , using the master - frame position fit from the f814w exposures .
we used the same psfs to measure fluxes for a number of other isolated point sources in the images , to tie this photometry to the vega magnitudes for stars in the images produced by dolphot @xcite .
the sn flux was then similarly scaled in each band ; the resulting magnitudes are listed in table 2 .
the uncertainties for the sn magnitudes in each band also include ( via quadrature sums ) the uncertainties in the scaling , although these are relatively small , compared to the uncertainties due to the smoothed model fitting . finally , we list in table 2 the resulting flux , @xmath32 , obtained from subtracting the sn flux from our aperture measurements of the echo flux
. the fractional contribution of the sn light to the total echo flux is @xmath33% in f275w , @xmath34% in f336w , @xmath35% in f438w , @xmath36% in f555w , and @xmath37% at f814w .
furthermore , we had @xmath38-supersampled and coadded the individual flc exposures in each band , with the sn flux removed , to produce images of just the light echo . in figure [ figsuper ]
we show the f275w , f438w , and f555w images summed together , in order to increase the overall signal - to - noise ratio of the fainter features of the echo .
we did this primarily to reveal any ring - like structure .
as one can see , the shape of the echo is still quite patchy and irregular , with the southeast enhancement still dominating the emission , although hints of a ring , or partial arcs , are possibly visible .
ccccc f275w & @xmath39 & @xmath40 & @xmath41 & @xmath42 + f336w & @xmath43 & @xmath44 & @xmath45 & @xmath46 + f438w & @xmath47 & @xmath48 & @xmath49 & @xmath50 + f555w & @xmath51 & @xmath52 & @xmath53 & @xmath54 + f814w & @xmath55 & @xmath56 & @xmath57 & @xmath58 +
as pointed out , it is not clear from figure [ figsuper ] that what we are seeing is a well - defined ring , with an expected central void indicating a lack of dust in the plane of the sn .
we therefore can not exclude at this point that the sn is actually immersed in a diffuse dust cloud . nonetheless , we determined the radius of any ring - like structure by taking cross - cut flux profiles through the echo from its center , avoiding the bright surface brightness enhancement .
see figure [ figcuts ] .
we estimated the radius to be 3.9 pixels , with a conservative uncertainty of @xmath59 pixel . at the uvis plate scale
, this radius corresponds to @xmath60 .
we show our estimate of a ring in figure [ figsuper ] .
we can approximate the light echo ellipsoid as a paraboloid , since the sn s distance from the earth , @xmath61 , is so much larger than the geometric dimensions of the echo ( e.g. , * ? ? ?
* ; * ? ? ?
the perpendicular linear distance of the line - of - sight to the sn from the line - of - sight to the echo is @xmath62 , where @xmath63 is the angular distance between the two lines - of - sight . at the distance @xmath61 to m95 , @xmath64 pc .
the distance from the sn to any scattering dust element along the echo , @xmath65 , can be obtained from @xmath66 , where @xmath67 is the distance from the sn to the echo along the line - of - sight @xcite . for @xmath68 pc ( 2.10 ly )
, we find that @xmath69 pc and @xmath70 pc , with an uncertainty of @xmath71 pc , resulting from the uncertainties in the estimates of both @xmath63 and @xmath61 .
the echo is therefore most likely due to interstellar dust .
we note that much or all of any pre - existing circumstellar dust was likely destroyed by the sn x - ray / uv flash @xcite . in modeling the echo , even with the uncertainties in the actual distribution of the diffuse interstellar dust
, we have assumed that the echo arises from single scattering in a thin dust sheet located between us and the sn , and that the sheet thickness is much smaller than the distance between the sheet and the sn .
the scattered flux @xmath72 at time @xmath73 from the echo at a given wavelength or bandpass is then @xmath74 where @xmath75 is the fluence of the sn at time @xmath76 , and @xmath77 is the impulse response function , i.e. , the fraction of light scattered by the echo toward us , in units of inverse time , which depends on the nature of the dust and the echo geometry @xcite .
the total sn light is treated as a short pulse over which the sn flux is constant , i.e. , @xmath78 @xcite .
the sn fluence is the integral of the light curves in each band with respect to time .
the light curves that we considered here are the combination of those presented in @xcite and @xcite in @xmath79 , which approximately match what we see from the echo in f336w , f438w , f555w , and f814w , and the _ swift _ @xmath80 data from @xcite , where @xmath80 is the most similar _ swift _ bandpass to f275w .
see figure [ latelc ] .
monitoring in @xmath81 began when the sn was at age @xmath2 2 days , while the @xmath82 and @xmath80 data commenced at @xmath2 4 days .
of course , at the very earliest ages the sn light is dominated by the shock breakout , which is particularly important in the uv as a luminous flash . however ,
the shock breakout was mostly not observed in the case of sn 2012aw .
so , we have simulated what might have occurred , via shock breakout models , specifically , the analytical expressions for rsgs from @xcite and the hydrodynamical model for a @xmath83 rsg from @xcite .
ehud nakar graciously provided his algorithms , which we ran assuming an initial mass @xmath84 , a radius . ]
@xmath85 , and an explosion energy of @xmath86 erg , at the central wavelengths of the five bandpasses .
the coefficients in the @xcite model have since been calibrated through numerical simulations of the explosion of an analytical star ( t. shussman et al .
, in preparation ) , and we used these updates .
nozomu tominaga also provided us with the output from his rsg model through the five filter bandpasses .
we have adjusted this model to @xmath87 and converted from ab magnitudes to vega magnitudes .
we show , specifically , the very early - time observed and model uv light curves in figure [ earlylc ] .
here we are showing the curves in flux units , after reddening correction for an assumed total extinction , @xmath88 mag ( see below ) , assuming the @xcite reddening law with @xmath1 .
the behavior of the two models is relatively similar , although they exhibit notable differences during the first @xmath89 d after explosion .
the post - explosion peak of the breakout in the uv is noticeably more pronounced in the @xcite model than in the @xcite model .
the secondary peak at @xmath35 d in the @xcite model is as luminous , if not somewhat more so , as the initial peak near @xmath90 d. the secondary peak emission for the @xcite model occurs earlier ( at @xmath91 d ) , and is more in line with the first observed datapoint , than for the @xcite model .
these secondary peaks and the declines thereafter undershoot ( for @xcite ) and overshoot ( for @xcite ) the observed uv light curve .
we therefore truncate both model uv curves at @xmath92 d. the behavior of the model light curves , relative to the observations , is quite similar in the other bands , although the initial post - explosion flash diminishes in brightness toward the longer wavelengths .
we merely needed to truncate the models before the first datapoint , as we did in the uv , to match the observations in the other bands .
the fact that both sets of shock breakout models tend to agree quite well in flux with the early observed data , requiring very little adjustment , implies that the input parameters for those models , i.e. , the progenitor initial mass and sn explosion energy , are reasonable representations of sn 2012aw and its progenitor star .
cccc @xmath80 & @xmath93 & @xmath94 & @xmath95 + @xmath82 & @xmath96 & @xmath97 & @xmath97 + @xmath98 & @xmath99 & @xmath100 & @xmath101 + @xmath102 & @xmath103 & @xmath104 & @xmath105 + @xmath106 & @xmath107 & @xmath107 & @xmath107 + we have also included in the observed light curves our estimates of the sn brightness in the 2014 images , as shown in figure [ latelc ] , where we assume f275w @xmath108 , f336w @xmath109 , f438w @xmath110 , f555w @xmath111 , and f814w @xmath112 .
all of the light curves at late times declined in a manner roughly consistent with radioactive decay from day @xmath113 to day @xmath114 , although the @xmath106 curve appears to have declined more steeply , while the @xmath80 curve somewhat less steeply .
after correction for the total extinction to the sn ( see below ) , we integrated , first , just the observed light curves in each band ( neglecting the comparatively small uncertainties in the observed sn photometry ) and , then , these light curves including the two early - time flash models . note that for the observed light curves shown in figure [ latelc ] we are missing the end of the plateau in the redder bands .
( in the uv no indication exists of a pronounced plateau , so this is not an issue for this band . ) for this reason , in particular , we interpolated the light curves in each band from the observed datapoints and estimated the approximate end of the plateau and commencement of the exponential decay phase of the light curves , based on the slope of the decline inferred from the initial data points , post - plateau , and the slope of the data points on the decline after day @xmath115 in each band .
the slopes of the decline in the light curves after the end of the observed data ( at day @xmath113 in @xmath79 , day @xmath116 in @xmath80 ) were interpolated up to our estimates on day @xmath114 .
our integrations , then , were over these interpolated light curves , rather than the actual observed data
. however , we consider the uncertainties in the total fluence introduced by this assumption to be small , since the behavior of our interpolated light curves are quite similar to that of the observed light curves of the prototypical , normal sn ii - p 1999em , for which the end of the plateau was observed ; see @xcite and @xcite .
we adopted vega as the flux zero point , obtained via synthetic photometry of vega s observed _ hst _ space telescope imaging spectrograph ( stis ) spectrum using stsdas / synphot filter transmission function from http://svo2.cab.inta - csic.es / theory / fps3/. ] in iraf .
we list the results in table 3 .
we see that the flash models in the uv increase the overall fluence by @xmath117% ( @xcite model ) to @xmath118% ( @xcite model ) .
this is in better agreement with the observations , since the flux from the echo is relatively high in the uv .
the early - time shock - breakout models contribute @xmath36% additional flux at @xmath82 , but only negligible additional amounts at @xmath81 . treating @xmath119 as the sn maximum flux in each band , we find that the sn pulse duration is 3.8 , 21.4 , 51.6 , 110.3 , and 141.2 days in @xmath80 , @xmath82 , @xmath98 , @xmath102 , and @xmath106 , respectively . the maximum value of @xmath120 , the effective pulse width @xcite , is then @xmath121 pc ( from @xmath106 ) .
the observed echo thickness , @xmath122 , is unresolved and is conservatively less than the fwhm of a single stellar profile , @xmath1232.0 wfc3 pixels , or , @xmath124 pc .
it follows from @xcite that the dust thickness , @xmath125 , is @xmath126 pc .
this value is clearly not less than our estimate of the dust sheet distance , @xmath67 .
however , since we can only place an upper limit on @xmath125 ( given the upper limit on @xmath122 ) , we infer that its actual value is likely far smaller .
our assumptions for the dust scattering , above , therefore likely still hold . following , e.g. , @xcite and @xcite @xmath127 where @xmath128 is the speed of light , @xmath129 is the h column density , @xmath130 is the scattering `` phase function , '' @xmath131 is the scattering cross section , and @xmath132 is the grain size distribution for grain radius @xmath133 .
the classical @xcite @xmath130 is not generally applicable in this case . as @xcite points out
, the phase function in the uv is strongly forward - scattering , particularly as the scattering angle becomes small ( @xmath134 ) .
@xcite advises tabulating the phase function at short wavelengths , rather than employing an analytical expression .
we have adopted his definition of the phase function , @xmath135 where @xmath136 is the scattering cross section and @xmath137 is the differential scattering cross section per h nucleon , both as a function of wavelength @xmath138 .
the scattering angle , @xmath63 , is obtained from @xmath139 / [ ( b / ct)^2 + 1]$ ] ( e.g. , * ? ? ?
the scattering angle is therefore @xmath140 for the sn 2012aw echo . to compute @xmath141 we used the
tabulated wavelength - dependent scattering properties for @xmath142 from @xcite for the @xcite galactic extinction model with @xmath1 .
furthermore , we also adopted @xmath131 from the @xcite dust model , with the update from @xcite , which also includes the relevant carbonaceous and silicate grain size distributions @xmath143 .
@xcite argued that the metallicity in the sn environment is consistent with solar metallicity , so assuming a galactic model for the dust in the sn interstellar environment is probably valid .
finally , we adopted the line - of - sight galactic foreground visual extinction toward sn 2012aw from @xcite , i.e. , @xmath144 mag , again , assuming the @xcite reddening law .
we show the results of our modeling of the echo in figure [ echosed ] .
the uncertainty in @xmath145 shown in the figure is entirely a result of the uncertainty in @xmath146 , the distance from the sn to any scattering dust element along the echo .
additional uncertainties which we have not included can arise from the scattering dust width , the inclination of the dust sheet with respect to the line - of - sight @xcite , and any asymmetries in the sn explosion ( @xcite ; early - time spectropolarimetry by @xcite may have indicated large intrinsic polarization and significant asymmetries in the outer sn ejecta ) .
the observations constrain @xmath129 in the echo - producing region to a narrow range of 3.55@xmath147 @xmath148 . in figure [ echosed ]
we show echo models with the average value , @xmath149 @xmath148 . following the relation between @xmath129 and @xmath150 from @xcite ,
this implies that the internal @xmath151 mag for the echo - producing dust . we would obtain the same estimate for @xmath150 based on the relation from @xcite .
we note that @xcite estimated the extinction within the host galaxy , based on the equivalent widths of the na i d features in a high - resolution spectrum of the sn near maximum light , to be @xmath152 mag ( see also @xcite ) , which is consistent with the implied extinction from the echo observations .
the sn occurred in a region of m95 seemingly devoid of any recent massive star formation , so natal molecular gas and dust may have already dispersed some time ago .
we find the close agreement in these estimates of @xmath150 to be sufficiently reassuring that our light echo models are reasonably correct . if we include the galactic foreground , the total extinction to sn 2012aw is @xmath88 mag .
the echo model is certainly not perfect : we have only included models for the very early - time light curve behavior for the sn and do not know the exact nature of the shock breakout .
we employed interpolated light curves in our computation of the sn fluences , since the observed photometric coverage in the five bands is not entirely complete .
we also may have incorrectly estimated the sn contribution to the total echo light .
possible faint positive residuals can still be seen in the sn - subtracted images of the echo .
our estimate of the echo radius may well be incorrect .
additionally , our assumed host galaxy distance may not be correct ; the nasa / ipac extragalactic database ( ned ) list 21 cepheid - based distance moduli which span , to within their uncertainties , a range of about 2.7 mpc in distance , although the uncertainty - weighted mean of these distance modulus estimates , 29.95 mag , is within the uncertainties of our assumed distance modulus .
if the echo radius were actually larger , implying @xmath146 would also be larger , then @xmath129 ( and , thus , @xmath150 ) would correspondingly also have to be larger ; the dust model itself would also change somewhat , as the scattering angle @xmath63 would be decreased . ( the reverse would be the case if the echo radius were smaller . )
the overall agreement with @xmath150 obtained from the early - time sn spectrum would no longer hold , although it is conceivable that that estimate is in error .
( @xcite estimated that the host extinction could be as high as @xmath1530.59 mag . ) furthermore , the echo - producing dust in m95 may not be as similar to milky way diffuse dust as we have assumed .
all of these factors would result in increasing the uncertainty in our overall model .
nonetheless , we have shown that , to within the uncertainties that we have estimated , we can fully and consistently account for the observed flux from the echo by assuming that the sn light has been scattered by diffuse interstellar dust .
the echo light can actually be fit adequately enough without addition of the flash models , although the model flux in the uv is somewhat low .
including the shock breakout flash , using either model , provides a superior representation of the observations in the blue and uv .
we know that a flash must occur , so it is logical to have included these models with the observed light curves .
all of the echo models tend to overpredict the flux slightly , relative to the observation , at @xmath106 .
we have discovered in the legus multi - wavelength _ hst _ imaging a dust - scattered light echo around the sn ii - p 2012aw in m95 , at @xmath24 d after explosion .
the superior _ hst _ angular resolution was essential ; at @xmath154 in diameter , observations from the ground under only truly exceptional conditions would have potentially partially resolved the echo .
the echo was quite bright at the time , at @xmath292122 mag in each of the five wfc3/uvis bands .
the echo is a result of the scattering of the total sn light from interstellar dust in the host galaxy , which we assume to have properties similar to milky way diffuse dust .
in essence , the light echo is a complete record of the sn s luminosity evolution .
we find from comparing a model of the echo to the observations that the amount of the dust extinction in the sn environment responsible for the echo is consistent with the value that was estimated from observations of the sn itself at early times .
an independent means of constraining @xmath129 might come from x - ray observations of sn 2012aw : the sn was observed on 2012 april 11 with the _ chandra x - ray observatory _ , and the data provide only a modest constraint , with the value of @xmath129 implied by the echo consistent with the limits allowed by the x - ray spectrum ( d. pooley , private communication ) .
the echo around sn 2012aw had become detectable in 2014 as the sn light had become sufficiently dim . under the assumption of a homogeneous thin dust sheet in front of the sn ,
the light echo is always present and essentially constant in luminosity , even at maximum light , although the echo luminosity is typically @xmath155 mag fainter than the sn at maximum ( see , e.g. , @xcite ) .
if one reasonably assumes that at @xmath114 d the total luminosity is dominated by the echo , it is @xmath156 mag less luminous at @xmath102 than the sn at maximum light , which is roughly in line with the prediction for typical interstellar light echoes . from the approximations in (
* his section 3 ) and adopting our estimates of the sn - sheet distance and the dust optical depth , one would obtain a sn - light echo contrast of @xmath36 mag .
this assumes standard parameters for the dust and a homogeneous distribution within a thin dust sheet the @xmath33 mag difference between the observed echo brightness and what is predicted could be explained by an inhomogeneous dust distribution , as apparently indicated by the asymmetric shape of the echo ( see figure [ figsuper ] ) .
one could conclude from the echo s appearance that there is more dust toward one side of the line - of - sight , so that the echo may also have increased in luminosity with time , as a result of increasing dust optical depth .
furthermore , we find that , if our estimate of the sn brightness in the f814w band is correct at @xmath157 mag ( see table 2 ) , then the sn at the time was fainter than the observed brightness of the sn progenitor star in that band , 23.39 mag @xcite .
we can not completely rule out that dust in the sn ejecta or immediate sn environment is at least partially responsible for the sn s faintness in 2014 .
however , it was evident that any circumstellar dust in the progenitor envelope was destroyed by the x - ray / uv flash @xcite . adopting a sn expansion velocity of 3631 km s@xmath158 for sn 2012aw @xcite , after 2.1
yr the sn would have already swept through @xmath159 , or @xmath160 au , most probably beyond the immediate circumstellar environment .
additionally , analysis of _ spitzer space telescope _
late - time data ( to be presented elsewhere ) shows that sn 2012aw was fading quite rapidly at 3.6 and 4.5 @xmath161 m up to day @xmath162 , when emission from the sn in these bands was barely detectable .
dust observed at these wavelengths would be relatively warm , and we can not discount that colder sn dust may be present .
also , relatively small newly - condensed dust masses have been found for sne ii - p ( e.g. , @xcite ; although see , e.g. , @xcite in the case of sn 1987a ) .
a more straightforward conclusion is that the progenitor star has vanished and that the rsg identified by @xcite and @xcite was almost certainly the progenitor of sn 2012aw .
finally , the sn 2012aw light echo should be further monitored with _ hst _
, particularly in the blue , to follow the evolution of its luminosity , as well as potentially to reveal new or evolving echo structures , and better constrain the nature of the echo geometry and of the scattering dust . continued observations will also place further constraints on the late - time sn brightness and disappearance of the progenitor .
analysis of any future polarimetric imaging of the echo ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , which would allow for an independent distance determination to the host galaxy , however , could be challenging , given the relatively compact nature of the sn 2012aw light echo .
we are grateful for the comments provided by the expert referee , which significantly improved this paper .
we appreciate helpful discussions we had with luc dessart and the generosities of both ehud nakar and nozomu tominaga .
we also appreciate david pooley sharing his analysis of the _ chandra _ data on sn 2012aw in advance of publication .
based on observations made with the nasa / esa _ hubble space telescope _
, obtained at the space telescope science institute , which is operated by the association of universities for research in astronomy , inc .
, under nasa contract nas 5 - 26555 .
these observations are associated with program # 13364 .
support for this program was provided by nasa through a grant from the space telescope science institute .
this research made use of the nasa / ipac extragalactic database ( ned ) , which is operated by the jet propulsion laboratory , california institute of technology , under contract with nasa . ah acknowledges funding by spanish mineco under grants aya2012 - 39364-c02 - 01 and sev 2011 - 0187 - 01 .
sedm acknowledges support by a marie sklodowska - curie reintegration fellowship ( h2020-msca - if-2014 , project i d 661502 ) awarded by the european commission .
bayless , a. j. , pritchard , t. a. , roming , p. w. a. , et al .
2013 , , 764 , l13 bond , h. e. , gilmozzi , r. , meakes , m. g. , et al . 1990 , , 354 , l49 bond , h. e. , henden , a. , levay , z. g. , et al .
2003 , , 422 , 405 bose , s. , kumar , b. , sutaria , f. , et al .
2013 , , 433 , 1871 calzetti , d. , lee , j. c. , sabbi , e. , et al .
2015 , , 149 , 51 cappellaro , e. , patat , f. , mazzali , p. a. , et al .
2001 , , 549 , l215 cardelli , j. a. , clayton , g. c. , & mathis , j. s. 1989 , , 345 , 245 chevalier , r. a. 1986 , , 308 , 225 couderc , p. 1939
, annales dastrophysique , 2 , 271 crotts , a. p. s. 1988 , , 333 , l51 crotts , a. 2014 , , submitted ( arxiv:1409.8671 ) crotts , a. p. s. , & kunkel , w. e. 1991 , , 366 , l73 crotts , a. p. s. , landsman , w. b. , bohlin , r. c. , et al . 1992 , , 395 , l25 dallora , m. , botticella , m. t. , pumo , m. l. , et al .
2014 , , 787 , 139 dolphin , a. e. 2000 , pasp , 112 , 1383 draine , b. t. 2003 , , 598 , 1017 drozdov , d. , leising , m. d. , milne , p. a. , et al . 2014 , arxiv:1410.8190 ekstrm , s. , georgy , c. , eggenberger , p. , et al .
2012 , , 537 , a146 eldridge , j. j. , izzard , r. g. , & tout , c. a. 2008 , , 384 , 1109 fraser , m. , maund , j. r. , smartt , s. j. , et al .
2012 , , 759 , l13 freedman , w. l. , madore , b. f. , gibson , b. k. , et al .
2001 , , 553 , 47 fruchter , a. s. , et al . 2010 , 2010 space telescope science institute calibration workshop , p. 382
- 387 gezari , s. , dessart , l. , basa , s. , et al .
2008 , , 683 , l131 gezari , s. , jones , d. o. , sanders , n. e. , et al .
2015 , , in press ( arxiv:1502.06964 ) gonzaga , s. , hack , w. , fruchter , a. , & mack , j. ( eds . ) 2012 , the drizzlepac handbook ( baltimore : stsci ) gouiffes , c. , rosa , m. , melnick , j. , et al .
1988 , , 198 , l9 gver , t. , & zel , f. 2009 , , 400 , 2050 hamuy , m. , pinto , p. a. , maza , j. , et al .
2001 , , 558 , 615 henyey , l. c. , & greenstein , j. l. 1941 , , 93 , 70 jerkstrand , a. , smartt , s. j. , fraser , m. , et al .
2014 , , 439 , 3694 kapteyn , j. c. 1901 , astronomische nachrichten , 157 , 201 kochanek , c. s. , khan , r. , & dai , x. 2012 , , 759 , 20 kotak , r. , meikle , w. p. s. , farrah , d. , et al .
2009 , , 704 , 306 krause , o. , rieke , g. h. , birkmann , s. m. , et al .
2005 , science , 308 , 1604 krause , o. , tanaka , m. , usuda , t. , et al .
2008 , , 456 , 617 leonard , d. c. , filippenko , a. v. , gates , e. l. , et al .
2002 , , 114 , 35 leonard , d. c. , pignata , g. , dessart , l. , et al .
2012 , the astronomer s telegram , 4033 , 1 liu , j .- f .
, bregman , j. n. , & seitzer , p. 2003
, , 582 , 919 matsuura , m. , dwek , e. , barlow , m. j. , et al .
2015 , , 800 , 50 mattila , s. , meikle , w. p. s. , lundqvist , p. , et al .
2008 , , 389 , 141 meikle , w. p. s. , mattila , s. , gerardy , c. l. , et al .
2006 , , 649 , 332 meikle , w. p. s. , mattila , s. , pastorello , a. , et al .
2007 , , 665 , 608 meikle , w. p. s. , kotak , r. , farrah , d. , et al .
2011 , , 732 , 109 miller , a. a. , smith , n. , li , w. , et al .
2010 , , 139 , 2218 nakar , e. , & sari , r. 2010 , , 725 , 904 patat , f. 2005 , , 357 , 1161 predehl , p. , & schmitt , j. h. m. m. 1995 , , 293 , 889 prieto , j. l. , rest , a. , bianco , f. b. , et al .
2014 , , 787 , ll8 quimby , r. m. , wheeler , j. c. , hflich , p. , et al .
2007 , , 666 , 1093 quinn , j. l. , garnavich , p. m. , li , w. , et al . 2006 , , 652 , 512 rest , a. , suntzeff , n. b. , olsen , k. , et al .
2005 , , 438 , 1132 rest , a. , matheson , t. , blondin , s. , et al . 2008a , , 680 , 1137 rest , a. , welch , d. l. , suntzeff , n. b. , et al .
2008b , , 681 , l81 rest , a. , sinnott , b. , welch , d. l. , et al .
2011a , , 732 , 2 rest , a. , foley , r. j. , sinnott , b. , et al . 2011b , , 732 , 3 rest , a. , prieto , j. l. , walborn , n. r. , et al .
2012a , , 482 , 375 rest , a. , sinnott , b. , & welch , d. l. 2012b , pasa , 29 , 466 ritchey , g. w. 1901 , , 14 , 293 sana , h. , de mink , s. e. , de koter , a. , et al .
2012 , science , 337 , 444 schaefer , b. e. 1987 , , 323 , l47 schawinski , k. , justham , s. , wolf , c. , et al .
2008 , science , 321 , 223 schlafly , e. f. , & finkbeiner , d. p. 2011
, , 737 , 103 sinnott , b. , welch , d. l. , rest , a. , sutherland , p. g. , & bergmann , m. 2013 , , 767 , 45 sokoloski , j. l. , crotts , a. p. s. , lawrence , s. , & uthas , h. 2013 , , 770 , l33 sparks , w. b. 1994 , , 433 , 19 sparks , w. b. 1996 , , 470 , 195 sparks , w. b. , macchetto , f. , panagia , n. , et al . 1999 , , 523 , 585 sugerman , b. e. k. 2003 , , 126 , 1939 sugerman , b. e. k. 2005 , , 632 , l17 sugerman , b. e. k. , andrews , j. e. , barlow , m. j. , et al .
2012 , , 749 , 170 sugerman , b. e. k. , & crotts , a. p. s. 2002 , , 581 , l97 tamburro , d. , rix , h .- w . , walter , f. , et al . 2008 , , 136 , 2872 tominaga , n. , morokuma , t. , blinnikov , s. i. , et al .
2011 , , 193 , 20 van dyk , s. d. 2013 , , 146 , 24 van dyk , s. d. , li , w. , & filippenko , a. v. 2006 , , 118 , 351 van dyk , s. d. , cenko , s. b. , poznanski , d. , et al .
2012 , , 756 , 131 wang , x. , li , w. , filippenko , a. v. , et al .
2008 , , 677 , 1060 weingartner , j. c. , & draine , b. t. 2001 , , 548 , 29 welch , d. l. , clayton , g. c. , campbell , a. , et al .
2007 , , 669 , 525 xu , j. , crotts , a. p. s. , & kunkel , w. e. 1994 , , 435 , 274 xu , j. , crotts , a. p. s. , & kunkel , w. e. 1995 , , 451 , 806 ( erratum : 1996 , , 463 , 391 ) yadav , n. , ray , a. , chakraborti , s. , et al .
2014 , , 782 , 30 | we have discovered a luminous light echo around the normal type ii - plateau supernova ( sn ) 2012aw in messier 95 ( m95 ; ngc 3351 ) , detected in images obtained approximately two years after explosion with the wide field channel 3 on - board the _ hubble space telescope _ ( _ hst _ ) by the legacy extragalactic ultraviolet survey ( legus ) .
the multi - band observations span from the near - ultraviolet through the optical ( f275w , f336w , f438w , f555w , and f814w ) .
the apparent brightness of the echo at the time was @xmath022 mag in all of these bands .
the echo appears circular , although less obviously as a ring , with an inhomogeneous surface brightness , in particular , a prominent enhanced brightness to the southeast .
the sn itself was still detectable , particularly in the redder bands .
we are able to model the light echo as the time - integrated sn light scattered off of diffuse interstellar dust in the sn environment .
we have assumed that this dust is analogous to that in the milky way with @xmath1 .
the sn light curves that we consider also include models of the unobserved early burst of light from the sn shock breakout .
our analysis of the echo suggests that the distance from the sn to the scattering dust elements along the echo is @xmath245 pc .
the implied visual extinction for the echo - producing dust is consistent with estimates made previously from the sn itself .
finally , our estimate of the sn brightness in f814w is fainter than that measured for the red supergiant star at the precise sn location in pre - sn images , possibly indicating that the star has vanished and confirming it as the likely sn progenitor . |
in recent years , considerable experimental and theoretical efforts have been undertaken to understand the physics of hadrons containing a heavy quark @xcite.the isgur - wise function @xcite is an important quantity in this area of hadron physics.it is in this spirit , that this function has been studied in various quark models @xcite@xcite besides qcd sum rule approach @xcite , the mit bag model @xcite and the skryme model @xcite . + since one of the basic ingredients of the iw function is the hadron wavefunction involving heavy quark @xcite@xcite , it is therefore meaningful to test any specific qcd inspired quark model by calculating the iw function and study it phenomenologically .
+ sometimes back , a specific qcd inspired quark model was proposed by us @xcite which had later been used to calculate iw function as well @xcite .
+ one of drawback of the model is that significant confinement effect could not be accomodated in the model@xcite due to perturbative constraints coming from using the dalgarno s method @xcite.only recently @xcite , standard confinement effect @xmath1 @xcite was accomodated in the improved version of qcd inspired quark model brought through the introduction of parameter @xmath0 in the potential : @xmath3 taking @xmath4 @xmath5 @xmath6 as its natural scale fixing and @xmath7 = @xmath8 where @xmath7 is an undetermined factor appearing in the series solution of the schrdinger equation [ eq.(8 ) of ref.19].in earlier works@xcite , the unknown coefficient @xmath9 occured in the wavefunction was set to zero .
+ one of the drawback of work [ 19]was the adhoc enhancement of strong coupling constant was needed to take into account of the slope and curvature of @xmath2,@xmath10 and @xmath11 mesons .
+ in this work , we take an alternative strategy to remove this adhoc enhancement.we use the wavefunction at the origin involving the unknown coefficient @xmath9 and fix it from the experimental values of masses and decay constants directly.the reality constraint on @xmath9 will then yeild lower bounds on the strong coupling constant @xmath12 , which would lead to the upper bounds on the slope and curvature of the iw function . +
the rest of the paper is organised as follows : section 2 contains the theory of the improved qcd inspired quark model , section 3 encloses the results and in section 4 we draw conclusion and remarks .
the spin independent fermi - breit hamiltonian for ground state ( @xmath13),neglecting the contact term proportional to @xmath14 is @xcite : @xmath15 where @xmath12 is the running coupling constant , @xmath16 is the confinememt parameter and @xmath4 is another parameter whose significance will be cleared later .
+ in this work , the @xmath12 values are taken from the v -scheme @xcite as done in @xcite which are large as compared to those of @xmath17-scheme.it is necessary as large @xmath12 values lead to better results for slope and curvature of isgur - wise function in earlier work @xcite .
+ in earlier work @xcite , large confinement was found to be inconvenient in the calculation of slope and curvature of isgur -wise function which was a limitation of the model.so , our aim has been focussed in the inclusion of larger confinement and hence we retain the same choice of @xmath1 @xcite to investigate whether this approach leads to better results or not.it is worth notable that inclusion of @xmath4 in the analysis @xcite allows larger @xmath16 useful , so we do nt think another choice of @xmath16 .
+ with , @xmath18 as the parent hamiltonian and @xmath19 as the perturbed hamiltonian , we obtain a ground state wavefunction upto the first order correction using the dalgarno method @xcite of stationary state perturbation theory as : + @xmath20 where @xmath21 is the unknown coefficient appearing in the series solution of the dalgarno method + including the relativistic effect @xcite , the wavefunction is : + @xmath22 here @xmath23 is given by : + @xmath24 and + @xmath25 @xmath26 and @xmath27 are the normalization constants given by : + @xmath28 and @xmath29 where @xmath30 is given in appendix a. + we note that the equations ( 2),(3),(6)and ( 7 ) are obtained from eq.(4),(6),(5 ) and ( 7 ) of ref.[19 ] exhibiting explicit dependence of @xmath9 in them . the wavefunction at the origin ( wfo ) , is related to the decay constant @xmath32 and the mass of the pseudoscalar meson @xmath33 through the relation @xcite : + @xmath34 again from equation ( 2 ) , we have : + @xmath35\ ] ] using equation(6 ) , we arrive at the quadratic equation for @xmath9 : @xmath36 where + @xmath37 @xmath38 and @xmath39-\frac{1}{\pi a_{0}^{3}}\ ] ] using the experimental values of @xmath32 and @xmath33 @xcite , we determine @xmath40 from equation(9 ) which in turn will yeild two solutions for @xmath9 in equation ( 10 ) : @xmath41 which will depend on @xmath42,@xmath43,@xmath44 and @xmath12.the solution corresponding to the + ve(-ve ) sign of equation(15)will be termed as + ve(-ve ) solution hereafter.it will be shown numerically that for a given @xmath42 , @xmath43,and @xmath44 , @xmath12 reaches the minimum value when the following condition is satisfied : @xmath45 the formalism involving eq.(5)-(16 ) is strictly valid only without relativistic effect as the wavefunction at the origin with such effect [ eq.(3 ) ] is not well defined due to its singularity at the origin.for a subsequent analysis , we assume that @xmath9 does not deviate significantly from its non - relativistic value so that it can be used to calculate the slope and curvature of the iw function even without relativistic effect
. + the isgur - wise function is written as @xcite : @xmath46 where @xmath47 and @xmath48 and @xmath49 being the four velocity of the heavy meson before and after the decay.the quantity @xmath50 is the slope of i - w function at @xmath51 and known as charge radius : + @xmath52 the second order derivative is the curvature of the i - w function known as convexity parameter : + @xmath53\ ] ] for the heavy - light flavor mesons the i - w function can also be written as @xcite : @xmath54 where @xmath55 equation ( 21 ) holds good for both relativistic and nonrelativistic case.the wavefunction @xmath56 takes different form for both the cases . without relativistic effect ,
it is given by equation(2 ) and with relativistic effect it is given by ( 3 ) .
+ with the wavefunction(2)in equation(10 ) i.e. including confinement only the charge radius @xmath57 and convexity parameter @xmath58 are respectively given by : + @xmath59}{8\pi c^{2}a_{0}^{2}a_{0}^{3}+8 + 45\mu^{2}b^{2}a_{0}^{6}+16ca_{0}\sqrt{\pi a_{0}^{3}}-24\mu bca_{0}\sqrt{\pi a_{0}^{3}}-24\mu ba_{0}^{3}}\ ] ] and : @xmath60}{16\pi c^{2}a_{0}^{2}a_{0}^{3}+16 + 90\mu^{2}b^{2}a_{0}^{6}+32ca_{0}\sqrt{\pi a_{0}^{3}}-48\mu bca_{0}\sqrt{\pi a_{0}^{3}}-48\mu ba_{0}^{3}}\ ] ] with the wavefunction ( 3)in equation ( 10 ) i.e. including both relavistic and confinement effect the charge radius @xmath61 and convexity parameter @xmath62 are given by : + @xmath63}{4[x_{2}]}\ ] ] and @xmath64}{96[x_{2}]}\ ] ] where @xmath30,@xmath65 and @xmath66 are given in appendix .
+ we note that equations ( 25 ) and ( 26 ) are equivalent to equations ( 18 ) and ( 19)of ref[19 ] exhibiting explicit @xmath9 dependence .
as noted earlier , @xmath9 depends on @xmath42,@xmath43,@xmath44 and @xmath12.in fig.1(a - e ) we plot @xmath9 vs @xmath12 for @xmath67,@xmath68,@xmath2,@xmath10 and @xmath11 mesons .it shows that @xmath12 tends to reach the minimum value when two solutions of eq.(11 ) almost merge satisfying the condition(16).this feature is true for any set of the parameters @xmath42 , @xmath32 and @xmath33 .
in table 1 , we give the lower bounds on @xmath12 for mesons having @xmath4 and @xmath16 quarks .
+ the dependence of @xmath9 on @xmath12 and @xmath42 can be noted as follows : + with constant @xmath42 , @xmath9 decreases with @xmath12 values rising and vice-versa.on the other hand , with constant @xmath12 , @xmath9 increases(decreases ) with increase ( decrease ) in @xmath42 .
+ vs @xmath12 for @xmath67 meson.the + ve ( -ve ) solution of eq.15 corresponds to the dashed ( solid ) line and the two lines nearly coincide at @xmath12 @xmath5 @xmath69 , the lower bound on @xmath12 corresponding to the solution of eq.16 for d meson .
, width=432 ] vs @xmath12 for @xmath68 meson.the + ve ( -ve ) solution of eq.15 corresponds to the dashed ( solid ) line and the two lines nearly coincide at @xmath12 @xmath5 @xmath70 , the lower bound on @xmath12 corresponding to the solution of eq.16 for @xmath68 meson.,width=432 ] vs @xmath12 for @xmath2 meson.the + ve ( -ve ) solution of eq.15 corresponds to the dashed ( solid ) line and the two lines nearly coincide at @xmath12 @xmath5 @xmath71 , the lower bound on @xmath12 corresponding to the solution of eq.16 for @xmath2 meson.,width=432 ] vs @xmath12 for @xmath10 meson.the + ve ( -ve ) solution of eq.15 corresponds to the dashed ( solid ) line and the two lines nearly coincide at @xmath12 @xmath5 @xmath72 , the lower bound on @xmath12 corresponding to the solution of eq.16 for @xmath10 meson.,width=432 ] vs @xmath12 for @xmath11 meson.the + ve ( -ve ) solution of eq.15 corresponds to the dashed ( solid ) line and the two lines nearly coincide at @xmath12 @xmath5 @xmath73 , the lower bound on @xmath12 corresponding to the solution of eq.16 for @xmath11 meson.,width=432 ] .
lower bounds on @xmath12 [ cols="^,^,^,^,^,^,^ " , ]
in this paper , we have shown that the reality bound on @xmath9 puts lower limit on @xmath12 and correspondingly upper limit on @xmath50 and @xmath74 .
+ furthermore , with @xmath9 , the upper bounds on @xmath50 and @xmath74 decrease which is evident from the above list of bounds[table-2 ] .
the estimated upper bounds on @xmath50 and @xmath74 for all the mesons are found to be consistent with other models and data [ table-3 ] without making any adhoc enhancement of the strong coupling constant as had been done in ref ( 19)[table-4 ] .from the phenomenological point of view we note that in the nonrelativistic limit , the universal form factor and isgur - wise function for semileptonic decay @xmath75 are identical when subleading terms in velocity and terms of order @xmath76 are neglected with @xmath77 as the binding energy and @xmath78 as the mass of heavy quark @xcite.however even if we make calculation for the universal form factor for finite mass , we obtain to first order in @xmath79 as @xmath80 -@xmath81 which seems to be satisfactory @xcite .
+ it is worth notable that in the limit @xmath82 , there will be no bounds on @xmath12 as well as on @xmath50 and @xmath74 ; rather fixed values of @xmath12 have to be used to get definite set of @xmath50 and @xmath74.so , in that case , the analysis will turn to that of ref [ 17,18 ] where large confinement could not be ( i.e.@xmath1 ) incorporated e.g. tables -(1,3 ) of ref[17 ] and tables -(2,3 ) of ref[18 ] .
+ we conclude this paper with a comment on the physical significance of the factor @xmath0 that has become so crucial for our analysis of bounds on slope and curvature . + it is common wisdom that a constant potential like @xmath0 just scales the energies and doesnot affect the wavefunction nor does it change physics.this can be seen from the hydrogen atom problem with the potential @xmath83.however , if one uses @xmath0 as the perturbation instead of as parent in the dalgarno method of perturbation theory @xcite , the wavefunction for the @xmath84-atom becomes : + @xmath85 + to be compared with the wavefunction with @xmath0 as parent : + @xmath86 + where the normalization constant : + @xmath87 + thus , the perturbative child @xmath0 rather than the parent @xmath0 plays the crucial role in the present analysis . + 1 m. neubert , phys rep 245,259 ( 1994 ) + n isgur and m b wise phys lett b 232,113 ( 1989 ) + b.holdom,m .
sutherland and j mureika , phys rev d 49,2359 ( 1994 ) + ivanov m a , lyubouvitskij v e , krner l g , kroll p , phy rev d 56,348(1997 ) + knig b , krner j g , krmer , kroll p , phy rev 56,4282(1997 ) + f e close and a wambach , nucl phys b 412,169 ( 1994 ) + h w huang , phys rev d 56,1579 ( 1979 ) + d melikhov , phys rev d 53,2460 ( 1996 ) + m r ahmady , r r mendel and j d talman , phys rev d 52,254 ( 1995 ) + m g olsson and s veseli , fermilab - pub-96/418-t + m g olsson and s veseli , phys rev d 51,2224 ( 1995 ) + a le yaouanc , l oliver and j c raynal ; phy rev d 69,094022 ( 2004 ) + y b dai , c s huang , m q huang and c liu;phys lett b,387,379 ( 1996 ) + m sadzikowski and k zalewski ; z phys c 59 , 667 ( 1993 ) + e jenkins , a manohar and m b wise ; nucl phys b 396,38(1996 ) + d k choudhury , p das , d d goswami and j
n sharma ; pramana j.phys.44,519(1995 ) + d k choudhury and n s bordoloi ; ijmpa , vol.15,no.23,3667(2000 ) + d k choudhury and n s bordoloi ; mpla , vol.17,no.29,1909 ( 2002 ) + d k choudhury and n s bordoloi ; mpla , vol.24,no.6,443(2009 ) + a ghatak and s lokanathan in `` quanum mechanics'',mcgraw hill,1997,pp-291 + e eichten ; phys.rev.d 17,3090 ( 1978 ) + j j sakurai in `` advanced quantum mechanics '' ( massachusetts , addison - willey publishing company,1986 ) pp-128 + c itzykson and j zuber in `` quantum field theory '' ( international student edition , mcgraw hill , singapore,1986 ) pp-79 + v o galkin , a yu mishurov and r n faustov;sov j nucl phys 53,1026(1991 ) + d e groom et al(particle data group);eur phy j c,15,1(2000 ) + y schrder ; phys lett b 447,321(1999 ) + y schrder ; nucl phys proc suppl 86,525(2000 ) + a le yaouanc , l oliver , o pene and j
c raynal ; phys.lett.b 365 ; 319 ( 1996 ) + j l rosner ; phys rev d 44,3732 ( 1990 ) + t mannel , w roberts and z ryzak;phys rev d 44,r18 ( 1991 ) + t mannel , w roberts and z ryzak;phys lett b 255,593(1993 ) + m neubert ; phys lett b 264,455 ( 1991 ) + f jugeau , a le yaouanc , l oliver , j c raynal;phys rev d 70,114020(2004 ) + f e close , a wambach;ral-94 - 041 , outp-94 09p;april 1994 + | the qcd inspired potential model persued by us earlier has been recently modified to incorporate an additional factor @xmath0 in the linear cum coulomb potential .
while it felicitates the inclusion of standard confinement parameter @xmath1 , unlike in previous work , it still falls short of explaining the isgur - wise function for the @xmath2 mesons without adhoc adjustment of the strong coupling constant .
+ in this work , we determine the factor @xmath0 from the experimental values of decay constants and masses and show that the reality constraint on @xmath0 yeilds bounds on the strong coupling constant as well on slope and curvature of isgur - wise function allowing more flexibility to the model .
+ keywords : dalgarno method , isgur - wise function , slope , curvature . +
pacs nos .
12.39.-x ; 12.39.jh ; 12.39.pn |
low momentum jets are estimated to produce 50% of transverse energy in rhic heavy ion collisions and 80% at the lhc @xcite . despite the large role these _ minijets _ play ,
they have received little attention from the general community since the start of the rhic experimental program . as minijet abundance increases at higher energies
, the dynamics of minijet interactions are becoming essential to understanding heavy ion collisions .
while low momentum jets are not individually resolvable , their combined effect generates an observable correlation . in a theoretical context minijets
are typically defined within the range of applicability of pqcd by specifying a low hadron @xmath6 cutoff around 2 gev / c , even though qcd interactions continue to lower @xmath6 .
we experimentally define minijets based on correlation structure rather than an _ a priori _ @xmath6 range .
this requires a _ minimum - bias _ two - particle correlation analysis where every possible pair of particles is considered instead of selecting a few trigger / associated pairs .
minijets are distinguished from other sources by decomposing the unique correlations .
previous analyses have used this technique to reveal large minijet contributions in transverse @xcite and axial ( @xmath4 , @xmath5 ) @xcite spaces at 130 gev at four centralities . here
we report the detailed energy and centrality dependence of minijet angular correlations at rhic .
charged particle tracks detected in the star tpc with @xmath7 gev / c , @xmath8 , and full @xmath9 azimuth were analyzed from 1.2 m minbias triggered 200 gev au+au and 6.7 m 62 gev au+au events .
pair densities @xmath10 were measured as number of pairs per unit area on relative angles ( @xmath11 , @xmath12 ) for all possible unique particle pairs .
particles within the same event form _ sibling _ pair densities @xmath13 , while mixing particles from different events measures the uncorrelated _ reference _ @xmath14 .
these are formed into a normalized covariance to produce a correlation measure .
the difference @xmath15 measures the covariance in number of pairs between histogram bins , and the normalization is provided by bin - wise division of @xmath16 .
thus we use the notation @xmath17 for a _ per - particle _ correlation measure , shown in figure [ fig1 ] for selected centralities .
proton - proton collisions provide a reference for measuring the contributions to these structures .
analysis of minimum - bias correlations @xcite and single particle @xmath6 spectra @xcite show that p+p collisions are well described by a two - component soft and semi - hard scattering model , as commonly used in event generators such as pythia .
the soft component represents longitudinal fragmentation in unlike - sign pairs and produces a 1d gaussian correlation centered along @xmath18=0 .
the semi - hard component contains a same - side peak , modeled as a 2d gaussian at the @xmath19 origin , and an away - side ridge centered at @xmath20=@xmath21 . for an inclusive @xmath6 range the away - side
is completely represented by function @xmath22 that approximates a wide gaussian which narrows with increasing @xmath6 @xcite .
the final component necessary to describe p+p data is a 2d exponential at the origin containing contributions from hbt in like - sign pairs and conversion @xmath23 in unlike - sign pairs .
to ensure the simplest possible fit function for au+au collisions , we use these components from p+p collisions with only one additional @xmath24 quadrupole term to account for correlations conventionally attributed to elliptic flow @xcite . the eleven parameter fit function used for the correlation structures in figure [ fig1 ]
is then : + [ eq : fit ] f = a _ _ ( _ ) + a_2_(2 _ ) + g_s(@xmath18 : a_0 , _ 0 ) + g_h(@xmath18 , @xmath20 : a_1 , _ , _ ) + + e(@xmath18 , @xmath20 : a_2 , w _ , w _ ) + a_3 where @xmath25 and @xmath26 are the soft and hard gaussian terms and @xmath27 is an exponential function with parameters listed after the colon .
an example of this fit is shown in figure [ fig2 ] .
figure [ fig3 ] shows the measured fit parameters for the same - side peak amplitude , @xmath18 width , and volume ( @xmath28 ) .
fitting errors are shown and systematic error is estimated to be @xmath299% of the correlation amplitude and at most a few percent of the widths .
the dashed lines show the binary scaling reference expected from independent nucleon - nucleon collisions . using the kharzeev and nardi two - component model @xcite and path length @xmath30 , the minijet amplitude in au+au collisions
is expected to scale as @xmath31 $ ] from the p+p value .
peripheral collisions follow the binary scaling reference closely , deviating only by small increases in @xmath18 and decreases in @xmath20 widths .
the data show a sharp transition at approximately 55% centrality for 200 gev and 40% for 62 gev where the amplitude and @xmath18 widths increase dramatically while the @xmath20 widths continue to decrease slightly .
centrality in figure [ fig3 ] is represented by transverse particle density calculated as @xmath32 with initial collision overlap area @xmath33 from monte carlo glauber .
transverse density brings the transition points for the two energies to coincidence , whereas conventional centrality measures displace the transition points and tend to compress the peripheral data
. width , and volume fits . points
show eleven centrality bins for each energy ( 84 - 93% , 74 - 84% , 65 - 74% , 55 - 65% , 46 - 55% , 37 - 46% , 28 - 37% , 19 - 28% , 9 - 19% , 5 - 9% , and 0 - 5% ) transformed to tranvserse density . ]
the correlation structures are modified at the transition , but are still likely to be associated with minijets for several reasons .
first , these results , particularly when taken with a similar analysis of @xmath6 correlations @xcite , show that contributions from a new physical mechanism unrelated to minijets are unlikely .
any such hypothetical process must have @xmath20 widths and @xmath6 correlations that match seamlessly with minijets , which would be a remarkable coincidence .
second , the amplitude and @xmath18 width increases are consistent with further minijet interactions , which may be possible due to path - length considerations @xcite . finally , it is possible that the new correlation structures are due to changes in minijet fragmentation .
the trends in the data also suggest a lower @xmath6 manifestation of the `` ridge '' @xcite , and these results may help to discriminate among the many competing models of ridge formation .
the same - side peak volume gives the total number of correlated pairs , though finding the particle yield requires estimating the average number of correlated structures per event .
assuming each structure originates with a semi - hard parton and that semi - hard scattering follows binary scaling , we estimate that 30% of all final - state hadrons in central 200 gev au+au collisions are associated with this same - side correlation . as a source of correlated low momentum particles
, minijets provide an extremely sensitive probe of the collision system .
the binary scaling reference represents one extreme limit of a transparent medium , while the other extreme is a completely thermalized system opaque to minijets @xcite .
these results call into question the existence of the latter system at rhic energies .
wang x - n and gyulassy m 1991 _ phys . rev .
d _ * 44 * 3501 adams j ( star collaboration ) 2007 _ j. phys .
phys . _ * 34 * 799 adams j ( star collaboration ) 2006 _ phys . rev .
c _ * 73 * 064907 porter r j and trainor t a 2005 _ j. phys .
* 27 * 98 adams j ( star collaboration ) 2006 _ phys .
rev d _ * 74 * 032006 trainor t a and kettler d t 2007 _ preprint _
[ hep - ph ] kharzeev d and nardi m 2001 _ phys .
b _ * 507 * 121 adams j ( star collaboration ) 2006 _ j. phys .
g : nucl . part .
phys . _ * 32 * l37 kajantie k 1987 _ phys .
lett . _ * 59 * 2527 putschke j ( star collaboration ) 2007 _ preprint _ nucl - ex/0701074 nayak g c 2001 _ nucl .
phys . a _ * 687 * 457 ; shin g r and mller 2003 _ j. phys .
g : nucl . part .
phys . _ * 29 * 2485 | we have measured 2d autocorrelations for all charged hadrons in star with @xmath0 @xmath1 0.15 gev / c and @xmath2 @xmath3 1 from au+au collisions at 62 and 200 gev .
the correlation structure is dominated by a peak centered at zero relative opening angles on @xmath4 and @xmath5 which we hypothesize is caused by minimum - bias jets ( minijets ) .
we observe a large excess of minijet correlations in more - central au - au collisions relative to binary - collision scaling ( more correlated pairs than expected from surface emission or even volume emission ) . we also observe a sudden increase of the minijet peak amplitude and @xmath4 width relative to binary - collision scaling of scattered partons which occurs at an energy - dependent centrality point .
there is a possible scaling of the transition point with transverse particle density .
the large minijet correlations bring into question the degree of thermalization in rhic collisions . |
the long standing solar neutrino deficit @xcite , the atmospheric neutrino anomaly @xcite , and the results from the lsnd experiment on @xmath12 neutrinos from @xmath13 decay and neutrinos from @xmath14 decay @xcite can each be understood in terms of oscillations between two neutrino species @xcite .
interestingly , the solar , atmospheric , and terrestrial ( lsnd ) neutrino oscillations have different @xmath15 and therefore require different neutrino mass squared differences @xmath16 to properly describe all features of the data . for example , if the atmospheric and lsnd @xmath5 scales are the same @xcite , one forfeits the recently reported zenith - angle dependence and up / down asymmetry of the atmospheric neutrino flux @xcite .
alternatively , if the solar and atmospheric @xmath5 scales @xcite are the same , the reduction in the solar neutrino flux is energy - independent , contrary to the three solar experiments which infer different oscillation probabilities in different neutrino energy regions @xcite .
since three distinct mass - squared differences can not be constructed from just three neutrino masses , the collective data thus argue provocatively for more than three oscillating flavors . an alternative but
less compelling possibility is to introduce new lepton
flavor changing operators with coefficients small enough to evade present exclusion limits , but large enough to explain the small lsnd amplitude @xcite .
if all of the existing observations are confirmed , a viable solution is to invoke one or more additional species of sterile light neutrino @xcite , thereby introducing another independent mass scale to the theory .
the additional neutrino must be sterile , i.e. without standard model gauge interactions , to be consistent with lep measurements of @xmath17 @xcite .
the introduction of a sterile neutrino to complement the three active neutrinos has had some phenomenological success @xcite .
in this paper we propose and study mass matrices for four neutrino models ( three active plus one sterile ) that can accommodate all the present data . once a fourth neutrino
is admitted to the spectrum , it is no longer mandatory that the @xmath1 mix with the @xmath18 at the atmospheric scale .
the @xmath1 may instead mix with the sterile @xmath19 , or with some linear combination of @xmath19 and @xmath18 .
similarly , the @xmath6 may mix with a linear combination of @xmath19 and @xmath18 . at first sight
the mixing of a sterile neutrino with active flavor neutrinos seems to be stringently constrained by big bang nucleosynthesis ( bbn ) physics .
the bound @xmath20 on the mass - squared difference @xmath5 and the mixing angle of the sterile neutrino was inferred to avoid thermal overpopulation of the `` extra '' sterile neutrino species@xcite . however , there are significant caveats to this bound .
one is the fact that some recent estimates of @xmath21 using higher abundances of @xmath11he yield considerably weaker bounds @xcite .
another is that a small asymmetry @xmath22 of flavor neutrinos ( but large compared to the present baryon asymmetry @xmath23 ) at @xmath24 s is enough to suppress @xmath25 oscillations and then the bound of eq .
( [ bbn ] ) does not apply @xcite .
such asymmetries , in fact , can be generated with the kind of model parameters considered herein ( as shown in ref .
@xcite ) . in light of this observation that bbn may allow sizeable mixing between sterile and active neutrinos , we consider both the small and large mixing with sterile neutrinos in this work .
we review all existing data that indicate neutrino oscillations , and then perform a model
independent analysis of the data using four neutrino unitarity constraints .
a very useful tool for this unitarity analysis is the set of probability rectangles , which we explain and exploit .
we draw several model
independent conclusions for the four neutrino universe .
we design a five parameter neutrino mass matrix which can account for each of the three viable solar solutions and accommodate the atmospheric and lsnd observations .
the three solar possibilities are the small - angle matter - enhanced ( sam ) @xcite , large - angle matter enhanced ( lam ) @xcite and large angle vacuum long wavelength ( vlw ) @xcite explanations of the solar neutrino deficit .
our mass matrix yields maximal oscillations of atmospheric @xmath1 .
we consider the possibility that the solar data is explained by @xmath26 or @xmath27 oscillations , in which case the atmospheric neutrino data is explained by either @xmath28 or @xmath29 oscillations , respectively .
we also consider the possibility that both atmospheric and solar neutrino oscillations have @xmath19 and @xmath18 components .
lack of @xmath19@xmath18 discrimination in the present data is the major source of ambiguity in the four neutrino model .
we discuss how future experiments can resolve this ambiguity . in sec .
2 we summarize the oscillation probability formulas and utilize a probability formalism , based on unitarity of the mixing matrix , which permits a simple visual represention of mixing . in sec .
3 we begin with a brief discussion of the three classes of experiments and the neutrino mass and mixing parameters needed to explain them .
we then use probability rectangles to display the inferences from the data for any four neutrino scheme . in sec .
4 we employ the probability rectangles to argue against a neutrino mass spectrum with one eigenstate separated from three other nearly degenerate states ( which we will refer to as the 1 + 3 spectrum ) in favor of two nearly degenerate mass pairs ( which we will refer to as the 2 + 2 spectrum ) .
we also derive some new relations among elements of the mixing matrix that result from data and unitarity which are satisfied in a four neutrino model for certain ranges of the parameters . then in sec .
5 we present a mass matrix whose eigenvalues consist of a nearly degenerate neutrino pair at @xmath30 ev and another nearly degenerate pair at low mass , as illustrated in fig . 1 .
we show how the existing data almost uniquely fixes the model parameters ( once a solar scenario is specified ) and strictly determines what new phenomenology the model predicts . in sec .
6 we derive expressions for the oscillation probabilities in our models in terms of the current neutrino experimental observables .
we present the model predictions in sec .
the new observable signature for the model is @xmath31 or @xmath32 oscillations for @xmath33 km / gev , depending on whether the atmospheric oscillations are @xmath28 or @xmath29 , respectively .
section 8 contains some discussion , and a summary .
to simplify the analysis of the available data , we will ignore possible cp violation and work with a real valued mixing matrix @xmath34 .
accordingly , the general formula for the vacuum oscillation probabilities becomes @xcite @xmath35 where @xmath36 , @xmath37 , and the sum is over all @xmath38 and @xmath39 subject to @xmath40 . for oscillations of two neutrinos ,
the oscillation amplitude ( i.e. , the coefficient of the @xmath41 term ) is given by @xmath42 , where @xmath43 is the mixing angle between the two neutrino states .
more generally for an arbitrary number of neutrinos , the amplitude of the @xmath44 to @xmath45 oscillation in the absence of cp violation is seen to be a^=-4_k < j u_ju_ju_ku_k , , [ amp ] where the sum is over mass states with mass - squared differences appropriate for the @xmath15 of the particular experiment .
we note that the oscillation amplitudes defined here are only for those oscillations at a particular @xmath46 scale in eq .
( [ oscprob ] ) .
we will use subscript labels on the amplitude to identify the @xmath46 scale ( which is determined by the relevant @xmath5 and @xmath15 ) for the particular experiment : `` sbl '' will denote short baseline experiments such as lsnd , `` atm '' will denote atmospheric and long baseline experiments , and `` sun '' will denote extraterrestrial experiments , especially those with solar neutrinos .
we will use superscripts on the amplitude to identify the oscillation flavors , unless it is obvious from the context ; in the absence of cp violation , @xmath47 . with four neutrino states ,
@xmath34 is a @xmath48 mixing matrix .
we also define the amplitude for @xmath44 disappearance a^ _ a^ , [ ampnot ] where @xmath49 represents a sum over neutrino flavor eigenstates other than @xmath44 . the mixing
matrix elements @xmath34 , and therefore the amplitudes @xmath50 , depend on the environment , e.g. , matter vs. vacuum . throughout this paper
we will quote values for the oscillation amplitudes _ in vacuum_. the `` probability rectangles '' used by liu and smirnov @xcite visually illustrate the mixing of the flavor eigenstates among the mass eigenstates . to construct the probability rectangles
, we introduce the notation p_j |u_j|^2 , [ prob ] such that @xmath51 is the probability that the @xmath52 flavor state is found in the @xmath53 mass state , or , alternatively , the probability that the @xmath53 mass state is contained in the @xmath52 flavor state . therefore , when cp
violation is neglected , the real mixing
matrix elements are determined by the probabilities up to a sign : @xmath54 . in principle , these signs may be determined by arranging for orthogonality of the rows , and columns , in the unitary mixing matrix @xmath34 . by unitarity of @xmath34
we have _ p_j = 1 [ uni1 ] for each mass state @xmath38 , and
_ j p_j = 1 [ uni2 ] for each flavor state @xmath55 .
thus , if each mass state is represented as a rectangle of unit area , then the fractional area assigned to @xmath51 within the rectangle is a graphical representation of the value of @xmath51 .
the probabilities @xmath51 depend on whether the environment is vacuum or matter . for consistency
, we will always display vacuum probabilities in the rectangles .
when the probability rectangles are displayed along a vertical axis labeled with mass squared , the @xmath5 values relevant for the various experiments are readily visualized .
figure 2 gives an example of the probability rectangles for a four neutrino model .
an inverted 2 + 2 mass spectrum , where the solar @xmath6 oscillation is driven by the separation of the heavier two states and the atmospheric @xmath1 oscillation is driven by the separation of the lighter two states , may also be possible , but is not considered . the following mini - theorem will prove to be useful : + _ in the absence of matter effects , the amplitude @xmath56 is independent of how the @xmath57 probabilities are partitioned among the mass eigenstates . _
+ the proof of this statement relies on the insertion of @xmath58 into + @xmath59 , to get a^=4_j > k p_j p_k , [ theorem ] where , as in eq .
( [ amp ] ) , the sum in eq .
( [ theorem ] ) is over all mass states with mass - squared differences appropriate for the @xmath15 of the particular experiment . in eq .
( [ theorem ] ) , @xmath60 is manifestly independent of the partitioning of the @xmath57 probabilities since it involves only the @xmath51 .
this theorem demonstrates the limitations on information derivable from disappearance experiments .
the minitheorem fails in the presence of matter effects because the partitioning of the flavor probabilities including @xmath51 are altered .
that is , with matter effects the amplitude @xmath61 _ does _ depend on how the @xmath57 are partitioned among the mass states .
matter will also alter the oscillation wavelength , causing further changes in the phenomenology of experiments sensitive to the oscillations rather than their averages .
matter effects have the potential to resolve the @xmath19@xmath62 ambiguity , as do some other measurements .
we discuss these possibilities in sec .
the lsnd experiment @xcite reports positive appearance results for @xmath63 oscillations from @xmath13 decay at rest ( dar ) and for @xmath7 oscillations from @xmath14 decay in flight ( dif ) .
the dar data has higher statistics , but the allowed regions for the two processes are in good agreement .
there are also restrictions from the null results of the bnl e-776
@xcite and karmen @xcite @xmath7 oscillation search experiments .
the combined data suggest @xmath7 vacuum oscillation parameters that lie approximately along the line segment described by @xmath64 however , values for @xmath65 as high as 10 ev@xmath9 are also allowed for @xmath66 , although values above 3 ev@xmath9 are disfavored by the r process mechanism of heavy element nucleosynthesis in supernovae @xcite .
there are also relevant data from the bugey reactor experiment which searches for @xmath67 disappearance @xcite , and from the cdhs@xcite and ccfr@xcite experiments which set bounds on disappearance .
the combined short baseline data set for @xmath68 , @xmath69 , and @xmath70 will be used in sec . 4 to argue against a hierarchical neutrino mass spectrum in favor of two pairs of nearly degenerate masses in the four neutrino spectrum .
the atmospheric neutrino experiments measure @xmath1 and @xmath6 ( and their antineutrinos ) created when cosmic rays interact with the earth s atmosphere .
one expects roughly twice as many muon neutrinos as electron neutrinos from the resulting cascade of pion and other meson decays .
several experiments @xcite obtain a @xmath71 ratio that is about 0.6 of the value expected from detailed theoretical calculations of the flux @xcite .
the super - kamiokande ( superk ) experiment has collected the most data and analysis @xcite indicates that their results for contained events can be explained as @xmath28 oscillations with @xcite @xmath72 the high end of each range is favored .
independent of flux normalization considerations , the @xmath7 oscillation channel is strongly disfavored by the zenith angle distributions of the data @xcite and by the up / down asymmetry separated into `` muon like '' ( ) and `` electron like '' ( ) events @xcite , which yield an up to down ratio of @xmath73 for @xmath74like events and @xmath75 for @xmath76like events ( the expected values are close to unity ) .
furthermore , the recent chooz @xmath67 disappearance experiment excludes @xmath77 oscillations with large mixing @xmath78 for @xmath79 @xcite . in a four neutrino context ,
another possibility for the atmospheric neutrino oscillations is @xmath29 .
oscillations of this type in principle could be affected by matter due to the different neutral current interactions of @xmath1 and @xmath19 . however , for the contained events ( with lower energy ) these effects are small , especially for larger values of @xmath5 @xcite ; hence , the allowed regions for @xmath29 should be similar to those for @xmath28 . for events at higher energies the matter effects could begin to be appreciable ; a definitive test requires more data .
the solar neutrino experiments @xcite measure @xmath6 created in the sun .
there are three types of experiments , @xmath6 capture in cl in the homestake mine , @xmath80 scattering at kamiokande and super - kamiokande , and @xmath6 capture in ga at sage and gallex ; each is sensitive to different ranges of the solar neutrino spectrum and measures a suppression from the expectations of the standard solar model ( ssm)@xcite . for @xmath26 oscillations in the sun ( in which case atmospheric neutrino oscillations are @xmath28 in our model ) the allowed parameter ranges at 95% c.l .
@xcite for the small - angle matter - enhanced solution are given in table [ sunsol ] .
the solution is based on the ssm fluxes in ref .
approximate parameters for the large - angle matter - enhanced @xcite and vacuum long wavelength solutions @xcite for @xmath26 oscillations of solar neutrinos are also shown in table [ sunsol ] .
if the solar neutrino deficit is caused instead by @xmath27 oscillations ( and the atmospheric oscillations are @xmath29 ) , then the allowed solar parameter ranges for the three solar cases are slightly different @xcite ; see table [ sunsol ] .
the exact values of the parameters may change as new data from superk @xcite become available and when fits are made with the new solar flux calculations . in any of the matter
enhanced scenarios it is also necessary that the eigenmass @xmath81 associated predominantly with @xmath6 be lighter than the eigenmass @xmath82 associated predominantly with the neutrino into which the @xmath6 is oscillating ( i.e. , or @xmath18 ) , so that it is @xmath6 rather than @xmath83 that is resonant in the sun . for the vacuum solutions the ordering of @xmath82 and @xmath81 does
not matter .
alternate scenarios where the @xmath6 is predominantly associated with the heavier two states and @xmath1 is predominantly associated with the lighter two states are also viable . for @xmath27 oscillations in the two - neutrino approximation
the propagation equation for the neutrino states in the charge - current basis is @xcite i d dt ( c _ e + _ ) = 14e ( cc 42 g_f e n_e & m^2 2 + m^2 2 & 2 m^2 2 ) ( c _ e + _ ) , where @xmath84 is the electron number density . for @xmath26 oscillations the propagation equation is instead @xcite i d dt ( c _ e + _ s ) = 14e ( cc 42 g_f e ( n_e-12n_n ) & m^2 2 + m^2 2 & 2 m^2 2 ) ( c _ e + _ s ) , where @xmath85 is the neutron number density . for the small angle matter enhanced case the non
adiabatic approximate solution for neutrino propagation is appropriate and the oscillation probability for a neutrino of energy @xmath86 is @xmath87 where in this case @xmath88 labels either @xmath89 or @xmath90 , and @xmath91 \,\ ] ] is the landau
zener transition probability and @xmath92 is either @xmath84 ( for @xmath27 oscillations ) or @xmath93 ( for @xmath26 oscillations ) .
the quantity @xmath94 is the appropriate logarithmic density gradient in the sun at @xmath95 , the critical density where maximal oscillations ( resonance ) occur .
for the large angle matter
enhanced case , the neutrino propagation is adiabatic and @xmath96 \,\ ] ] assuming the neutrinos are created where the electron density is well above the critical density .
for the vacuum long wavelength solution the oscillation probability is just given by the usual vacuum expressions . in neutrino oscillation descriptions of the solar , atmospheric , and lsnd data , a distinct oscillation wavelength and oscillation amplitude is required for each of the three data sets .
experimental uncertainties allow for some latitude in these amplitudes and wavelengths , and for the solar data , there are three isolated islands of viability in the @xmath16amplitude plane @xcite ; see fig . 3 .
night asymmetry measurement , found to be small in the recent superk data , removed about half of the previously viable solar regions @xcite . the vacuum oscillation wavelength is linear in the neutrino energy , allowing further possibilities that are summarized in table [ amplength ] .
the chosen neutrino energies are typical for solar , and reactor sources ( 5 mev ) , pion facilities ( 100 mev ) , for contained ( 2 gev ) , partially contained ( 10 gev ) , and throughgoing ( 100 gev ) neutrino events in underground detectors , and for astrophysical sources ( 1 tev ) .
although full oscillation wavelengths are also listed in table [ amplength ] , oscillation effects may well be measurable for a fraction of an oscillation wavelength or as an average over many oscillation wavelengths .
throughgoing and partially contained atmospheric neutrinos may show nodes as a function of l / e .
further possibilities arise when the matter effect of the earth is included in the oscillation physics .
we consider earth
matter effects in sec
. 7.6 .
we consider first the probability rectangles for the atmospheric and chooz data .
the atmospheric data indicate @xmath97 and nearly maximal flavor changing mixing of @xmath1 with @xmath98 .
the present data do not distinguish between @xmath18 or @xmath19 as the dominant state into which @xmath1 mixes .
the probability rectangles for the atmospheric scale are displayed in fig .
we label the two masses defining the atmospheric scale as and , with @xmath99 . because of the `` @xmath18@xmath19 '' ambiguity we show the union @xmath100 rather than the partitions into @xmath101 and @xmath102 . for maximal -@xmath98 mixing
, one must choose @xmath103 .
next we consider the pair of mass eigenstates whose mass squared difference is fixed by the solar scale .
we provisionally investigate a four neutrino mass spectrum that consists of two pairs of nearly degenerate neutrinos separated by the lsnd scale @xmath104 ( to explain the lsnd result in terms of oscillations ) .
we argue in sec . 4 that the data favor this spectrum over a spectrum with one mass separated from three relatively degenerate masses .
we label the second pair of mass states as and , and define @xmath105 . since @xmath106 and @xmath107 sum to near unity , @xmath108 and @xmath109 must be small .
thus the probability rectangles for the @xmath110 and @xmath111 states appear as shown in fig .
accordingly , the lsnd amplitude for @xmath112 , a^e_sbl=-4 [ u_3 u_e 3 + u_2 u_e 2 ] [ u_1 u_e 1 + u_0 u_e 0 ] , [ lsnd1 ] is necessarily small .
we emphasize that the smallness of the lsnd @xmath6@xmath1 mixing is an inevitable consequence of the large mixing of to @xmath98 at the atmospheric scale and the constraints of unitarity , independent of particular model considerations including rearrangement of the neutrino mass spectrum .
four neutrino unitarity may be used to rewrite eqn .
( [ lsnd1 ] ) as a^e_sbl = 4 |u_3 u_e3 + u_2 u_e2|^2 = 4 |u_1 u_e1 + u_0 u_e0|^2 .
[ lsnd2 ] written this way , it is clear that the lsnd data is blind to the partitioning of and in the probability rectangles of mass states and .
this flavor ambiguity is shown in fig .
4a . with the identification @xmath113
, we may use eq .
( [ theorem ] ) to write the solar disappearance amplitude as a^ee_sun=4p_e0 p_e1 .
[ asun ] because matter in the sun may exert a significant effect on propagating neutrinos , the values of @xmath114 and @xmath115 for the sun have some sensitivity to the state or @xmath19 into which @xmath6 oscillates . however , the sensitivity of present data to this difference is weak , and there is considerable freedom in assigning or or a linear combination thereof as the mixing partner to @xmath6 .
this @xmath18@xmath19 ambiguity for mixing at the solar scale is complementary to the @xmath19@xmath18 ambiguity for mixing at the atmospheric scale .
potential measurements to resolve the @xmath19@xmath18 ambiguity at the solar scale will be discussed in sec .
the @xmath114 and @xmath115 partitioning specifies whether the solar model is a small angle model or a large angle model . as can be inferred from eq .
( [ asun ] ) , with nearly equal partitioning of @xmath114 and @xmath115 , the mixing amplitude is near maximal ( large angle ) . with highly nonequal partitioning
, i.e. , @xmath116 or @xmath117 , the mixing amplitude is small . of the three viable solar neutrino options
, sam falls into the small angle category , while lam and vlw fall into the large angle category .
the probability rectangles for the small and large angle classes of models are shown in figs .
4b and 4c .
recall that in order to obtain the msw resonant enhancement required for the sam and lam solutions , it is necessary that the state which is predominantly be the lighter of the two mass states , .
qualitatively , lam and vlw are distinguishable in their probability rectangles only by the choice of value for @xmath118 .
quantitatively , the two solutions and the vlw solution are distinguishable in ways which are discussed in sec .
if the active sterile mixing is small , then all ambiguities in the probability rectangles are resolved : the large atmospheric mixing must be @xmath1@xmath18 , and the solar solution must be small angle sam with @xmath6@xmath19 mixing .
the probability rectangles for this model are shown in fig .
this particular solution has recently been analyzed in the context of the minimal four neutrino mass matrix @xcite .
it has been shown by bilenky , giunti and grimus @xcite that a hierarchical ordering of the four neutrino spectrum ( implying one dominant mass ) is disfavored by the data when the null results of reactor and accelerator disappearance experiments are included .
we will refer to this spectrum as the 1 + 3 spectrum , defined as one heavier mass state separated from three lighter , nearly - degenerate states , or vice versa .
we demonstrate the argument with a set of logical steps similar to theirs .
assume a mass spectrum with one heavy mass well separated from three other nearly - degenerate states and let the heavy mass state be labeled as @xmath119 .
then the lsnd mass - squared scale is @xmath120 and the lsnd amplitude is a^e_sbl=-4u_e3u_3 [ _ j3 u_ju_ej ] = 4p_e3 p_3 , p_e3+p_31 .
[ emu1 ] on the other hand , the and disappearance experiments at reactors and accelerators are also sensitive to the lsnd scale .
these experiments measure the disappearance amplitudes a^ee_sbl=4 p_e3 [ _ j3 p_e j ] = 4p_e3 [ 1-p_e3 ] [ enote ] and a^_sbl = 4p_3 [ 1-p_3 ] .
[ munotmu ] the second equalities in eqs .
( [ emu1 ] ) and ( [ enote ] ) ( see eq .
( [ theorem ] ) ) follow from unitarity of the mixing matrix .
the three amplitudes in eqs .
( [ emu1])([munotmu ] ) depend on just two parameters , and so are interrelated .
all three of these amplitudes are constrained by experiments to be small . _ a priori _
then , @xmath121 and @xmath107 may both be small , or one ( but not both ) may be near unity with the other small .
the fact that @xmath122 is an _ appearance observation _ rather than a bound means that if @xmath121 and @xmath107 are both small , they can not be too small . in the 1 + 3 model ,
the atmospheric scale does not involve the heavy state @xmath119 . without loss of generality we label the state which determines the atmospheric scale as @xmath123 . then from eq .
( [ theorem ] ) the atmospheric @xmath1 disappearance oscillation amplitude is given by a^_atm = 4 p_2 ( p_0 + p_1 ) ( 1-p_3)^2 , [ atmbound ] where the inequality comes from maximizing @xmath124 subject to the constraint @xmath125 .
the superk data indicate that is maximally mixed at the @xmath126 scale , i.e. , there is little content available to the @xmath119 state .
quantitatively we have @xmath127 , which implies @xmath128 . since @xmath107 is small , eq .
( [ munotmu ] ) becomes a^_sbl4p_3 < < 1 .
[ munotmu2 ] the probability rectangles for the 1 + 3 model with small @xmath107 are presented in fig . 5 .
note that it is the zenith
angle , or up / down asymmetry data , which really establishes @xmath126 as different from @xmath65 , that is crucial for the argument @xcite .
we are left with the possibilities of @xmath121 being small or near unity . as can be seen in fig . 5 ,
if @xmath121 is near unity , then there is little @xmath129 to distribute over the three lighter mass states . in particular ,
the solar amplitude @xmath130 , where the solar scale is @xmath105 , is second order in small quantities , too small for even the sam solution ( @xmath131 ) to the solar flux .
this may be easily quantified . if @xmath121 were near unity , we would have a^ee_sbl4[1-p_e 3 ] < < 1 . [ enote2 ] together with unitarity , this in turn bounds the magnitude of the solar amplitude : a^ee_sun=4p_e0 p_e1 ( 1-p_e3)^2 ( a^ee_sbl)^2 , [ sunbound ] where the inequality in eq .
( [ sunbound ] ) comes from maximizing @xmath132 subject to the constraint @xmath133 .
the experimental upper limit on @xmath134 from the bugey experiment @xcite is about 0.1 for @xmath135 ev@xmath9 , which disallows even the small angle solar solution .
we conclude that @xmath121 is small , in which case a^ee_sbl 4p_e3 < < 1 . [ enote3 ]
thus , both @xmath121 and @xmath107 must be small in the 1 + 3 model , and from eqs .
( [ emu1 ] ) , ( [ munotmu2 ] ) , and ( [ enote3 ] ) , we infer the relation a^e_sbla^_sbl a^ee_sbl .
[ aaa ] however , the experimental upper bounds on the disappearance amplitudes @xmath134 @xcite and @xmath136 @xcite and the measured appearance result for @xmath122 @xcite are not compatible with eq .
( [ aaa ] ) , thereby disfavoring the 1 + 3 model .
for example , for @xmath137 ev@xmath9 , @xmath138 from bugey , @xmath139 from cdhs , which implies @xmath140 ; however , for this value of @xmath141 , the lsnd data indicate @xmath142 .
the lsnd results are presented in terms of maximum likelihood rather than confidence level limits , so it is not straightforward to state an exclusion probability .
put another way , @xmath122 is large enough that the bugey and cdhs limits force one of @xmath143 and @xmath107 to be small and the other to be large , but this is ruled out by the solar and atmospheric data .
the constraints on @xmath121 and @xmath107 from the three short baseline amplitudes @xmath144 , @xmath134 , and @xmath136 , the atmospheric amplitude @xmath145 and the solar amplitude @xmath146 ( from eqs .
( [ emu1 ] ) , ( [ enote ] ) , ( [ munotmu ] ) , ( [ atmbound ] ) , and ( [ sunbound ] ) , respectively ) are conveniently summarized in fig .
6 for two different values of @xmath65 .
the measured values and bounds for the short baseline appearance and disappearance amplitudes depend on the magnitude of @xmath65 .
( there is effectively a suppressed @xmath65 third axis in our fig . 6 , which samples only two particular values of @xmath141 . ) for certain allowed values of @xmath5 ( e.g. , at 1.7 @xmath147 and 0.25 @xmath147 according to fig . 2 of @xcite ) the violation of eq .
( [ aaa ] ) is mild , and the 1 + 3 model is just barely incompatible with the data ; see , e.g. , fig .
the argument against the 1 + 3 model does not depend on the sign of @xmath65 .
this means that the inverted 3 + 1 model with the three nearly degenerate mass states heavier than the remaining state is equally disfavored .
we now turn to the favored class of four neutrino models , namely those with two nearly degenerate mass pairs separated by the lsnd scale as displayed in fig
it is interesting to see how this `` pair of pairs '' mass spectrum of four neutrino models realizes the dependency among @xmath122 , @xmath136 , and @xmath134 which conflicted with the 1 + 3 model .
let @xmath110 and @xmath111 label the pair of the nearly degenerate mass eigenstates responsible for the solar oscillations , and @xmath123 and @xmath119 label the pair of the nearly degenerate mass eigenstates responsible for the atmospheric oscillations .
the expressions for the oscillation amplitudes are a^e_sbl= 4|u_e2u_2+u_e3u_3|^2 = 4|u_e0u_0+u_e1u_1|^2 , [ emupairs ]
a^ee_sbl = 4(p_e3p_e1+p_e3p_e0+p_e2p_e1+p_e2p_e0 ) = 4_e(1-_e ) , [ enotepairs ] and a^_sbl = 4(p_3p_1 + p_3p_0 + p_2p_1 + p_2p_0 ) = 4_(1- _ ) , [ munotmupairs ] with
_ |u_2|^2+|u_3|^2 = p_2 + p_3 .
[ sigma ] the schwartz vector inequality @xmath148 applied to the vectors @xmath149 and @xmath150 then gives a^e_sbl 4_e_. [ schwartz ] furthermore , in the 2 + 2 model the solar oscillation amplitude is a^ee_sun = 4 p_e0 p_e1 ( 1 - _ e)^2 , [ sunboundpairs ] and the atmospheric oscillation amplitude is a^_atm = 4 p_2 p_3 _ ^2 , [ atmboundpairs ] where the inequalities in eqs .
( [ sunboundpairs ] ) and ( [ atmboundpairs ] ) come from maximizing the expressions subject to the constraints @xmath151 and @xmath152 , respectively . if the vector inequality in eq .
( [ schwartz ] ) is saturated , then @xmath153 , @xmath134 , and @xmath136 each has the same functional dependence on two parameters as it did in the 1 + 3 model ( @xmath154 has replaced @xmath121 and @xmath155 has replaced @xmath107 ) .
then the previous argument that the lsnd , bugey and cdhs data require one parameter to be small ( @xmath156 ) and the other large ( @xmath157 ) applies .
the argument is unaffected if the vector inequality is not saturated . as before in the 1 + 3 case , the solar constraint indicates that @xmath154 must be small .
this time however , unlike the 1 + 3 case , the atmospheric constraint involves @xmath158 and not @xmath159 , and can be met if @xmath155 is large ( @xmath0 ) .
therefore the constraints of the data can be satisfied by assigning @xmath6 dominantly to one pair of mass states and @xmath1 dominantly to the other pair . instead of eq .
( [ aaa ] ) pertinent to the 1 + 3 spectrum , we obtain for the 2 + 2 spectrum a^e_sbl a^ee_sbl .
[ aa ] this bound is linear in the small disappearance amplitudes , and is easily satisfied by the data .
for example , the tightest constraint on @xmath134 is about 0.02 for @xmath160 ev@xmath9 , while the lsnd data indicate @xmath122 can be as low as 0.009 for this value of @xmath65 . in fig .
7 we have drawn the @xmath154@xmath155 plot for the 2 + 2 model , analogous to the @xmath121@xmath107 plot for the 1 + 3 model , for @xmath161 ev@xmath9 .
the allowed regions with @xmath155 near unity ( implying near maximal mixing of @xmath1 in the @xmath123@xmath119 pair ) and @xmath154 small ( implying almost no mixing of @xmath6 into the @xmath123@xmath119 pair ) show that the 2 + 2 model can comfortably accommodate the data . since only mass - squared differences are important for oscillations , the inverted 2 + 2 model , where the solar oscillations are driven by the mass - squared difference of the upper mass pair and the atmospheric oscillations are driven by the mass - squared difference of the lower mass pair , is equally viable .
two features of the data are especially noteworthy .
the first is the remarkably high degree of isolation of @xmath6 into one mass pair and @xmath1 into the other mass pair , as inferred from the bounds on the disappearance amplitudes .
the second is the near saturation of the vector inequality in eq .
( [ schwartz ] ) by the lsnd appearance amplitude @xmath122 for @xmath162 ev@xmath9 .
equations ( [ enotepairs ] ) and ( [ munotmupairs ] ) bound the degree to which and are found in opposite pairs of mass eigenstates . without loss of generality
we assume that @xmath6 is predominantly associated with @xmath110 and @xmath111 , and that @xmath1 is predominantly associated with @xmath123 and @xmath119 . then from the bugey and cdhs data we find the constraints _ e a^ee_sbl 0.016 ( 0.009 ) , [ sigesmall ] and 1- _ a^_sbl 0.013 ( 0.2 ) , [ sigmusmall ] respectively , for @xmath163 ev@xmath9 ( @xmath164 ev@xmath9 ) .
we then deduce a^ee_sbl 0.065 ( 0.04 ) , which can be compared to the lsnd data a^e_sbl 0.0025 ( 0.04 ) , for these two values of @xmath65 .
the near saturation of the inequality in eq .
( [ aa ] ) for @xmath165 ev@xmath9 has very interesting implications .
it means that @xmath166 is nearly parallel or antiparallel to @xmath167 , which in turn indicates that |u_e2/u_e3| |u_2/u_3| .
[ parallel23 ] this is a new result .
furthermore , the superk data suggest that is maximally mixed in the mass pair with mass - squared difference @xmath126 , so for this pair , called @xmath123 and @xmath119 , that |u_2| |u_3| .
[ umu23 ] then eq .
( [ parallel23 ] ) implies |u_e2| parameters are indeed near the limits of the bugey bound , the four neutrino mixing matrix in the 2 + 2 model must satisfy eq .
( [ parallel23 ] ) , which implies eq .
( [ ue23 ] ) if the atmospheric @xmath1 mixing is maximal .
we can derive additional constraints by considering @xmath168 and @xmath169 rather than @xmath154 and @xmath155 .
the data requires @xmath170 to be large ( @xmath157 ) and @xmath171 small , and the schwartz inequality reduces to a^e_sbl a^_sbl , [ schwartz2 ] where the cdhs bound is a^_sbl 0.05 ( 0.8 ) , at @xmath172 ev@xmath9 ( @xmath164 ev@xmath9 ) . because the inequality in eq .
( [ schwartz2 ] ) is not saturated by the data for any @xmath65 , a relation similar to eq .
( [ parallel23 ] ) for @xmath173 , @xmath174 , @xmath175 , and @xmath176 is not required in the 2 + 2 model .
however , the explicit mass matrices we consider do have such additional relations ; see sec .
finally , we mention a curiosity @xcite in the data which occurs for the pair of pairs mass spectrum with the matter
enhanced solar solutions ( sam and lam ) .
the _ linear _ mass splitting at the heavier pair is @xmath177 . if this pair is associated with the atmospheric scale , we have @xmath178 . on the other hand , if the lighter mass pair is associated with the matter
enhanced solar scale , and @xmath179 , then the linear mass splitting of this pair is @xmath180 .
the two linear mass splittings within the pairs are nearly identical . while squared masses enter into the oscillation formulae for relativistic neutrinos , the more fundamental constructs of field theory , such as the lagrangian and the resulting equations of motion , are linear in fermion masses ( and quadratic in boson masses , these powers of mass being related to the dimensionality of the fermion field vs. the boson field ) .
thus it is a worthy enterprise to attempt to deduce linear neutrino mass relations whenever possible .
to describe the above oscillation phenomena in the scenario where the solar neutrino deficit is described by @xmath26 oscillations and the atmospheric data by @xmath28 , we consider the neutrino mass matrix ansatz @xmath181 presented in the ( @xmath182 ) basis ( i.e. the basis where the charged lepton mass matrix is diagonal ) . by considering the field redefinitions @xmath183 and
@xmath184 one realizes that @xmath185 , and at least one of @xmath186 and @xmath187 , and at least one of @xmath188 and @xmath189 , may be taken as positive ; we will take @xmath185 and all four @xmath190 to be positive for simplicity .
the mass matrix @xmath191 contains five parameters ( @xmath185 , @xmath186 , @xmath188 , @xmath189 , @xmath187 ) , just enough to incorporate the required three mass - squared differences and the oscillation amplitudes for solar and lsnd neutrinos .
the large amplitude for atmospheric oscillations does not require a sixth parameter in our model because the structure of the mass matrix naturally gives maximal mixing of @xmath1 with @xmath18 ( or with @xmath19 if @xmath18 and @xmath19 are interchanged ) . for simplicity
, we have taken the mass matrix to be real and symmetric .
the choice of a symmetric neutrino mass matrix is well motivated in the context of oscillations , for what is measured in neutrino oscillations are the differences of squared masses , which are eigenvalues of the hermitian matrix @xmath192 , which is itself symmetric when cp conservation is assumed .
@xmath191 is diagonalized by an orthogonal matrix @xmath34 ( real ) and there is no cp violation .
the @xmath190 are assumed to be small compared to unity , but not all necessarily of the same order of magnitude .
the zero terms in the mass matrix could be taken as nonzero without changing the phenomenology discussed here as long as they are small compared to the terms shown .
also , the @xmath193 term could be chosen different from @xmath187 while still giving maximal mixing of and since maximal mixing results from the large value of the @xmath194 matrix element relative to the diagonal @xmath193 and @xmath195 elements , without any need for fine tuning of the difference @xmath196 . here
we choose to take the minimal form for @xmath191 needed to describe the data and then derive the associated consequences . to a good approximation , the two large eigenvalues of the mass matrix in eq .
( [ m ] ) are @xmath197 the values of the two small eigenvalues depend on the hierarchy of the @xmath198 .
for the three solar cases we have : @xmath199 the two small eigenvalues are then approximately given by @xmath200 \ , \
, , \label{caseb}\\ { \rm vlw } : & m_{0,1 } \simeq m \left[\pm \eps_2 + \eps_3 ^ 2\eps_4/2 \right ] \ , . \label{casec}\end{aligned}\ ] ] these approximate expressions for the eigenvalues have been obtained by multiplying each @xmath190 by powers of a hypothetical parameter @xmath201 , where the number of powers of @xmath201 assigned to each @xmath190 depends upon the ordering in eqs .
( [ ordera])-([orderc ] ) .
for example , in the sam case @xmath189 are multiplied by @xmath201 , @xmath186 and @xmath187 by @xmath202 , and @xmath188 by @xmath203 . then each eigenvalue is written as an expansion in powers of @xmath201 , the coefficients of which may be solved for by requiring that the expression @xmath204 reproduces the eigenvalue equation for the mass matrix order by order in @xmath201 .
once the coefficients are found , @xmath201 is set equal to unity .
the eigenvalues in all cases have the desired hierarchy @xmath205 , which gives the mass spectrum of the 2 + 2 model described in sec . 4.2 and depicted in fig . 1
. the small relative mass splitting of the heavier masses @xmath206 is governed entirely by the parameter @xmath187 : @xmath207 .
the lsnd @xmath208 oscillations are driven by the scale @xmath209 , the atmospheric @xmath1 oscillations are determined by @xmath210 , and the solar @xmath211 oscillations are determined by @xmath212 , the approximate expression for which can be obtained by eqs .
( [ casea])-([casec ] ) .
the charged - current eigenstates are approximately related to the mass eigenstates by @xmath213 where @xmath214 .
the dots indicate nonzero terms that are much smaller than the terms shown .
it is their smallness that suppresses mixing between and .
the mixing matrix @xmath34 depends on just three of the original five parameters ; it is independent of @xmath187 and the overall mass scale parameter @xmath185 .
note that @xmath110 and @xmath111 couple predominantly to @xmath19 and @xmath6 .
the nearly
degenerate @xmath123 and @xmath119 are seen to consist primarily of nearly equal mixtures of @xmath1 and @xmath18 .
these results , illustrated in fig . 2 ,
conform to the qualitative arguments of sec . 3 based on probability rectangles .
it is noted that this mixing matrix not only satisfies the approximate equalities of eqs .
( [ parallel23])([ue23 ] ) , but in fact replaces the approximate equalities , derived from parameter
independent arguments , with exact equalities to first order in @xmath190 .
inspection of the mixing matrix reveals that our model predicts saturation of eqs .
( [ aa ] ) and ( [ schwartz2 ] ) to this order , i.e. , @xmath215 . a small improvement in the measurement of @xmath134 or a modest improvement in the measurement of @xmath136
is predicted to show a positive disappearance signal .
another scenario , with solar @xmath27 and atmospheric @xmath29 oscillations , is readily obtained by interchanging @xmath216 and @xmath217 .
the mass matrix in the ( @xmath19 , @xmath6 , @xmath1 , @xmath18 ) basis is then @xmath218 the eigenvalues and parameter hierarchies are still given by eqs .
( [ m23])-([casec ] ) .
the mixing matrix is then given by @xmath219 where again @xmath220 . in the vlw case ,
the parameter @xmath186 is negligibly small if the solar oscillations are maximal , and can be taken as zero without affecting the phenomenology .
if this is done , then reference to the mass matrix shows that both @xmath6 and @xmath18 derive their masses entirely from flavor non diagonal couplings ,
and they are maximally mixed ( analogous to the @xmath1@xmath19 system ) .
also , if @xmath186 is taken as zero , then there are only four independent parameters needed in the mass matrix , and just two in the mixing matrix .
the derived @xmath43 parameter becomes @xmath221 , and the mixing matrix becomes very simple : @xmath222 a more general scenario which is a mixture of the previous two is for solar neutrinos to undergo both @xmath26 and @xmath27 oscillations .
this is easily parameterized by replacing the @xmath19 and @xmath18 states in eq .
( [ u ] ) with the rotated states @xmath223 and @xmath224 and defining ( c _
^ ) = ( cc & - + & ) ( c _ s + _ ) .
[ urot ] then the mass matrix in the ( @xmath19 , @xmath6 , @xmath1 , @xmath18 ) basis becomes @xmath225 and the matrix which diagonalizes @xmath191 is @xmath226 where @xmath220 as before .
for the mixing in eq .
( [ u ] ) ( when the solar oscillations are @xmath26 ) , the off - diagonal vacuum oscillation probabilities obtained from eq .
( [ oscprob ] ) , to leading order in @xmath190 for each @xmath227 and ignoring amplitudes smaller than @xmath228 , are @xmath229 \ ; , \label{prob1 } \\
p(\nu_e\leftrightarrow\nu_\tau ) & \simeq & \eps_3 ^ 2 \sin^2\delta_{32 } \ ; , \label{prob2 } \\
p(\nu_e\leftrightarrow\nu_s ) & \simeq & \sin^22\theta \sin^2\delta_{01 } \ ; , \label{prob3 } \\
p(\nu_\mu\leftrightarrow\nu_\tau ) & \simeq & \sin^2\delta_{32 } \ ; , \label{prob4 } \\
p(\nu_\mu\leftrightarrow\nu_s ) & \simeq & \sin^22\theta\eps_3 ^ 2 \sin^2\delta_{01 } \ ; , \label{prob5}\end{aligned}\ ] ] where @xmath230 due to the neutrino mass spectrum . in our model , only the
oscillation is suppressed beyond @xmath228 . for small @xmath15
only the leading oscillations @xmath231 contribute , and the only appreciable oscillation probability is @xmath232 where @xmath233 . from eq .
( [ lsndosc ] ) we can fix two model parameters @xmath234 since the only short baseline oscillation is @xmath235 , these models predict the equality of the @xmath6 disappearance probability , the @xmath1 disappearance probability , and the lsnd @xmath7 appearance probability in short baseline experiments . for @xmath15 typical to atmospheric or long baseline neutrino experiments ,
the oscillations in @xmath46 assume their average values .
the @xmath236 oscillation is now evident , and the non - negligible oscillation probabilities in vacuum are @xmath237 @xmath238 @xmath239 from eq .
( [ lb1 ] ) @xmath240 which determines another parameter of the model .
the model automatically gives maximal oscillations for atmospheric @xmath1 s , while oscillations in other channels are suppressed .
the @xmath1 maximal mixing is natural in the sense that it results from the large value of the @xmath194 matrix element relative to the diagonal @xmath193 and @xmath195 elements , without any need for fine tuning of the difference @xmath241 . finally , for very large @xmath242 km / gev , @xmath243 averages to @xmath244 and the appreciable oscillations in vacuum are @xmath245\ , , \\
p(\nu_e\leftrightarrow\nu_\tau ) \simeq & { 1\over2 } \eps_3 ^ 2 \ , , \\
p(\nu_\mu\leftrightarrow\nu_\tau ) \simeq & { 1\over2 } \;.\end{aligned}\ ] ] the solar data can then be explained if the parameters in vacuum satisfy @xmath246 in the three cases . in any of these scenarios in sec .
5.1 - 5.3 , the parameters @xmath185 , @xmath188 , @xmath189 , @xmath187 , and @xmath186 are obtained from the data in exactly the same way , i.e. , via eqs .
( [ lsndcon ] ) , ( [ atmpar ] ) , and ( [ sampar])-([vlwpar ] ) .
this is a consequence of the @xmath19@xmath18 ambiguity . in all cases ,
the parameters @xmath185 , @xmath189 , and @xmath187 are related to the observables by @xmath247 in the solar sector we have @xmath248 for the specific values @xmath249 and @xmath250 , @xmath251 and @xmath252 , and @xmath253 and @xmath254 , the model parameters are given in table [ params1 ] .
if we take instead @xmath255 and @xmath256 ( which gives the smallest value of @xmath65 allowed by the data ) , we get the model parameters shown in table [ params2 ] . in either of these two examples , the @xmath5 scale for the atmospheric neutrino oscillation
can be adjusted simply by varying @xmath187 . also in either case , the two heaviest masses provide relic neutrino targets for a mechanism that may generate the cosmic ray air showers observed above @xmath257 ev @xcite .
we note that the model parameters in tables [ params1 ] and [ params2 ] obey the hierarchies described in eqs .
( [ ordera])-([orderc ] ) .
if the solar oscillations are @xmath26 as described in sec . 5.1 , then our four - neutrino model predicts that the atmospheric oscillations are @xmath28 . on the other hand if the solar oscillations are @xmath27 as in sec .
5.2 , the atmospheric oscillations are @xmath29 .
several possibilities have been discussed to resolve the ambiguous assignment of @xmath18 and @xmath19 as the oscillation partners of the @xmath6 s in the sun and the @xmath1 s in the atmosphere .
the solar neutrino observatory ( sno ) @xcite , which can measure both charge - current ( cc ) and neutral - current ( nc ) interactions , will be able to test whether the solar @xmath6 s oscillate to sterile or active neutrinos : in the sterile case the cc / nc ratio in sno will be unity and both cc and nc rates will be suppressed from the ssm predictions , while in the active case only the cc rate is suppressed . of course if the cc measurement is consistent with the nc , one needs additional evidence to rule out the possibility that the ssm is in error . for instance , superkamiokande and sno can also accurately measure the shape of the @xmath258b neutrino spectrum , which would be distorted by oscillations .
also , a measurement of lower energy neutrinos , such as by the borexino experiment @xcite , could also be used to detect deviations from the ssm spectrum . turning to the atmospheric data , the possibilities to resolve the ambiguity center around the earth
matter effects which are possible in the - oscillation channel but not in the - channel ; there is a relative phase difference between @xmath1 and @xmath19 due to neutral current forward scattering , but there is no phase difference between @xmath1 and @xmath18 .
the analytical analysis of matter effects involving active and sterile neutrinos can be somewhat complicated @xcite , but the schrdinger like evolution equations can always be solved numerically @xcite .
other tests have been proposed recently to resolve the @xmath18@xmath19 ambiguity .
one test is to measure the asymmetry between downward - going and upward - coming events , for electrons and muons separately @xcite .
various oscillation scenarios give rise to dramatically differing trajectories of the asymmetries versus energy for muons and electrons .
the preliminary data from superk for the individual muon and electron asymmetries suggests again that the atmospheric anomaly is primarily due to @xmath1 oscillating into either @xmath18 or @xmath19 , but not @xmath6 . by eventually measuring an up - down asymmetry for neutral current ( nc ) events ( e.g. @xmath259 ) ,
the ambiguity can be resolved : for the @xmath18 case there is no nc asymmetry , whereas for the @xmath19 case there is a large nc asymmetry , as shown in ref .
the ratio of the rates of nc events relative to the charged current ( cc ) events can be also used to the same end @xcite , as can multi - ring events @xcite .
searches for muon - less events which come from @xmath18 , in association with a @xmath1 disappearance measurement , can also in principle distinguish between @xmath18 and @xmath19 @xcite . assuming that the solar oscillations are @xmath26 , we can determine the new oscillation signals predicted by the model . given the order of magnitude of the @xmath260 and @xmath261 , observable new phenomenology occurs for @xmath33 km / gev in the oscillation channels @xmath262 where @xmath263 is the oscillation amplitude which describes the lsnd results and @xmath264 is the oscillation argument which describes the atmospheric neutrino data .
we emphasize the new predictions in the @xmath265 and @xmath31 channels : long baseline oscillations with common oscillation length determined by the _
atmospheric _ @xmath266 and common amplitude given by @xmath267 times the _ lsnd _ amplitude @xmath122 .
these oscillations are in addition to the @xmath268 oscillations due to @xmath269 in eq .
( [ lsndosc ] ) , which average to the value of @xmath270 in a long baseline experiment .
the amplitudes and lengths of these new oscillations complement the set in table [ amplength ] , which are inevitable , given the present data , and are therefore required in any model . how can the oscillation probabilities in eqs .
( [ emu ] ) and ( [ etau ] ) be tested ?
a list of experiments currently underway or being planned to test neutrino oscillation hypotheses is given in table [ oscexp ] @xcite .
in each case the oscillation channel and the parameters which are expected to be tested are shown .
the minos experiment @xcite can detect @xmath7 or @xmath28 oscillations and is sensitive down to @xmath271 ev@xmath9 and a mixing amplitude of @xmath272 , which partially overlaps the region of interest ; see fig .
if the minos experiment can increase its sensitivity , it will provide an even better test of this new phenomenology .
long baseline experiments with an intense @xmath6 or @xmath67 neutrino beam and which can detect @xmath89 s can see the @xmath27 oscillations in eq .
( [ etau ] ) and provide a definitive test of the new phenomenology predicted by the model .
high intensity muon sources @xcite can provide simultaneous high intensity @xmath1 and @xmath67 ( or @xmath273 and @xmath6 for antimuons ) beams with well determined fluxes , which could then be aimed at a neutrino detector at a distant site .
it is expected that @xmath89 s will be detected through their @xmath74 decay mode and that a charge determination can be made , so that one can tell if the @xmath89 originated from @xmath274 or @xmath275 oscillations .
current proposals @xcite consider soudan ( @xmath276 732 km ) or gran sasso ( @xmath276 9900 km ) as the far site from an intense muon source at fermilab ( mc ) .
these experiments could also observe @xmath277 oscillations via detection of `` wrong sign '' muons , i.e. , those with sign opposite to that expected from the @xmath1 or @xmath273 source .
the neutrino energies are in the 1050 gev range . assuming that low backgrounds can be achieved , the sensitivity to @xmath5 is roughly proportional to the inverse square root of the detector size ( given the same neutrino energy spectrum at the source ) ; the @xmath16 sensitivity does not depend on detector distance @xmath278 because although the flux in the detector falls off with @xmath279 , the oscillation argument grows with @xmath279 for small @xmath280 . for 20 gev muons at fermilab and a 10 kt detector at either soudan or gran sasso , the single
event @xmath5 sensitivity for @xmath2 oscillations is about @xmath281 ev@xmath9 for maximal mixing @xcite . for large @xmath5 , the oscillation amplitude single
event sensitivity is roughly inversely proportional to the neutrino flux at the detector divided by the detector size ; about @xmath282 for soudan and @xmath272 for gran sasso @xcite . in general , the closer detector has comparable @xmath5 sensitivity but better @xmath50 sensitivity .
the model predicts @xmath27 oscillations with amplitude @xmath283 ( which ranges from 0.0006 to 0.01 ) and mass squared difference of @xmath126 ( which ranges from @xmath284 to @xmath285 ev@xmath9 ) .
the region of possible @xmath27 oscillations in our model and the regions which can be tested at the soudan and gran sasso sites with a neutrino beam from a high - intensity muon source at fermilab are shown schematically in fig .
8 , along with the favored parameters for the lsnd , atmospheric neutrino , and solar neutrino oscillations .
such experiments would be sensitive to some of the @xmath286 region , though they may not cover the low mass , small amplitude part .
these searches would also be able to test the @xmath277 oscillations in eq .
( [ emu ] ) and the atmospheric @xmath274 oscillations . additionally , long baseline experiments to the amanda detector @xcite from fermilab ( @xmath287 11700 km ) or kek ( @xmath287 11300 km ) may be useful in probing oscillations with small @xmath5 . if the solar oscillations are @xmath27 , then the oscillations of atmospheric neutrinos are @xmath29 and the new oscillations in eq .
( [ etau ] ) are instead @xmath26
. neither of these signals would be detectable in long baseline experiments since the signal is @xmath6 or @xmath1 disappearance at the few percent level or less .
the only measurable signal of the model in this case is the @xmath288 oscillations in eq .
( [ emu ] ) . if the solar neutrinos oscillate into both @xmath19 and @xmath18 as given by the mixing in eq .
( [ uprime ] ) , any vacuum oscillation in secs . 6.1 - 6.4 which has @xmath19 as the final state is replaced by oscillations to @xmath19 with relative probability @xmath289 and to @xmath18 with relative probability @xmath290 .
conversely , any oscillation which has @xmath18 as the final state is replaced by oscillations to @xmath19 with relative probability @xmath290 and to @xmath18 with relative probability @xmath289 . in particular , a solar @xmath6 oscillates into a mixture of @xmath19 and @xmath18 with relative probability @xmath289 and @xmath290 , respectively , in a vacuum , and an atmospheric @xmath1 oscillates into a mixture of @xmath19 and @xmath18 with relative probability @xmath290 and @xmath289 , respectively , in a vacuum . the new oscillation signal in eq .
( [ etau ] ) for long baseline experiments is replaced by @xmath291 also , there are new oscillations between @xmath19 and @xmath18 , with the general probability for any baseline given by @xmath292 \;.\end{aligned}\ ] ] since the @xmath294 mass matrix element is zero in eq .
( [ m ] ) , there is no neutrinoless double-@xmath293 decay at tree level in our model .
the present limit on @xmath294 from this process is @xmath2950.5 ev @xcite .
new experiments are under development which may measure @xmath294 down as low as 0.01 ev@xcite . if a nonzero @xmath294 is found at these levels , it would be incompatible with the solar solutions in our models .
if @xmath6 is primarily associated with the lighter pair in the 2 + 2 model , and @xmath205 , then there will be no measurable effect in the endpoint of the tritium decay spectrum .
since only mass - squared differences are important for oscillations , the inverted 2 + 2 model , where @xmath296 , is equally viable , although it is not derivable from the mass matrix in sec .
then the @xmath6 will have an effective mass of @xmath297 ev , which is just below the current limit @xcite . for @xmath298 1.4 ev , approximately the largest value allowed by the lsnd data ,
@xmath299 ev , which according to recent work on early universe formation of the largest structures provides an ideal hot dark matter component @xcite . for @xmath298 0.55
ev , the contribution to hot dark matter is much smaller .
the contribution of the neutrinos to the mass density of the universe is given by @xmath300 ev ) , where @xmath301 is the hubble expansion parameter in units of 100 km / s / mpc @xcite ; with @xmath302 our model implies @xmath303 .
an interesting test of neutrino masses is the power spectrum of early galaxy sizes , to be provided by the sloan digital sky survey ( sdss ) @xcite . for two
nearly degenerate massive neutrino species , sensitivity down to about 0.2 to 0.9 ev ( depending on @xmath304 and @xmath301 ) is expected , providing coverage of all or part of the lsnd allowed range ( @xmath3050.55 to 1.4 ev in our model ) .
also , the future map @xcite and planck @xcite satellite missions , which will measure the cosmic microwave background radiation , should be sensitive to neutrino densities to high precision @xcite , and in particular to the @xmath29 atmospheric or the @xmath26 large - angle msw neutrino mixing solutions @xcite . the curves in fig .
8 ( in the scenario where the solar oscillations are @xmath26 ) assume vacuum oscillations in the earth . in general
, large corrections to oscillations involving @xmath6 and @xmath19 are possible due to matter .
the @xmath6 diagonal element in the effective mass - squared matrix receives an additional term @xmath306 from its forward elastic cc interaction , and the @xmath19 diagonal element receives the contribution @xmath307 ( relative to the active neutrinos ) because it does not have nc interactions . here ,
@xmath84 and @xmath85 denote the electron and neutron number density , respectively . in table
[ eres ] we display the resonant energies in the earth for the various vacuum @xmath5 values suggested by the available lsnd , atmospheric , and solar data . for neutrinos with energies significantly above @xmath308 , oscillations are suppressed ; for neutrinos with energies significantly below @xmath308 , the matter effect is negligible ; at the resonant energy , @xmath309 and the oscillation length is increased by @xmath310 .
some of the resonant energy values in table [ eres ] are of particular interest .
upcoming neutrinos from the atmosphere or astrophysical sources , with mass at the lower end of the lsnd range , can have their oscillations resonantly enhanced by the earth s mantle and/or core .
atmospheric neutrinos below a few gev and the sam and lam @xmath311 neutrinos from the sun also appear to be near resonance in the earth s matter .
day night modulation of the solar flux due to earth
matter effects is expected to discriminate between and solar fluxes @xcite , while a precise measurement of zenith angle dependence may discriminate between and atmospheric fluxes @xcite .
oscillation wavelengths commensurate with the size of the earth s mantle and/or core are especially sensitive . in our model of sec . 5.1 with @xmath28 atmospheric neutrino oscillations
, however , these corrections do not significantly affect the large @xmath312 and @xmath313 mass eigenvalues as long as @xmath314 1 tev , and hence only modify the @xmath315 oscillation argument .
we have verified this by explicit diagonalization of the mass matrix when matter effects are included .
hence we conclude that the matter corrections to the mass matrix in eq .
( [ m ] ) probably have no observable consequences in all terrestrial experiments . for large @xmath15 , such as when @xmath316 mev for solar neutrinos , the only significant effect of matter is the usual msw enhancement of @xmath317 that leads to the solar neutrino suppression of the flux ; in all other oscillation channels the matter enhanced amplitudes are at the @xmath318 level or smaller . in the models discussed in secs .
5.2 and 5.3 , where there is a @xmath319 component to the atmospheric oscillation , matter effects as discussed here may be important in terrestrial experiments @xcite .
the vlw solar solution may be discriminated from the two msw solutions by a careful measurement of the solar neutrino spectrum by superk and borexino @xcite , or by determining the amount of seasonal variation of the @xmath320be and @xmath321 neutrinos @xcite , which can be measured by the borexino experiment .
the @xmath258b neutrino spectrum as measured in superk and sno will also be useful in discriminating between the sam and lam solutions @xcite .
the heron and hellaz @xcite experiments would be able to measure the @xmath311 neutrino energy spectrum , which would also be useful in differentiating the three scenarios .
the general case with - mixing is described in sec .
5.3 . the unmixed cases eqs . ( [ m ] ) and ( [ m2 ] )
are obtained in the limits @xmath322 and @xmath323 , respectively ; distinguishing between the unmixed scenarios is discussed in sec 7.1
. how might non - trivial mixing of and be observed , and how might the mixing angle be deduced ?
a mixed model would generally have a signature intermediate between the two unmixed signatures @xcite ; e.g. , experiments measuring neutral current events for solar and atmospheric neutrinos would find a result between those expected for @xmath18 and @xmath19 .
an analysis of all the available data ( short baseline lsnd , reactor and accelerator , long baseline atmospheric , and extraterrestrial length solar ) in terms of neutrino oscillations leads to the conclusion that three independent oscillation lengths are contributing .
this then further requires mixing of at least four light mass neutrinos .
for a four light mass neutrino universe , we draw the following model independent conclusions : + ( i ) the 1 + 3 ( or 3 + 1 ) mass spectrum with a separated mass is disfavored when all the data ( lsnd , reactor , accelerator , solar , and atmospheric ) are considered , leaving a spectrum with two nearly degenerate pairs as preferred ; + ( ii ) the neutrino mixing matrix elements satisfy @xmath324 if @xmath325 ev@xmath9 , i.e. , the parameters lie near the bugey @xmath6 disappearance limit ; + ( iii ) the relation @xmath326 is inferred from the near maximal mixing of atmospheric @xmath1 s measured by superk , which together with ( ii ) implies @xmath327 . based upon the apparent need for more than three light neutrinos , we have presented four - neutrino models with three active neutrinos and one sterile neutrino .
the models naturally have maximal @xmath28 ( or @xmath29 ) oscillations of atmospheric neutrinos and can also explain the solar neutrino and lsnd results .
the solar solutions can be @xmath26 or @xmath27 , and can be small - angle matter - enhanced , large - angle matter - enhanced , or vacuum long wavelength oscillations ; the increased statistics on the electron energy distribution and day / night differences of the superk data @xcite may further clarify the allowed regions for the solar solutions .
the models predict @xmath328 ( or @xmath26 ) and @xmath265 oscillations in long baseline experiments with @xmath33 km / gev with amplitudes that are determined by the lsnd oscillation amplitude and @xmath5 scale determined by the oscillation scale of atmospheric neutrinos . for the @xmath27 case
, these oscillations might be seen by experiments based on neutrino beams from an intense muon source at fermilab with a detector at the soudan or gran sasso sites .
the @xmath7 oscillations might be seen by the minos experiment or at kek with detectors at kamiokande and superkamiokande .
the models also predict the equality of the @xmath6 disappearance probability , the @xmath1 disappearance probability , and the lsnd @xmath7 appearance probability in short baseline experiments .
we thank d. caldwell , r. foot , p. krastev , j.g . learned , s. petcov , g. raffelt , j. stone , and r. volkas for comments and discussions .
this work was supported in part by the u.s .
department of energy , division of high energy physics , under grants no .
de - fg02 - 94er40817 , no .
de - fg05 - 85er40226 , and no .
de - fg02 - 95er40896 , and in part by the university of wisconsin research committee with funds granted by the wisconsin alumni research foundation , the vanderbilt university research council , the university of hawaii , and the max planck institute for physics , munich .
99 j.n . bahcall and m.h .
pinsonneault , rev .
phys . * 67 * , 781 ( 1995 ) ; j.n .
bahcall , s. basu , and m.h .
pinsonneault , astro - ph/9805135 ; the new solar flux calculation in the second reference has not yet been implemented in the solar neutrino analyses .
et al . _ ,
b ( proc . suppl . ) * 38 * , 47 ( 1995 ) ; kamiokande collaboration , y. fukuda _ et al . _ ,
lett , * 77 * , 1683 ( 1996 ) ; gallex collaboration , w. hampel _
et al . _ ,
b388 * , 384 ( 1996 ) ; sage collaboration , j.n .
et al . _ ,
lett . * 77 * , 4708 ( 1996 ) .
kamiokande collaboration , k.s .
et al . _ ,
b280 * , 146 ( 1992 ) ; y. fukuda _ et al .
b335 * , 237 ( 1994 ) ; imb collaboration , r. becker - szendy _ et al .
. suppl . *
38b * , 331 ( 1995 ) ; soudan-2 collaboration , w.w.m .
et al . _ ,
b391 * , 491 ( 1997 ) .
the super kamiokande collaboration , talk by e. kearns in ref .
@xcite and hep ex/9803007 ; y. fukuda et al .
, hep ex/9803006 and phys
* b * , to appear . the super kamiokande collaboration , y. fukuda et al . ,
hep ex/9805006 .
j.g . learned , s. pakvasa , and t.j .
weiler , phys .
lett . * b207 * , 79 ( 1988 ) ; v. barger and k. whisnant , phys . lett . *
b209 * , 365 ( 1988 ) ; k. hidaka , m. honda , and s. midorikawa , phys .
* 61 * , 1537 ( 1988 ) .
liquid scintillator neutrino detector ( lsnd ) collaboration , c. athanassopoulos _ et al .
_ , phys . rev
. lett . * 75 * , 2650 ( 1995 ) ; _ ibid . _ * 77 * , 3082 ( 1996 ) ; nucl - ex/9706006 . for a recent review , see talks at the itp conference on solar neutrinos : news about snus , santa barbara , december 1997 at + http://doug-pc.itp.ucsb.edu/online/snu/schedule.html c.y .
cardall and g.m .
fuller , nucl .
. suppl . *
51b * , 259 ( 1996 ) ; g.l .
fogli , e. lisi , d. montanino , and g. scioscia , phys .
* d56 * , 4365 ( 1997 ) .
a. acker , j.g .
learned , s. pakvasa and t.j .
weiler , phys .
b298 * , 149 ( 1992 ) ; a. acker and s. pakvasa , phys . lett .
* b397 * , 209 ( 1997 ) ; p.i . krastev and s.t .
petcov , phys . lett . *
b395 * , 69 ( 1997 ) .
johnson and d. mckay , hep ph/9805311
. v. barger , p. langacker , j. leveille , and s. pakvasa , phys .
45 * , 692 ( 1980 ) ; z.g .
berezhiani and r.n .
mohapatra , phys . rev . *
d52 * , 6607 ( 1995 ) ; e.j .
chun , a.s .
joshipura and a.y .
smirnov , phys .
b357 * , 608 ( 1995 ) , and phys . rev . * d54 * , 4654 ( 1996 ) ; k. benakli and a.y .
smirnov , phys .
lett * 79 * , 4314 ( 1997 ) ; j.r .
espinosa , hep
ph/9707541 ; g. cleaver , m. cvetic , j.r .
espinosa , l. everett , and p. langacker , phys .
rev . * d57 * , 2701 ( 1998 ) ; p. langacker , hep ph/9805281 .
lep electroweak working group and sld heavy flavor group , d. abbaneo _ et al .
_ , cern - ppe-96 - 183 , december 1996 .
caldwell and r.n .
mohapatra , phys . rev . *
d 48 * , 3259 ( 1993 ) ; j.t .
peltoniemi and j.w.f .
valle , nucl .
* b 406 * , 409 ( 1993 ) ; r. foot and r.r .
volkas , phys . rev . * d 52 * , 6595 ( 1995 ) ; ernest ma and probir roy , phys . rev . *
d 52 * , r4780 ( 1995 ) ; j.j .
gomez - cadenas and m.c .
gonzalez - garcia , z. phys .
* c 71 * , 443 ( 1996 ) ; n. okada and o. yasuda , int .
phys . * a 12 * , 3669 ( 1997 ) ; s.m .
bilenky , c. giunti , and w. grimus , hep ph/9711416 ; r.n . mohapatra , hep ph/9711444 ; n. gaur , a. ghosal , ernest ma , and probir roy , hep - ph/9806272 ; particularly strong enthusiasm for the sterile neutrino is expressed by d.o .
caldwell , hep
r. barbieri and a. dolgov , phys .
b237 * , 440 ( 1990 ) ; k. enqvist , k. kainulainen , and m. thomson , nucl . phys . *
b373 * , 498 ( 1992 ) ; x. shi , d.n .
schramm , and b.d .
fields , phys . rev . *
d 48 * , 2563 ( 1993 ) ; c.y .
cardall and g.m .
fuller , phys . rev . *
d 54 * , 1260 ( 1996 ) ; d.p .
kirilova and m.v .
chizhov , hep
ph/9707282 ; s.m .
bilenky , c. giunti , w. grimus and t. schwetz , hep ph/9804421 .
p.j . kernan and s. sarkar , phys . rev . *
54 * , r3681 ( 1996 ) ; s. sarkar , _ reports on progress in physics _ * 59 * , 1 ( 1996 ) ; k.a .
olive , talk at 5th international workshop on topics in astroparticle and underground physics ( taup 97 ) , gran sasso , italy , 1997 ; k.a .
olive , proc . of 5th international conference on physics beyond the standard model , balholm , norway , 1997 ; c.j .
copi , d.n .
schramm and m.s .
turner , phys .
75 * , 3981 ( 1995 ) ; k.a .
olive and g. steigman , phys .
b354 * , 357 ( 1995 ) .
r. foot and r.r .
volkas , phys .
lett . * 75 * , 4350 ( 1995 ) ; r. foot , m.j .
thompson , and r.r .
volkas , phys . rev . * d 53 * , 5349 ( 1996 ) ; n.f .
bell , r. foot , and r.r .
volkas , hep ph/9805259 .
r. foot and r.r .
volkas , phys . rev . *
d55 * , 5147 ( 1997 ) ; q.y .
liu and a. yu .
smirnov , hep
l. wolfenstein , phys . rev .
* d 17 * , 2369 ( 1978 ) ; s.p .
mikheyev and a. smirnov , yad .
fiz . * 42 * , 1441 ( 1985 ) ; nuovo cim . * 9c * , 17 ( 1986 )
rosen and j.m .
gelb , phys .
d 34 * , 969 ( 1986 ) .
parke , phys .
lett . * 57 * , 1275 ( 1986 ) ; s.j .
parke and t.p .
walker , phys .
57 * , 2322 ( 1986 ) ; w.c .
haxton , phys .
lett . * 57 * , 1271 ( 1986 ) ; t.k . kuo and j. pantaleone ,
61 * , 937 ( 1989 ) .
h. bethe , phys .
* 56 * , 1305 ( 1986 ) ; v. barger , r.j.n .
phillips , and k. whisnant , phys .
d 34 * , 980 ( 1986 ) ; v. barger , r.j.n .
phillips , and k. whisnant , phys . rev .
* d 24 * , 538 ( 1981 ) .
s.l . glashow and l.m .
krauss , phys .
b190 * , 199 ( 1987 ) .
v. barger , k. whisnant , d. cline , and r.j.n .
phillips , phys .
b93 * , 194 ( 1980 ) .
l. borodovsky _
et al . _ ,
lett . * 68 * , 274 ( 1992 ) .
karmen collaboration , b. bodmann _
* a553 * , 831c ( 1993 ) ; j. kleinfeller , nucl . phys . *
b48 * ( proc
. suppl . ) 207 ( 1996 ) ; talk by b. armbruster at 33rd rencontres de moriond : electroweak interactions and unified theories , les arcs , france , march 1998 .
71 * , 1965 ( 1993 ) .
y. declais _
et al . _ ,
b434 * , 503 ( 1995 ) .
f. dydak et al . , phys . lett . *
b134 * , 281 ( 1984 ) .
i.e. stockdale et al . ,
* 52 * , 1384 ( 1984 ) .
g. barr , t.k .
gaisser , and t. stanev , phys . rev . *
d 39 * , 3532 ( 1989 ) ; m. honda , t. kajita , k. kasahara , and s. midorikawa , phys . rev . * d52 * , 4985 ( 1995 ) ; v. agrawal , t.k .
gaisser , p. lipari , and t. stanev , phys . rev . * d 53 * , 1314 ( 1996 ) ; t.k .
et al . _ , phys .
d 54 * , 5578 ( 1996 ) .
garcia , h. nunokawa , o. peres , t. stanev , and j.w.f .
valle , hep - ph/9801368 .
chooz collaboration , m. apollonio _
et al . _ , hep - ex/9711002 .
e. akhmedov , p. lipari and m. lusignoli , phys .
* b300 * , 128 ( 1993 ) ; p. lipari and m. lusignoli , hep
ph/9712278 , and hep ph/9803440 ; r. foot , r.r .
volkas and o. yasuda , hep ph/9801431 .
n. hata and p. langacker , phys .
* d56 * , 6107 ( 1997 ) .
p. krastev and s.t .
petcov , phys .
lett . * 72 * , 1960 ( 1994 ) ; nucl . phys .
* b449 * , 605 ( 1995 ) . y. totsuka , talk at _ neutrino-98 _ , takayama , japan , june 1998 . v. barger , k. whisnant , s. pakvasa , and r.j.n .
phillips , phys . rev .
* d 22 * , 2718 ( 1981 ) .
p. langacker , j.p .
leveille , and j. sheiman , phys .
* d 27 * , 1228 ( 1983 ) . v. barger , n. deshpande , p.b .
pal , r.j.n .
phillips , and k. whisnant , phys .
d 43 * , 1759 ( 1991 ) ; s. bludman , d.c .
kennedy , and p. langacker , nucl .
b374 * , 373 ( 1992 ) . c. yanagisawa ,
talk given at _
pheno 98 _ , madison , wi , april 1998 .
v. barger , k. whisnant , and t.j .
weiler , hep
ph/9712495 and phys . lett .
* b * , to appear .
gibbons , r.n .
mohapatra , s. nandi , and a. raychaudhuri , hep
s. m. bilenky , c. giunti and w. grimus , eur .
j. c1 , 247 ( 1998 ) and hep ph/9607372 .
this was pointed out to us by g. raffelt .
weiler , hep
ph/9710431 , to appear in astropart .
e. norman _
_ , solar neutrino observatory ( sno ) collaboration , in proc . of _ the fermilab conference
: dpf 92 _ ,
november 1992 , batavia , il , ed .
by c. h. albright , p.h .
kasper , r. raja , and j. yoh ( world scientific , singapore , 1993 ) , p. 1450 .
borexino collaboration , c. arpesella _ et al .
_ , `` infn borexino proposal , '' vols .
i and ii , edited by g. bellini , r. raghavan _ et al . _ ( univ . of milan , 1992 ) ; j. benziger , f.p .
_ , `` proposal for participation in the borexino solar neutrino experiment , '' ( princeton university , 1996 ) .
balantekin , j.f .
beacom , and j.m .
fetter , hep - ph/9712390 .
j. flanagan , j. g. learned and s. pakvasa , phys .
d57 , r2649(1998 ) ; j. bunn , r. foot and r.r .
volkas , phys .
b413,109(1997 ) . j. g. learned , s. pakvasa and j. stone , hep
f. vissani and a. smirnov , hep
l.j . hall and h. murayama , hep
ph/9806218 .
a. curioni et al . , hep
ph/9805249 . for world
wide web links to more information on these and other neutrino oscillation experiments , see the neutrino oscillation industry web page at http://www.hep.anl.gov/ndk/hypertext/nuindustry.html .
minos collaboration , `` neutrino oscillation physics at fermilab : the numi - minos project , '' numi - l-375 , may 1998 .
s. geer , hep
s. barwick _
_ , amanda collaboration , in proc .
xxvith international conference on high energy physics , dallas tx , august 1992 , ed . by james r. sanford ( aip , new york , 1993 ) , p. 1250 ; f. halzen , astro
kleingrothaus , hep
ph/9802007 ; j. phys . *
g24 * , 483 ( 1998 ) .
a.i . belesev _
et al . _ ,
b350 * , 263 ( 1995 ) ; v.m .
lobashev et al . , in _ proc . of neutrino
98 _ , ed . by k. enqvist , k. huitu , and j. maalampi ( world scientific , singapore , 1997 ) . for a recent discussion
see j. primack , astro ph/9707285 .
kolb and m.s .
turner , _ the early universe _ ( addison - wesley , reading , 1990 )
. w. hu , d.j .
eisenstein , and m. tegmark , astro
see http://map.gsfc.nasa.gov for information on map .
see http://astro.estec.esa.nl/planck/ for information on planck .
ma and e. bertschinger , astrophys . j. * 455 * , 7 ( 1995 ) ; s. dodelson , e. gates , and a. stebbins , astrophys .
j. * 467 * , 10 ( 1996 ) ; r.e .
lopez , s. dodelson , a. heckler , and m.s .
turner , astro
s. hannestad and g. raffelt , astro ph/9805223 .
m. maris and s.t .
petcov , hep
ph/9803244 ; s.t petcov , hep ph/9805262 .
akhmedov , hep ph/9805272 .
liu , s.p .
mikheyev , and a.y .
smirnov , hep
b. faid , g.l .
fogli , e. lisi , and d. montanino , hep ph/9805293 .
v. barger , r.j.n .
phillips , and k. whisnant , phys .
lett . * 65 * , 3084 ( 1990 ) .
fogli , e. lisi , and d. montanino , hep ph/9803309 .
hellaz collaboration , f. arzarello et al .
, cern - laa/94 - 19 .
y. suzuki , talk at _ neutrino-98 _ , takayama , japan , june 1998 .
.resonant energies @xmath329 tev in the earth s core and mantle for @xmath6@xmath330 oscillations , for some typical vacuum values for the @xmath5 s and amplitudes suggested by the data . here @xmath50 is the oscillation amplitude and @xmath331 is avagadro s number .
we have taken the core electron density to be 4.5 to 6.0 @xmath332 , and the mantle density to be 1.6 to 2.6 @xmath332 .
resonant energies for @xmath6@xmath19 oscillations are @xmath333 times larger than for @xmath6@xmath330 oscillations , resonant energies for @xmath330@xmath19 oscillations are @xmath334 times larger , but resonance occurs for @xmath335 rather than @xmath336 since the phase difference due to matter has the opposite sign . [ cols="^,^,^,^,^ " , ] | we make a model independent analysis of all available data that indicate neutrino oscillations . using probability diagrams
, we confirm that a mass spectrum with two nearly degenerate pairs of neutrinos separated by a mass gap of @xmath0 ev is preferred over a spectrum with one mass eigenstate separated from the others .
we derive some new relations among the four neutrino mixing matrix elements .
we design four - neutrino mass matrices with three active neutrinos and one sterile neutrino that naturally incorporate maximal oscillations of atmospheric @xmath1 and explain the solar neutrino and lsnd results .
the models allow either a large or small angle msw or vacuum oscillation description of the solar neutrino deficit .
the models predict ( i ) oscillations of either @xmath2 or @xmath3 in long baseline experiments at @xmath4 km / gev , with amplitude determined by the lsnd oscillation amplitude and argument given by the atmospheric @xmath5 , and ( ii ) the equality of the @xmath6 disappearance probability , the @xmath1 disappearance probability , and the lsnd @xmath7 appearance probability in short baseline experiments .
6.5 in = cmssbx10 scaled 2 to * variations on four neutrino oscillations * + 0.7 cm v. barger@xmath8 , s. pakvasa@xmath9 , t.j .
weiler@xmath10 , and k. whisnant@xmath11 + @xmath8_department of physics , university of wisconsin , madison , wi 53706 , usa _
+ @xmath9_department of physics and astronomy , university of hawaii , manoa , hi 96822 , usa _ + @xmath10_department of physics and astronomy , vanderbilt university , nashville , tn 37235 , usa _ + @xmath11_department of physics and astronomy , iowa state university , ames , ia 50011 , usa _ + |
semileptonic @xmath3 ( @xmath4 ) decays provide the most precise determination of the cabibbo - kobayashi - maskawa ( ckm ) matrix element @xmath2 @xcite .
presently , the dominant uncertainty in the experimentally determined quantity @xmath5 arises from the unknown shape of the hadronic form factor @xmath6 , as a function of momentum transfer @xmath7 @xcite .
the data can be fit to both simplified pole models and series expansions , but the uncertainty inherent in these simplifications is difficult to estimate , and the difference between the resulting @xmath5 determinations represents a systematic error . constraining this shape from first principles and providing rigorous error estimates is an important problem .
the remaining ( and presently , dominant ) error in the @xmath4 determination of @xmath2 arises from the normalization , @xmath8 .
@xmath9 data ( specifically @xmath10 ) also constrain the shape of the scalar form factor , @xmath11 .
this can potentially reduce dominant errors in the theory normalization , either by avoiding an extrapolation from zero recoil in lattice determinations @xcite , or by fixing low - energy constants of chiral perturbation theory @xcite . at present , the comparison between theory and experiment , and between different experiments , is complicated by uncertainties in the form factor parameterization .
again , extracting as much information as possible from the experimental data without model assumptions is an important task .
this paper establishes rigorous bounds for the expansion coefficients appearing in a general parameterization of the semileptonic @xmath0 form factors .
the framework for the analysis is based on familiar arguments invoking analyticity and crossing symmetry @xcite .
however , in contrast to conventional arguments appealing to unitarity and the evaluation of an operator product expansion ( ope ) , bounds for the vector form factor are derived directly from experimental data for hadronic tau decays , @xmath1 .
the result is more stringent than can be obtained from the ope analysis , and applies to a more general class of parameterizations .
+ _ analyticity and convergence . _
form factors are defined as usual by the matrix element of the weak vector current @xmath12 : ( @xmath13 , @xmath14 ) @xmath15 @xmath16 and @xmath17 can be extended to analytic functions throughout the complex @xmath18 plane , except along a branch cut on the positive real axis starting at @xmath19 production threshold @xmath20 [ @xmath21 .
the cut plane is mapped onto the unit circle by @xmath22 where @xmath23 is the point mapping onto @xmath24 . and
an overall phase see e.g. @xcite . ]
the form factors are analytic in @xmath25 , and so may be expanded in a convergent power series : @xmath26 where @xmath27 is an as - yet arbitrary function analytic in @xmath25 , which may depend on one or more parameters , denoted generically ( and for reasons to become clear ) by @xmath28 .
the function @xmath29 sums an infinite number of terms , transforming the original series , naively an expansion involving @xmath30 , into a series with a much smaller expansion parameter .
for example , the choice @xmath31 minimizes the maximum value of @xmath32 occurring in the semileptonic region , and for this choice @xmath33 .
the function @xmath27 and the number @xmath34 may be regarded as defining a `` scheme '' for the expansion .
the expansion parameter @xmath32 and coefficients @xmath35 are then `` scheme - dependent '' quantities , with the scheme dependence dropping out in physical observables such as @xmath36 . neglecting terms beyond @xmath37 in ( [ eq : fexpand ] ) introduces a relative normalization error @xmath38 .
similarly , the error on the relative slope involves terms of order @xmath39 . since @xmath40 and @xmath41 , simple power counting in @xmath32 yields strong constraints on the impact of higher - order terms in the expansion , provided that the coefficients @xmath42 are well - behaved .
+ _ crossing symmetry and form factor bounds . _
it is important in practice to determine whether large `` order unity '' coefficients @xmath42 could upset the formal power counting in @xmath32 . to address this question , a norm may be defined via @xmath43 by crossing symmetry , the norm can be evaluated using form factors for the related process of @xmath19 production .
the following sections investigate bounds on the integral appearing on the right hand side of ( [ eq : norm ] ) .
to compare with unitarity predictions , and to motivate a choice of @xmath27 in ( [ eq : fexpand ] ) , we consider the correlation function , @xmath44 an unsubtracted dispersion relation can be written for the quantity : ( @xmath45 ) @xmath46 = { 1\over \pi } \int_0^\infty\ !
dt\ , { t { \rm im}\pi_1(t ) \over ( t+q^2)^3 } \,.\ ] ] assuming isospin symmetry , the contribution of all @xmath19 states to the ( positive ) spectral function @xmath47 is @xmath48^{3/2}\over t^3 } @xmath49 along the contour in ( [ eq : norm ] ) ] @xmath50 then yields the inequality : defined in @xcite is normalized to the leading ope prediction : @xmath51 . ]
[ note that @xmath52 @xmath53 for @xmath54 , @xmath55 can be reliably calculated using the ope in ( [ eq : ope ] ) .
collecting results from the literature @xcite , we have at renormalization scale @xmath56 : light flavors , in the modified minimal subtraction scheme . ]
@xmath57 \\ & & + { 8\pi^2\over q^4}\bigg [ -m_s\langle \bar{u}u \rangle - { \alpha_s\over 12\pi } \langle g^2\rangle + \dots \bigg ] + \dots \bigg\ } \ , , \nonumber\end{aligned}\ ] ] where corrections of order @xmath58 are neglected , and @xmath59 @xcite , @xmath60 @xcite .
the light band in fig . [ fig : vector ] shows the resulting bound on the quantity @xmath61 , setting @xmath62 and @xmath63 .
the perturbative uncertainty is estimated by varying @xmath64 from @xmath65 to @xmath66 , and allowing for higher - order contributions of relative size @xmath67 .
uncertainties from perturbative and power corrections are small above @xmath68 , but become significant below this scale .
the width of the band represents a @xmath69 contour obtained by adding uncertainties in perturbative and power corrections linearly .
= 7.7 cm = 7.7 cm although the ope breaks down at small @xmath70 , the norm in ( [ eq : norm ] ) remains perfectly well defined .
the dark band in fig .
[ fig : vector ] shows the result of evaluating the integral in ( [ eq : norm ] ) using @xmath1 decay data @xcite for the region @xmath71 , and estimating the remaining contribution from @xmath72 using perturbative qcd .
uncertainties from the @xmath73 decay data are conservatively estimated by taking the maximum ( minimum ) value of the weight function in ( [ eq : norm ] ) for each bin , and multiplying by the corresponding upper ( lower ) bound of the @xmath69 interval for the @xmath19 component of the spectral function measured in @xcite .
it may be noted that for the present purpose , there is no need to resolve the underlying resonance structure of the spectral function , nor to subtract the ( small ) @xmath74-wave contribution .
the remaining perturbative contribution is estimated using @xmath75 @xcite , conservatively setting @xmath76 , and assigning @xmath77 uncertainty to the result .
the effect is @xmath78 for all values of @xmath79 , and is insignificant when added in quadrature to the much larger contribution from the resonance region . as illustrated by fig .
[ fig : t0vector ] , the dependence of @xmath80 on @xmath34 is very mild .
here we again set @xmath81 ; this is strictly an overestimate , and accounts for much of the @xmath34 variation in the figure .
the norm for all schemes ( [ eq : phiplus ] ) with @xmath82 and @xmath83 satisfies @xmath84 more refined values for @xmath80 can be obtained from figs .
[ fig : vector ] and [ fig : t0vector ] at particular values of @xmath34 and @xmath70 .
the result from the ope represents an _ upper bound _ for @xmath80 . for large @xmath70 ,
the leading contributions to @xmath80 are @xmath85 , coming from the regions @xmath86 , and @xmath87 in ( [ eq : norm ] ) .
gives @xmath88 .
since @xmath89 is not very small , this does not lead to a large enhancement . for @xmath90
the result is simply @xmath91 , with dominant contributions from @xmath92 . ]
the ope overestimates @xmath80 by a factor @xmath93 . as fig .
[ fig : vector ] shows , although the ope evaluation becomes exceedingly precise , the resulting unitarity bound begins to wildly overestimate the norm of the form factor already at @xmath94 .
= 7.7 cm in the scalar case , an unsubtracted dispersion relation can be written for the quantity @xmath95 = { 1\over \pi}\int_0^\infty\!dt\ , { t { \rm im}\pi_0(t ) \over ( t+q^2)^2 } \,.\ ] ] noticing that @xmath96^{1/2}\over t^3 } |f_0(t)|^2 \theta(t - t_+ ) \,,\ ] ] and defining @xmath97 yields the inequality @xmath98 again , @xmath99 can be reliably calculated when @xmath100 @xcite : @xmath101 { \nonumber\\ } & & + { 1 \over q^4}\big[m_s\left ( \langle\bar{u } u\rangle - \langle \bar{s } s \rangle \right ) + \dots \big ] + \dots \,,\end{aligned}\ ] ] where @xmath102 @xcite , @xmath103 @xcite . the light band in fig . [ fig : scalar ] shows the @xmath69 contour obtained by adding errors from all sources linearly , as described after ( [ eq : chi1 ] ) ; in this case , higher - order contributions of relative size @xmath104 are included .
, the result for @xmath105 is consistent with a related study in @xcite , where however uncertainties due to the poor convergence of the perturbation series , and the effects of power corrections were not considered . ]
the behavior of the unitarity bound with respect to @xmath70 is milder than in the vector case . for large @xmath70 ,
the leading contributions to @xmath106 are @xmath85 , coming from the regions @xmath86 and @xmath87 in ( [ eq : norm ] ) . , @xmath107 , with dominant contributions from @xmath108 .
] again , it should be remembered that the ope gives an _ upper bound _ to @xmath106 , although for this case the bound is overestimated by only a single power of @xmath109 .
while it is not possible to rigorously bound the norm of @xmath110 using the ope for schemes with @xmath111 , there is no parametric enhancement of the integral in ( [ eq : azerodef ] ) for these @xmath70 values .
a conservative bound on the norm for all schemes with @xmath82 and @xmath83 is @xmath112 in principle , this bound could be improved by extracting the @xmath19 component of the scalar spectral function from @xmath1 .
to implement the expansion ( [ eq : fexpand ] ) we must choose a `` scheme '' specified by the functional form of @xmath27 , and by the values of @xmath28 and @xmath34 .
possible candidates for @xmath27 are @xmath113 , or @xmath27 given by ( [ eq : phiplus ] ) and ( [ eq : phizero ] ) .
the latter choice ensures that the norm in ( [ eq : norm ] ) is independent of @xmath34 and relates in a simple way to the spectral functions extracted from @xmath73 decay ; with , say @xmath114 , this choice also allows comparison to ope bounds
. in ( [ eq : phiplus ] ) and ( [ eq : phizero ] ) also ensures that @xmath115 does not scale as some power of the heavy - quark mass @xcite . ] the value of @xmath34 can be selected according to various criteria .
for example , the choice @xmath31 minimizes the maximum value of @xmath32 throughout the semileptonic range , while @xmath62 simplifies the translation between coefficients @xmath35 and the derivatives of the form factor at @xmath116 .
alternatively , the choice of @xmath34 can be used to minimize correlations between shape parameters .
when only a single form factor contributes ( as in @xmath117 decays ) , the error correlations between expansion coefficients @xmath118 measured in an experiment with near ideal acceptance and resolution are completely determined without a priori knowledge of the form factor shape . with @xmath119 where @xmath120 are partial rates measured in intervals of @xmath18
, it follows that the error matrix @xmath121 determined from a @xmath122 fit is : ( here @xmath123 is the pion momentum in the kaon rest frame ) @xmath124 if the @xmath117 data is fit to a normalization proportional to @xmath8 and two shape parameters @xmath125 and @xmath126 , the correlation between @xmath127 and @xmath128 vanishes in the ideal case for @xmath129 ( with @xmath130 as in ( [ eq : phiplus ] ) and @xmath131 ) .
is insensitive to the precise form factor shape , e.g. whether @xmath132 is assumed constant , or is input from experimental data . since ref .
@xcite measures a distribution in transverse momentum , @xmath133 , the partial rates @xmath134 involve a smearing over different @xmath18 values . taking this into account
, the correlations remain very insensitive to the form factor shape , and are minimized in the ideal case for @xmath135 ( at @xmath136 . ]
only the vector form factor is relevant for the electron mode . from ( [ eq : vbound ] ) and @xmath137 [ at @xmath31 ] , the quadratic term , @xmath139 in ( [ eq : fexpand ] ) , can affect the normalization by an amount not greater than @xmath140 . to establish a precision significantly smaller than this , @xmath141 must be constrained by data .
the cubic term is bounded by @xmath142 .
these errors can also be robustly estimated by including higher order terms in fits to the experimental data , under the constraint ( [ eq : vbound ] ) .
extraction of other observables , such as the form factor slope and curvature at @xmath116 can be treated similarly .
the simple taylor expansion in @xmath18 converges much more slowly than the @xmath32 expansion in ( [ eq : fexpand ] ) , and it is difficult to reliably bound the errors associated with higher - order terms in fits to the truncated @xmath18 series . the systematic approach advocated here should remedy this situation , and make a more definitive comparison between different experiments possible .
-decay data in @xcite ( and a negligible perturbative contribution ) to obtain @xmath143 . ]
the scalar form factor can be probed in the muon mode . with enough precision , constraining
the slope of this form factor could aid lattice and chiral perturbation theory estimates of @xmath144 by eliminating extrapolation errors @xcite , or by determining higher - order constants in the chiral lagrangian @xcite . in a general scheme : ( the prime denotes derivative with respect to @xmath18 ) @xmath145 present experiments do not strongly constrain the coefficients beyond the linear @xmath146 term in ( [ eq : fexpand ] ) . from ( [ eq : sbound ] ) and @xmath137 it is straightforward to estimate the resulting uncertainty on the slope : @xmath147 \approx a_{f_0 } \times 0.001 < 0.005 $ ] .
this error can also be robustly estimated by including higher order terms in fits to the experimental data , under the constraint ( [ eq : sbound ] ) . to further constrain the slope
requires either more precise semileptonic data , or independent constraints on the form factor .
one type of constraint appeals to symmetry principles , in particular at the callan - treiman point where @xmath148 with a controllably small correction , @xmath149 @xcite .
unfortunately , this point is too far removed from @xmath116 to have a very strong influence on the extraction of @xmath150 . )
yields a more precise value for the slope than is obtained from combining unitarity with the callan - treiman point in the absence of data , cf . @xcite . ] as a second type of constraint , the phase of the form factor above @xmath19 threshold can be determined from @xmath19 scattering .
even supposing arbitrarily precise scattering data were available up to the inelastic threshold , @xmath151 , it is easy to see that dramatic constraints are not expected in the semileptonic region . since the additional data relates to the form factor at points along the cut in the @xmath18-plane , where @xmath152 , constraints from the scattering data can be satisfied by tuning higher - order coefficients @xmath35 in ( [ eq : fexpand ] ) , which have relatively minor impact in the semileptonic region .
more quantitatively , with arbitrarily precise data , a suitable phase redefinition of the form factor ( a particular choice of @xmath27 ) can postpone the branch point in @xmath153 until @xmath154 . replacing @xmath155 with @xmath154 in ( [ eq : zdef ] )
, we see that the maximum size of the effective expansion parameter can then be made as small as @xmath156 throughout the semileptonic region , compared to @xmath157 when the branch point appears at @xmath155 .
thus while some improvement is possible by directly incorporating the experimental scattering data , more dramatic gains in precision require further dynamical inputs or assumptions @xcite .
the general parameterization ( [ eq : fexpand ] ) , with constraints ( [ eq : vbound ] ) and ( [ eq : sbound ] ) , can be used to systematically analyze the @xmath9 data without model assumptions .
this should eliminate a dominant uncertainty in the experimental determination of @xmath5 , and allow for a more definitive comparison of shape observables measured by different experiments .
@xmath4 decays provide a unique opportunity to investigate the convergence properties of the expansion ( [ eq : fexpand ] ) .
precision data can be used to directly constrain the first few coefficients , and the existence of a heavy lepton , @xmath158 , makes it possible to establish bounds on the dominant vector form factor that are both more stringent and more generally applicable than those obtained from an ope analysis .
this provides a direct test of scaling arguments that apply also in the charm and bottom systems @xcite .
_ acknowledgements . _
the author acknowledges a number of fruitful discussions with a. glazov and r. kessler .
thanks also to t. becher for comments on the manuscript , and m. davier for a useful remark concerning the details of ref .
fermilab is operated by universities research association inc . under contract
de - ac02 - 76ch03000 with the u.s .
department of energy . for
a review see : e. blucher _ et al .
_ , hep - ph/0512039
. t. alexopoulos _ et al . _
[ ktev collaboration ] , phys .
d * 70 * , 092007 ( 2004 ) .
a. lai _ et al . _ [ na48 collaboration ] , phys .
b * 604 * , 1 ( 2004 ) .
o. p. yushchenko _ et al .
_ , phys . lett .
b * 589 * , 111 ( 2004 ) .
f. ambrosino _ et al . _
[ kloe collaboration ] , phys .
b * 636 * , 166 ( 2006 ) .
d. becirevic _ et al .
_ , nucl .
b * 705 * , 339 ( 2005 ) .
j. bijnens and p. talavera , nucl . phys .
b * 669 * , 341 ( 2003 ) .
for further references , see : j. bijnens , hep - ph/0604043 .
see e.g. : c. bourrely , b. machet and e. de rafael , nucl .
b * 189 * , 157 ( 1981 ) , and references therein .
early applications to heavy - meson decays and related developments can be found in : c. g. boyd , b. grinstein and r. f. lebed , phys .
lett . * 74 * , 4603 ( 1995 ) ; l. lellouch , nucl .
b * 479 * , 353 ( 1996 ) ; i. caprini and m. neubert , phys .
b * 380 * , 376 ( 1996 ) .
l. v. ahlfors , _ complex analysis , _ mcgraw - hill , 1953 .
t. becher and r. j. hill , phys .
b * 633 * , 61 ( 2006 ) .
s. g. gorishnii , a. l. kataev and s. a. larin , phys .
b * 259 * , 144 ( 1991 ) .
l. r. surguladze and m. a. samuel , phys . rev .
lett . * 66 * , 560 ( 1991 ) [ erratum - ibid .
* 66 * , 2416 ( 1991 ) ] . k. g. chetyrkin , phys .
b * 391 * , 402 ( 1997 ) .
s. c. generalis and d. j. broadhurst , phys .
b * 139 * , 85 ( 1984 ) .
q. mason , h. d. trottier , r. horgan , c. t. h. davies and g. p. lepage [ hpqcd collaboration ] , phys .
d * 73 * , 114501 ( 2006 ) .
s. narison , hep - ph/0202200 .
r. barate _
[ aleph collaboration ] , eur .
j. c * 11 * , 599 ( 1999 ) .
the spectral function in fig . 9 of this reference is related to @xmath159 in ( [ eq : pi1 ] ) by @xmath160 .
g. p. lepage and s. j. brodsky , phys .
d * 22 * , 2157 ( 1980 ) . k. g. chetyrkin , phys .
b * 390 * , 309 ( 1997 ) .
r. j. hill , phys .
d * 73 * , 014012 ( 2006 ) .
j. gasser and h. leutwyler , nucl .
b * 250 * , 517 ( 1985 ) .
for some recent analysis along these lines , see : m. jamin , j. a. oller and a. pich , hep - ph/0605095 . v. bernard , m. oertel , e. passemar and j. stern , hep - ph/0603202 . c. bourrely and i. caprini , nucl .
b * 722 * , 149 ( 2005 ) . | rigorous bounds are established for the expansion coefficients governing the shape of semileptonic @xmath0 form factors .
the constraints enforced by experimental data from @xmath1 eliminate uncertainties associated with model parameterizations in the determination of @xmath2 .
the results support the validity of a powerful expansion that can be applied to other semileptonic transitions . |
in @xcite and @xcite topological invariants for closed oriented three manifolds and cobordisms between them are defined by using a construction from symplectic geometry .
the resulting floer homology package has many properties of a topological quantum field theory .
the construction of heegard floer homology is more combinatorial in flavor than the corresponding gauge theoretical constructions of donaldson - floer ( see @xcite ) and seiberg - witten theories ( see @xcite , @xcite , @xcite ) .
however , the construction still depends on profoundly analytic objects - holomorphic disks . as a result ,
one is tempted to consider classes of manifolds which allow a completely combinatorial description of their heegaard floer homologies . in @xcite a class of plumbing three - manifolds is studied .
it is proved there that @xmath0 of these manifolds can be expressed in terms of equivariant maps .
we extend the results of this paper to get a similar description for negative - semidefinite plumbed manifolds with one bad vertex and @xmath2 . to formulate the main theorems we first give required preliminaries .
let g be a weighted graph and let @xmath3 and @xmath4 be respectively the weight and the degree of the vertex @xmath5 .
we denote by @xmath6 the four - manifold with boundary having @xmath7 as its plumbing diagram .
let @xmath8 be the oriented three - manifold which is the boundary of @xmath6 . for @xmath9 ,
the group @xmath10 is the lattice freely spanned by vertices of @xmath7 .
denoting by @xmath11 $ ] the homology class in @xmath10 corresponding to the vertex @xmath5 of @xmath7 , the values of the intersection form of @xmath12 on the basis are given by @xmath11{\cdot}[v ] = m(v)$ ] ; @xmath11{\cdot}[w ] = 1 $ ] if @xmath13 is an edge of @xmath7 and @xmath11{\cdot}[w]=0 $ ] otherwise . the graph @xmath7 is called _ negative-(semi)definite _ if the form is negative-(semi)definite .
a vertex @xmath5 is said to be a _
bad vertex _ of @xmath7 if @xmath14 .
let @xmath15 denote the @xmath16$]-module which is the quotient of @xmath17 $ ] by the submodule @xmath18 $ ] , graded so that the element @xmath19 ( for @xmath20 ) is supported in degree @xmath21 .
define @xmath22 $ ] . denoting by @xmath23 the set of characteristic vectors for
the intersection form define @xmath24 to be the set of finitely supported functions satisfying the following relations for all characteristic vectors @xmath25 and vertices @xmath5 : @xmath26 ) = \phi(k),\ ] ] if @xmath27 and @xmath28 ) = u^{-n}{\cdot}\phi(k)\ ] ] for @xmath29 .
we can decompose @xmath30 according to @xmath31 structures over @xmath32 .
note first that the first chern class gives an identification of the set of @xmath31 structures over @xmath9 with the set of characteristic vectors @xmath23 .
observe that the image of @xmath33 in @xmath34 is spanned by the poincar duals of the spheres corresponding to the vertices .
using the restriction to boundary , it is easy to see that the set of @xmath31 structures over @xmath32 is identified with the set of @xmath35-orbits in @xmath23 . fix a @xmath31 structure @xmath36 over @xmath32 .
let @xmath37 denote the set of characteristic vectors for @xmath12 which are first chern classes of @xmath31 structures @xmath38 whose restriction to the boundary is @xmath36 .
similarly , we let @xmath39 be the subset of maps which are supported on the subset of characteristic vectors @xmath40 .
we have a direct sum splitting : @xmath41 for a negative definite graph @xmath7 a grading on @xmath30 is introduced as follows : we say that an element @xmath42 is homogeneous of degree @xmath43 if for each characteristic vector @xmath25 with @xmath44 , @xmath45 is a homogeneous element with : @xmath46 the following theorem is proved in @xcite : [ intro : someplumbings ] let @xmath7 be a negative - definite weighted forest with at most one bad vertex .
then , for each @xmath31 structure @xmath36 over @xmath47 , there is an isomorphism of graded @xmath16 $ ] modules , @xmath48 we formulate a similar statement for negative - semidefinite trees with one bad point . in this case @xmath2 and so we have both torsion and non - torsion @xmath31 structures on @xmath32 .
note that for @xmath49 with @xmath36 torsion , @xmath50 is well defined . as a result
, we can introduce a grading on @xmath51 as before with replaced by @xmath52 for a torsion @xmath31 structure ozsvth - szab floer homology groups of manifolds with @xmath2 have absolute grading by half integers .
this grading lifts absolute @xmath53 grading of the same groups : @xmath54 modulo @xmath55 of the first correspond to the even degree of the latter and @xmath56 modulo @xmath55 corresspond to the odd degree .
the following theorem is an analogue of theorem [ intro : someplumbings ] for the negative - semidefinite case .
[ intro : main ] let @xmath7 be a negative - semidefinite weighted forest with at most one bad vertex .
then , for each @xmath31 structure @xmath36 over @xmath47 , there is an isomorphism of graded @xmath16 $ ] modules , @xmath57 for non - torsion @xmath31 structure @xmath36 we have @xmath58 while if @xmath36 is torsion then @xmath59 for some @xmath60 .
let @xmath36 be a torsion @xmath31 structure on a three - manifold @xmath32 with @xmath61 .
correction terms @xmath62 are defined as the minimal grading of any non - torsion element in the image of @xmath63 in @xmath64 with grading @xmath65 modulo 2 . for the type of manifolds we are considering we see that coditions of being non - torsion and lying in the image of @xmath66 are superficial .
an important property of these correction terms is given by @xmath67 see @xcite , section 4.2 .
it follows from theorem [ intro : main ] that @xmath68 and @xmath69 the above result gives a practical calculation of @xmath70 : for a given @xmath71 , it is easy to see that the minimum in question is always achieved among the finitely many characteristic vectors @xmath72 with latexmath:[\[|k{\cdot}v| \leq in section 3.1 . ) for our calculations of @xmath0 to be completed we need to specify @xmath74 for each torsion @xmath31 structure @xmath36 .
sometimes one can use truncated euler characteristic of @xmath0 , calculated from turaev torsion , see @xcite , section 10.6 .
another way is to consider a negative - semidefinite forest @xmath75 corresponding to the manifold @xmath76 .
if @xmath75 has only one bad vertex , then one can use properties of correction terms to finish the calculation .
the above results are proved in section 2 . in section 3
we discuss the calculation of @xmath30 .
we end this paper with several sample calculations .
* acknowledgments*no words can express my gratitude to my advisor zoltn szab who kindly and patiently taught almost everything i know on heegaard floer homology and beyond .
i would like to thank paul seymour for his invaluable support and encouragement .
let us postpone consideration of @xmath78 coefficients and concentrate on the case of @xmath79 coefficients . thus ,
unless stated otherwise , all floer homology groups in this subsection are meant to be taken with @xmath80 coefficients(which we supress from the notation ) .
in particular , @xmath15 now denotes the quotient of @xmath81 $ ] by the submodule @xmath82 $ ] .
define a map @xmath83 as follows .
the plumbing diagram gives a cobordism from @xmath47 to @xmath84 .
now let @xmath85 be the map given by @xmath86 where @xmath87 is the @xmath31 structure whose first chern class is @xmath25 , and @xmath88 denotes the four - dimensional cobodism invariant defined in @xcite .
now we state a more precise statement of the first part of therem [ intro : main ] .
[ thm : odd ] let @xmath7 be a negative - semidefinite tree with at most one bad vertex .
then , @xmath89 induces a grading - preserving isomorphism : @xmath90 the following proposition is proved in the same way as in @xcite .
[ prop : setup ] the map @xmath91 induces an @xmath92$]-equivariant , degree - preserving map from @xmath93 to @xmath94 whose image lies in @xmath95 if @xmath7 is a weighted graph with a distinguished vertex @xmath96 , let @xmath97 be a new graph formed by introducing one new vertex @xmath98 labelled with weight @xmath99 , and connected to only one other vertex , @xmath5 .
let @xmath100 denote the weighted graph whose underlying graph agrees with @xmath7 , but whose weight at @xmath5 is increased by one ( and the weight stays the same for all other vertices ) .
the two three - manifolds @xmath101 and @xmath102 are diffeomorphic .
the following long exact sequence follows from theorem [ holdisktwo : thm : generalsurgery ] of @xcite @xmath103 here the maps @xmath104 , @xmath105 , and @xmath106 are induced by two - handle additions .
note that @xmath104 changes @xmath53 grading , while the other two maps preserve it .
without loss of generality we can suppose that our negative - semidefinite forest with one bad vertex is @xmath100 for negative - definite forest @xmath7 with ( at most ) one bad vertex .
corollary 1.4 of @xcite states that for negative - definite trees with at most one bad vertex @xmath0 is supported in even degrees , and so , for example , @xmath107 is 0 . from previous exact sequence
we get @xmath108 corressponding to maps @xmath104 , @xmath105 maps @xmath109 and @xmath110 defined in @xcite so that the following diagram commutes : @xmath111 moreover @xmath109 is injective , and @xmath112 the proof of these facts proceeds in the same way as the proof of lemma 2.10 of @xcite .
gathering all of these together we have the diagram : @xmath113 where the maps are indicated as isomorphisms when it follows from @xcite . from the diagram it follows that @xmath114 is an isomorphism .
this concludes the proof of theorem [ thm : odd ] . to complete the proof of the main theorem we need to calculate @xmath115
this will be done separately for the torsion and non - torsion case in following two lemmas .
if @xmath36 is a non - torsion @xmath31 structure then @xmath116 let @xmath5 be the bad vertex .
we have a surjection @xmath106 from @xmath117 to @xmath118 .
since @xmath119 does not have any bad vertices its homology is given by the direct sum of @xmath120 s , with one summand with its own @xmath43 for each @xmath31 structure on @xmath121 .
this clearly means that for any integer @xmath122 and an element @xmath123 there exists another element @xmath124 so that @xmath125 . since @xmath36 is non - torsion , then there exists an integer @xmath126 such that action of @xmath127 on @xmath128 is zero . taking into account
that @xmath106 is @xmath129-equivariant and surjective proves the lemma .
if @xmath36 is torsion then @xmath130 for some integer d. given @xmath2 we know the group @xmath66 and so here it is enough to show that for any integer @xmath122 and an element @xmath131 there exists another element @xmath124 so that @xmath132 this again follows from the equivariance and surjectivity of @xmath106 .
since the sequence [ eq : exactseq ] holds with @xmath78 coefficients then the previous two lemmas are also true over @xmath78 , with the proofs being the same .
it remains to figure out @xmath133 .
@xmath134 over @xmath78 is torsion free .
the lemma easily follows from the sequence [ eq : exactseq ] over @xmath78 and the fact that @xmath135 is torsion free , see @xcite .
for calculational purposes it is helpful to adopt a dual point of view .
let @xmath136 be the set of equivalence classes of elements of @xmath137 ( and we write @xmath138 for the pair @xmath139 ) under the following equivalence relation .
for any vertex @xmath5 let @xmath140 if @xmath141 , then @xmath142 ) \sim u^m\otimes k,\ ] ] while if @xmath143 , then @xmath144 ) \sim u^{m - n}\otimes k.\ ] ] starting with a map @xmath145 consider an induced map @xmath146 defined by @xmath147 clearly , the set of finitely - supported functions @xmath148 whose induced map @xmath149 descends to @xmath136 is precisely @xmath30 .
a _ basic element _ or _ basic vector _ of @xmath136 is one whose equivalence class does not contain any element of form @xmath138 with @xmath150 .
given two non - equivalent basic elements @xmath151 and @xmath152 in the same @xmath31 structure , one can find positive integers @xmath122 and @xmath153 such that @xmath154 if , moreover , the numbers @xmath122 and @xmath153 are minimal then this relation will be called the _ minimal relationship _ between @xmath155 and @xmath156 .
[ rem : matrix ] let @xmath157 be the incidence matrix of our graph , with the diagonal elements equal to the weights of the corresponding vertices . note that @xmath25 is torsion if and only if it is in the range of @xmath157 . as a result , it is possible to set ( without ambiguity ) @xmath158 for any @xmath159 satisfying @xmath160 non - torsion vectors lie in the complement of the range of @xmath157 .
it is an exercise in linear algebra to show that for any vector @xmath25 corressponding to a non - torsion @xmath31 structure there exist non - negative integers @xmath122 and @xmath153 , @xmath161 such that @xmath162 such relations with minimal @xmath153 and @xmath122 for non - torsion basic vectors @xmath25 are also included among the minimal relationships ; we define the _ height _ of @xmath25 to be minimal such @xmath122 . on
can see that @xmath136 is specified as soon as one finds its basic elements and all of the minimal relationships .
we describe now a modification of the algorithm for calculating the basic elements given in @xcite to the case of negative - semidefinite trees .
construct a sequence of vectors @xmath164 , where @xmath165 is obtained from @xmath166 by choosing any vertex @xmath167 with @xmath168 and then letting @xmath169.\ ] ] note that any two vectors in this sequence are equivalent .
if the sequence is infinite then it can be made periodic .
one just have to choose an ordering on the vertices of @xmath7 .
the ordering is used to eliminate the indeterminacy when there are several choices for @xmath173 while constructing the sequence .
the proof is similar to the proof of proposition 3.2 of @xcite .
the only new ingredient is that the sequence can become periodic because of the tree being semidefinite . in this case , since one can also subtract @xmath179 s instead of adding them , it follows that the sequence is purely periodic . as a consequence ,
the initial vector is basic . to find the non - torsion basic elements we consider all @xmath25 coressponding to non - torsion @xmath31 structures with @xmath180 then construct a sequence of vectors in the same way as for torsion case .
this time the sequence can not be periodic and it will terminate in one of the two ways described above .
the equivalence classes in @xmath181 which have no representative of the form @xmath178 with @xmath150 are in one - to - one correspondence with non - torsion vectors @xmath25 satisfying inequality for which the previous sequence terminates in a characteristic vector @xmath175 satisfying inequality and for which there is no positive integer @xmath182 with @xmath183 the proof is similar to the proof of proposition 3.2 of @xcite . however , since @xmath50 can not be defined in this case , the last part of the cited proof can not be carried out . as a result , to get a one - one correspondence one has to exclude initial vectors @xmath25 which satisfy @xmath184 for some positive integer @xmath182 .
* 0-surgery on @xmath185 torus knot*we give here another calculation of @xmath186 for a class of negative - semidefinite plumbed manifolds @xmath187 , where @xmath122 is a positive integer and @xmath187 is obtained as 0-surgery on @xmath185 torus knot , compare @xcite .
the plumbing diagram of @xmath188 is depicted in figure [ fig : example1 ] .
the diagram for @xmath187 is similar , except it has @xmath189 of @xmath190 s in the middle strand , and instead of @xmath191 we have @xmath192 .
we will write the elements of @xmath30 as row vectors with the first four coordinates corresponding to the vertices with weights @xmath99 , @xmath190 , @xmath193 and @xmath194 respectively , and all remaining entries corresponding to @xmath190 s on the middle strand ordered by the distance from the root starting with the closest one .
let us start eith @xmath0 for @xmath188 .
one calculates the basic vectors to be @xmath202 @xmath203 @xmath204 @xmath205 is the only torsion vector , while @xmath206 and @xmath155 correspond to @xmath207 .
one can show that @xmath208 , i.e. the height of @xmath209 is equal to 1 .
it follows at once that @xmath210/u^1,\ ] ] supported in the odd degree , and @xmath211 for all other non - torsion @xmath36 .
we have @xmath212 from the second part of the main theorem it follows that @xmath213 for some @xmath43 .
since @xmath188 is a 0-surgery on ( 2,5 ) torus knot , we can easily calculate the truncated euler characteristic of @xmath214 from the alexander polynomial , see @xcite .
in fact , @xmath215 and so @xmath216 .
[ lemma : rel ] the following relationships are satisfied for non - torsion basic vectors : @xmath221 @xmath222 where @xmath223 .
these relationships are minimal , i.e. height of @xmath224 is equal to @xmath201 .
* another example*let us consider the plumbing diagram depicted in the figure [ fig : example2 ] .
we will use shorthand @xmath232 . note that there is another negative - semidefinite forest @xmath75 with one bad vertex satisfying @xmath233 .
picture of @xmath75 is also star shaped , with @xmath193 in the center , and four strands coming out of the center , first strand contains one @xmath190 , second - two @xmath190 s , third - six @xmath190 s and the last one with forty one @xmath190 s .
we will use @xmath75 to calculate @xmath43 .
we will again write the elements of @xmath30 as row vectors with the coordinates corresponding to the vertices with weights @xmath99 , @xmath190 , @xmath193 , @xmath234 and @xmath235 in the order given .
let @xmath195 be a generator of @xmath196 and let @xmath197 be the @xmath31 structure with first chern class equal to @xmath198 .
basic vectors are @xmath236 and @xmath237 for @xmath238 and @xmath239 .
moreover , @xmath240 lies in the @xmath31 structure @xmath197 , and @xmath241 lies in @xmath242 .
note that since everything is symmetric under @xmath197 goes to @xmath243 , we need to consider only @xmath244 .
let @xmath245 for @xmath246 .
it turns out that heights of @xmath240 and @xmath247 are both equal to @xmath248 for @xmath249 .
the minimal relations are given by @xmath250 for @xmath251 .
one calculates @xmath252 .
we conclude that @xmath253 @xmath254/u^{h_i},\mathrm{for \qua } 16\leq i \leq 21,\ ] ] @xmath254/u^{h_i } \oplus { \mathbb{z}}[u]/u^{h_i},\mathrm{for \qua } 1 \leq
i \leq 15,\ ] ] @xmath255 | in this paper we extend the results of @xcite to calculate ozsvth - szab floer homology group @xmath0 for a class of negative - semidefinite plumbings with @xmath1 . |
relativistic ejections ( jets ) are a common consequence of accretion processes onto black holes in active galactic nuclei ( agn ) as well as onto stellar mass black holes in x - ray binaries ( xrbs , see fender 2010 for a review ) . in the low / hard state ( lhs ) and probably in the quiescent state of black hole candidates ( bhcs ) a compact
, steady jet is present ( fender 2001 ; gallo et al .
the characteristic signature of compact - steady jets ( blandford & knigl 1979 ) is a flat / slightly inverted spectrum ( @xmath2 , @xmath3 ) observed in the radio band and sometimes extending to infrared ( ir ) and possibly optical frequencies ( e.g. hynes et al .
2000 ; brocksopp et al .
the jet power dominates over the accretion - powered luminosity in the lhs at @xmath4 ( @xmath5 and @xmath6 are the x - ray and the eddington luminosities , respectively ; fender , gallo & jonker 2003 , migliari & fender 2006 ) .
there is strong evidence that the jet is highly quenched in the high / soft state ( hss ) of bhcs ( tananbaum et al .
1972 ; fender et al . 1999 ) .
krding , jester & fender ( 2006 ) showed that a generalization of the accretion states used to describe bhcs could also be applied to agn .
this suggests that despite the different masses involved , systems that contain a black hole display similar accretion states and jet properties .
hannikainen et al .
( 1998 ) and corbel et al .
( 2003 ) found that the radio flux of the bhc gx 339 - 4 in the lhs correlates over several orders of magnitude with the x - ray flux .
gallo , fender & pooley ( 2003 ) included other sources in the sample and proposed that a correlation of the form @xmath7 ( where @xmath8 is the radio luminosity ) with @xmath9 ( gallo et al .
2006 ) could be universal and also valid for sources in quiescence .
this indicated that the mechanisms responsible for the ejection of the outflows are closely coupled to the properties of the accretion flow .
russell et al .
( 2006 ) verified that an empirical correlation between the x - ray luminosity and the optical / ir luminosities also holds ( @xmath10 , where @xmath11 is the optical / ir luminosity ) for bhcs in the lhs and in quiescence .
there is evidence that in most cases the near - ir emission is jet - dominated while the optical emission is not dominated by the jet but by the reprocessing of the x - rays in the outer regions of the accretion disc ( russell et al .
2006 ) .
c merloni , heinz & di matteo ( 2003 ) and falcke , krding & markoff ( 2004 ) have independently shown that the same x - ray / radio scaling found for bhcs also holds for supermassive black holes in agn , if the mass of the compact object is taken into account .
this suggests that similar mechanisms couple accretion and ejection processes to / from black holes hold over @xmath12 9 orders of magnitude in mass .
the existence of the radio / x - ray correlation has broad implications .
for example , the small scatter around it has been used as an argument by heinz & merloni ( 2004 ) to infer that jets from bhcs and agn ( once a mass - correction factor is introduced ) are characterized by similar bulk velocities , unless they are all non relativistic .
l c c c c c c c r + & & & + source & radio & near ir & @xmath13 ( hr ) & @xmath14(@xmath15 & q & @xmath16(@xmath17 ) & i ( @xmath18 ) & d ( kpc ) + h 1743 - 322 & 28.16(1,2 ) & - & - & - & - & - & - & 7.5(1 ) + gx 339 - 4 & 28.81 & 33.74 & 42.1 & @xmath195.8 & @xmath20 & 6.69 & @xmath21(3 ) & @xmath22 + xte j1118 + 480 & 28.51 & 33.20 & 4.1 & @xmath23 & 27.2 & 1.61 & @xmath24 & @xmath25 + grs 1915 + 105 & 29.26 & 33.28 & 846 & @xmath26 & 17.28 & 64.25 & @xmath27 & @xmath28(3 ) + v404 cyg & 28.80 & - & 155.3 & @xmath29 & 15.38 & 19.55 & @xmath30 & @xmath31(4 ) + a0620 - 00 & 29.00 & - & 7.75 & @xmath32 & 14.86 & 2.73 & @xmath33 & @xmath34 + gro j0422 + 32 & 27.63 & - & 5.09 & @xmath35 & 8.63 & 1.39 & @xmath36 & @xmath37 + gs 1354 - 64 & 28.85 & - & 61.1(5 ) & @xmath38(5 ) & 7.68(5 ) & 8.99 & @xmath39 & @xmath40(5 ) + 4u 1543 - 47 & 29.18 & 34.03(6 ) & 26.8 & @xmath41 & 3.84 & 5.13 & @xmath42 & @xmath43 + xte j1550 - 564 & 27.92 & 33.16(6 ) & 36.96 & @xmath44 & 7.41 & 7.16 & @xmath45 & @xmath46(7 ) + gro j1655 - 40 & 27.86(8 ) & 33.77 & 62.9 & @xmath47 & 2.99 & 7.98 & @xmath27 & @xmath48(9 ) + xte j1650 - 500 & 27.55 & - & 7.63 & @xmath49(10 ) & @xmath50(10 ) & 2.26 & @xmath51(10 ) & @xmath52(11 ) + swift j1753.5 - 0127 & 27.90(12 ) & 32.59(12 ) & 3.2(13 ) & @xmath53(13 ) & @xmath54(13 ) & 0.91 & @xmath55(14 ) & @xmath56(13,15 ) + cyg x-1 & 28.22 & - & 134.4(16 ) & @xmath57(16 ) & 0.57(16 ) & 13.26 & @xmath58(3 ) & @xmath59 + xte j1720 - 318 & 27.49 & - & - & - & - & - & - & @xmath60(17 ) + 1e1740.7 - 2942 & 27.99 & - & 305.52(18 ) & - & - & - & - & @xmath61(3 ) + grs 1758 - 258 & 27.34 & - & 442.8(18 ) & - & - & - & - & @xmath61(3 ) +
+ ( 1 ) jonker et al .
( 2010 ) ; ( 2 ) mcclintock et al .
( 2009 ) ; ( 3 ) gallo et al .
( 2003 ) ; ( 4 ) miller - jones et al . ( 2009 ) ; ( 5 ) casares et al .
( 2009 ) ; ( 6 ) russell et al .
( 2007 ) ; ( 7 ) hannikainen et al .
( 2009 ) ; ( 8) migliari et al . ( 2007a ) ; ( 9 ) foellmi et al . ( 2006 ) ; ( 10 ) orosz et al .
( 2004 ) ; ( 11 ) homan et al .
( 2006 ) ; ( 12 ) soleri et al .
( 2010 ) ; ( 13 ) zurita et al .
( 2008 ) ; ( 14 ) hiemstra et al .
( 2009 ) ; ( 15 ) cadolle bel et al . (
2007 ) ; ( 16 ) herrero et al .
( 1995 ) ; ( 17 ) chaty & bessolaz ( 2006 ) ; ( 18 ) smith et al .
( 2002 ) however , in the past few years , the supposed universality of the radio / x - ray correlation has been doubted ( xue & cui 2007 ) and several `` radio quiet '' outliers have been found ( gallo 2007 ) .
these sources seem to feature similar x - ray luminosities to other bhcs but are characterized by a radio emission that , given a certain x - ray luminosity , is fainter than expected from the radio / x - ray correlation .
it is possible that a correlation with similar slope but lower normalization than the other bhcs could describe this discrepancy , at least in a few sources ( e.g. corbel et al .
2004 ; gallo 2007 ; soleri et al . 2010 ) .
if confirmed , this would suggest that some other parameters might be tuning the accretion - ejection coupling , allowing accretion flows with similar radiative efficiency to produce a broad range of outflows .
garcia et al . ( 2003 ) investigated the dependence of the jet power on the orbital period of the binary .
they noted that , among 14 dynamically confirmed bhcs , we can spatially resolve a powerful jet in 4 systems characterized by long orbital periods .
although these radio jets are most likely blobs launched during major ejection events ( and not compact - steady jets typical of bhcs in the lhs , see fender , belloni & gallo 2004 ) , this could suggest that the orbital period might play a role in powering jets from bhcs .
c peer & casella ( 2009 ) presented a model for the emission from jets in xrbs in which the electrons are accelerated only once at the base of the jet ( at variance with other models , in which multiple accelerations occur ; see e.g. maitra et al .
2009 , jamil , fender & kaiser 2009 ) . in the model ,
the jet magnetic field is a parameter that can cause a quenching of the radio emission ( when above a critical value @xmath62 ) , without influencing the energy output in the x rays ( casella & peer 2009 ) . + the dependence of the jet power on the spin of the black hole has recently been investigated by fender , gallo & russell ( 2010 ) .
they inferred the jet power from the normalizations of the radio / x - ray and near - ir / x - ray correlations found by gallo et al .
( 2006 ) and russell et al .
they concluded that , if our measures of the spin and the estimates of the jet power are correct , the spin does not play any role in powering jets from bhcs . in agn , on the other hand , the most powerful jets have been associated with high - spin black holes ( sikora , stawarz & lasota 2007 ) . in this paper
we investigate whether there is any connection between the values of some fixed binary parameters and properties of the outburst of bhcs and the compact steady - jet power .
we discuss how doppler de - boosting effects could qualitatively explain the scatter around the radio / x - ray correlation , when a particular dependence of the bulk lorentz factor of the jet on the accretion - powered luminosity is taken into account .
we also compare the `` radio quiet '' bhcs to the neutron star ( ns ) xrbs , since these systems are usually fainter in radio ( given a certain x - ray luminosity ) than bhcs ( see e.g. migliari & fender 2006 ) .
in this paper we will follow the approach presented in fender et al .
( 2010 ) to use the normalizations of the radio / x - ray and oir / x - ray correlations as a proxy for the jet power .
we will consider the slopes of the correlations as @xmath63 . although some sources have been found to follow a correlation with a different slope ( e.g. h 1743 - 322 , @xmath64 ; jonker et al .
2010 ) and this parameter is often badly constrained by fitting the data points ( e.g. swift j1753.5 - 0127 , soleri et al .
2010 ) , we can consider it universal ( see [ par : discussion ] for a discussion on the applicability of this method ) , in order to have a rough estimate of the jet power .
for more details on this method we refer the reader to fender et al .
( 2010 ) .
table [ tab : log_sources ] lists the bhcs considered in this paper .
unless otherwise specified , the normalizations have been calculated using the data from gallo et al .
( 2003 , 2006 ) , gallo ( 2007 ) and russell et al .
( 2006 , 2007 ) .
for the radio data , we always considered an observing frequency of 8.5 ghz .
since the optical emission is not dominated by the jet , we used only the near - ir data from russell et al .
( 2006 , 2007 ; j , k and h bands ) . for each source
we fitted the x - ray and radio data using a relation of the form @xmath65 , considering the normalizations @xmath66 as free parameters . since the x - ray and radio data that we fitted do not have an error , we considered the root mean square ( rms ) of the residuals ( the dispersion of the data points around the best fit relation ) as an estimate of the uncertainty of @xmath66 .
sources that can not be well fitted using a slope @xmath67 will give high values of the rms of the residuals .
we applied the same method to obtain the normalizations @xmath68 and the rms of the residuals from the x - ray and near - ir data . since gallo et al .
( 2003 ) and russell et al .
( 2006 ) in some cases adopted different distances to the same source , in this paper we use the most recent estimates .
for the bhc cyg x-1 we do not include the data points that show evidence for suppression of the radio emission as the source enters softer x - ray states ( see figure 3 of gallo et al .
2003 for the details ) .
although we are only considering bhcs in the lhs , we include in our sample grs 1915 + 105 ( which spends all its time in the intermediate states ) , using data from the radio - bright plateau state , which is approximately an analogous to the lhs ( see fender & belloni 2004 for a review on the source ) .
the properties of the bhcs are reported in table [ tab : log_sources ] , as well as the normalizations .
( @xmath69 and @xmath68 ) used as a proxy for the jet power .
the rms of the residuals and the number of data points used for each fit are reported in table [ tab : rms_residuals ] , while the properties of the outbursts of our bhcs are listed in table [ tab : log_outbursts ] .
.rms of the residuals obtained fitting the x - ray and radio data with a relation of the form @xmath65 , leaving @xmath66 as a free parameter .
the same has been done for the near - ir and the x - ray data .
the table also shows the number of data points used for each fit .
for some bhcs we only had one data point , hence the rms of the residuals could not be calculated .
instead , for these sources , we report the average of the rms of the other systems .
[ cols="<,^,^,^,^ " , ] + ( 1 ) capitanio et al .
( 2009 ) ; ( 2 ) fender & belloni ( 2004 ) ; ( 3 ) @xcite ; ( 4 ) cadolle bel et al . ( 2007 ) ; ( 5 ) negoro et al .
( 2009 ) ; ( 6 ) gallo et al .
( 2003 ) c since the accretion disc occupies @xmath70 of the roche lobe of the black hole ( frank , king & raine 2002 ) , we calculated the size of the roche lobe of the accretor as a measure of the disc size . following frank , king & raine ( 2002 ) , the orbital separation @xmath71 is given by : @xmath72 where @xmath14 is the mass of the accretor ( in @xmath73 units )
, @xmath74 is the mass ratio ( @xmath75 , @xmath76 is the mass of the donor ) and @xmath13 the orbital period ( in hours ) .
the size of the roche lobe of the compact object @xmath77 can be calculated as follows : @xmath78 figure [ fig : powervsorbital_param ] shows the radio and near - ir normalizations as a function of the size of the roche lobe of the black hole and the orbital period of the binary .
the lower panels of figure [ fig : powervsorbital_param ] might suggest that the near - ir normalization increases with the size of the roche lobe of the accretor and with the orbital period , although the sample contains only 7 bhcs .
two of them do not follow this possible trend : xte j1550 - 564 and grs 1915 + 105 . to test whether there is any correlation between the jet power and these two orbital parameters , we calculated the spearman rank correlation coefficients for the data points in figure [ fig : powervsorbital_param ] .
the values of the correlation coefficient @xmath79 , as well as the null hypothesis probabilities ( the probability that the data are not correlated ) , are reported in table [ tab : log_spearman ] .
clearly no correlation is present .
l c c c + normalization & number of points & spearman coefficient @xmath79 & probability for the null hypothesis ( % ) + + radio & 13 & 0.5 & 11.0 + near ir & 7 & 0.4 & 37.9 + + radio & 15 & 0.2 & 54.2 + near ir & 7 & 0.5 & 25.4 + + radio & 13 & -0.2 & 44.7 + near ir & 7 & -0.9 & 3.0 + + normalization & number of points & spearman coefficient @xmath79 & probability for the null hypothesis ( % ) + + radio & 8 & 0.5 & 20.8 + near ir & 4 & @xmath80 & 99.2 + + radio & 8 & 0.4 & 28.0 + near ir & 4 & @xmath80 & 99.2 + + radio & 8 & -0.2 & 61.0 + near ir & 4 & -0.8 & 16.5 + + normalization & number of points & spearman coefficient @xmath79 & probability for the null hypothesis ( % ) + + radio & 8 & 0.3 & 44.7 + + radio & 8 & 0.2 & 52.9 + + radio & 8 & -0.4 & 31.2 + we also investigate the dependence of the radio and near - ir normalizations on the orbital inclinations of the bhcs .
although there is evidence that in some sources the binary inclination does not coincide with the inclination between the jet axis and the line of sight ( e.g. the misalignment has been estimated to be @xmath81 in gro j1655 - 40 , see maccarone 2002 and references therein ) , in this paper we will consider the orbital inclination as a proxy for the jet axis inclination to the line of sight .
we will refer to this angle @xmath82 as either inclination or viewing angle .
compact - steady jets in the lhs are thought to be mildly relativistic ( with bulk lorentz factor @xmath83 ; gallo et al .
2003 , fender et al .
2004 ) . that suggests that de - boosting effects should not be relevant and the jets , except for high inclination angles , should not be de - boosted . however , new results cast doubts on this fundamental assumption .
casella et al ( 2010 ) recently observed the bhc gx 339 - 4 in the lhs at low x - ray luminosity ( @xmath84 , considering a distance to the source and a mass @xmath14 as in table [ tab : log_sources ] ) , with coordinated x - ray and ir observations at high - time resolution .
from the analysis of the cross - correlation function , they inferred a lower limit on the bulk lorentz factor of the jet @xmath85 ( at @xmath86 confidence level ) .
this result suggests that de - boosting effects can become important , not only at high viewing angles . assuming that the x - ray emission is un - beamed , jets with @xmath87 and not pointing towards us , should result less luminous than expected from the empirical radio / x - ray and oir / x - ray correlations .
however , the possibility that the x rays are coming from the base of the jet can not be discarded ( see markoff , nowak & wilms 2005 for a theoretical model ) , since there is now evidence that , at least in the bhc xte j1550 - 564 , the synchrotron emission from the compact - steady jet dominates the x - ray emission ( in the luminosity range @xmath88 , russell et al .
2010 ) . bearing this in mind , in this paper
we will consider the x - ray emission as un - beamed .
figure [ fig : powervsinclination ] shows the radio and near - ir normalizations as a function of the inclination angles @xmath82 . from the upper panel , no obvious dependence can be found . in the lower panel
, the distribution of the data points suggests that bhcs characterized by a high inclination could have a low near - ir normalization . to test
if a correlation exists , we calculated the spearman coefficient @xmath79 for the data points .
we show them in table [ tab : log_spearman ] . in the case of the near - ir normalizations , we obtained @xmath89 , with a probability for the null hypothesis of @xmath90 .
this represents marginal evidence for an anticorrelation between the inclination angle and the near - ir normalization .
however , the lack of data points ( compared to the upper panel of figure [ fig : powervsinclination ] , in which no correlation can be found ) might have biassed this result .
in this sub - section we repeat the same analysis done in [ par : binary ] on a restricted sample of sources .
we will consider only those bhcs for which both mass and distance have been measured : xte j1118 + 480 , grs 1915 + 105 , v404 cyg , a0620 - 00 , gro j0422 + 32 , 4u 1543 - 47 , xte j1550 - 564 and cyg x-1 .
table [ tab : log_spearman ] reports the sperman correlation coefficients as well as the probabilities of the null hypothesis .
the radio normalizations are not correlated to the size of the roche lobe , the orbital period and the orbital inclination .
the near - ir jet power does not show any correlation either .
however , in the near - ir sample only 4 bhcs are present .
we also calculated new radio luminosities for the sources in the restricted sample by applying a mass correction .
considering the fundamental plane of black hole activity of merloni et al .
( 2003 ) , the mass corrected radio luminosity is @xmath91 , where @xmath92 is the mass of the compact object .
using @xmath93 we calculated the new mass - corrected radio normalizations .
the spearman correlation coefficients are reported in table [ tab : log_spearman ] : the mass corrected radio normalizations and the 3 orbital parameters examined in these paper are not correlated . during an outburst ,
bhcs usually show a transition to the hss ( see belloni 2009 and references therein ) .
the transition to the soft states is characterized by sudden changes in the jet properties ( tananbaum et al .
1972 ; fender et al . 1999 ) and it is possibly associated with the emission of highly relativistic jets ( with @xmath85 ; fender et al .
furthermore , russell et al . ( 2007 ) and corbel et al .
( in preparation ) found a dependence of the near - ir and radio normalizations on the phase of the outburst in which the bhc is observed ( e.g. xte j1550 - 564 ) .
more specifically , the normalizations measured at the outburst rise ( in the lhs ) are different from the normalizations measured at the outburst decay , after the source has transited back from the hss to the lhs . for these reasons
, we could imagine that even in the lhs , the jet properties ( e.g. its bulk velocity and power ) are influenced by the transition to the soft states .
however , some sources spend the whole outburst in the lhs , without transiting to the soft states ( brocksopp , bandyopadhyay & fender 2004 ) .
since it is not clear why some bhcs do not soften , it is important to ascertain whether lhs - only outbursts feature different jet properties . table [ tab : log_outbursts ] presents the properties of the outbursts for 10 bhcs in our sample .
sources that had only `` normal '' outbursts ( with `` normal '' we refer to bhcs that showed a transition to the soft states ) have not been listed . to see whether the type of outburst ( lhs only or with a transition to the soft states ) affects the jet power , in figure [ fig : power_outburst ] we reported the radio and near - ir normalizations of our sample of bhcs , dividing it according to the properties of the outburst .
no obvious dependence of the jet power on the type of the outburst can be found .
here we define a jet - toy model .
the aim is to test whether a dependence of the bulk lorentz factor of the jet @xmath94 on the accretion powered x - ray luminosity might qualitatively describe the scatter around the radio / x - ray correlation and the `` radio quiet '' bhcs population .
as mentioned above , we are assuming that the x - ray emission is un - beamed ( but see markoff et al .
2005 and russell et al .
2010 ) .
we will consider a @xmath94 lorentz factor that becomes larger than @xmath95 above @xmath96 of the eddington luminosity @xmath6 ( see figure [ fig : de_boosting ] , left - hand panel ) .
this assumption is based on the fact that compact steady jets are thought to be mildly relativistic ( @xmath97 ; fender et al .
2004 , but see casella et al .
2010 ) in the lhs below about @xmath98 of @xmath6 ( e.g. fender et al .
2003 , migliari & fender 2006 ) while major relativistic ejections ( @xmath99 ) are tentatively associated with the transition from the hard to the soft states .
these transitions occur at a variable @xmath5 but usually above a few per cent of @xmath6 ( fender et al .
we calculated the doppler boosting factor @xmath100 ( @xmath101 ; @xmath102 is the bulk velocity of the emitting material ; @xmath103 ) for 10 possible inclinations @xmath82 . since the orientation of the approaching jet is random on a hemisphere of @xmath104 sr , @xmath105 is uniformly distributed between 0 and 1 .
this does not imply that @xmath82 is uniformly distributed in the range @xmath106 .
we considered a uniform distribution of 10 values of @xmath105 between 0 and 1 .
figure [ fig : de_boosting ] ( right - hand panel ) shows the evolution of @xmath107 as a function of @xmath5 . at @xmath108 , only for one inclination angle ( of the ten considered ) the jet will be boosted ( @xmath109 ) and not de - boosted .
if the the jet power is well traced by the radio luminosity @xmath8 and @xmath110 , for each inclination @xmath82 we can determine how the doppler de - boosting will affect the radio - jet luminosity @xmath8 .
figure [ fig : toy_model ] illustrates the results from our toy model : it clearly results in a distribution in the ( @xmath111 ) plane which broadens at higher luminosities .
the data points represents the bhcs sample used to infer the radio normalizations reported in table [ tab : log_sources ] .
the scatter around the best - fit correlation of gallo et al .
( 2006 , dashed line ) can be partially reproduced .
c l c c + & * distance ( kpc ) * & * references * + + 4u 1728 - 34 & 4.6 & ( 1 ) + ser x-1 & 12.7 & - + 4u 1820 & 7.6 & - + mxb 1730 - 335 & 8.8 & - + 4u 0614 - 091 & 3.2 & ( 2,3 ) + 4u 1608 - 52 & 5.8 & ( 4 ) + + aql x-1 & 5.2 & ( 5 ) + igr j00291 + 5934 & 3 & ( 6,7 ) + sax j1808.4 - 3658 & 3.5 &
( 8) + + sco x-1 & 2.8 & - + gx 17 + 2 & 14 & - + gx 349 + 2 & 5 & - + cyg x-2 & 13.3 & - + gx 5 - 1 & 9.2 & - + gx 340 + 0 & 11 & - + gx 13 + 1 & 7 & - + + ( 1 ) falanga et al .
( 2006 ) ; ( 2 ) migliari et al . ( 2010 ) ; ( 3 ) kuulkers et al .
( 2009 ) ; ( 4 ) gver et al .
( 2010 ) ; ( 5 ) tudose et al .
( 2009 ) , ( 6 ) torres et al .
( 2008 ) ; ( 7 ) lewis et al .
( 2010 ) ; ( 8) galloway & cumming ( 2006 ) ;
nss are known to be fainter in radio than bhcs , given a certain x - ray luminosity , by a factor @xmath112 ( see migliari & fender 2006 and references therein ) .
this difference in radio power can be reduced to a factor @xmath113 if a mass correction from the fundamental plane of black hole activity of merloni et al .
( 2003 ) is applied ( see [ par : binary_restricted ] ) .
we will now compare the `` radio quiet '' bhcs to the population of nss that have been detected in radio .
our sample of nss is presented in table [ tab : log_ns ] : we considered the same data points as in migliari & fender ( 2006 ) with the addition of points from recent observations of aql x-1 , 4u 0614 - 091 and igr j00291 + 5934 .
the luminosities have been calculated using the latest estimates of the distances .
figure [ fig : bh_ns ] shows the mass corrected radio luminosity for our sample of bhcs and nss as a function of the x - ray luminosity .
the ns points seem to overlap to the `` radio quiet '' bhc points . to test whether these two groups of data points are statistically distinguishable
, we performed a two - dimensional kolmogorov - smirnov ( k - s ) test ( peacock 1983 ; fasano & fanceschini 1987 ) , excluding all the upper limits .
our results are reported in table [ tab : ks ] .
the k - s test shows that the probability that the `` radio quiet '' bhcs and the nss are statistically indistinguishable ( i.e. the probability of the null hypothesis ) is different from 0 , despite being small ( @xmath114 ) .
if we do not apply a mass correction , the probability of the null hypothesis is consistent with 0 : the two groups constitute two different populations .
the two - dimensional k - s test gives similar results considering only the atoll sources ( including the accreting millisecond x - ray pulsars ) , instead of all the nss , both applying or not a mass correction factor .
in this paper we tested whether there is a connection between the values of some characteristic binary parameters of bhcs , as well as the properties of their outbursts , and the energy output in the form of a jet .
our discussion is based on the assumption that the jet power can be traced by the radio and near - ir normalizations of the radio / x - ray and oir / x - ray correlations ( fender et al .
however , some sources might not follow a correlation with slope @xmath115 .
table [ tab : rms_residuals ] reports the rms of the residuals obtained from fitting the x - ray / radio and near - ir / x - ray data .
figure [ fig : normalizations_vs_rms ] shows the rms of the residuals as a function of the radio and near - ir normalizations . in the left - hand panel , the bhcs h 1743 - 322 and gro j1655 - 40 are characterized by a much higher rms than all the other sources ( a factor 5.3 and 4.3 higher than the rms of the residuals of the prototypical system gx 339 - 4 ) .
this suggests that a fit with a slope @xmath115 is probably not satisfactory and the data points have a high dispersion around the best fit relation ( see jonker et al .
2010 for h 1743 - 322 )
. however , other `` radio quiet '' bhcs for which more than 1 data point ( in x - ray / radio ) is available are characterized by a rms higher than the one of gx 339 - 4 only by a factor 1.7 ( swift j1753.5 - 0127 ) , 1.5 ( xte j1650 - 500 ) or even lower ( xte j1720 - 318 ) . in the right - hand panel ,
the `` radio quiet '' bhc xte j1550 - 564 has the highest rms , a factor 2 higher than the second highest , xte j1118 + 480 .
the only other `` radio quiet '' bhc in the plot , swift j1753.5 - 0127 , has actually the lowest rms ( but there are only 7 sources in the sample ) .
these considerations suggest that fitting the x - ray / radio and x - ray / near - ir data fixing the slope to the value @xmath115 constitutes a good method or at least does not particularly affect the result for the `` radio quiet '' sources .
figure [ fig : powervsorbital_param ] and the spearman correlation coefficients in table [ tab : log_spearman ] show that there is no connection between two ( non independent ) parameters of the binary system ( the size of the roche lobe of the black hole and the orbital period ) and the radio and near - ir normalizations .
this result is not unexpected : theoretical works predict that a powerful jet should be formed when a thick accretion flow is present ( as it is thought the case in the lhs ; livio et al .
1999 ; meier 2001 ) .
considering that the thick accretion flow does not extend to the outer regions of the accretion disc ( although the details of its geometry are unknown , see gilfanov 2009 and references therein ) , it seems unlikely that the jet power is affected by the size of the roche lobe and the orbital period .
we repeated the same analysis using a restricted sample of bhcs for which both the mass and the distance have been measured ( [ par : binary_restricted ] ) .
our conclusions do not change , suggesting that our results are probably not strongly biassed by the large uncertainties in the measures of the mass and the distance .
our conclusions are unvaried even if we apply the mass correction from the fundamental plane of black hole activity . in [ par : outburst ]
we examined the spectral properties of the outbursts occurred by the bhcs in our sample .
our results suggest that the energy output in the jet in the lhs is not regulated by the fact that the bhc transits or not to the soft states .
it is worth to note that the three of the four known bhcs with the shortest orbital periods ( swift j1753.5 - 0127 , xte j1118 + 480 and gro j0422 + 32 ) only had hard outbursts ( but see negoro et al .
2009 for a possible softening of swift j1753.5 - 0127 and brocksopp et al .
2010 for the last outburst of xte j1118 + 480 ) .
recently , negoro et al . (
2010 ) discovered a new bhc , maxi j1659 - 152 .
this source might be the bhc with the shortest orbital period ( 2 hours and 25 minutes , see belloni et al .
2010a and kuulkers et al . 2010 ) . however , this system had a normal outburst , with a transition to the soft states ( see belloni et al .
2010b and shaposhnikov & yamaoka 2010 ) .
this suggests that a bigger sample of bhcs characterized by a short orbital period will certainly help to clarify this issue .
c from binary evolution calculations , bhcs can in principle have orbital periods as short as @xmath192 hours and evolutionary models actually predict that short - period systems might form the majority of them , similarly to what is observed in cataclysmic variables ( yungelson et al .
the fact that 16 bhcs in table [ tab : log_sources ] have an orbital period @xmath116 hr is quite puzzling .
a possible explanation is that in short - period systems , the mass transfer from the companion star might be interrupted by resonances within the primary s roche lobe , if the mass ratio is @xmath117 ( yungelson et al .
zurita et al ( 2008 ) suggested that the mass ratio in swift j1753.5 - 0127 is @xmath118 and could be as high as approximately @xmath119 , thus making swift j1753.5 - 0.127 the first bhc detected nearly in this regime .
however , even if the mass transfer in swift j1753.5 - 0127 is partially interrupted ( because of the high mass ratio ) , this does not explain why the source is less luminous than expected in the radio band ( in other words why its radio and near - ir normalizations are low compared to the majority of the bhcs in our sample ) , although its x - ray luminosity is comparable to other bhcs ( soleri et al .
the behaviour of xte j1118 + 480 is different : although it is the known bhc with the second shortest orbital period ( @xmath120 hr ) and its mass ratio is rather high compared the other sources in the sample ( @xmath121 ) , its radio normalization is higher than the one estimated for 7 other bhcs with longer orbital periods .
this suggests that it features jet properties that are not different from the majority of the bhcs .
c in [ par : binary ] we investigated the connection between the inclination angles of the bhcs and the radio and near - ir normalizations .
the upper panel of figure [ fig : power_outburst ] does not show any correlation between the radio normalization and the inclination angles ( see also table [ tab : log_spearman ] ) .
the near - ir normalization instead has a fairly high probability to be anticorrelated to the inclination angle .
this would be consistent with a scenario in which the inner jet ( responsible for the ir emission , blandford & knigl 1979 ) has a higher lorentz factor @xmath94 ( and therefore it is more de - boosted ) than the the outer jet ( where the radio emission originates ) , i.e. a decelerating jet .
if we restrict our analysis on a sample of sources for which both the mass and the distance have been measured ( [ par : binary_restricted ] ) neither the radio nor the near - ir normalizations are correlated to the inclination angle . however
, this result might have been strongly biassed by the limited number of data points .
the radio jet power does not appear to be correlated to the viewing angle even if the radio normalizations are calculated using the mass - corrected radio luminosities .
to further investigate the decelerating jet scenario , in figure [ fig : radio_vs_jhk_normalization ] we reported the near - ir normalizations versus the radio normalizations . for a decelerating jet , we do not expect the data to be linearly correlated , especially for high - inclination sources .
considering all the 7 points , we calculates a spearman rank coefficient @xmath122 , with a probability for the null hypothesis of @xmath123 .
if we exclude the data point for the bhc gro j1655 - 40 and xte j1550 - 564 ( which have very high rms of the residuals in x - ray / radio and x - ray / near - ir fits , respectively ; see figure [ fig : normalizations_vs_rms ] left - hand panel ) the spearman rank correlation coefficient is @xmath124 , with a probability for the null hypothesis of @xmath125 . if we consider only high - inclination systems ( above @xmath126
; in this case we exclude the bhcs 4u 1543 - 47 and gx 339 - 4 ) the spearman rank correlation coefficient is @xmath127 , with a probability for the null hypothesis of @xmath128 . despite the number of data points
is extremely limited ( 7 or 5 ) , our results could support the decelerating jet scenario .
more coordinated x - ray , near - ir and radio observations are needed to enlarge the sample . to further test if doppler de - boosting effects play a role in populating the sample of bhcs characterized by a faint jet , in [ par : boosting ] we defined a jet - toy model .
this model results in a distribution in the ( @xmath111 ) plane which broadens at higher luminosities . considering a range of viewing angles ( figure [ fig : toy_model ] ) , the model can qualitatively describe the scatter around the radio / x - ray correlation .
the line for @xmath129 ( @xmath130 ) increases its slope at @xmath131 ( figure [ fig : toy_model ] ) , since the jet is boosted . for higher inclination angles
, the model predicts that the radio / x - ray correlation , because of de - boosting effect , should become flatter at high @xmath5
. this might be the case for the bhc h 1743 - 322 ( for which @xmath132 ; jonker et al .
2010 ) , in agreement with the high rms of the residuals that we obtained for this source .
unfortunately , the viewing angle of h 1743 - 322 system is unknown .
however , the jet - toy model has strong limitations in explaining the behaviour of other systems ( e.g. xte j1650 - 500 and swift j1753.5 - 0127 ) .
these `` radio quiet '' bhcs feature a slope of the x - ray / radio correlation steeper than @xmath67 ( even if figure [ fig : normalizations_vs_rms ] seem to suggest that fitting their data point with a slope fixed to @xmath67 is a reasonable approximation ) , and they are probably characterized by high inclination angles ( @xmath133 and @xmath134 , respectively ) .
we also note that the bhc v404 cyg lies approximately on the line for @xmath129 .
this does not match the measured value of @xmath135 ( table [ tab : log_sources ] ) .
the same happens for other bhcs in our sample , which are actually scattered over several lines .
our considerations suggest that , although the model can qualitatively describe the scatter in the ( @xmath136 ) plane , it should be only seen as a viable possibility to describe the `` radio quiet '' population . in fact , at the moment we do not know whether the `` radio quietness '' is an intrinsic property or it might change with time .
in [ par : comparison_ns ] we compared the `` radio quiet '' bhcs to the nss .
a two - dimensional k - s test can not completely rule out the possibility that the two families are statistically indistinguishable in the x - ray : radio plane , if a mass correction is applied .
interestingly , including or not the ns z sources ( that usually feature a disc - jet coupling more similar to bhcs than to ns atoll sources , migliari & fender 2006 ) does not substantially change the results of the k - s test . since for several outliers the mass of the compact object has not been dynamically measured ( e.g. grs 1758 - 258 and xte j1720 - 318 ) , our results seem to suggest that some of the outliers could actually be neutron stars .
another possibility is that the disc - jet coupling in the `` radio quiet '' bhcs is more similar to the one in nss .
c c c c c + first sample & second sample & mass correction & k - s statistic ( d ) & prob ( d @xmath137 observed ) + `` radio quiet '' bhcs & nss & yes & 0.46 & @xmath138 + `` radio quiet '' bhcs & nss & no & 0.63 & @xmath139 + `` radio quiet '' bhcs & atoll sources & yes & 0.55 & @xmath140 + `` radio quiet '' bhcs & atoll sources & no & 0.70 & @xmath141 + table [ tab : ks ] also shows that if a mass correction from the fundamental plane of merloni et al .
( 2003 ) is not applied , `` radio quiet '' bhcs and nss clearly do not belong to the same population .
it is interesting to note that the mass correction works for black holes of different masses as well as nss : this implies that the correction may not be dependent on black - hole features like for example the presence of an event horizon .
we examined three characteristic parameters of bhcs and the properties of their outbursts to test whether they regulate the energy output in the form of a jet .
this has been motivated by the fact that a growing population of sources seems to feature similar accretion flows to the majority of the bhcs ( e.g. similar radiative efficiency ) but fainter jets than expected .
garcia et al . ( 2003 ) suggested that spatially - resolved powerful jets ( so discrete ejections and not compact - steady jets in the lhs ) might be associated with long period systems .
if our estimates of the jet power are correct , both the orbital period and the size of the accretion disc are not related to the radio and near - ir jet power .
we retrieved the properties of the outbursts occurred by the bhcs in our sample to see if lhs - only outbursts feature different jet properties .
we did not find any association between the jet power and the fact that a bhc transits to the soft states during an outburst .
we also considered the inclination angles for our sample of bhcs . a recent result shows that compact - steady jets in the lhs might have bulk lorentz factor @xmath85 .
this suggests that not only jets with a high inclination can suffer de - boosting effects .
however , we did not find any association between the viewing angles and the jet power inferred from radio observations . the jet power obtained from near - ir measurements
decreases when the inclination angle increases .
although this result might be biassed by the small number of bhcs for which we have ir data , it could favour a scenario in which the jet decelerates moving from the ir - emitting to the radio - emitting part .
we defined a jet - toy model in which the bulk lorentz factor @xmath94 becomes larger than @xmath0 above @xmath142 .
considering an uniform distribution of viewing angles in the @xmath143 space , the model results in a distribution in the ( @xmath111 ) plane which broadens at higher luminosities .
the model can qualitatively reproduce the scatter around the radio / x - ray correlation .
although this result is quite promising , we stress that the toy - model has several limitations , for instance it can not reproduce the slope of the x - ray / radio correlation for some `` radio quiet '' bhcs characterized by high inclination angles .
nevertheless we think that it suggests a valid possibility that theoretical models should explore in more detail .
we finally compared the `` radio quiet '' bhcs to the nss .
a two - dimensional k - s test can not completely rule out the possibility that the two families are statistically indistinguishable in the x - ray : radio plane , if a mass correction from the fundamental plane of black hole activity is applied .
this result suggests that some `` radio quiet '' bhcs could actually be nss ; alternatively it suggests that some bhcs feature a disc - jet coupling more similar to nss rather than to the majority of the bhcs .
ps acknowledges support from nwo ( netherlands foundation for scientific research ) .
we thank elena gallo and david russell for providing the original data used to calculate the radio and near - ir normalizations and michiel van der klis for very useful comments on the early drafts of the manuscript .
the authors also thank simone migliari for providing the neutron star data and the referee for giving very useful comments that certainly improved the quality of this manuscript .
ps would also like to thank piergiorgio casella , alessandro patruno , tomaso belloni , andrea sanna , mariano mndez and monica colpi for useful discussion .
cadolle bel , m. , rib , m. , rodriguez , j. , chaty , s. , corbel , s. , goldwurm , a. , frontera , f. , farinelli , r. , davanzo , p. , tarana , a.,ubertini , p. , laurent , p. , goldoni , p. , mirabel , i. f. , 2007 , apj , 659 , 549 casella , p. , maccarone , t. j. , obrien , k. , fender , r. p. , russell , d. m. , van der klis , m. , peer , a. , maitra , d. , altamirano , d. , belloni , t. , kanbach , g. , klein - wolt , m. , mason , e. , soleri , p. , stefanescu , a. , wiersema , k. , wijnands , r. , 2010 , mnras , 404 , l21 charles , p. a. & coe , m. j. , 2006 , in `` compact stellar x - ray sources '' , eds .
w.h.g . lewin and m. van der klis , cambridge university press , cambridge corbel , s. , nowak , m. a. , fender , r. p. , tzioumis , a. k. & markoff , s. , 2003 , a&a , 400 , 1007 jonker , p. g. , miller - jones , j. , homan , j. , gallo , e. , rupen , m. , tomsick , j. , fender , r. p. , kaaret , p. , steeghs , d. t. h. , torres , m. a. p. , wijnands , r. , markoff , s. , lewin , w. h. g. , 2010 , mnras , 401 , 1255 kuulkers , e. , in t zand , j. j. m. , atteia , j. , levine , a. m. , brandt , s. , smith , d. a. , linares , m. , falanga , m. , sanchez - fernandez , c. , markwardt , c. b. , strohmayer , t. e. , cumming , a.,suzuki , m. , 2009 , a&a , 514 , a65 lewis , f. , russell , d. m. , jonker , p. g. , linares , m. , tudose , v. , roche , p. , clark , j. s. , torres , m. a. p. , maitra , d. , bassa , c. g. , steeghs , d. , patruno , a. , migliari , s. , wijnands , r. , nelemans , g. , kewley , l. j. , stroud , v. e. , modjaz , m. , bloom , j. s. , blake , c. h. , starr , d. , 2010 , a&a in press soleri , p. , fender , r. p. , tudose , v. , maitra , d. , bell , m. , linares , m. , altamirano , d. , wijnands , r. , belloni , t. , casella , p. , miller - jones , j.c.a . , muxlow , t. , klein - wolt , m. , garrett , m. , van der klis , m. , 2010 , mnras in press torres , m. a. p. , jonker , p. g. , steeghs , d. , roelofs , g. h. a. , bloom , j. s. , casares , j. , falco , e. e. , garcia , m. r. , marsh , t. r. , mndez , m. , miller , j. m. , nelemans , g. , rodrguez - gil , p. , 2008
, apj , 672 , 107 | the coupling between accretion processes and ejection mechanisms in accreting black holes in binary systems can be investigated by empirical relations between the x - ray / radio and x - ray / optical - infrared luminosities .
these correlations are valid over several orders of magnitude and were initially thought to be universal .
however , recently many black hole binaries have been found to produce jets that , given certain accretion - powered luminosities , are fainter than expected from the earlier correlations .
this shows that black holes with similar accretion flows can produce a broad range of outflows in power , suggesting that some other parameters or factors might be tuning the accretion / ejection coupling .
recent work has already shown that this jet power does not correlate with the reported black hole spin measurements . here
we discuss whether fixed parameters of the binary system ( orbital period , disc size , inclination ) , as well as the properties of the outburst , produce any effect on the energy output in the jet . no obvious dependence is found .
we also show that there is no systematic variation of the slope of the radio : x - ray correlation with normalization .
we define a jet - toy model in which the bulk lorentz factor becomes larger than @xmath0 above @xmath1 of the eddington luminosity . with this model , if we assume random inclination angles which result in highly variable boosting at large eddington ratios , we are able to reproduce qualitatively the scatter of the x - ray / radio correlation and the `` radio quiet '' population
. however the model seems to be at odds with some other observed properties of the systems .
we also compare the `` radio quiet '' black holes with the neutron stars .
we show that if a mass correction from the fundamental plane is applied , the possibility that they are statistically indistinguishable in the x - ray : radio plane can not be completely ruled out .
this result suggests that some of the outliers could actually be neutron stars , or that the disc - jet coupling in the `` radio quiet '' black holes is more similar to the one in neutron stars .
[ firstpage ] x - rays : binaries ism : jets and outflows accretion , accretion discs . |
isostructural and isovalent laofep and laofeas are layered conductors , the first being superconducting at t@xmath3=2.5 k@xcite while the second becomes antiferromagnetically ordered at t@xmath4 @xmath5 140 k@xcite and is not superconducting .
the discovery of superconductivity at 26 k in carrier - doped laofeas@xcite , followed by rapid improvement now up to t@xmath3=55 k@xcite in this class , makes these superconductors second only to the cuprates in critical temperature .
several dozen preprints appeared within the two months after the original publication , and many hundred since , making this the most active field of new materials study in recent years ( since the discovery in mgb@xmath6 , at least ) . a host of models and ideas about the `` new physics '' that must be operating in this class of compounds is appearing , pointing out the need to establish a clear underpinning of the basic electronic ( and magnetic ) structure of the system .
the materials are strongly layered , quasi - two - dimensional in their electronic structure , by consensus .
the electronic structure of laofep was described by lebgue,@xcite with the electronic structure and its neighboring magnetic instabilities of laofeas being provided by singh and du@xcite .
several illuminating papers have appeared since , outlining various aspects of the electronic and magnetic structure of laofeas .
the extant electronic structure work has provided a great deal of necessary information , but still leaves many questions unanswered , and indeed some important questions are unaddressed so far . in this paper
we address some of these questions more specifically .
stoichiometric laofeas is afm ; then @xmath70.05 carriers / fe doping of either sign destroys magnetic order and impressive superconductivity arises , with @xmath8 seemingly depending little on the carrier concentration . another question is : with the nonmagnetic electronic structure of laofep and laofeas being so similar , why is the former superconducting while the latter is ( antiferro)magnetic ? surely this difference must be understood and built into bare - bones models , or else such models risk explaining nothing , or explaining anything .
another question is the effect of the structure .
unusual sensitivity to the as height @xmath0(as ) has been noted@xcite ; t@xmath3 is reported to increase with applied pressure@xcite ( reduction in volume ) for low values of doping ( up to @xmath9 in lao@xmath10f@xmath11feas , which is reported as the amount of f for optimal doping ) ; there are increases in t@xmath3 due to replacement of la with other rare earth ions , and the variation in size of the rare earth is often a dominant factor in the observed trends in their compounds . very important also is the magnetism in these materials , as magnetism is a central feature in the cuprate superconductors and in correlated electron superconductors .
another important question is : what can be expected if other pnictide atoms can be incorporated into this system : sb ( or even bi ) on the large atom side , or n on the small atom end . in this paper
we address these questions .
the members of the family of the new fe - based superconductors crystallize in the zrcusias type structure@xcite ( space group p4/nmm , z = 2 ) .
for instance , laofeas is made of alternating lao and feas layers , as presented in fig .
[ fig : structure ] . ,
scaledwidth=48.0% ] the fe and o atoms lie in planes , while the as and la atoms are distributed on each side of these planes following a chessboard pattern .
the crystal structure is fully described by the a and c lattice parameters , together with the internal coordinates of la and as .
experimentally , @xmath12 and @xmath13 , while z(la ) = 0.14154 and z(as ) = 0.6512 . magnetic structure of the fe - as substructure of laofeas , showing alternating chains of fe spin up ( red circles with black dots ) and fe spin down ( blue circles with black crosses ) .
the as atoms above ( below ) the fe plane are represented as large ( small ) squares .
the @xmath14 cell is represented in full lines , while the @xmath15 cell is in dashed lines .
[ fig : afm ] , scaledwidth=28.0% ] however to describe correctly the antiferromagnetic structure , a @xmath14 cell must be used , with four fe atoms per cell , as shown in full lines in fig .
[ fig : afm ]
. we will refer to this antiferromagnetic order as the q@xmath16 afm order , or equivalently as ( @xmath17 ) , while the q@xmath18 afm order corresponds to an antiferromagnetic order of the original cell ( dashed lines in fig .
[ fig : afm ] ) with two fe atoms .
also , fm will refer to a ferromagnetic arrangement of the spins , while nm means non - magnetic .
to calculate the relevant quantities , we have used density functional theory ( dft ) @xcite , as implemented in three different electronic structure codes .
the full potential local orbital ( fplo ) code@xcite was mainly used , while we double checked some of the calculations with wien2k code @xcite . for most of the fplo and lapw calculations ,
the perdew and wang 1992 ( pw92)@xcite exchange - correlation ( xc ) functional was used , but the effect of xc functional was checked using also lsda(pz)@xcite , the pbe ( perdew _ et al . _ 1996)@xcite , and another gga ( perdew _ et al . _
1992)@xcite xc functionals . at each constant volume ,
the crystal structure was fully relaxed , i.e. , @xmath19 , @xmath0(la ) and @xmath0(pn ) were relaxed , where pn is the pnictogen atom .
the errors were estimated to be within 0.5@xmath20 for c / a , and 1.0@xmath20 for z(la ) and z(pn ) .
the relaxation was performed in the q@xmath16 afm structure , with 132 irreducible k points in the bz .
we double checked the total energy with a finer mesh with 320 irreducible k points in the bz , and the difference is very small .
after relaxation , all calculations were performed using dense meshes , with 320 , 1027 , and 637 irreducible k points in the bz of the q@xmath16 afm , q@xmath18 afm and nm structure , respectively . in the q@xmath16 afm structure , we used 464 irreducible k points in the bz to double check the result , without any noticeable difference in the dos nor band structure . as for the results presented in sect .
[ sect - pipipi ] , we used the paw ( projector augmented waves ) method @xcite as implemented in the code vasp ( vienna ab - initio simulation package)@xcite .
the perdew burke ernzerhof@xcite variant of the generalized gradient approximation ( gga ) was used for the exchange - correlation potential . a cut - off of @xmath21 ev
was used for the plane - wave expansion of the wave function to converge the relevant quantities .
for brillouin zone integrations , a mesh of @xmath22 @xmath23-points@xcite was used within the modified tetrahedron method@xcite .
this mesh was decreased to @xmath24 for the cell doubled along the @xmath25 axis .
laofeas has a tetragonal structure ( as described in sect . [
sec : two ] ) at room temperature@xcite .
although it undergoes a structural phase transition at lower temperature@xcite ( see section [ sect - pipipi ] ) , the doped ( and superconducting ) material lao@xmath10f@xmath11feas remains in this structure down to low temperature , so the study of laofeas in the high symmetry structure is a necessary step towards the understanding of the electronic structure of the whole family of compounds .
first , we studied the electronic structure of laofeas in the experimental ( tetragonal ) crystal structure for different magnetic states ( q@xmath16 afm , q@xmath18 afm , fm and nm ) using two different codes ( fplo@xmath26 and wien@xmath27k ) and different exchange - correlation functionals .
this is necessary in view of the large number of theoretical papers@xcite which appeared recently and often contain strong disagreements .
this was partly studied by mazin
_ et al._.@xcite table [ as - mag - xc ] summarizes the results : the magnetic moment on the fe atom together with the total energy differences for each magnetic state studied here .
independent of the code or the xc functional used , the q@xmath16 afm state is always found to be the ground state , which confirms our earlier report@xcite .
the magnetic moment for both afm orders are considerably larger than the ordered moment reported from neutron diffraction and muon spin relaxation experiments , while the one for the fm order is much smaller .
for this last case , fplo7 gives zero which indicates no magnetism with both pz and pw92 xc functional ; wien2k gives about 0.36 @xmath28@xmath29 with gga and pbe and 0.13 @xmath28@xmath29 with pw92 .
it appears therefore that the magnetic moment of fe for the same state with different xc functionals varies by up to 0.5 @xmath28@xmath29 , which is unexpectedly large , although gga is known to enhance magnetism.@xcite the difference between fplo7 and wien2k in predicting the fe magnetic moment for each state may explain the total energy differences among them .
virtual doping ( see subsection b ) by 0.1 e@xmath30/fe enhances the fe magnetic moment in the q@xmath16 afm state but reduces it in the fm state for all the xc functionals used . in the structural optimization ( performed in the q@xmath16 state ) , fplo7 with pw92 ( lda ) functional gives reasonable c / a and z(la ) in good agreement with experiment , but it predicted z(as ) @xmath7 0.139 , which is 0.011 off the experimental value , about 0.1 in length .
however , wien2k with pbe(gga ) xc functional gives an optimized z(as ) @xmath7 0.149 , which agrees well with experimental z(as ) .
similar results are found in the xfe@xmath6as@xmath6 family ( x = ba , sr , ca ) too .
it suggests that , gga ( pbe ) xc functional optimizes the feas - based system much better than lda ( pw92 ) xc functional . and gga should have better performance in dealing with the structure ( including c / a , equilibrium volume and z(as ) ) under pressure of this feas family
this is probably due to the layered structure of the feas family which results in large density gradient between layers , thus gga has better description of the potential .
but in the meantime , gga ( pbe ) further overestimates the magnetic moment of fe , which is already overestimated by lda ( pw92 ) .
.calculated magnetic moment of fe , the amounts of total energy per fe lie below nonmagnetic state of fm , q@xmath18 afm and q@xmath16 afm states from fplo7 and wien2k with different xc functionals of laofeas with experimental structure .
positive @xmath31 ee means lower total energy than nm state . [ cols="^,^,^,^,^,^,^,^",options="header " , ]
we have investigated in some detail the electronic structure and magnetic properties of the laofeas class of novel superconductors using ab - initio methods .
the effects of the fe - as distance , of doping , and of pressure , as well as calculations of the efgs have been reported .
it was found that ( approximate ) electron - hole symmetry versus doping , and strong magnetophonon coupling are primary characteristics of the laofeas system , and are two of the ingredients that need to be understood to proceed toward the discovery the mechanism of superconducting pairing .
we studied effects of the structural distortion and of the ( @xmath32 ) magnetic order , finding that experiments can be reproduced fairly well by our calculations .
finally , the related materials laofep , laofesb , and laofen were investigated and their properties compared to those of laofeas . from these comparisons , it appears that laofep is significantly different from the other materials studied here ; this difference might explain why , at stoichiometry , laofep is superconducting while laofeas is antiferromagnetic .
also , in view of their similarities with laofeas , either pure or doped laofesb and laofen are potential candidates as superconductors .
we acknowledge financial support from anr pnano grant n@xmath33 anr-06-nano-053 - 02 and n@xmath33 anr - blan07 - 1 - 186138 ( s.l . ) and from doe grant de - fg03 - 01er45876 ( w.e.p . ) .
w.e.p . is happy to acknowledge a grant from the france berkeley fund that enabled the initiation of this project . | the occurrence of high temperature superconductivity , and the competition with magnetism , in stoichiometric and doped laofeas and isostructural iron - oxypnictides is raising many fundamental questions about the electronic structure and magnetic interactions in this class of materials .
there are now sufficient experimental data that it may be possible to identify the important issues whose resolution will lead to the understanding of this system . in this paper
we address a number of the important issues .
one important characteristic is the fe - as distance ( or more abstractly the pnictogen ( pn ) height @xmath0(pn ) ) ; we present results for the effect of @xmath0(pn ) on the electronic structure , energetics , and fe magnetic moment
. we also study laofeas under pressure , and investigate the effects of both electron and hole doping within the virtual crystal approximation .
the electric field gradients for all atoms in the laofeas compound are presented ( undoped and doped ) and compared with available data .
the observed @xmath1 magnetic order is studied and compared with the computationally simpler @xmath2 order which is probably a very good model in most respects .
we investigate the crucial role of the pnictogen atom in this class , and predict the structures and properties of the n and sb counterparts that have not yet been reported experimentally . at a certain volume
a gap opens at the fermi level in laofen , separating bonding from antibonding bands and suggesting directions for a better simple understanding of the seemingly intricate electronic structure of this system .
finally , we address briefly on the possible effects of post - lanthanum rare earths , which have been observed to enhance the superconducting critical temperature substantially . |
we present an analytical and phenomenological model for chemical evolution of galaxies based on hierarchical structure formation .
this model provides a possible explanation for the metal abundances observed in damped ly@xmath2 ( dla ) systems and the metallicity distribution of stars in galaxies .
it also suggests a possible mechanism for the formation of globular clusters .
measurement of `` metallicities '' of dla systems by lu et al .
( 1996 ) and further studies by prochaska & wolfe ( 2000 , 2002 ) showed that these systems have a large dispersion in [ fe / h ] @xmath5 for a given redshift @xmath6 and that there appears to be a baseline enrichment at the level of [ fe / h ] @xmath7 over a wide range of redshift @xmath8 .
the work by prochaska et al .
( 2003 ) provided a much more extensive data base over a wider range of redshift @xmath3 .
all these data exhibited the same behavior .
we earlier found that the data on dla systems could be explained by a simple closed - system model ( wasserburg & qian 2000b and fig . 1 here ) , where the number abundance ( fe / h ) of fe relative to h atoms in a dla system is determined by @xmath9 .
\label{cdla}\ ] ] in the above equation , ( fe / h)@xmath10 ( [ fe / h]@xmath11 ) is the initial fe abundance corresponding to a postulated prompt inventory of metals in the intergalactic medium ( igm ) , @xmath12 is the rate of fe production by type ii supernovae ( sne ii ) per h atom in the gas of the dla system and is taken to be @xmath13 based on estimates for the galaxy ( see 2 ) , @xmath14 is the age of the universe at the redshift @xmath6 of the system , and @xmath15 is the `` turn - on '' time corresponding to the onset of astration and fe production in the system .
it was found that the available data on [ fe / h ] for essentially all the dla systems lie between the prompt inventory value and the upper bound obtained for @xmath16 .
the large dispersion in [ fe / h ] at a given @xmath6 could be explained by different @xmath15 for different dla systems .
it was found that for the above closed - system model , most systems at a given @xmath6 have @xmath15 in a range close to @xmath14 ( i.e. , they `` turned on '' long after the big bang so that their baryonic matter had to be stored in the igm for this long period ) .
the astrophysical implications of the late turn - on of most dla systems in the above closed - system model were explored by us in a recent study based on the cold dark matter ( cdm ) model of hierarchical structure formation ( qian & wasserburg 2003b ) .
we there associated the evolution of a dla system with the growth of a halo through infall of cdm and gas .
the late turn - on was reinterpreted in terms of two factors .
first , a halo must achieve some threshold mass before star formation could occur .
as a halo associated with a low-@xmath17 density fluctuation reaches a given mass at a later time , fe production in such a halo would be delayed simply due to the later onset of star formation .
second , subsequent to the onset of star formation , the growth of ( fe / h ) in the gas of a halo is determined by the competition between production by sne ii and dilution by gas infall . the uniform growth with time
does not set in until infall ceases .
thus , the late `` turn - on '' does not require the introduction of an arbitrary delay time @xmath15 but is a direct result of the infall rates in the cdm model of hierarchical structure formation .
we here extend our model for chemical evolution of dla systems to a more general discussion of galactic chemical evolution . in
2 we review the consequences of astration concurrent with infall for the chemical evolution of gas in a halo that is not disrupted by sne ii .
we discuss the effects of infall cessation , particularly the resultant metallicity distribution of stars in 3 .
the connection with the cdm model of hierarchical structure formation is described in 4 .
a comparison of our chemical evolution model with the data on dla systems is presented in 5 .
possible scenarios of globular cluster formation in connection with this model are discussed and conclusions given in 6 .
we first consider the chemical evolution of gas in a halo that grows according to hierarchical structure formation .
we assume that cdm and gas are fed into the halo at a fixed mass ratio through infall . for simplicity
, the infalling gas from the igm is assumed to be pure big bang debris ( i.e. , metal - free ) .
observations show that the igm is significantly enriched with `` metallicities '' corresponding to [ fe / h ] @xmath7 over a wide range of redshift @xmath18 ( e.g. , songaila 2001 ; pettini et al .
2003 ; see also simcoe , sargent , & rauch 2002 , 2004 ; schaye et al .
2003 ; aguirre et al .
thus , the assumption of metal - free infalling gas is not necessarily valid .
however , possible metallicities at the level of [ fe / h ] @xmath7 for the infalling gas are soon overwhelmed by enrichments associated with astration in the halo .
insofar as the evolution of the halo substantially enhances the metal content of its gas to well above [ fe / h ] @xmath7 , neglecting the metallicity of the infalling gas is justified .
the history of star formation in the halo is crucial to the chemical evolution of the gas .
the issue of what the first generation of stars were like has long been a matter of intense interest .
numerical simulations of collapse and fragmentation of metal - free gas clouds indicate that the first stars ( or baryonic aggregates ) were very massive .
formation of normal stars may not have occurred until a `` metallicity '' corresponding to [ fe / h ] @xmath19 was reached in the igm ( see bromm & larson 2004 , and references therein ) .
we do not consider the effect of this threshold metallicity that may possibly be required for normal astration .
we assume that normal astration proceeds at a constant rate per h atom in the gas once the halo reaches some critical mass ( see 4 ) . from the discovery of a low - mass galactic halo star with [ fe / h ] @xmath20 ( christlieb et al .
2002 ) , it is evident that some low - mass stars were formed when the level of heavy metals in the igm was extremely low , far below the threshold metallicity for normal astration suggested by theoretical models ( bromm et al .
2001 ) .
it has long been recognized that stars in the galactic halo form a distinct population .
extensive observational work has established that these stars have distinctively low [ fe / h ] ranging from typically @xmath21 ( e.g , carney et al .
1996 ) down to @xmath22 ( see fig . 5 in christlieb 2003 ) .
there is a clear break at [ fe / h ] @xmath23 observed in the galactic evolution of [ ba / fe ] and [ o / fe ] with [ fe / h ] . for [ ba / fe ] , this break is due to the lack of major @xmath24-process contributions from asymptotic giant branch stars to the ba in the interstellar medium ( ism ) for [ fe / h ] @xmath0 ( cf . ,
busso , gallino , & wasserburg 1999 ) , while for [ o / fe ] , it is due to the lack of major fe contributions from sne ia for [ fe / h ] @xmath0 ( tinsley 1979 ; see also kobayashi et al .
in addition , the abundance patterns of the elements from the fe group down to o observed in galactic halo stars are characteristic of the nucleosynthetic yields of sne ii .
thus , the overall abundances of `` metals '' in the early galaxy corresponding to [ fe / h ] @xmath0 in the ism were clearly dominated by contributions from sne ii with massive ( @xmath25@xmath26 ) progenitors .
we note that the marked change at [ fe / h ] @xmath7 observed for the abundances of heavy @xmath27-process nuclei in galactic halo stars ( e.g. , mcwilliam et al .
1995 ; sneden et al . 1996
; westin et al . 2000 ; burris et al .
2000 ) appears to require a source other than sne ii for these nuclei ( qian & wasserburg 2003a ) .
however , this @xmath27-process source can not produce any significant amount of the elements from the fe group down to o ( e.g. , qian & wasserburg 2002 ) , and therefore , does not affect our discussion of the chemical evolution of these elements here .
based on the above discussion , we only consider metal production by sne ii subsequent to the onset of astration during the early evolution of a halo .
the general effects of sne ia that contribute mainly the fe group elements at [ fe / h ] @xmath28 will be discussed in 3 . for convenience , we use fe as a representative metal element . at [ fe / h ]
@xmath0 , the other elements ( from the fe group down to o ) essentially evolve in proportion to fe as specified by the effective yield pattern of sne ii ( see table 1 in qian & wasserburg 2002 ) .
if gas in the halo is homogeneous and is not expelled by explosions of sne ii , the basic equation governing its chemical evolution ( qian & wasserburg 2003b ) may be written as : @xmath29 here ( fe / h ) is the number ratio of fe to h atoms in the gas , @xmath30 is the net rate of fe production by sne ii , ( h ) is the number of h atoms in the gas , and @xmath31 is the rate of infall of h atoms into the halo .
the time @xmath32 corresponds to the big bang .
the equation governing ( h ) is : @xmath33 where @xmath34 is the astration rate . during
the infall dominated regime @xmath35 so that @xmath36 . then for this regime equation ( [ feh ] ) can be rewritten as : @xmath37 where @xmath38 , @xmath39 $ ] , and @xmath40 .
note that for @xmath41 to increase rapidly , @xmath42 must be significantly smaller than @xmath12 .
the general solution to equation ( [ zfe ] ) for a specific halo is @xmath43 \int^t_{t_0}\lambda_{\rm fe}(t'')\exp\left[\int^{t''}_{t_0 } \lambda_{\rm in}(t''')dt'''\right]dt '' , \label{zfet}\ ] ] where @xmath44 is the onset of astration and fe production [ i.e. , @xmath45 for @xmath46 . the time @xmath47 can be related to the mass of the halo at that time based on the density fluctuation associated with the halo ( see fig . 2a and 4 ) .
given the rate of infall @xmath48 and the rate of fe production per h atom in the gas @xmath12 , @xmath49 only depends upon the time @xmath44 for onset of astration and the time @xmath47 . for the infall phase
, we explicitly assume that @xmath50 is constant and take @xmath51 .
this value of @xmath12 corresponds to the average rate of fe production by sne ii in the galaxy [ @xmath52 of ( fe / h)@xmath53 over a period of @xmath54 gyr ] .
the infall rate @xmath48 is not an arbitrary function of @xmath47 but is directly determined from the model of hierarchical structure formation discussed in barkana & loeb ( 2001 ) ( see below ) .
the evolution of [ fe / h ] in halos with infall of big bang debris but without gas expulsion by sne ii can then be calculated in a straightforward manner .
the rate of gas infall governs the growth of the baryonic mass of a halo .
we assume that a constant fraction @xmath55 of the total halo mass is in baryonic matter , where @xmath56 and @xmath57 are the fractions of the present critical density contributed by baryonic and all matter , respectively .
the infall rate is then related to the total halo mass @xmath58 ( mostly in cdm ) as @xmath59 for a flat universe , we have @xmath60 where @xmath61 km
s@xmath62 mpc@xmath62 is the present hubble parameter and @xmath63 is the fraction of the present critical density contributed by the cosmological constant .
we take @xmath64 , @xmath65 , @xmath66 , and @xmath67 .
based on the model of hierarchical structure formation ( see 4 ) , we show @xmath58 as a function of @xmath6 for halos associated with @xmath68 and @xmath69 density fluctuations ( @xmath68 and @xmath69 halos ) , respectively , in figure 2a . using the function @xmath70 for a specific halo , we calculate @xmath71 from equations ( [ inf ] ) and ( [ dzdt ] ) . assuming that astration starts ( see fig .
2a and 4 ) when a @xmath68 halo reaches @xmath72 corresponding to @xmath73 ( @xmath74 gyr ) or when a @xmath69 halo reaches @xmath75 corresponding to @xmath76 ( @xmath77 gyr ) , we numerically integrate equation ( [ zfe ] ) to calculate the evolution of [ fe / h ] @xmath78 as a function of @xmath6 until @xmath79 [ @xmath80 gyr , @xmath81 for the @xmath68 halo or until @xmath82 [ @xmath83 gyr , @xmath81 for the @xmath69 halo .
the results are shown in figure 2b . if we consider smaller values of @xmath84 , say @xmath85 , then the onset of astration will be moved to higher redshift .
the evolution of [ fe / h ] for a @xmath69 halo with @xmath86 corresponding to @xmath87 ( @xmath88 gyr ) is shown as the dot - dashed curve in figure 2c , where the case for @xmath75 corresponding to @xmath76 is also shown as the solid curve .
it can be seen that the dot - dashed curve merges with the solid curve at @xmath89 .
this is because [ fe / h ] quickly approaches a quasi - steady state subsequent to the onset of astration ( see below ) .
inspection of equation ( [ zfe ] ) suggests that if @xmath90 , then a quasi - steady state ( qss ) exists such that @xmath91 the approach of @xmath49 to @xmath92 may be approximated as @xmath93 .
\label{ztoqss}\ ] ] for @xmath94^{-1}$ ] , the term containing the exponential in equation ( [ ztoqss ] ) is insignificant and @xmath49 approximately equals the qss value @xmath95 ( see qian & wasserburg 2003b ) . when we numerically integrate equation ( [ zfe ] ) to calculate the evolution of [ fe / h ] @xmath96 as a function of @xmath6 , we always use the actual function @xmath70 for a specific halo to obtain the corresponding @xmath71 . for a simple estimate of @xmath71
, we note that the slopes of the @xmath70 curves are @xmath97 within a factor of a few for both @xmath68 and @xmath69 halos ( see fig .
we also note that for @xmath98 , @xmath63 can be ignored in equation ( [ dzdt ] ) to give @xmath99 using @xmath97 and equations ( [ inf ] ) and ( [ tz ] ) , we obtain @xmath100 as a rough estimate for both @xmath68 and @xmath69 halos .
this gives for these halos @xmath101 or @xmath102}_{\rm qss}=\log z_{\rm fe}^{\rm qss}(t ) \sim -2.9 - 2.5\log\left(\frac{1+z}{10}\right ) .
\label{fehq}\ ] ] the above estimate for [ fe / h]@xmath103 is shown as the solid curve labeled `` quasi - steady state estimate '' in figure 2b .
it can be seen from figure 2b that for both @xmath68 and @xmath69 halos , [ fe / h ] of the gas initially rises very rapidly and approaches the qss value , which ranges from [ fe / h]@xmath104 at @xmath105 to @xmath106 at @xmath107 . for comparison ,
the case of evolution from an initial metal - free state without infall , @xmath108 ( prior to the onset of sne ia ) , is also shown in figure 2b .
this should give an upper bound to @xmath41 for all halos and appears to be a reasonable bound for the observational data on dla systems ( wasserburg & qian 2000b ; qian & wasserburg 2003b ; see fig .
compared to the case of no infall , the fe enrichment for @xmath68 and @xmath69 halos is suppressed due to dilution by infall by a factor of @xmath109 , which ranges from a factor of @xmath110 for @xmath111 gyr ( @xmath107 ) to a factor of @xmath112 for @xmath113 gyr ( @xmath105 ) .
we now consider that infall of gas ceases at some time for a halo and investigate the subsequent chemical evolution of gas in the halo .
the metallicity distribution of stars formed before and after infall cessation ( ic ) is studied .
the general effects of sne ia on the chemical evolution of gas and stars in a halo are also discussed at the end of this section .
we first consider the chemical evolution of gas in a halo after continuous infall ceases sharply at some time @xmath114 .
the governing equations in this case are still equations ( [ feh ] ) , ( [ ht ] ) , and ( [ zfe ] ) .
but now @xmath115 , @xmath116 , and the new initial state has @xmath117 as determined by the prior evolution with infall . thus , for @xmath118 but prior to the onset of sne ia ( see 3.3 ) , @xmath119 taking @xmath120 gyr ( @xmath121 ) for a @xmath68 halo and @xmath122 gyr ( @xmath123 ) for a @xmath69 halo , we show the above evolution of [ fe / h ] @xmath96 for these two halos in figure 2b .
it can be seen that the evolution subsequent to infall cessation departs markedly from the qss case . for the @xmath69 halo ,
the evolution clearly begins to approach the case of no infall . in these calculations
, we have assumed that infall ceases sharply at time @xmath114 corresponding to redshift @xmath124 when the mass of a halo reaches @xmath125 ( see fig . 2a and 4 ) .
if we consider larger values of @xmath126 , say @xmath127 , this will only delay the onset of uniform growth of @xmath41 to lower redshift .
the general evolution of [ fe / h ] outlined above will be unaffected .
the evolution of [ fe / h ] for a @xmath69 halo with @xmath127 corresponding to @xmath128 gyr ( @xmath129 ) is shown as the dashed curve in figure 2c .
in addition to the effect on the evolution of @xmath41 discussed above , the instant cessation of infall has another consequence . during the infall phase
, the amount of gas in a halo rapidly increases as ( h ) @xmath130 for @xmath131 .
subsequent to instant infall cessation , ( h ) steadily decreases due to astration .
thus , ( h ) is maximal at time @xmath114 . as the rate of star formation increases with ( h ) , this rate also reaches a maximum at time @xmath114 .
the general history of star formation before and after infall cessation is reflected by the metallicity distribution of stars in a halo .
this distribution is defined as the number @xmath132 of stars formed per unit interval of [ fe / h ] @xmath96 , @xmath133}=-(\ln 10)\frac{z_{\rm fe } } { ( { \rm h})_\star}\frac{d({\rm
h})_{\rm as}/dt}{dz_{\rm fe}/dt } , \label{md}\ ] ] where @xmath134 is the average number of h atoms in a star [ note that @xmath34 ] .
the metallicity distribution of stars formed at @xmath118 is simple to obtain . in this case ,
@xmath136 and @xmath137 $ ] ( cf . , eqs . [
[ ht ] ] and [ [ zfe ] ] ) .
assuming that the sn ii rate is the same as the rate at which massive sn ii progenitors are formed ( the instantaneous recycling approximation ) , we can relate @xmath30 to the astration rate @xmath138 as @xmath139 where @xmath2 is a positive dimensionless constant .
thus , for @xmath118 , @xmath140 the above equation can be solved to give @xmath141 , \label{hic}\ ] ] where ( h)@xmath142 is the number of h atoms in the gas at @xmath143 .
the metallicity distribution at @xmath118 is then @xmath133}=-(\ln 10)\frac{z_{\rm fe}}{({\rm h})_\star } \frac{d{\rm ( h)}}{dz_{\rm fe}}=\frac{(\ln 10)}{\alpha } \frac{{\rm ( h)}_{\rm ic}}{({\rm h})_\star}z_{\rm fe}\exp \left[-\frac{z_{\rm fe}-z_{\rm fe}(t_{\rm ic})}{\alpha}\right ] . \label{mdic}\ ] ] the above metallicity distribution assumes that only sne ii provide fe .
note that this distribution does not depend on the specific forms of @xmath138 and @xmath30 as functions of time but only on the ratio of these two quantities .
this is true of any closed system as recognized by hartwick ( 1976 ) long ago .
in contrast , the metallicity distribution of stars formed during the infall phase depends on the specific forms of @xmath144 and @xmath30 . in general
, the evolution of @xmath145 and ( h ) during the infall phase must be solved together as explicit functions of time from equations ( [ feh ] ) and ( [ ht ] ) after @xmath138 , @xmath30 , and the infall rate @xmath31 are specified for a given halo .
we again assume @xmath146=\lambda_{\rm fe}= ( 30\ { \rm gyr})^{-1}$ ] as in 2 . using this assumption and equation ( [ pfeas ] )
gives @xmath147 as a numerical example of estimating @xmath2 , we assume that sne ii with progenitors of @xmath148@xmath149 produce @xmath150 of fe as in the case of sn 1987a but no other sne ii produce fe . for a salpeter initial mass function @xmath151 over a range of stellar masses @xmath152@xmath26 , a total of @xmath153 of stars are formed for each fe - producing sn ii . from equation ( [ pfeas ] ) we obtain @xmath154 , where we have used an fe mass fraction of @xmath155 corresponding to the number abundance ( fe / h)@xmath53 . noting that the fe yields of sne ii , especially the dependence of these yields on the progenitor masses , are rather uncertain
, we take @xmath156 as a nominal value and examine the effects of a higher value of @xmath157 later . for a halo with @xmath158 of gas and
@xmath159 , the astration rate in equation ( [ has ] ) is 6 and @xmath160 yr@xmath62 for @xmath156 and 1/6 , respectively .
these astration rates are typically much smaller than the infall rate as assumed in 2
. however , to be more accurate , we can solve @xmath41 and ( h ) together as functions of time from equations ( [ feh ] ) , ( [ ht ] ) , and ( [ has ] ) .
these two functions can then be used to obtain the metallicity distribution at @xmath161 from equations ( [ md ] ) and ( [ has ] ) .
the metallicity distribution over the whole time domain @xmath161 and @xmath118 is shown for a @xmath69 halo with @xmath156 , @xmath77 gyr ( @xmath76 ) , and instant infall cessation at @xmath122 gyr ( @xmath123 ) in figure 3a ( solid curve ) .
it can be seen that this distribution has a sharp peak at [ fe / h ] @xmath162 and a broad peak at [ fe / h ] @xmath163 .
the broad peak occurs in the region @xmath118 and is easy to understand from equation ( [ mdic ] ) .
this equation shows that @xmath164}$ ] has a peak at @xmath165 , which corresponds to [ fe / h ] @xmath166 for @xmath156 .
the sharp peak in the metallicity distribution of stars occurs at @xmath143 , the time of infall cessation .
the origin of this peak can be understood by considering equation ( [ md ] ) in the immediate neighborhood of @xmath114 ( @xmath167 ) .
note that @xmath168 has a discontinuity at @xmath143 but @xmath41 and @xmath169 are the same on both sides of this discontinuity .
this gives @xmath170)_{t_{\rm ic}^- } } { ( dn / d[{\rm fe / h}])_{t_{\rm ic}^+ } } = \frac{(dz_{\rm fe}/dt)_{t_{\rm ic}^+ } } { ( dz_{\rm fe}/dt)_{t_{\rm ic}^- } } = \frac{\lambda_{\rm fe}}{(dz_{\rm fe}/dt)_{t_{\rm ic}^-}}. \label{ric}\ ] ] as the qss approximation @xmath171 holds during the infall phase ( @xmath161 ) , @xmath172 to zeroth order and @xmath173 is singular ( see eq . [ [ zfe ] ] ) .
to first order , @xmath174 _ { t_{\rm ic}}\sim\frac{\lambda_{\rm fe}}{6 } \left(\frac{t_{\rm ic}}{\rm gyr}\right)^{2/3 } , \label{dzfeic}\ ] ] where we have used the estimate for @xmath48 in equation ( [ infa ] ) for @xmath68 and @xmath69 halos to obtain the last result .
it follows from equations ( [ ric ] ) and ( [ dzfeic ] ) that the ratio @xmath175 for @xmath68 and @xmath69 halos is much greater than unity for a wide range of @xmath114 .
this leads to a sudden drop in @xmath164}$ ] at @xmath143 and produces a sharp peak in the metallicity distribution .
the value of [ fe / h ] at the sharp peak depends on @xmath114 . for a @xmath69 halo with @xmath122 gyr ( @xmath123 ) , this peak is at [ fe / h ] @xmath176 ( see eq . [ [ zfeqa ] ] and fig .
3a ) . to compare the effect of instant infall cessation as assumed above with that of a smooth transition for a @xmath69 halo , we have calculated models where @xmath177 decays smoothly starting at time @xmath178 with the functional form @xmath179 $ ] for @xmath118 .
the results are shown in figure 3a for @xmath180 and 0.2 gyr . as can be seen
, the main effect of changing @xmath181 from 0 ( instant infall cessation ) to a few tenths of a gyr is to increase the fraction of stars formed at @xmath118 .
this is because there is more gas available for star formation in the case where the infall rate decays smoothly over an extended time .
however , the width of the sharp peak is increased only slightly from @xmath182 dex for @xmath183 to @xmath184 dex for @xmath185 gyr .
concerning the total population of stars in the metallicity distribution , we note that only 10% of all the stars are in the sharp peak for @xmath186 and this fraction is even smaller for a finite @xmath181 . for a slowly - decaying infall rate with @xmath187 gyr ,
the sharp peak at @xmath143 is still present but contains only @xmath188 of the stars and becomes a small feature of an otherwise smooth metallicity distribution ( see fig .
3b for the case of @xmath189 gyr ) .
the bimodal form of the metallicity distribution of stars discussed above applies to all halos in general .
one critical aspect of this bimodal behavior is the value of [ fe / h ] at the sharp peak .
this value is governed by the time @xmath114 for infall cessation , which depends on the density fluctuation associated with a halo ( see 4 ) .
the metallicity distributions for a @xmath68 halo with instant infall cessation ( @xmath186 ) and smooth transition to exponential decay of infall ( @xmath189 gyr ) at a relatively late time of @xmath120 gyr ( @xmath121 ) are shown in figure 3c [ @xmath156 and @xmath74 gyr ( @xmath73 ) are used ] . compared with figures 3a and 3b for a @xmath69 halo with @xmath122 gyr ( @xmath123 ) , the sharp peak for the @xmath68 halo
is shifted to a higher [ fe / h ] of @xmath190 . to obtain
a peak with a lower [ fe / h ] of @xmath191 would require @xmath192 gyr corresponding to @xmath193 ( see eq . [ [ fehq ] ] ) .
comparison of figures 3a and 3b with 3c also shows that the sharp peak for the @xmath69 halo contains a much smaller fraction of stars than that for the @xmath68 halo .
consider the case of instant infall cessation as a specific example .
for the @xmath69 halo , the infall rate at high redshift corresponding to @xmath161 is large compared to the astration rate and only 10% of stars are formed during the infall phase .
in contrast , for the @xmath68 halo , the astration rate is no longer small compared to the infall rate at lower redshift near the end of the infall phase and @xmath194 of stars are formed during this phase . from the model presented here it follows that galaxies in general should have a bimodal distribution of stars with a relatively small number of stars defining a sharp peak at low [ fe / h ] .
this peak reflects the state when gas infall ceased or greatly decreased .
note that we have not addressed the location of stars in a galaxy when discussing their metallicity distribution .
the bimodal metallicity distribution is more likely to describe stars in the halo than those in the disk of a galaxy .
as discussed above , the metallicity distribution of stars in a halo is calculated by specifying @xmath12 , @xmath2 , @xmath44 , @xmath114 , and the form of infall cessation .
the infall rate @xmath48 used in the calculation for @xmath161 is determined from the model of hierarchical struture formation as in 2 .
the effects of the form of infall cessation ( instant or exponential decay ) on the metallicity distribution have been discussed above . to examine the effects of @xmath2 and @xmath114 ,
we have calculated the metallicity distribution for a @xmath69 halo with @xmath157 , @xmath77 gyr ( @xmath76 ) , and instant infall cessation at @xmath128 gyr ( @xmath129 ) .
this is shown as the solid curve in figure 3d .
compared to the solid curve in figure 3a [ @xmath156 and @xmath122 gyr ( @xmath123 ) ] , the sharp peak in figure 3d is shifted to [ fe / h ] @xmath195 due to a later @xmath114 and the broad peak ( solid curve ) in figure 3d is shifted to [ fe / h ] @xmath196 due to a larger @xmath2 . in addition , @xmath197 of all the stars represented by the solid curve in figure 3d have @xmath198}\lesssim 0 $ ] while the corresponding fraction is only @xmath199 for the solid curve in figure 3a . for completeness , we note that @xmath44 for the onset of astration has little effect on the metallicity distribution as most of the stars are formed close to @xmath143 for @xmath161 .
it is evident from inspection of figure 3 that if a protogalactic halo were to expel its gas immediately following the instant cessation of infall , an assemblage of stars with nearly identical [ fe / h ] would be left behind .
this assemblage of stars may be accreted by a larger system and become a globular cluster of the larger system . in this model
, the sharp distribution of [ fe / h ] for the stars in a globular cluster would not be the result of a rapid burst of star formation but rather the result of the following two factors .
first , for a protogalactic halo with ongoing infall and astration over an extended period , the rate of star formation increases rapidly with time and reaches a maximum at the time @xmath114 for instant infall cessation when there is the largest amount of gas in the halo .
this causes nearly all the stars formed during the infall phase to have the metallicity of the gas at @xmath143 .
second , the rapid loss of gas subsequent to infall cessation terminates further astration .
our model relates the value of [ fe / h ] for the globular cluster to @xmath143 , and therefore , can be used to estimate the formation time ( relative to the big bang ) of the cluster based on its [ fe / h ] if the initial metallicity of the infalling gas is negligible . to illustrate the general effects of sne ia
, we assume that these supernovae start to contribute fe when [ fe / h ] @xmath200 is reached at time @xmath201 . the rate of fe production by sne ia is taken to be twice that for sne ii so that a solar fe abundance can be achieved over a period of @xmath54 gyr as in the galaxy .
the total fe production rate per h atom in the gas then increases from @xmath12 at @xmath202 to @xmath203 at @xmath204 .
consider halos with instant infall cessation occurring at @xmath143 prior to the onset of sne ia ( i.e. , @xmath205 ) .
the evolution of @xmath41 at @xmath204 for these halos is given by @xmath206 where @xmath207 ( [ fe / h ] @xmath200 ) is taken to correspond to the onset of sne ia .
the above evolution of [ fe / h ] @xmath96 is shown for a @xmath68 halo with @xmath74 gyr ( @xmath73 ) and @xmath120 gyr ( @xmath121 ) and for a @xmath69 halo with @xmath77 gyr ( @xmath76 ) and @xmath122 gyr ( @xmath123 ) in figure 2b .
the evolution of [ fe / h ] at @xmath204 for the case of no infall is also given by equation ( [ zia ] ) and shown in figure 2b .
the only effect of sne ia on the evolution of [ fe / h ] for all halos is to increase the rate of fe enrichment at @xmath204 . as in the case of infall cessation ,
the sudden increase of the fe production rate per h atom in the gas at the onset of sne ia leads to a sharp drop of @xmath164}$ ] at [ fe / h ] @xmath200 .
this is shown for a @xmath69 halo with @xmath156 and smooth transition to exponential decay of infall ( @xmath189 gyr ) at @xmath122 gyr ( @xmath123 ) in figure 3b and for another @xmath69 halo with @xmath157 and instant infall cessation at @xmath128 gyr ( @xmath129 ) in figure 3d .
for the cases where only sne ii provide fe , the metallicity distribution of stars has a broad peak at [ fe / h ] @xmath166 for figure 3b and @xmath208 for figure 3d ( solid curves ) . if we include sn ia contributions for [ fe / h ] @xmath1 , this broad peak is truncated and a new broad peak is formed at [ fe / h ] @xmath209 for figure 3b and @xmath210 for figure 3d ( dashed curves ) .
this change to the metallicity distribution is caused by the more rapid increase of [ fe / h ] with additional contributions from sne ia .
the above effect of sne ia on the metallicity distribution of stars should apply to all halos . for completeness
, we note that the final halo mass for the @xmath69 halo shown in figure 3b is @xmath211 corresponding to a baryonic mass of @xmath212 . in this case @xmath213 of the baryonic matter
has been consumed in star formation prior to the onset of sne ia . in comparison ,
the final halo mass for the @xmath69 halo shown in figure 3d is @xmath214 corresponding to a baryonic mass of @xmath215 . in this case @xmath216 of the baryonic matter has been consumed in star formation prior to the onset of sne ia .
in the formal discussion of 2 and 3 on the chemical evolution of a halo , we used the times for the onset of astration ( @xmath44 ) and infall cessation ( @xmath114 ) and the infall rate ( @xmath48 ) , all of which are related to the growth of the halo as determined by the cdm model of hierarchical structure formation .
we now explicitly explore how this cosmological input is determined .
the physical conditions governing the general evolution of a halo are closely related to the mass ( mostly in cdm ) of the halo . for a halo associated with an @xmath217 density fluctuation ( an @xmath217 halo )
, its mass @xmath58 at redshift @xmath98 can be estimated from @xmath218 where @xmath219 is the standard deviation of the present power spectrum of overdensity on mass scale @xmath58 . using the version of equation ( [ nsig ] ) presented in the appendix and taking @xmath219 from figure 5 of barkana & loeb ( 2001 )
, we show @xmath58 as a function of @xmath6 for @xmath68 and @xmath69 halos , respectively , in figure 2a .
the function @xmath70 for a specific halo was used to calculate the infall rate @xmath48 from equations ( [ inf ] ) and ( [ dzdt ] ) in 2 .
in addition , when discussing the effects of infall cessation in 3 , we assumed that infall ceases at time @xmath114 corresponding to redshift @xmath124 at which the mass of a halo reaches @xmath220 . for @xmath221 , @xmath121 or 4.3 for a @xmath68 or @xmath69 halo , respectively ( see figure 2a ) .
if @xmath222 is assumed , then infall cessation is delayed to @xmath223 or 2.8 for a @xmath68 or @xmath69 halo , respectively .
the redshifts at which infall cessation occurs for other @xmath217 halos are given for @xmath224 and @xmath225 , respectively , by the corresponding curves in figure 4 .
the range of @xmath226 to @xmath225 is in accord with the halo masses of typical galaxies .
a critical physical parameter for the onset of astration in a halo is the virial temperature @xmath227 of its gas . for @xmath98 this temperature
is related to the halo mass @xmath58 as @xmath228 where @xmath229 is the mean molecular weight and @xmath230 or 0.6 for a neutral or ionized gas of primordial composition , respectively .
following barkana & loeb ( 2001 ) , we assume that astration starts in a halo when the gas reaches a threshold virial temperature @xmath231 . at a given redshift @xmath6 ,
the mass @xmath232 that is required for a halo to reach @xmath233 can be obtained from equation ( [ tvir ] ) . using this equation ( see appendix ) ,
we show @xmath234 as a function of @xmath6 for @xmath235 k ( @xmath230 ) and @xmath236 k ( @xmath237 ) , respectively , in figure 2a .
the redshift @xmath238 at which the mass @xmath70 of an @xmath217 halo matches @xmath239 then corresponds to the onset of astration ( @xmath44 ) in this halo . for @xmath236
k , @xmath73 or 9.7 for a @xmath68 or @xmath69 halo , respectively ( see figure 2a ) . if @xmath235 k is assumed , then the onset of astration occurs earlier at @xmath240 or 16.8 for a @xmath68 or @xmath69 halo , respectively .
the redshifts at which astration starts for other @xmath217 halos are given for @xmath235 or @xmath241 k , respectively , by the corresponding curves in figure 4 .
the value of @xmath231 for the onset of astration depends crucially on the available coolant . in the presence of cooling by h@xmath242 molecules ,
@xmath235 k is appropriate .
if h@xmath242 molecules are dissociated ( e.g. , couchman & rees 1986 ; haiman , rees , & loeb 1997 ; ciardi , ferrara , & abel 2000 ) , then @xmath236 k is required for efficient cooling by atomic species . while the choice of @xmath235 or @xmath241 k does not affect the long - term chemical evolution of a halo ( see fig .
2c ) , it can make an important difference in the early evolution of the halo .
this difference concerns the possibility for a single sn ii to expel all the gas , thereby disrupting the chemical evolution of a halo .
this possibility can be assessed by comparing the explosion energy of @xmath243 erg for an sn ii with the binding energy @xmath244 of the gas in the halo .
for @xmath98 this binding energy is related to the halo mass @xmath58 as @xmath245 we consider a simple scenario where all the gas is expelled when @xmath244 for a halo is less than @xmath246 erg . using equation ( [ ebgas ] ) ( see appendix ) ,
we show the halo mass @xmath247 for @xmath248 erg as a function of @xmath6 in figure 2a .
the redshifts at which different @xmath217 halos reach @xmath248 erg are given by the corresponding curve in figure 4 .
as can be seen from this figure , if @xmath235 k is assumed for the onset of astration , then the chemical evolution of all halos will be disrupted by sne ii until @xmath248 erg is reached .
in contrast , for @xmath236 k , such disruption can not occur as @xmath248 erg is always reached prior to the onset of astration .
we now compare our chemical evolution model with the data on [ fe / h ] for dla systems . using @xmath236 k for the onset of astration , @xmath221 for instant infall cessation , and [ fe / h ] @xmath200 for the onset of sne ia
, we show the evolution of [ fe / h ] as a function of @xmath6 for @xmath249 , @xmath68 , and @xmath69 halos along with the data on dla systems ( prochaska et al .
2003 ) in figure 5 ( see also qian & wasserburg 2003b ) . the case of no infall is also shown in this figure .
it can be seen that the chemical evolution model presented here provides a good description of the data . for any given @xmath6
the model gives a wide range of [ fe / h ] corresponding to three different evolutionary stages of dla systems : sharp initial rise after the onset of astration , quasi - steady state due to competition between fe production and dilution by infall , and uniform growth subsequent to infall cessation .
the evolutionary stage of a dla system at a given @xmath6 depends on the density fluctuation associated with the system . the large dispersion in [ fe / h ] for dla systems at any given @xmath6 is then accounted for by the different density fluctuations associated with these systems .
a detailed quantitative assessment of the model would require many more dla systems with accurate measurements of [ fe / h ] at a given @xmath6 ( see qian & wasserburg 2003b for an illustrative example ) .
the model also provides an explanation for the baseline enrichment of [ fe / h ] @xmath7 observed for dla systems .
subsequent to the onset of astration , it takes only @xmath250 myr to reach [ fe / h ] @xmath7 for @xmath159 . due to this rapid initial rise of [ fe / h ]
, the probability to find dla systems with [ fe / h ] below @xmath7 is very low .
the baseline enrichment of [ fe / h ] @xmath7 for dla systems is similar to the abundance level in the igm that was attributed to a prompt inventory of metals suggested by other arguments ( e.g. , wasserburg & qian 2000a ; qian & wasserburg 2002 ; qian , sargent , & wasserburg 2002 ) . based on the model presented here , this postulated prompt
inventory is not required to explain the data on dla systems although it is also not in conflict with these data .
we presented a model for the chemical evolution of the gas and stars in halos as a function of redshift .
this model assumes that a normal population of stars including sn ii progenitors can form from big bang debris and that sne ii will provide metals such as the elements from the fe group down to o. if there is a prompt inventory of metals corresponding to [ fe / h ] @xmath4 in the igm , this will simply provide a baseline enrichment for the initial state and will not alter the results .
the infall of gas into a halo is assumed to closely follow that of cdm .
the infall rate and the general evolution of the halo are prescribed by the model of hierarchical structure formation based on the density fluctuation associated with the halo . for systems that are not disrupted by sn ii explosions ,
the chemical evolution of the gas and stars is explicitly determined by the infall rate as compared with the astration rate and the corresponding rate of metal production by sne ii per h atom in the gas .
once the astration rate and the metal production rate are specified , the only `` free '' parameters in this model are the times ( or states ) at which astration starts and infall ceases for a halo . for all systems not disrupted by sn ii explosions ,
if there is a cessation of gas infall , the metallicities of stars follow a bimodal distribution .
this distribution is characterized by a sharp peak at the value of [ fe / h ] corresponding to the time of infall cessation and by a broad peak at a higher value of [ fe / h ] corresponding to the subsequent period of astration during which the bulk of the remaining gas forms stars .
the fraction of stars in the sharp peak depends on how rapidly the infall cessation takes place .
while it is not clear how observational bias may affect the proper metallicity distribution of stars , we note that the distribution observed for the galactic halo stars ( e.g. , fig . 5 in christlieb 2003 ) and
the bimodal distribution presented here show a strong resemblance . based on our model , a distinct narrow peak at low [ fe / h ] for
the galactic halo stars would signify the time when infall started to decrease for the galaxy .
we also discussed the general effects of sne ia on the chemical evolution of the gas and stars in halos . for [ fe / h ]
@xmath28 the production rate for the fe group elements is increased due to additional contributions from sne ia .
consequently , the broad peak in the metallicity distribution of stars corresponding to the case where only sne ii provide fe is truncated at [ fe / h ] @xmath23 and a new broad peak is formed at a higher [ fe / h ] value .
we also considered the physical conditions governing the evolution of a halo .
the mass @xmath58 of a halo is an important parameter and its growth as a function of redshift is determined by the model of hierarchical structure formation based on the @xmath217 density fluctuation associated with the halo .
we assumed that the cessation of infall for both baryonic and dark matter corresponds to the redshift at which the halo mass reaches @xmath226 to @xmath214 .
this range of @xmath251 is in accord with halo masses of typical galaxies .
the other two important parameters of a halo are the virial temperature @xmath227 and the binding energy @xmath244 of the gas .
we assumed that astration starts in a halo when a critical virial temperature @xmath231 corresponding to a halo mass @xmath234 is reached . in the presence of cooling by h@xmath242 molecules ,
@xmath235 k is appropriate and @xmath234 ranges from @xmath252 to @xmath253 for different @xmath217 halos .
if h@xmath242 molecules are dissociated , then @xmath236 k is required and @xmath234 ranges from @xmath254 to @xmath255 .
while the assumed value of @xmath231 does not affect the long - term chemical evolution of a halo , it makes an important difference concerning the effect of sn ii explosions on the early evolution of the halo . for @xmath236
k , astration starts in a halo after the halo mass reaches @xmath256 corresponding to a gas binding energy of @xmath248 erg . in this case , the chemical evolution of the gas and stars in the halo will not be disrupted by sn ii explosions and is summarized in the preceding paragraph .
in contrast , for @xmath235 k , astration starts in a halo before the halo mass reaches @xmath256 and all low - mass halos with @xmath257 will have their gas dispersed into the igm by a single sn ii .
if such low - mass halos indeed form , then their gas expulsion by sne ii may significantly enrich the igm at very high redshift .
for example , at @xmath258 , halos with @xmath257 are associated with 1.31.7@xmath17 density fluctuations ( see fig .
4 ) and a fraction @xmath259 of all baryonic matter is being processed through such halos . taking @xmath260 ( see eq . [ [ zfeqa ] ] ) for the gas expelled from such halos , we estimate that their gas expulsion would result in an average metallicity of @xmath261 corresponding to [ fe / h]@xmath262 in the igm at @xmath258 .
we expect that @xmath231 for the onset of astration would increase from 300 to @xmath241 k as various radiation sources turned on during the evolution of the universe . so the range of low - mass halos that would be disrupted by sn ii explosions would decrease with time and eventually vanish .
the question of whether such low - mass halos may actually provide significant metal enrichment of the igm at very high redshift will be discussed in a subsequent paper that considers aspects regarding dissociation of h@xmath242 molecules .
we proposed a possible mechanism of globular cluster formation based on the sharp peak in the metallicity distribution of stars .
this peak corresponds to the time of infall cessation .
if infall ceases rather suddenly for a protogalactic halo and its gas is rapidly expelled afterwards , an assemblage of stars with nearly identical [ fe / h ] will be left behind .
this assemblage of stars may be accreted by a larger system and become a globular cluster of the larger system . to estimate the masses and metallicities of such globular clusters , we assume @xmath236 k for the onset of astration . in this case
the gas in a halo will be enriched by sne ii without disruption subsequent to the onset of astration . during the infall
phase the amount of infalling gas increases rapidly on a timescale of @xmath263^{5/2}$ ] gyr ( see eq . [ [ infa ] ] ) . for an astration rate of @xmath264 with @xmath159 and @xmath156 ,
the fraction of the halo mass @xmath58 in stars at redshift @xmath6 is @xmath265 and can be estimated as @xmath266^{5/2}$ ] .
if infall ceases at redshift @xmath6 , then the sharp peak in the metallicity distribution contains a total mass @xmath267 of stars with essentially the same [ fe / h ] as given approximately by [ fe / h]@xmath103 for this @xmath6 ( see eq . [ [ fehq ] ] ) .
the assumed value of @xmath236 k corresponds to @xmath268 and @xmath269 for @xmath258 and 4 , respectively ( see fig .
so in the above scenario , the globular clusters would have masses exceeding @xmath270 and [ fe / h ] @xmath7 if they were formed at @xmath258 while they would have masses exceeding @xmath271 and [ fe / h ] @xmath272 if they were formed at @xmath273 .
no small globular clusters with masses of @xmath274 are formed in this scenario unless the larger aggregates are fragmented .
we note that the masses derived for galactic globular clusters are rather uncertain and range from @xmath275 to @xmath276 with a mean of @xmath277 ( mandushev , spassova , & staneva 1991 ; richer et al .
the metallicity distribution of galactic globular clusters is bimodal with peaks at [ fe / h ] @xmath278 and @xmath279 , respectively ( armandroff & zinn 1988 ) .
approximately 10% of the clusters have [ fe / h ] @xmath280 .
the masses and metallicities of globular clusters in our model are broadly consistent with those of the galactic population .
a detailed comparison requires considerations of possible ranges for @xmath12 , @xmath48 , and @xmath2 .
we also note that stars in an individual galactic globular cluster typically have a very narrow range of [ fe / h ] , which can be explained by our model
. however , there is a considerable range in [ fe / h ] for the most massive galactic globular cluster @xmath281 centauri ( suntzeff & kraft 1996 ; norris , freeman , & mighell 1996 ) with a mass of @xmath282 ( richer et al .
the range of [ fe / h ] for @xmath281 centauri may have resulted from further astration subsequent to infall cessation based on our model . as an sn ii can not unbind the gas in the protogalactic halo associated with such a large cluster and some additional mechanism for gas loss must be involved , a more extensive evolution of [ fe / h ] would be expected .
the gas loss may have to occur by e.g. , passing through the galactic disk .
if we assume @xmath235 k for the onset of astration , then a different scenario of globular cluster formation may occur .
this scenario is possible if h@xmath242 molecules are not dissociated and the igm is not reionized so we here only consider it for very high redshift , say @xmath283 . for @xmath258 , @xmath284 for the assumed value of @xmath235 k and @xmath285 ( see fig .
a low - mass halo with @xmath257 has started astration but its gas will be expelled by a single sn ii . as the metals produced by the sn ii are expelled along with the gas , any metals in the stars already formed can not be due to this sn ii but must have come from the infalling gas . in other words ,
the metallicities of these stars represent the metal inventory of [ fe / h ] @xmath4 provided to the igm at @xmath283 by sources at even higher redshift .
the assemblage of stars after the sn ii explosion resembles a globular cluster with a mass of @xmath286 to @xmath287 .
we note that galactic globular clusters in this mass range typically have [ fe / h ] much higher than @xmath191 .
whether globular clusters in the same mass range but with [ fe / h ] @xmath4 exist remains to be seen .
the possibility that low - mass halos might form at @xmath288 from an enriched igm also remains to be explored .
we would like to acknowledge the anonymous referee for one of our earlier studies on cosmochemical evolution ( qian & wasserburg 2003a ) for being supportive and for suggesting that we might pay attention to infall .
good advice sometimes does fall in .
we also want to thank the anonymous referee for the present work and the scientific editor , brad gibson , for helpful suggestions .
this work was supported in part by doe grants de - fg02 - 87er40328 , de - fg02 - 00er41149 ( y. z. q. ) and de - fg03 - 88er13851 ( g. j. w. ) , caltech division contribution 8909(1111 ) .
in this appendix , we summarize the equations that are used to calculate the mass @xmath58 , the virial temperature @xmath227 , and the gas binding energy @xmath244 of a halo for any redshift @xmath6 . while these equations were discussed in many works , our discussion here closely follows that of barkana & loeb ( 2001 ) .
we consider a flat universe with @xmath66 , @xmath65 , @xmath67 , and @xmath64 .
the evolution of the mass @xmath58 for an @xmath217 halo is prescribed by @xmath289 where @xmath290 is the growth factor ( peebles 1980 ) and @xmath219 is the standard deviation of the present power spectrum of overdensity on mass scale @xmath58 .
the growth factor can be approximated as ( e.g. , eisenstein & hu 1998 ) @xmath291^{4/7 } -\omega_\lambda^z+[1+(\omega_m^z/2)][1+(\omega_\lambda^z/70 ) ] } , \label{gf}\ ] ] where @xmath292 note that the proportionality relation in equation ( [ gf ] ) is sufficient to determine the ratio @xmath293 in equation ( [ d0z ] ) . for @xmath98 , equation ( [ d0z ] )
can be simplified to give equation ( [ nsig ] ) for the adopted cosmological parameters .
the virial temperature @xmath227 of the gas in a halo of mass @xmath58 at redshift @xmath6 is @xmath294 where @xmath295 ( cf . bryan & norman 1998 ) .
the binding energy @xmath244 of the gas is @xmath296 for @xmath98 , equations ( [ tvirz ] ) and ( [ ebgasz ] ) can be simplified to give equations ( [ tvir ] ) and ( [ ebgas ] ) , respectively , for the adopted cosmological parameters . | we present an analytical and phenomenological model for metal enrichment in halos based on hierarchical structure formation .
this model assumes that astration of normal stellar populations along with type ii supernovae ( sne ii ) already occurs at very high redshift .
it focuses on the regime of [ fe / h ] @xmath0 where sne ia are not major contributors .
the regime of [ fe / h ] @xmath1 where sne ia are major contributors is also discussed in general . for halos that are not disrupted by sn ii explosions ,
the chemical evolution of the gas and stars is explicitly determined by the rate of gas infall as compared with the astration rate and the corresponding rate of metal production by sne ii per h atom in the gas .
this model provides a good description of the data on [ fe / h ] for damped ly@xmath2 systems over a wide range of redshift @xmath3 . for all halos not disrupted by sn ii explosions ,
if there is a cessation of gas infall , the metallicities of stars follow a bimodal distribution .
this distribution is characterized by a sharp peak at the value of [ fe / h ] corresponding to the time of infall cessation and by a broad peak at a higher value of [ fe / h ] corresponding to the subsequent period of astration during which the bulk of the remaining gas forms stars . such a distribution may be compared to that observed for the galactic halo stars . if the gas in a halo is rapidly lost upon cessation of infall , then an assemblage of stars with a very sharply - defined [ fe / h ] value will be left behind .
this assemblage of stars may be accreted by a larger system and become a globular cluster of the larger system .
we also discuss the masses and metallicities of the globular clusters in this model considering the criterion for the onset of astration and possible disruption of low - mass halos by sn ii explosions .
the implications of possible globular clusters with very low metallicities of [ fe / h ] @xmath4 are discussed . |
stars and brown dwarfs in the immediate solar neighborhood ( @xmath11 pc ; @xcite ) are ideal targets for detailed investigations of the structural , atmospheric and populative properties of these objects .
this is particularly true for the very lowest - mass ( vlm ) and lowest - luminosity dwarfs the late - m , l , t and y dwarfs whose recent discovery has been facilitated by wide - area red and infrared imaging surveys ( e.g. , 2mass , denis , sdss , ukidss , wise ) and multi - epoch red and infrared astrometry ( e.g. , lspm , supercosmos , allwise ) . yet , despite their apparent brightness and high proper motion , up to @xmath1020% of our nearest ( @xmath1220 pc ) vlm neighbors remain `` missing '' , particularly toward the galactic plane @xcite , as exemplified by the very recent discovery of the third ( luhman 16ab ; @xcite ) and fourth ( wise j085510.83 - 071442.5 ; @xcite ) closest systems to the sun , comprised of l , t and y dwarfs .
several very low - temperature t and y dwarfs and even m dwarfs ( e.g. , ) have also been uncovered within 5 pc of the sun in the past decade .
one of the most recent nearby discoveries is wise [email protected] ( hereafter wise j0720@xmath00846 ) , a candidate m / l dwarf identified by in the wise survey . with a parallax distance measurement of only [email protected] pc , this source had been `` hiding '' in the galactic plane , its modest proper motion ( 123@xmath12 mas yr@xmath2 ) preventing it from being picked up in earlier astrometric surveys . estimated a photometric classificaiton of m9@xmath11 , which was confirmed in near - infrared spectroscopy reported by @xcite . given its proximity to the sun , wise j0720@xmath00846 is an important new system for investigating the m dwarf / l dwarf and star / brown dwarf transitions at high spatial , spectral and temporal resolution . in this article , we report new observations that confirm wise j0720@xmath00846 to be a nearby ( 6 pc ) , late - m dwarf , and show it to have old disk kinematics , magnetic activity ( including flares ) , and rapid rotation .
we further identify it as a spectral binary system with a t - type brown dwarf companion , possibly resolved at a separation of @xmath100.8 au and estimated orbital period of 24 yr . in section 2
we describe our observations of the system , including low- and high - resolution optical and near - infrared spectroscopy , high - resolution imaging , and red optical monitoring . in section
3 we analyze these data , determining the optical and near - infrared classifications , improved astrometry , spatial kinematics , rotation , magnetic emission and limits on the photometric variability of this source . in section 4
we identify the binary nature of wise j0720@xmath00846 , both as a spectral binary and a potentially resolved system , and characterize its orbital properties from imaging and multi - epoch radial velocity measurements .
in section 5 we review the physical properties of this source , and examine it in the context of multiplicity among late m dwarfs in the immediate vicinity of the sun .
results are summarized in section 6 .
wise j0720@xmath00846 was observed on 2013 december 30 ( ut ) with the ritchey - chretien spectrograph ( rc spec ) on the kitt peak national observatory s 4 m mayall telescope .
rc spec was equipped with the t2ka ccd , and we used the bl420 grating blazed at 7800 ( first order ) and gg-475 order blocking filter with the 1@xmath85 @xmath13 98@xmath14 slit aligned north to south , yielding spectral data spanning 63009000 at a resolving power of @xmath103 .
the source was observed near transit at an airmass of 1.32 in a single integration of 1000 s. we also observed the spectrophotometric standard hr 3454 @xcite at an airmass of 1.15 for flux calibration , as well as flat field and henear arc lamps .
data were reduced using custom interactive data langauge ( idl ) routines which corrected images for bad pixels and cosmic rays , performed bias - subtraction , and flat - fielding .
after sky background subtraction and spectral extraction ( summing along the spatial axis ) , we calibrated the wavelength scale to air wavelengths using the arclamp spectrum , and corrected for throughput losses using the standard star observation .
no correction for telluric absorption was attempted .
the reduced spectrum is shown in figure [ fig : optspec ] and discussed further in section 3 .
a low - resolution near - infrared spectrum of wise j0720@xmath00846 was obtained on 2013 december 5 ( ut ) using the spex spectrograph mounted on the 3 m nasa infrared telescope facility ( irtf ; @xcite ) .
conditions were clear with 1@xmath14 seeing at @xmath15-band .
we used the spex prism mode with the 0@xmath85 slit aligned with the parallactic angle , yielding 0.82.45 @xmath16 spectra with an average resolution @xmath17 @xmath18 120 .
six exposures of 30 s each were obtained at an airmass of 1.14 , followed by observations of the a0 v star hd 56525 ( @xmath19 = 7.19 ) at an airmass of 1.20 .
henear arc lamps and quartz lamp exposures were also obtained for dispersion and pixel response calibration .
data were reduced using the spextool package version 3.4 @xcite following standard procedures for point - source extraction .
these data are similar to those reported by @xcite .
the reduced spectrum is shown in figure [ fig : spex ] and discussed further in section 4 .
wise j0720@xmath00846 was observed with the hamilton echelle spectrograph @xcite on the lick observatory shane 3 m telescope over 7 nights between 21 december 2013 and 26 february 2014 ( ut ; table [ tab : hamilton ] ) .
conditions varied from clear to overcast , with seeing between 0@xmath88 - 1@xmath85 .
we used the 640 @xmath16 slit , 31.5 lines mm@xmath2 grating , and ubk ( crown glass ) cross - dispersing prisms to obtain 350010000 spectroscopy over 107 orders , at a resolution of @xmath17 = 62,000 as measured from the full - width at half - maximum ( fwhm ) of single arclines in lamp spectra .
multiple integrations of 40 min were obtained while the source was above an airmass of 2.0 . on each night , halogen lamp flat field frames were obtained for pixel response calibration , and tiar arclamp spectra were obtained for wavelength calibration .
the m4 radial velocity standard gj 251 @xcite was observed each night following wisej0720@xmath00846 for radial velocity calibration .
llccccc 2013 dec 21 & 56648.02218 & 5 & + [email protected] & + 76@xmath15 & 2@xmath11 & @[email protected] + 2013 dec 22 & 56648.93341 & 15 & + [email protected] & + 82@xmath15 & 1@xmath11 & @[email protected] + 2014 jan 03 & 56660.96342 & 15 & + [email protected] & + 78@xmath15 & 6@xmath11 & @[email protected] + 2014 jan 21 & 56678.92950 & 5 & + [email protected] & + 79@xmath15 & 5@xmath13 & @[email protected] + 2014 feb 23 & 56711.82038 & 25 & + [email protected] & + 79@xmath15 & 13@xmath15 & @[email protected] + 2014 feb 24 & 56712.80845 & 15 & + [email protected] & + 80@xmath15 & 5@xmath12 & @[email protected] + 2014 feb 25 & 56713.75047 & 40 & + [email protected] & + 80@xmath15 & 2@xmath11 & @[email protected] + data were reduced in the iraf environment following @xcite .
briefly , data were bias subtracted , flat - fielded , box - car extracted , and wavelength calibrated using ti and ar lines from arclamp exposures @xcite .
these data are discussed further in section 3 .
high resolution near - infrared @xmath20-band spectra of wise j0720@xmath00846 were obtained with the nirspec echelle spectrograph on the keck ii telescope @xcite on four nights , 2014 january 19 and 20 , march 10 and april 12 ( ut ; table [ tab : nirspec ] ) . for each night
we used the n7 order - sorting filter and 0@xmath8432-wide slit to obtain 2.002.39 @xmath16 spectra over orders 3238 with @xmath17 = 20,000 ( @xmath21 = 15 km s@xmath2 ) and dispersion of 0.315 pixel@xmath2 .
two dithered exposures of 360600 s each were obtained , along with observations of the nearby a0 v stars hd 65102 ( night 1 , @xmath19 = 6.83 ) and hd 65158 ( nights 2 - 4 , @xmath19 = 7.16 ) .
flat field and dark frames were obtained at the start of each night for detector calibration .
llcc 2014 jan 19 & 56676.51 & + [email protected] & [email protected] + 2014 jan 20 & 56677.50 & + [email protected] & [email protected] + 2014 mar 10 & 56726.22 & + [email protected] & [email protected] + 2014 apr 12 & 56759.24 & + [email protected] & [email protected] + data were reduced using a modified version of the redspec package , which took the rectified images produced by that routine and optimally extracted the source spectrum with background subtraction .
we focused exclusively on order 33 ( 2.292.33 @xmath16 ) which samples co @xmath22 = 2 - 0 transitions and telluric co , h@xmath23o and ch@xmath24 features . the optimally extracted spectra , which had signals - to - noise exceeding 100 on each night , were forward modeled as described in section 3 .
wise j0720@xmath00846 was observed with the sodium laser guide star adaptive optics system ( lgsao ; @xcite ) and facility near - infrared camera 2 ( nirc2 ) on the 10 m keck ii telescope on 19 january 2014 ( ut ) . conditions were clear , dry and windy with slightly poor seeing ( @xmath101 ) . the narrow field - of - view ( fov ) camera of nirc2 was utilized , providing an image scale of @xmath25 mas pixel@xmath2 @xcite over a 10@xmath82 @xmath13 10@xmath82 area .
we used the mko @xmath26-band filter , and obtained six 60 s integrations using a three - position dither pattern with variable step size . while the lgs provided the wavefront reference source for ao correction , tip - tilt aberrations and slow variations were measured by monitoring the @xmath27 mag field star usno 0812 - 0137390 located 20@xmath14 northeast from wise j0720@xmath00846 .
as this tip - tilt star is faint , and conditions were marginal , we were only able to achieve a strehl of 1.4% with these data .
data were reduced using custom routines to perform flat - fielding , background subtraction , bad - pixel correction , and shifting - and - stacking .
the reduced image is shown in figure [ fig : image ] .
the poor strehl is evident in the broad wings of the point spread function ( psf ) , which has a azimuthally - averaged fwhm of 0@xmath823 , and there is a slight elongation perpendicular to the direction of the tip - tilt star .
these data are analyzed further in section 4 .
wise j0720@xmath00846 was monitored for 13 non - consecutive days between 30 december 2013 and 16 february 2014 ( ut ) with the transiting planets and planetesimals small telescope ( trappist ; @xcite ) , a 0.6 m robotic telescope located at la silla observatory in chile .
the telescope is equipped with a thermoelectrically - cooled 2k@xmath132k ccd camera with a 0@xmath865 pixel scale and a 22@xmath2822@xmath29 field of view .
light is passed through a broad - band @xmath30 + @xmath31 filter with @xmath3290% transmission from 0.751.1 @xmath16 , the long - wavelength cutoff set by the quantum efficiency of the ccd detector .
individual exposures of 40 s each were obtained for continuous periods ranging from 4.8 to 8.5 hours , for a total of 73.6 hr on source .
data were reduced as described in . after a standard pre - reduction ( bias , dark , flatfield correction ) , aperture photometry was performed using iraf / daophot2 @xcite .
differential photometry was then determined by comparison to a grid of non - varying background stars , and the overall light curve was normalized to its mean value .
relative light curves for wise j0720@xmath00846 and a nearby comparison star 2mass j07200688@xmath00846504 ( @[email protected] , @xmath33 = [email protected] ) are shown in figure [ fig : trappist ] and [ fig : trappist_comp ] as a function of modified heliocentric julian day ( hjd ) .
these lightcurves are discussed in detail in section 3 .
in addition , nightly astrometry measurements for wise j0720@xmath00846 were determined from the december through february data , as well as a short sequence of 23 frames obtained on 16 september 2014 ( ut ) . to register the astrometric frames
, we matched 150 bright , red sources to the ppmxl catalogue @xcite , using a third - order polynomial ( with cross - terms ) to map pixels to position .
observations on either side of meridian transit had root mean square deviations of order 15 mas ; however , a larger shift of order 30 - 60 mas was found between frames taken before and after meridian .
we therefore adopted the mean of pre- and post - meridian measurements as nightly values , and used the cross - meridian shift as an estimate of uncertainty .
these values ( equinox j2000 ) are listed in table [ tab : astrometry ] and discussed further in section 3 .
llccccc 2014 jan 4 & 56662.5 & 07@xmath3420@xmath3503@xmath362108 & 31 & @xmath008@xmath3746@xmath2951@xmath8830 & 28 & 573 + 2014 jan 5 & 56663.5 & 07@xmath3420@xmath3503@xmath362108 & 14 & @xmath008@xmath3746@xmath2951@xmath8832 & 34 & 575 + 2014 jan 6 & 56664.5 & 07@xmath3420@xmath3503@xmath362107 & 23 & @xmath008@xmath3746@xmath2951@xmath8832 & 36 & 573 + 2014 jan 9 & 56667.5 & 07@xmath3420@xmath3503@xmath362100 & 17 & @xmath008@xmath3746@xmath2951@xmath8832 & 33 & 586 + 2014 jan 10 & 56668.5 & 07@xmath3420@xmath3503@xmath362102 & 25 & @xmath008@xmath3746@xmath2951@xmath8825 & 36 & 492 + 2014 jan 17 & 56675.5 & 07@xmath3420@xmath3503@xmath362091 & 16 & @xmath008@xmath3746@xmath2951@xmath8827 & 43 & 325 + 2014 feb 6 & 56695.5 & 07@xmath3420@xmath3503@xmath362094 & 11 & @xmath008@xmath3746@xmath2951@xmath8821 & 55 & 506 + 2014 feb 7 & 56696.5 & 07@xmath3420@xmath3503@xmath362060 & 19 & @xmath008@xmath3746@xmath2951@xmath8813 & 28 & 506 + 2014 feb 9 & 56698.5 & 07@xmath3420@xmath3503@xmath362058 & 41 & @xmath008@xmath3746@xmath2951@xmath8815 & 28 & 486 + 2014 feb 13 & 56702.5 & 07@xmath3420@xmath3503@xmath362052 & 29 & @xmath008@xmath3746@xmath2951@xmath8809 & 29 & 483 + 2014 feb 14 & 56703.5 & 07@xmath3420@xmath3503@xmath362063 & 48 & @xmath008@xmath3746@xmath2951@xmath8822 & 57 & 480 + 2014 feb 16 & 56705.5 & 07@xmath3420@xmath3503@xmath362045 & 34 & @xmath008@xmath3746@xmath2951@xmath8809 & 29 & 472 + 2014 sep 17 & 56908.9 & 07@xmath3420@xmath3503@xmath362214 & 16 & @xmath008@xmath3746@xmath2951@xmath8750 & 16 & 23 +
with 13 nights of precision trappist astrometry we attempted to validate and improve the parallax and proper motion measurements reported by .
as our new measurements span less than a year , we combined the survey data reported by scholz with our trappist measurements , rejecting only the supercosmos sky survey h@xmath3 and short - r survey measurements @xcite for which the psfs were reported to be elliptical .
a parallax solution was determined using a monte carlo markov chain ( mcmc ) analysis with the metropolis - hastings sampling algorithm .
starting from the parameters determined by scholz ( parallax , proper motion ) and the average position of the target , we performed a 10@xmath38-step random walk , at each step varying the astrometric parameters using normal distributions .
we compared the f - distribution probability distribution function for @xmath39 residual values computed before and after each step to a uniform distribution as our acceptance ratio . ignoring the first 10% of the chain , we calculated the mean and standard deviation of the parameters in the remaining steps , weighting each solution by the f - distribution .
our results , displayed in figures [ fig : astrometry1 ] and [ fig : astrometry2 ] and listed in table [ tab : properties ] , are in agreement with but with improved uncertainties , particularly in proper motion . our derived parallax distance , [email protected] pc , is more in line with the spectrophotometric estimates reported by scholz , and the improvement is driven largely by the more accurate trappist astrometry .
we aim to continue monitoring this source for further improvement .
lcl @xmath3 ( j2000 ) & 07@xmath3420@xmath3502@xmath3619 + @xmath40 ( j2000 ) & @xmath008@xmath3746@xmath2959@xmath853 + optical spt & m9.5 + nir spt & m9+t5 + @xmath41 ( mas ) & 166@xmath128 + @xmath42 ( pc ) & [email protected] + @xmath43 ( mas yr@xmath2 ) & @[email protected] + @xmath44 ( mas yr@xmath2 ) & @[email protected] + @xmath45 ( km s@xmath2 ) & [email protected] + rv ( km s@xmath2 ) & + [email protected] + @xmath46 ( km s@xmath2 ) & @[email protected] + @xmath19 ( km s@xmath2 ) & @[email protected] + @xmath47 ( km s@xmath2 ) & [email protected] + @xmath9 ( km s@xmath2 ) & [email protected] + @xmath4@xmath5@xmath6 & @xmath48 + @xmath49 & [email protected] + @xmath7 & [email protected] + @xmath50 & [email protected] + @xmath51 ( mas ) & 139@xmath114 + @xmath51 ( au ) & [email protected] + the reduced rc spec spectrum of wise j0720@xmath00846 is compared to an m9.5 spectral template in figure [ fig : optspec ] , produced by merging the m9 and l0 sdss templates from @xcite .
this hybrid template provides the best overall match to the optical spectral shape of wise j0720@xmath00846 .
spectral indices from @xcite ; and @xcite are also consistent with this classification and indicate solar metallicity ( table [ tab : indices ] ) . we detect no 6708 li i absorption in these data or the co - added high resolution spectrum , to a 3@xmath52 equivalent width limit of @xmath53 0.15 ;
we also find no indication of low surface gravity in the strengths of alkali lines or tio / vo bands @xcite .
both observations imply a mass @xmath540.06 m@xmath55 and age @xmath54100 myr for this source @xcite .
h@xmath3 emission is detected , as described below .
lcc + tio5 & 0.463@xmath1 0.007 & + cah1 & [email protected] & + cah2 & [email protected] & + cah3 & [email protected] & + + crh - a & [email protected] & m9/l0 + rb - b / tio - b & [email protected] & m9/l0 + cs - a / vo - b & [email protected] & m9/l0 + + vo1 & [email protected] & m9.7 + vo2 & [email protected] & m8.6 + tio6 & [email protected] & + tio7 & [email protected] & m8.0 + color - m & [email protected] & m9.1 + + @xmath56 & [email protected] & dm + for our near - infrared data , we compared the spex spectrum to 727 equivalent spectra of optically - classified m and l dwarfs drawn from the spex prism library ( spl ; @xcite ) . following the fitting methodology described in @xcite
, we found a best match to the m9 dwarf lhs 2924 ( @xcite ; figure [ fig : spex ] ) and a f - distribution weighted mean to all templates of [email protected] .
subtle variances between the spectrum of wise j0720@xmath00846 and lhs 2924 are discussed in detail below .
these results confirm the near - infrared classification derived by @xcite and establish congruence between the optical and near - infrared spectral morphologies .
the small proper motion of wise j0720@xmath00846 , [email protected] mas yr@xmath2 , translates into a relatively low tangential velocity , [email protected] km s@xmath2 .
we determined the radial velocity of this source from our high - resolution optical and near - infrared data .
for the optical data , we cross - correlated each of the spectra across several orders in the 66009000 range to the rv standard gl 251 ( @xmath57 = [email protected] km s@xmath2 ; @xcite ) to determine a relative offset .
the widths of the cross - correlation peak in each order yield velocity uncertainties of 0.51.0 km s@xmath2 ( @xmath101/5 of the resolution ) , which average down to 0.30.5 km s@xmath2 after combining all orders .
the nirspec data were forward - modeled using a custom mcmc implementation of the method described in @xcite .
the solar atlas of @xcite was used to model telluric absorption and the bt - settl atmosphere models @xcite were used to model the spectrum of wise j0720@xmath00846 .
spectral models with t@xmath58 = 23002500 k and @xmath59 = 5.5 ( cgs ) provided the best fits .
figure [ fig : nirspec ] shows that this procedure provides an accurate fit to the data , with residuals dominated by uncorrected fringing . radial velocity and rotational broadening
were among the parameters modeled , and the mean and standard deviation of their marginalized distributions in the mcmc chain were used as estimates of their measured values and uncertainties . for the radial velocities , uncertainties range over 0.30.7 km s@xmath2 .
tables [ tab : hamilton ] and [ tab : nirspec ] list all radial velocity measurements over the 11 epochs observed .
both datasets are consistent with a large radial motion for wise j0720@xmath00846 , with averages of + [email protected] km s@xmath2 from the hamilton data and + [email protected] km s@xmath2 from the nirspec data , ) between these mean values to a systematic shift between the instruments ; see @xcite . ] in clear contrast to its small tangential motion .
the total spatial velocity vector in the local standard of rest pointed radially away from the galactic center , @xmath19 in the direction of galactic rotation , and @xmath47 toward the galactic north pole .
we also adopt a local standard of rest motion of ( @xmath46,@xmath19,@xmath47)@xmath60 = ( 11.1 , 12.24 , 7.25)@xmath1(0.7 , 0.5 , 0.4 ) km s@xmath2 @xcite .
] is ( @xmath46,@xmath19,@xmath47 ) = ( @xmath047.5 , @xmath047.6 , 8.0)@xmath1(0.4 , 0.4 , 0.5 ) km s@xmath2 , which falls well outside the 1@xmath52 distribution of local late - type m dwarfs , ( @xmath61,@xmath62,@xmath63 ) = ( 32,20,17 ) km s@xmath2 @xcite .
its galactic orbit , computed as described in @xcite , is modestly eccentric , spanning galactic radii of 49 kpc ( assuming a solar radius of 8.5 kpc ) . using the criteria of @xcite , we assign this source to the old disk kinematic population , suggesting a `` mature '' age of 0.55 gyr @xcite .
h@xmath3 emission is detected in both the rc spec and hamilton spectrograph data . in the low - resolution data , we measure an equivalent width @xmath64 = @[email protected] . using a @xmath65 from @xcite , we determine @xmath5 = @xmath66 = @xmath48 , which is typical for m9 dwarfs in the vicinity of the sun @xcite .
figure [ fig : halpha ] displays our hamilton spectral data of wise j0720@xmath00846 around the 6563 h@xmath3 line .
continuum emission is weakly detected at red wavelengths in these data , but the h@xmath3 line in unambiguous , with both broad and highly variable emission .
equivalent widths ( absolute values ) range from @xmath671 to 13@xmath15 , corresponding to @xmath5 @xmath68 ( inactive ) to @xmath69 ( @xmath322@xmath13 `` normal '' activity ) .
these lines are also well - resolved , with fwhm of @xmath104060 km s@xmath2 , over 5 times the nirspec - derived rotational velocity ( see below ) .
the h@xmath3 line centers are nevertheless aligned with absorption line doppler shifts to within 5 km s@xmath2 , ruling out jets or accretion projected along the line of sight .
in addition to emission line variability , the trappist light curve shows several impulsive bursts on hjd 6662 , 6664 and 6667 , visible in figure [ fig : trappist ] and shown in detail in figure [ fig : flare ] .
these have relatively low power , with peak fluxes 48% above the local continuum .
three exhibit classic flare profiles , with impulsive rises ( @xmath12 1 min ) followed by exponential declines ( 25 min ) .
in contrast , the burst on hjd 6664.5859 , one of two in a 2 hr period , shows a more complex temporal structure , with a broad peak that persists for over 5 min followed by a 15 min decline with possible secondary bursts after 10 min .
this may reflect an associated chain of flare events over an extended region , or a massive flare rotating out of view .
these events indicate a 0.8% duty cycle of flare emission at the level of 2% above quiescence , somewhat below the typical duty cycles of late m dwarfs ( 3@xmath11% ; @xcite ) . hence , while wise j0720@xmath00846 exhibits numerous signs of activity , including an unusually broad h@xmath3 line , the strength of its persistent emission and frequency of flaring are at or below the average of equivalently classified dwarfs .
this is consistent with its old kinematic age and slow rotation rate ( see below ) , although it should be noted that age - activity and rotation - activity trends are not well established in the late m and l dwarf regime @xcite .
in addition to radial motion , our nirspec analysis provides multi - epoch measurements of rotational broadening . we measure a consistent @xmath9 = [email protected] km s@xmath2 , which is on the low end of , but consistent with , the rotational velocities of equivalently classified dwarfs @xcite . assuming a radius of 0.1 @xmath70 @xcite , this velocity corresponds to a maximum rotation period of 15@xmath11 hr , consistent with period measurements for equivalent - mass objects @xcite . despite being on the low end of the @xmath9 distribution of late - type m dwarfs ,
this source is still a rapid rotator , a likely explanation for its magnetic activity . to confirm the slow rotation of wise j0720@xmath00846 , we searched for rotational modulation of surface features in the trappist lightcurve .
as shown in figure [ fig : trappist ] , the red optical brightness of this source exhibits considerable night - to - night variation , particularly prior to hjd 6695 , the same period we detect flaring bursts . from hjd
6695 onward , wise j0720@xmath00846 is far more stable .
hereafter , we refer to these periods as the `` active '' and `` quiescent '' phases . for comparison ,
the relative fluxes of the nearby comparison star ( 0.06 mag fainter in the trappist bandpass ) are steady throughout the observing period ( figure [ fig : trappist_comp ] ) .
the nightly variations appear to be episodic , and we could find no clear period associated with them .
we therefore normalized each night s light curve to search for persistent low - level variability .
there is additional structure in the normalized lightcurves that arises from differential color extinction at large airmass @xcite .
wise j0720@xmath00846 has a distinct spectral energy distribution in the 0.75 - 1 @xmath16 trappist band compared to its neighboring sources , so color - dependent extinction is not corrected in relative photometry .
figure [ fig : airmass ] shows source flux as a function of airmass .
a statistically significant trend is present , reaching a deviation of 0.5% at an airmass of 2.0 .
in contrast , no significant trend is found in the comparison source .
this airmass effect was corrected in the lightcurve of wise j0720@xmath00846 by dividing by a second order polynomial fit to airmass ; we also rejected all measurements at airmasses greater than 2.0 to minimize residual bias .
this reduced our total number of samples to 5895 .
we found no significant trends with airmass or seeing after the airmass correction was applied .
the corrected and nightly - normalized light curve was analyzed using the phase dispersion minimization technique ( * ? ? ?
* ; * ? ? ?
our implementation of this method is described in appendix a. we examined periods between 4 hr and 16 hr , the lower limit consistent with the most rapidly rotating brown dwarfs and the upper limit chosen to be just above our rotational velocity limit .
figures [ fig : pdm ] and [ fig : pdm - comp ] display distributions of the pdm statistic @xmath71 as a function of period for wise j0720@xmath00846 and the comparison star , respectively .
the pdm of wise j0720@xmath00846 shows considerable structure , but none of the features exceed our significance threshold of 90% .
the phased light curve of the strongest period at [email protected] hr displays a compelling , double - peaked pattern with an amplitude of [email protected]% ( 2.6@xmath52 ) .
this period is consistent with a rotation axis inclination angle of 71@xmath37@xmath18@xmath37 .
however , the @xmath39 of the phased photometric residuals relative to the smoothed lightcurve is formally consistent with a lightcurve without phasing .
several other statistics were investigated , including lomb - scargle analysis , and trial sinusoidal fits to the data were also insufficient for identifying a robust period .
we therefore set a 3@xmath52 upper limit of 1.5% on periodic variability in this source , although larger episodic variations are clearly present .
note that the comparison source also exhibits no significant period . as a function of period for the flare - cleaned
, airmass - corrected , and daily - normalized trappist lightcurve of wise j0720@xmath00846 . a significance threshold of 90% based on @xcite
is indicated .
none of the features exceed this threshold .
( bottom ) : phased light curve for the strongest candidate period at [email protected] hr .
individual measurements are indicated by small circles , where different shades of red indicate different cycles ( 17 in total ) .
the black histogram with error bars ( scatter per phase bin ) delineates the mean phased light curve .
, title="fig:",scaledwidth=45.0% ] + as a function of period for the flare - cleaned , airmass - corrected , and daily - normalized trappist lightcurve of wise j0720@xmath00846 . a significance threshold of 90% based on @xcite
is indicated .
none of the features exceed this threshold .
( bottom ) : phased light curve for the strongest candidate period at [email protected] hr .
individual measurements are indicated by small circles , where different shades of red indicate different cycles ( 17 in total ) .
the black histogram with error bars ( scatter per phase bin ) delineates the mean phased light curve .
, title="fig:",scaledwidth=45.0% ] + but for the nearby comparison star 2mass j07200688@xmath00846504 .
the minima in @xmath71 reflect the sampling window of the measurement ( 12 hr and harmonics ) and are not significant.,title="fig:",scaledwidth=45.0% ] + but for the nearby comparison star 2mass j07200688@xmath00846504 .
the minima in @xmath71 reflect the sampling window of the measurement ( 12 hr and harmonics ) and are not significant.,title="fig:",scaledwidth=45.0% ] +
as described above , the spex spectrum of wise j0720@xmath00846 is best matched to that of the m9 spectral standard lhs 2924 .
however , close inspection reveals specific discrepancies near 1.3 @xmath16 and 1.6 @xmath16 , and a subtle `` dip '' feature at 1.62 @xmath16 .
such features have been previously noted in the combined - light spectra of very low mass spectral binary systems with late - type m or l dwarf primaries and t dwarf secondaries ( e.g. , @xcite ) . to assess whether wise j0720@xmath00846 is such a system
, we used the fitting method described in @xcite , comparing the spectrum of wise j0720@xmath00846 to 699 single star templates and 107,646 binary templates from the spl .
the latter were constructed using spectra of m7-l4 dwarfs for the primary and l9-t7 dwarfs for the secondary , with component spectra scaled to absolute magnitudes using the @xmath72/spectral type relation of @xcite .
figure [ fig : spex ] displays the best fitting binary template , which consists of lhs 2924 paired to the t5 2mass j04070885 + 1514565 @xcite .
this combination reproduces the excess flux at 1.3 @xmath16 , 1.6 @xmath16 and 2.1 @xmath16 , as well as the shape of the 1.62 @xmath16 dip feature , and is a statistically significant better match to the data based on an f - test comparison ( @xmath3299% confidence ) .
marginalizing over all fits weighted by the f - distribution , we infer component types of [email protected] and [email protected] , and @xmath7 = [email protected] , implying that the secondary contributes only 2% of the combined light in this band .
this system is similar in composition to the previously confirmed m8.5 + t5 spectral binaries 2mass j03202839@xmath00446358 ( hereafter 2mass j0320@xmath00446 ; @xcite ) and sdss [email protected] ( hereafter sdss j0006@xmath00852 ; @xcite ) .
we searched for this potential companion using our nirc2 @xmath26-band images .
given the somewhat poor observing conditions , the wings of the psf reach 2% relative brightness at a relatively large radius of 0@xmath86 ( 4.2 au ; figure [ fig : image ] ) .
only one point source is seen at larger radii , 4@xmath885 to the southwest of wise j0720@xmath00846 with @xmath73 .
however , a stationary optical and near - infrared counterpart to this source is seen in both digital sky survey and 2mass images , the former going back to 1955 , so it is likely an unrelated background star . to probe tighter separations ,
we self - subtracted the combined image after rotating by 180@xmath37 and offsetting to minimize the total squared deviation , effectively using the source psf as its own model .
figure [ fig : image ] shows that a faint source emerges upon this subtraction , 139@xmath114 mas from the psf center at a position angle of 262@xmath742@xmath37 , corresponding to a projected separation of [email protected] au at the [email protected] pc distance of this system .
the peak flux of this source relative to that of wise j0720@xmath00846 corresponds to a relative magnitude of [email protected] mag , consistent with that predicted from the spex spectral analysis , albeit with considerable uncertainty .
this source is therefore a promising candidate for the brown dwarf companion , although both its validity and physical association with wise j0720@xmath00846 must be confirmed .
we can constrain the presence of an unresolved companion from the radial velocity measurements obtained over 3 months with keck / nirspec and over 2 months with lick / hamilton spectrograph . in both sets of data
, we find the measured radial velocities are constant within the uncertainties . for the nirspec data , we measure a @xmath39 = 5.01 for 3 degrees of freedom , corresponding to a 17% false alarm probability ( marginal signficance ) ; for the hamilton data , we measure a @xmath39 = 5.21 for 6 degrees of freedom , corresponding to a 52% false alarm probability ( no significance ) .
the marginal significance in the nirspec data may be indicative of an additional @xmath120.5 km s@xmath2 systematic error not included in our forward modeling analysis ( see @xcite ) . using the simulations described in appendix b
, we ascertained the range of semi - major axes ( @xmath75 ) ruled out assuming no detection of radial velocity variability .
given the slight offset between the nirspec and hamilton measurements , we treated the datasets separately .
we first converted the component spectral types to t@xmath58s using the empirical relations of @xcite , and then converted these to masses using the evolutionary models of @xcite for ages 0.5 - 10 gyr ( table [ tab : rvsim ] ) .
the nirspec measurements rule out the presence of a companion with @xmath76 0.460.76 au ( 0.260.44 au ) at the 50% ( 80% ) confidence level , the range reflecting young to old ages ( figure [ fig : rvsim ] , table [ tab : rvsim ] ) .
the hamilton data , given its shorter time coverage , provides a less stringent constraint .
note that even the closest separations can not be completely ruled out given the possibilities of a face - on orbit , an eccentric orbit observed at apoapse , or poor synching of observations and orbital inflection points .
+ ccccccccc 0.5 & 0.066 & 0.027 & 0.41 & 0.26 & 0.46 & 0.771.2 & 2.24.3 & 1.22.4 + 1.0 & 0.077 & 0.036 & 0.47 & 0.30 & 0.53 & 0.781.3 & 2.03.9 & 1.42.9 + 5.0 & 0.084 & 0.065 & 0.77 & 0.41 & 0.72 & 0.771.2 & 1.83.4 & 2.24.6 + 10.0 & 0.084 & 0.073 & 0.87 & 0.44 & 0.76 & 0.861.4 & 1.73.3 & 2.45.0 + these limits are consistent with the location of the possible resolved companion .
we performed a second set of simulations that used both the radial velocity measurements and the observed separation to assess the distributions of probable semi - major axes , periods and primary radial velocity variability amplitudes .
these are constrained to be in the range 0.771.4 au , 1.74.3 yr and 1.25.0 km s@xmath2 , respectively .
both period and primary radial velocity amplitude are sensitive to the age and component masses of the system .
given the reasonably short time scale , maximum angular separation ( 0@xmath8130@xmath823 ) , and significant radial velocity perturbations , this system is an excellent target for individual component mass measurements in the near term .
the inferred properties of wise j0720@xmath00846 are summarized in table [ tab : properties ] .
based on the analysis described above , we conclude that wise j0720@xmath00846a is a relatively old , magnetically active , low - mass field star .
its degree of activity ( e.g. , flaring rate ) and its rotation frequency are both somewhat below those of other late - type m dwarfs , suggesting some long - term angular momentum loss may have occurred for this source .
the putative t dwarf companion wise j0720@xmath00846b must be substellar , and its time - dependent cooling permits a model - dependent constraint on the age of the system if the component masses can be determined .
we can already estimate that the mass ratio of this system @xmath77 m@xmath23/m@xmath78 is highly age - dependent , with 0.47 @xmath79 0.87 for 1 @xmath80 10 gyr . if the possible source detected in our nirc2 images is the companion , an inertial orbit measurement should be achieveable in 25 yr by combining astrometric and radial velocity monitoring .
note that a robust detection of photometric variability would allow us to measure the rotational axis inclination of wise j0720@xmath00846a and assess spin - orbit alignment in this system , a critical test of binary formation that has only been examined in one vlm binary to date .
it is worth noting the striking similarities between wise j0720@xmath00846ab and two other late - m plus t spectral binaries sdss j0006@xmath00852ab and 2mass j0320@xmath00446ab , both m8.5 + t5 systems .
the common classification of the secondary in all of these systems is likely a selection bias , as this subtype lies at the peak of the so - called `` @xmath15-band bump '' , the 1 @xmath16 brightening from late - l to mid - t likely caused by the depletion of photospheric clouds at the l dwarf / t dwarf transition @xcite .
t5 companions are simply more readily detectable in blended light spectra due to their brighter magnitudes .
all three systems also have separations @xmath671 au , below the @xmath1047 au peak of the separation distribution of resolved vlm binaries @xcite .
this agreement supports evidence that the spectral binary method is uncovering a significant population of tight binaries @xcite . finally , all three systems appear to be relatively mature .
sdss j0006@xmath00852ab is @xmath547 gyr based on the inactivity of its widely - separated m7 tertiary ; 2mass j0320@xmath00446ab is @xmath542 gyr based on the mass and evolutionary state of its companion ; wise j0720@xmath00846ab may be @xmath105 gyr based on its old disk kinematics . this congruence may reflect the known preference for resolved vlm binaries to have nearly - equal mass components ( @xcite ; however , see below ) . a brown dwarf closer to the hydrogen burning minimum mass ( hbmm ) must cool longer to reach a t@xmath58 @xmath18 1200 k consistent with a mid - t dwarf . from these few examples , we speculate that late - m dwarf plus t dwarf spectral binary systems specifically probe an old , tightly - bound population of vlm binaries , which are particularly useful systems for orbital mass measurements . wise j0720@xmath00846 joins the m8.5 scr 1845@xmath06357ab system as one of two late - m plus t dwarf binaries in the immediate vicinity of the sun . remarkably , these are the _ only _ binary systems among the 13 m7-m9.5 dwarf primaries known within 10 pc ( table [ tab : latem ] ) .
this fact is particularly surprising given the numerous efforts to identify faint companions to cool stars close to the sun @xcite , and the apparent preference for high mass - ratio systems among vlm binaries . as it stands , this volume - limited sample has a brown dwarf companion fraction , @xmath81 = 15@xmath82% , that is marginally higher ( but consistent with ) its stellar companion fraction limit @xmath83 @xmath1212% ( 1@xmath52 binomial uncertainties ) .
llcccl scr 1845 - 6357ab & m8.5+t6 & [email protected] & 10.58 & 8.89 & 1,10 + denis j1048 - 3956 & m8.5 & [email protected] & 10.44 & 8.88 & 1 + lsr j1835 + 3259 & m8.5 & [email protected] & 10.41 & 9.10 & 2 + * wise j0720@xmath00846ab * & m9.5+t5 & 166@xmath128 & 10.5 & & 3 , 4 + lp 944 - 20 & m9 & [email protected] & 10.51 & 9.15 & 5 + gj 3877 &
m7 & 152@xmath12 & 9.84 & 8.02 & 5 , 7 + scr 1546 - 5534 & m8 : & 149@xmath140 & 10.0 & & 6 + sips 1259 - 4336 & m8 : & @xmath10128 & @xmath1010.1 & & 9 , 11 + lehpm 1 - 3396 & m9 & 121@xmath14 & 10.80 & & 12 , 13 + lhs 2065 & m9 & [email protected] & 10.30 & 9.00 & 5 , 7 + 1rxs j1159 - 5247 & m9 & [email protected] & 10.44 & & 14 , 15 + lp 655 - 48 & m7 & 106@xmath13 & 9.66 & 8.31 & 1 , 16 + lp 647 - 13 & m9 & 104@xmath12 & 10.52 & 8.84 & 8 + is the nature of multiples in the local late - m dwarf population simply due to small number statistics ? to examine this question , we performed a population simulation similar to those described in @xcite , combined with a random draw experiment to determine the likelihood of various binary configurations ; see appendix c for details . from our simulations
we find that t dwarfs are remarkably common companions to late m dwarfs , comprising @xmath1025% of secondaries ( figure [ fig : binfrac ] ) .
their relative abundance derives from two factors .
first , the typical masses of m7-m9.5 dwarfs in our simulation , [email protected] m@xmath55 , are close to the hbmm , so stellar companions exist over a relatively narrow range of masses .
brown dwarf companions to late - m primaries consistently outnumber stellar companions in our simulation ; in contrast , only @xmath109% of systems with mid - m ( m4m7 ) primaries host t dwarf companions .
second , there is the well - known `` pile - up '' of t dwarfs in field brown dwarf populations due to their slow cooling rates @xcite .
t dwarfs are the most common companion to the oldest systems in our simulation ( @xmath325 gyr ) .
this is consistent with the old ages inferred for sdss j0006@xmath00852ab , 2mass j0320@xmath00446ab and wise j0720@xmath00846ab .
even with the prevalence of t dwarf companions to late - m primaries , the incidence of finding 2 or more t dwarf companions in the nearby sample and only t dwarf companions remains low , plateauing around 1.4% for a total binary fraction 10% @xmath12 @xmath84 @xmath12 25% . the local sample would seem to be an anomaly if the underlying multiplicity distributions are accurately characterized .
it is possible that more ( and less ) massive companions to other late m dwarfs in the nearby sample have yet to be found , and may be uncovered astrometrically with the current gaia mission .
we also note that while this specific configuration is rare , there are roughly 10@xmath85 10-pc `` bubbles '' in the 312 kpc galactic disk , making our sample of m dwarfs one of a million .
finally , we note that the peak probability of exclusively t dwarf companions to late m dwarfs occurs in a range of @xmath84 consistent with previous determinations of the vlm binary fraction @xcite .
this is significant , as the decline in frequency at high fractions is caused by increased contamination by earlier - type secondaries .
the lack of such companions in the nearby sample may be a consequence of the low binary fraction of vlm dwarfs compared to more massive stars .
the companionship of late - m dwarfs appears to be determined by the confluence of four key statistics of vlm dwarfs : the mass function across the substellar limit , the cooling rate of brown dwarfs , the mass ratio distribution of vlm dwarfs , and the underlying multiplicity fraction .
the importance of these statistics for brown dwarf formation and interior theories motivates a complete assessment of the multiplicity properties of late m dwarfs in a much larger volume .
we have conducted a detailed investigation of the recently identified , nearby vlm dwarf wise j0720@xmath00846 . from optical and infrared imaging and spectroscopic investigations , we have determined the optical and near - infrared classifications of this source , improved its astrometry , measured its spatial kinematics and rotational velocity , identified persistent and flaring magnetic activity , and found evidence for a brown dwarf companion .
the reduced level of activity in this source compared to other late - m dwarfs , and its old disk kinematics , both suggest that this is a relatively mature system .
the putative t dwarf companion , identified in combined - light spectroscopy and possibly resolved at 1 au projected separation , will aid in constraining the age , as astrometric and radial velocity monitoring over the next few years should allow us to map the orbit of the system and extract individual component masses .
remarkably , this is one of only two binaries among late m dwarfs in the immediate vicinity of the sun , and both have t dwarf companions .
we argue that , while rare , this may reflect a combination of the proximity of the primary to the hbmm , the evolutionary properties of brown dwarfs , and the underlying binary fraction , making multiplicity studies of late m dwarfs a potentially useful window into brown dwarf formation and evolution theories .
the authors thank bill golisch and john rayner at irtf ; wayne earthman , erik kovacs , donnie redel and pavl zachary at lick observatory ; diane harmer and krissy reetz at kpno ; scott dahm , greg doppmann , heather hershley , gary punawai , luca rizzi , and terry stickel at keck for their assistance with the observations .
we also acknowledge useful discussions with gregg hallinan and stuart p. littlefair on m dwarf magnetic activity ; and john gizis , todd henry , j. davy kirkpatrick and i. neill reid on the 10 pc sample .
we thank our referee , r. scholz , for his very helpful comments that allowed us to considerably improve the manuscript .
this research has made use of the simbad database , operated at cds , strasbourg , france ; the m , l , t , and y dwarf compendium housed at dwarfarchives.org ; and the spex prism spectral libraries at http://www.browndwarfs.org/spexprism .
c.m . acknowledges support from the national science foundation under award no .
trappist is a project funded by the belgian fund for scientific research ( f.r.s .- fnrs ) under grant frfc 2.5.594.09.f , with the participation of the swiss national science foundation .
m. gillon and e. jehin are f.r.s .- fnrs research associates .
l. delrez and j. manfroid acknowledge the support of the f.r.s .-
fnrs for their phd theses .
the authors wish to recognize and acknowledge the very significant cultural role and reverence that the summit of mauna kea has always had within the indigenous hawaiian community .
we are most fortunate to have the opportunity to conduct observations from this mountain .
the phase dispersion minimization ( pdm ) technique identifies periodic signals in time - series data by searching for minimal dispersion about a mean phased signal @xcite .
the frequency sampling is set by the fundamental frequency @xmath86 @xmath87 1/@xmath88 = 0.021 cycles day@xmath2 , where @xmath88 = 1160 hr is the full monitoring period @xcite .
we examined 1081 periods in the period range 4 hr @xmath89 16 hr , sampled evenly in frequency space in steps of @xmath90 . for each period , we phase - folded the lightcurve , computed a mean curve sampled at @xmath91 = 100 linearly - spaced phase points , then computed the @xmath39 deviation of the phased data relative to the mean curve .
the statistic of merit is the ratio @xmath92 where @xmath93 is measured from the unphased lightcurve , and @xmath94 = 5895 is the total number of data points .
following @xcite , we assessed the significance of minima in @xmath95 using the regularized incomplete beta function @xmath30 : @xmath96^{1/n_f}\ ] ] here , pr@xmath97 is the false - alarm probability when @xmath71 is less than some value 0 @xmath98 1 by chance ; 1-pr gives the significance of a period corresponding to a minimum in @xmath71 .
the 1/@xmath99 exponent is the band penalty incurred when searching for multiple periods , where we have used @xmath100 as the number of independent frequencies sampled @xcite .
periods with 1-pr @xmath32 90% were deemed significant , but no features in the cleaned and normalized light curve satisfied this limit .
to quantify the companion detection limits from our radial velocity observations , and determine whether the candidate companion identified in nirc2 imaging data is consistent with these limits , we performed a pair of monte carlo orbit simulations .
the first simulation aimed to determine the range of orbital semimajor axes that would have been detected given the radial velocity measurement uncertainties and sampling .
we generated a large number ( 10@xmath101 ) of hypothetical orbits , uniformly sampling semi - major axes 0 @xmath102 2.5 au , eccentricities 0 @xmath103 0.6 @xcite , inclinations 0 @xmath104 @xmath41 , longitude of ascending node 0 @xmath105 2@xmath41 , argument of periapse 0 @xmath106 2@xmath41 , and mean anomaly angle 0 @xmath107 2@xmath41 for the nirc2 imaging epoch .
we then solved for the maximum radial velocity amplitude of the primary component ( e.g. , @xcite ) : @xmath108 with component masses m@xmath78 and m@xmath23 estimated from the evolutionary models of @xcite , system ages of 0.5 , 1 , 5 and 10 gyr , and component t@xmath58s of 2300 k and 1100 k for wise j0720@xmath00846a and b based on the t@xmath58/spectral type relation of @xcite ( table [ tab : rvsim ] ) .
the observed radial motion of the primary at each nirspec or hamilton epoch @xmath109 ( these datasets were modeled separately ) was then calculated as @xmath110 where @xmath111 relates the true anomaly @xmath88 to the eccentric anomaly @xmath112 , which is in turn related to the mean anomaly through kepler s equation : @xmath113 here , @xmath114 is the period of the orbit , determined from the estimated masses , input semimajor axis , and kepler s period law ; and @xmath115 is the time since periastron passage . setting @xmath116 = 0 and @xmath117 = 0 ( the time of the nirc2 image ) , we can set @xmath118 and solve for the primary velocities numerically . for each simulated orbit , we calculated @xmath39 for the radial velocity epochs relative to their mean , using the corresponding observational uncertainties .
we then determined the fraction of orbits that exceeded our measured @xmath39 as a function of semi - major axis .
figure [ fig : rvsim ] displays the results of this calculation for the nirspec data and a system age of 1 gyr . due to inclination variations and sampling , detection probabilities never reach 100% ; we therefore use 50% and 80% probabilities of detection as our thresholds . to add in the constraint of a possible detection of the companion at a projected separation of @xmath119
, we used the same orbital parameters to assess the range of semi - major axes consistent with this separation .
cartesian positions in the plane of the sky for an orbit of unit semi - major axis are @xmath120 were @xmath121 , @xmath122 , @xmath123 and @xmath124 are the thiele - innes constants @xcite : @xmath125 the total projected separation @xmath126 can then be compared to the observed projected separation ( @xmath51d ) to constrain the semimajor axis for a given orbit @xmath127 : @xmath128 figure [ fig : rvsim ] displays the distributions of semi - major axes and primary radial velocity variability amplitudes consistent both with the possible detection and lack of detectable radial velocity variation .
we performed another monte carlo simulation to assess the likelihood of finding two or more t dwarf companions ( exclusively ) in the sample of 13 m7m9.5 dwarfs within 10 pc of the sun ( table [ tab : latem ] ) .
we started with a sample of @xmath94 = 10@xmath38 primaries with uniform ages spanning 0.57 gyr ( corresponding to the typical ages of disk stars ) and masses 0.013 m@xmath55 @xmath129 m @xmath129 0.2 m@xmath55 distributed as a power - law mass function @xmath130 @xcite .
for each system we assigned a secondary mass using the power - law mass ratio distribution of @xcite , @xmath131 , based on bayesian analysis of vlm imaging samples ; this distribution favors equal - mass systems .
the component masses and system ages were transformed to bolometric luminosities using the evolutionary models of @xcite , and these converted into spectral types by combining the spectral type / absolute @xmath15-band magnitude relation of @xcite with the spectral type/@xmath15-band bolometric correction relation of @xcite .
systems with component spectral types outside the range m5-t9 were rejected .
we selected a subset of m7-m9.5 sources ( 4626 systems ) and used these as the primaries for our random - draw experiment .
the distribution of companion types for this sample is shown in figure [ fig : binfrac ] . flagging a randomly - assigned subset @xmath132 of these systems as actual binaries , with @xmath133 [ 0.05,0.5 ]
, we repeatedly ( 10@xmath38 times ) drew 13 systems from this collection and determined the fraction that contained at least two companions with spectral types t2t7 ( detectable by the spectral binary method ; @xcite ) and no other companion types .
figure [ fig : binfrac ] displays the frequency of this outcome as a function of overall binary fraction . , f. , homeier , d. , & freytag , b. 2011 , in astronomical society of the pacific conference series , vol .
448 , 16th cambridge workshop on cool stars , stellar systems , and the sun , ed . c. johns - krull , m. k. browning , & a. a. west , 91 , i. s. , graham , j. r. , becklin , e. e. , et al .
2000 , in society of photo - optical instrumentation engineers ( spie ) conference series , vol . 4008 , society of photo - optical instrumentation engineers ( spie ) conference series , ed . m. iye & a. f. moorwood , 10481055 , d. 1986 , in society of photo - optical instrumentation engineers ( spie ) conference series , vol .
627 , society of photo - optical instrumentation engineers ( spie ) conference series , ed .
d. l. crawford , 733+ | we report observations of the recently discovered , nearby late - m dwarf wise [email protected] .
new astrometric measurements obtained with the trappist telescope improve the distance measurement to [email protected] pc and confirm the low tangential velocity ( [email protected] km s@xmath2 ) reported by scholz .
low - resolution optical spectroscopy indicates a spectral type of m9.5 and prominent h@xmath3 emission ( @xmath4@xmath5@xmath6 = @[email protected] ) , but no evidence of subsolar metallicity or li i absorption . near - infrared spectroscopy reveals subtle peculiarities that can be explained by the presence of a t5 binary companion , and high - resolution laser guide star adaptive optics imaging reveals a faint ( @xmath7 = 4.1 ) candidate source 0@xmath814 ( 0.8 au ) from the primary . with high - resolution optical and near - infrared spectroscopy
, we measure a stable radial velocity of + [email protected] km s@xmath2 , indicative of old disk kinematics and consistent with the angular separation of the possible companion .
we measure a projected rotational velocity of @xmath9 = [email protected] km s@xmath2 and find evidence of low - level variabilty ( @xmath101.5% ) in a 13-day trappist lightcurve , but can not robustly constrain the rotational period .
we also observe episodic changes in brightness ( 1 - 2% ) and occasional flare bursts ( 48% ) with a 0.8% duty cycle , and order - of - magnitude variations in h@xmath3 line strength .
combined , these observations reveal wise j0720@xmath00846 to be an old , very low - mass binary whose components straddle the hydrogen burning minimum mass , and whose primary is a relatively rapid rotator and magnetically active .
it is one of only two known binaries among late m dwarfs within 10 pc of the sun , both of which harbor a mid t - type brown dwarf companion .
we show that while this specific configuration is rare ( 1.4% probability ) , roughly 25% of binary companions to late - type m dwarfs in the local population are likely low - temperature t or y brown dwarfs . |
the entropy power inequality ( epi ) for statistically independent real _ continuous_-valued random variables @xmath0 and @xmath3 , @xmath6 was first stated by shannon @xcite , where @xmath7 is the _ entropy power _ of @xmath0 , with @xmath8 being the differential entropy of @xmath0 .
equality in holds if and only if @xmath0 and @xmath3 are gaussian . using @xmath9 the entropy of a gaussian random variable with variance @xmath10 , we get @xmath11 . in other words ,
the entropy power of @xmath0 is the variance of a gaussian random variable that has the same differential entropy as that of @xmath0 .
the epi was subsequently proved by stam @xcite and blachman @xcite .
the epi has proven to be a powerful tool in information theory .
it has found use in converse proofs of various gaussian channel coding theorems involving continuous - valued sources and channels , especially in the respective converse proofs .
some prominent examples are : the capacity region of the scalar gaussian broadcast channel @xcite , its generalization to the mimo case @xcite , the private capacity of the gaussian wiretap channel @xcite , capacity region of the gaussian interference channel @xcite , gaussian source multiple - description problem @xcite , and rate distortion region for multi - terminal gaussian source coding @xcite . due to this success of shannon s epi ,
it is natural to speculate if there is an epi for statistically - independent discrete - valued random variables @xmath0 and @xmath3 defined on the set of natural numbers @xmath12 , which has analogous implications as above in discrete - valued information and coding theorems .
there have been several attempts to provide an epi for discrete random variables .
it is simple to see that does not hold as is when @xmath0 and @xmath3 are arbitrary discrete random variables and the differential entropy @xmath13 is replaced by the discrete shannon entropy @xmath1 ; a simple counterexample occurring for when @xmath0 and @xmath3 are deterministic .
epis with bernoulli sources were proven where the @xmath14 " operation in was taken modulo @xmath15
harremo " es and vignat retained the use of @xmath14 " as integer addition , and proved that holds when @xmath0 and @xmath3 are independent binomial @xmath16 and @xmath17 variables , upon redefining @xmath18 in as @xmath19 , with @xmath1 being the discrete entropy @xcite .
a discrete epi was proven for arbitrary i.i.d .
integer - valued random variables , albeit of a non - standard asymmetric form @xcite .
yu and johnson developed a natural form of the discrete epi but one that holds only for ultra - log concave ( ulc ) distributed random variables @xcite .
recently , woo and madiman showed that an inequality very similar to holds for uniform distributions over finite subsets of the integers @xcite . in this paper , we present an axiomatic framework to develop a natural generalization of the epi for arbitrary discrete random variables and an associated central limit theorem ( clt ) .
we begin with an overview of shannon s epi in section [ sec : review ] .
section [ sec : discreteepi ] provides an intuitive development of our discrete epi . in section
[ sec : quantum ] , we provide proofs leveraging the algebra of quantum optics , and a close connection of our discrete epi to a quantum epi that naturally emerges in capacity proofs for transmitting classical information over bosonic channels @xcite .
using the following scaling identity for entropy power : @xmath20 it is straightforward to see that ineq . can be restated as : @xmath21 @xmath22 is a scaled addition operation .
the ` linear ' form of the epi , @xmath23 i.e. , the fact that differential entropy is concave with respect to normalized linear combinations , was first stated by lieb @xcite , a rigorous proof of which , and the proof of the fact that is equivalent to , was later provided by dembo , cover and thomas @xcite . that , hence implies follows simply from the application of @xmath24 to both sides of , and using the concavity of the logarithm function .
the converse implication requires the entropy of a scaled random variable , @xmath25 @xcite .
consider random variables @xmath26 and @xmath27 , with @xmath28 .
using the expression for @xmath29 stated above , it readily follows that @xmath30 . we then have , using , @xmath31 applying @xmath32 to both sides , and using , yields the epi .
scaling plays an important role in many proofs of the epi .
the fact that the epi can be stated in terms of scaled random variables as in inequality was implicit in verd ' u and guo s proof of the epi @xcite , but johnson and yu later realized the significance of this form to construct an epi for ulc discrete random variables , as we will describe in section [ sec : discreteepi ] @xcite .
artstein , ball , barthe and naor proved a stronger form of the epi for sums of independent continuous random variables @xcite , a special case of which was the first rigorous proof of _ monotonicity _ of the convergence of differential entropy in the clt , i.e. , for i.i.d .
@xmath33 , the entropy of the normalized sum @xmath34 , with @xmath35 , is monotone increasing in @xmath36 , and converges to the entropy of the gaussian ( which has the maximum entropy for a given variance ) as @xmath37 .
note that @xmath38 for i.i.d . @xmath39 and @xmath40 follows from , repeated application of which shows that @xmath41 is nondecreasing in @xmath42 .
this cruder version of monotonicity is already sufficient to prove the clt @xcite .
alternative proofs of monotonicity of @xmath43 were later given by tulino and verd ' u @xcite , and by madiman and barron @xcite .
the main ingredient needed for a natural discrete - variable generalization of shannon s epi , i.e. , and is a : * * scaled addition * : an appropriate definition of @xmath2 for @xmath0 and @xmath3 both defined on @xmath44 , and @xmath4 .
further , the definition of @xmath45 should be extendable to a random vector @xmath46 , i.e. , @xmath47 , where @xmath48 with @xmath49 and @xmath50 , such that : ( 1 ) it reduces to the bivariate case for @xmath51 , viz .
, @xmath52 ; ( 2 ) it is commutative in the sense that @xmath47 is invariant under an arbitrary ( yet identical ) permutation of the entries of @xmath53 and @xmath54 ; and ( 3 ) it is well behaved under a clt : * * _ limiting _ distribution * : a distribution @xmath55 $ ] , @xmath56 should exist that can be defined solely as a function of its mean @xmath57 , and is the limiting distribution in a clt under @xmath58 addition . in other words ,
@xmath55 $ ] is the distribution of @xmath59 , as @xmath37 , for i.i.d .
arbitrarily - distributed @xmath57-mean random variables @xmath60 . for a complete analogy with shannon s epi
, one would want @xmath61 , the entropy of @xmath62 , to be monotonically increasing in @xmath63 .
thus , @xmath64 should be the distribution with the maximum entropy for a given mean @xmath57 .
we already know that the geometric distribution @xmath65 = \left(1+\lambda\right)^{-1}\left(\lambda/(1+\lambda)\right)^k$ ] , @xmath56 , has this property .
so , we would like the @xmath58 operation for which the above clt holds with the geometric distribution being the limiting distribution .
one would then define : * * entropy power * : @xmath66 of @xmath0 as the mean of a random variable with distribution @xmath67 that has entropy @xmath1 .
note that in order for the above to make sense , the entropy of the limiting distribution @xmath67 should be monotonic ( increasing ) in its mean @xmath57 . with the above
, the following should hold : * * entropy power inequality * : the epi for discrete random variables , i.e. , and should hold with the shannon entropy power @xmath18 replaced by the discrete entropy power @xmath66 and the differential entropy @xmath13 replaced by the discrete shannon entropy @xmath1 .
further , equality in both aforesaid forms of the discrete epi should hold when @xmath0 and @xmath3 both are distributed according to the limiting distribution @xmath67 ( possibly with different means ) .
the above discussion suggests that once we have the ` correct ' definition of the @xmath58 operation for discrete random variables ( i.e. , one that is well behaved under the clt with the limiting distribution being geometric ) , one would immediately obtain the natural discrete generalization of shannon s epi . in the above framework for discrete - valued random variables , we chose to peg the definitions to the mean as opposed to the variance ( as in the case of continuous random variables ) . in the discrete case , the law of small numbers ( see footnote [ footnote : losn ] ) and the corresponding maximum entropy property both require ` thinning ' the mean , whereas in the continuous case
, the central limit theorem requires the thinning of the variance , which is achieved by multiplication by @xmath68 .
many have realized that the r ' enyi thinning operation @xmath69 is the natural discrete - variable equivalent of multiplication by @xmath68 , and has the desirable effect of thinning the mean by a factor @xmath70 @xcite . given @xmath4 , and @xmath71 , the random variable @xmath72 , obtained by @xmath70-thinning of @xmath0 ( denoted , @xmath73 ) , has the distribution of the sum @xmath74 , where @xmath75 are binary @xmath76 valued i.i.d .
bernoulli(@xmath77 ) random variables that are independent of @xmath0 . the p.m.f .
of @xmath3 is : @xmath78 = \sum_{k = n}^\infty p_x[k ] \binom{k}{n}\eta^n(1-\eta)^{k - n}$ ] .
yu and johnson developed a promising line of approach to the discrete epi for ultra - log concave ( ulc ) random variables ( i.e. , those with p.m.f.s @xmath79 , n \in { \mathbb n}_0 $ ] , for which @xmath80/p_x[n-1]$ ] is decreasing as @xmath36 increases ) .
they defined the scaled addition @xmath81 which extends naturally to multiple variables and is well behaved under the clt with the associated limiting distribution being poisson . for i.i.d .
mean-@xmath57 @xmath82 with p.m.f .
@xmath79 $ ] , @xmath83 converges to the poisson distribution of mean @xmath57 as @xmath37 the distribution of @xmath84 is the @xmath36-fold convolution of @xmath79 $ ] .
when @xmath85 bernoulli(@xmath86 ) , @xmath87 binomial(@xmath88 ) and @xmath89 binomial(@xmath90 ) @xmath91 poisson(@xmath86 ) as @xmath37 .
this special case is the classical binomial - to - poisson convergence , known as the law of small numbers " @xcite . ] .
since the limiting distribution is not geometric , the @xmath45 operation in does not satisfy the criteria in section [ sec : axioms ] .
however , within the class of all ulc random variables of mean @xmath57 , the poisson(@xmath57 ) random variable maximizes the entropy @xcite .
furthermore , @xmath61 increases monotonically in @xmath36 if @xmath79 $ ] is ulc @xcite . motivated by this , yu and johnson defined entropy power @xmath92 in terms of the entropy @xmath93 of a poisson(@xmath57 ) random variable , with the hope that the straightforward equivalents , and would hold , with @xmath0 and @xmath3 restricted to ulc random variables .
they proved that the linear form ( concavity of entropy ) holds @xcite , i.e. , @xmath94 for all ulc independent @xmath0 and @xmath3 .
this was the first major step towards a discrete epi .
the equivalents of and , @xmath95 were naturally conjectured @xcite , but were shown later _ not _ to hold in general , even for ulc @xmath0 and @xmath3 @xcite .
the key step in going from the epi to the linear form was the scaling identity for the entropy power .
if such an identity were to hold for any ulc variable @xmath0 , i.e. , @xmath96 , then would imply the conjectured discrete epi .
but since the conjecture was found to be false , the scaling identity above can not be true .
however , the following one - sided version of it was proved for ulc @xmath0 : @xmath97 which was termed the restricted discrete epi @xcite .
even though does not hold in general , the restated form does hold for discrete ulc variables @xcite : given ulc independent @xmath0 and @xmath3 , if @xmath98 @xmath99 and @xmath100 s.t .
@xmath101 and @xmath102 for some @xmath4 , s.t .
@xmath103 , then @xmath104 with equality iff @xmath0 and @xmath3 are poisson .
the reason why holds but not , is that finding @xmath70 , @xmath99 and @xmath100 that satisfy the aforesaid conditions is not always possible .
this is unlike the continuous version , for which @xmath70 , @xmath99 and @xmath100 can always be constructed that satisfy the conditions , as shown in section [ sec : review ] . and @xmath105 combined on a beamsplitter with transmissivity @xmath70 results in a number - diagonal state @xmath106 at the output .
the distribution of the photon number at the output is a particular scaled addition of the two input distributions , explained in section [ sec : discreteepi_gsg ] , which is at the core of our discrete epi . ]
our goal is to develop a @xmath45 operation that : ( i ) is well behaved under a clt with the geometric distribution as the limiting distribution ; and ( ii ) satisfies ineqs . and for @xmath0 and @xmath3 with _
arbitrary _ discrete distributions , and with the entropy power @xmath107 defined in terms of @xmath108 , the entropy of the geometric distribution with mean @xmath57 .
let us consider the following @xmath109-to-@xmath109 transform between a discrete distribution @xmath79 $ ] , @xmath110 and a circularly - symmetric continuous distribution @xmath111 , @xmath112 ( proof in section [ sec : quantum ] ) : @xmath113 \frac{e^{-\left| { \boldsymbol r } \right|^2 } \left| { \boldsymbol r } \right|^{2n}}{n ! } , { \text{and}}\label{eq : t}\\ p_x[n ] & = & \frac{1}{\pi}\int p_{{\boldsymbol x}_c}(\boldsymbol r ) { \cal l}_n\left(\left| { \boldsymbol s } \right|^2\right ) e^{{\boldsymbol r}{\boldsymbol s}^ * - { \boldsymbol r}^*{\boldsymbol s}}d^2{\boldsymbol r}d^2{\boldsymbol s } , \label{eq : inverset}\end{aligned}\ ] ] where @xmath114 is the @xmath36-th laguerre polynomial .
let us denote @xmath115 and @xmath116 is not a function that maps the random variable @xmath0 to the random variable @xmath117 .
it is a transform that takes a _ function _ , a discrete p.m.f .
@xmath79 $ ] , to another function , a continuous p.d.f . @xmath111 . ] .
we define scaled addition @xmath45 for discrete random variables @xmath0 and @xmath3 as follows , but in section [ sec : discreteepi_gsg ] onwards , the definition in eq . will be implied for @xmath45 . ] : @xmath118 where the @xmath45 on the rhs of is the usual continuous - variable definition as in .
the following is simple to verify : if @xmath78 = \delta[n]$ ] is a delta function at @xmath119 , @xmath120 .
similarly , if @xmath79 = \delta[n]$ ] , @xmath121 .
[ thm : thinning_pureloss ] our definition of @xmath45 coincides with the yu - johnson definition when one of the two variables is @xmath122 with probability @xmath109 , but instead of directly adding the integer - valued scaled random variables @xmath123 and @xmath124 , we add the real - valued scaled random variables @xmath125 and @xmath126 ( and transform back to the integer domain using ) , where @xmath115 and @xmath127 .
generalizing our @xmath45 operation for multiple variables is straightforward . following the steps outlined in section [ sec : axioms ]
, we consider a clt under this scaled addition . for i.i.d .
arbitrarily - distributed @xmath57-mean random variables @xmath60 , the p.m.f . of @xmath59 converges to the geometric distribution of mean @xmath57 as @xmath37 .
[ thm : clt_gsg ] we have the following conjecture on monotonicity of entropy : @xmath61 increases monotonically with @xmath63 and converges to @xmath128 , as @xmath37 .
[ conj : monotonicity ] this is analogous to harremo " es _ et al .
_ s law of small numbers @xcite and johnson - yu s discrete entropic - monotonicity result @xcite , but no longer restricted to ulc random variables .
[ thm : epi_gsg_linear ] the linear form of the discrete epi holds for arbitrary independent discrete - valued random variables @xmath0 , @xmath3 : @xmath129 we will show in section [ sec : quantum ] that follows as a special case of a recent result on a quantum version of the epi @xcite .
inequality with @xmath130 implies @xmath131 , which is sufficient to prove theorem [ thm : clt_gsg ] but only proves conjecture [ conj : monotonicity ] for @xmath36 increasing in power - of-@xmath15 steps .
[ conj : epi_gsg_scaled ] the following is the natural discrete generalization of shannon s epi , which holds true for arbitrary independent discrete - valued random variables @xmath0 and @xmath3 : @xmath132 even though we do not provide a proof of , we will show in section [ sec : quantum ] that is a simple special case of the yet - unproven entropy photon number inequality ( epni ) @xcite .
de palma _ et al .
_ recently proved the following restricted one - sided version of , analogous to yu and johnson s ineq .
: @xmath133 see theorem @xmath134 of ref .
@xcite for a proof .
there has been a suite of recent progress in quantum versions of the epi @xcite , and an eventual proof of the epni will imply the validity of the discrete epi in conjecture [ conj : epi_gsg_scaled ] .
finally , we have : using a ( somewhat unnatural ) definition of entropy power , @xmath135in analogy with the continuous entropy power @xmath18the epi statement in holds true for arbitrary independent discrete random variables @xmath0 , @xmath3 : @xmath136 we show in section [ sec : quantum ] that this result follows as a special case of a quantum epi result recently proved in @xcite .
in this section , we will provide proofs of the various statements in section [ sec : discreteepi_gsg ] . even though it is possible to prove them directly ,
it is much easier to leverage the mathematics of quantum optics .
let us consider a beamsplitter of transmissivity @xmath4 that mixes two modes whose annihilation operators are @xmath137 and @xmath138 , to produce an output mode @xmath139 .
assume that the quantum states of the two input modes , @xmath140 are statistically independent .
the density operators @xmath141 and @xmath105 are infinite - dimensional , unit - trace , positive , hermitian matrices , whose matrix elements we will express in the complete orthonormal _ fock _ ( or photon - number ) basis @xmath142 , @xmath143 .
the von neumann entropy of @xmath141 , @xmath144 , where @xmath145 are the eigenvalues of @xmath141 . for a number diagonal state
@xmath146|n\rangle \langle n|$ ] , @xmath147 , shannon entropy of the discrete random variable @xmath0 .
consider independent @xmath0 and @xmath3 with p.m.f.s @xmath79 $ ] and @xmath78 $ ] , @xmath110 .
consider mixing independent number - diagonal states @xmath146|n\rangle \langle n|$ ] and @xmath148|n\rangle \langle n|$ ] on a beamsplitter of transmissivity @xmath70 .
then the output state is also number - diagonal , i.e. , @xmath149|n\rangle \langle n|$ ] , and the output number distribution @xmath150 $ ] is that of the random variable @xmath151 , with our definition of @xmath45 given in . in other words , the scaled addition in eq .
has a physical interpretation
it is how a beamsplitter of transmissivity @xmath70 ` adds ' photon number distributions of two independent number - diagonal input states .
the husimi function of a quantum state @xmath141 , @xmath152 , @xmath112 , can be interpreted as the p.d.f . of a continuous - valued random variable @xmath153 . here , @xmath154 is the _ coherent state _ of complex amplitude @xmath155 , an eigenstate of the annihilation operator @xmath156 .
when @xmath146|n\rangle \langle n|$ ] is number - diagonal , its husimi function @xmath111 is given by eq .
, which is a circularly - symmetric function in the phase space .
the anti - normally - ordered characteristic function of @xmath157 is then @xmath158 . using the operator fourier inverse to express the state @xmath159 , writing down the number - basis diagonal elements @xmath160 $ ] , and using the identity @xmath161
, we obtain eq . .
so , we have the @xmath109-to-@xmath109 transform , @xmath115 and @xmath116 . physically , @xmath0 and @xmath162 are random variables corresponding to the outcomes of ( ideal ) photon number measurement and ( ideal ) optical heterodyne detection measurement respectively , on the state @xmath141 .
because the husimi function of a quantum state is unique , the above transform relation implies that the husimi function of a state is circularly symmetric if and only if it is diagonal in the number basis .
next , we observe that the husimi function of the output state @xmath106 is given by the scaled convolution @xmath163 , which implies that the respective transformed variables are related by the scaled addition , @xmath164 @xcite .
given @xmath141 and @xmath105 are both number diagonal ( and hence have circularly - symmetric husimi functions ) , @xmath106 must also have a circularly - symmetric husimi function , and hence be number diagonal .
therefore , the number distribution of @xmath106 is that of a random variable @xmath165 .
it is simple to verify that , for number - diagonal @xmath141 , if @xmath78 = \delta[0]$ ] , the output state s number distribution @xmath150 $ ] is the @xmath70-thinned version of @xmath79 $ ] , @xmath166 .
similarly , with @xmath79 = \delta[0]$ ] and number - diagonal @xmath105 , @xmath167 .
interestingly however , if neither @xmath141 nor @xmath105 is the vacuum state , the output number distribution is _ not _ a simple addition of the thinned input distributions , i.e. , @xmath168 ( the rhs is yu and johnson s @xmath45 ) ; rather it is given by .
proof of the discrete clt in theorem [ thm : clt_gsg ] follows from theorem 5.10 of @xcite when restricted to number - diagonal inputs and using the definition of @xmath45 in eq . .
let us consider the following two definitions of entropy power for a quantum state , and associated entropy power inequalities for each definition : \(a ) _ entropy photon number _
, @xmath169 , where @xmath170 is the von neumann entropy of a thermal state @xmath171|n\rangle \langle n|$ ] of mean photon number @xmath57 , with @xmath172 = \left(1+\lambda\right)^{-1}\left(\lambda/(1+\lambda)\right)^n$ ] . since @xmath173 is diagonal in the number basis , @xmath128 is also the shannon entropy of the geometric distribution of mean @xmath57 , as used in section [ sec : discreteepi_gsg ] .
\(b ) _ quantum entropy power _ , @xmath174 , where @xmath175 , defined in analogy with the gaussian entropy function @xmath176 used to define the classical entropy power @xmath18 that appears in the original epi .
_ entropy photon number inequality _
( epni ) @xcite for a pair of independent states @xmath141 and
@xmath105 input to a beamsplitter of transmissivity @xmath70 , producing the state @xmath106 at the output , guha , shapiro and erkmen conjectured the following @xcite : @xmath177 proving that equality occurs if @xmath141 and @xmath105 are thermal states .
the epni plays a role analogous to the epi in converse proofs of capacities of gaussian bosonic channels .
a general proof of epni remains open , but it was proved for gaussian - state inputs @xcite .
it was shown that a special case of the epni , @xmath178 a statement analogous to , would complete the converse proofs of the capacity region of the multi - user bosonic broadcast channel @xcite , and the triple tradeoff region of the pure - loss bosonic channel @xcite .
inequality was proved recently by de palma _
et al . _
@xcite , which when applied to number - diagonal inputs , implies as a special case . _ quantum entropy power inequality _ ( q - epi )
@xcite koenig and smith conjectured ( and provided a partial proof for ) the following , by replacing the @xmath179 in the epni by @xmath180 @xcite : @xmath181 a complete proof of the q - epi , and a multi - input generalization thereof , were provided by de palma _ et al . _
however , unlike the epni , which emerges naturally in the bosonic gaussian channel capacity converses , the q - epi only implies upper bounds ( or outer bounds , in case of broadcast channels ) to the respective capacities @xcite ( or capacity regions , for broadcast channels @xcite ) .
the q - epi implies the following linear form , analogous to , by applying @xmath182 on both sides of and using the concavity of the log function : @xmath183 if the epni were proven true , the concavity of @xmath184 would also imply .
the linear form when applied to number - diagonal inputs implies in theorem [ thm : epi_gsg_linear ] .
similarly , the q - epi implies the discrete epi . by employing for @xmath130 recursively on pairs of identical input states ,
we get : @xmath185 which provides a partial proof of a conjecture by guha and shapiro on entropic monotonicity in a quantum clt ( theorem 5.10 of @xcite ) , with @xmath36 increasing in power - of-@xmath15 increments
. a complete proof of the aforesaid conjecture , when applied to identical independent number - diagonal states , will imply conjecture [ conj : monotonicity ] as a special case . finally ,
a proof of the epni would imply the natural discrete - variable generalization of shannon s epi , which would satisfy all the desirable properties stated in section [ sec : axioms ] and would hold for all discrete random variables and not be restricted to ulc distributed random variables as in the yu - johnson discrete epi .
shannon s entropy power inequality ( epi ) found many applications in proving coding theorem converses for many gaussian channel and source coding problems .
the entropy photon - number inequality ( epni ) was shown to assume a role analogous to shannon s epi in capacity converse proofs for transmitting classical information over gaussian bosonic ( quantum ) channels . even though the general form of the epni remains unproven , several special cases of it have been proven in the recent years .
many attempts have been made to find the most natural discrete - variable version of shannon s entropy power inequality ( epi ) . in this paper , we developed an axiomatic framework from which we deduced the natural form of a discrete - variable epi and an associated entropic monotonicity in a discrete - variable central limit theorem . in this discrete epi , the geometric distribution , which has the maximum entropy among all discrete distributions with a given mean , assumes a role analogous to the gaussian distribution in shannon s epi , and the thermal state in the epni .
we defined the entropy power of a discrete random variable @xmath0 as the mean of a geometric random variable with entropy @xmath1 .
the crux of our construction is a discrete - variable version of lieb s scaled addition @xmath2 of two discrete random variables @xmath0 and @xmath3 with @xmath4 .
we discussed the relationship of our discrete epi with recent work of yu and johnson who developed an epi for a restricted class of random variables that have ultra - log - concave ( ulc ) distributions , and pegged their definition of the entropy power to the entropy function of the poisson distribution , which attains the maximum entropy for a given mean , among the class of ulc random variables .
even though we left open the proof of the aforesaid natural form of the discrete epi that we conjectured in this paper , we showed that this discrete epi holds true for discrete random variables with arbitrary discrete distributions when the entropy power is redefined as @xmath5 in analogy with the continuous version .
finally , we showed that our conjectured discrete epi is a special case of the epni , corresponding to the case when two input quantum states to the epni are independent number - diagonal states .
acknowledge the organizers of the third beyond i.i.d .
workshop held at birs , banff in july 2015 , where this work took seed , and thank rupert l. frank and elliott lieb for inspiring discussions on the epi at that workshop .
the authors thank vittorio giovannetti and seth lloyd for useful discussions .
sg would like to acknowledge the darpa quiness program funded under us army research office award number w31p4q-12 - 1 - 0019 .
j.h.s . acknowledges support from the air force office of scientific research ( afosr ) grant number fa9550 - 14 - 1 - 0052 .
is research associate of the f.r.s .- fnrs . c. e. shannon , a mathematical theory of communication , " bell syst .
j. , * 27 * , pp . 379423 and 623656 ( 1948 ) .
a. j. stam , some inequalities satisfied by the quantities of information of fisher and shannon , " inform .
contr . , * 2 * , 101112 ( 1959 )
. n. m. blachman , the convolution inequality for entropy powers , " ieee trans .
inform . theory ,
it-*11 * , 267271 ( 1965 ) .
p. bergmans , a simple converse for broadcast channels with additive white gaussian noise , " ieee transactions on information theory , vol .
2 , 279280 ( 1974 ) .
h. weingarten , y. steinberg , and s. shamai ( shitz ) , the capacity region of the gaussian multiple - input multiple - output broadcast channel , " ieee transactions on information theory , vol .
52 , no . 9 , 39363964 ( 2006 ) .
m. mohseni and j. m. cioffi , a proof of the converse for the capacity of gaussian mimo broadcast channels , " in proceedings of the ieee international symposium on information theory , seattle , 881885 ( 2006 ) .
s. leung - yan - cheong and m. hellman , the gaussian wire - tap channel , " ieee transactions on information theory , vol .
4 , 451456 ( 1978 ) .
m. h. m. costa , on the gaussian interference channel , " ieee transactions on information theory , vol .
31 , no . 5 , 607615 ( 1985 ) .
r. zamir ,
gaussian codes and shannon bounds for multiple descriptions , " ieee transactions on information theory , vol .
45 , no . 7 , 26292636 ( 1999 ) .
l. ozarow , on a source - coding problem with two channels and three receivers , " bell system technical journal , vol .
10 , 19091921 ( 1980 ) .
y. oohama ,
gaussian multiterminal source coding , " ieee transactions on information theory , vol .
43 , no . 6 , 19121923 ( 1997 )
a. wyner and j. ziv , a theorem on the entropy of certain binary sequences and applications : part i , and part ii " ieee trans .
theory , it-*19 * , 769772 ; 772777 ( 1973 ) .
witsenhausen , entropy inequalities for discrete channels , " ieee trans .
theory , it-*20 * , 610616 ( 1974 ) .
r. ahlswede and j. k " oruer , on the connection between the entropies of input and output distributions of discrete memoryless channels , " proc . of the fifth conference on probability theory , brasov ( 1974 ) .
s. shamai and a. wyner , a binary analog to the entropy - power inequality , " ieee trans .
theory , it-*36 * , 14281430 ( 1990 ) .
p. harremoe " es and c. vignat , an entropy power inequality for the binomial family , " jour . of inequalities in pure and applied math .
, vol . * 4 * , issue 5 , article 93 ( 2003 ) .
s. haghighatshoar , e. abbe , and e. telatar , a new entropy power inequality for integer - valued random variables " , proc .
isit , 589593 ( 2013 ) .
o. t. johnson , log - concavity and the maximum entropy property of the poisson distribution , " stoch .
* 117 * , 6 791802 ( 2007 ) .
y. yu and o. t. johnson , concavity of entropy under thinning , " proc .
isit , 144148 ( 2009 ) .
y. yu , monotonic convergence in an information - theoretic law of small numbers , " ieee trans .
theory , * 55 * , 12 54125422 ( 2009 ) .
o. johnson and y. yu , monotonicity , thinning and discrete versions of the entropy power inequality , " ieee trans . on inform .
theory , vol . * 56 * , no .
11 , 53875395 ( 2010 ) .
j. o. woo and m. madiman , a discrete entropy power inequality for uniform distributions " , proc .
isit , 16251629 ( 2015 ) .
s. guha , j. h. shapiro , and b. i. erkmen , capacity of the bosonic wiretap channel and the entropy photon - number inequality , " proc .
isit , 9195 ( 2008 ) .
e. lieb , proof of an entropy conjecture of wehrl , " comm .
phys . , vol .
* 62 * , pp . 3541 ( 1978 ) .
a. dembo , t. m. cover , and j. a. thomas , information theoretic inequalities , " ieee trans .
* 37 * , 6 , 15011518 ( 1991 ) .
s. verd ' u and d. guo , a simple proof of the entropy - power inequality , " ieee trans .
theory , vol . * 52 * , no .
5 , pp . 21652166 ( 2006 ) .
s. artstein , k. m. ball , f. barthe , and a. naor , solution of shannon s problem on the monotonicity of entropy , " j. amer .
* 17 * , no .
4 , 975982 ( 2004 ) .
r. shimizu , on fisher s amount of information for location family " , statistical distributions in scientific work , g. p. patil _
et al . _ ,
dordrecht , the netherlands : reidel , vol . * 3 * , 305312 ( 1975 ) .
a. tulino and s. verd ' u , monotonic decrease of the non - gaussianness of the sum of independent random variables : a simple proof , " ieee trans .
theory , vol .
* 52 * , no .
9 , pp . 42954297 ( 2006 ) .
m. madiman and a. barron , generalized entropy power inequalities and monotonicity properties of information , " ieee trans .
theory , vol . * 53 * , no . 7 ,
23172329 ( 2007 ) .
p. harremo " es , o. johnson , and i. kontoyiannis , thinning and the law of small numbers , " proc .
isit , 14911495 ( 2007 ) .
r. koenig and g. smith , the entropy power inequality for quantum systems , " ieee trans . inf .
, * 60 * , 3 , 15361548 ( 2014 ) .
s. guha , j. h. shapiro , and b. i. erkmen , classical capacity of bosonic broadcast communication and a new minimum output entropy conjecture , " phys .
a * 76 * , 032303 ( 2007 ) .
g. de palma , d. trevisan , and v. giovannetti ,
gaussian states minimize the output entropy of the one - mode quantum attenuator " , arxiv:1605.00441 [ quant - ph ] ( 2016 ) .
s. guha , multiple - user quantum information theory for optical communication channels " , mit ph.d .
thesis ( 2008 ) .
v. giovannetti , s. guha , s. lloyd , l. maccone , and j. h. shapiro , minimum output entropy of bosonic channels : a conjecture , " phys .
a * 70 * , 032315 ( 2004 ) .
v. giovannetti , a. s. holevo and r. garc ' ia - patr ' on s ' anchez , a solution of gaussian optimizer conjecture for quantum channels , " commun .
august 2014 .
g. de palma , a. mari , s. lloyd , and v. giovannetti , the multi - mode quantum entropy power inequality , " phys .
rev . a * 91 * , 032320 ( 2015 ) .
g. de palma , a. mari , and v. giovannetti , a generalization of the entropy power inequality to bosonic quantum systems , " nat .
photonics * 8 * , 958964 ( 2014 ) .
g. de palma , d. trevisan , and v. giovannetti , passive states optimize the output of bosonic gaussian quantum channels , " arxiv:1511.00293 [ quant - ph ] ( 2015 ) . m. m. wilde , p. hayden and s. guha , information trade - offs for optical quantum communication " , phys .
lett . * 108 * , 140501 ( 2012 ) .
r. koenig and g. smith , limits on classical communication from quantum entropy power inequalities , " nat .
photonics * 7 * , 142146 ( 2013 ) . | many partially - successful attempts have been made to find the most natural discrete - variable version of shannon s entropy power inequality ( epi ) .
we develop an axiomatic framework from which we deduce the natural form of a discrete - variable epi and an associated entropic monotonicity in a discrete - variable central limit theorem . in this discrete epi , the geometric distribution , which has the maximum entropy among all discrete distributions with a given mean , assumes a role analogous to the gaussian distribution in shannon s epi .
the entropy power of @xmath0 is defined as the mean of a geometric random variable with entropy @xmath1 .
the crux of our construction is a discrete - variable version of lieb s scaled addition @xmath2 of two discrete random variables @xmath0 and @xmath3 with @xmath4 .
we discuss the relationship of our discrete epi with recent work of yu and johnson who developed an epi for a restricted class of random variables that have ultra - log - concave ( ulc ) distributions .
even though we leave open the proof of the aforesaid natural form of the discrete epi , we show that this discrete epi holds true for variables with arbitrary discrete distributions when the entropy power is redefined as @xmath5 in analogy with the continuous version .
finally , we show that our conjectured discrete epi is a special case of the yet - unproven entropy photon - number inequality ( epni ) , which assumes a role analogous to shannon s epi in capacity proofs for gaussian bosonic ( quantum ) channels . |
long - lived @xmath17 states are especially well suited for testing qcd via lattice calculations @xcite and effective theories of strong interactions , like potential models @xcite . the narrow triplet @xmath18 states , @xmath7 , @xmath19 and @xmath1 , were discovered in 1977 in proton - nucleus collisions at fermilab @xcite .
later , they were better resolved and studied at various @xmath8 storage rings .
six triplet @xmath20 states , @xmath21 and @xmath22 with @xmath23 , were discovered in radiative decays of the @xmath1 and @xmath19 in 1982 @xcite and 1983 @xcite , respectively .
there has been no observations of new narrow @xmath24 states since then despite the large number of such states predicted below the open flavor threshold ( see fig . [
fig : level ] ) . in this paper , we present the first observation of the @xmath2 states . except for the yet unobserved @xmath25 states , these are the only long - lived @xmath26 mesons in nature .
they are produced in a two - photon cascade starting from the @xmath1 resonance : @xmath27 , @xmath28 .
to suppress photon backgrounds from @xmath29s , which are copiously produced in gluonic annihilation of the @xmath17 states , we select events with two more subsequent photon transitions , @xmath30 , @xmath31 , followed by the @xmath7 annihilation to either @xmath8 or @xmath9 ( see fig . [
fig : level ] ) .
the product branching ratio for these five decays in sequence was predicted by godfrey and rosner @xcite to be @xmath32 .
the cleo iii experiment has recently accumulated @xmath0 @xmath1 decays , which makes the detection of such rare processes possible .
the cleo iii sample constitutes roughly a ten - fold increase in @xmath1 statistics compared to the cleo ii data set @xcite , and roughly a four - fold increase compared to the integrated cusb data set @xcite .
thanks to the good granularity and the large solid angle of the cleo csi(tl ) calorimeter , the cleo iii detection efficiency for these final states is about 4.5 times larger than in the cusb detector . even though the cleo iii calorimeter is essentially the same as in the cleo ii detector @xcite , the photon selection efficiency and detector resolution were improved in the endcaps and in part of the barrel calorimeter by the installation of a new , lower - mass , tracking system @xcite .
another change was the replacement of the time - of - flight system by a rich detector in the barrel part .
finally , the calorimeter endcaps were restacked and moved farther away from the interaction point to accommodate a new , higher luminosity , interaction point optics .
we select events with exactly four photons and two oppositely charged leptons .
the leptons must have momenta of at least 3.75 gev/@xmath33 .
we distinguish between electrons and muons by their energy deposition in the calorimeter .
electrons must have a high ratio of energy observed in the calorimeter to the momentum measured in the tracking system ( @xmath34 ) .
muons are identified as minimum ionizing particles , and required to leave @xmath35 mev of energy in the calorimeter .
stricter muon identification does not reduce background in the final sample , since all significant background sources contain muons .
each photon must have at least 60 mev of energy .
we also ignore all photons below 180 mev in the calorimeter region closest to the beam , because of the spurious photons generated by beam - related backgrounds .
the total momentum of all photons and leptons in each event must be balanced to within 300 mev/@xmath33 .
the invariant mass of the two leptons must be consistent with the @xmath7 mass within @xmath36 mev/@xmath14 .
much better identification of the @xmath7 resonance is obtained by measuring the mass of the system recoiling against the four photons .
the average resolution of the recoil mass is 17 mev/@xmath14 .
cuts on the recoil mass are described below .
the mass resolution of the produced @xmath2 state depends on the measurement of the energies of the two lowest energy photons in the event .
thus , we require that at least one of them is detected in the barrel part of the calorimeter , where the energy resolution is best .
the energy resolution as determined by the fit to the @xmath27 photon lines is @xmath37 mev for @xmath38 mev @xcite .
the most prominent signal among @xmath39 decays are @xmath16 transitions . to identify these decays
, we try all three combinations of four photons into two - photon pairs .
we find the minimal @xmath29 mass deviation chi - squared among these combinations : @xmath40 ^ 2 + [ ( m_{\gamma\gamma}^{kl}-m_{\pi^0})/\sigma_{m_{\gamma\gamma}^{kl}}]^2\right\}$ ] . after the requirement @xmath41 , the deviation of the four - photon recoil mass from the @xmath7 mass , @xmath42 , is found .
the distribution of the ratio , @xmath43 , where @xmath44 is the expected recoil mass resolution , is plotted in fig . [
fig : mdrpi0pi0 ] . as can be seen from the sidebands of this distribution , backgrounds from non-@xmath45 transitions ( e.g. radiative bhabha and @xmath46pair events )
are small .
we subtract this background by fitting the recoil mass distribution with the signal shape extracted from the monte carlo simulations . the fit is also shown in fig .
[ fig : mdrpi0pi0 ] .
we observe @xmath47 events with an efficiency of 13.6% .
the efficiency was determined from the monte carlo simulations .
this sample contains a small fraction , 0.9% , of photon cascade background that we estimate from the cross - efficiency determined with the monte carlo ( below 1% ) and the branching ratios measured from the data , as described below . after the background subtraction ,
we obtain the following result for the product branching ratio : @xmath48 .
the first error is statistical and the second is systematic .
the systematic error is dominated by uncertainties in the monte carlo simulation of the selection efficiency .
it was determined by varying the selection criteria . here
, @xmath49 means either @xmath50 or @xmath51 , since these two branching ratios should be equal from lepton universality . using the world average value @xcite , @xmath52
, we can unfold our result for @xmath53 .
this result is consistent with , but more precise , than previous measurements by cleo ii @xcite , @xmath54 , and cusb @xcite , @xmath55 .
the efficiency - corrected invariant mass distribution of the two - pion system is shown in fig .
[ fig : mpi0pi0 ] . as in the previous measurements , some dynamical structure
is observed in this distribution .
the determination of @xmath56 presented above serves as a good cross - check of our experimental techniques . in a separate paper submitted to this conference @xcite
, we also present the analysis of the two - photon cascades between the @xmath1 and the @xmath7 or @xmath19 . branching ratios for two - photon cascades via @xmath57 and @xmath58 states
were determined with improved precision compared to previous experiments .
these data , together with the @xmath59 transitions , were used to obtain the calibration of absolute photon energies in the @xmath60 mev range . using the well known masses of the @xmath7 , @xmath19 and @xmath1 @xcite , together with
the recoil mass distributions like the one shown in fig .
[ fig : mdrpi0pi0 ] , we calibrated the photon energies to a precision of @xmath61 .
the photon energies observed in the two - photon cascade data for @xmath27 agreed well with the previous measurements .
thus , the systematic effects in both rate and photon energy measurements are well under control .
to look for four - photon cascades between the @xmath1 and @xmath7 , we must suppress backgrounds from the @xmath59 transitions .
we first require , @xmath62 . in addition
, every photon pair must have an invariant mass at least two standard deviations away from the nominal @xmath29 mass . to look for @xmath2 events , we constrain events to be consistent with a photon cascade from the @xmath1 to the @xmath7 via one of the @xmath21 and one of the @xmath22 states ( @xmath63 ) .
the masses of the @xmath20 states are well known from other measurements @xcite . for each @xmath64 ,
@xmath65 combination we calculate a chi - squared : @xmath66 where @xmath67 are the measured photon energies ; @xmath68 are the expected photon energies from the masses of the @xmath17 states and the measured photon directions in the cascade .
the masses of the @xmath2 states are not know .
therefore , we minimize the above chi - squared with respect to @xmath69 which is allowed to vary for each event .
the above formalism requires that we know how to order the four photons in the cascade . while the highest energy photon must be due to the fourth transition , and the second highest energy photon must be due to the third transition , there is sometimes an ambiguity in the assignment of the two lowest energy photons to the first two transitions , since the range of photon energies in the @xmath27 decay overlaps the possible photon energy range in the @xmath28 transition .
we choose the combination that minimizes the above chi - squared .
there are nine possible combinations of @xmath64 , @xmath65 spins .
we try all of them and choose the one that produces the smallest chi - squared , @xmath70 .
in addition to the four - photon cascade via the @xmath2 states , our data contain events with the four - photon cascade via the @xmath19 state : @xmath27 , @xmath71 , @xmath72 , @xmath31 , @xmath73 ( see fig . [
fig : level ] ) .
this is a coincidence of the two - photon cascades , @xmath74 and @xmath75 , which were experimentally observed in the two - photon cascades detected in the @xmath1 @xcite and @xmath19 data @xcite .
the product branching ratio for this entire decay sequence ( including @xmath73 ) is predicted by godfrey and rosner @xcite to be @xmath76 , thus comparable to the predicted @xmath2 production rate . combining the measured rates for @xmath77 @xcite and @xmath78 @xcite with the measurement of @xmath79 @xcite
, we obtain a factor @xmath80 higher value . in these events , the second highest energy photon is due to the second photon transition ( see fig . [
fig : level ] ) .
unfortunately , these events can be sometimes confused with the @xmath2 events due to the limited experimental energy resolution .
the second and third photon transitions in the @xmath19 cascade sequence can be mistaken for the third and second transition in the @xmath2 cascade sequence .
therefore , it is important to suppress the @xmath19 cascades . we achieve this by finding the @xmath64 , @xmath65 combination that minimizes the associated chi - squared for the @xmath19 hypothesis , @xmath81 , where @xmath82 is exactly analogous to @xmath83 with the @xmath69 replaced with @xmath84 .
we then require @xmath85 .
notice that the masses of all intermediate states are known for the @xmath19 cascade , thus this variable is more constraining than @xmath83 . to further suppress the @xmath19 cascade events , we construct a quasi - chi - squared variable , @xmath86 , that sums in quadrature only positive deviations of the measured photon energies from their expected values .
this variable is less sensitive to fluctuations in the longitudinal and transverse energy leakage in photon showers that sometimes produce large negative energy deviations and correspondingly a large @xmath82 value . with the additional cut @xmath87 ,
the cross - efficiency for @xmath19 events is reduced to 0.7% and is further suppressed by a factor of two when a cut of @xmath88 is required . the data @xmath83 distribution , after we require the four - photon recoil mass to be between @xmath89 and @xmath90 standard deviations from the nominal @xmath7 mass , is shown in fig .
[ fig : chiddata ] .
a narrow peak near zero is observed , just as expected for @xmath2 events . the signal monte carlo distribution for @xmath2 events is shown in fig .
[ fig : chidmc ] .
the background monte carlo distribution for the @xmath19 cascades , after a factor of 10 enhancement relative to the @xmath2 normalization , is also shown for comparison .
clearly , the @xmath19 cascade background can not produce as narrow a peak as observed in the data .
the estimate of the number of @xmath19 cascade events passing the @xmath88 requirement is @xmath91 events ( depending on the assumed @xmath19 product branching ratio ) .
we observe 44 events after this cut in the data .
the cross - efficiency for @xmath59 events is @xmath92 in this sample , which gives a negligible contribution to the signal region .
a flat tail in the @xmath83 distribution for the data , not observed for the signal monte carlo , indicates that there is still some background present from other processes like radiative qed events .
since such processes can satisfy the multiple mass constraints built into the @xmath83 variable only by a random coincidence , the distribution of that background varies slowly in @xmath83 except for the bias towards lower values introduced by the @xmath83 minimization with respect to @xmath69 , @xmath64 and @xmath65 .
we use the @xmath16 monte carlo ( with the @xmath29 veto cuts removed ) as a model of the @xmath83 distribution for this background component ( see fig . [
fig : chidp0p0fit ] ) .
normalization of this background is fixed from the observed tail of the @xmath83 distribution .
this shape is likely to overestimate the amount of bias in @xmath83 towards the lower values since , unlike the @xmath16 process , the qed backgrounds do not automatically satisfy the @xmath93 mass constraint built into the chi - squared variable .
therefore , we also used a linear background model , with the line parameters fixed by a fit to the tail of the @xmath83 distribution . depending on the background shape model , the @xmath88 sample has a @xmath94 background rate ( including the @xmath19 cascade background ) .
a fit of the @xmath2 signal , on top of the backgrounds , is shown in fig .
[ fig : chidfit ] .
the @xmath19 background contribution was fixed from the monte carlo simulations .
normalization of the other backgrounds was left as a free parameter in the fit .
the fit gives a signal amplitude of @xmath95 events .
if we try to fit the data with the background contribution alone , the likelihood value of the fit drops dramatically . from the change of the likelihood we estimate that the signal has a significance of 9.7 standard deviations .
significant signals are observed in the @xmath96 and @xmath97 subsamples taken separately ( @xmath98 and @xmath99 , respectively ) .
to turn the observed signal yield into a signal branching ratio , we need to estimate the signal efficiency .
even though the signal shape in the @xmath83 variable is fairly independent of the masses and spins of the produced @xmath2 states , the selection efficiency does depend on these parameters via the cuts used to suppress the @xmath19 cascade background .
this is illustrated in fig .
[ fig : effvsmass ] , where we plot the selection efficiency as a function of the mass of the @xmath2 state .
three different curves are presented , for @xmath100 and @xmath101 .
the @xmath102 state is predicted by godfrey and rosner @xcite to be produced with the highest rate , @xmath103 .
the @xmath104 , @xmath105 combination is predicted to be the main production and decay path for this state .
the selection efficiency for the @xmath102 state gradually increases with its mass , but is reasonably high in the entire search window .
the @xmath106 state , predicted with a @xmath107 rate , can only be produced via the @xmath108 , @xmath109 combination due to angular momentum conservation .
this happens to be an unfortunate combination , since for a wide range of masses such a production fakes the @xmath19 cascade , and the detection efficiency after the @xmath19 suppression cuts is very low .
we become sensitive to this state only if its mass exceeds about .
finally , the @xmath110 state is predicted by godfrey and rosner to have the lowest product branching ratio , @xmath111 .
this state is again predicted to be produced mostly via the @xmath104 , @xmath105 combination , and therefore our efficiency for this state is similar to the @xmath102 efficiency . a straightforward way to measure
the mass of the produced @xmath2 state is to calculate the mass of the system recoiling against the two lowest energy photons in the event .
the distribution of the difference between the center - of - mass energy and this recoil mass , @xmath112 , which should peak at @xmath113 , is shown for the monte carlo simulation of a @xmath15 state with a mass of 10160 mev/@xmath14 in fig .
[ fig : m3s1dmc ] .
this distribution can be well described with a gaussian that has a power - law tail on the low side ( a so - called crystal ball line shape ) , as illustrated in fig .
[ fig : m3s1dmc ] .
the asymmetric tail is caused by transverse and longitudinal shower energy leakage in the calorimeter .
the gaussian part has a resolution of @xmath114 mev/@xmath14 . the mass value that minimizes @xmath83 is also an estimate of the true @xmath2 mass . the distribution of this mass for the same monte carlo sample as above is shown in fig .
[ fig : m1dfmc ] .
a narrow peak at the correct mass is observed with a resolution of @xmath115 mev/@xmath14 .
the mass resolution is improved in this method due to the constraints from the well measured masses of the @xmath20 states built into the @xmath83 variable . unfortunately , there are also two satellite peaks , one on each side of the main peak .
these false peaks originate from photon energy fluctuations which can make a wrong @xmath64 , @xmath65 combination produce the smallest chi - squared value .
the distributions of @xmath112 and the @xmath2 mass which minimizes @xmath83 for each event ( called the most - likely mass ) are shown for the data in figs .
[ fig : m3s1ddata]-[fig : m1dfdata ] after the @xmath88 cut .
the background fraction for this sample is in the @xmath116 range .
two peaks are observed in the distribution of the most likely mass , around 10160 and 10175 mev/@xmath14 .
the recoil mass distribution does not show a double - peak structure , but the mass resolution is worse in this variable , as discussed above .
to analyze the most likely mass distribution , which has a better mass resolution , we fit the data to either a two - peak or a one - peak ( positioned either around 10160 or 10175 mev/@xmath14 ) hypothesis . the background is assumed to have a flat distribution .
since the backgrounds are not significant , their parameterization is rather unimportant . because , in principle , up to three triplet @xmath2 states can contribute to our sample , we also tried to force a third peak into the fit function .
however , the fit either makes the amplitude of this peak zero , with no effect on the amplitudes of the other two peaks , or the third peak collapses into one of the other two peaks .
the two - peak fit , displayed in fig .
[ fig : m1dfit2 ] , gives the best confidence level ( @xmath117 ) .
the amplitude ( mass ) of the first peak is @xmath118 events ( @xmath119 mev/@xmath14 ) , and of the second peak @xmath120 events ( @xmath121 mev/@xmath14 ) .
the assumption that there is no peak around 10160 mev/@xmath14 produces a very low confidence level ( @xmath122 ) and can be ruled out . from the change of likelihood between these two fits ,
the significance of the first peak is 6.8 standard deviations .
the hypothesis of just one peak around 10160 mev/@xmath14 also produces a good confidence level of @xmath123 ( see fig .
[ fig : m1dfit1 ] ) , with @xmath124 events in the peak ( @xmath125 mev/@xmath14 mass )
. furthermore , while the recoil mass distribution is consistent with the production of two states with the amplitudes and masses as determined by the fit above ( confidence level @xmath126 ) , it prefers to allocate most of the events to the lower mass state , @xmath127 , with only @xmath128 events attributed to the higher mass state , when the amplitudes are let free in the fit .
this fit is displayed in fig .
[ fig : m3s1dfit2 ] and has a confidence level of @xmath129 . the single - peak interpretation of the recoil mass distribution also has a high confidence level ( @xmath130 ) , with @xmath131 events at a mass of @xmath132 mev/@xmath14 ( fig .
[ fig : m3s1dfit1 ] ) . in the region of the first peak in the most likely mass distribution ,
@xmath133%of the events have the @xmath104 , @xmath105 combination as the most likely .
this is consistent with the monte carlo simulations for both the @xmath15 ( @xmath134 ) and @xmath135 ( @xmath136 ) states .
furthermore , we have very low efficiency for the @xmath137 state at this mass ( fig . [
fig : effvsmass ] ) .
when we remove the cuts suppressing the @xmath19 cascades , which are responsible for the efficiency loss for the @xmath137 state , the amplitude of this peak does not increase dramatically .
therefore , this can not be the @xmath106 state .
the @xmath15 state is predicted to be produced with a 6 times larger rate than the @xmath135 state @xcite .
the clebsch - gordan coefficients play a big role in this ratio , thus the model dependence of this prediction is weak .
therefore , we take the @xmath15 state as the most likely interpretation of the first peak . taking an average over the different fits described above
, we measure mass of this state to be @xmath13 mev/@xmath14 .
the error is dominated by the dependence on the fit approach ( @xmath138 mev/@xmath14 ) , with smaller contributions from the statistical error ( @xmath139 mev/@xmath14 ) , photon energy scale error ( @xmath139 mev/@xmath14 ) and the uncertainty on the @xmath1 mass ( @xmath140 mev/@xmath14 ) .
predictions of the most successful potential models predict the center - of - gravity of the @xmath2 triplet to be around this mass @xcite .
calculations of the fine splitting of the @xmath2 states predict the @xmath102 state to be close to the center - of - gravity mass @xcite .
our mass measurement is consistent with these predictions .
only 3 events have @xmath108 , @xmath109 as the most likely combination . if the @xmath106 state was produced with the amplitude of 12 events , as suggested by the two - peak fit to the distribution of the most likely mass , we would expect 9.4 such events , including 1.5 events of feed - down from the @xmath15 state at 10162 mev/@xmath14 . as described previously ,
the recoil mass distribution shows no evidence for a state at 10174 mev/@xmath14 either .
thus , there is no strong evidence for the second state in our data .
more data , which will be accumulated by cleo iii at the @xmath1 in the future , will help to clarify the interpretation of these results .
to calculate the signal efficiency in the fit to the @xmath83 distribution ( fig .
[ fig : chidfit ] ) , we weight the @xmath15 efficiency for a state with a mass of @xmath141 mev/@xmath14 and the @xmath137 efficiency for a state with a mass of @xmath142 mev/@xmath14 in a ratio proportional to the observed peak amplitudes in the two - peak fit to the observed most likely mass distribution .
we use the difference between this weighted efficiency and the individual efficiencies for each state as an estimate of the systematic error .
this gives an efficiency of @xmath143 . converting the number of @xmath2 events determined by the fit to the @xmath83 distribution to a product branching ratio , we obtain @xmath144
this is an inclusive rate that sums over all possible @xmath2 states .
this rate is in good agreement with the @xmath32 prediction by godfrey and rosner @xcite .
in summary , we present evidence at the 9.7 standard deviation significance for production of @xmath2 states in four - photon cascades from the @xmath1 .
the inclusively measured product branching ratio , @xmath145 @xmath146 @xmath147 @xmath148 @xmath49 @xmath149 , is in good agreement with the theoretical predictions by godfrey and rosner @xcite , where @xmath49 above means the average branching ratio over @xmath8 and @xmath9 . from the observed mass distribution
we see a 6.8 standard deviation signal for a state with a mass of @xmath13 mev/@xmath14 .
this is likely to be the @xmath102 state , though we can not rule out the @xmath110 hypothesis .
however , the @xmath110 hypothesis is strongly disfavored by theoretical expectations .
evidence for a second heavier state in our data is so far inconclusive .
this is the first discovery of an @xmath26 quarkonium state below the open flavor threshold .
we gratefully acknowledge the effort of the cesr staff in providing us with excellent luminosity and running conditions .
we thank j. l. rosner for the useful discussions concerning the @xmath2 states .
m. selen thanks the pff program of the nsf and the research corporation , and a.h .
mahmood thanks the texas advanced research program .
this work was supported by the national science foundation , and the u.s .
department of energy . | the cleo iii experiment has recently accumulated a large statistics sample of @xmath0 @xmath1 decays .
we present the first evidence for the production of the triplet @xmath2 states in the four - photon cascade , @xmath3 , @xmath4 , @xmath5 , @xmath6 , followed by the @xmath7 annihilation to @xmath8 or @xmath9 .
the signal has a significance of @xmath10 standard deviations .
the measured product branching ratio for these five decays , @xmath11 , is consistent with the theoretical estimates . we see a @xmath12 standard deviation signal for a state with a mass of @xmath13 mev/@xmath14 , consistent with the @xmath15 assignment .
we also present improved measurements of the @xmath16 branching ratio and the associated di - pion mass distribution . #
1#2#3#4 + # 1#2#3#4#5 + # 1#2#3#4#5#6#7 + # 1#2#3#4#5#6#7#8#9 + + international conference on high energy physics , july 2002 , amsterdam ] |
it is of great importance to elucidate the symmetry of a superconducting pair potential is of great importance when studying the cooper pairing mechanism in unconventional superconductors ( scs ) .
the pair potential is composed of spin and orbital wave functions .
the orbital wave function is characterized by its amplitude and phase ( sign of the wave function ) . in the past decade , experimental techniques for the field - angle resolved specific heat and thermal conductivity measurements have developed to identify the cooper pairing symmetry in various superconducting systems.@xcite these angle - resolved mesurements are powerful techniques that can detect the anisotropy of the pair potential amplitude . the theory proposed by vorontsov and vekhter
has successfully explained these experiments for cecoin@xmath2 assuming @xmath3-wave cooper pairing.@xcite however , these field - angle resolved experiments can not probe the sign change in the pair potential .
that is , they are not phase - sensitive probes .
in addition to detecting the anisotropy of the pair potential , it is crucial to probe the phase of the cooper pair in order to discriminate unconventional scs , including iron - based scs , from conventional ones . until now
, only a few phase - sensitive probes have been developed and succeeded , e.g. , the half - flux quantum observation in the tricrystal geometry by a scanning squid ( superconducting quantum interference device ) microscope,@xcite and detecting the quasiparticle interference pattern by scanning tunneling spectroscopy ( sts).@xcite another phase - sensitive probe is measuring bound states at an interface by point - contact spectroscopy or sts experiments . if both sides of a sc / sc junction are of the same pair potential amplitude but with opposite signs , the quasiparticle ( qp ) path through the interface acquires a @xmath4 phase shift and generating bound states around the interface .
this situation is similar to a vortex line in superconductors .
however , there is the difficulty of fabricating a junction in terms of nono - structured processing techniques .
actually , the phase - sensitive test using sc / sc junctions succeeds only for cuprate superconductors.@xcite in addition to these existing experiments , a new phase - sensitive test is highly desired . in this paper
, we propose a new experiment that can detect the phase ( sign - change ) of the pair potential free from fabricating a sc / sc junction .
this is the great advantage of the phase - sensitive test proposed in the present work .
we theoretically study the in - plane field - angle dependence of the flux - flow resistivity @xmath0 for typical gap functions and fermi surface ( fs ) . from our numerical calculations , we show that the phase - sensitive qp scattering inside a vortex core leads to different behavior of @xmath0 between conventional and unconventional cooper pairing .
in addition , we show that @xmath0 has its maximum when the applied magnetic field @xmath5 is parallel to the gap - node directions .
our results show that the field - angle dependence of the flux - flow resistivity can detect both the sign change of the pair potential and the direction of the gap nodes .
the flux - flow resistivity @xmath1 is dominated by the quasiparticle within a vortex core .
we assume the system belongs to the moderately clean regime , in which there are two important contributions to @xmath1 .
one is the qp scattering rate @xmath6 inside a vortex core , and the other is the momentum - dependent quantum level spacing of the vortex bound states @xmath7.@xcite here , the qp scattering is due to non - magnetic impurities randomly distributed in the system .
we attribute the flux - flow resistivity @xmath1 to the energy dissipation of the vortex bound states due to the impurity scattering inside a vortex core.@xcite @xmath1 is characterized by the two quantities mentioned above,@xcite @xmath8^{-1 } , \label{flux - flow_resistivity}\ ] ] where @xmath9 is the impurity scattering rate in the normal state and @xmath10 is the bulk amplitude of the pair potential .
we assume that the temperature @xmath11 dependence of @xmath1 comes predominantly from @xmath6 with the qp energy @xmath12 . here
, we have made a rough estimate . actually , the qps distributed with the energy width @xmath13 contribute to @xmath6 .
the total density of states on a fs is @xmath14 , with @xmath15 being an area element on the fs , the fermi velocity @xmath16 , and the fermi wave vector @xmath17 .
@xmath18 is the energy dispersion of electrons .
@xmath20 is the azimuthal ( polar ) angle on the fs .
@xmath21 , @xmath22 , and @xmath23 denote orthogonal unit vectors spanning crystal coordinates .
we use the unit system in which @xmath24 .
the momentum - dependent inter - level spacing of the vortex bound states @xmath7 is obtained analytically as @xcite @xmath25 using the quasiclassical green s function method and the kramer - pesch approximation.@xcite @xmath26 indicates the anisotropy of pair potential and the vector with @xmath27 denotes the vector component projected onto the plane perpendicular to @xmath28 .
we treat the non - magnetic impurity scattering by means of the born approximation.@xcite the quasiparticle scattering rate for the qps with the energy @xmath29 inside a vortex core is obtained as @xcite @xmath30 \cos\theta , \label{coherence - factor } \\
% % % d(\bm{k}_{\rm f},\bm{k}^\prime_{\rm f } ) & = \frac{1}{\vert\sin\theta\vert } , \\
% % % f(\varepsilon,\bm{k}_{\rm f},\bm{k}^\prime_{\rm f } ) & = \frac{\vert \bm{v}_{\rm f\perp}(\bm{k}^\prime_{\rm f } ) \vert}{\vert \bm{v}_{\rm f\perp}(\bm{k}_{\rm f } ) \vert } \frac{\vert d(\bm{k}_{\rm f } ) \vert}{\vert d(\bm{k}^\prime_{\rm f})\vert } e^{-u(s_0,\bm{k}_{\rm f } ) } e^{-u(s^\prime_0,\bm{k}^\prime_{\rm f})}. \label{f}\end{aligned}\ ] ] here , @xmath31 , @xmath32 for the scattering angle [ see fig . [
fig : fig1 ] ] .
@xmath6 has the decay factor @xmath33 $ ] with @xmath34 we adopt @xmath35 as the spatial variation of the pair potential .
the coherence length is defined by @xmath36 with @xmath37 .
here , we define the field - angular dependent effective coherence length for the later discussions as @xmath38 . \label{xi_eff}\ ] ] this length scale @xmath39 characterizes the size of the bound states of the qp with the momentum @xmath40 .
figure [ fig : fig1 ] shows the qp trajectories on the plane perpendicular to the magnetic field @xmath28 .
the quantities with a prime are those after scattering . @xmath41 and @xmath42 denote the length between the point that is the nearest from the vortex center on the qp trajectory and the scattering point.@xcite
p80 mm + in this study , we consider the case in which @xmath28 is applied parallel to the @xmath43-@xmath44 plane and rotated .
the field angle measured from @xmath43 axis is @xmath45 . here ,
@xmath43 , @xmath44 , and @xmath46 are crystal axes .
when calculating the dependence of @xmath1 on the magnetic field angle @xmath45 , we need a coordinate system fixed to @xmath28 which is spanned by @xmath47 , @xmath48 , and @xmath49 ( vortex coordinate system ) . here , these axes are orthogonal unit vectors with @xmath49 set parallel to @xmath28 ( @xmath50 ) .
@xmath51 , @xmath52 , and those with a prime are defined in the vortex coordinates .
however , @xmath40 and @xmath53 are identified by @xmath54 on a fs in the crystal coordinates spanned by @xmath21 , @xmath22 and @xmath23 , which characterize the crystal axes . in order to calculate the field - angle @xmath45 dependence of @xmath1
, we need to derive the relation between @xmath55 , @xmath52 , and @xmath56 , @xmath57 @xcite . then , the component of @xmath58 projected onto the plane perpendicular to @xmath28 is finally obtained as @xmath59 thus , the relation between the vortex coordinate and the crystal coordinate is derived . here
we give the expression of the projected fermi velocity for an arbitrary anisotropic fs ( see ref .
for the expression for a uniaxially anisotropic fs ) .
we have now reached the position where we can perform the numerical integration on fs and calculate the field - angle dependence of eq .
( [ flux - flow_resistivity ] ) .
we consider the following two types of the simple pair potential model .
one is a line - node @xmath60-wave pair : @xmath61 .
the other is a @xmath62-wave one : @xmath63 .
each one has gap nodes from the north pole of the fs to the south one in the @xmath64 $ ] directions ( gap - node directions ) with @xmath65 .
@xmath66 $ ] directions correspond to anti - node directions . in the momentum space ,
these two pair potentials have the same anisotropy in their amplitude @xmath67 . however , only the @xmath3-wave pair has the sign change and the @xmath60-wave pair does not .
our calculations are performed for a uniaxially anisotropic fs with the mass anisotropy parameter @xmath68 @xcite . @xmath69 and @xmath70 are masses characterizing charge transport along the @xmath46-axis and within the @xmath71 plane , respectively .
p70mmp70 mm + [ cols= " < , < , < , < " , ] [ table:1 ]
first of all , we consider @xmath6 , which is a part of the contribution to @xmath1 as shown in eq .. in the case of the line - node @xmath60-wave pair , taking into account the characteristics of the qp scattering ( i ) ( iii ) , we can explain qualitatively the behavior of @xmath72 . in the line - node @xmath60-wave pair , @xmath72 has its minimum when @xmath28 is applied to the gap - node direction @xcite . in this case
, according to the factor ( iii ) , the effect of the coherence factor on @xmath73 is small .
in addition , in the line - node @xmath60-wave pair , we confirmed that the coherence factor has no field - angle dependence .
hence , we can neglect the factor ( iii ) in the line - node @xmath60-wave pair . when @xmath28 is parallel to the gap - node direction , the weight of @xmath73 shifts toward the field direction ( i.e. , the gap - node direction ) due to the factor ( ii ) .
therefore , the qps around the gap node become easier to be scattered .
however , according to the factor ( i ) , the contribution of the qps in the vicinity of the gap nodes to @xmath73 is small .
hence , the large contribution to @xmath73 due to the factor ( ii ) gets small due to the factor ( i ) . on the other hand , when @xmath28 is parallel to the anti - node direction , the qps in the direction of @xmath28 ( i.e. , the anti - node direction ) , which have small @xmath74 , can give a large contribution to @xmath73 due to the factor ( ii ) .
in this case , contrary to the case of @xmath28 parallel to the node direction , @xmath6 remains large due to the factor ( i ) . as a result of the above consideration
, the minimum of @xmath72 appears when @xmath28 is parallel to the gap - node direction .
p70 mm in the case of the @xmath3-wave pair , the backward scattering is suppressed ( see table [ table:1 ] ) since the coherence factor becomes zero and the scattering point is far from the vortex core . however , the forward scattering with the sign change of pair potential is enhanced ( see table [ table:1 ] ) .
so the coherence factor gives the large contribution to @xmath73 . in fig .
[ fig : fig5 ] , we show the field - angle dependences of @xmath75 , which is the part of @xmath76 containing the coherence factor @xmath77 .
we also calculate the field - angle dependences of @xmath78 , which does not contain the coherence factor .
when @xmath28 is parallel to the node direction , @xmath79 shows a sharp maximum . on the other hand
, @xmath78 shows little field - angle dependence .
this sharp maximum reproduces the behavior of @xmath72 in the @xmath3-wave case.@xcite let us explain the physical picture of the qp scattering rate in the @xmath3-wave case .
first of all , we should note that the intensity of the forward scattering is the important factor of the qp scattering around a vortex , since the forward scatterings occur when the scattering point is near the vortex center ( i.e. , the qp scattering occurs inside a vortex core ) , as shown in fig .
[ fig : fig1 ] .
thus , we consider the field - angle dependence of the intensity of the forward scattering .
we note that the intensity of the forward scattering becomes larger upon decreasing the effective coherence length @xmath39 .
the effective coherence length @xmath39 defined by eq . is proportional to the projected fermi velocity @xmath80 and is inversely proportional to the amplitude of a pair potential @xmath81 .
the minimum effective coherence length is zero when the fermi velocity @xmath82 is parallel to @xmath28 [ i.e. , @xmath80 becomes zero ] .
as shown in fig .
[ fig : fig3 ] , the intensity of @xmath83 becomes large in the region where @xmath80 becomes small .
this is the reason for the factor ( ii ) .
the most important factor of the intensity of the forward scattering is the factor ( iii ) .
as shown in table [ table:1 ] , the sign - conserved forward scattering is suppressed even when the effective coherence length becomes small .
therefore , the sign - reversed forward scatterings with the small effective coherence length dominantly contribute to the qp scattering rate in the @xmath3-wave case .
when @xmath28 is parallel to the gap - node direction , we have confirmed numerically that the forward scattering is realized by calculating the contribution of anti - nodal qps to @xmath84 .
as shown in fig .
[ fig : fig6 ] , the forward scattering occurs through the qp scattering in the parallel direction of @xmath28 when @xmath28 is parallel to the gap - node direction .
moreover , this qp scattering process is sign - reversing , since the quasiparticles with @xmath56 are scattered across the gap - node perpendicular to @xmath28 .
hence , the sign - reversed forward scattering occurs when @xmath28 is parallel to the gap - node direction even in a single - band superconductor . on the other hand , when @xmath28 is parallel to the anti - node direction , it was revealed through the same analysis that although the sign - reversed scattering occurs , not only the forward scattering but also the backward scattering occurs . as a result of the above discussion , the quasiparticle scattering rate is enhanced when @xmath28 is parallel to the gap - node direction in the @xmath3-wave case .
p50 mm + the other contribution to @xmath1 is @xmath7 .
when @xmath85 , the field angle dependence of @xmath7 for a spheroidal fs is not qualitatively different from that for an isotropic fs.@xcite when @xmath86 , @xmath87 . in the @xmath3-wave case
, the dependence of @xmath7 on the field - angle makes the behavior of @xmath72 stand out . as a result
, @xmath0 has a sharp maximum when @xmath28 is parallel to the gap - node direction . on the other hand , in the @xmath60-wave case , the field angle dependence of @xmath7 makes the oscillation amplitude of @xmath72 inverted and @xmath0 has its broad maximum when @xmath28 is parallel to the gap - node direction .
finally , we comment on the experimental condition for measuring the flux - flow resistivity under a rotating magnetic field .
our theory is based on the vortex bound states formed inside a vortex core .
therefore , an extremely two - dimensional system , in which a josephson vortex is formed parallel to the layer , is beyond our theoretical framework .
however , we should note that our method can be applied to iron - pnictides , which have a warped cylindrical fermi surface such as that found in 11-compounds ( fese or fete ) and 122-compounds [ bafe@xmath88(as@xmath89p@xmath90)@xmath88 ] , since the angular - resolved specific heat and thermal conductivity measurements have successfully detected the gap minima in fese@xmath92te@xmath93 @xcite and the position of the gap - nodes in bafe@xmath88(as@xmath94p@xmath95)@xmath88,@xcite respectively , under a rotated magnetic field within the basal plane . in layered organic compounds @xmath96-(et)@xmath88cu(ncs)@xmath88 , in - plane field angular dependence of the josephson - vortex flow resistance
has already been measured by yasuzuka _
et al_.,@xcite but in a three - dimensional system , measurement of the flux - flow resistivity has not been performed yet .
we calculate @xmath0 also in the case of the in - plane anisotropic fs.@xcite in this case , the behavior of @xmath0 is not qualitatively different from that in the isotropic fs case .
when considering multiband superconductors such as iron - based superconductors , we need to take into account the contribution from holelike fs to the flux - flow resistivity in addition to that from electronlike fs.@xcite in this study , we consider the contribution only from electronlike fs . the multiband effect on the flux - flow resistivity is left for future study .
in conclusion , we theoretically studied the in - plane magnetic field - angle dependence of the flux - flow resistivity for a uniaxially anisotropic fs . we showed that the measurement of the flux - flow resistivity changing the field direction within the @xmath43-@xmath44 plane can detect both the position of the gap nodes and the sign change of the pairing potential .
one can estimate the flux - flow resistivity by means of microwaves . instead of fabricating a junction , one can obtain the information on the phase of the pair potential by measuring the microwave surface impedance under a rotating magnetic field .
the authors thank t. okada and s. yasuzuka for helpful discussions .
99 t. sakakibara , a. yamada , j. custers , k. yano , t. tayama , h. aoki , and k. machida , j. phys .
. jpn . * 76 * , 051004 ( 2007 ) ; y. matsuda , k. izawa , and i. vekhter , j. phys . : condens .
matter * 18 * , r705 ( 2006 ) .
g. e. volovik , pisma zh .
. fiz . * 58 * , 444 ( 1993 ) [ jetp lett . * 58 * , 455 ( 1993 ) ] ; * 70 * , 601 ( 1999 ) [ * 70 * , 609 ( 1999 ) ] ; n. b. kopnin and g. e. volovik , _ ibid_. * 64 * , 641 ( 1996 ) [ * 64 * , 690 ( 1996 ) ] .
y. nagai and n. hayashi , phys .
* 101 * ( 2008 ) 097001 ; y. nagai , y. ueno , y. kato , and n. hayashi , j. phys .
* 75 * ( 2006 ) 104701 , and references therein .
l. j. buchholtz and g. zwicknagl , phys .
b * 23 * , 5788 ( 1981 ) ; e. v. thuneberg , j. kurkijrvi , and d. rainer , _
ibid_. * 29 * , 3913 ( 1984 ) ; n. hayashi , y. kato , and m. sigrist , j. low temp . phys . *
139 * , 79 ( 2005 ) .
s. yasuzuka , s. uji , h. satsukawa , m. kimata , t. terashima , h. koga , y. yamamura , k. saito , h. akutsu , and j. yamada , physica b * 405 * , s288 ( 2010 ) ; s. yasuzuka , k. saito , s. uji , m. kimata , h. satsukawa , t. terashima , and j. yamada , j. phys .
* 82 * , 064716 ( 2013 ) . | we theoretically investigate the applied magnetic field - angle dependence of the flux - flow resistivity @xmath0 for a uniaxially anisotropic fermi surface .
@xmath1 is related to the quasiparticle scattering rate inside a vortex core , which reflects the sign change in the superconducting pair potential .
we find that @xmath0 is sensitive to the sign change in the pair potential and has its maximum when the magnetic field is parallel to the gap - node direction .
we propose the measurement of the field - angle dependent oscillation of @xmath0 as a phase - sensitive field - angle resolved experiment . |
dimension reduction is ubiquitous in many areas ranging from pattern recognition , clustering , classification , to fast numerical simulation of complicated physical phenomena .
the fundamental question to address is how to approximate a @xmath4-dimensional space by a @xmath5-dimensional one with @xmath6 .
specifically , we are given a set of high - dimensional data @xmath7 \in { \mathbb r}^{m \times n},\ ] ] and the goal is to find its low - dimensional approximation @xmath8 \in { \mathbb r}^{m \times d}\ ] ] with reasonable accuracy .
there are two types of dimension reduction methods .
the first category consists of `` projective '' ones .
these are the linear methods that are _ global _ in nature , and that explicitly transform the data matrix @xmath0 into a low - dimensional one by @xmath9 .
the leading examples are the principal component analysis ( pca ) and its variants .
the methods in the second category act locally and are inherently nonlinear . for each sample in the high - dimensional space
( e.g. each column of @xmath0 ) , they directly find their low - dimensional approximations by preserving certain locality or affinity between nearby points . in this paper , inspired by the reduced basis method ( rbm )
, we propose a linear method called `` reduced basis decomposition ( rbd ) '' .
it is much faster than pca / svd - based techniques .
moreover , its low - dimensional vectors are equipped with error estimator indicating how close they are approximating the high - dimensional data .
rbm is a relative recent approach to speed up the numerical simulation of parametric partial differential equations ( pdes ) @xcite .
it utilizes an offline online computational decomposition strategy to produce surrogate solution ( of dimension @xmath10 ) in a time that is of orders of magnitude shorter than what is needed by the underlying numerical solver of dimension @xmath11 ( called _ truth _ solver hereafter ) .
the rbm relies on a projection onto a low dimensional space spanned by truth approximations at an optimally sampled set of parameter values @xcite .
this low - dimensional manifold is generated by a greedy algorithm making use of a rigorous _ a posteriori _ error bounds for the field variable and associated functional outputs of interest which also guarantees the fidelity of the surrogate solution in approximating the truth approximation .
the rbd method acts in a similar fashion .
given the data matrix @xmath0 as in , it iteratively builds up @xmath2 whose column space approximates that of @xmath0 .
it starts with a randomly selected column of @xmath0 ( or a user input if existent ) . at each step where we have @xmath12 vectors @xmath13
, the next vector @xmath14 is found by scanning the columns of @xmath0 and locating the one whose error of projection into the current space @xmath15 is the largest .
this process is continued until the maximum projection / compression error is small enough or until the limit on the size of the reduced space is reached .
an important feature is an offline - online decomposition that allows the computation of the compression error , and thus the cost of locating @xmath14 , to be independent of ( the potentially large ) @xmath16 .
this paper is organized as follows . in section
[ sec : background ] , we review the background material , mainly the rbm .
section [ sec : rbc ] describes the reduced basis decomposition algorithm and discuss its properties .
numerical validations are presented in section [ sec : numerical ] , and finally some concluding remarks are offered in section [ sec : conclusion ] .
the reduced basis method was developed for use with finite element methods to numerically solve pdes .
we assume , for simplicity , that the problems ( usually parametric partial differential equations ( pde ) ) to simulate are written in the weak form : find @xmath17 in an hilbert space @xmath0 such that @xmath18 where @xmath19 is an input parameter .
these simulations need to be performed for many values of @xmath19 chosen in a given parameter set @xmath20 . in this problem
@xmath21 and @xmath22 are bilinear and linear forms , respectively , associated to the pde ( with @xmath23 and @xmath24 denoting their numerical counterparts ) .
we assume that there is a numerical method to solve this problem and the solution @xmath25 , called the `` truth approximation '' or `` snapshot '' , is accurate enough for all @xmath26 .
the fundamental observation utilized by rbm is that the parameter dependent solution @xmath27 is not simply an arbitrary member of the infinite - dimensional space associated with the pde . instead
, the solution manifold @xmath28 can typically be well approximated by a low - dimensional vector space .
the idea is then to propose an approximation of @xmath29 by @xmath30 where , @xmath31 are @xmath10 @xmath32 pre - computed truth approximations corresponding to the parameters @xmath33 judiciously selected according to a sampling strategy @xcite . for a given @xmath19
, we now solve in @xmath34 for the reduced solution @xmath35 .
the online computation is @xmath36-independent , thanks to the assumption that the ( bi)linear forms are affine @xmath37 and the fact that they can be approximated by affine ( bi)linear forms when they are nonaffine @xcite . hence , the online part is very efficient . in order to be able to `` optimally ''
find the @xmath10 parameters and to assure the fidelity of the reduced basis solution @xmath35 to approximate the truth solution @xmath27 , we need an _ a posteriori _ error estimator @xmath38 which involves the residual @xmath39 and stability information of the bilinear form @xcite . with this estimator , we can describe briefly the classical * greedy algorithm * used to find the @xmath10 parameters @xmath40 and the space @xmath34 .
we first randomly select one parameter value and compute the associated truth approximation .
next , we scan the entire ( discrete ) parameter space and for each parameter in this space compute its rb approximation @xmath41 and the error estimator @xmath42 .
the next parameter value we select , @xmath43 , is the one corresponding to the largest error estimator .
we then compute the truth approximation and thus have a new basis set consisting of two elements .
this process is repeated until the maximum of the error estimators is sufficiently small .
the reduced basis method typically has exponential convergence with respect to the number of pre - computed solutions @xcite .
this means that the number of pre - computed solutions can be small , thus the computational cost reduced significantly , for the reduced basis solution to approximate the finite element solution reasonably well .
the author and his collaborators showed @xcite that it works well even for a complicated geometric electromagnetic scattering problem that efficiently reveals a very sensitive angle dependence ( the object being stealthy with a particular configuration ) .
in this section , we detail our proposed methodology by stating the algorithm , studying the error evaluation , and pinpointing the computational cost .
set @xmath44 , @xmath45 , and @xmath46 a random integer between @xmath47 and @xmath4 . at the heart of the method stated in algorithm [ alg : c_greedy ] is a greedy algorithm similar to that used by rbm .
it builds the reduced space dimension - by - dimension . at each step , the _ greedy _ decision for the best next dimension to pursue in the space corresponding to the data is made by examining an error indicator quantifying the discrepancy between the uncompressed data and the one compressed into the current ( reduced ) space . in the context of the rbm
, we view each column ( or row if we are compressing the row space ) of the matrix as the fine solution of certain ( virtual ) parametric pde with the ( imaginary ) parameter taking a particular value . since this solution is _ explicitly _ given already by the data , the fact that the pde and the parameter are absent does not matter .
once this _ common mechanism _
satisfied by each column ( or row ) is identified , the greedy algorithm still relies on an accurate and efficient estimate quantifying the error between the original data and the compressed one .
this will be the topic of the next subsection . to state the algorithm
, we assume that we are given a data matrix @xmath48 , the largest dimension @xmath49 that the practitioner wants to retain , and a tolerance @xmath50 capping the discrepancy between the original and the compressed data .
the output is the set of bases for the compressed data ( a low - dimensional approximation of the original data ) @xmath51 and the transformation matrix @xmath52 . here
, @xmath53 is the actual dimension of the compressed data . with this output
, we can * compress . * : : we represent any data entry @xmath54 , the @xmath55 column of @xmath56 , by the @xmath55 column of @xmath3 , @xmath57 , with usually @xmath58 . * uncompress . * : : an _ approximation _ of the data is reconstructed by @xmath59 * evaluate the compression of out - of - sample data . * : : given any @xmath60 that is not equal to any column of @xmath0 , its compressed representation in @xmath61 is @xmath62 a critical part to facilitate the greedy algorithm and make the algorithm realistic is an efficient mechanism measuring ( or estimating ) the error @xmath63 under certain norm , @xmath64 , in step @xmath65 of the algorithm . in this work , we are using the @xmath66norm defined as follows . for a given symmetric and positive definite matrix @xmath67 ,
the @xmath66norm of a vector @xmath60 is defined by @xmath68 for @xmath69 being any column of the data matrix @xmath0 and @xmath70 its low - dimensional approximation @xmath71 , it is easy to see that @xmath72 the choice of @xmath73 reflects the criteria of the data compression .
typical examples are : * * identity : * equal weights are assigned to each component of the data entry .
this makes the quality of compression uniform . in this case , the evaluation of is greatly simplified and the algorithm is the fastest as shown below by the numerical results . * * general diagonal matrix : * this setting can be used if part of each data entry needs to be preserved better and other parts can afford less fidelity . * * general spd matrix : * this most general case can be helpful if the goal is to preserve data across different entries anisotropiclly .
the goal is then to evaluate the error through as efficiently as possible for any given @xmath74 .
this is achieved by employing an offline - online decomposition strategy where the @xmath74-independent parts are evaluated beforehand ( offline ) enabling a quick turnaround time for any given @xmath74 encountered online .
the specifics are given in the next subsection .
the offline - online decomposition of the computations and their complexities are as follows . here
, we use @xmath75 to denote the number of nonzero entries of a sparse matrix @xmath73 .
* offline * : : the total cost is of order @xmath76 + * offline mgs * ; ; every basis needs to orthogonalized against the current set of bases .
the total cost is of order @xmath77 .
* offline calculation of errors * ; ; the next basis is located by comparing each column with its compressed version into the current space . to enable that
, we encounter the following computational cost : + * pre - computation * : : of @xmath78 ( for @xmath79 in ) and @xmath80 ( for @xmath81 in ) .
the cost is of order @xmath82 . * expansion * : : of @xmath83 and @xmath84 .
the former takes time of order @xmath85 , and the latter of order @xmath86 . * offline searching * ; ; after these calculations , the comparison between the original and compressed data is then only dependent on the size of @xmath74 ( which is also the number of columns for @xmath2 ) .
the complexity is of order @xmath87 .
it will be repeated for up to @xmath4 times in the searching process of step 2.3 of the algorithm for each of the up to @xmath88 basis elements .
the total cost is at the level of @xmath89 . *
online * : : given any ( possibly out - of - sample ) data @xmath60 , its coefficients in the compressed space is obtained by evaluating @xmath90 .
the cost is of order @xmath91 .
the decoding ( @xmath92 ) can be done with the same cost .
the online computation has complexity of order @xmath93 we remark that , if the actual practice does not requires forming @xmath70 ( e.g. clustering and classification etc ) and so we only work with the coordinates @xmath74 of @xmath69 in the compressed space , then the online cost will be independent of @xmath16 and thus much smaller .
in this section , we test the reduced basis decomposition on image compression , and data compression .
lastly , we devise a simple face recognition algorithm based on rbd and test it on a database of @xmath94 images while comparing rbd with @xmath95 other face recognition algorithms
. the computation is done , and thus the speedup numbers reported herein should be understood as , in matlab 2014a on a 2011 imac with a @xmath96 ghz intel core i7 processor . [
sec : numerical ] we first test it on compressing two standard images lena and mandrill in figure [ fig : original ] .
they both have an original resolution of @xmath97 .
we take @xmath98 and test the algorithm . for each component of every image
, we run the algorithm with @xmath99 which implies a compression ratio of @xmath100 , @xmath101 , and @xmath102 respectively .
the resulting images ( formed by multiplying the corresponding @xmath2 and @xmath3 together ) are shown on the @xmath103 and @xmath104 row of figures [ fig : lenaresult ] . as a comparison ,
we run svd and obtain the reconstructed matrices with the first @xmath88 singular values accordingly .
the resulting images are on the second and last row .
clearly , svd provides the best quality pictures among all possible algorithms ( and thus better than what rbd provides ) .
however , we see that the rbd pictures are only slightly blurrier .
moreover , it takes much less time .
in fact , we show the comparison in time between svd and rbd in table [ tab : lenatime ] .
we see that , when @xmath105 , rbd is three times faster than svd and seven times faster when @xmath106 . here
the svd time is the shorter between those taken by ` vd } and { \verb ` vds commands in matlab . from left to right .
the first and third row are from reduced basis decomposition , and the second and fourth are from singular value decomposition.,title="fig:",scaledwidth=32.0% ] from left to right . the first and third row
are from reduced basis decomposition , and the second and fourth are from singular value decomposition.,title="fig:",scaledwidth=32.0% ] from left to right .
the first and third row are from reduced basis decomposition , and the second and fourth are from singular value decomposition.,title="fig:",scaledwidth=32.0% ] + from left to right . the first and third row
are from reduced basis decomposition , and the second and fourth are from singular value decomposition.,title="fig:",scaledwidth=32.0% ] from left to right . the first and third row
are from reduced basis decomposition , and the second and fourth are from singular value decomposition.,title="fig:",scaledwidth=32.0% ] from left to right .
the first and third row are from reduced basis decomposition , and the second and fourth are from singular value decomposition.,title="fig:",scaledwidth=32.0% ] + from left to right . the first and third row
are from reduced basis decomposition , and the second and fourth are from singular value decomposition.,title="fig:",scaledwidth=32.0% ] from left to right . the first and third row
are from reduced basis decomposition , and the second and fourth are from singular value decomposition.,title="fig:",scaledwidth=32.0% ] from left to right . the first and third row
are from reduced basis decomposition , and the second and fourth are from singular value decomposition.,title="fig:",scaledwidth=32.0% ] + from left to right . the first and third row
are from reduced basis decomposition , and the second and fourth are from singular value decomposition.,title="fig:",scaledwidth=32.0% ] from left to right .
the first and third row are from reduced basis decomposition , and the second and fourth are from singular value decomposition.,title="fig:",scaledwidth=32.0% ] from left to right . the first and third row
are from reduced basis decomposition , and the second and fourth are from singular value decomposition.,title="fig:",scaledwidth=32.0% ] .relative computational time for image compression . [ cols="^,^,^",options="header " , ] we use the umist database @xcite that is publicly available on roweis web page .
table [ tab : datasetinfo ] summarizes its characteristics : it contains @xmath107 people under different poses .
the number of different views per subject varies from @xmath108 to @xmath109 .
we use the cropped version whose snapshot is shown in figure [ fig : umist ] . as in @xcite , we randomly choose @xmath110 views from each class to form a training set .
the rest of the samples ( @xmath111 of them ) are used as testing images .
we show the average classification error rates in figure [ fig : fr_result ] left .
these averages are computed over @xmath112 random formations of the training and test sets . shown in the middle
are the results of six traditional dimension reduction techniques taken from @xcite .
clearly , our method has similar performance as the pca method , and outperforms three of the other five methods . however , rbd is much faster than pca and other methods since they all involves solving eigenproblems @xcite .
a speedup factor as a function of the number of bases is plotted in figure [ fig : fr_result ] right which demonstrates a speedup factor of larger than two for this particular test when we reach the asymptotic region ( around when the number of basis vectors is @xmath113 ) .
this paper presents and tests an extremely efficient dimension reduction algorithm for data processing .
it is multiple times faster than the svd / pca - based algorithms .
what makes this possible is a greedy algorithm that iteratively builds up the reduced space of basis vectors .
each time , the next dimension is located by exploring the errors of compression into the current space for all data entries .
thanks to an offline - online decomposition mechanism , this searching is independent of the size of each entry . numerical results including one concerning a real world face recognition problem confirm these findings .
m. barrault , n. c. nguyen , y. maday , and a. t. patera , _ an `` empirical interpolation '' method : application to efficient reduced - basis discretization of partial differential equations _ , c. r. acad .
paris , srie i * 339 * ( 2004 ) , 667672 .
a. buffa , y. maday , a. t. patera , c. prudhomme , and g. turinici , _ a priori convergence of the greedy algorithm for the parametrized reduced basis _ , esaim - math . model .
( 2011 ) , special issue in honor of david gottlieb .
j. p. fink and w. c. rheinboldt , _ on the error behavior of the reduced basis technique for nonlinear finite element approximations _ , z. angew .
* 63 * ( 1983 ) , no . 1 , 2128 .
mr mr701832 ( 85e:73047 ) d. b graham and n. m allinson , _ characterizing virtual eigensignatures for general purpose face recognition _ , face recognition : from theory to applications ( h. wechsler , p. j. phillips , v. bruce , f. fogelman - soulie , and t. s. huang , eds . ) , nato asi series f , computer and systems sciences , vol .
163 , 1998 , pp . 446456 .
m. a. grepl , y. maday , n. c. nguyen , and a. t. patera , _ efficient reduced - basis treatment of nonaffine and nonlinear partial differential equations _
, mathematical modelling and numerical analysis * 41 * ( 2007 ) , no . 3 , 575605 .
l. machiels , y. maday , i. b. oliveira , a. t. patera , and d. v. rovas , _ output bounds for reduced - basis approximations of symmetric positive definite eigenvalue problems _ , c. r. acad .
* 331 * ( 2000 ) , no . 2 , 153158
. mr mr1781533 ( 2001d:65148 ) y. maday , _ reduced basis method for the rapid and reliable solution of partial differential equations _ , international congress of mathematicians . vol .
iii , eur .
soc . , zrich , 2006 , pp .
mr 2275727 ( 2007m:65099 ) y. maday , a. t. patera , and d. v. rovas , _ a blackbox reduced - basis output bound method for noncoercive linear problems _ , nonlinear partial differential equations and their applications .
collge de france seminar , vol .
xiv ( paris , 1997/1998 ) , stud .
31 , north - holland , amsterdam , 2002 , pp .
mr mr1936009 ( 2003j:65120 ) y. maday , a. t. patera , and g. turinici , _ a priori convergence theory for reduced - basis approximations of single - parameter elliptic partial differential equations _
* 17 * ( 2002 ) , 437446 .
nguyen , k. veroy , and a. t. patera , _ certified real - time solution of parametrized partial differential equations _
, handbook of materials modeling ( sidney yip , ed . ) , springer netherlands , 2005 , pp .
15291564 ( english ) . c. prudhomme , d. rovas , k. veroy , y. maday , a. t. patera , and g. turinici , _ reliable real - time solution of parametrized partial differential equations : reduced - basis output bound methods _ , journal of fluids engineering * 124 * ( 2002 ) , no . 1 , 7080 . g. rozza , d.b.p .
huynh , and a.t .
patera , _ reduced basis approximation and a posteriori error estimation for affinely parametrized elliptic coercive partial differential equations : application to transport and continuum mechanics _ , arch comput methods eng * 15 * ( 2008 ) , no . 3 , 229275 .
s. sen , k. veroy , d.b.p .
huynh , s. deparis , n.c .
nguyen , and a.t .
`` natural norm '' a posteriori error estimators for reduced basis approximations _ , j. comput .
* 217 * ( 2006 ) , no . 1 , 37 62 . | dimension reduction is often needed in the area of data mining .
the goal of these methods is to map the given high - dimensional data into a low - dimensional space preserving certain properties of the initial data .
there are two kinds of techniques for this purpose .
the first , projective methods , builds an explicit linear projection from the high - dimensional space to the low - dimensional one . on the other hand ,
the nonlinear methods utilizes nonlinear and implicit mapping between the two spaces . in both cases ,
the methods considered in literature have usually relied on computationally very intensive matrix factorizations , frequently the singular value decomposition ( svd ) .
the computational burden of svd quickly renders these dimension reduction methods infeasible thanks to the ever - increasing sizes of the practical datasets . in this paper , we present a new decomposition strategy , reduced basis decomposition ( rbd ) , which is inspired by the reduced basis method ( rbm ) .
given @xmath0 the high - dimensional data , the method approximates it by @xmath1 with @xmath2 being the low - dimensional surrogate and @xmath3 the transformation matrix .
@xmath2 is obtained through a greedy algorithm thus extremely efficient .
in fact , it is significantly faster than svd with comparable accuracy .
@xmath3 can be computed on the fly . moreover , unlike many compression algorithms , it easily finds the mapping for an arbitrary `` out - of - sample '' vector and it comes with an `` error indicator '' certifying the accuracy of the compression .
numerical results are shown validating these claims
. data mining , lossy compression , reduced basis method , singular value decomposition , greedy algorithm |
the exploration of cp violation , which was discovered in 1964 through the observation of @xmath20 decays @xcite , is one of the hot topics in particle physics . in this decade , studies of @xmath0-meson decays will provide various stringent tests of the kobayashi maskawa mechanism of cp violation @xcite , allowing us to accommodate this phenomenon in the standard model of electroweak interactions . at present , the experimental stage is governed by the asymmetric @xmath21 @xmath0 factories operating at the @xmath22 resonance , with their detectors babar ( slac ) and belle ( kek ) .
interesting @xmath0-physics results are also soon expected from run ii of the tevatron @xcite .
several strategies can only be fully exploited in the era of the lhc @xcite , in particular by lhcb ( cern ) and btev ( fermilab ) .
= 5.0truecm the central target for the exploration of cp violation is the unitarity triangle of the cabibbo
maskawa ( ckm ) matrix illustrated in fig .
[ fig : ut ] , which can be determined through measurements of its sides on the one hand , and direct measurements of its angles through cp - violating effects in @xmath0 decays on the other hand ( for a detailed review , see @xcite ) .
if we take into account next - to - leading order terms in the wolfenstein expansion of the ckm matrix @xcite , following the prescription given in @xcite , we obtain the following coordinates for the apex of the unitarity triangle : @xmath23 where @xmath24 .
the goal is to overconstrain the unitarity triangle as much as possible . in this respect ,
a first important milestone has already been achieved by the babar and belle collaborations , who have announced the observation of cp violation in the @xmath0-meson system in the summer of 2001 @xcite .
this discovery was made possible by @xmath25 and similar channels .
as is well known , these modes allow us to determine @xmath26 , where @xmath27 denotes the cp - violating weak @xmath28@xmath29 mixing phase , which is given by @xmath30 in the standard model . using the present world average @xmath31 @xcite ,
we obtain the following twofold solution for @xmath27 itself : @xmath32 where the former would be in perfect agreement with the `` indirect '' range implied by the ckm fits , @xmath33 @xcite , while the latter would correspond to new physics .
the two solutions can be distinguished through a measurement of the sign of @xmath34 .
several strategies to accomplish this important task were proposed @xcite .
for example , in the @xmath35 system , @xmath36 can be extracted with the help of the time - dependent angular distribution of the @xmath37 k^\ast[\to\pi^0k_{\rm s}]$ ] decay products if the sign of a hadronic parameter @xmath38 , involving a strong phase @xmath39 , is fixed through factorization @xcite .
this analysis is already in progress at the @xmath0 factories @xcite .
a remarkable link between the two solutions for @xmath27 and ranges for @xmath15 is also suggested by recent data on the cp - violating @xmath40 observables @xcite .
interesting tools for the exploration of cp violation are provided by @xmath3 , @xmath4 and @xmath6 , @xmath7 modes , which originate from @xmath41 and @xmath42 transitions , respectively , and receive only contributions from tree - diagram - like topologies @xcite@xcite .
the neutral @xmath9 mesons may be observed through decays into cp eigenstates @xmath12 , or through decays into cp non - eigenstates @xmath13 . in a recent paper @xcite , we pointed out that @xmath43 and @xmath44 decays , where @xmath45 and @xmath46 denote the cp - even and cp - odd eigenstates of the neutral @xmath9-meson system , respectively , offer very efficient , theoretically clean determinations of the angle @xmath15 of the unitarity triangle . in this strategy , only `` untagged '' and mixing - induced cp - violating observables are employed , which satisfy a very simple relation , allowing us to determine @xmath47 . using a plausible dynamical assumption ,
@xmath15 can be fixed in an essentially unambiguous manner .
the corresponding formalism can also be applied to @xmath48 and @xmath49 decays .
although these modes appear less attractive for the extraction of @xmath15 , they provide interesting determinations of @xmath50 . in comparison with the conventional @xmath51 and @xmath52 methods , these extractions
do not suffer from any penguin uncertainties , and are theoretically cleaner by one order of magnitude . in the present paper ,
we give a more detailed discussion of the @xmath53 case by having a closer look at @xmath10 processes , and focus on the cp non - eigenstate @xmath11 case . here
we have to deal with interference effects between cabibbo - favoured and doubly cabibbo - suppressed @xmath9-meson decays , which lead to certain complications @xcite .
we point out that @xmath15 can nevertheless be determined in a powerful and essentially unambiguous way from the time - dependent @xmath54k_{\rm s(l)}$ ] , @xmath55\eta^ { ( ' ) } , d[\to f_{\rm ne}]\phi , ... $ ] rates with the help of simple `` untagged '' measurements of the corresponding @xmath43 , @xmath44 processes .
moreover , we may also extract interesting hadronic @xmath0- and @xmath9-decay parameters .
this strategy can be implemented with the help of transparent formulae in an _ analytical _ manner .
decays of the kind @xmath56 , @xmath57 also exhibit interesting features , but they appear not to be as attractive for the extraction of @xmath15 as their @xmath58 counterparts , since they have a different ckm structure . for the testing of the kobayashi
maskawa mechanism of cp violation @xcite , it is essential to measure @xmath15 in a variety of ways .
in particular , it will be very interesting to compare the theoretically clean values for @xmath15 obtained with the help of the new strategies proposed here with the results that will be provided by other approaches . in this context , @xmath59 and @xmath60 ,
@xmath40 modes are important examples , allowing the extraction of @xmath15 through @xmath61 flavour - symmetry arguments ( for overviews , see @xcite ) .
several theoretically clean strategies were also proposed , employing various different pure tree - diagram - like @xmath0 decays ( see , for instance , @xcite and @xcite@xcite ) . in general , these approaches are either conceptually simple but experimentally very challenging , or do not allow transparent extractions of @xmath15 through simple analytical expressions , thereby also complicating things considerably and yielding multiple discrete ambiguities .
alternative approaches to extract weak phases from neutral @xmath58 modes can be found in @xcite@xcite .
the outline of this paper is as follows : in section [ sec : evol - ampl ] , we introduce a convenient notation and discuss the time evolution and amplitude structure of neutral @xmath0-meson decays caused by @xmath41 and @xmath62 transitions . in section [ sec : cp ]
, we then turn to the case where the neutral @xmath9 mesons are observed through decays into cp eigenstates @xmath12 , whereas we shall focus on the interesting opposite case , where the @xmath9 mesons are observed through decays into cp non - eigenstates @xmath13 , in section [ sec : cpne ] .
the main approach of this paper is presented in section [ sec : strat ] , and is also illustrated there with the help of a few instructive examples . finally , we conclude in section [ sec : concl ] .
= 5.0truecm = 5.0truecm
the decays we shall consider in the following analysis may be written generically as @xmath63 , @xmath64 , where @xmath65 and @xmath66 , as can be seen in fig .
[ fig : bqdf ] .
note that the @xmath67 quark is contained in @xmath68 ; in order not to complicate the notation unnecessarily , we do not indicate this explicitly through another label @xmath67 that should , in principle , be added to @xmath68 .
if we require that @xmath68 be a cp - self - conjugate state , i.e. satisfies @xmath69 @xmath70 and @xmath71 may both decay into the same state @xmath72 , thereby leading to interference effects between @xmath70@xmath71 mixing and decay processes .
let us illustrate this abstract but quite efficient notation by giving a few examples : * @xmath73 : @xmath3 @xmath74 ; @xmath4 @xmath75 . * @xmath76 : @xmath6 @xmath75 ; @xmath7 @xmath74 .
all these decays are governed by the colour - suppressed tree - diagram - like topologies shown in fig .
[ fig : bqdf ] . in the case of @xmath77 ,
also exchange topologies may contribute .
however , these contributions do not modify the phase structure of the corresponding decay amplitudes discussed in subsection [ subsec : ampl ] , since they enter with the same ckm factors as the colour - suppressed tree topologies .
the @xmath78 mode has already been observed by the belle , cleo and babar collaborations , with branching ratios at the @xmath79 level @xcite .
interestingly , also the observation of the @xmath80 channel was very recently announced by the belle collaboration @xcite , measuring the branching ratio @xmath81 .
if both a @xmath70 and a @xmath71 meson may decay into the same final state @xmath82 , as is the case for the modes specified above , we obtain the following time - dependent rate asymmetry @xcite : @xmath83,\label{ee6}\end{aligned}\ ] ] where @xmath84 and @xmath85 denote the time - dependent decay rates for initially , i.e. at time @xmath86 , present @xmath70 and @xmath71 states , and @xmath87 and @xmath88 are the mass and decay width differences of the @xmath89 mass eigenstates , respectively .
the observables @xmath90 and @xmath91 are given by @xmath92 respectively , where @xmath93 measures the strength of the interference effects between the @xmath70@xmath71 mixing and decay processes . here
@xmath94 is the cp - violating weak @xmath70@xmath71 mixing phase , which is within the standard model negligibly small in the @xmath95-meson system ; the convention - dependent phase @xmath96 is defined through @xmath97 as we will see below , @xmath96 is cancelled through the amplitude ratio in ( [ xi - def ] ) .
the width difference @xmath98 provides another observable , @xmath99 which is , however , not independent from @xmath90 and @xmath91 , satisfying @xmath100 ^ 2 + \bigl[s(b_q\to f)\bigr]^2 + \bigl[{\cal a}_{\delta\gamma}(b_q\to f)\bigr]^2=1.\ ] ] this observable could be determined from the `` untagged '' rate @xmath101\nonumber\\ & & \times\left[\cosh(\delta\gamma_qt/2)-{\cal a}_{\rm \delta\gamma}(b_q\to f ) \,\sinh(\delta\gamma_qt/2)\right]e^{-\gamma_qt},\label{untagged}\end{aligned}\ ] ] where the oscillatory @xmath102 and @xmath103 terms cancel , and @xmath104 denotes the average decay width @xcite .
however , in the case of the @xmath105-meson system , the width difference is negligibly small , so that the time evolution of ( [ untagged ] ) is essentially given by the well - known exponential @xmath106 . on the other hand , the width difference @xmath107 of the @xmath95-meson system may be as large as @xmath108 ( for a recent review , see @xcite ) . in the discussion given below
, we shall see as in @xcite that observables provided by the untagged rates specified in ( [ untagged ] ) are a key element for an efficient determination of @xmath15 from @xmath109 modes . in contrast to the strategies proposed in @xcite , we do _ not _ have to rely on a sizeable width difference @xmath98 ; the untagged measurements are only required to extract the `` unevolved '' , i.e. time - independent , untagged rates @xmath110 consequently , we will not discuss the effects that may be caused by a possible sizeable value of @xmath111 in further detail .
they could be taken into account straightforwardly .
concerning the time evolution of the @xmath112 modes analysed in this paper , the subsequent decay of the neutral @xmath9-meson state plays a crucial rle .
let us assume that this decay , @xmath113 , occurs at time @xmath114 after the decay of the @xmath89 meson into @xmath115 . in this case , we have to consider decay amplitudes of the kind @xmath116f_r),\ ] ] and have in principle to deal with @xmath117@xmath118 mixing for @xmath119 .
however , within the standard model , this phenomenon is very small .
neglecting hence the mixing effects , the time evolution of the neutral @xmath120 states is given by @xmath121f_r)=e^{-\gamma_dt ' } \gamma(b^0_q(t)\to d[\to f_d]f_r),\ ] ] where @xmath122f_r)$ ] corresponds to the rates entering ( [ ee6 ] ) , and @xmath123 denotes the decay width of the neutral @xmath9 mesons . as is well known , @xmath117@xmath118 mixing may be enhanced through new physics .
should this actually be the case , the mixing effects could be included through a measurement of the @xmath117@xmath118 mixing parameters @xcite .
note that @xmath117@xmath118 mixing does not lead to any effects in the case of @xmath124 , even if this phenomenon should be made sizeable by the presence of new physics . in the following considerations
, we shall neglect @xmath117@xmath118 mixing for simplicity . moreover
, we shall assume that cp violation in @xmath9 decays is also negligible , as in the standard model .
in analogy to @xmath117@xmath118 mixing , these effects could also be enhanced through new physics . in the following sections
, we will therefore emphasize clearly whenever this assumption enters .
note that also the tiny `` indirect '' cp violation in the neutral kaon system has at least in principle to be taken into account if @xmath125 or @xmath126 states are involved .
this could be done through the corresponding experimental @xmath127-decay data .
however , it appears very unlikely that the experimental accuracy of the strategies for the determination of @xmath15 discussed below will ever reach a level where these effects may play any rle from a practical point of view . in order to analyse the structure of the @xmath128 decay amplitudes , we employ appropriate low - energy effective hamiltonians
@xcite : @xmath129\label{heff - bar}\\ { \cal h}_{\mbox{{\scriptsize eff}}}(b^0_q\to d^0f_r)&= & \frac{g_{\rm f}}{\sqrt{2}}\,v_r^\ast \left[{\cal o}_1^{r\dagger}\ , { c}_1(\mu)+{\cal o}_2^{r\dagger}\ , { c}_2(\mu)\right],\label{heff}\end{aligned}\ ] ] where @xmath130 and @xmath131 denote current current operators , which are given by @xmath132 and @xmath133 @xmath134 are ckm factors , with ( for the numerical values , see @xcite ) @xmath135 @xmath136 using ( [ heff - bar ] ) , we may write @xmath137 where @xmath138 on the other hand , we have @xmath139 } \frac{g_{\rm f}}{\sqrt{2}}\,v_r^\ast m_{f_r},\end{aligned}\ ] ] where @xmath140 here we have employed @xmath141 @xmath142 as well as @xmath143 where @xmath144 denotes the angular momentum of the @xmath72 state , and @xmath145 in analogy to ( [ b - phase ] ) . in the following , we shall use the abbreviation @xmath146 for convenience , we have collected in table [ tab : cp ] the values of @xmath144 , @xmath147 and @xmath148 for various decay channels of the kind specified in subsection [ subsec : not ] .
.the values of @xmath144 , @xmath147 and @xmath148 for various @xmath149 modes . [ cols="<,<,<,<,<",options="header " , ] we observe that all observables in table [ tab : illu1 ] take large values , which is very promising from an experimental point of view .
moreover , @xmath150 and the observable combinations @xmath151 and @xmath152 , which play the key rle for the determination of @xmath15 and @xmath153 , respectively , are also favourably large . on the other hand , the observables @xmath154 , which allow the extraction of @xmath155 , are suppressed by the small deviation of this phase from @xmath156 .
consequently , in order to be able to probe @xmath157 , the experimental resolution has to reach the @xmath158 level , where we then must care about the @xmath159 corrections . in our examples , the observables
@xmath160 , allowing us to probe @xmath161 , take values at the @xmath162 level . for completeness , we have also included the values for the observables @xmath163 , which are negligibly small .
fortunately , they are not required for the strategy discussed above .
the first interesting information on @xmath15 can already be obtained from @xmath150 . using ( [ gam - bound ] ) , we obtain @xmath164 and @xmath165 for @xmath166 and @xmath167 , respectively , where we have also taken into account ( [ sgn - cos - gam ] ) .
the measurement of @xmath168 then allows us to _ determine _ the angle @xmath15 . in fig .
[ fig : splus - bd1 ] , we show the situation for @xmath169 . here
the horizontal dot - dashed line represents the `` measured '' value of @xmath170 , and the thin solid contours were calculated with the help of ( [ s - lo ] ) , neglecting the @xmath159 corrections .
we observe that the intersection with the horizontal line yields a twofold solution for @xmath15 , which is given by @xmath171 .
these values can also be obtained directly in a simple manner with the help of ( [ tan - gam - lo ] ) . using ( [ sgn - cos - gam ] )
, the negative value of @xmath150 implies @xmath172 . as @xmath173 and @xmath174
, we may hence exclude the @xmath175 solution .
alternatively , we could already restrict the range for @xmath15 in fig .
[ fig : splus - bd1 ] to @xmath164 , which corresponds to the @xmath150 bound ( [ gam - bound ] ) discussed above . however , it is more instructive for our purpose to consider the whole @xmath15 range in this figure .
it is interesting to note that @xmath176 differs by only @xmath177 from our input value of @xmath178 . in order to take into account the @xmath159 corrections ,
the contours in the @xmath15@xmath170 plane are a key element , as we have seen above . using ( [ s - nlo ] )
, we obtain the thick solid contours shown in fig .
[ fig : splus - bd1 ] .
taking again into account the @xmath172 constraint , their intersection with the horizontal dot - dashed line gives the single solution @xmath179 , which is in perfect agreement with our input value . in table
[ tab : illu3 ] , we collect also the results for the extracted hadronic parameters .
it is interesting to note that the value of @xmath161 determined from ( [ rd - sind ] ) is in good agreement with our input parameter .
moreover , the value of @xmath153 extracted from ( [ x - det - lo ] ) agrees also well with the `` true '' one of 0.4 , whereas the value of @xmath155 following from ( [ tan - del - det - lo ] ) suffers from large @xmath159 corrections .
however , taking them into account through ( [ smnlo ] ) , we obtain a result that differs by only @xmath180 from our input value of @xmath181 . the interesting feature
that already the simple expressions ( [ tan - gam - lo ] ) and ( [ x - det - lo ] ) give values of @xmath15 and @xmath153 that are very close to their `` true '' values is due to the fact that the relevant observables , @xmath182 and @xmath183 , are much larger than @xmath184 . on the other hand , this is not the case for @xmath185 , which plays the key rle for the extraction of @xmath155 . in figs .
[ fig : splus - bd2 ] and [ fig : splus - bs ] , we repeat this exercise for @xmath186 and @xmath187 , respectively . these figures are complemented by the numerical values given in table [ tab : illu3 ] . the @xmath95-meson case shown in fig .
[ fig : splus - bs ] is particularly interesting , as the @xmath159 effects would there be essentially negligible , and would also affect , to a smaller extent , the extraction of @xmath155 . a remarkable feature of the contours shown in figs . [ fig : splus - bd1][fig : splus - bs ]
is that already measurements of the @xmath168 observables with rather large uncertainties would allow us to obtain impressive determinations of @xmath15 .
let us now finally turn to the @xmath76 case , i.e. to decays of the kind @xmath6 and @xmath7 , where the interference effects between the @xmath188 and @xmath189 decay paths are doubly cabibbo - suppressed , as reflected by @xmath190 as was pointed out in @xcite , the @xmath191 case provides interesting determinations of @xmath50 , which are in comparison with the conventional @xmath51 and @xmath52 methods theoretically cleaner by one order of magnitude ( see also subsection [ subsec : appl - cp ] ) . let us here focus again on the case where the neutral @xmath9-mesons are observed through decays into cp non - eigenstates , having in particular @xmath192 in mind .
because of ( [ x - f - d ] ) and @xmath193 , we keep only the leading - order terms in @xmath194 and @xmath159 , yielding @xmath195 + 2\,r_d\cos\delta_d\sin\phi_q + ... \label{spd}\\ \eta_{f_d}\langle \tilde s_{f_d } \rangle_-&=-&\gamma_{+-}^{f_d } \left[\cos\phi_q-\tan\gamma\sin\phi_q\right]\tan\delta_{f_d}- 2\,r_d\sin\delta_d\cos\phi_q + ... , \label{smd}\end{aligned}\ ] ] where the dots are an abbreviation for the neglected @xmath196 , @xmath197 and @xmath198 terms .
note that these observables are expected to take small values , which are not promising from an experimental point of view .
moreover , we have @xmath199 so that this quantity does no longer allow us to fix @xmath200 and @xmath201
. however , we may instead consider both a @xmath95 and a @xmath105 mode , for example @xmath55k_{\rm s(l)}$ ] and @xmath54\rho^0 $ ] . applying ( [ spd ] ) to these channels and assuming that @xmath202 and @xmath27 are known unambiguously
, their observables @xmath203 allow us to determine straightforwardly @xmath47 and @xmath201 .
furthermore , we may then apply ( [ smd ] ) to extract the corresponding strong phases @xmath204 , and may use @xmath205 to determine the hadronic parameters @xmath194 .
note that we have @xmath206 so that these observables do not provide useful information in this case .
if @xmath202 should take its negligibly small standard - model value , implying @xmath207 and @xmath208 , ( [ spd ] ) would simplify even further : @xmath209 thereby allowing a direct determination of @xmath47 .
interestingly , because of ( [ x - f - d ] ) , the @xmath210f_d$ ] observables may be dominated by the @xmath159 terms . in this case , ( [ spd ] ) and ( [ smd ] ) would lose their sensitivity on @xmath15 , but would instead allow simple determinations of @xmath201 and @xmath200 , respectively , thereby yielding @xmath159 and @xmath161 by using only @xmath0-decay measurements .
these parameters could then also be employed as an input for the @xmath210f_s$ ] analysis discussed above .
in this paper , which complements @xcite , we have performed a comprehensive analysis of the @xmath112 modes , where the neutral @xmath9 mesons may be observed through decays into cp eigenstates @xmath12 , or into cp non - eigenstates @xmath13 . after a detailed discussion of the formalism to calculate the observables provided by the @xmath211 f_r$ ] and @xmath210 f_r$ ] transitions , we have focused on the latter case , since the exploration of cp violation with @xmath212 modes was already discussed in detail in @xcite . because of interference effects between cabibbo - favoured and doubly cabibbo - suppressed @xmath9-meson decays , which are described by a parameter @xmath159 determined experimentally from @xmath9-decay measurements , the situation is more complicated in @xmath210 f_r$ ] .
however , as we have pointed out in this paper , we may nevertheless extract @xmath15 and interesting hadronic @xmath0- and @xmath9-decay parameters in an efficient and essentially unambiguous manner : * for the extraction of @xmath15 , the @xmath73 modes , i.e. the @xmath3 , @xmath4 channels , play the key rle , as certain interference effects between the @xmath213 and @xmath214 decay paths are employed to this end . whereas these interference effects are governed by @xmath215 , i.e. are favourably large , their @xmath76 counterparts are doubly cabibbo - suppressed . in order to determine @xmath15
, we use the @xmath70@xmath71 mixing phase @xmath216 as an input , which can be determined through well - known strategies . * an interesting bound on @xmath15 can already be obtained from the observable @xmath150 , as @xmath217 .
complementing @xmath150 through a measurement of the mixing - induced observable @xmath218 , @xmath15 can be determined in an elegant manner with the help of a simple relation .
measuring , in addition , @xmath219 , we may also extract the hadronic parameter @xmath153 .
the theoretical accuracy of these efficient determinations of @xmath15 and @xmath153 is limited by @xmath220 corrections . however , since the relevant observables @xmath218 and @xmath219 are expected to be large , which is also very promising from an experimental point of view , these effects should play a minor rle , as we have also seen in our numerical studies . on the other hand , since the strong phase @xmath155 is expected to be close to @xmath156 , its determination from the correspondingly suppressed observable @xmath221 would require the inclusion of the @xmath159 corrections .
* following the strategy proposed in this paper , the @xmath159 effects can be taken into account straightforwardly with the help of the observables @xmath160 , @xmath221 and @xmath219 . to this end , it is very convenient to consider contours in the @xmath15@xmath218 plane . in the case of our numerical examples , these contours would allow impressive determinations of @xmath15 , even if the experimental value for @xmath218 should suffer from a sizeable uncertainty , and the @xmath159 corrections would have a small impact .
the observables @xmath219 , @xmath221 and @xmath160 provide interesting extractions of @xmath153 , @xmath155 and @xmath161 .
the numerical studies suggest , furthermore , that the @xmath95-meson decays @xmath4 may be particularly promising , thereby representing a nice new playground for hadronic @xmath0 experiments of the lhc era
. moreover , there are various interesting ways to combine the information provided by @xmath95- and @xmath105-meson decays . *
if we assume that @xmath216 will be known unambiguously by the time these measurements can be performed in practice , our strategy gives a twofold solution @xmath222 , with @xmath223 $ ] and @xmath224 $ ] .
using the plausible assumption that @xmath225 , the sign of @xmath150 allows us to distinguish between these solutions , thereby fixing @xmath15 _ unambiguously_. in particular , we may check whether this angle is actually smaller than @xmath156 , as implied by the standard - model interpretation of @xmath226 . on the other hand , if we assume as is usually done that @xmath227 $ ] ,
in accordance with @xmath226 , a knowledge of @xmath50 is sufficient for our strategy .
the assumption of @xmath225 then allows us to determine the sign of @xmath228 , thereby fixing @xmath216 unambiguously , and to obtain also an unambiguous value of @xmath15 . following these lines
, we could distinguish between the two solutions @xmath229 and @xmath230 , which are suggested by a recent analysis of cp violation in @xmath40 .
similar comments apply to the @xmath231 analysis performed in @xcite . *
although the @xmath76 modes @xmath232 and @xmath233 provide interesting determinations of @xmath234 and @xmath26 , respectively , as pointed out in @xcite , decays of this kind are not very attractive for the determination of @xmath15 . in the case of @xmath210 f_d$ ] , we have furthermore to deal with the @xmath159 effects .
it is nevertheless important to make efforts to measure the corresponding observables .
in particular , they may even be dominated by the @xmath220 terms , and would then allow us to extract @xmath159 and @xmath161 in a very simple manner .
these quantities could then be used as an input for the extraction of @xmath15 from the @xmath210 f_s$ ] channels . following these lines
, @xmath128 decays provide very interesting tools to explore cp violation .
moreover , we may obtain valuable insights into the hadron dynamics of the corresponding @xmath0- and @xmath9-meson decays .
consequently , we strongly encourage detailed experimental studies .
since we are dealing with colour - suppressed @xmath0 decays , exhibiting branching ratios at the @xmath235 level in the @xmath73 case , these strategies appear to be particularly interesting for the next generation of dedicated @xmath0 experiments , lhcb and btev , as well as those at super-@xmath0 factories .
however , first important steps may already be achieved at the present @xmath0-decay facilities , i.e. at babar , belle and run ii of the tevatron .
_ , _ phys . rev
. lett . _ * 13 * ( 1964 ) 138 . m. kobayashi and t. maskawa , _ prog
. theor .
phys . _ * 49 * ( 1973 ) 652
. k. anikeev _
et al . _ , fermilab - pub-01/197 [ hep - ph/0201071 ] .
et al . _ , cern - th/2000 - 101 [ hep - ph/0003238 ] .
r. fleischer , _ phys .
rep . _ * 370 * ( 2002 ) 537 .
l. wolfenstein , _ phys .
_ * 51 * ( 1983 ) 1945
buras , m.e .
lautenbacher and g. ostermaier , _ phys .
_ * d50 * ( 1994 ) 3433 .
babar collaboration ( b. aubert _ et al .
_ ) , _ phys . rev .
lett . _ * 87 * ( 2001 ) 091801 ; + belle collaboration ( k. abe _
_ ) , _ phys .
lett . _ * 87 * ( 2001 ) 091802
. y. nir , wis-35 - 02-dpp [ hep - ph/0208080 ] .
buras , tum - hep-435 - 01 [ hep - ph/0109197 ] ; + a. ali and d. london , _ eur .
j. _ * c18 * ( 2001 ) 665 ; + d. atwood and a. soni , _ phys .
_ * b508 * ( 2001 ) 17 ; + m. ciuchini _ et al .
_ , _ jhep _ * 0107 * ( 2001 ) 013 ; + a. hcker _
_ , _ eur .
j. _ * c21 * ( 2001 ) 225 .
azimov , v.l . rappoport and v.v .
sarantsev , _ z. phys .
_ * a356 * ( 1997 ) 437 ; + y. grossman and h.r .
quinn , _ phys .
_ * d56 * ( 1997 ) 7259 ; + j. charles _ et al .
_ , _ phys .
_ * b425 * ( 1998 ) 375 ; + b. kayser and d. london , _ phys .
_ * d61 * ( 2000 ) 116012 ; + h.r .
et al . _ , _ phys .
lett . _ * 85 * ( 2000 ) 5284 .
dighe , i. dunietz and r. fleischer , _ phys .
_ * b433 * ( 1998 ) 147 .
i. dunietz , r. fleischer and u. nierste , _ phys .
_ * d63 * ( 2001 ) 114015 .
r. itoh , kek - preprint-2002 - 106 [ hep - ex/0210025 ] .
r. fleischer and j. matias , _ phys .
_ * d66 * ( 2002 ) 054009 .
m. gronau and d. london , _ phys .
_ * b253 * ( 1991 ) 483 .
b. kayser and d. london , _ phys .
_ * d61 * ( 2000 ) 116013 .
d. atwood and a. soni , bnl - het-02 - 12 [ hep - ph/0206045 ] .
r. fleischer , cern - th/2003 - 010 [ hep - ph/0301255 ] .
d. atwood , i. dunietz and a. soni , _ phys .
_ * 78 * ( 1997 ) 3257
. m. neubert , clns-02 - 1794 [ hep - ph/0207327 ] ; + r. fleischer , cern - th/2002 - 293 [ hep - ph/0210323 ] .
m. gronau and d. wyler , _ phys .
_ * b265 * ( 1991 ) 172 .
i. dunietz , _ phys .
_ * b270 * ( 1991 ) 75 .
r. aleksan , i. dunietz and b. kayser , _ z. phys .
_ * c54 * ( 1992 ) 653 .
i. dunietz , _ phys .
_ * b427 * ( 1998 ) 179 .
m. gronau , _ phys .
_ * d58 * ( 1998 ) 037301 .
r. fleischer and d. wyler , _ phys .
_ * d62 * ( 2000 ) 057503 .
r. aleksan , t.c . petersen and a.
soffer , lal-02 - 83 [ hep - ph/0209194 ] .
y. grossman , z. ligeti and a. soffer , lbnl-51630 [ hep - ph/0210433 ] .
m. gronau , cern - th/2002 - 331 [ hep - ph/0211282 ] .
belle collaboration ( k. abe _
_ ) , _ phys .
lett . _ * 88 * ( 2002 ) 052002 ; + cleo collaboration ( t.e .
_ ) , _ phys . rev .
lett . _ * 88 * ( 2002 ) 062001 ; + babar collaboration ( b. aubert _ et al .
_ ) , babar - conf-02/17 [ hep - ex/0207092 ] . belle collaboration ( p. krokovny _
_ ) , hep - ex/0212066 .
i. dunietz , _ phys .
_ * d52 * ( 1995 ) 3048 .
m. beneke and a. lenz , _
j. phys .
_ * g27 * ( 2001 ) 1219 .
r. fleischer and i. dunietz , _ phys .
_ * d55 * ( 1997 ) 259 ; _ phys .
_ * b387 * ( 1996 ) 361 .
j.p . silva and a.
soffer , _ phys .
_ * d61 * ( 2000 ) 112001 ; + d. atwood , i. dunietz and a. soni , _ phys .
_ * d63 * ( 2001 ) 036005 .
buras , tum - hep-489 - 02 [ hep - ph/0210291 ] . m. neubert and a.a .
petrov , _ phys .
_ * b519 * ( 2001 ) 50 ; + c .- k .
chua , w .- s . hou and k .- c .
yang , _ phys .
_ * d65 * ( 2002 ) 096007 .
m. beneke , g. buchalla , m. neubert and c.t .
sachrajda , _ nucl .
_ * b591 * ( 2000 ) 313 .
particle data group ( k. hagiwara _ et al .
_ ) , _ phys . rev .
_ * d66 * ( 2002 ) 010001 .
z. ligeti , _ aip conf .
_ * 618 * ( 2002 ) 298 .
l. wolfenstein , _ phys .
lett . _ * 75 * ( 1995 ) 2460 .
m. gronau and j.l .
rosner , _ phys .
_ * b500 * ( 2001 ) 247 .
m. bauer , b. stech and m. wirbel , _ z. phys .
_ * c34 * ( 1987 ) 103 .
m. ablikim , d .- s . du and m .- z .
yang , bihep - th-2002 - 50 [ hep - ph/0211413 ] .
chau and h.y .
cheng , _ phys .
_ * b333 * ( 1994 ) 514 ; + t.e .
browder and s. pakvasa , _ phys .
_ * b383 * ( 1996 ) 475 ; + a.f .
falk , y. nir and a.a .
petrov , _ jhep _ * 9912 * ( 1999 ) 019 ; + s. bergmann _
_ , _ phys .
_ * b486 * ( 2000 ) 418 .
m. gronau , y. grossman and j.l .
rosner , _ phys .
_ * b508 * ( 2001 ) 37 . y. grossman , b. kayser and y. nir , _ phys .
lett . _ * b415 * ( 1997 ) 90 . | decays of neutral @xmath0 mesons originating from @xmath1 @xmath2 processes , i.e. @xmath3 , @xmath4 @xmath5 and @xmath6 , @xmath7 @xmath8 modes , offer valuable insights into cp violation . here
the neutral @xmath9 mesons may be observed through @xmath10 or @xmath11 , where @xmath12 and @xmath13 are cp eigenstates and cp non - eigenstates , respectively .
recently , we pointed out that `` untagged '' and mixing - induced cp - violating @xmath14 observables provide a very efficient , theoretically clean determination of the angle @xmath15 of the unitarity triangle . here
we perform a more detailed analysis of the @xmath16f_s$ ] channels , and focus on @xmath17 f_s$ ] , where interference effects between cabibbo - favoured and doubly cabibbo - suppressed @xmath9-meson decays lead to complications .
we show that @xmath15 can nevertheless be determined in an elegant and essentially unambiguous manner from the @xmath17f_s$ ] observables with the help of the corresponding `` untagged '' @xmath16f_s$ ] measurements .
moreover , we may also extract hadronic @xmath0- and @xmath9-decay parameters .
several interesting features of decays of the kind @xmath18 are also pointed out .
cern - th/2003 - 011 + hep - ph/0301256 * a closer look at @xmath19 decays and * novel avenues to determine @xmath15 robert fleischer + _ theory division , cern , ch-1211 geneva 23 , switzerland _
cern - th/2003 - 011 + january 2003 |
single - dish surveys have shown that molecular outflows are ubiquitous in high - mass star forming regions @xcite . according to the standard paradigm of low - mass star formation
this provides a statistical argument that high - mass stars may form via accretion processes .
however , given the typical large distances ( @xmath15 kpc ) and clustering environment of massive star formation , single - dish observations of @xmath16 resolutions can not resolve the detailed structures of massive molecular outflows or identify their driving sources .
recently there have been an increasing number of interferometric studies which provide new insights into the morphology and kinematics of massive molecular outflows ( e.g. , * ? ? ?
* and references therein ) , but the statistics is still poor and many basic questions , e.g. , how massive molecular outflows systematically differ from their low - mass counterparts , how they are driven , and how they affect their surroundings and host clouds , remain open . to address these questions , which are essential to understand massive star formation , a large set of high - angular - resolution observations covering a wide range of luminosity and evolutionary stages is needed . toward these goals we study g240.31 + 0.07 ( hereafter g240 ) , an active high - mass star - forming region with a far - infrared luminosity of @xmath0 l@xmath1 at a distance of 6.4 kpc @xcite .
it harbors an ultra - compact ( uc ) hii region and is associated with h@xmath17o and oh masers @xcite .
@xcite reported a parsec - scale velocity gradient in c@xmath18o ( 21 ) with a @xmath19 beam and the detection of near - infrared h@xmath17 emission in this region .
@xcite presented an interferometric study at 654 ghz of the central core and resolved the dust continuum emission into two clumps both associated with c@xmath18o ( 6 - 5 ) emission . in the meantime
, molecular outflow gas in this region has been mapped using single - dish telescopes : @xcite reported a co ( 10 ) bipolar outflow with a @xmath20 beam ; @xcite mapped the co ( 32 ) emission with a @xmath21 beam and proposed two overlapping outflows both in the northwest - southeast orientation .
however , limited by low angular resolutions , neither of the two single - dish studies provided a detailed analysis of the outflow morphology and kinematics .
high - angular - resolution observations are crucial to unveil the outflow nature of this luminous source . here
we present a @xmath3co , @xmath4co ( 21 ) study of the g240 outflow , using the submillimeter array ( sma ) observations .
complementary short spacings are obtained from the caltech submillimeter observatory ( cso ) 10.4 m telescope .
the archival _
mid - infrared imaging is also presented for comparison .
the 1.3 mm continuum observations are used to pinpoint the outflow powering source .
the sma observations were carried out on 2008 february 25 with 8 antennas in the compact configuration and on 2008 february 16 with 7 antennas in the extended configuration .
the @xmath22ghz correlator was configured to cover @xmath3co ( 21 ) in the upper sideband and @xmath4co ( 21 ) in the lower sideband . and the continuum was constructed from line - free channels .
lines of @xmath3co , @xmath4co ( 21 ) and the 1.3 mm continuum , which are relevant to the outflow study , are presented in this work .
many other lines covered in the same setup will be presented in a future paper . to ensure the coverage of the entire outflow we observed two fields centered at ( r.a . ,
decl.)@xmath23 and ( r.a . ,
decl.)@xmath24 , respectively .
we used titan as the primary flux calibrator and 3c273 as the bandpass calibrator . the time dependent gain was monitored by observing quasars j0730 - 116 and j0826 - 225 every 23 mins .
visibilities were calibrated using the idl mir package and then output to miriad for imaging . with natural weighting
the synthesized beams in the compact and extended configurations are about @xmath10 and @xmath12 , respectively .
the shortest baseline in our sma observations is 16.5 m , corresponding to a spatial scale of @xmath21 for observations at 225 ghz .
thus spatial structures more extended than @xmath21 were not sampled in the sma observations .
this spatial filtering can significantly affect the @xmath3co , @xmath4co ( 21 ) maps at velocities close to the cloud velocity . to recover the missing short spacing information we observed the @xmath3co , @xmath4co ( 21 ) emission with the cso 10.4 m telescope on 2008 february 12 . during the observation the weather condition was excellent for 1 mm waveband with @xmath25 .
the observations were carried out in the on - the - fly mode centered on ( r.a .
, decl.)@xmath26 .
we obtained a @xmath27 grid in @xmath3co and a @xmath28 grid in @xmath4co , with an integration of about 5 s on each cell .
the @xmath29 cell spacing is about @xmath30 of the cso beam , which is @xmath31 in @xmath3co ( 21 ) . by observing mars and saturn we derived a beam efficiency of @xmath32 .
the spectrometer used has 1024 channels throughout the 50 mhz bandwidth , resulting in a spectral resolution of 0.0488 mhz ( about 0.065 km s@xmath8 ) per channel .
the final maps were smoothed to 2 km s@xmath8 per channel .
the data were reduced using the standard class package .
we combined the sma compact and cso @xmath3co , @xmath4co data in miriad following a procedure outlined in @xcite .
the deep _ spitzer _ irac images were obtained from the _ spitzer _ archive ( ) .
we found no obvious artifacts from a visual inspection , thus adopted s14.0 post basic calibrated data products provided by the _
spitzer _ science center .
in figure [ continuum]a , the 1.3 mm continuum emission observed with the sma compact array shows an elongated envelope with its major axis in the northeast - southwest direction , as well as a minor peak to the northeast . at a higher angular resolution of @xmath12 the central part of the structure
is resolved into three dusty cores , namely mm1 , mm2 , and mm3 from north to south ( figure [ continuum]b ) .
assuming optically thin dust emission , the dust mass can be estimated according to @xmath33 where @xmath34 is the dust mass , @xmath35 the continuum flux at frequency @xmath36 , @xmath37 the source distance , @xmath38 the planck function at dust temperature @xmath39 , and @xmath40 in @xmath41g@xmath8 the dust opacity @xcite .
based on sma observations toward this region @xcite derived a gas rotational temperature of @xmath42 k by comparing the brightness of two h@xmath17co transitions , and derived an opacity index @xmath43 of @xmath44 by comparing flux densities at 223 and 654 ghz .
we then adopt @xmath45 k , @xmath46 , and a gas - to - dust ratio of 100 , to estimate the mass of the larger envelope and the combined mass of the three dusty cores , which amount to 150 m@xmath1 and 40 m@xmath1 , respectively .
the uncertainty of this estimate mainly comes from the determinations of @xmath43 and @xmath39 ; for example , deviations of @xmath47 in @xmath43 or @xmath48 k in @xmath39 will induce @xmath49% uncertainty in the mass estimates .
figure [ flow_int ] shows the integrated @xmath3co , @xmath4co ( 21 ) emission observed with the sma , as well as the _
irac imaging for comparison . because the interferometric filtering to extended emission helps to distinguish the outflow from ambient gas , we present the sma observations without merging with the single - dish data .
the cloud velocity ( @xmath50 ) with respect to the local standard of rest is about 67.5 km s@xmath8 , which is adopted from the cso c@xmath18o ( 21 ) observations by @xcite and consistent with our sma c@xmath18o ( 21 ) observations .
the low - velocity channels , i.e. , 5864 km s@xmath8 for the blueshifted lobe and 7480 km s@xmath8 for the redshifted lobe in figure [ flow_int]a , are chosen to highlight wide - angle structures and to avoid contamination from diffuse gas in the surrounding , and the high - velocity channels ( figure [ flow_int]b ) are those in the outer line wings .
the low - velocity @xmath3co emission clearly reveals a northwest - southeast ( nw - se ) bipolar outflow centered at the group of the mm peaks and extending about 1.3 pc from end to end ( figure [ flow_int]a ) . the outflow shell highlighted by limb - brightening outlines a quasi - parabolic structure with a wide opening angle .
both the spatial separation of the two lobes of such a wide - angle outflow and the partial overlap of blue- and redshifted emission in the nw lobe suggest that the flow axis is moderately inclined to the plane of sky .
we suggest the molecular outflow seen in the @xmath3co emission as ambient gas swept - up by an underlying wide - angle wind ( see [ comparison ] ) . in this context , the nw wind is overall moving away from us , sweeping up ambient gas to create the redshifted @xmath3co outflow , but could be partly moving toward us in its near side , causing the blueshifted @xmath3co emission .
the patchy appearance of the blueshifted emission might be attributed to the localized density structure of the medium .
the se counterpart of the wind may be expanding into a medium which is less dense in the far - side and would not create appreciable redshifted @xmath3co emission . immediately to the southeast of the bipolar outflow , there is a minor extension with overlapping red- and blueshifted emission , which might arise from another outflow .
meanwhile , the high - velocity @xmath3co emission coincides with the inner part of the low - velocity bipolar outflow , with the bulk blueshifted emission revealing a clump in the middle of the low - velocity se lobe and the redshifted emission a conical structure narrower than the low - velocity nw lobe ( figure [ flow_int]b ) .
the bipolar outflow is also detected in the @xmath4co emission with a morphology similar to but slightly less extended than the @xmath3co outflow ( figure [ flow_int]c ) .
_ spitzer _ irac images often reveal outflow cavities through scatted light in the 3.6 @xmath5 m band and molecular outflow driving agents ( e.g. , underlying jets or winds ) through shocked h@xmath17 emission in the 4.5 @xmath5 m band @xcite . in figure
[ flow_int]d , the irac three - color composite image of the g240 central area yields a nw - se nebula in the 3.6 and 4.5 @xmath5 m emission .
the north to south shell of the nebula appears prominent in the 3.6 @xmath5 m band ( _ blue _ ) and coincides well with the north to south edge of the @xmath3co outflow , suggesting its emission being dominated by scattered light from the outflow cavity wall , whereas the nebula structure coincident with the main nw - se @xmath3co lobe shows diffuse 3.6 and 4.5 @xmath5 m emission , which might be dominated by scattered light and shocked h@xmath17 emission .
figures [ sma12ch ] and [ smacso12ch ] show channel maps of the @xmath3co emission from the sma observations alone and from the combined sma and cso data at low velocities , respectively .
it is immediately evident that for channels around @xmath50 the combined data effectively recover the extended emission filtered out by the interferometer and significantly improve the side - lobes of the interferometric observations .
in particular the diffuse gas emissions at 66 and 68 km s@xmath8 are mostly filtered out by the interferometer and recovered with the combined data . from a comparison among the spectra constructed from the cso , sma , and
the combined data ( figure [ spec ] ) , the sma compact array observations recover less than @xmath51 of the cso flux at the line center , and about @xmath52 of the cso flux at the line wing of 54 and 80 km s@xmath8 , whereas the combined sma and cso data recover more than 85% of the cso flux all the way from 54 to 80 km s@xmath8 ; at the outer line wing of @xmath5352 km s@xmath8 and @xmath5482 km s@xmath8 , there is little missing flux in the sma observations .
outflow features emerge at 64 and 70 km s@xmath8 ; the se wide - angle shell at 64 km s@xmath8 appears less prominent in figure [ smacso12ch ] than that in figure [ sma12ch ] due to the confusion from diffuse gas .
emissions from outflow structures appear predominant in channels of @xmath55 km s@xmath8 and @xmath56 km s@xmath8 .
the two lobes of the wide - angle outflow accord with each other in orientation and have similar opening angles at the base .
they both appear more collimated at higher velocities : the nw lobe has a fan - shaped morphology at low velocities ( most prominent at 72 and 74 km s@xmath8 ) and a narrower , _
v_-shaped appearance at high velocities ( most prominent at 84 to 88 km s@xmath8 ) ; the se lobe shows the wide - angle shell at low velocities ( most prominent at 62 and 64 km s@xmath8 ) and a clump away from the mm peaks at high velocities ( most prominent at 48 to 52 km s@xmath8 ) .
again the difference in detailed velocity structures of the two lobes could be at least in part attributed to the localized density structure of the environment within which the underlying wind is expanding . as seen in the integrated emission , a possible secondary outflow to the southeast of the mm peaks can be found in both blue- and redshifted channels .
channel maps of the @xmath4co emission are shown in figures [ sma13ch ] and [ smacso13ch ] , for the sma alone and the combined sma and cso data around @xmath50 , respectively .
channels of 6470 km s@xmath8 are dominated by surrounding diffuse gas but also show extension along the outflow orientation .
there is an embedded dense clump showing a northeast - southwest velocity gradient , which is also detected in other dense gas tracers .
emission from the bipolar outflow dominates channel of @xmath5762 km s@xmath8 and @xmath5872 km s@xmath8 . the @xmath4co emission appears confined within 10 km s@xmath8 from @xmath50 , for both blue- and redshifted lobes .
we use the combined sma and cso data to estimate the gas mass of the outflow , following @xmath59 where @xmath60 , @xmath61 , @xmath37 , @xmath62 , and @xmath63 are the outflow gas mass in m@xmath1 , excitation temperature of the @xmath3co ( 21 ) transition , source distance in kpc , optical depth for @xmath3co ( 21 ) , and line flux in jy , respectively .
we adopt an excitation temperature of 30 k , a @xmath3co - to - h@xmath17 abundance of @xmath64 , and a mean gas atomic weight of 1.36 . assuming the same excitation temperature for @xmath3co ( 21 ) and @xmath4co ( 21 ) , we calculate @xmath62 according to @xmath65 where @xmath66 is the abundance ratio of @xmath3co to @xmath4co , which is estimated to be 103 at a galactocentric distance of 12.7 kpc for g240 @xcite .
many @xmath3co outflow studies estimate outflow mass either by assuming that @xmath3co emission is optically thin at all outflow velocities or by correcting for a constant @xmath3co opacity .
recently a few interferometric outflow studies use single - dish @xmath3co/@xmath4co line ratios to correct for @xmath3co opacity as a function of velocity @xcite .
our interferometric , beam - matching observations of @xmath3co and @xmath4co emission , both complemented by short spacing information from single - dish observations , enable a measurement of spatial and spectral distribution of @xmath3co/@xmath4co line ratios , thus a more sophisticated correction for the @xmath3co opacity . we assume the @xmath3co emission is optically thin where the @xmath4co is not detected . the momentum and energy of the outflow is then derived following @xmath67 and @xmath68 respectively , where ( and hereafter ) @xmath69 denotes the outflow velocity relative to @xmath50 . to avoid contamination from the ambient molecular gas , we use channels of @xmath70 km s@xmath8 in the blueshifted outflow and channels of @xmath71 km s@xmath8 in the redshifted outflow , although channels of 64 and 70 km s@xmath8 show contributions from the outflow ( see figures [ sma12ch ] and [ smacso12ch ] ) .
the area taken into account is defined by a polygon encompassing the outflow structure shown in figure [ flow_int]a .
the dynamical timescale of the outflow is derived following @xmath72 , where @xmath73 pc and @xmath74 km s@xmath8 are the half length of the end - to - end flow extension and the maximum flow velocity relative to @xmath50 , respectively .
consequently the mass loss rate of the outflow can be derived following @xmath75 .
the derived outflow parameters are listed in table [ table1 ] .
from figure [ flow_int ] , it is evident that one of the three mm continuum cores harbors the driving source of the bipolar outflow . from the maser and cm continuum observations in the literature
@xcite , mm1 roughly coincides with an uc hii region and a h@xmath17o maser ; mm2 is associated with h@xmath17o and oh masers but shows no cm emission down to 0.5 mjy at 6 cm or 2 mjy at 3.6 cm ; mm3 has no apparent signature of active star formation .
the 1.3 mm continuum emission from mm3 appears much weaker than that from mm1 and mm2 , implying a mass significantly less than 10 m@xmath76 .
its relatively faint dust emission and the apparent non - association with active star formation suggest that mm3 does not harbor the driving source of the bipolar outflow . of the remaining two mm cores
, mm2 appears closer to the geometric center of the bipolar outflow , thus is more likely to be the parent core of the outflow driving source .
this is compatible with the relatively short outflow dynamical timescale , @xmath9 yr , in comparison with @xmath77 yr for massive molecular outflows from uc hii regions @xcite .
low - mass molecular outflows are known to exhibit highly collimated jetlike structures and/or wide - angle quasi - conical or quasi - parabolic structures .
it is still unclear whether massive molecular outflows have an essentially similar appearance as low - mass outflows .
high - angular - resolution interferometric observations have resolved a few single - dish bipolar outflows into multiple jetlike outflows , which were suggested as indirect evidence that associated high - mass stars may form via disk accretion ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . on the other hand , very few well - defined wide - angle molecular outflows have been found in high - mass star - forming regions . @xcite
reported bipolar conical outflow cavities in _
mid - ir imaging of iras 20126 + 4104 and hh 80/81 , but the former is subject to ambiguity about whether it is caused by a precessing jet and the latter lacks kinematical information due to the lack of high - angular - resolution mm line observations .
the g192.16 - 3.82 outflow , originating from an uc hii region , may be representative of a bipolar wide - angle outflow with limb - brightened shells in a @xmath78 l@xmath1 star forming region @xcite .
our observations of the g240 outflow reveal a bipolar wide - angle molecular outflow with a quasi - parabolic shell . to our knowledge
this is hitherto the best case of a well - defined wide - angle molecular outflow in a @xmath14 l@xmath1 star forming region .
the outflow has a morphology similar to protostellar low - mass molecular outflows , e.g. , l1228 outflow @xcite , b335 outflow @xcite .
the behavior of higher collimation at higher velocities is also often seen in low - mass molecular outflows . given the striking similarity between the g240 outflow and low - mass wide - angle outflows ,
the g240 outflow could be driven by an underlying wide - angle wind , as commonly suggested for low - mass wide - angle outflows .
@xcite argued that neither accumulated stellar winds nor entrained interstellar matter in bipolar jets is able to account for @xmath79 m@xmath1 of gas masses measured in massive molecular outflows typically at @xmath80 yrs old .
it is true that the g240 molecular outflow is unlikely to primarily contain accumulated stellar winds , because stellar winds from ob stars are radiation driven and their momentum supply rates can not be higher than @xmath81 , while the g240 outflow has a mechanical force , @xmath82 , about 25 times higher than @xmath81 .
the mass limit of entrained interstellar matter argued by @xcite was based on the turbulent jet entrainment model @xcite . in this model
a bipolar jet entrains aside interstellar matter through a turbulent mixing layer . because the environment can not react to the rarefaction by the entrainment at velocities faster than the sound speed @xmath83
( typically an order of 1 km s@xmath8 ) , the entrainment rate per unit area is given by @xmath84 , where @xmath85 is the entrainment efficiency and @xmath86 the density of the ambient medium .
on the other hand , @xcite proposed an analytical model for bipolar molecular outflows , where an underlying wide - angle wind radially sweeps up ambient gas into an expanding shell to form a molecular outflow .
further analytical calculations and simulations of the model , by adopting an angle - dependent wind blowing into an angle - dependent molecular cloud core , appear generally consistent with observations @xcite . in this model molecular outflows are assumed to be _ momentum driven _ and the entrainment rate per unit area is simply proportional to @xmath87 , where @xmath88 is the shell expanding speed ( i.e. , the molecular outflow speed observed in @xmath3co , typically an order of 10 km s@xmath8 ) and determined by the momentum supply rate of the underlying wind .
the mass loading limit in the turbulent entrainment model does not seem to be applicable to the swept - up shell model . even in the turbulent entrainment model , a factor of 10 higher mass loading
could be achieved if the entrainment process works at alfven speeds , which are a factor of 10 higher than @xmath83 , when the magnetic fields are important .
another molecular outflow model that can readily account for high mass loading is the so - called `` circulation model '' @xcite .
molecular outflow materials in this model are those deflected from infalling matter , instead of entrained or swept - up ambient gas by an underlying jet or wind . however , in at least one case , the g192.16 outflow , there is strong evidence for the presence of a wide - angle wind being the driving agent of the wide - angle @xmath3co outflow ( * ? ? ?
* and references therein ) .
the 3.6 and 4.5 @xmath5 m emission coincident with the g240 @xmath3co outflow lobes shown in figure [ flow_int]d , is likely excited by an underlying fast wind .
therefore , we suggest that the g240 outflow is more likely to be driven / swept - up by an underlying wide - angle wind .
the wind itself might be accretion driven and somewhat analogous to _ x_-wind or disk - wind in low - mass stars @xcite , although it is unknown whether _ x_-wind or disk - wind can be scaled up and applied to high - mass stars . figure
[ pv ] shows the position - velocity ( pv ) diagram of the g240 @xmath3co outflow cut roughly along the outflow major axis .
the nw redshifted lobe shows a feature extending to @xmath89 from the driving source at a velocity of @xmath90 km s@xmath8 and another feature extending to @xmath91 from the source at a velocity of @xmath92 km s@xmath8 .
connecting these two features following the gradient of the contour levels in figure [ pv ] one can get an approximately parabolic trajectory opening outward from the driving source .
this kind of pv structure is very similar to that of the northern lobe of the vla 05487 outflow , a good example of a wide - angle wind driven low - mass outflow , and can be reasonably reproduced by the hydrodynamic simulation of a modified _ x_-wind model @xcite .
according to the model the pv structure can be qualitatively understood in the context of a radially expanding shell with a hubble - law velocity structure and a moderate inclination angle , where the far - side shell has a smaller projected extension and higher line - of - sight velocities , while the near - side shell has a larger projected extension and lower line - of - sight velocities .
this parabolic pv structure supports the scenario that the g240 molecular outflow is driven by a wide - angle wind .
it is difficult to identify a similar pv structure in the se blueshifted lobe , probably because this lobe is highly asymmetric at most velocities ( the western part of the shell can be only seen at 62 and 64 km s@xmath8 in figure [ sma12ch ] ) and contaminated by a secondary outflow to the west .
its pv structure is dominated by a high velocity feature @xmath93 away from the source , corresponding to a clump shown in figure [ flow_int]b . while molecular outflows associated with both low- and high - mass ( proto)stars are found to exhibit a `` mass spectrum '' , @xmath94 ,
with the power - law index @xmath95 around 2 for low velocity gas and often steepening at velocities higher than @xmath96 km s@xmath8 , its physical origin is not well established @xcite .
figure [ mv ] shows the mass - velocity diagram of the g240 outflow , where both lobes can be fitted as a broken power - law , with @xmath95 steepening from @xmath97 to @xmath98 at 13.5 km s@xmath8 for the blueshifted lobe and from @xmath99 to @xmath100 at 10.5 km s@xmath8 for the redshifted lobe .
it appears that the g240 outflow has fairly large values of @xmath95 at high velocities , comparable to that of some other @xmath3co outflows associated with both high- and low - mass ( proto)stars @xcite . such large values of @xmath95 have been an unsolved issue to both jet bow - shock and wide - angle wind models since numerical simulations of these models predict @xmath101 ; future simulations that run for flows more than @xmath102 yr old might be able to address whether it can be explained by outflow evolution ( see a review by * ? ? ?
moreover , as shown by colored lines in figure [ mv ] , the mass spectra of high velocity channels of both lobes can be further fitted with a broken power - law rather than a single power - law , and in those narrower velocity ranges @xmath95 steepens from @xmath103 to @xmath104 at 19.5 km s@xmath8 for the blueshifted lobe and from @xmath105 to @xmath106 at 20.5 km s@xmath8 for the redshifted lobe .
this may simply reflect a continuing decrease of mass entrainment efficiency with increasing outflow velocities in this outflow .
the total gas mass of the g240 outflow , estimated based on combined interferometer and single - dish data with sophisticated opacity correction , amounts to 98 m@xmath1 , making it one of the most massive ( proto)stellar outflows known . given the unknown inclination angle of the outflow , the estimated outflow momentum of 620 m@xmath1 km s@xmath8 and energy of @xmath107 erg can be treated as lower limits accordingly . based on a velocity gradient in their cso c@xmath18o ( 21 ) observations and assuming the gas is in virial equilibrium , @xcite derived a gas mass of 625 m@xmath1 . considering
the cloud is likely infalling to feed the accretion onto the central ( proto)stars and thus deviates from virial equilibrium , this estimate may provide a lower limit of the dense cloud mass .
the gravitational binding energy of the cloud , @xmath108 , where @xmath109 pc @xcite , is then @xmath110 erg , which is comparable to the outflow energy ( table [ table1 ] ) .
although there are many uncertainties in calculations of both the gravitational binding energy and the outflow energy , the g240 outflow seems to play a critical role in disrupting or unbinding its host cloud .
the mass loss rate of @xmath111 m@xmath1 yr@xmath8 is about two to three orders of magnitude higher than that of typical low - mass outflows . assuming momentum conservation between the swept - up molecular outflow and an underlying primary wind , i.e. , @xmath112 ( also see eq .
8 of * ? ? ?
* ) , where @xmath113 , @xmath114 are the wind mass loss rate and wind speed , respectively , and adopting a wind speed of 500 km s@xmath8 , the mass loss rate of the wind amounts to @xmath115 m@xmath1 yr@xmath8 .
furthermore , adopting a ratio of @xmath30 for the wind mass loss rate to mass accretion rate @xcite , we get an accretion rate of @xmath116 m@xmath1 yr@xmath8 , high enough to overcome radiation pressure from the central object to form a high - mass star @xcite . again considering the underestimate to @xmath117 ,
this estimate may represent a lower limit to the actual accretion rate .
we have presented a high - angular - resolution study of the molecular outflow in a @xmath0 l@xmath1 star forming region , using the @xmath3co and @xmath4co ( 21 ) observations from the sma and cso , as well as the _ spitzer _
irac color - composite image .
the sma 1.3 mm dust continuum images are used to pinpoint the outflow powering source . a bipolar , wide - angle , quasi - parabolic molecular outflow is revealed in the @xmath3co and @xmath4co emission at velocities more than 3 km s@xmath8 apart from the cloud velocity . the outflow is most likely powered by a deeply embedded high - mass protostar prior to forming a significant uc hii region . from its morphological similarity to some low - mass molecular outflows and parabolic pv structure in the redshifted lobe , the g240 outflow is likely driven / swept - up by an underlying wide - angle wind .
for both lobes the outflow mass - velocity relation can be fitted with broken power - laws with the indices steeping at high velocities .
the outflow has @xmath6 m@xmath1 gas mass and a dynamical timescale of @xmath9 yr , thus a mass loss rate of @xmath7 m@xmath1 yr@xmath8 . without correction for the unknown inclination angle the momentum and energy of the outflow amount to 620 m@xmath1 km s@xmath8 and @xmath107 erg , respectively .
our observations of the g240 outflow reveal a well - defined wide - angle molecular outflow with its morphology and kinematics similar to low - mass molecular outflows , but with two to three orders of magnitude greater mass and energetics suggestive of a high accretion rate .
these findings appear to be consistent with the scenario that high - mass stars of early - b to late - o types can form in a similar way as low - mass star formation , even in a clustering environment .
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 the national aeronautics and space administration ( nasa ) .
arce , h. g. & goodman , a. a. 2001 , , 554 , 132 arce , h. g. & sargent , a. i. 2004 , apj , 612 , 342 arce , h. g. , shepherd , d. s. , gueth , f. , lee , c .- f .
, bachiller , r. , rosen , a. , & beuther , h. 2007 , in protostars and planets v. , eds .
b. reipurth , d. jewitt , & k. keil , 245 beuther , h. , schilke , p. , gueth , f. , mccaughrean m. , andersen , m. , sridharan , t. k. , & menten , k. m. 2002b , , 387 , 931 beuther , h. , schilke , p. , sridharan , t. k. , menten , k. m. , walmsley , c. m. , & wyrowski , f. 2002a , , 383 , 892 cant , j. & raga , a. c. 1991 , , 372 , 646 caswell , j. l. 1997 , , 289 , 203 chen , h .-
, su , y .-
n . , liu , s .- y . , hunter , t. r. , wilner , d. j. , zhang , q. , lim , j. , ho , p. t. p. , ohashi , n. , & hirano , n. 2007 , , 654 , l87 churchwell , e. 1997 , , 479 , l59 hildebrand , r. h. 1983 , , 24 , 267 hirano , n. , kameya , o. , nakayama , n. , & takakubo , k. 1988 , , 327 , l69 hughes , v. a. & macleod , g. c. 1993 , , 105 , 1495 hunter , t. r. 1997 , ph.d .
thesis , caltech jijina , j. & adams , f. c. 1996 , , 462 , 874 jrgensen , j. k. , bourke , t. l. , myers , p. c. , di francesco , j. , van dishoeck , e. f. , lee , c .- f . , ohashi , n. , schier , f. l. , takakuwa , s. , wilner , d. j. , & zhang , q. 2007 , , 659 , 479 lee c .- f . , stone , j. m. , ostriker , e. c. , & mundy , l. e. 2001 , , 557 , 429 migenes et al . 1999 , , 123 , 487 kumar , m. s. n. , fernandes , a. j. l. , hunter , t. r. , davis , c. j. , & kurtz , s. 2003 , , 412 , 175 lery , t. , henriksen , r. n. , fiege , j. d. , ray , t. p. , frank , a. , & bacciotti , f. 2002 , , 387 , 187 li , z .- y . & shu , f. h. 1996 , , 472 , 211 macloed , v. a. , scalise , jr .
e. , saedt , s. , galt , j. a. , & gaylard , m. j. 1998 , , 116 , 1897 pudritz , r. e. , ouyed , r. , fendt , c. , & brandenburg , a. 2007 , in protostars and planets v , eds .
b. reipurth , d. jewitt , & k. keil , 277 qiu , k. , zhang , q. , beuther , h. , & yang , j. 2007 , , 654 , 361 qiu , k. , zhang , q. , megeath , s. t. , gutermuth , r. a. , beuther , h. , shepherd , d. s. , sridharan , t. k. , testi , l. , & de pree , c. g. 2008 , , 685 , 1005 richer , j. , shepherd , d. , cabrit , s , bachiller , r. , & churchwell , e. 2000 , in protostars and planets iv , eds . v. mannings , a. boss , & s. russell ( tucson : univ .
arizona press ) , 867 shang , h. , allen , a. li , z .- y . , liu , c .- f . ,
chou , m .- y . , anderson , j. 2006 , , 649 , 845 shepherd , d. s. & churchwell , e. 1996 , , 457 , 267 shepherd , d. s. , testi , l. , & stark , d. p. 2003 , 584 , 882 shepherd , d. s. , watson , a. m. , sargent , a. i. , & churchwell , e. 1998 , , 507 , 861 shu , f. h. , ruden , s. p. , lada , c. j. , & lizano , s. 1991 , , 370 , l31 shu , f. h. , najita , j. r. , shang , h. , & li , z .- y .
2000 , in protostars and planets iv , 789 su , y .- n . , zhang , q. , & lim , j. 2004 , , 604 , 258 tomisaka , k. 1998 , , 502 , l163 wilson , t. l. & rood , r. 1994 , , 191 wolfire , m. g. & cassinelli , j. p. 1987
, , 319 , 850 yorke , h. w. & sonnhalter , c. 2002 , , 569 , 846 yu , k. c. , billawala , y. , smith , m. d. , bally , j. , & butner , h. m. 2000 , , 120 , 1974 zhang , q. , ho . , p. t. p. , wright , m. c. h. , & wilner , d. j. 1995 , , 451 , l71 zhang , q. , hunter , t. r. , beuther , h. , sridharan , t. k. , liu , s .- y .
, su , y .-
n . , chen , h .-
, & chen , y. 2007a , , 658 , 1152 zhang , q. , hunter , t. r. , brand , j. , sridharan , t. k. , molinari , s. , kramer , m. a. , & cesaroni , r. 2001 , , 552 , l167 zhang , q. , sridharan , t. k. , hunter , t. r. , chen , y. , beuther , h. , & wyrowski , f. 2007b , , 470 , 269 | we present submillimeter array observations toward the @xmath0 l@xmath1 star - forming region g240.31 + 0.07 , in the @xmath21 transition of @xmath3co and @xmath4co and at 1.3 mm continuum , as well as the @xmath3co and @xmath4co observations from the caltech submillimeter observatory to recover the extended emission filtered out by the interferometer .
maps of the @xmath3co and @xmath4co emission show a bipolar , wide - angle , quasi - parabolic molecular outflow , roughly coincident with an ir nebula revealed by the _ spitzer _ 3.6 and 4.5 @xmath5 m emission .
the outflow has @xmath6 m@xmath1 molecular gas , making it one of the most massive molecular outflows known , and resulting in a very high mass - loss rate of @xmath7 m@xmath1yr@xmath8 over a dynamical timescale of @xmath9 yr .
the 1.3 mm continuum observations with a @xmath10 beam reveal a flattened dusty envelope of @xmath11 m@xmath1 , which is further resolved with a @xmath12 beam into three dense cores with a total mass of @xmath13 m@xmath1 .
the central mm core , showing evidence of active star formation , approximately coincides with the geometric center of the bipolar outflow thus most likely harbors the powering source of the outflow .
overall our observations provide the best case to date of a well - defined wide - angle molecular outflow in a @xmath14 l@xmath1 star - forming region .
the outflow is morphologically and kinematically similar to low - mass protostellar outflows but has two to three orders of magnitude greater mass , momentum , and energy , and is apparently driven by an underlying wide - angle wind , hence further supports that high - mass stars up to late - o types , even in a crowded clustering environment , can form as a scaled - up version of low - mass star formation . |
goudsmit and uhlenbeck discovered electron spin in 1925 , while trying to find out what causes very close parallel lines to appear in the optical spectrum of hydrogen .
they were able to show that these lines could be due to the fact that the electron possesses a natural angular impulse and bipolar magnetic momentum , whose component must be equal to @xmath1 or @xmath8 . when describing the circumstances of the discovery , uhlenbeck commented that both he and goudsmit imagined that the electron would be configured as a sphere in permanent rotation , but that they realised that , given the mass of the electron a spin momentum of @xmath1 would require the tangential velocity at its equator to be many times the speed of light .
two years later , the experiments of phipps and taylor showed clearly that the electron does in fact possess a natural angular impulse , @xmath9 , whose value is @xmath1 , and soon after , dirac used schrdinger s equation while substituting the classic equation @xmath10 by its relativistic form @xmath11 , to show that the electron must have a spin of @xmath12 .
robert eisberg hailed dirac s achievement as a great victory for the theory of relativity , since it established electron spin on very sound theoretical foundations , and showed that it is intimately connected to relativity " .
326 - 27 . in 1916
nernst suggested that the quantic fluctuations of space must cause an electromagnetic radiation which would therefore be inherent to space and , consequently , have a relativistically invariant spectrum . in 1958
sparnaay found this radiation when he was measuring the casimir effect at temperatures close to absolute zero .
he detected some radiation , which was independent of temperature and whose spectrum was such that the intensities of its flows are inversely proportional to the cubes of the wavelengths , which is a necessary condition for the radiation to be relativistically invariant . in 1997 , s.k .
lamoreaux carried out new measurements of the intensity of the energy flow of zero - point radiation , using a different method , and reached the same measurements as sparnaay s .
a function of spectral distribution which is inversely proportional to the cubes of the wavelengths implies a distribution of energies which is inversely proportional to the @xmath13 power of the wavelengths , because the energies of the photons are inversely proportional to their wavelengths . in 1969 , timothy h. boyer showed that the spectral density function of zero - point radiation is : @xmath14 where @xmath15 is the number giving the measurement of the wavelength @xmath16 .
this function produces the next for the corresponding energies .
@xmath17 for @xmath18
there must be , therefore a threshold for @xmath16 , which hereafter will be designated by the symbol @xmath20 .
* this paper is closely related to
interactions between zero - point radiation and electrons " ( [ 2 ] in references ) and we include here the following extracts from it , to facilitate the subsequent argument .
zero - point radiation proceeds equally from all directions of space , and its interactions with electrons could play the role of the poincar tensions " , preventing the electrons from shattering as a result of the repulsion of their charge against itself . for this to be the case , there must operate the equation : @xmath21_m;\eqno{(1)}\vspace{-6pt}\ ] ] equation ( 17 ) in [ 2 ] , where : @xmath22 : : measurement of the wavelength of the photons with the greatest energy in zero - point radiation , expressed in @xmath20 ( quanta of wavelength ) .
@xmath23 : : measurement of the unit of length , @xmath24 , in the @xmath25 system , expressed in @xmath20 .
@xmath26 : : measurement of the radius of the electron , expressed in @xmath24 .
@xmath27_m=\que7{48}b-\que{11}{50}b^2+\cdots+t_mb^m;\ ] ] where @xmath28 and @xmath29 \que1{m+3},\vspace*{-9pt}\ ] ] p.9 in [ 2 ] .
the hypothesis that zero - point radiation is also the effective cause of gravitational attraction between two electrons leads to the equation : @xmath30_m}{g_e};\eqno{(2)}\ ] ] equation ( 20 ) in [ 2 ] .
in [ 2 ] it was also deduced that @xmath31 @xmath32 where @xmath33 is the numerical coefficient of the gravitational constant in the @xmath25 system @xmath34 , where @xmath35 is the planck length .
@xmath36 x=5.257601\ti10^{27}\end{array}\right\}\eqno{(4)}\ ] ] this paper uses the @xmath25 system of units in which the basic magnitudes are @xmath37 electron charge , @xmath38 , electron mass , @xmath7 , and the speed of light , @xmath39 .
the unit of length is the classic radius of the electron @xmath40 , the unit of time is @xmath41 , and that of quantity of action is @xmath42 .
therefore @xmath43 , but he symbol @xmath44 will not be used .
we know that photons have the following properties : 1 .
they are energy packets " which possess the quantity of action @xmath45 , and can behave as particles .
they move at the speed of light along rectilinear trajectories . these are geodesic lines , in a space without curvature . within the space defined by the 3-dimensional spherical surface @xmath46 , the geodesic lines are great circles , but in the case of our universe
, the immense size of @xmath47 makes it very difficult not to confound them with straight lines .
their wave - motion follows planes perpendicular to their trajectories , producing a spin of @xmath48 .
they only differ from each other in the quantity of energy , @xmath49 , which is contained in the packets in question .
this quantity is given by the equations @xmath50 , in which @xmath51 is the frequency ( number of photons per time - unit ) and @xmath16 is the wavelength . since @xmath51 and @xmath16 are linked by the relation @xmath52 , we need to consider only one of these variables .
our aim , to investigate possible configurations of photons , implies that it is preferable to characterise them according to their wavelength @xmath16 .
all photons of wavelength @xmath16 are identical .
we must make a brief digression here .
the wave - motion of photons is very different from that of the waves of the sea . except where they break
, the wave - motion of the sea waves is produced as a vertical movement , and is propagated through a medium which does not move except in that direction .
this wave - motion does not carry with it anything of substance , being merely a wave - motion within a uniform substance . in the case of photons ,
their wave - motion does carry substance with it , that of the energy packets"having a quantity of action @xmath53 which is what they are , while they are not limited in their movement to the wave - motion which gives rise to their spin @xmath0 , but advance in a straight line at the speed of light .
we will now analyse the following suggested configuration for photons : the substance of the photons undulates along cylindrical helices , in such a way that the advance of one wavelength along its rectilinear trajectory is accompanied by an advance of an equal distance over a circumference of radius @xmath54 , having its centre on the said trajectory , and situated on a plane perpendicular to it , so that for every complete rotation over that circumference , there will be an advance of one wavelength over the rectilinear trajectory . * to help the analysis of the suggested configuration , there are shown in fig .
1 the projections over the planes @xmath55 and @xmath56 of a cylindrical helix , whose equations are : @xmath57 z = r\sin\left(\que xr\right)\\ \\[-8pt ] y = r\cos\left(\que xr\right)\end{array}\right\}\eqno{(5)}\ ] ] during the time elapsed between @xmath58 and @xmath59 , the point on the helix has advanced a distance of @xmath60 along the axis @xmath61 , and has made one complete rotation around that axis , which presupposes a journey of @xmath60 over the circumference of the projection on the plane @xmath56 . both during this journey and during that over @xmath61 , the velocity has been @xmath39 ( obviously , these speeds are not to be added together , but seen as independent of each other )
. the advance of @xmath62 over the axis @xmath61 , i.e. one wavelength , @xmath16 , is matched by a length of the helix traced over the cylinder of rotation with a radius of @xmath63 , which is given by @xmath64 @xmath65 the length of the cylindrical helix corresponding to a distance of one wavelength is equal to @xmath66 .
seeing that electromagnetic wave - motion must run over a cylindrical rotational surface , it must do so following a cylindrical helix , since for cylinders , whether rotational or not , their helices are their geodesic lines .
for every distance of one wavelength , @xmath16 , along @xmath61 , there can be only one corresponding photon .
a photon whose energy is equal to @xmath67 ( the energy equivalent of the mass of the electron ) has a wavelength of @xmath68 , to which there corresponds @xmath69 , and its spin momentum will be @xmath70 which is the right value for photons .
in order to develop the hypothesis proposed in this paper , it will be better to limit ourselves initially to considering the electron , and then move on to the remaining elementary particles .
the basic suppositions for all of them are the following : 1 . the transformation of a photon into a particle does not require any change in its substance , but only a change in the organisation of that same substance .
it continues to be the same packet of energy with @xmath71 quantity of action " which , instead of moving lineally and forming part of a train of waves moving at the speed of light , has turned round on itself to bite its tail " , creating an isolated entity rotating in such a way that at any point on it , the tangential velocity is @xmath39 .
2 . the photon does not lose its essentially linear nature in favour of a spherical surface .
what happens is that it rolls itself around a sphere of radius @xmath47 , in such a way that during the period @xmath72 it makes the journey of @xmath73 around the great circle @xmath74 of radius @xmath47 , which meanwhile completes a rotation of @xmath75 radians around the axis @xmath76 ( see fig .
2 ) . at the end of the said period @xmath72 ,
the point which was initially at @xmath77 has moved to @xmath78 , whose coordinates are : @xmath79 y = on\cos\va\sin\psi\\[+4pt ] z = on\sin\va\end{array}\ ] ] + and where @xmath80 ; @xmath81 : @xmath82 y = r\cos\va\sin\va\\[+4pt ] z = r\sin\va\end{array}\right\}\eqno{(6)}\ ] ] the journey of the photon over curve ( 6 ) may be considered as being made over a great circle which turns around the axis @xmath83 in such a way that the angles @xmath84 and @xmath85 are equal .
the journey over the great circle is equivalent to the photon s journey over @xmath86 in fig .
1 . both have a length of @xmath62 , and neither of them generates angular momentum . during its journey over the great circle ,
the angular momentum would follow a diameter at the equator . however when
it has completed a rotation around every point on the equator , it will be balanced by its opposite going in a contrary direction so that the sum of the two is zero .
the rotation around @xmath83 generates at every point a momentum determined by the tangential velocity @xmath87 and by the radius @xmath88 . in other words : @xmath89 where the value @xmath90 is determined by the equal distribution of the energy of the photon in a journey of @xmath91 radians
. therefore we arrive at : @xmath92 for @xmath93 , @xmath94 . from equations ( 6 ) we obtain : @xmath95 dy = r(\cos^2\va-\sin^2\va)d\va = r\cos2\va\,d\va\\[+4pt ] dz = r\cos\va\,d\va\end{array}\ ] ] therefore , @xmath96 , whence @xmath97 and , finally : @xmath98 which is somewhat greater than the distance travelled along the great circle between @xmath99 and @xmath100 .
the length which corresponds to the wave motion is @xmath101 compared with @xmath102 , which means a relation 1.216 . in fig .
1 the relation was @xmath103 for a photon with energy @xmath67 , we obtain @xmath104 , @xmath69 .
the radius of the electron is @xmath24 and its circumference @xmath105 .
therefore a `` rolling up '' of @xmath106 would mean making @xmath107 turns around a circumference of radius @xmath24 , i.e. a length of @xmath108 , which would be covered in a period of @xmath109 if a speed of @xmath39 is maintained .
we do not need to introduce a new diagram or new equations ; but need only imagine fig .
2 at a scale of @xmath110 , i.e. 1/137.0360 and with the wavelength of the photon , @xmath106 , rolled up 137.0360 times around the sphere whose radius will now be @xmath24 . in equations ( 6 ) , it is sufficient to replace @xmath69 by @xmath111 .
the energy @xmath67 will be distributed equally over the length of the wavelength @xmath112 , so that the value of the spin momentum will still be @xmath1 .
if at any time the `` rolling - up '' should come undone while the `` tail - biting '' remains , centrifugal force would cause it to change into a circumference of radius @xmath113 , rotating at the same tangential velocity @xmath39 and generating the angular momentum @xmath0 . a great circle of radius @xmath24 , making one turn around the axis @xmath76 every @xmath114
, i.e. every @xmath115 seconds , would be very hard to distinguish from a spherical surface of radius @xmath24 . as noted above , goudsmit and uhlenbeck discovered in 1925 that the electron must have a spin of @xmath1 , and
deduced that if it were configured as a spherical surface , the radius of that surface would have to be greater than that of an entire hydrogen atom ; otherwise , the tangential velocity at its equator would have to be much greater than that of light .
this is perhaps the only possible experimental support for the conjectural `` rolling - up '' of the wavelength @xmath106 around the sphere of radius @xmath3 and the wave equations ( 5 ) , with @xmath111 instead of @xmath69 . the value of @xmath110 , which implies the reduction of scale between the world of electromagnetic waves and the world of particles possessing mass , is due to the interaction of the latter particles with zero - point radiation which , thanks to the compton effect , compresses them until equilibrium is reached with electrostatic repulsion , which in the case of the electron happens when @xmath111 . for the photon which has an energy equal to @xmath67
, we have @xmath116 , @xmath113 and @xmath111 .
the decrease from @xmath117 , which is a characteristic of the wavelength @xmath118 , to @xmath119 , which is a characteristic of the particle , is always the same ; for any elementary particle with mass @xmath120 , @xmath119 is always equal to @xmath121 .
the fine structure constant @xmath110 , the wavelength @xmath20 of the most energetic photon and the speed of the light , @xmath39 , are three fundamental and invariable physical entities which modulate the architecture of our universe .
the suggested configuration implies that the whole of the substance of the photon transformed into an electron maintains its coherence , and revolves around the centre of the sphere of radius @xmath24 , over whose surface it moves , in such a way that the tangential velocity at all its points is equal to that of the original photon , i.e. that of light .
the `` energy packet '' of dimensions @xmath122 , which in the configuration described here revolves as a single entity around a point , from which all its substance is distant by @xmath24 , in such a way that its tangential velocity is always @xmath39 , implies a mass of @xmath123 and a centrifugal force of @xmath124 .
this force is exactly equal to that of the electrostatic repulsion of the electron s charge against itself on the surface of the particle , i.e. to : @xmath125 the equivalence just stated leads us to suggest that the electrostatic field determined by the electron s charge , could proceed from the interaction between the centrifugal force inherent to its spin and zero - point radiation [ 2 ] .
figure 3 shows a section of the electron with centre at @xmath126 , following approximately a spire of the path given by equations ( 6 ) .
the arrow @xmath127 shows the tangential velocity `` @xmath39 '' , constant along its length , while the other arrow @xmath128 shows the centrifugal force inherent to the mass of the electron , its radius and the said tangential velocity .
if we imagine that the dimension of the radius of the electron extends as far as the surface of the electron with centre at @xmath129 , its centrifugal force at the point of contact @xmath130 will be @xmath131 i.e. will be equal to the electrostatic repulsion between the two electrons . when the particle with centre at @xmath126 and the particle with centre at @xmath129 have identical spins , the arrow which follows the circle with centre at @xmath126 and radius @xmath132 faces at @xmath130 an opposed arrow following the circle with centre at @xmath129 and radius @xmath3 . when their spins are opposed , those arrows are in agreement . in the first case ,
the force @xmath133 is one of repulsion , in the second one of attraction . for this to be possible
, the spin of the elementary particles must be always around the same direction , which can be only the centre @xmath134 of the surface @xmath46 , which is the configuration of the universe .
this is not impossible , as can be seen if we imagine minor circles on the surface @xmath135 , and assume that they spin around axes pointing towards the centre of that surface . in fig .
3 , we can also see how all this could be achieved through the interactions of zero - point radiation and electrons .
the photons which fall on any point on the surface of the electron have three times the energy necessary to balance the centrifugal energy generated by the rotation of the mass @xmath136 around @xmath134 with tangential velocity `` @xmath39 '' , see [ 2 ] pag .
10 , and the sum of their projections over any given direction is equal to a third of that energy .
the radius of every electron measures @xmath3 , and the line parallel to @xmath137 from @xmath138 on the surface of the electron with centre at @xmath126 reaches a point @xmath139 on the surface of the electron with centre at @xmath129 . the photons which reach @xmath138 disperse from that point in every direction , after having generated the centripetal force required to balance the centrifugal force at that point , and the intensity of the energy flows emerging from @xmath138 is @xmath140 , at the distance @xmath141 from @xmath138 .
the sum of all the energy flows which emerge from the hemisphere of the electron with centre at @xmath126 facing the electron with centre at @xmath129 is @xmath142 , for each @xmath143 , at the distance @xmath3 from @xmath126 and @xmath144 , for each @xmath143 at the distance @xmath132 from @xmath126 .
the sum of their projections over the direction @xmath137 is @xmath145 for each @xmath143 , which generates a force @xmath146 in this direction , when reaching the points @xmath147 at the surface of the electron with centre at @xmath129 .
the presence of the electron with centre at @xmath129 has a symmetrical effect on the electron with centre at @xmath126 , whence the apparent force of rejection @xmath148 .
the centrifugal forces , @xmath150 , inherent to the spin of the electron with the configuration suggested in section 3 of this paper , are equal to the centripetal forces , @xmath151 , resulting from the interactions of the electron with the zero - point radiation , whose intensity given by equation ( 15 ) in [ 2 ] is independent from the characteristics of the said particle . in other words
@xmath152 the quantic threshold @xmath153 , where `` @xmath120 '' is the measurement of the mass of the particle expressed in @xmath7 , and `` @xmath119 '' is that of the radius of the same particle expressed in @xmath24 , allows us to write : @xmath154 for @xmath155 , @xmath156 , so that the total of the centrifugal forces will be less than the total of the centripetal ones , and these centripetal forces will squeeze the particle , whose radius will decrease to @xmath157 .
however , its mass will remain the same , so that we have @xmath158 , which violates the said quantic threshold .
therefore , there can not exist elementary particles of spin @xmath1 with a mass of less than 1 @xmath7 , or with a radius greater than 1 @xmath24 .
the spherical surface of radius @xmath111 is the geometrical location at which there is produced an equilibrium between the centrifugal forces inherent to elementary particles of spin @xmath1 , and the centripetal forces resulting from the particles interactions with zero - point radiation .
such an equilibrium only happens for particles of mass @xmath7 , i.e. electrons , and this shows clearly their singular character as fundamental components of the architecture of the universe . for @xmath159 ,
the quantic threshold @xmath160 prevents a particle of mass @xmath120 from expanding beyond a spherical surface of radius @xmath161 , and the centrifugal force determined by a mass @xmath162 spinning with a radius of @xmath163 and a tangential velocity `` @xmath39 '' is : @xmath164 for the proton , @xmath165 ; @xmath166 , and therefore @xmath167 .
faced with this centrifugal force , more than 3 million times greater than the centripetal , a particle of mass @xmath159 could only reach equilibrium by expanding its radius to measure @xmath168 .
this is impossible because the quantic threshold @xmath160 prevents it .
the enormous difference between @xmath169 and @xmath151 at the limit imposed by this quantic threshold gives rise to new quantic fluctuations of space , distinct from those inherent to its curvature which are the cause of zero - point radiation .
these new fluctuations produce flows of photons in the direction of the centre , @xmath126 , of the particle in question , and their intensity is that which is needed in order to balance out the centrifugal and the centripetal forces , i.e. a force of : @xmath170 for the entire particle . in the case of the proton , @xmath171 and we can assume that @xmath172 , allowing for a relative error @xmath173 . in the case of the electron @xmath174 , and
this does not allow for the appearance of any new quantic fluctuations of space .
in [ 2 ] it was stated that , in the case of the electron , the intensity of the energy flow per @xmath175 is that which is required to generate the force : @xmath176.}\ ] ] the mass @xmath2 spinning at a tangential velocity `` @xmath39 '' and with a radius @xmath3 causes a centrifugal force of @xmath177 . as remarked above this force
is equal to the electrostatic repulsion against itself of the electrical charge `` @xmath38 '' confined into a spherical surface of radius @xmath24
. we must here remember that the quantic threshold @xmath160 does not only imply the spin @xmath1 , but also that the electrostatic repulsion @xmath178 is equal to the centrifugal force at @xmath179 . on the spherical surface of radius @xmath111 ,
i.e. the basic spherical surface , the centrifugal force corresponding to the mass @xmath162 spinning at a tangential velocity `` @xmath39 '' is : @xmath180 where the symbols into brackets mean units of the @xmath181 system . this same mass , @xmath162 , when spinning at a tangential velocity `` @xmath39 '' and radius of spin @xmath182 , gives rise to the centrifugal force : @xmath183 because @xmath184
. it therefore follows that between a point a distance of @xmath3 from the centre of the particle , and a point at a distance of @xmath185 from that centre , the centrifugal force of the elementary particles is inversely proportional to the distance to their centres , whilst the intensities per @xmath175 of the energy flows around to balance the centrifugal and the centripetal forces at a distance @xmath186 from @xmath134 , would concentrate from @xmath187 each @xmath143 , at the distance @xmath188 , to @xmath189 each @xmath143 , at the distance @xmath190 from @xmath134 , which means intensities inversely proportional to the squares of the distances . in the following lines we will suggest a hypothesis to explain this . * we have supposed that the flows of photons which arise to balance the centrifugal force on the surface of nucleons , are directed towards the centres of these particles . in this respect , they act differently from zero - point radiation , which is made up of photons which can follow any trajectory .
the radius of the electron is extremely small , @xmath191 cm , but the elementary areas @xmath175 which compose its surface are even smaller .
they number @xmath192 and the distance between the centres of contiguous areas is @xmath193 , where @xmath35 is planck s length .
we have no experience of the behaviour of the photons flows which , separated by distances of this order of magnitude , are almost in parallel , but we can suppose that they can interreact with each other , with the photons behaving as particles , instead of doing so through the interference of their respective waves .
if so , the number of collisions per unit of time will be directly proportional to the densities of the flows , and inversely proportional to the squares of the distances to the centre towards which they converge .
it is logical to suppose that these collisions between particles will produce changes of direction , which will surely begin at a distance from the centre which is much greater than the radius of the electron .
initially , the frequency of these collisions will be insignificant .
we know that the intensity of the flows of photons which converge towards the centre of the proton is , at a distance @xmath194 from that centre , that which is needed for their interactions with the particle to produce a centripetal force per @xmath175 which equals the centrifugal force generated by the particle s spin , of radius @xmath195 and tangential velocity @xmath39 ; in other words , equal to @xmath196 , where @xmath4 is the mass of the proton , and @xmath197 , the number which results from measuring the radius of the proton in units of length @xmath24 .
if the intensities of the photon flows diminish , starting at a distance @xmath195 from the centre of the proton , @xmath134 , in inverse proportion to the square of the distance to it , we obtain , for the distance @xmath198 : @xmath199 for @xmath200 , we obtain : @xmath201 however , this intensity has been dispersed over the @xmath202 of the surface of the electron , from the @xmath203 of the surface of the proton , so that the intensity of the flow on the surface of the electron must be multiplied by @xmath204 , with the result that at a distance of @xmath24 from @xmath134 , the flow is of @xmath205 , which could cause a force equal and opposite to that of the mass of the proton rotating with a radius of @xmath24 and a tangential velocity of @xmath39 , since @xmath206 the photons which have been dispersed from their trajectories towards
@xmath134"continue along different ones , but all the photons which have entered the spherical surface of radius @xmath207 will exit from it , and the sum of the possible exchanges of energy between them will be nil , thus producing an exact equality between the energies which enter and exit . for any distance @xmath208 , we can write : @xmath209 for a distance @xmath24 from @xmath134 , we must multiply by @xmath210 , from which results @xmath211 , whence @xmath212 ; that is , equal to @xmath213 multiplied by the inverse of the relation of distance from @xmath134 " . * * figure 4 shows two elementary particles which are @xmath214 apart , the first having mass @xmath162 and centre at @xmath126 , and the second having mass @xmath215 and centre at @xmath129 .
the energy flow aroused by the inferiority of the centripetal forces to the centrifugal ones on the surface of the particle with centre at @xmath126 and mass @xmath120 , has at @xmath216 , which is @xmath3 distant from @xmath126 , an intensity of : @xmath217 at a distance of @xmath218 from @xmath126 , this flow , which proceeds from the @xmath219 areas @xmath175 which are @xmath220 distant from @xmath126 , is spread over @xmath221 , with a resulting diffusion of @xmath222 .
the lessening of the intensity of the flows @xmath223 which flow in from each of the @xmath219 areas of @xmath224 on the spherical surface of radius @xmath220 , is not due to a spatial diffusion of the demand from the area @xmath225 to the area @xmath226 ; each of these maintains its identity from one of the @xmath227 areas @xmath175 , which are @xmath228 distant from @xmath126 , up to the @xmath219 areas @xmath175 on the said spherical surface .
what is produced is a temporal diffusion . during the lapse of time `` @xmath143 '' the flows onto each @xmath175 on the spherical surface of radius @xmath220 proceed from any one of those which belong to the spherical segment of one base which is determined by the intersection of the cone with vertex at @xmath126 and base on that @xmath175 , with the spherical surface of radius @xmath214 and centre at @xmath126 . it may be any one of those which exist at the said intersection , and the probability of being any particular one during the lapse @xmath229 is @xmath230
therefore , the intensity of the flow which arrives from any one of them at one of those situated at a distance of @xmath220 from @xmath126 is : @xmath231 the flow @xmath232 per @xmath175 each @xmath143 , aroused at the area @xmath233 of the spherical segment of one base , height @xmath234 and centre at @xmath216 , falls every @xmath235 on the analogous spherical segment with centre at @xmath236 , on the spherical surface with centre at @xmath129 and radius @xmath220 . from here , and in response to the demand provoked by the inferiority of the centripetal forces to the centrifugal ones on the surface of the elementary particle of mass @xmath237 and centre on @xmath129 , it converges towards @xmath129 , suffering interferences which causes its intensity per @xmath175 at @xmath238 to be : @xmath239 and it causes there to be , on the @xmath175 situated at @xmath238 a force along @xmath240 with an intensity of : @xmath241 the presence at @xmath126 of a particle of mass @xmath120 determines that there exists at @xmath242 , on @xmath243 and at a distance @xmath119 from @xmath126 , a force of equal intensity directed towards @xmath129 , so that there results an apparent attraction between the two particles given by : @xmath244 in the @xmath25 system , the value of the numerical coefficient of the gravitational constant is @xmath245 .
therefore the force @xmath246 is equal to the gravitational attraction between two particles which are @xmath132 apart and whose masses are @xmath120 and @xmath237 respectively . *
the photon flows of zero - point radiation which fall incessantly on every @xmath175 of the surface of an elementary particle of mass @xmath120 and radius @xmath247 , with the intensity needed to generate a force equal and opposite to the centrifugal force : @xmath248 must necessarily oppose that force at all times .
the spin @xmath1 is a quantic threshold ; the flows can not be converted into mass added to that of the particle , and since they can not accumulate in it they must pass out of it .
they can do this along a tangent to the line which follows the photon rolled up around the spherical surface of radius @xmath119 ( see the hypothesis suggested in section 3 of this paper ) producing a configuration of photons with spheres of increasing radii , and flows whose intensities decrease in inverse proportion to the squares of those radii .
their intensity at the spherical surface of radius @xmath119 is that which is needed to equal the electrostatic repulsion against itself of the charge @xmath38 " , distributed evenly over that surface , in other words , to equal the force @xmath249 at a distance of @xmath3 from @xmath126 , the intensity @xmath250 is that which is needed to equal the force @xmath251 , and at a distance @xmath141 , the intensity @xmath252 is that which is needed to equal the force @xmath253 the revolutions along the said spheres are in the direction in which the photon is rolled up into the particle . if they meet with another particle which has the photon rolled up in the same direction , they will go in the opposite direction to it ( see fig . 5 ) and , since their uninterrupted flows can not have any effect on the particle in question , they must exit from it . and if , previously , flows directed towards the centre of the particle must produce exits along tangents to the lines of rolling - up , the flows which follow the tangents will now give rise to a single flow towards @xmath126 , having an intensity of @xmath254 , where @xmath255 is the distance @xmath240 .
since we are now dealing with all the flows which strike upon the hemisphere of the particle or which are emitted from it , the value of the force of repulsion directed along @xmath240 is : @xmath256 and , since the presence of the particle at @xmath129 has the same effect on the particle whose centre is at @xmath126 , there results an apparent repulsion between the two , which is equal to the quotient @xmath257 , which is @xmath258 times more intense than gravitational attraction .
if the particle at @xmath129 is configured as a photon rolled up in the direction opposite to that of the particle whose centre is at @xmath126 , there will be produced an attraction instead of a repulsion .
the hypothesis set out in section 3 of this paper enables us to explain how a spin of @xmath1 can be produced in the elementary particles . through the help of that hypothesis and suggestions made in the present section
, we can also explain the causes of electrostatic attraction and repulsion and of the existence of positive and negative electric charges , and how gravitational attraction and the electrostatic forces are due to one and the same cause ; this compels the intensity of the electrostatic forces to be @xmath259 times that of gravitational attraction , if @xmath260 where @xmath24 is the classical radius " of the electron , and @xmath35 the planck length .
this simple relation between the gravitational constant , @xmath261 the fine structure constant @xmath110 , the classical radius " of the electron , @xmath3 , and the planck length @xmath262 , can be admired as a fascinating harmony within the world of the elementary particles and fundamental forces .
the title of this section does not mean that we mean to propose a model for atomic nuclei , but that , much less ambitiously , we want to show that the suggestions made in the previous sections lead to the conclusion that if they fit with physical reality , the energy flows resulting from the presence of the protons and neutrons which make up atomic nuclei , would possess the intensity needed to ensure the cohesion of those nuclei .
the reasoning developed in order to reach this aim rests on a basis consisting of the said suggestions , through the following premises , taken from [ 4 ] , pp .
48 - 50 : 1 .
the number of nucleons , @xmath77 , per unit of volume is roughly constant , @xmath263 constant , whence @xmath264 , where @xmath47 is the radius of the nucleus .
2 . from electron scattering measurements
it is concluded that : @xmath265 3 .
the density charge changes very little from the lightest nuclei to the heaviest up to a certain point , and then declines slowly over a certain distance , over which the charge falls from 90% to 10% of its central value .
the length of this distance is approximately @xmath266 .
4 . for light elements ,
the number of protons @xmath267 is approximately equal to the number of neutrons @xmath268 . for heavy elements , @xmath269 ( see [ 4 ] , p. 70 ) .
the binding force for each nucleon is roughly equal to @xmath270 ( see [ 4 ] , p. 67 ) .
the nuclear force must be spin - dependent ( see [ 4 ] , p. 92 ) . 7 . in [ 4 ] , fig .
3.4 on p. 49 , we can see the curves which express the radial distribution of the charges for certain nuclei .
the lightest is that of c@xmath271 , with 6 protons and 6 neutrons @xmath272 , and the heaviest is that of pb@xmath273 , with 82 protons and 126 neutrons @xmath274
@xmath275 @xmath276 @xmath277 .
+ in the nucleus of c@xmath271 , the radius of the nucleus measures @xmath278 , and 90% of its charge is within a sphere with its centre at the centre of the particle , and with a radius of @xmath279 . in the case of pb@xmath273 ,
the radius of the nucleus measures @xmath280 , and 90% of the particle lies within a sphere with the same centre and a radius of @xmath281 .
figure 6 presents the case of two elementary particles , one of them with centre at @xmath126 , a mass of @xmath162 , and a radius of @xmath247 .
the other has its centre at @xmath129 , at a distance of @xmath282 from @xmath126 , a mass of @xmath215 , and a radius of @xmath283 . in this figure ,
the symbol @xmath284 signifies the intensity at @xmath77 of the energy flow required in order to generate the force @xmath285 per @xmath175 .
the presence of the particle with centre at @xmath129 and radius @xmath286 prevents the flows aroused in the spherical segment of one base @xmath287 , on the spherical surface of radius @xmath288 , from arriving at the particle with centre at @xmath126 .
the area of this segment is : @xmath289 @xmath290 where @xmath291 whence @xmath292(q_\la)^2.\ ] ] the blocking of the flows @xmath293 , aroused in the areas @xmath294 of the spherical segment of one base @xmath295 , prevents them from becoming concentrated on the spherical segment which corresponds to the spherical surface of radius @xmath119 and centre at @xmath126 , whose surface is @xmath296(q_\la)^2.\ ] ] such a concentration would cause their intensity per @xmath175 on the surface of radius @xmath119 and centre at @xmath126 to be @xmath297 times @xmath298 ; however , the interference between flows , as explained above , causes that intensity per @xmath175 to be only @xmath299 , which implies that the total of the energy flows which do not arrive from the surface of the spherical segment of one base which has an area of @xmath294 , to the spherical segment which has an area of @xmath300 , is that required to generate a force of : @xmath301 \que{m_x}{m_y}\left(\que{m_el_e}{t_e^2}\right)\ ] ] the electrostatic repulsion at a distance @xmath302 from @xmath126 determines a bundle of forces equal and opposed to the forces @xmath303 , and when they are not counteracted by the latter forces , they move towards the exterior .
the sum of the projections of these forces along the line @xmath240 determines a force along @xmath304 .
figure 7 shows the generation of the spherical surface with centre at @xmath126 and radius @xmath186 , through differential surface elements , consisting of rings of radius @xmath305 and width @xmath306 orientated along the tangent at @xmath147 .
in effect , @xmath307 whence @xmath308 the area of the spherical surface of radius @xmath163 is : @xmath309 the force of intensity @xmath310 , directed along the radius @xmath311 , can be broken down into two : @xmath312 , its projection along the radius @xmath313 , and @xmath314 , which is cancelled out when added to the equal and opposing force which corresponds to @xmath315 along the radius @xmath316 , which is symmetrical in @xmath313 with respect to @xmath311 .
this leads to the result : @xmath317 ; whence @xmath318 , and : @xmath319^\va_0 \\ \\[-4pt ] \disp = f_{xx}\pi(r_xl_e)^2\sin^2\va,\end{array}\eqno{(12)}\ ] ] fig .
8 shows that @xmath291 , and also @xmath320 by introducing these values into ( 12 ) , we obtain : @xmath321 remembering that @xmath188 , we finally arrive at : @xmath322 where @xmath26 radius of the proton , and @xmath323 radius of the neutron , @xmath324 ; @xmath325 and when both particles are equal @xmath326 .
when we consider the reciprocal effect of the particle with centre at @xmath126 , on the particle with centre at @xmath129 , and the sum of the two , we obtain : for the pairs proton - proton " and neutron - neutron " : @xmath327 for the pairs proton - neutron " : @xmath328 for heavy nuclei , the number of protons @xmath329 is approximately equal to @xmath330 .
we must therefore combine the electrostatic force of repulsion 0.41 times with that of gravitational attraction , i.e. a mutual attraction of @xmath331 , with a mutual repulsion of @xmath332 ; this implies a slight predominance of the centripetal forces , which can be compensated by the centrifugal forces of the orbits of the nucleons within the atomic nucleus .
this complies with the proposition initially stated , which was to show that the energy flows which arise to balance out the centripetal and centrifugal forces in elementary particles of mass @xmath159 have sufficient intensity to generate the strong interaction needed to ensure the cohesion of atomic nuclei .
* it will be useful to consider another method of showing that the energy flows which arise to balance out the enormous differences between the centripetal and centrifugal forces on the surface of the nucleons , are sufficient to generate a gravitational attraction with an intensity analogous to the electrostatic repulsion between two protons whose nuclei are separated by a distance of less than @xmath3 . to make a first approximation , we can firstly consider the mean distance between the particles .
premises 1 , 2 and 3 taken from [ 4 ] lead to the equation : @xmath333 whence @xmath334 .
the number of protons is approximately equal to @xmath335 for heavy nuclei , where @xmath77 is the number of nucleons , as it is said in premise 7 taken from [ 4 ] . within a sphere of radius @xmath3 around the centre , @xmath126 , of a proton , the intensities of gravitational attraction decrease in inverse proportion to the distance from @xmath126 , but the intensities of electrostatic repulsion decrease in inverse proportion to the square of this distance .
a proton of mass @xmath4 and radius @xmath336 possesses @xmath337 areas @xmath175 on its surface .
all of them play a part in electrostatic repulsion , but on the gravitational attraction between the protons with centre at @xmath126 and another nucleon with centre at @xmath129 , less than @xmath288 away from @xmath126 , the only influence is that exerted by the energy flows intercepted by the nucleon with centre at @xmath129 and direction @xmath240 .
the centripetal force corresponding to the mass @xmath4 spinning at a tangential velocity of @xmath39 " and a radius of spin of @xmath288 is @xmath338 and the centrifugal force per unit of surface is : @xmath339 because @xmath340 .
the energy flows which could cause gravitational attraction between the particle of mass @xmath4 , radius @xmath195 and centre at @xmath126 and the similar particle with centre at @xmath129 , are those which , being directed towards @xmath126 fall on the spherical segment of one base , which belongs to the sphere with centre at @xmath126 and radius @xmath288 , whose height is @xmath341 and which has its centre at @xmath77 .
the area of this spherical segment is @xmath342 , and the projection of this area over the plane perpendicular to @xmath343 at @xmath126 does not differ significantly from @xmath344 .
the sum of the flows which fall on the half spherical surface with centre at @xmath126 and radius @xmath345 is proportional to @xmath346 and the projection of this surface over the said plane perpendicular to @xmath343 at @xmath126 has the area @xmath347 .
the quotient @xmath348 is the factor of area , @xmath349 , by we must multiply the force per unit of surface at the distance @xmath350 from @xmath126 , in order to obtain half the intensity of the energy flow which generates the gravitational force .
this flow arrives at a distance @xmath132 from @xmath126 multiplied by the factor @xmath351 , because up to distance @xmath288 from @xmath126 the said flow decrease in intensity in inverse ratio to the distances from @xmath126 , and not to the square of those distances . taking all of this together ,
we obtain @xmath352 .
the presence of the particle with center at @xmath126 exerts an equal and opposite effect on the particle with center at @xmath129 , so that we finally obtain : @xmath353 @xmath354 @xmath355 trying for a better approximation we can write : @xmath356 @xmath357,\ ] ] in which @xmath358 , are the distances between a proton and the nucleons which are at a distance of less than @xmath3 from its centre . premise 3 in [ 4 ] implies a uniform distribution of charges in the heart of the nucleus , and a mean distance of @xmath359 determines that in such a case there are within a sphere of radius @xmath3 around the proton : @xmath360 for @xmath361 , @xmath362 ; @xmath363 and we obtain @xmath364 whence : @xmath365 which is a very good approximation to the true value @xmath366
[ 1 ] r. eisberg y r. resmic : _ fsica cuntica_. ed .
limus s.a .
1983 , pp .
326 - 328 . | in this paper we suggest a configuration of photons consistent with a spin @xmath0 , and a configuration of the fermions coherent with a spin @xmath1 .
these suggested configurations open the way to further analyses which lead to the following conclusions : * there can not exist elementary particles of spin @xmath1 with a mass inferior to @xmath2 or with a radius greater than @xmath3 . *
the electrostatic force derives from the centrifugal forces inherent to the spin and are propagated by photons . *
the derivation of the electrostatic force explains the existence of positive and negative charges and coulomb s law . *
the enormous differences between the centrifugal forces and the centripetal forces at the surface of the protons give rise to quantic fluctuations of space which generate the energy flows necessary for equilibrium .
these energy flows can explain gravitation and the strong force .
* the mass of the proton , @xmath4 , and the mass of the neutron , @xmath5 , must each have a concrete value required for the cohesion of the atomic nuclei .
the suggested hypotheses show that the relation @xmath6 must be very slightly greater than 1.00 , and that , by a first approximation , both @xmath5 and @xmath4 must be slightly less than 1851 @xmath7 .
* there are no `` gravitons '' and no `` gluons '' ; the `` messenger particles '' are always the very real photons .
= = = = = = = = = = = = = = = = = = = = = = = = = = = # 1#2 = = = = = |
we are given a random variable @xmath0 with outcomes in a set of @xmath1 of source words with associated probabilities @xmath2 with @xmath3 , and a code word subset of @xmath4 , the set of finite binary strings .
let @xmath5 denote the _ length _ ( number of bits ) in @xmath6 .
let @xmath7 .
the _ optimal source coding problem _ is to find a 1:1 mapping @xmath8 , satisfying @xmath9 is not a proper prefix of @xmath10 for every pair @xmath11 , such that @xmath12 is minimal among all such mappings .
codes satisfying the prefix condition are called _ prefix codes _ or _
instantaneous codes_. this problem is solved theoretically up to 1 bit by shannon s noiseless coding theorem @xcite , and exactly and practically by a well - known greedy algorithm due to huffman @xcite , which for @xmath13 source words runs in @xmath14 steps , or @xmath15 steps if the @xmath16 s are not sorted in advance .
if @xmath17 achieves the desired minimum , then denote @xmath18 .
we study the far more general question of length restrictions on the individual code words , possibly different for each code word .
this problem has not been considered before .
the primary problem in this setting is the problem with equality lengths restrictions , where we want to find the minimal expected code - word length under the restriction of individually prescribed code - word lengths for a subset of the code words .
apart from being a natural question it is practically motivated by the desire to save some part of the code tree for future code words , or restrict the lengths of the code words for certain source words to particular values .
for example , in micro - processor design we may want to reserve code - word lengths for future extensions of the instruction set .
no polynomial time algorithm was known for this problem .
initially , we suspected it to be np - hard . here , we show an @xmath19 dynamic programming algorithm .
this method allows us to solve an integer programming problem that may be of independent interest .
the key idea is that among the optimal solutions , some necessarily exhibit structure that makes the problem tractable .
this enables us to develop an algorithm that finds those solutions among the many possible solutions that otherwise exhibit no such structure .
formally , we are given _ length restrictions _
@xmath20 , where the @xmath21 s are positive integer values , or the _ dummy _ @xmath22 , and we require that the coding mapping @xmath23 satisfies @xmath24 for every @xmath25 with @xmath26 .
for example the length restrictions @xmath27 mean that we have to set @xmath28 and @xmath29 , say @xmath30 and @xmath31 .
then , for the remaining @xmath32 s the coding mapping @xmath23 can use only code words that start with @xmath33 .
we assume that the length restrictions satisfy below , the kraft s inequality @xcite , @xmath34 where we take @xmath35 for @xmath36 , since otherwise there does not exist a prefix code as required .
* related work : * in @xcite a variant of this question is studied by bounding the maximal code - word length , which results in a certain redundancy ( non - optimality ) of the resulting codes . in @xcite both
the maximal code - word length and minimal code - word length are prescribed .
shannon s noiseless coding theorem @xcite states that if @xmath37 is the entropy of the source , then @xmath38 .
the standard proof exhibits the shannon - fano code achieving this optimum by encoding @xmath32 by a code word @xmath9 of length @xmath39 ( @xmath25 ) . ignoring the upper rounding to integer values for the moment
, we see that @xmath40 for a code that codes @xmath32 by a code word of length @xmath41 .
this suggests the following approach .
suppose we are given length restrictions @xmath42 .
let @xmath43 be the set of equality length restrictions , and let @xmath44 be the minimal expected code - word length under these restrictions given the probabilities .
similar to shannon s noiseless coding theorem , we aim to bound the minimal expected code - word length under equality restrictions below by an entropy equivalent @xmath45 where @xmath46 corresponds to the best possible coding with real - valued code - word lengths
. define @xmath47 if we define @xmath48 s also for the @xmath32 s with @xmath49 such that @xmath50 , then altogether we obtain a new probability assignment @xmath48 for every @xmath32 ( @xmath25 ) , which has a corresponding shannon - fano code with code lengths @xmath51 for the @xmath32 s .
moreover , with respect to the probabilities induced by the original random variable @xmath0 , and simultaneously respecting the length restrictions , the minimum expected code word length of such a @xmath52-based shannon - fano code is obtained by a partition of @xmath53 into @xmath48 s ( @xmath49 ) such that @xmath54 is minimized .
clearly , the part @xmath55 can not be improved .
thus we need to minimize @xmath56 over all partitions of @xmath57 into @xmath48 s .
the partition that reaches the minimum @xmath58 does not change by linear scaling of the @xmath16 s .
hence we can argue as follows .
consider @xmath59 such that @xmath60 is the entropy of the set of probabilities @xmath61 .
denote @xmath62 and define @xmath63 then , both @xmath58 and @xmath60 with the @xmath64 ( @xmath49 ) reaches their minimum for this partition of @xmath65 .
[ lem.1 ] assume the above notation with with @xmath66 determined as above .
the minimal expected prefix code length under given length restrictions is achieved by encoding @xmath32 with code length @xmath67 for all @xmath49 .
let us compare the optimal expected code length under length constraints with the unconstrained case .
the difference in code length is @xmath68 the kulback - leibler divergence @xmath69 between the @xmath70-distribution and the @xmath52-distribution @xcite .
the kl - divergence is always nonnegative , and is 0 only if @xmath71 for all @xmath25 . for the optimum @xmath52-distribution determined in lemma [ lem.1 ] for the index set @xmath72 we can compute it explicitly : @xmath73 [ lem.2 ] given a random source @xmath0 with probabilities @xmath2 ( @xmath25 ) , length restrictions @xmath20 and with @xmath66 determined as above .
then , the minimum expected constrained code length is @xmath74 which equals the minimal expected unconstrained code word length @xmath75 only when @xmath76 for all @xmath25 .
thus , the redundancy induced by the equality length restrictions is @xmath77 note that , just like in the unconstrained case we can find a prefix code with code word lengths @xmath78 , showing that the minimal expected integer prefix - code word length is in between the entropy @xmath79 and @xmath80 , the same holds for the constrained case .
there , we constructed a new set of probabilities with entropy @xmath46 , and for this set of probabilities the minimal expected integer prefix - code word length is in between the entropy @xmath46 and @xmath81 by the usual argument .
let us look at an example with probabilities @xmath82 and length restrictions @xmath83 .
the entropy @xmath84 bits , and the , non - unique , huffman code , without the length restrictions , is @xmath85 , which shows the minimal integer code - word length of @xmath86 bits , which is @xmath87 bits above the noninteger lower bound @xmath79 .
the redundancy excess induced by the equality length restrictions is @xmath88 bits , which shows that the integer minimal average code - word length @xmath44 is in between @xmath89 bits and @xmath90 bits .
the actual optimal equality restricted code , given by algorithm a below , is @xmath91 with @xmath92 , which is @xmath93 bits above the noninteger lower bound @xmath46 .
above we have ignored the fact that real shannon - fano codes have code - word length @xmath94 rather than @xmath95 .
this is the reason that @xmath38 in the unconstrained case , leaving a slack of 1 bit for the minimal expected code word length .
the huffman code is an on - line method to obtain a code achieving @xmath96 .
gallager @xcite has proved an upper bound on the redundancy of a huffman code , @xmath97 of @xmath98 $ ] which is approximately @xmath99 , where @xmath100 is the probability of the least likely source message .
this is slightly improved in @xcite .
our task below is to find a huffman - like method to achieve the minimal expected code - word length @xmath44 in the length - constrained setting .
our goal is to come as close to the optimum in lemmas [ lem.1 ] , [ lem.2 ] as is possible .
input is the set of source words @xmath101 with probabilities @xmath102 and length restrictions @xmath103 that should be satisfied by the target prefix code @xmath23 in the sense that @xmath104 for all @xmath105 except for the @xmath106 s with @xmath36 for which @xmath106 s there are no code word length restrictions .
let @xmath107 , @xmath108 , and @xmath109 .
denote @xmath110 for convenience in notation , assume that the source words @xmath101 are indexed such that @xmath111 with @xmath112 .
if @xmath113 for some @xmath106 we set @xmath114 ; @xmath115 ; @xmath116 ; @xmath117 .
we repeat this process until there are no equal lengths left , and finish with @xmath118 with @xmath119 s the resulting lengths .
that is , we just iteratively merge two nodes which are at the same level in the tree .
therefore , the problem is reduced to considering a code tree with forbidden code - word lengths @xmath119 ( @xmath120 , the largest forbidden length @xmath121 leading to a forbidden node with a free sibling node at the end of a path at the same length from the root .
that is , a code word tree with a single available free node at each level @xmath122 , satisfying @xmath123 and @xmath124 ( @xmath125 ) .
each @xmath122 corresponds to a path of length @xmath122 leading from the root to a node @xmath126 corresponding to a code - word prefix that is as yet unused .
we call such a node ( and the path leading to it or the corresponding code - word prefix ) a _
free stub_. denote the set of levels of these free stubs @xmath126 by @xmath127 , and let @xmath128 . without loss of generality ,
@xmath129 with @xmath130 .
we now have to the find a code - word tree using only the free stubs , such that the expected code - word length is minimized .
we can do this in the straightforward manner , dividing the probabilities in @xmath131 among the @xmath132 free stubs , and computing the minimal expected code - word length tree for the probabilities for every stub for each of those divisions , and determining the division giving the least expected code - word length .
we can use huffman s construction since it does nt depend on the probabilities summing to 1 .
there are @xmath133 possible divisions , so this process involves computing @xmath133 huffman trees exponentially many unless @xmath134 which is the unrestricted common huffman case .
let @xmath135 denote the number of edges in a path from the leaf node labeled by @xmath70 to the root .
so @xmath136 .
then , if a tree is optimal ( has least expected code - word length ) , then @xmath137 since otherwise the expected code - word length can be decreased by interchanging @xmath70 and @xmath52 .
if @xmath138 is a prefix - code word tree for the source words under the given length restrictions , then the subtree @xmath139 is the subtree of @xmath138 with the free stub @xmath140 as its root ( @xmath141 ) .
suppose the contrary : for every optimal tree @xmath138 , there are @xmath145 with @xmath70 in @xmath139 and @xmath52 in @xmath144 for some @xmath142 .
fix any such @xmath138 . by ,
let @xmath147 be the subtree with root at @xmath148 containing the leaf @xmath70 .
then we can interchange @xmath52 and @xmath147 without changing the expected code - word length represented by the tree .
this idea leads to the following sorting procedure : repeat until impossible : find a least level probability , and if there are more than one of them a largest one , that violates the condition in the lemma , and interchange with a subtree as above .
since no probability changes level the expected code - word length stays invariant . in each operation a least level violating probability moves to a lower indexed subtree , and the subtree it is interchanged with does not introduce new violating probabilities at that level .
the transformation is iteratively made from the top to the bottom of the tree .
this process must terminate with an overall tree satisfying the lemma , since there are only a given number of probabilities and indexed subtrees .
for ease of notation we now assume that the source words @xmath149 are indexed such that the unrestricted source words are indexed @xmath150 with probabilities @xmath151 .
the fact that we just have to look for a partition of the ordered list of probabilities into @xmath132 segments , rather than considering every choice of @xmath132 subsets of the set of probabilities , considerably reduces the running time to find an optimal prefix code .
consider the ordered list @xmath152 .
partition it into @xmath132 contiguous segments , possibly empty , which gives @xmath153 partitions .
we can reduce this number by noting that in an optimal tree the free stubs are at different heights , and therefore each of them must have a tree of at least two elements until they have empty trees from some level down .
otherwise the tree is not optimal since it can be improved by rearranging the probabilities .
therefore , we can restrict attention to partitions into @xmath154 segements , which contain at least two elements . there are at most @xmath155 such partitions . for each choice , for each set of probabilities corresponding to a @xmath106th segment construct the huffman tree and attach it to the @xmath106th free stub , and compute the expected code - word length for that choice .
a tree associated with the least expected codeword length is an optimal tree .
thus , we have to construct at most @xmath155 huffman trees , which is polynomial in @xmath156 for fixed @xmath132 , and also polynomial in @xmath156 for either @xmath132 or @xmath157 bounded by a constant . from each partition of @xmath158 into @xmath159 segments ( possibly empty ) consisting of probabilities @xmath160 for the @xmath161th segment , we can construct trees @xmath162 with @xmath163 having the probabilities of @xmath160 as the leaves , and free stub @xmath164 as the root .
clearly , if @xmath138 is an overall tree with minimum expected code - word length , then each subtree with the free stub @xmath164 as root considered in isolation , achieves minimal expected code - word length over the probabilities involved .
we want to find the optimal partition with a minimum amount of work .
note that , from some @xmath165 on , every subtree @xmath163 with @xmath166 may be empty . for every tree @xmath138 , not necessarily
optimal , let @xmath167 denote the expected code - word length for the probabilities in @xmath131 according to tree @xmath138 .
define @xmath168 $ ] to be the _ minimal _ expected code - word length of the leaves of a tree
@xmath169 $ ] constructed from probabilities @xmath170 ( @xmath171 ) and with a singlefold path from the root of @xmath138 to the free stub node @xmath164 , and subsequently branching out to encode the source words ( probabilities ) concerned .
then , @xmath172 = \sum_{i \leq r \leq j } p_r ( h_k+l(p_r))\ ] ] each probability @xmath173 labeling a leaf at the end of a path of length @xmath174 from the root of @xmath138 , the first part of length @xmath175 to the free stub @xmath164 , and the second part of length @xmath176 from @xmath164 to the leaf concerned . for a partition of the probability index sequence @xmath177 into @xmath132 ( possibly empty ) contiguous segments
@xmath178 $ ] ( @xmath179 ) , inducing subtrees @xmath180 $ ] using free stubs @xmath175 accounting for expected code - word length @xmath181 $ ] , we obtain a total expected code - word length for the overall tree @xmath138 of @xmath182.\ ] ] let us now consider the expected code word length of a tree @xmath183 which consists of tree @xmath138 with a subset of subtrees @xmath163 removed and the corresponding probabilities from the overall probability set @xmath131 . removing subtree @xmath163 is equivalent to removing the corresponding free stub @xmath164 , and turning it into a length restriction .
let @xmath138 , @xmath131 and @xmath183 be defined as above .
let @xmath138 has minimal total code - word length for @xmath131 then the total code - word length of every @xmath183 as above can not be improved by another partition of the probabilities involved among its subtrees .
( _ if _ ) if we could improve the total code word length of @xmath138 by a redistribution of probabilities among the subtrees attached to the free stubs then some @xmath183 would not have minimal total code - word length before this redistribution .
( _ only if _ ) if we could improve the total code - word length of any @xmath183 by redistribution of the probabilities among its subtrees attached to the free stubs involved , then we could also do this in the overall tree @xmath138 and improve its overall total code - word length , contradicting minimality .
this suggests a way to construct an optimal @xmath138 by examining every @xmath161-partition corresponding to a candidate set of @xmath161 subtrees ( for @xmath184 ) , of every initial segment @xmath185 of the probability sequence ( for @xmath186 ) .
the minimal expected code - word length tree for the @xmath161th partition element is attached to the @xmath161th free stub .
the crucial observation is that by corollary [ cor.seqopt1 ] the minimal total code word length for probabilities @xmath187 using free stub levels @xmath188 is reached for a binary split in the ordered probabilities and free stubs involved , consisting of the minimal total code - word length solution for probabilities @xmath189 using stub levels @xmath190 and probabilities @xmath191 using free stub level @xmath192 . computing the optimal minimum code - word lengths of initial probability segments and initial free stub level segments in increasing order , this way we find each successive optimum by using previously computed optima .
this type of computation of a global optimum is called dynamic programming . the following algorithm
a gives the precise computation . at termination of the algorithm ,
the array @xmath193 $ ] will contain the minimal expected code word length of a tree @xmath194 $ ] using the largest @xmath195 ( @xmath196 ) probabilities @xmath197 , optimally divided into subtrees attached to the least level @xmath161 ( @xmath198 ) free stubs .
thus , @xmath194 $ ] contains subtrees @xmath199 $ ] ( @xmath200 , @xmath201 ) each such subtree with @xmath202 as the root ( @xmath200 ) .
thus , on termination @xmath203 $ ] contains the minimal expected code - word length of the desired optimal tree @xmath138 with subtrees @xmath204 with @xmath205 rooted at the root of @xmath138 and using free stubs @xmath206 , respectively .
( note , @xmath205 consists of the previously defined subtree @xmath163 , rooted at free stub @xmath164 , plus the path from @xmath164 to the root of @xmath138 . )
we can reconstruct these subtrees , and hence the desired code words , from the values of the array @xmath207 on termination .
denote @xmath208 , @xmath209 , @xmath210 ( @xmath211 ) , and @xmath212 . then , @xmath213 $ ] and @xmath214 $ ] equals the expected code - word length of the source words encoded in subtrees @xmath215 ( @xmath179 ) .
thus , the array values of @xmath207 give the desired partition of the ordered probability sequence @xmath152 , and we can trivially construct the tree @xmath138 and the code - words achieving minimal expected code - word length by huffman s construction on the subtrees . *
input : * given @xmath13 source words with ordered probabilities and equality length restrictions , first check whether is satisfied with @xmath216 , otherwise return `` impossible '' and quit .
compute free stub levels @xmath217 .
probabilities @xmath152 as above .
* step 1 : * = compute = @xmath168 $ ] = as in , = for all @xmath218 and @xmath161 ( @xmath219 , @xmath179 ) .
+ set @xmath220 : = h[1,j,1]$ ] ( @xmath221 ) .
+ @xmath222 * do * + @xmath223 * do * + @xmath193 : = \min_{i : 1 \leq
i \leq j } \ { f[i , k-1 ] $ ] @xmath224 \}$ ] ; + @xmath225 : = i_0 $ ] , with @xmath226 the least @xmath106 achieving the minimum + the complexity of computing the @xmath217 and @xmath158 is @xmath227 .
step 1 of the algorithm takes @xmath228 steps .
first , compute for every @xmath218 the quantities @xmath229 : = \sum_{i = r}^j l(p_r)$ ] , @xmath230 : = \sum_{i = r}^j p_r l(_r)$ ] .
there are @xmath231 such quantities and each computation takes @xmath232 steps .
second , for every @xmath161 compute @xmath168 = l[i , j]+h_k p[i , j]$ ] .
there are @xmath132 such quantities and each computation takes @xmath233 steps .
step 2 of the algorithm takes @xmath232 steps .
step 3 of the algorithm involves a outer loop of length @xmath234 , an inner loop of length @xmath156 , and inside the nesting the determining of the minimum of @xmath235 possibilities ; overall @xmath236 steps .
the running time of the algorithm is therefore @xmath237 steps . since @xmath238 this shows the stated running time . | we study the new problem of huffman - like codes subject to individual restrictions on the code - word lengths of a subset of the source words . these are prefix codes with minimal expected code - word length for a random source where additionally the code - word lengths of a subset of the source words is prescribed , possibly differently for every such source word .
based on a structural analysis of properties of optimal solutions , we construct an efficient dynamic programming algorithm for this problem , and for an integer programming problem that may be of independent interest . |
table i : transformation property of fermions and higgs multiplets under s@xmath37-symmetry .
the @xmath129-dimensional multiplets are chosen to be complex ..
1 : @xmath130 and @xmath131 are plotted as a function of @xmath82 for @xmath132 .
the scale of the vertical axes is 8.37 @xmath2 10@xmath133 ev for @xmath56 = 10@xmath134 gev and @xmath135 . | the simplest scenario for the three known light neutrinos that fits the solar and atmospheric neutrino deficit and a mixed dark matter ( mdm ) picture of the universe requires them to be highly degenerate with @xmath0 1 - 2 ev .
we propose an so(10 ) grand unified model with an s@xmath1-horizontal symmetry that leads naturally to such a scenario .
an explicit numerical analysis of the quark and lepton sector of the model shows that it can lead to desired mass differences to fit all data only for the small angle non - adiabatic msw solution to the solar neutrino puzzle .
= 3ex = .5 cm = 4ex = = .35 in = .35 in umd - pp-94 - 95 * an so(10 ) @xmath2 s@xmath1 scenario for naturally degenerate neutrinos * + [ 6 mm ] * dae - gyu lee and r. n. mohapatra * + _ department of physics , university of maryland + college park , maryland 20742 _ + [ 4 mm ] in the two recent papers , it has been pointed out by caldwell and one of the authors ( r.n.m.)@xcite that if the existing data on solar@xcite and atmospheric neutrino@xcite deficit is confirmed and if the presently popular mixed dark matter ( mdm)@xcite model of the universe requiring a few ev neutrino as its hot component is taken seriously , then the simplest three - light neutrino ( @xmath3 , @xmath4 , @xmath5 ) mass matrix that fits all data has the following form : @xmath6 where @xmath7=2 ev , s@xmath8 , s@xmath9 , @xmath10 ev and @xmath11 .2 to .002 ev .
here we have assumed that the solar neutrino puzzle is solved by the small angle msw ( mikheyev - smirnov - wolfenstein)@xcite solution@xcite .
note the high degree of degeneracy among the three neutrino species .
an immediate implication of this is that if the neutrinos are majorana particles , the neutrinoless double beta decay would be measurable in the current generation experiments involving @xmath12@xcite and @xmath13@xcite thereby providing a test of the degenerate neutrino hypothesis .
it is then perhaps not premature to search for gauge models which can generate this highly degenerate neutrino spectrum@xcite in a technically natural manner .
such degeneracy is suggestive of a horizontal symmetry , which will contain all three neutrinos in one irreducible representation .
this symmetry however must be broken in the charged lepton sector .
since the see - saw mechanism@xcite generally connects the neutrino masses with the charged fermion masses , the horizontal symmetry breaking in the charged fermion sector will cause mass differences between the neutrinos .
such mass differences are of course required and from equation ( 1 ) , we see that these must be very tiny .
a simple scaling argument then says that the see - saw scale must be in the range of @xmath14 gev or so .
such mass scales have their natural place in grand unified theories and we will consider so(10 ) as the flavor grand unifying theory in order to understand the required neutrino spectrum . turning now to the horizontal symmetry , an obvious choice is to consider it to be @xmath15 as has been done in several recent papers@xcite . in this letter , we consider a somewhat more economical group based on the permutation group @xmath16 as our horizontal symmetry .
we will assume that the symmetry is softly broken so that there is no domain wall problem .
we find that in this model , the fermion sector is completely specified by fifteen arbitrary parameters , twelve of which are fixed by the charged fermion sector ( i.e. by six quark masses , three ckm angles and three charged lepton masses ) .
the structure of the neutrino mass matrix is then such that only the small angle msw solution to the solar neutrino problem , before presenting the details of the model , let us first review the basic strategy given in ref .
@xcite .
it is well - known that when the conventional see - saw mechanism for neutrino - masses@xcite is implemented in gauge models such as so(10 ) or the left - right symmetric models , it gets modified to the following form@xcite @xmath17 where @xmath18 @xmath19 is the scale of su(2)@xmath20-breaking and @xmath21 is breaking scale of parity . therefore , unless special care is taken to break parity symmetry at a scale higher than the su(2)@xmath20 or u(1)@xmath22 , @xmath23 ( since @xmath24 ) .
the light neutrino masses are then given by : @xmath25 recall that the conventional see - saw formula omits the first term ( which , as just mentioned , is justified only under special circumstances ) leading to an approximate quadratic scaling relation between neutrino and up - quark mass ( or in some instances charged lepton masses ) .
we will however keep both the terms in the present discussion .
now notice that if due to some symmetry reasons , @xmath26 , then a degenerate neutrino spectrum emerges . before discussing the fermion masses ,
let us briefly review some of the discussions of ref .
@xcite .
we will consider the breaking of so(10 ) @xmath27
su(2)@xmath28 @xmath2 su(2)@xmath29 @xmath2 su(4)@xmath30 @xmath2 p
( denoted by @xmath31 ) by means of a @xmath32-dim .
higgs multiplet .
this symmetry is subsequently broken down to the standard model by a @xmath33-dim .
higgs multiplet .
detailed two - loop analysis of the mass scales in this model@xcite leads to @xmath34 gev .
so that for @xmath35 , we get @xmath36 ev , as desired .
let us now turn to the implication of s@xmath37-symmetry for the charged fermion and neutrino masses , which is the main contribution of this paper .
s@xmath37-symmetry has been used before in the discussion of charged fermion masses at the electroweak level@xcite .
it has irreducible representations with dimension @xmath38 , @xmath39 , @xmath40 , @xmath41 , and @xmath42 .
our assignment of fermions and higgs multiplets to irreducible representations of s@xmath37 are shown in table i. the s@xmath37-invariant yukawa coupling can be written symbolically as @xmath43 \nonumber \\ & & + { f_3 \over \sqrt{2 } } \left [ ( \psi_1 \psi_3 + \psi_3
\psi_1 ) h_{3 } + ( \psi_2 \psi_1 + \psi_1
\psi_2 ) h_4 + ( \psi_3 \psi_3 - \psi_2 \psi_2 ) h_5 \right ] \nonumber \\ & & + h.~c.\end{aligned}\ ] ] we then assume that the s@xmath37-symmetry is softly broken by the masses of @xmath44 ( @xmath45 ) and @xmath46 ( @xmath47 ) , so that their vacuum expectation values are arbitrary .
we also assume a softly broken u(1)@xmath48 symmetry so that the complex @xmath49 s have only one coupling to the fermions .
the @xmath44 s therefore have vev s given by @xmath50 and @xmath51 .
turning now to the @xmath52 s , we choose only the ( mass)@xmath53 of @xmath54 negative and large so that the ( @xmath55 ) submultiplet of it ( the numbers denote representation under the group @xmath31 ) acquire a vev @xmath56 that breaks @xmath31 down to the standard model .
the remaining two @xmath52 s ( @xmath57 ) have large ( @xmath58 ) positive masses so that their ( * 2,2,15 * ) components acquire induced vev s of order of the electroweak scale without any fine tuning@xcite due to the presence of s@xmath59 u(1)@xmath48 invariant terms such as @xmath60 in their potential ( where i=1,2 ) .
a term of the form @xmath61 also induces non - zero vev s to the ( * 2,2,15 * ) submultiplets of @xmath62 .
we denote these @xmath52-vev s by @xmath63 , @xmath64 ( @xmath47 ) .
note that the @xmath54 coupling in eq .
( 5 ) leads after symmetry breaking to the degenerate neutrino masses .
the charged fermion and dirac - neutrino mass matrices can now be written as follows .
@xmath65 where @xmath66 where @xmath67 s and @xmath68 s are products of type @xmath69 and @xmath70 respectively .
similar matrices can be written down for @xmath71 and @xmath72 by replacing @xmath67 s by @xmath73 s , @xmath74 by @xmath75 and @xmath68 s by @xmath76 s .
this way of writing makes it clear that there are a total of 18 parameters .
however , we can choose a basis in which the up - quark mass matrix is diagonal , thereby getting rid of three parameters and we are left with fifteen parameters .
we will now fit these parameters so that the six quark masses , three charged lepton masses and three ckm angles are reproduced . in order to proceed ,
we first extrapolate these parameters from the electro - weak scale to the scale @xmath77 , where their values are : @xmath78 we have chosen all masses to be positive .
we do the fitting as follows . in the basis where @xmath79 is diagonal
, we have @xmath80 .
the first choice leads to six constraints on the original ( before choosing the up - quark basis ) nine parameters leaving three arbitrary .
we choose them to be @xmath81 , @xmath82 , and @xmath83 .
similarly the @xmath84 equation leaves three arbitrary parameters chosen to be @xmath85 , @xmath86 , and @xmath87 . the @xmath85 , @xmath86 , and @xmath87
are then determined by three equations that relate them to @xmath88 , @xmath89 and @xmath90 which of course are easily expressed in terms of the @xmath91 , @xmath92 , and @xmath93 .
the matrix @xmath94 is then completely determined .
diagonalizing it as usual i.e. @xmath95 , where @xmath96 diag(@xmath91 , @xmath92 , @xmath93 ) , the three angles that parameterize @xmath97 are predicted .
in particular , note that , in the basis in which @xmath79 is diagonal , @xmath98 mixing arises predominantly from the ( @xmath99 ) sector due to the fact that @xmath100 is block diagonal and fitting @xmath91 , @xmath92 , and @xmath93 leads a prediction for @xmath101 .
we find that for negative @xmath102 , two sets of solutions for the mixing matrix @xmath103 correspond to the correct charged lepton masses .
only one of them leads to the desired @xmath104-@xmath4 mixing angle and we choose this one ; the parameters @xmath81 , @xmath82 , and @xmath83 not fixed by the up - quark sector , are then varied to obtain the required mass difference squares between @xmath105 , @xmath106 , and @xmath107 to fit the data .
let us now briefly discuss the dependence of our solutions on the three parameters @xmath81,@xmath82 and @xmath83 .
requiring @xmath108 implies that , @xmath109 ; if we further assume that the non - degenarate part of the contribution to the neutrino masses have a hierarchical structure between the first and the second generation leading to @xmath110 , this leads to @xmath111 and @xmath112 . in figure 1
, we have shown the dependence of the @xmath104- @xmath4 mass difference on @xmath82 for the above choices of @xmath83 and @xmath81 and we see that the desired range comes only for @xmath112 .
we give below the detailed solution for the neutrino masses and mixings for this case .
( we have written @xmath113 , where @xmath114 is the direct @xmath115 contribution . )
@xmath116 and @xmath117 note that , for @xmath118 gev and @xmath119 , this predicts @xmath120 ev@xmath53 for @xmath121 2 ev , @xmath122 ev@xmath53 , which are in the range required to solve both the solar and atmospheric neutrino deficit for the values of @xmath123 and @xmath124 given above . in particular , we wish to note the preference of theory for the small angle msw solution to the solar neutrino problem .
this is quite interesting in view of the recent conclusions@xcite that the large angle msw solutions have a discouragingly large @xmath125-fit .
we wish to remark that our solutions do not depend on the choice of sign for the up quark masses .
we have checked that we lose the desired values of masses and mixings if we choose @xmath126 to be negative for both signs of the cabibbo angle . in conclusion , we have constructed an so(10 ) @xmath2 s@xmath37 model , which leads to a degenerate neutrino mass scenario including correct mixing angles required to fit atmospheric and solar neutrino data as well as the hot dark matter in the universe .
a critical test of the model is the observation of a non - zero signal in neutrinoless double data decay in the current generation of @xmath127ge and @xmath128te experiments .
also our model prefers only the small angle msw solution to the solar neutrino puzzle , a result which is already indicated in recent analyses and will surely be tested once more data accumulates . 10 d.o .
caldwell and r.n .
mohapatra , ( 1993 ) ; and ( 1994 ) .
r. davis et al.,_proceedings of the 21st international cosmic ray conference _ ,
vol .
12 , edited by r.j .
protheroe ( univ . of adelaide press , adelaide , 1990 ) , p. 143 ; k.s .
hirata et al . , _ phys . rev .
* d 44 * , 2241 _ ( 1991 ) ; a.i .
abazov et al . , _ phys .
rev .
lett . *
67 * , 3332 _
( 1991 ) and t. bowles , invited talk at icnapp , bangalore ( 1994 ) ; p. anselmann et al .
, _ phys
. lett . *
b285 * , 376 _ ( 1992 ) and _ * 314b * , 445 _ ( 1993 ) .
t. kirsten , invited talk at icnapp , bangalore ( 1994 ) .
for a theorectical discussion , see j.n .
bahcall and m.h .
pinsonneault , _ rev .
mod .
phys . * 64 * , 885 _ ( 1992 ) . k.s .
hirata et al . , _ phys . lett . *
b280 * , 146 _ ( 1992 ) ; r. becker - szendy et al .
, _ phys
. rev . *
d 46 * , 3720 _ ( 1992 ) ; p.j .
litchfield in international europhysics conference on high energy physics , marseille , france , 1993 ( unpublished ) .
ch .
berger et al . , _ phys . lett . *
b245 * , 305 _ ( 1990 ) ; _ * b227 * , 489 _ ( 1989 ) .
m. aglietta et al . , _ europhys . lett . * 15 * , 559 _ ( 1991 ) .
w. frati et al . , _ phys .
rev .
* d 48 * , 1140 _ ( 1993 ) .
e.l .
wright et al . , _ astrophys .
j. * 396 * , l13 _ ( 1992 ) ; m. davis , f.j .
summers , and d. schagel , _ nature * 359 * , 393 _ ( 1992 ) ; a.n .
taylor and m. rowan - robinson , _ ibid .
, * 359 * , 396 _ ( 1992 ) ; r.k .
schaefer and q. shafi , _ report no .
ba-92 - 28 _ ( 1992 , unpublished ) and _ nature 359 , 199 _ ( 1992 ) ; j.a . holtzman and j.r .
primack , _ astrophys .
j. * 405 * , 428 _ ( 1993 ) ; a. klypin et al .
, _ astrophys .
j. * 416 * , 1 _ ( 1993 ) .
q. shafi and f .. stecker , _ phys .
rev .
lett . * 53 * 1292 _ ( 1984 ) .
s.p .
mikheyev and a.yu .
smirnov , _ yad .
fiz .
, * 42 * , 1441 _ ( 1985 ) ; l. wolfenstein , _ phys . rev .
* d 17 * , 2369 _ ( 1978 ) ; _ * d * 2634 _ ( 1979 ) . n. hata and p. langacker ,
_ univ . of pensilvenia report upr-0592 t _
( 1993 , to be published in _ phys . rev . *
d * _ ) ; p.i .
krastev and s.t .
petcov , _ report no.sissa 177/93/ep _ ( 1993 , unpublished and g. fogli , e. lisi and d. montanino , cern - th.6944/93 .
h. klapdor - kleingrothaus , invited talk at the erice summer school , 1993 to appear in _ prog . in particle and nuclear physics _ , edited by a. faessler ( 1994 ) .
e. garcia , talk at taup93 .
further implications of and models for the degenerate scenario have been considered in : s. petcov and a. y. smirnov , trieste preprint ( 1993 ) .
p. barmert and c. burgess , mcgill-94 - 07 ( 1994 ) ; a. joshipura , prl preprint ( 1994 ) ; k.s .
babu and s. pakvasa , _ phys .
lett . *
b172 * , 360 _ ( 1986 ) .
a. ioanyssian and j. w. f. valle , valencia preprint ( 1994 ) .
m. gell - mann , p. ramond , and r. slansky .
, edited by d. freedman et .
al .
( north - holland , amsterdam , 1980 ) ; t. yanagida .
( unpublished ) ; r. n. mohapatra and g. senjanovic .
( 1980 ) .
r. n. mohapatra and g. senjanovic .
( 1981 ) .
n. g. deshpande , e. keith , and p. b. pal .
( 1992 ) ; d. g. lee , university of maryland preprint , ( 1993 ) .
s@xmath37-symmetries in the context of charged fermion masses have been considered by s. pakvasa and h. sugawara .
( 1979 )
. k. s. babu and r. n. mohapatra .
( 1993 ) .
p. anselman et .
al .
gallex preprint 44 - 1994 . |
frequency - stable laser - local - oscillators ( llo ) are critical components of the new generation of optical atomic clocks @xcite
. the stability of such clocks is presently limited by thermomechanical noise in the cavity - stabilized llos @xcite .
quieter llos would enable optical clocks with improved stability , yielding faster comparison measurements among different clocks and more precise searches for the variation of fundamental constants @xcite .
quieter llos may also yield lower - phase - noise microwave oscillators by frequency - division via femtosecond laser frequency combs @xcite .
spectral - hole - burning ( shb ) laser frequency stabilization @xcite in eu@xmath0-doped y@xmath1sio@xmath2 has recently approached the performance of state - of - the - art cavity - stabilized llos @xcite and indirect measurements suggest that the fundamental noise floor may be significantly lower . because of the long lifetime of spectral holes ( of order @xmath3 s at 3.7 k @xcite ) , the large number of dopant ions within a crystal ( of order @xmath4/@xmath5 ) , and the narrow linewidth of each absorber , such crystals can serve as precise memories for laser frequency .
comparison of the current laser frequency with the frequencies imprinted in the memory enables frequency stabilization . inside the crystal ,
the narrow @xmath6f@xmath7@xmath8@xmath9d@xmath7 optical transition of each eu@xmath0 ion is shifted by the local crystal field , which is slightly random , such that each eu@xmath0 dopant has a fixed transition frequency within the typical 1 - 30 ghz inhomogeneous linewidth .
the isolated nature of rare - earth ion inner - shell transitions has allowed dephasing rates below 500 hz to be observed in this type of crystal @xcite . in eu@xmath0 ,
the @xmath6f@xmath7 ground - state comprises several long - lived hyperfine levels ( nuclear spin i=5/2 ) .
the hyperfine level splittings are of order 100 mhz , and for any given laser frequency within the inhomogeneous linewidth , particular combinations of crystal - fields and hyperfine levels are resonant .
the resonant population of eu@xmath0 ions is excited by the laser , and a significant fraction decays into different hyperfine levels , causing a transparency , or spectral - hole , at the laser frequency .
high - resolution spectroscopy has shown that the environmental sensitivity of eu@xmath0:y@xmath1sio@xmath2 spectral holes to temperature , pressure , and accelerations are all smaller than that of fabry - prot cavities @xcite , and frequency drift rates can be lower than those of cavity - stabilized lasers @xcite .
these results suggest that spectral holes in eu@xmath0:y@xmath1sio@xmath2 could form a frequency - stable flywheel for optical atomic clocks and potentially extend the coherence time of atom - laser interactions . unlike fabry - prot cavities , whose oscillation period is directly proportional to an artificial length , the optical frequency of spectral holes depends primarily on the atomic hamiltonian of dopant ions .
spectral holes are less sensitive to mechanical perturbations , which couple only weakly through the field of the host crystal , but may be more sensitive to ambient electric and magnetic fields . moreover , probing of the spectral holes causes additional burning that can modify the hole shape and frequency .
this work aims to determine the extent to which such frequency shifts will limit the performance of shb llos , and we describe measurements of eu@xmath0:y@xmath1sio@xmath2 spectral - hole properties including the magnetic field sensitivity , perturbations to the spectral - hole pattern due to side - holes and anti - holes , and frequency shifts due to fluctuations in the optical probe power .
temperature - dependent shifts are also measured , and it is shown that the linear dependence of the hole frequency on temperature can be eliminated by placing the crystal inside a sealed chamber that is filled with a particular pressure of helium gas . with the current level of environmental stability , fluctuations in these frequency shifts are small enough to allow for laser stabilization with a fractional - frequency noise floor below @xmath10 .
this would represent an order of magnitude improvement over state - of - the - art laser frequency stabilization using fabry - prot cavities @xcite .
this paper proceeds as follows : section [ sec : setup ] describes the experimental setup
. measurements of eu@xmath0:y@xmath1sio@xmath2 spectral - hole properties are presented in sections [ sec : magnetic ] ( magnetic field sensitivity ) and [ sec : tempandpower ] ( temperature and optical probe power sensitivity ) .
perturbations to spectral - hole patterns due to side - holes and anti - holes are discussed in section [ sec : sideholes ] .
a summary of the frequency shifts of optical frequency references based on spectral - hole burning in eu@xmath0:y@xmath1sio@xmath2 is provided in section [ sec : shifts ] .
a schematic of the experimental setup is shown in figure [ fig1 ] .
spectral holes in cryogenically cooled eu@xmath0:y@xmath1sio@xmath2 are burned and probed by a dye laser that is pre - stabilized to a spherical fabry - prot reference cavity @xcite .
the laser frequency can be tuned coarsely by stepping among longitudinal modes of the reference cavity ( 3 ghz free spectral range ) and finely by shifting the frequency of an acousto - optic modulator ( aom3 ) located between the laser and the cavity .
the pre - stabilized laser has a fractional frequency stability of @xmath11 between 0.5 s and 12 s and a typical drift rate of @xmath12/s .
eu@xmath0:y@xmath1sio@xmath2 exhibits two inhomogeneously broadened absorption features at 580.039 nm ( site 1 ) and 580.211 nm ( site 2 ) that result from different crystal field shifts at the two distinct positions within the y@xmath1sio@xmath2 unit cell where eu@xmath0 can substitute for y@xmath0 ; in this work we focus primarily on site 1 spectral holes because they display a smaller temperature shift .
all of the measurements presented in section [ sec : tempandpower ] were performed on a matching pair of cylindrical samples of 1.0 atomic ( at . )
% eu@xmath0:y@xmath1sio@xmath2 with a diameter of 10.0 mm and a length of 4.9 mm . for each sample
the laser propagates parallel to the axis of the cylinder ( b - axis of the crystal @xcite ) , whose ends are wedged at a 2@xmath13 angle to prevent etalon effects , and is polarized along the d1 crystal axis .
the samples are cooled to 3.5 to 8.8 k in a closed - cycle cryostat that allows optical access . inside the cryostat ,
the two crystals are enclosed in a single hermetically sealed chamber that is filled with helium gas to provide independent control of temperature and pressure .
the pressure is controlled by a thin tube that connects the sealed chamber to a room - temperature gas manifold and each crystal is mounted in a way that reduces the sensitivity to vibrations @xcite .
the measurements presented in secs . [ sec : magnetic ] and [ sec : sideholes ] were made with samples of 0.1 and 0.5 at .
% eu@xmath0:y@xmath1sio@xmath2 in an optical flow cryostat described previously @xcite .
all of the crystals used in this work were grown with a natural abundance of europium isotopes ( 47.8% @xmath14eu and 52.2% @xmath15eu ) .
experimental setup for measuring the properties of eu@xmath0:y@xmath1sio@xmath2 spectral holes .
a 580 nm dye laser is pre - stabilized to a spherical fabry - prot reference cavity .
the laser burns and probes spectral holes in two eu@xmath0:y@xmath1sio@xmath2 crystals located in a single closed cycle cryostat .
spectroscopy is controlled by a field - programmable gate array ( fpga ) .
aom : acousto - optic modulator , eom : electro - optic modulator , det : detector . ] spectroscopy of eu@xmath0:y@xmath1sio@xmath2 spectral holes is controlled by a field - programmable gate array ( fpga ) that includes an embedded microprocessor .
the laser beams ( 5 mm diameter ) incident on each crystal can be tuned independently over a 600 mhz frequency range and 0.1 to 20 @xmath16w/@xmath17 intensity by adjusting the radio - frequency ( rf ) drives of aom 1 and 2 .
phase modulation sidebands are imprinted on the lasers for pound - drever - hall ( pdh ) laser frequency stabilization @xcite by electro - optic modulators .
detectors i1 and i2 measure the power incident on the crystals , which is stabilized at the desired level by feedback to the amplitudes of the rf drives of aom 1 and aom 2 .
detectors t1 and t2 measure the transmission of the laser through crystal 1 and crystal 2 , respectively .
the temperature - sensitivity measurement is based on locking the laser frequency to a pattern of several spectral holes by the pdh method , and observing frequency changes with respect to the fabry - prot reference cavity .
measurements of magnetic field effects and side - features rely on fitting lorentzian line - shapes to the measured transmission spectrum of a single spectral hole .
the spectral - hole - burning mechanism in eu@xmath0:y@xmath1sio@xmath2 is based on population redistribution among long - lived hyperfine levels , and changes in the magnetic field environment may modify the frequencies and lineshapes of spectral holes due to the zeeman effect . however , the @xmath6f@xmath7@xmath8@xmath9d@xmath7 transition of eu@xmath0 has small magnetic moments in both the ground and excited states @xcite , because only the nuclear angular - momentum is non - zero . the linear zeeman effect in 0.1 at .
% eu@xmath0:y@xmath1sio@xmath2 site 1 spectral holes at 4.5 k. ( a ) a magnetic field applied to spectral holes after they have been written splits the spectral hole into a symmetric set of spectral components . in this example
a spectral hole is written in an applied magnetic field of 0 g and measured in an applied magnetic field of 37 g along each of the crystal axes .
the three measurements are offset vertically for clarity .
( b ) frequency shift of the most prominent spectral components ( indicated by vertical arrows in fig .
[ fig2]a ) as a function of applied magnetic field along the crystal b axis . ]
figure [ fig2]a shows the transmission spectrum of a spectral hole written using a 200 ms burn pulse with an intensity of 6.6 @xmath16w/@xmath17 at zero applied field and subsequently measured in a magnetic field of 37 g ( 1 g = 0.1 mt ) that is sequentially applied along each of the three crystal axes . here
the spectral hole , which was a single 1.1 khz fwhm peak at zero applied field , splits into a symmetric set of transmission peaks corresponding to transitions between pairs of kramer s doublets .
the zeeman shifts are different for magnetic fields along the different crystal axes .
[ fig2]b shows the frequency shifts of the most prominent spectral components as a function of applied magnetic field along the crystal b axis . because of the symmetry of the kramer s doublets , the linear zeeman shift averages to zero for each spectral hole , but it can contribute to broadening of the holes .
the linear zeeman shift of the most sensitive states is 3.8 khz / g ( magnetic field along the crystal b axis ) . to avoid broadening spectral holes at the 1 khz fwhm level
, the magnetic field fluctuations must be less than 0.1 g. the quadratic zeeman shift in 0.1 at .
% eu@xmath0:y@xmath1sio@xmath2 site 1 spectral holes at 4.5 k for magnetic fields applied along the b - axis , d2-axis , and d1-axis .
] unlike the linear zeeman effect , the quadratic zeeman effect causes a frequency shift of spectral holes .
figure [ fig3 ] shows the area - weighted average ( over the most prominent transmission peaks , indicated in figure [ fig2]a by vertical arrows ) frequency shift of a spectral hole written at zero applied field and measured in a nonzero applied magnetic field .
we measure a quadratic frequency sensitivity of 1.4(1.4 ) hz / g@xmath18 for magnetic fields along the crystal b axis , which is again the most sensitive direction .
note that we use the measured value as an estimate of the uncertainty because the atomic population of the original spectral hole is nearly twice the total population of the discernible zeeman components .
the remaining population is contained in spectral features whose center frequency is not readily discernible .
the measured sensitivity can be compared to the expected value of 0.64 hz / g@xmath18 in free eu@xmath0 atoms @xcite . in order to reach a fractional frequency instability of @xmath19 the magnetic field
must be held at 0 g with an instability better than 60 mg , or in a laboratory field of 0.5 g the magnetic field instability must be below 7 mg .
at low temperature , the frequency shift ( @xmath20 ) of spectral holes in eu@xmath0:y@xmath1sio@xmath2 exhibits a fourth - order temperature ( @xmath21 ) dependence @xcite .
we measured @xmath22 with @xmath23 hz / k@xmath24 for site 1 and @xmath25
hz / k@xmath24 for site 2 over a temperature range of 2.55.5 k in a previously described apparatus @xcite . at 3.65 k ,
the slope is @xmath26 khz / k for site 1 spectral holes and such a high sensitivity could make eu@xmath0:y@xmath1sio@xmath2 crystals unsuitable for laser - frequency stabilization .
here we immerse the shb crystal in a sealed helium - gas environment to mitigate @xcite the temperature shift @xcite .
temperature increases cause the pressure of the helium gas to increase , which compresses the crystal .
because the spectral - hole frequency shift due to compression is negative and that due to a rise in temperature is positive , the slopes of the two shifts can be made to cancel at a particular pressure of helium gas .
note , however , that this cancellation works only when the crystal is in thermal equilibrium with the surrounding gas . in our apparatus a tube connects the sealed sample chamber to a room - temperature gas manifold through which helium can be added and removed . to reduce time - dependent frequency shifts associated with the thermalization time of the two - reservoir system ,
we have minimized the volume of the room - temperature portion such that it is approximately equal to the volume of the cryogenic portion . in the limit @xmath27 where @xmath28 ( @xmath29 ) and @xmath30 ( @xmath21 ) are the volume and temperature of the room - temperature ( cryogenic ) portion of the helium gas , the pressure at which first order temperature shifts cancel is @xmath31 , where @xmath32 is the operating temperature and the spectral - hole pressure shift has been measured @xcite to be @xmath33
hz / pa for site 1 and @xmath34 hz / pa for site 2 . the residual quadratic temperature sensitivity is @xmath35 calibration of the temperature insensitive point of 1.0 at .
% eu@xmath0:y@xmath1sio@xmath2 site 1 spectral holes in 270 pa of helium gas .
( a ) time - trace of the relative frequency between spectral - hole patterns in crystals 1 and 2 and the pre - stabilization reference cavity .
the temperature is stepped in the vicinity of the insensitive temperature causing shifts in the frequency of the hole pattern .
( b ) fits of the frequency shifts reveal both the insensitive temperature and the quadratic temperature sensitivity .
( c ) temperature stability ( allan deviation ) of the sample chamber .
the total measurement duration is @xmath36 s. ] figures [ fig4]a and [ fig4]b show the temperature - dependent frequency shift in 270 pa of helium gas . the frequency difference between the pre - stabilization fabry - prot cavity and a laser that is locked to a pattern of spectral holes is recorded as a function of time while the temperature is changed in discrete steps near the insensitive point .
figure [ fig4]b shows the frequency shift as a function of the crystal temperature .
the residual quadratic temperature shift at 3.65 k for site 1 spectral holes is 6.5 khz / k@xmath18 , which implies that in order to reach @xmath19 fractional frequency instability , the temperature instability needs to be below 1 mk .
if the temperature is 5 mk away from the insensitive temperature , a stability better than 80 @xmath16k is required .
figure [ fig4]c shows the temperature stability of the sample chamber inside the closed - cycle cryostat whose cold head is thermally stabilized by a feedback loop . the stability shown here
is measured by an out - of - loop sensor located on the sample chamber .
the chamber is thermally decoupled from the cold head by brass washers , which improves the temperature stability of the chamber at short times but degrades it at long times .
temperature stability is better than 1 mk for the entire range of measurement times and better than 80 @xmath16k for durations between 250 s and 50000 s. local heating of the crystal by the laser may cause an additional thermal frequency shift .
however , this effect is reduced because the helium gas inside the chamber thermalizes the crystal with the chamber walls .
we have measured an upper bound for the spectral - hole frequency shift due to a step change in laser power .
we locked two beamlines of the laser separately to a pattern of site 1 spectral holes in each crystal at 2.0 @xmath16w/@xmath17 laser intensity for 400 s , then increased the laser intensity incident on crystal 2 to 6.6 @xmath16w/@xmath17 for 50 s ( leaving the laser intensity incident on crystal 1 unchanged ) .
the relative frequency shift was less than 1 hz , corresponding to a laser intensity sensitivity of less than 0.2 hz/(@xmath16w/@xmath17 ) .
current intensity stability is 1 % , so laser intensity fluctuations are unlikely to limit the performance of this system at the @xmath19 fractional frequency stability level .
side - hole and anti - hole spectrum due to the eu@xmath0:y@xmath1sio@xmath2 site 1 spectral - hole hyperfine structure as a function of spectral - hole width and eu@xmath0 doping . in each transmission measurement , a single spectral hole is burned to a peak amplitude of roughly 600 in the arbitrary transmission units of the figure .
the legend labels the transmission measurements by specifying the eu@xmath0doping level , as well as the crystal temperature and the fwhm of the spectral hole .
the amplitude of the side - holes and anti - holes is reduced by writing narrower spectral holes because the side - holes and anti - holes exhibit inhomogeneous broadening . ]
side - holes and anti - holes @xcite perturb broadband patterns of spectral - holes , which may be the best way to implement a shb llo that can operate continuously .
figure [ fig5 ] shows the side - hole and anti - hole spectrum of a single spectral hole for several spectral - hole widths and two eu@xmath0doping levels .
inhomogeneous broadening in the eu@xmath0:y@xmath1sio@xmath2 hyperfine structure leads to side - holes and anti - holes that are broader than the main spectral hole feature . as a result
, the amplitude of the side - holes and anti - holes can be reduced by writing narrow spectral holes . as the eu@xmath0doping level increases , the inhomogeneous broadening in the eu@xmath0:y@xmath1sio@xmath2 hyperfine structure increases .
figure [ fig6 ] shows the widths and amplitudes of the side - holes and anti - holes of a single spectral hole for two eu@xmath0doping levels . increasing the doping from 0.1 % to 0.5 % results in a mean increase in the widths of the side - holes and anti - holes by a factor of 3.8 .
perturbations from side - holes and anti - holes to broadband patterns can therefore be reduced by choosing the highest dopant concentration that is compatible with the requirements for spectral - hole lifetime and coherence .
the effect of side - holes and anti - holes on the performance of a shb llo can be estimated by considering how much the center frequency of a spectral hole will be shifted by an overlapping side - hole or anti - hole . if two 2 khz fwhm spectral holes of equal depth are burned in 0.5 at .
% eu@xmath0:y@xmath1sio@xmath2 with worst - case frequency spacing , then side - features of the second hole will shift the center frequency of the first hole by a fractional - frequency of less than @xmath19 with respect to the optical transition . fwhm and amplitude of the side - holes ( positive ) and anti - holes ( negative ) due to the eu@xmath0:y@xmath1sio@xmath2 site 1 spectral - hole hyperfine structure as a function of eu@xmath0 doping . in each transmission measurement , a single 11.6 khz ( 0.1% doping at 7.3 k ) or 83 khz ( 0.5% doping at 8.8 k ) fwhm spectral hole was burned to full transparency ; the transmission amplitude is normalized to the transmission of the main hole . ]
@lrrl parameter & & & units + wavelength & 580.039 & 580.211 & nm + bulk modulus @xcite & 134.8 & 134.8 & gpa + volume @xcite & 0.055 & 0.014 & @xmath37 + magnetic field & 1.4(1.4 ) & - & hz / g@xmath18 + electric field @xcite & @xmath38 & - & hz/(v / m)@xmath18 + temperature@xmath39 & 6.2 & 21 & khz / k@xmath18 + pressure @xcite & -211.4(4 ) & -52.0(7 ) & hz / pa + laser intensity & @xmath38 & - & hz/(@xmath16w/@xmath17 ) + acceleration @xcite & @xmath40 & - & @xmath41 + @xmath39 at 3.7 k and 270 pa ( 3600 pa ) helium gas pressure for site 1 ( 2 ) .
table [ tab : systematics ] summarizes properties of eu@xmath0:y@xmath1sio@xmath2 relevant to shb laser frequency stabilization .
we expect the frequency noise of site 1 spectral holes due to fluctuations in magnetic field , temperature , pressure , and laser intensity to be below @xmath19 in our experimental setup .
acceleration induced frequency shifts due to vibrations of the closed cycle cryostat may contribute to the spectral hole frequency instability at short averaging time , but will not limit performance at long averaging time .
while the linear stark coefficient of eu@xmath0:y@xmath1sio@xmath2 site 1 spectral holes has been measured @xcite to be 350(50 ) hz/(v / m ) , the more relevant quadratic stark effect has not yet been measured .
similar to the zeeman effect , when the amplitude of the electric field is small , the linear stark effect causes spectral - hole broadening and the quadratic stark effect causes the center frequency of the spectral holes to shift .
the linear stark effect data of graf @xcite are consistent with a zero quadratic stark effect with a large uncertainty of @xmath42 hz/(v / m)@xmath18 for eu@xmath0:y@xmath1sio@xmath2 site 1 spectral holes .
the rate of spectral - hole frequency drift due to background electric and magnetic field drift can be estimated using the linear and quadratic field sensitivities together with a measurement of the rate of spectral - hole broadening .
we have measured the rate of eu@xmath0:y@xmath1sio@xmath2 site 1 spectral - hole broadening in the absence of broadening due to continuous probing by burning a 2.1 khz fwhm spectral hole and measuring the width every 30 minutes .
the observed growth rate of the spectral - hole fwhm , 0.04 hz / s , indicates that the rate of change of background fields is likely below @xmath43 ( v / m)/s and @xmath44 g / s . in a laboratory magnetic field of 0.5 g
, such a magnetic field drift would cause less than @xmath45 hz / s of spectral - hole frequency drift .
spectral - hole burning in eu@xmath0:y@xmath1sio@xmath2 shows much promise for improved llos .
the systematic frequency shifts of eu@xmath0:y@xmath1sio@xmath2 spectral holes due to magnetic field , temperature , helium gas pressure , and optical probe power fluctuations are all small enough to allow laser frequency stabilization at the @xmath19 level in the current apparatus .
we have shown qualitatively that perturbations to the spectral - hole pattern due to side - holes and anti - holes can be reduced by using narrower spectral holes and higher doping levels , but more work is required to determine the contribution of side - holes and anti - holes to shb llo frequency noise .
further studies of the quadratic stark shift and the thermomechanical noise floor are also needed , but we are optimistic that spectral - hole burning can reduce the frequency noise of lasers and extend their coherence time .
we thank r. l. cone , j. c. bergquist and d. j. wineland for useful discussions , and s. cook and c. w. oates for critical readings of this manuscript .
this work is supported by darpa quasar and onr , and is not subject to us copyright .
10 url # 1#1urlprefix[2][]#2 chou c w , hume d b , koelemeij j c j , wineland d j and rosenband t 2010 _ phys .
lett . _ * 104 * 070802 rosenband t , hume d b , schmidt p o , chou c w , brusch a , lorini l , oskay w h , drullinger r e , fortier t m , stalnaker j e , diddams s a , swann w c , newbury n r , itano w m , wineland d j and bergquist j c 2008 _ science _ * 319 * 180812 blatt s , ludlow a d , campbell g k , thomsen j w , zelevinsky t , boyd m m , ye j , baillard x , fouch m , le targat r , brusch a , lemonde p , takamoto m , hong f l , katori h and flambaum v v 2008 _ phys .
lett . _ * 100 * 140801 | several properties of eu@xmath0:y@xmath1sio@xmath2 spectral holes are measured , to assess the suitability of broad - band hole - patterns for use as laser - frequency references .
we measure frequency shifts due to magnetic fields , side - features of neighboring spectral holes , and changing optical probe power
. a precise calibration of a temperature insensitive point is also performed , where the temperature - induced frequency shift is canceled to first order by the pressure - induced shift from the crystal s helium - gas environment . |
in this paper , we present the theory of completely regular codes in the hamming graph enjoying the property that the eigenvalues of the code are in arithmetic progression .
we call these _ arithmetic completely regular codes _ and propose their full classification , modulo the classification of perfect codes of hamming type
. our results are strongest when the hamming graph admits a completely regular partition into such codes ( e.g. , the partition into cosets of some additive completely regular code ) , since it is known that the quotient graph obtained from any such partition is distance - regular . using leonard s theorem , the list of possible quotients is determined , with a few special cases left as open problems . in the case of linear arithmetic completely regular codes ,
more can be said .
the techniques employed are mainly combinatorial and products of codes as well as decompositions of codes into `` reduced '' codes play a fundamental role .
the results have impact in algebraic coding theory , but more so in the theory of distance - regular graphs .
since , for large minimum distance , the coset graph of any such linear code is a distance - regular graph locally isomorphic to a hamming graph , the analysis can be viewed as a special case of a fairly difficult open classification problem in the theory of distance - regular graphs .
what is interesting here is how the combinatorics of the hamming graph gives leverage in the combinatorial analysis of its quotients , a tool not available in the unrestricted problem .
it is important here to draw a connection between the present paper and a companion paper , currently in preparation . in @xcite
, we will introduce the concept of _ leonard completely regular codes _ and we will develop the basic theory for them
. we will also give several important families of leonard completely regular codes in the classical distance - regular graphs , demonstrating their fundamental role in structural questions about these graphs .
we conjecture that all completely regular codes in the hamming graphs with large enough covering radius are in fact leonard .
the class of arithmetic codes we introduce in this paper is perhaps the most important subclass of the leonard completely regular codes in the hamming graphs and something similar is likely true for the other classical families , but this investigation is left as an open problem .
the layout of the paper is as follows .
after an introductory section outlining the required background , we explore products of completely regular codes in hamming graphs . first noting ( prop .
[ pcxccrimplesbothcr ] ) that a completely regular product must arise from completely regular constituents , we determine in proposition [ pprodcrc ] exactly when the product of two completely regular codes in two hamming graphs is completely regular . at this point , the role of the arithmetic property becomes clear and we present a generic form for the quotient matrix of such a code in lemma [ ltridiageigenvalues ] . with this preparatory material out of the way , we are then ready to present the main results in section [ subsec : classification ] . applying the celebrated theorem of leonard , theorem [ tarithmeticquo ]
provides powerful limitations on the combinatorial structure of a quotient of a hamming graph when the underlying completely regular partition is composed of arithmetic codes .
stronger results are obtained in proposition [ pcliques ] , theorem [ tbreakdown ] and proposition [ pnotdoob ] when one makes additional assumptions about the minimum distance of the codes or the specific structure of the quotient .
when @xmath0 is a linear completely regular code with the arithmetic property and @xmath0 has minimum distance at least three and covering radius at most two , we show that @xmath0 is closely related to some hamming code .
these results are summarized in theorem [ tproductrho=1 ] , which gives a full classification of possible codes and quotients in the linear case ( always assuming the arithmetic and completely regular properties ) and corollary [ chammingquo ] which characterizes hamming quotients of hamming graphs .
in this section , we summarize the background material necessary to understand our results .
most of what is covered here is based on chapter 11 in the monograph @xcite by brouwer , cohen and neumaier .
the theory of codes in distance - regular graphs began with delsarte @xcite .
the theory of association schemes is also introduced in the book @xcite of bannai and ito while connections between these and related material ( especially equitable partitions ) can be found in godsil @xcite .
suppose that @xmath1 is a finite , undirected , connected graph with vertex set @xmath2 . for vertices
@xmath3 and @xmath4 in @xmath2 , let @xmath5 denote the distance between @xmath3 and @xmath4 , i.e. , the length of a shortest path connecting @xmath3 and @xmath4 in @xmath1 .
let @xmath6 denote the diameter of @xmath1 ; i.e. , the maximal distance between any two vertices in @xmath2 . for @xmath7 and @xmath8 ,
let @xmath9 and put @xmath10 , @xmath11 .
the graph @xmath1 is called _ distance - regular _ whenever it is regular of valency @xmath12 , and there are integers @xmath13 so that for any two vertices @xmath3 and @xmath4 in @xmath2 at distance @xmath14 , there are precisely @xmath15 neighbors of @xmath4 in @xmath16 and @xmath17 neighbors of @xmath4 in @xmath18 it follows that there are exactly @xmath19 neighbors of @xmath4 in @xmath20 .
the numbers @xmath15 , @xmath17 and @xmath21 are called the _ intersection numbers _ of @xmath1 and we observe that @xmath22 , @xmath23 , @xmath24 , @xmath25 and @xmath26 .
the array @xmath27 is called the _ intersection array _ of @xmath1 . from now on , assume @xmath1 is a distance - regular graph of valency @xmath28 and diameter @xmath29 .
define @xmath30 to be the square matrix of size @xmath31 whose rows and columns are indexed by @xmath2 with entries @xmath32 we refer to @xmath30 as the _ @xmath33 distance matrix _ of @xmath1 .
we abbreviate @xmath34 and call this the _ adjacency matrix _ of @xmath1 . since @xmath1 is distance - regular , we have for @xmath35 @xmath36 so that @xmath37 for some polynomial @xmath38 of degree @xmath14 .
let @xmath39 be the _ bose - mesner algebra _
, the matrix algebra over @xmath40 generated by @xmath41 .
then @xmath42 and @xmath43 is a basis for @xmath39 . as @xmath39 is semi - simple and commutative
, @xmath39 has also a basis of pairwise orthogonal idempotents @xmath44 .
we call these matrices the _ primitive idempotents _ of @xmath1 . as @xmath39 is closed under the entry - wise ( or hadamard ) product @xmath45 , there exist real numbers @xmath46 , called the _ krein parameters _ , such that @xmath47 the graph @xmath1 is called
_ @xmath48-polynomial _ if there exists an ordering @xmath49 of the primitive idempotents and there exist polynomials @xmath50 of degree @xmath14 such that @xmath51 , where the polynomial @xmath52 is applied entrywise to @xmath53 .
we recall that the distance - regular graph @xmath1 is @xmath48-polynomial with respect to the ordering @xmath54 of its primitive idempotents provided its krein parameters satisfy * @xmath55 unless @xmath56 ; * @xmath57 whenever @xmath58 or @xmath59 . by an _ eigenvalue _ of @xmath1 , we mean an eigenvalue of @xmath60 .
since @xmath1 has diameter @xmath6 , it has at least @xmath61 eigenvalues ; but since @xmath1 is distance - regular , it has exactly @xmath61 eigenvalues .
we denote these eigenvalues by @xmath62 and , aside from the convention that @xmath63 , the valency of @xmath1 , we make no further assumptions at this point about the eigenvalues except that they are distinct .
let @xmath1 be a distance - regular graph with distinct eigenvalues @xmath64 . by a _ code _ in @xmath1 , we simply mean any nonempty subset @xmath0 of @xmath65 .
we call @xmath0 _ trivial _ if @xmath66 or @xmath67 and _ non - trivial _ otherwise . for @xmath68 , the _ minimum distance _ of @xmath0 , @xmath69 ,
is defined as @xmath70 and for any @xmath71 the distance @xmath72 from @xmath3 to @xmath0 is defined as @xmath73 the number @xmath74 is called the _ covering radius _ of @xmath0 .
for @xmath0 a nonempty subset of @xmath2 and for @xmath75 , define @xmath76 then @xmath77 is the _ distance partition _ of @xmath65 with respect to code @xmath0 . a partition @xmath78 of @xmath2 is called _ equitable _ if , for all @xmath14 and @xmath79 , the number of neighbors a vertex in @xmath80 has in @xmath81 is independent of the choice of vertex in @xmath82 following neumaier @xcite , we say a code @xmath0 in @xmath1 is _ completely regular _ if this distance partition @xmath83 is equitable . in this case
the following quantities are well - defined : @xmath84 where @xmath3 is chosen from @xmath85 .
the numbers @xmath86 are called the _ intersection numbers _ of code @xmath0 .
observe that a graph @xmath1 is distance - regular if and only if each vertex is a completely regular code and these @xmath31 codes all have the same intersection numbers
. set the tridiagonal matrix @xmath87 for @xmath0 a completely regular code in @xmath1 , we say that @xmath88 is an _ eigenvalue of @xmath0 _ if @xmath88 is an eigenvalue of the quotient matrix @xmath89 defined above .
by @xmath90 , we denote the set of eigenvalues of @xmath0 .
note that , since @xmath91 for all @xmath14 , @xmath63 belongs to @xmath90 . given a partition @xmath92 of the vertex set of a graph @xmath1 ( into nonempty sets )
, we define the _ quotient graph _
@xmath93 on the classes of @xmath92 by calling two classes @xmath94 adjacent if @xmath95 and @xmath1 contains an edge joining some vertex of @xmath0 to some vertex of @xmath96 .
a partition @xmath92 of @xmath2 is _ completely regular _ if it is an equitable partition of @xmath1 and all @xmath97 are completely regular codes with the same intersection numbers .
if @xmath98 is a completely regular partition we write @xmath99 and we say @xmath100 is the covering radius of @xmath92 if it is the covering radius of @xmath101 . [ pbcnquo ] let @xmath102 be a + completely regular partition of any distance - regular graph @xmath1 such that each @xmath103 has intersection numbers @xmath104 , @xmath105 and @xmath106 ( @xmath107 ) . then @xmath93 is a distance - regular graph with intersection array @xmath108 remaining intersection numbers @xmath109 , and eigenvalues @xmath110 for @xmath111 .
all of these lie among the eigenvalues of the matrix @xmath112 .
+ [ psmallesteig ] let @xmath92 be a non - trivial completely regular partition of a distance - regular graph @xmath1 and assume that @xmath113 . then @xmath114 by proposition [ pbcnquo ] , the eigenvalues of @xmath93 are @xmath115 .
as @xmath93 has at least one edge , it follows that its smallest eigenvalue is at most @xmath116 .
hence @xmath117 .
+ @xmath118 let @xmath0 be a non - trivial completely regular code in a distance - regular graph @xmath1 .
let @xmath119 .
is it true that @xmath120 ?
let @xmath121 be a finite abelian group .
a _ translation distance - regular graph _ on @xmath121 is a distance - regular graph @xmath1 with vertex set @xmath121 such that if @xmath3 and @xmath4 are adjacent then @xmath122 and @xmath123 are adjacent for all @xmath124 .
a code @xmath125 is called _ additive _ if for all @xmath126 , also @xmath127 ; i.e. , @xmath0 is a subgroup of @xmath121 . if @xmath0 is an additive code in a translation distance - regular graph on @xmath121 , then we obtain the usual coset partition @xmath128 of @xmath121 ; whenever @xmath0 is a completely regular code , it is easy to see that @xmath129 is a completely regular partition . for any additive code @xmath0 in a translation distance - regular graph @xmath1 on vertex set @xmath121 ,
the _ coset graph _ of @xmath0 in @xmath1 is the graph with vertex set @xmath130 and an edge joining coset @xmath96 to coset @xmath131 whenever @xmath1 has an edge with one end in @xmath96 and the other in @xmath131 .
an important result of brouwer , cohen and neumaier ( * ? ? ? * p353 ) states that every translation distance - regular graph of diameter at least three defined on an elementary abelian group @xmath121 is necessarily a coset graph of some additive completely regular code in some hamming graph . of course , the hamming graph itself is a translation graph .
let @xmath48 be an abelian group with @xmath132 .
then we may identify the vertex set of the hamming graph @xmath133 with the group @xmath134 ; so the hamming graph can be viewed as a translation distance - regular graph in a variety of ways . for the remainder of this paper
, we will consider @xmath135-ary codes of length @xmath136 as subsets of the vertex set of the hamming graph @xmath133 . in this section
we will focus on @xmath135-ary completely regular codes , i.e. @xmath135-ary codes of length @xmath136 which are completely regular in @xmath133 .
we now recall the cartesian product of two graphs @xmath1 and @xmath137 .
this new graph has vertex set @xmath138 and adjacency @xmath139 defined by @xmath140 precisely when either @xmath141 and @xmath142 in @xmath137 or @xmath143 in @xmath1 and @xmath144 .
now let @xmath0 be a nonempty subset of @xmath2 and let @xmath96 be a nonempty subset of @xmath145 .
the cartesian product of @xmath0 and @xmath96 is the code defined by @xmath146 we are interested in the cartesian product of codes in the hamming graphs @xmath133 and @xmath147 .
note that if @xmath0 and @xmath96 are additive codes then @xmath148 is also additive .
in particular , if @xmath0 is a completely regular code in @xmath133 with vertex set @xmath149 , then @xmath150 is a completely regular code in @xmath151 .
also if @xmath152 is a completely regular partition of @xmath133 then @xmath153 is a completely regular partition of @xmath154 we say a completely regular code @xmath155 is _ non - reduced _
if @xmath156 for some @xmath157 , and _ reduced _ otherwise . in similar fashion we say that a completely regular partition is _ non - reduced _ or _
reduced_. the next three results will determine exactly when the cartesian product of two arbitrary codes in two hamming graphs is completely regular .
[ pcxccrimplesbothcr ] let @xmath0 and @xmath96 be non - trivial codes in the hamming graphs @xmath133 and @xmath147 respectively .
if the cartesian product @xmath148 is completely regular in @xmath158 , then both @xmath0 and @xmath96 themselves must be completely regular codes in their respective graphs .
if the distance partition of @xmath133 with respect to @xmath0 is @xmath159 , then we easily see that every vertex @xmath160 of @xmath161 is at distance @xmath14 from @xmath162 in the product graph .
moreover , the neighbors of this vertex which lie at distance @xmath163 from the product code are precisely @xmath164 ; so the size of the set @xmath165 must be independent of the choice of @xmath166 .
this shows that the intersection numbers @xmath104 are well - defined for @xmath0 .
an almost identical argument gives us the intersection numbers @xmath106 .
swapping the roles of @xmath0 and @xmath96 , we find that each of these is a completely regular code .
+ note that , if @xmath167 , then the product graph is not distance - regular .
although this case is not of primary interest , there are examples where @xmath148 can still be a completely regular code ( in the sense of neumaier ) in such a graph .
for instance , suppose @xmath0 is a perfect code of covering radius one in @xmath133 and @xmath96 is a perfect code of covering radius one in @xmath147 . if we happen to have @xmath168 , then , by the above proposition , @xmath148 is a completely regular code in @xmath169 with intersection numbers @xmath170 and @xmath171 , @xmath172 and @xmath173 .
it is well - known that equitable partitions are preserved under products .
if @xmath92 is an equitable partition in any graph @xmath1 and @xmath174 is an equitable partition in another graph @xmath137 , then @xmath175 is an equitable partition in the cartesian product graph @xmath176 .
the following special case will prove useful in our next proposition .
[ lequitprod ] consider completely regular codes @xmath0 in @xmath133 and @xmath96 in + @xmath147 with distance partitions @xmath177 and @xmath178 respectively .
* the partition @xmath179 is an equitable partition in the product graph @xmath180 . *
a vertex in @xmath181 has all its neighbors in @xmath182 \cup \left [ c_i \times \left ( c'_{j-1 } \cup c'_j \cup c'_{j+1 } \right ) \right].\ ] ] * if @xmath0 has intersection numbers @xmath183 and @xmath96 has intersection numbers @xmath184 , then in the product graph , a vertex in @xmath181 has : @xmath104 neighbors in @xmath185 ; @xmath106 neighbors in @xmath186 ; @xmath187 neighbors in @xmath188 ; @xmath189 neighbors in @xmath190 ; and @xmath191 neighbors in @xmath181 .
straightforward .
+ [ pprodcrc ] let @xmath0 be a non - trivial completely regular code in @xmath133 with @xmath192 and intersection numbers @xmath193 and @xmath104 ( @xmath107 ) .
let @xmath96 be a non - trivial completely regular code in @xmath147 with @xmath194 and intersection numbers @xmath195 and @xmath196 ( @xmath197 ) .
assume , without loss , that @xmath198 .
then @xmath148 is a completely regular code in @xmath169 if and only if there exist integers @xmath199 and @xmath200 satisfying ( a ) and ( b ) : * @xmath201 for @xmath107 and @xmath202 for @xmath197 ; * @xmath203 for @xmath107 and @xmath204 for @xmath197 . in this case , @xmath148 has covering radius @xmath205 and intersection numbers @xmath206 and @xmath207 for @xmath208 .
@xmath209 suppose first that @xmath148 is a completely regular code . from lemma [ lequitprod ] , we see that @xmath162 has covering radius @xmath210 . for @xmath211 , let @xmath212 then it follows easily from the second statement in lemma [ lequitprod ] that @xmath213 moreover a vertex in @xmath214 has @xmath215 neighbors in @xmath216 and @xmath217 neighbors in @xmath218 . for @xmath219 ,
this forces @xmath220 .
assume inductively that @xmath221 and @xmath222 for @xmath223 .
then , considering a vertex in @xmath224 ( where @xmath225 and @xmath226 ) , we find @xmath227 for @xmath228 , this gives @xmath229 . for @xmath230 , we deduce , @xmath231 . so we have ( a ) by induction . a symmetrical argument
establishes part ( b ) .
@xmath232 considering the same partition of @xmath233 into cells of the form @xmath234 , we obtain the converse in a straightforward manner : if @xmath0 and @xmath96 have intersection numbers given by ( a ) and ( b ) , then their cartesian product @xmath148 is completely regular in the product graph .
+ [ ltridiageigenvalues ] let @xmath235 and @xmath100 be positive integers .
the tridiagonal matrix @xmath236 where @xmath237 , has eigenvalues @xmath238 where @xmath239 . by direct verification .
( see , for example , terwilliger ( * ? ? ?
* example 5.13 ) . )
+ from now on we will look at a completely regular partition @xmath92 with @xmath240 in @xmath133 . as a direct consequence of proposition [ psmallesteig ]
, we have the following proposition .
let @xmath92 be a non - trivial completely regular partition of the hamming graph @xmath133 such that @xmath92 has covering radius @xmath100 and @xmath240 for some @xmath241 .
then @xmath242 + now we will classify the possible quotients @xmath93 where @xmath92 is a completely regular partition of @xmath243 with @xmath240 for some @xmath241 .
[ tarithmeticquo ] let @xmath1 be the hamming graph @xmath133 and let @xmath92 be a completely regular partition of its vertices .
assume that @xmath92 has covering radius @xmath244 and eigenvalues @xmath240 for some @xmath241 .
then @xmath93 is has diameter @xmath100 and is isomorphic to one of the following : * a folded cube ; * a hamming graph ; * a doob graph ( i.e. , the cartesian product of some number of 4-cliques and at least one shrikhande graph ) ; * a distance - regular graph with intersection array @xmath245 .
let @xmath92 be a completely regular partition with covering radius @xmath244 and eigenvalues @xmath246 for some @xmath241 .
denote @xmath247 where , @xmath248 and @xmath249 .
then by proposition [ pbcnquo ] , @xmath93 is a distance - regular graph with eigenvalues @xmath250 and diameter @xmath251 by ( * ? ? ? * proposition 3.3 , lemma 4.3 ) , @xmath93 is @xmath48-polynomial with respect to @xmath252 where @xmath253 is the primitive idempotent corresponding to @xmath254 leonard ( 1982 ) showed that @xmath48-polynomial distance regular - graphs with diameter @xmath6 at least @xmath255 , fall into nine types , namely , ( ) , ( ) , ( ) , ( ) , ( ) , ( ) , ( ) , ( ) and ( ) , where we follow the notation of bannai and ito ( * ? ?
type ( ) is not possible as it only occurs if @xmath256 as the eigenvalues of @xmath93 are in arithmetic progression and @xmath244 , neither type ( ) , ( ) , ( ) , ( ) , ( ) nor ( ) can occur .
the only remaining possibilities are types ( ) and ( ) .
terwilliger @xcite showed that a @xmath48-polynomial distance - regular graph @xmath137 of type ( ) with diameter @xmath257 is either the antipodal quotient of the @xmath258-cube , or has the same intersection numbers as the antipodal quotient of the @xmath259-cube . by ( * ? ? ?
* theorem 9.2.7 ) , it follows that @xmath93 is either the folded graph of a @xmath136-cube for @xmath260 or has intersection array @xmath245 . for type ( ) , we have ( in the notation of bannai and ito ( * ? ? ?
* p270 ) ) @xmath261 @xmath262 @xmath263 @xmath264 since @xmath265 @xmath266 and @xmath267 hence @xmath93 has the same parameters as a hamming graph .
but egawa @xcite showed that a distance - regular graph with same parameters as a hamming graph must be a hamming graph or a doob graph .
hence the theorem is proved .
+ a. : : theorem [ tarithmeticquo ] also holds when @xmath1 is a doob graph .
there are completely regular partitions of doob graphs that are hamming graphs with @xmath268 for example , let @xmath269 be an integer . in any doob graph @xmath1 of diameter @xmath270 , there exist an additive completely regular code , say @xmath0 , with covering radius @xmath271 _ @xcite_. let @xmath272 be an integer .
then @xmath273 is an additive completely regular code with covering radius @xmath136 in @xmath274 recall @xmath275 then @xmath276 .
b. : : there are 3 nonisomorphic graphs with the same intersection array , @xmath277 @xmath278 , as the folded 6-cube .
they are the point - block incident graphs of the 2-(16,6,2)-designs .
we do not know any example in which the quotient graph has intersection array @xmath245 , but is not the folded 6-cube .
c. : : if @xmath1 is a hamming graph and @xmath0 is an additive completely regular code in @xmath1 satisfying the eigenvalue conditions of the above theorem , then @xmath279 is always a hamming graph , a doob graph or a folded cube .
we do not know any example where @xmath1 is a hamming graph and @xmath92 is a completely regular partition such that @xmath93 is a doob graph .
but we will prove below in proposition [ pnotdoob ] that this can not occur when @xmath280 for some linear completely regular code @xmath0 .
d. : : we wonder whether it is true that if the quotient graph @xmath93 in the above theorem [ tarithmeticquo ] is isomorphic to a hamming graph or a doob graph , then each part in @xmath92 can be expressed as a cartesian product of completely regular codes with covering radius at most 2
. we will show ( theorem [ tproductrho=1 ] , below ) that this holds when @xmath92 is the coset partition of a linear completely regular code satisfying the eigenvalue conditions of the above theorem .
[ pcliques ] let @xmath1 be the hamming graph @xmath133 .
let @xmath92 be any completely regular partition of @xmath1 where each code @xmath0 in @xmath92 has minimum distance @xmath281 . * if @xmath282 then @xmath283 * if @xmath284 then @xmath93 is not isomorphic to any doob graph . *
if @xmath285 then @xmath93 can not have the same intersection array as any folded cube of diameter at least two ( including the array @xmath286 ) .
* suppose further that @xmath280 for some additive code @xmath0 and @xmath93 has intersection array @xmath245 .
then @xmath287 and @xmath93 is the folded @xmath288-cube .
since each @xmath0 in @xmath92 satisfies @xmath289 , the vertices in any clique in @xmath1 belong to pairwise distinct classes in @xmath92 .
so the quotient @xmath93 has a clique of size at least @xmath135 , the clique number of @xmath133 .
this immediately implies ( a)-(c ) as : @xmath290 has no clique of size larger than @xmath291 ; any doob graph has maximum clique size three or four ; a folded cube other than @xmath292 has maximum clique size two , as does any graph with intersection array @xmath245 . for part ( d ) , we clearly need only consider the case where @xmath287 . as @xmath281 , it is easy to see that every claw @xmath293 in @xmath93 belongs to a unique 3-cube so that @xmath93 must be the folded @xmath288-cube ( * ? ? ?
* corollary 4.3.8 ) .
+ [ eh24h22 ] let @xmath1 be the hamming graph @xmath294 and let @xmath295 , where @xmath296 @xmath297 @xmath298 and @xmath299 then the quotient graph @xmath93 is the hamming @xmath300 as depicted in figure 1 .
observe that each class in partition @xmath92 is expressible as a cartesian product : @xmath301 @xmath302 finally note that this example can be easily extended to give @xmath133 as a quotient of @xmath303 , @xmath304 from example [ eh24h22 ] , we see that the @xmath136-cube and the folded @xmath136-cube can be quotients of @xmath305 .
so proposition [ pcliques ] does not hold when @xmath306 , even for additive codes .
we next show that , if @xmath92 is the coset partition of @xmath133 with respect to a linear code with a quotient of hamming type , then each class in @xmath92 may be expressed as a cartesian product of completely regular codes with covering radius one .
[ tbreakdown ] let @xmath1 be the hamming graph @xmath133 .
let @xmath0 be an additive completely regular code with minimum distance @xmath281 in @xmath1 and let @xmath129 be the coset partition of @xmath1 with respect to @xmath0 .
suppose that @xmath307 then @xmath308 divides @xmath136 and @xmath309 , where each @xmath103 is a @xmath135-ary completely regular code with covering radius @xmath271 and length @xmath310 .
let @xmath311 denote the codeword with @xmath33 position @xmath271 and all other positions @xmath312 . since @xmath0 is additive , the relation @xmath313 defined by @xmath314 if and only if @xmath315 is an equivalence relation on @xmath316 .
let @xmath317 denote the equivalence class containing @xmath14 .
note that @xmath318 since @xmath319 and @xmath320 .
since @xmath321 and @xmath322 have @xmath323 common neighbors in @xmath1 while @xmath0 and @xmath324 have @xmath325 common neighbors in @xmath279 , we find that @xmath326 .
this gives @xmath327 for @xmath328 .
so there are exactly @xmath308 equivalence classes and they all have the same size .
let @xmath329 denote the set of vertices @xmath330 of @xmath133 satisfying @xmath331 for all @xmath332 and let @xmath137 be the subgraph of @xmath1 induced by @xmath333 so @xmath334 .
note that @xmath137 has the property @xmath335 : if @xmath336 and @xmath337 then the common neighbors of @xmath330 and @xmath338 in @xmath1 are also in @xmath339 .
let @xmath340 .
* _ claim : _ * @xmath103 is a @xmath135-ary completely regular code with covering radius 1 in @xmath137 with @xmath341 where @xmath342 .
+ _ proof of claim : _ we show the following two statements ( i ) and ( ii ) .
+ ( i ) for any @xmath343 , distance @xmath344 : + this is obvious by induction on the weight of @xmath345 , as @xmath137 has property @xmath335 .
+ ( ii ) @xmath103 is a @xmath135-ary completely regular code of length @xmath310 : + assume that for @xmath343 , @xmath346 .
then clearly @xmath347 by ( i ) and @xmath340 .
this gives us the claim . +
to finish the proof of the theorem note that @xmath348 so that @xmath349 as @xmath0 is additive . by proposition 3.4
, @xmath350 is a completely regular code with the same quotient matrix as @xmath0 .
so @xmath309 .
this shows the theorem .
+ [ ldelta>=2 ] assume that @xmath135 is a prime power .
let @xmath0 be a non - trivial reduced linear completely regular code over @xmath351 of length @xmath136 .
then the minimum distance of @xmath0 , @xmath69 , is at least @xmath352 .
+ [ pnotdoob ] if @xmath0 is a linear @xmath135-ary completely regular code of length @xmath136 , then @xmath353 is never isomorphic to a doob graph . without loss of generality
, we may assume @xmath0 is reduced .
since every doob graph has the shrikhande graph as a quotient , we may suppose that @xmath354 , the shrikhande graph . since @xmath0 is non - trivial and reduced , by lemma [ ldelta>=2 ] , it has minimum distance @xmath281 and by proposition [ pcliques](b ) , we have @xmath355 .
the shrikhande graph is locally a @xmath288-gon ; denote by @xmath356 the six vertices adjacent to vertex @xmath0 in the coset graph , with the understanding that @xmath357 is adjacent to @xmath358 ( @xmath359 ) in @xmath137 .
let @xmath311 denote the codeword with @xmath33 position @xmath271 and all other positions @xmath312 .
consider first the case @xmath287 .
with an appropriate ordering of the coordinates , we may assume that @xmath360 , @xmath361 and @xmath362 .
but then @xmath363 must lie in @xmath364 and a contradiction is obtained by considering which coset contains @xmath365 : this coset must be distinct from @xmath0 , @xmath364 and @xmath366 ( as @xmath289 ) and yet adjacent to all three . but
@xmath137 contains no 4-clique .
it remains to consider @xmath367 .
we may suppose , without loss that @xmath368 and @xmath369 .
now some neighbor of the zero vector we may call it @xmath365 lies in @xmath370 .
the word @xmath371 must lie in some coset adjacent to both @xmath0 and @xmath370 since @xmath289 .
it can not be @xmath366 since otherwise @xmath372 forcing @xmath373 and @xmath365 into the same coset .
so @xmath374 .
but now we obtain a contradiction by considering the coset containing @xmath375 . in @xmath137
, this coset must be adjacent to both @xmath366 and @xmath376 .
if @xmath377 , then @xmath378 and @xmath365 must belong to the same coset , contrary to the way @xmath365 was chosen .
likewise , @xmath379 is not possible since @xmath373 and @xmath365 were chosen to belong to distinct cosets .
+ [ lequivreln ] let @xmath0 be a non - trivial , reduced , linear completely regular code over @xmath351 in the @xmath133 with parity check matrix @xmath380 .
let @xmath381 ( @xmath382 ) denote columns of @xmath380 and consider the relation @xmath383 on @xmath384 given by @xmath385 if @xmath381 and @xmath386 are linearly dependent .
then @xmath383 is an equivalence relation and its each equivalence class has size @xmath387 it is easy to check that the relation @xmath383 is reflexive , symmetric and transitive and hence an equivalence relation . as @xmath388 has @xmath389 neighbors in @xmath0 ,
there are @xmath390 two - weight codewords in @xmath0 and each of these codewords is orthogonal to every row of @xmath380 .
this means immediately that each equivalence class has size @xmath389 .
+ let @xmath0 be a non - trivial , reduced , linear completely regular code over @xmath351 in @xmath133 .
let @xmath391 such that each equivalence class of @xmath383 defined in lemma [ lequivreln ] contains exactly one element from @xmath392 .
let @xmath393 then @xmath6 is linear and has minimum distance @xmath394 moreover , if @xmath0 has @xmath395 and @xmath6 is completely regular , then @xmath6 has @xmath396 , where @xmath397 .
the following result is probably known but we could not find it in the literature .
let @xmath0 be a linear completely regular code over @xmath351 with minimum distance @xmath398 in @xmath133 .
* if @xmath0 has @xmath399 for some @xmath400 , then @xmath0 is a hamming code . *
if @xmath0 has @xmath401 for some @xmath400 , then @xmath402 , where @xmath6 is a hamming code or @xmath287 and @xmath0 is an extended hamming code .
\(a ) since @xmath0 has covering radius one and minimum distance three , it follows immediately that @xmath0 has to be perfect .
\(b ) let @xmath403 be the graph having as vertices the codewords of @xmath404 in which two codewords @xmath405 and @xmath345 are joined if and only if @xmath406 .
then by ( * ? ? ?
* thm 5.7 ) , @xmath403 is strongly regular . with @xmath407 denoting the number of common neighbors of two nonadjacent vertices in @xmath403 , we first aim to show that @xmath408 when @xmath409 .
let @xmath410 be codewords of @xmath404 such that @xmath411 , @xmath412 and @xmath413 .
let us consider the common neighbors of @xmath321 and @xmath414 . if @xmath415 is a common neighbor of these two which is distinct from @xmath416 and @xmath417 , then @xmath418 , @xmath419 and @xmath420 .
if @xmath421 , then @xmath422 must have form @xmath423 for some @xmath424 since otherwise , there exist @xmath425 for which the codeword @xmath426 has weight lying strictly between @xmath312 and @xmath241 . but
@xmath427 is also impossible as @xmath428 and not @xmath429 . if @xmath430 , then for @xmath431 , the codewords @xmath432 and @xmath433 have distinct weights both less than @xmath429 .
so whenever @xmath409 , we must have @xmath408 . in this case , @xmath403 must be an @xmath434-grid graph , i.e , @xmath435 , where @xmath436 and , with the same reasoning as in theorem [ tbreakdown ] , the result follows .
suppose now that @xmath287 and @xmath437 .
let us consider the induced subgraph @xmath438 of @xmath403 on the common neighbors of @xmath321 and @xmath439 then @xmath438 is a cocktail party graph and by lemma [ ltridiageigenvalues ] , @xmath440 for @xmath441 , let @xmath414 and @xmath422 be codewords in @xmath404 such that @xmath442 .
then as @xmath404 is linear , @xmath443 is either @xmath444 or @xmath241 .
but in both cases , we must have @xmath445 a contradiction . for @xmath446 , there exist only one codeword @xmath330 with @xmath447 . since @xmath403 is vertex transitive , @xmath403 must be a cocktail party graph ( i.e. , the complement of a perfect matching ) .
this implies @xmath404 is the dual of an extended hamming code so @xmath0 is an extended hamming code .
hence the result follows .
+ [ tproductrho=1 ] let @xmath0 be a non - trivial , reduced , linear completely regular code over @xmath351 .
suppose further that @xmath0 has covering radius @xmath100 in @xmath133 and @xmath448 for some @xmath241 .
then one of the following holds : * @xmath287 and @xmath449 where @xmath450 is a parity check matrix for a binary repetition code , @xmath397 and the quotient is the folded cube ; * @xmath451 and @xmath452 where @xmath380 is a parity check matrix for some hamming code and @xmath397 ; * @xmath453 , @xmath287 and @xmath454 where @xmath455 is a parity check matrix for a fixed extended hamming code and @xmath397 ; * @xmath456 , @xmath457 and @xmath458 where @xmath459 is a completely regular code with covering radius 1 .
the reader can easily check that examples ( a)(d ) are all completely regular .
( see also bier @xcite . ) that they are the only completely regular codes possible follows directly from proposition 3.9 , theorem 3.11 , proposition 3.13 and theorem 3.15 .
+ note that the above result includes the following generalization of the result of bier @xcite concerning coset graphs which are isomorphic to hamming graphs ( cf .
* p354 ) ) : [ chammingquo ] let @xmath0 be a linear completely regular code in @xmath133 whose coset graph is a hamming graph @xmath290 .
then one of the following holds : * @xmath451 and @xmath452 where @xmath380 is a parity check matrix for some hamming code and @xmath397 ; * @xmath453 , @xmath287 and @xmath454 where @xmath455 is a parity check matrix for the extended hamming code and @xmath397 . * @xmath456 , @xmath457 and @xmath458 where @xmath459 is a completely regular code with covering radius 1 .
assume @xmath460 .
the eigenvalues of the hamming graph @xmath279 are @xmath461 ( @xmath462 ) but are also obtained from proposition [ pbcnquo ] giving @xmath463 this in turn gives @xmath464 for @xmath462
. so theorem [ tproductrho=1 ] applies .
part of this work was completed while the third author was visiting pohang institute of science and technology ( postech ) .
wjm wishes to thank the department of mathematics at postech for their hospitality and com@xmath465mac for financial support .
jhk and lws are partially supported by the basic science research program through the national research foundation of korea ( nrf ) funded by the ministry of education , science and technology ( grant number 2009 - 0089826 ) .
jhk was also partially supported by a grant of the korea research foundation funded by the korean government ( moehrd ) under grant number krf-2007 - 412-j02302 .
wjm wishes to thank the us national security agency for financial support under grant number h98230 - 07 - 1 - 0025 .
99 , _ algebraic combinatorics i : association _ _ schemes _ , benjamin - cummings lecture note ser .
58 , benjamin / cummings , london , 1984 .
, _ a family of nonbinary linear codes _ , discrete math .
65 no . 1 ( 1987 ) ,
41 - 51 . , _ distance - regular graphs _ ,
springer , heidelberg , 1989 .
, _ the geometry of two - weight codes _ , bull .
london math .
. 2 ( 1986 ) , pp .
97 - 122 . , _ an algebraic approach to the association schemes of coding theory _ , philips res .
no . 10 ( 1973 ) , vi+97 . ,
_ characterization of h(n , q ) by the parameters _ , j. combinatorial theory ( a ) , vol . 31 ( 1981 ) ,
108 - 125 . ,
_ algebraic combinatorics _ , chapman & hall , new york , 1993 . , _ algebraic properties of completely regular codes _ , in preparation . ,
tight @xmath352-designs and perfect @xmath271-codes in doob graphs _ , special issue in honor of professor ralph stanton
. j. statist .
inference 86 , no . 2 ( 2000 ) ,
505 - 513 . ,
_ completely regular codes , in : a collection of contributions in honour of jack van lint _
, discrete math .
, 106/107 ( 1992 ) , pp .
353 - 360 .
, _ the classification of distance - regular graphs of type iib _ , combinatorica vol . 8 ( 1988 ) , pp .
125 - 132 .
, _ two linear transformations each tridiagonal with respect to an eigenbasis of the other ; comments on the parameter array _ , des
. codes cryptogr . , 34 ( 2005 ) , no . 2 - 3 , pp .
307 - 332 . | in this paper , we explore completely regular codes in the hamming graphs and related graphs .
we focus on cartesian products of completely regular codes and products of their corresponding coset graphs in the additive case .
connections to the theory of distance - regular graphs are explored and several open questions are posed . |
before addressing the convergence theory , we would like to discuss stochastic noise modeling and its intrinsic conflict with the deterministic model . here and throughout the rest of the paper ,
assume @xmath31 to be a complete probability space with a set @xmath32 of outcomes of the stochastic event , @xmath33 the corresponding @xmath34-algebra and @xmath35 a probability measure , @xmath36 $ ] .
we restrict ourselves here to probability measures for the sake of simplicity .
extensions to more general measures are straightforward . in the hilbert - space
setting , the noise is typically modeled as follows , see for example @xcite .
let @xmath37 be a stochastic process .
then for @xmath1 @xmath38 defines a real - valued random variable . assuming that @xmath39 for all @xmath40 and that this expectation is continuous in @xmath41
, @xmath42 defines a continuous , symmetric nonlinear bilinearform .
in particular , there exists the _ covariance operator _
@xmath43 with @xmath44 for the stochastic analysis of infinite dimensional problems via deterministic results , is problematic .
namely , if @xmath45 is an orthonormal basis in @xmath46 , the set @xmath47 consists of infinitely many identically distributed random variables with @xmath48 @xcite .
thus @xmath49 is almost surely infinite and a realization of the noise is an element of the hilbert space @xmath46 with probability zero .
let us take the common example of gaussian white noise which can be modeled via the above construction .
namely , with @xmath50 and the covariance operator @xmath51 where @xmath52 is the identity and @xmath53 the variance parameter , the gaussian white noise is described @xcite . as consequence of ( [ eq : eerror ] ) and
explained for example in @xcite , a realization of such a gaussian random variable is with probability zero an element of an infinite dimensional @xmath54-space .
it is therefore inappropriate to use an @xmath54-norm for the residual in case of an infinite dimensional problem . since in this case
a realization of gaussian white noise only lies ( almost surely ) in any sobolev space @xmath55 with @xmath56 where @xmath57 is the dimension of the domain , one should adjust the norm for the residual accordingly . except for the paper @xcite
this issue seems not to have been addressed in the literature . a main reason for this might be that for the practical solution of the inverse problem this is not a severe issue since in reality the measurements are finite dimensional and , in order to use a computer to solve the problem , a finite dimensional approximation of the unknown object has to be used .
in this case the sum in is finite and the noise lies almost surely in the finite dimensional space .
however , difficulties arise whenever one seeks to investigate convergence of the discretized problem to its underlying infinite dimensional problem
. we will not address this issue and assume throughout the whole work that @xmath58 or use the slightly weaker bound on the ky - fan metric ( see section [ ssec : kyfan ] ) . in order to handle the ky fan metric we need to be able to evaluate probabilities @xmath59 , @xmath60 , which is only meaningful if @xmath61 .
assuming that @xmath46 is finite dimensional , then this is clear . for infinite dimensional problems , however , we have to assume that the noise is smooth enough for the sum in ( [ eq : eerror ] ) to converge .
examples for this are brownian noise ( @xmath62-noise ) or pink noise ( @xmath63-noise ) , see e.g. @xcite . at this point
we would also like to mention that as a consequence of our rather generic noise model we might not make use of some specific properties of the noise as would be possible when focusing on a particular distribution of the noise .
however , we are able to show convergence for a large variety of regularization methods .
in order to measure the magnitude of the stochastic noise and the quality of the reconstructions , we need metrics that incorporate the stochastic nature of the problem
. one such metric , which will be the the main tool for our stochastic convergence analysis , is the ky fan metric ( cf .
it is defined as follows .
let @xmath64 and @xmath65 be random variables in a probability space @xmath66 with values in a metric space @xmath67 .
the distance between @xmath64 and @xmath65 in the _ ky fan metric _ is defined as @xmath68 we will often drop the explicit reference to @xmath69 . this metric essentially allows to lift results from a metric space to the space of random variables as the connection to the deterministic setting is inherent via the metric @xmath70 used in its definition .
the deterministic metric is often induced by a norm @xmath71 .
we will implicitly assume that equation is scaled appropriately since @xmath72 @xmath73 by definition .
note that one can use definition also if @xmath74 is a more general distance function than a metric .
then the construction itself is no longer a metric , however , the techniques used in later parts of the paper can readily be expanded to this setting .
an immediate consequence of ( [ def : kyfan ] ) is that @xmath75 if and only if @xmath76 almost surely . convergence in the ky fan metric is equivalent to convergence in probability , i.e. , for a sequence @xmath77 and @xmath78 one has @xmath79 hence convergence in the ky fan metric also leads to pointwise ( almost sure ) convergence of certain subsequences in the metric @xmath70 @xcite .
a somewhat more intuitive and more frequently used stochastic metric is the expectation , or more general , a ( stochastic ) @xmath80 metric .
for random variables @xmath81 and @xmath82 with values in a metric space @xmath83 , @xmath84 defines the @xmath85-th moment of @xmath86 for @xmath87 , assuming the existence of the integral .
we will use @xmath88 and refer to it as _ convergence in expectation_. note that since the variance is defined as @xmath89 one always has @xmath90 we will show later that for parameter choice rules the expectation of the noise has to be slightly overestimated , hence estimating @xmath91 via the popular and often easier to compute @xmath54-norm @xmath92 with is not problematic . while the main part of our analysis is based on the description of the noise and the reconstruction quality in the ky fan metric
, we will also allow the expectation as measure of the stochastic noise and partially show convergence of the reconstructed solutions in expectation . to this end
, we comment in the following on the connection between those two metrics .
it is well - known that convergence in expectation implies convergence in probability , see for example @xcite .
hence , convergence in the ky fan metric is implied by convergence in expectation ( and also by convergence of higher moments ) .
namely , with markovs inequality one has , for an arbitrary nonnegative random variable @xmath93 with @xmath94 and @xmath95 @xmath96 under an additional assumption one can show conversely that convergence in probability implies convergence in expectation .
we have the following definition .
[ def : uniform_int ] let @xmath66 be a complete probability space .
a family @xmath97 is called _ uniformly integrable _
if @xmath98 [ thm : uniform_int ] let @xmath99 be a sequence convergent almost everywhere ( or in probability ) to a function @xmath7 .
if the sequence @xmath100 is uniformly integrable , then it converges to @xmath7 in the norm of @xmath101 . from a practical point of view
, uniform integrability of a sequence of regularized solutions to an inverse problem is a rather natural condition . since inverse problems typically arise from some real - world application , it is to be expected that the true solution is bounded .
for example , in computer tomography , the density of the tissue inside the body can not be arbitrarily high .
although for an inverse problem with a stochastic noise model , boundedness of the regularized solutions can not be guaranteed due to the possibly huge measurement error , one can enforce the condition from a priori knowledge of the solution .
assume that the true solution @xmath102 fulfills @xmath103 and @xmath104 globally for some fixed @xmath105 . under this assumption ,
let @xmath106 be a sequence of regularized solution for noisy data with variance @xmath107 .
let @xmath108 and define @xmath109 then the sequence @xmath110 is uniformly integrable . in other words , by discarding solutions that must be far away from the true solution in regard of a priori knowledge , convergence in the ky fan metric implies convergence in expectation . to close this section ,
let us remark on the computation of the ky fan distance .
it can be estimated via the moments of the noise .
let @xmath111 be random variables in a complete probability space @xmath66 and @xmath112 for some @xmath113
. then @xmath114{\mathbb{e}(d_\mathcal{y}(y_1,y_2)^s)}\ ] ] one has , due to markov s inequality and the monotonicity of the mapping @xmath115 for @xmath116 , @xmath117 for @xmath118 . solving @xmath119 for @xmath120 yields the claim .
note that even if moments exist for all @xmath113 @xmath121{\mathbb{e}(d_\mathcal{y}(y_1,y_2)^s)}\neq \mathbb{e}(d_\mathcal{y}(y_1,y_2)),\ ] ] see @xcite , due to the tail of the distributions .
in the gaussian case , a direct estimate has been derived in @xcite .
we present it in proposition [ prop : kyfandistance ] .
as mentioned previously , the intention of this paper is to show convergence for inverse problems under a stochastic noise model using results from the deterministic setting .
assume we have at hand a deterministic regularization method of our liking for the solution of under the noisy data where now @xmath122 for some @xmath9 . by regularization method
we understand ( possibly nonlinear ) mappings @xmath123 where @xmath124 is the regularized solution to the regularization parameter @xmath16 given the data @xmath22 .
often , @xmath125 is obtained via the minimization of functionals of the type . in order to deserve the name regularization we require @xmath126 to fulfill @xmath127 under a certain choice of the regularization parameter @xmath16 chosen either a priori @xmath128 or a posteriori @xmath129 . in our notation @xmath102
is the true solution , usually the minimum norm solution with respect to the penalty @xmath130 in , i.e. , @xmath131 note that , in particular for nonlinear problems , @xmath102 does not need to be unique . in @xcite
it was pointed out that this is problematic for the lifting arguments . a standard argument in the deterministic theory is to prove convergence of subsequences to the desired solution , and then deduces convergence of the whole series of regularized solutions , if possible . in the stochastic
setting , this is not possible in general since subsequences for different @xmath69 do not have to be related
. a constructed example for this behavior can be found in section 4.1 . of @xcite . in order to lift general deterministic regularization methods into the stochastic setting
we must therefore require that @xmath102 is unique .
we formulate our convergence results assuming the noise to be bounded with respect to the ky fan metric or in expectation .
as we will see , in the latter case we have to `` inflate '' the expectation for decreasing variance @xmath18 in order to obtain convergence . for the analysis we mainly use a lifting argument using deterministic theory . in (
* ; * ? ? ?
* theorem 4.1 ) , it was proved how by means of the ky fan metric deterministic results can be lifted to the space of random variables for nonlinear tikhonov regularization .
since the theorem is based solely on the fact that there is a deterministic regularization theory and that the probability space @xmath32 can be decomposed into a part where the deterministic theory holds and a small part where it does not , it is easily generalized .
before we state the theorem , we need the following lemmata .
( @xcite , see also @xcite)[lem : egoroff ] let @xmath66 be a complete probability space .
let @xmath132 and @xmath7 be measurable functions from @xmath32 into a metric space @xmath133 with metric @xmath70 .
suppose @xmath134 for @xmath35-almost all @xmath135.then for any @xmath136 there is a set @xmath137 with @xmath138 such that @xmath139 uniformly on @xmath137 , that is @xmath140 [ prop : kyfan ] let @xmath100 be a sequence of random variables that converges to @xmath7 in the ky fan metric .
then for any @xmath141 and @xmath136 there exist @xmath142 , @xmath143 , and a subsequence @xmath144 with @xmath145 furthermore there exists a subsequence that converges to @xmath7 almost surely .
we give a sketch of the proof for the first statement taken from @xcite .
+ set @xmath146 . by definition of the ky fan metric ( [ def : kyfan ] ) , for given @xmath147 , there exists a set @xmath148 with @xmath149 and @xmath150 such that @xmath151 . for arbitrary @xmath136 and @xmath152
we pick a subsequence @xmath153 with @xmath154 and introduce the set @xmath155
. one can check that @xmath156 . since @xmath137 is a subset of every @xmath157 we have @xmath158 which proves he first statement .
the second one follows since convergence in ky fan metric is equivalent to convergence in probability , which itself implies almost - sure convergence of a subsequence , cf @xcite . with this
, we are ready for the convergence theorem which we shall split in two parts , one for the ky fan metric as error measure and one for the expectation .
[ thm : lifting_convergence_kyfan ] let @xmath126 be a regularization method for the solution of in the deterministic setting under a suitable choice of the regularization parameter .
let now @xmath159 where @xmath160 is a stochastic error such that @xmath161 as @xmath162 . then , assuming has a unique solution @xmath102 and all necessary assumptions for the deterministic theory ( except the bound on the noise ) hold with probability one , the regularization method @xmath126 fulfills @xmath163 under the same parameter choice rule as in the deterministic setting with @xmath164 replaced by @xmath165 .
if the regularized solutions are defined by , then it holds that @xmath166 denote @xmath167 .
define @xmath168 .
( note that @xmath169 due to the properties of the ky fan metric ) .
we show in the following that for arbitrary @xmath136 we have @xmath170 and hence @xmath171 as a first step we pick a worst case " subsequence @xmath172 of @xmath173 , a subsequence for which the corresponding solutions satisfy @xmath174 .
we now show that even from this worst case " sequence we can pick a subsequence @xmath175 for which we have @xmath176 for arbitrary @xmath136 .
+ let @xmath136 . according to lemma [ prop : kyfan ]
we can pick a subsequence @xmath175 and a set @xmath137 with @xmath177 as well as @xmath178 , @xmath141 arbitrarily small , on @xmath137 . for all @xmath179
, the noise tends to zero .
we can therefore use the deterministic result with @xmath180 and deduce that @xmath181 converges to the unique solution @xmath182 for @xmath183 , @xmath179 where in the choice of the regularization parameter @xmath164 is substituted by @xmath184 .
the convergence is not uniform in @xmath69 ; nevertheless , pointwise convergence implies uniform convergence except on sets of small measure according to lemma [ lem : egoroff ] .
therefore there exist @xmath185 , @xmath186 and @xmath187 such that @xmath188 @xmath189 and @xmath190 .
we thus have @xmath191 since we split @xmath192 with @xmath193 , @xmath194 we have shown existence of a subsequence @xmath195 such that @xmath196 for @xmath195 sufficiently small .
this @xmath197 is by definition of the ky fan metric an upper bound for the distance between @xmath198 and @xmath102 .
therefore we have @xmath199 on the other hand , the original sequence satisfied @xmath200 . since @xmath201 it follows @xmath170 . because @xmath136 was arbitrary , this implies @xmath202 , which concludes the proof of convergence in the ky fan metric .
convergence in expectation follows from theorem [ thm : uniform_int ] noting that by the sequence of regularized solutions is uniformly integrable .
[ thm : lifting_convergence_exp ] let @xmath126 be a regularization method for the solution of in the deterministic setting under a suitable choice of the regularization parameter .
let now @xmath159 where @xmath160 is a stochastic error such that @xmath203 as @xmath162 . then
, assuming has a unique solution @xmath102 and all necessary assumptions for the deterministic theory ( except the bound on the noise ) hold with probability one , the regularization method @xmath126 fulfills @xmath163 under the same parameter choice rule as in the deterministic setting with @xmath164 replaced by @xmath204 where @xmath205 fulfills @xmath206 if the regularized solutions are defined by then it holds that @xmath166 as previously we pick a
worst case " subsequence @xmath172 of @xmath173 , a subsequence for which the corresponding solutions satisfy @xmath174 .
+ let @xmath136 .
we can now pick a subsequence which we again denote by @xmath175 fulfilling @xmath207 , where without loss of generality @xmath208 , such that @xmath209 this again defines , via the complement in @xmath32 , @xmath137 with @xmath177 on which @xmath210 .
as before , we can now apply the deterministic theory by substituting @xmath164 with @xmath211 .
the remainder of the proof is identical to the one of theorem [ thm : lifting_convergence_exp ] .
the theorems justify the use of deterministic algorithms under a stochastic noise model .
since the proof is solely based on relating the stochastic noise to a deterministic one on subsets of @xmath32 and does not use any specific properties of the regularization methods or the underlying spaces , it opens most of the deterministic methods for the a stochastic noise model .
in particular , the parameter choice rules from the deterministic setting are easily adapted .
as usual in deterministic literature , the general convergence theorem is followed by convergence rates which are obtained under additional assumptions .
often these conditions ensure at least local uniqueness of the true solution . if not , we have to require such a property for the same reason as previously .
[ thm : lifting_rates_kyfan ] let @xmath126 be a regularization method for the solution of in the deterministic setting such that , under a set of assumptions on the operator @xmath212 and the solutions @xmath102 and a suitable choice of the regularization parameter , @xmath213 with a monotonically increasing right - continuous function @xmath214 .
let now @xmath159 where @xmath160 is a stochastic error such that * @xmath161 or * @xmath203 as @xmath162 .
then , assuming all necessary assumptions for the deterministic theory ( except the bound on the noise ) hold with probability one and that there is ( either by the deterministic conditions or by additional assumption ) a ( locally ) unique solution @xmath102 to , the regularization method @xmath126 fulfills @xmath215 in case a ) or , respectively , in case b ) , @xmath216 under the same parameter choice rule as in the deterministic setting with @xmath164 replaced by @xmath165 ( case a ) ) or @xmath204 where @xmath205 fulfills ( case b ) ) .
we start again with the ky fan distance as noise measure .
since we have the deterministic theory at hand , we know that @xmath217 whenever @xmath218 . with @xmath219
we have , since @xmath214 is monotonically increasing and right continuous , @xmath220 and hence by definition @xmath221 .
if the expectation is used as measure for the data error , we have @xmath222 by markovs inequality . hence , with probability @xmath223 we are in the deterministic setting with @xmath224 and @xmath225 the convergence rate follows by the definition of the ky fan metric . for inverse problems ,
the convergence rates are most often given by functions which decay at most linearly fast , i.e. , @xmath226 hence in this case the convergence rates are preserved in the ky fan metric . for the expectation
this is not the case .
we have to gradually inflate the expectation by the parameter @xmath227 in order to obtain convergence ( and rates ) .
let us discuss the simple example of gaussian noise in the finite dimensional setting , i.e. @xmath5 from consists of @xmath228 i.i.d .
random variables @xmath229 with zero mean and variance @xmath230 .
then it has been shown in @xcite that for any @xmath231 @xmath232 with the gamma functions @xmath233 and @xmath234 defined as @xmath235 in particular , is independent of the variance @xmath230 . in order to to decrease the probability to zero
, we therefore have to link @xmath227 with the variance . for gaussian noise of the above kind
the following estimate for the ky fan distance between true and noisy data has been given in @xcite .
[ prop : kyfandistance ] let @xmath236 be a random variable with values in @xmath237 .
assume that the distribution of @xmath5 is @xmath238 with @xmath239 .
then it holds in @xmath240 that @xmath241 recall that @xmath242 see e.g. @xcite . comparing and
, one sees that @xmath243 and in particular the decay of @xmath244 slows down with decreasing @xmath18 . in other words ,
the artificial inflation we had to impose on the expectation is automatically included in the ky fan distance which we suppose is the reason why the convergence theory carries over in such a direct fashion for the ky fan metric . for many nonlinear inverse problems the requirement of a unique solution is too strong .
often one has several solutions of the same quality , in particular there exists more than one minimum norm solution . in this case ,
theorem [ thm : lifting_convergence_kyfan ] is not applicable . in the example (
* example 4.3 and 4.5 ) with two minimum norm solutions the noise was constructed such that , while the error in the data converges to zero , for each fixed @xmath135 the regularized solutions jump between both solutions such that no converging subsequence can be found .
the main problem there is that the ky fan distance can not incorporate the concept that all minimum norm solutions are equally acceptable .
we will now define a pseudo metric that resolves this issue .
let @xmath245 be a metric space . denote with @xmath246 the set of minimum - norm solutions to
. then @xmath247 measures the distance between an element @xmath2 and the set @xmath246 , in particular it is @xmath248 with this , one can define a pseudometric on @xmath66 via @xmath249 obviously is positive , symmetric and fulfills the triangle inequality .
however , @xmath250 does not imply @xmath251 a.e .
but instead @xmath252 which fixes the aforementioned issue of the ky fan metric and allows the following theorems . let @xmath126 be a regularization method for the solution of in the deterministic setting under a suitable choice of the regularization parameter .
let now @xmath159 where @xmath160 is a stochastic error such that * @xmath161 or * @xmath203 as @xmath162 . then , assuming all necessary assumptions for the deterministic theory ( except the bound on the noise ) hold with probability one , the regularization method @xmath126 fulfills @xmath253 under the same parameter choice rule as in the deterministic setting with @xmath164 replaced by @xmath165 ( case a ) ) or @xmath204 where @xmath205 fulfills ( case b ) ) .
in particular , the series of regularized solutions fulfills @xmath254 the proof follows the lines of the one of theorem [ thm : lifting_convergence_kyfan ] with @xmath255 replaced by @xmath256 . also lemma [ lem : egoroff ]
is easily adjusted to incorporate multiple solutions .
so far we assumed that only the noise is stochastic whereas the operator @xmath212 and the unknown @xmath7 were assumed to be deterministic . in @xcite
general stochastic inverse problems @xmath257 were considered .
it was shown how deterministic conditions such as source conditions can be incorporated into the stochastic setting by assuming that the deterministic conditions hold with a certain probability . however , additional conditions may occur when lifting these in order to ensure the deterministic requirements up to a certain probability . since this is easier
seen given an example , we move the discussion of the complete stochastic formulation in the next section . although we will address only one particular example , the technique can be applied to general approaches .
due to the possible multiplicity of stochastic conditions which might appear in this context it seems not possible to develop a lifting strategy in such a general fashion as in the previous section .
we will therefore consider two classical examples , namely nonlinear tikhonov regularization and landweber s method for nonlinear inverse problems .
the theory is taken completely from @xcite .
we seek the solution of a nonlinear ill - posed problem via the variational problem @xmath258 with a reference point @xmath259 and given noisy data @xmath22 according to where the stochastic distribution of the noise is assumed to be known .
we shall skip the general convergence theorem ( which follows as in the previous section ) and move to convergence rates directly . in the deterministic theory ,
i.e. when @xmath260 is the noisy data with @xmath261 , we have the following theorem from @xcite .
[ thm : tikh_nonlin_det ] let @xmath262 be convex , @xmath263 such that @xmath261 and @xmath102 denote the @xmath264-minimum norm solution of .
furthermore let the following conditions hold .
* @xmath212 is frchet - differentiable * there exists @xmath265 such that @xmath266 in a sufficiently large ball @xmath267 * @xmath268 satisfies the source condition @xmath269 for some @xmath270 . *
the source element satisfies @xmath271 .
then for the choice @xmath272 with some fixed @xmath273 we obtain @xmath274 as given in theorem 4.6 of @xcite , the following stochastic formulation of theorem [ thm : tikh_nonlin_det ] holds .
[ thm : tikh_nonlin_stoch ] let @xmath262 be convex , let @xmath22 be such that @xmath275 and @xmath102 denote the @xmath264-minimum norm solution of for almost all @xmath69 . furthermore let the following conditions hold . * @xmath276 is frechet - differentiable for almost all @xmath69 * @xmath277 satisfies @xmath278 in a sufficiently large ball @xmath279 * ( smoothness ) @xmath280 where @xmath281 * ( closedness ) @xmath282 * ( decay ) @xmath283 . then for the choice
@xmath284 we obtain @xmath285 we have @xmath286 with probability @xmath287 . now fix @xmath288 and @xmath289 .
then with probability @xmath290 conditions d ) and e ) are fulfilled .
thus for the corresponding values of @xmath69 we can apply theorem [ thm : tikh_nonlin_det ] and obtain @xmath291 or , fixing the parameter @xmath284 , @xmath292 this estimate holds on a set with probability greater or equal @xmath293 .
the ky fan distance can therefore be bounded as @xmath294 this estimate is valid for arbitrary choices of @xmath236 and @xmath227 above , therefore we may bound the ky fan distance of @xmath102 and @xmath295 by taking the infimum with respect to @xmath236 and @xmath227 .
the core principle of the lifting strategy is to ensure that there exists a subset @xmath296 such that all deterministic assumptions hold with probability one on @xmath137 .
this may lead to the introduction of new conditions such as the decay condition in theorem [ thm : tikh_nonlin_stoch ] .
namely , since @xmath297 and @xmath298 may vary with @xmath69 , it may be possible that for a sequence @xmath299 @xmath300 and @xmath301 such that still for all @xmath302 @xmath303 . in this case
the parameter @xmath227 can not be treated as a constant in the convergence rate , but it influences it to a significant degree . the decay condition had to be imposed in order to control the growth of @xmath227 .
it is , however , possible to avoid condition e ) by imposing other conditions .
for example , one could require that @xmath297 is bounded below by some @xmath304 . in this case condition
d ) implies e ) . a more detailed discussion is given in @xcite .
accordingly , in order to lift other deterministic convergence rate results into the fully stochastic setting , a careful examination of the conditions necessary for convergence in the stochastic setting , understanding their cross - connections and dependencies is important .
however , once the conditions have been translated to the stochastic setting , convergence rates follow immediately using the ky fan metric .
we will close this example by showing how particular choices of the stochastic parameters in theorem [ thm : tikh_nonlin_stoch ] influence the convergence rate . to this end , we cite remark 4.8 of @xcite .
let in the first examples the operator be deterministic , i.e. , @xmath305 where @xmath306 .
first consider the case that @xmath307 $ ] , i.e. , it is uniformly distributed on the interval @xmath308 $ ] .
we therefore have @xmath309 , as well as @xmath310 for @xmath231 .
thus theorem [ thm : tikh_nonlin_stoch ] implies @xmath311 which gives for @xmath284 the optimal rate @xmath312 for the second case suppose that @xmath313 for some exponent @xmath314 . since now we do not have @xmath315 , but @xmath316 we obtain @xmath317 since the right hand side does not converge to zero we do not obtain a convergence rate anymore .
however , convergence itself still follows from theorem [ thm : lifting_convergence_kyfan ] . finally , consider the case when both d ) and e ) from theorem [ thm : tikh_nonlin_stoch ] influence the convergence behavior , because @xmath212 is stochastic with varying @xmath297 .
for instance in the case the for some @xmath318 $ ] we have @xmath319 and @xmath320 , we find that @xmath309 and @xmath321 are compatible realizations of @xmath322 and @xmath323
. with this one can show @xmath324 under the parameter choice @xmath325 .
from the given examples it is evident that the convergence speed is heavily influenced by the conditions d ) and e ) in theorem [ thm : tikh_nonlin_stoch ] .
therefore , although the general formula for the convergence rate may suggest that the convergence rate is close to the deterministic one , it may be significantly slower due to the additional stochastic properties . as before we seek the solution of a nonlinear ill - posed problem given noisy data @xmath22 according to where the stochastic distribution of the noise is assumed to be known .
landweber s method can be seen as a descent algorithm for @xmath326 and is defined via the iteration @xmath327 where @xmath328 is an appropriately chosen stepsize and @xmath329 an initial guess .
landweber s method constitutes a regularization method if it is stopped early enough @xcite . in the deterministic theory ,
i.e. when @xmath260 is the noisy data with @xmath261 , we have the following theorem from @xcite for convergence rates of the landweber method .
[ thm : lw_nonlin_det ] let @xmath262 be convex , @xmath263 such that @xmath261 and @xmath102 denote the @xmath264-minimum norm solution of
. assume has a solution in @xmath330 .
furthermore let the following conditions hold on @xmath331 .
* @xmath212 is frechet - differentiable with @xmath332 and @xmath333 * @xmath334 where the bounded linear operators @xmath335 satisfy @xmath336 * @xmath268 satisfies the source condition @xmath337 for some @xmath270 and @xmath338 .
let @xmath339 be sufficiently small . then
, if the regularization parameter is stopped according to the discrepancy principle , i.e. , at the unique index @xmath340 for which for the first time @xmath341 with @xmath342 , we obtain @xmath343 we can obtain a stochastic version of theorem [ thm : lw_nonlin_det ] in the same way and with the same techniques used to show that theorem [ thm : tikh_nonlin_stoch ] followed from theorem [ thm : tikh_nonlin_det ] .
[ thm : lw_nonlin_stoch ] let @xmath262 be convex , @xmath344 be given with @xmath165 and let @xmath182 denote the @xmath264-minimum norm solution of .
assume has a solution in @xmath345 for almost all @xmath69 .
furthermore let the following conditions hold on @xmath331 .
* @xmath346 where for almost all @xmath69 the set @xmath347 describes a family of bounded linear operators with @xmath348 * @xmath268 satisfies the source condition @xmath349 for some @xmath350 and @xmath338 . *
@xmath351 * @xmath352 then , if the regularization parameter is stopped according to the discrepancy principle , i.e. , at the unique index @xmath340 for which for the first time @xmath353 with @xmath354 , we obtain for @xmath355 sufficiently small the rate @xmath356 where the constant @xmath357 depends on @xmath358 only . in the fully stochastic
setting , the source condition b ) from theorem [ thm : lw_nonlin_stoch ] need not hold with constant exponent @xmath358 for all @xmath135 .
there are at least two situations which lead to the power @xmath358 being a stochastic quantity as well , i.e. , it holds @xmath359 with @xmath360 . in the first case all solutions @xmath182 come from some initial element @xmath361 , with small @xmath46-norm .
some randomly smoothing operator is acting on this element and generates @xmath182 .
( one could for instance think of some kind of evolution process , e.g. , a diffusion process that is applied to some initial value @xmath362 ) .
the smoothness of @xmath182 is therefore random .
secondly , @xmath102 may be a deterministic element satisfying a certain smoothness condition .
the data @xmath363 is generated by applying a forward operator @xmath364 with random smoothness properties .
if the realization of @xmath364 is strongly smoothing , this corresponds to a source condition with small @xmath365 , if @xmath364 is weakly smoothing we have the source condition with larger @xmath365 .
the following proposition shows the convergence rate that results from the source condition for the case that @xmath365 is uniformly distributed on the interval @xmath366 $ ] .
let all conditions of theorem [ thm : lw_nonlin_stoch ] hold except for b ) and d ) .
let @xmath182 satisfy where @xmath298 is uniformly bounded and sufficiently small .
let @xmath365 be uniformly distributed on the interval @xmath366 $ ] , i.e. , @xmath367 then the approximations @xmath368 obtained by landweber s method satisfy the convergence rate @xmath369 where @xmath370 denotes the lambert w - function , defined by @xmath371 , see @xcite .
as can be seen from the proof of theorem 3.1 in @xcite , the requirement `` @xmath339 sufficiently small '' , becomes stronger , the larger @xmath358 is .
supposing that @xmath339 in is sufficiently small for the case @xmath372 , implies therefore that also the convergence conditions for @xmath373 are satisfied .
secondly we observe that the convergence rate in theorem [ thm : lw_nonlin_stoch ] contains a constant @xmath357 that depends on @xmath358 .
although it is difficult to state an explicit formula for @xmath357 , investigation of @xcite shows , that @xmath374 attains its maximum value when @xmath372 .
after these observations we start with the actual derivation of the convergence rate . for the sake of simplicity
we assume that all appearing constants are just equal to 1 .
furthermore we may assume that @xmath322 and @xmath323 both vanish .
asymptotically , for given @xmath69 we therefore have the estimate @xmath375 measuring the distance in the ky fan metric we must , since we assumed that @xmath365 is as in , solve the equation @xmath376 for @xmath358 .
we first consider the simplified equation @xmath377 which is solved by @xmath378 in the following we show that the above approximate solution is sufficiently accurate .
therefore we construct a better estimate via the ansatz @xmath379 .
the original equation then contains the term @xmath380 . neglecting the quadratic part
, we can replace this term with @xmath381 , and obtain an equation that mathematica can solve for @xmath382 .
the solution for the correction term is given as @xmath383 and tends to zero approximately linearly in @xmath384 .
thus this correction becomes small rather quickly , and we can consider the asymptotic bound in as sufficiently accurate due to the asymptotics of the lambert w - function .
we will now apply the theory developed in the previous section to selected deterministic regularization methods .
let @xmath385 be a linear compact operator between hilbert spaces @xmath386 and @xmath46 with singular system @xmath387 , see e.g. @xcite .
then , for @xmath388 , the generalized inverse @xmath389 to @xmath385 is given by @xmath390 since for compact operators the singular values approach zero , their inverse blows up and the generalized inverse yields a meaningless solution to for noisy data .
a popular class of regularization methods is based on the filtering of the generalized inverse . introducing an appropriate filter function @xmath391 depending on the regularization parameter @xmath16 that controls the growth of @xmath392 ,
the regularized solutions are defined by @xmath393 examples for filter based methods are for example the classical tikhonov regularization , truncated singular value decomposition or landwebers method @xcite .
the regularization properties are fully determined by the filter functions . in the deterministic setting ,
the conditions can be found in , e.g. , ( * ? ? ?
* theorem 3.3.3 . ) .
convergence rates can be obtained for a priori and a posteriori parameter choice rules under stricter conditions on the filter functions .
we will only comment on an a priori choice here in order to keep this section short .
an example of the discrepancy principle as a posteriori parameter choice is given in the next section in a different context .
using the smoothness condition @xmath394 the following theorem can be obtained .
* theorem 3.4.3)[thm : louis_filter_optimality ] let @xmath395 and @xmath396 .
assume that it holds @xmath397 and for @xmath398 , @xmath399 where @xmath400 and @xmath401 are constants independent of @xmath164 .
then with the a priori parameter choice @xmath402 the method induced by the filter @xmath403 is order optimal for all @xmath404 , i.e. , @xmath405 for some constant @xmath406 independent of @xmath164 and @xmath407 .
now we use theorem [ thm : lifting_rates_kyfan ] and obtain convergence rates in the ky fan metric .
let @xmath395 and @xmath165 be known .
assume that it holds @xmath397 and for @xmath398 , and hold .
then with the a priori parameter choice @xmath408 the method induced by the filter @xmath403 fulfills @xmath409 for some constant @xmath406 independent of @xmath164 and @xmath407 .
more about filter methods in the stochastic setting including numerical examples can be found in @xcite .
we consider an autoconvolution equation @xmath410(s)=\int_0^s x(s - t)x(t)\ , dt,\qquad 0\leq s\leq 1\ ] ] between hilbert spaces @xmath411 $ ] and @xmath412 $ ] where @xmath413 .
such an equation is of great interest in , for example , stochastics or spectroscopy and has been analyzed in detail in @xcite .
recently , a more complicated autoconvolution problem has emerged from a novel method to characterize ultra - short laser pulses @xcite . here
, we want to show the transition from the deterministic setting to the stochastic setting in a numerical example .
we base our results on the deterministic paper @xcite . using the haar - wavelet basis , the authors of @xcite reformulate as an equation from @xmath414 to @xmath414 by switching to the space of coefficients in the haar basis . in order to stabilize the inversion , an @xmath415 penalty term
is used such that the task is to minimize the functional @xmath416 the regularization parameter @xmath16 in is chosen according to the discrepancy principle . in @xcite ,
the following formulation is used : for @xmath417 choose @xmath129 such that @xmath418 holds .
the authors show that this leads to a convergence of the regularized solutions against a solution of with minimal @xmath415-norm of its coefficients .
it was also shown that the regularization parameter fulfills @xmath419 by courtesy of stephan anzengruber we were allowed to use the original code for the numerical simulation in @xcite .
we only changed the parts directly connected to the data noise .
namely , we replaced the deterministic error @xmath420 with i.i.d gaussian noise , @xmath421 @xmath422 .
the discretization is due to the truncation of the expansion of the functions in the haar - basis after @xmath423 elements .
the parameter choice was realized with @xmath164 replaced by @xmath424 in accordance with theorem [ thm : lifting_convergence_kyfan ] . instead of the correct expectation @xmath425
see @xcite , we used the upper bound @xmath426 since , as shown in this chapter , the expectation has to be `` blown up '' anyway . in a first experiment we let @xmath427 . in this case , the numerical results suggest that the regularization parameter decreases too fast , i.e. , @xmath428 does not converge to zero as the requirement in states ; see figure [ fig : conv_nonlin ] . for comparison ,
in a second run we chose @xmath429 where @xmath423 is the amount of data points .
this way , @xmath430 .
now @xmath428 converges to zero as it should be the case from [ eq : alpha_props ] , see figure [ fig : conv_nonlin ] . at this point
we would like to mention that the discrepancy principle using the ky fan distance and the deterministic one are not completely equivalent since a different way of measuring the noise is used .
typically the stochastic noise level will be smaller ( it need to bound 100% of the possible realizations ) and the iteration will be stopped later than in the deterministic setup . .
left : constant value of @xmath227 in the discrepancy principle with the expectation of the noise leads to the regularization parameter decreasing too fast .
right : increasing @xmath227 appropriately with decreasing variance resolves this issue.,title="fig : " ] . left : constant value of @xmath227 in the discrepancy principle with the expectation of the noise leads to the regularization parameter decreasing too fast .
right : increasing @xmath227 appropriately with decreasing variance resolves this issue.,title="fig : " ] in @xcite the lifting strategy was used in a slightly different way . in particular , the ky fan metric was used to obtain a novel parameter choice rule .
the convergence rates obtained there , however , can also be viewed in the framework of this work .
the scope of that paper was to transfer the deterministic convergence results from @xcite into the stochastic setting .
the seminal paper @xcite initiated the investigation of sparsity - promoting regularization for inverse problems .
looking for the solution of the linear ill - posed problem @xmath431 between hilbert spaces @xmath386 and @xmath46 with given noisy data @xmath432 , the regularization strategy was to obtain an approximation @xmath125 to @xmath102 via @xmath433 where @xmath434 is an appropriate index set , @xmath435 @xmath436 , @xmath437 a dictionary ( typically an orthonormal basis or frame ) in @xmath386 and @xmath438 .
choosing a sufficiently smooth wavelet basis for @xmath437 and setting @xmath439 with @xmath440 , the penalty term in corresponds to a norm in the besov space @xmath441 . formulating the problem of determining @xmath7 from noisy data @xmath442 , @xmath443 , in the bayesian setting with the distributions @xmath444 and @xmath445 and using the maximum a - posteriori solution lead to the formulation @xmath446 where @xmath18 is the variance of the noise and @xmath447
can roughly be described as the inverse variance of the prior .
the product @xmath448 gives the actual regularization parameter . in direct application of theorem [ thm : lifting_convergence_kyfan ] ,
the deterministic condition @xmath449 with @xmath164 replaced by @xmath165 from translates to the conditions @xmath450 leading to convergence of @xmath451 to the unique ( in case @xmath88 the operator is assumed to be injective ) solution @xmath102 of minimal norm @xmath452 in the ky fan metric .
the proof of convergence rates is based on two assumption : @xmath453 where @xmath454 and @xmath455 for some @xmath456 .
combining proposition 4.5 , proposition 4.6 , proposition 4.7 from @xcite it is @xmath457 where with @xmath458 , @xmath459 and @xmath460 we know that the deterministic rate is an upper bound to the reconstruction error whenever @xmath461 and @xmath462 .
hence , it is @xmath463 where @xmath464 and @xmath465 where the besov - space functions were truncated after the first @xmath466 basis functions . by definition of the ky fan
metric , it follows immediately from that @xmath467 since @xmath447 is a free parameter , we can balance the terms in , i.e. solve the nonlinear equation @xmath468 for @xmath447 . with this parameter choice rule one obtains by construction @xmath469 we can also apply the theory developed in this work to this problem . in the deterministic setting ,
see @xcite , it was proposed to chose the regularization parameter @xmath470
. combining ( * ? ? ?
* proposition 4.5 ) and ( * ? ? ?
* proposition 4.7 ) then yields the rate @xmath471 with @xmath472 from and some @xmath95 . theorem [ thm : lifting_rates_kyfan ] then yields in the stochastic setting the parameter choice @xmath473 and @xmath474 in the notation of it is for gaussian noise @xmath475 @xmath476 see proposition .
since @xmath477 , compare and , the convergence rate in is slightly slower than the one in , but they share the same order of convergence .
our goal was to demonstrate how convergence results for inverse problems in the deterministic setting can be carried over into the stochastic setting . using the ky fan metric , we have shown that , when only the noise is assumed to be stochastic whereas the other quantities are deterministic , this is is possible in a straight - forward way .
namely , assuming the knowledge of an estimate of @xmath165 , the convergence results and parameter choice follows from the deterministic setting by replacing @xmath164 , which originates from the basic deterministic assumption @xmath478 , with @xmath479 . we have shown that , under some slight modifications , it is possible to use the expectation as measure for the magnitude of the noise . in a fully stochastic situation , where additionally to the noise other objects might be of stochastic nature , the lifting of deterministic convergence results is possible as well .
however , careful analysis is necessary in order to carry the deterministic conditions over into the stochastic setting . | both for the theoretical and practical treatment of inverse problems , the modeling of the noise is a crucial part .
one either models the measurement via a deterministic worst - case error assumption or assumes a certain stochastic behavior of the noise . although some connections between both models are known , the communities develop rather independently . in this paper
we seek to bridge the gap between the deterministic and the stochastic approach and show convergence and convergence rates for inverse problems with stochastic noise by lifting the theory established in the deterministic setting into the stochastic one .
this opens the wide field of deterministic regularization methods for stochastic problems without having to do an individual stochastic analysis for each problem . in inverse problems ,
the model of the inevitable data noise is of utmost importance . in most cases , an additive noise model @xmath0
is assumed . in
, @xmath1 is the true data of the unknown @xmath2 under the action of the ( in general ) nonlinear operator @xmath3 , @xmath4 and @xmath5 in corresponds to the noise .
the spaces @xmath6 are assumed to be banach- or hilbert spaces . when speaking of inverse problems , we assume that is ill - posed .
in particular this means that solving for @xmath7 with noisy data is unstable in the sense that `` small '' errors in the data may lead to arbitrarily large errors in the solution .
hence , is not a sufficient description of the noise .
more information is needed in order to compute solutions from the data in a stable way . in the _ deterministic _ setting , one assumes @xmath8 for some @xmath9 where @xmath10 is an appropriate distance functional .
typically , @xmath11 is induced by a norm such that reads @xmath12 . here and further on we use the superscript @xmath13 to indicate the deterministic setting .
solutions of under the assumption , are often computed via a tikhonov - type variational approach @xmath14 where again @xmath11 is a distance function and @xmath15 is the penalty term used to stabilize the problem and to incorporate a - priori knowledge into the solution . the regularization parameter @xmath16 is used to balance between data misfit and the penalty and has to be chosen appropriately .
the literature in the deterministic setting is rich , at this point we only refer to the monographs @xcite for an overview .
the deterministic worst - case error stands in contrast to _
stochastic _ noise models where a certain distribution of the noise @xmath5 in is assumed .
we shall indicate the stochastic setting by the superscript @xmath17 . in this paper
, @xmath18 will be the parameter controlling the variance of the noise .
depending on the actual distribution of @xmath5 , @xmath19 may be arbitrarily large , but with low probability .
a very popular approach to find a solution of is the bayesian method . for more detailed information
, we refer to @xcite . in the bayesian
setting , the solution of the inverse problem is given as a distribution of the random variable of interest , the _ posterior distribution _
@xmath20 , determined by bayes formula @xmath21 that is , roughly spoken , all values @xmath7 are assigned a probability of being a solution to given the noisy data @xmath22 . in , the _ likelihood function _
@xmath23 represents the model for the measurement noise whereas the _ prior distribution _
@xmath24 represents a - priori information about the unknown .
the data distribution @xmath25 as well as the normalization constants are usually neglected since they only influence the normalization of the posterior distribution . in practice one is often more interested in finding a single representation as solution instead of the distribution itself .
popular point estimates are the _
conditional expectation _ ( conditional mean , cm ) @xmath26 and the _ maximum a - posteriori ( map ) _
solution @xmath27 i.e. , the most likely value for @xmath7 under the prior distribution given the data @xmath22 .
both point estimators are widely used .
the computation of the cm - solution is often slow since it requires repeated sampling of stochastic quantities and the evaluation of high - dimensional integrals .
the map - solution , however , essentially leads to a tikhonov - type problem .
namely , assuming @xmath28 and @xmath29 , one has @xmath30 analogously to . also non - bayesian approaches for inverse problems often seek to minimize a functional , see e.g. @xcite or use techniques known from deterministic theory such as filter methods @xcite .
finally , inverse problems appear in the context of statistics .
hence , the statistics community has developed methods to solve , partly again based on the minimization of .
we refer to @xcite for an overview . in summary ,
tikhonov - type functionals and other deterministic methods frequently appear also in the stochastic setting . from a practical point of view
, one would expect to be able to use deterministic regularization methods for even when the noise is stochastic .
indeed , the main question for the actual computation of the solution , given a particular sample of noisy data @xmath22 , is the choice of the regularization parameter .
a second question , mostly coming from the deterministic point of view , is the one of convergence of the solutions when the noise approaches zero . in the stochastic setting these questions
are answered often by a full stochastic analysis of the problem . in this paper
we present a framework that allows to find appropriate regularization parameters , prove convergence of regularization methods and find convergence rates for inverse problems with a stochastic noise model by directly using existing results from the deterministic theory .
the paper takes several ideas from the dissertation @xcite , which is only publicly available as book @xcite and not published elsewhere .
it is organized as follows . in section [ ssec : noisemodel ]
we discuss an issue occurring in the transition from deterministic to stochastic noise for infinite dimensional problems .
the ky fan metric , which will be the main ingredient of our analysis , and its relation to the expectation will be introduced in section [ ssec : kyfan ] .
we present our framework to lift convergence results from the deterministic setting into the stochastic setting in section [ ssec : conv_stoch ] .
examples for the lifting strategy are given in section [ sec : examples ] . |
the understanding of the bfkl pomeron has been demanding a considerable theoretical effort . its behavior in perturbative qcd
is determined by generating the integral equation @xcite .
that procedure consists of summing the leading logarithms on energy , @xmath4 , order by order from perturbation theory and the main result is that the @xmath5 for the @xmath6-exchange process is a power of @xmath7 . a priori
, bfkl is itself asymptotic and we may ask if at finite energies , i.e. non asymptotic regime , summing a finite number of terms from the bfkl series could describe the existent data . for
that we should use a truncated bfkl series , performing a finite sum of gluon ladders ( bearing in mind reggeized gluons and considering effective vertices ) .
the question that remains is how many orders should one take into account .
the lowest order two gluons exchange calculation leads to a @xmath5 constant on @xmath7 and the next contribution to the sum is the one rung gluon ladder , which provides a logarithmic growth . in order to avoid unitarity violation and
by simplicity we truncate our summation at this order . as a result a successfull fit to the @xmath0 with these two contributions is obtained and presented in sec .
2 . these results motivate to check the non - forward amplitude in order to obtain the prediction for the @xmath2 , which gives the behavior on the momentum transfer @xmath3 . in the bfkl framework
such analysis is dependent of the proton impact factor ( if ) input , which introduces certain uncertainty due to the presence of non - perturbative content .
the ifs determine the coupling of the pomeron to the color singlet hadrons and play a crucial role in the calculation of the non - forward amplitude .
we calculate the @xmath8 elastic amplitude at @xmath1 taking into account two distict anstze to the proton if : the dirac form factor , as proposed recently by balitsky and kuchina @xcite and the usual non - perturbative ansatz @xcite .
the main resulting features are discussed in sec . 3 , although considering that a more realistic ansatz to the proton if is still to be found .
in addition the @xmath2 is calculated in the small @xmath3 approximation and compared with the experimental data . in the last section we present our conclusions .
in the leading logarithm approximation ( lla ) , the pomeron is obtained considering the color singlet ladder diagrams whose vertical lines are reggeized gluons coupled to the rungs through the effective vertices .
the correspondent amplitude is purely imaginary and the coupling constant @xmath9 is considered frozen in some transverse momentum scale @xcite . for the elastic scattering of a hadron , the mellin transform of the scattering amplitude is given by @xcite : @xmath10 where the @xmath11 is the color factor for the color singlet exchange , @xmath12 and @xmath13 are the transverse momenta of the exchanged gluons in the @xmath3-channel and @xmath14 is the momentum transfer , with @xmath15 .
the function @xmath16 is the mellin transform of the bfkl kernel @xmath17 , which states the dynamics of the process and is obtained in perturbative qcd .
the main properties of the lo kernel are well known @xcite and the results arising from the nlo calculations have yielded intense debate in the literature recently @xcite . in the case of @xmath18(@xmath19 ) scattering , the factor @xmath20 is the proton if , which in the absence of a perturbative scale has a non - perturbative feature and furnishes the @xmath6-proton coupling .
this factor turns the amplitude infrared safe when the colliding particles are colorless . in the leading order of perturbation theory we have @xmath21 and in the next order @xmath22 \ , .
\end{aligned}\ ] ] here , @xmath23 , where @xmath24 is the color number and @xmath9 is the strong coupling constant . in order to perform a reliable calculation the convenient proton if should be introduced .
this is not an easy task , namely these hadronic processes are soft and there is no hard scale allowing to use perturbation theory .
in fact , we should know in details the parton wavefunction in the hadron to calculate the ifs properly . since this is not available ,
several models are proposed in order to calculate them . from the optical theorem ,
the lowest order contribution to the @xmath5 is a constant term , and the next order term is a logarithm of the energy , scaled by a typical gluon transverse momentum of the process ( arbitrary ) .
when considering @xmath25 there is no need to deal with both a specific form for the if and the transverse momentum integration .
this allows to consider @xmath7-independent factors in each term as free parameters and to obtain them from data .
the correct description at low energy requires the reggeon contribution , which is parameterized from regge theory .
our expression to the total cross section is then , @xmath26 hence we fix the constants @xmath27 and @xmath28 from data on @xmath19 , imposing the same contribution for both @xmath18 and @xmath19 .
this procedure is reasonable due to the higher energies reached on @xmath19 collision , where the pomeron dominates . on the other hand ,
@xmath18 data are predominantly at low energy , which is not strongly sensitive to the pomeron model , thus dominated by the reggeonic contribution .
a successful description of data is obtained for the whole range of energy .
the result is shown in the fig .
the parameters and a more detailed discussion can be found in ref .
@xcite .
the hypothesis of considering two orders from the bfkl series ( @xmath29 ) is phenomenologically corroborated by the well known dispersion relation fit @xcite .
an important additional advantage is that the total cross section obtained is consistent with the unitarity constraint , avoiding unitarization procedures . as a final remark , at the lhc energy ( @xmath30 tev ) the extrapolation of our results
will give @xmath31 .
in order to calculate the elastic amplitude at @xmath1 , information about the coupling between the proton and the @xmath3-channel reggeized gluons in the ladder is required .
namely , introducing a reliable proton if that has to be modeled since it can not be calculated from first principles due to the unknowledge on the wavefunction of the hadronic constituent partons . here
are analized two distinct models for the if : balitsky and kuchina proposed recently @xcite that at large momentum transfer the coupling of the bfkl pomeron to the nucleon is essentially equal to the dirac form factor of the nucleon .
their basic idea is that in the lowest order in perturbation theory there is no difference between the diagrams for the nucleon if and similar diagrams with two gluons replaced by two photons , in such a way that the amplitudes can be calculated without any model assumption .
this if , @xmath32 , is decoupled in the transverse momentum integration and presents an explicit dependence on @xmath3 , being similar to the usual pomeron - proton coupling used in regge phenomenology .
the expression is @xmath33 the choice for this proton if is useful when one analyzes near forward observables , for instance the elastic differential cross section .
however it does not play the role of a regulator of infrared divergences at @xmath8 process because clearly it does not vanish when the gluon transverse momenta goes to zero . in electron - proton process
the situation is different since the photon impact factor supplies that condition @xcite .
then the next step is to perform the gluon transverse momenta integrations .
in fact , such integrals are infrared divergent and should be regularized .
an usual way out is to introduce an infrared cut - off @xmath34 , temporally defining a small gluon mass , avoiding problems at the infrared region .
this procedure is quite similar as to take into account a non - perturbative massive gluon propagator ( see i.e. ref .
@xcite ) . the lowest order ( order @xmath35 )
contribution , using eqs .
( 1 - 2 ) , gives the following result : @xmath36 ^ 2 \frac{\pi}{(|t|-\lambda^2)}\,\ln \left(\frac{\lambda^2}{|t|}\right ) \ , .
\nonumber \end{aligned}\ ] ] the one rung gluon ladder has two components ( order @xmath37 ) , given by the following expression : @xmath38 ^ 2 \,s\ , \ln\left(\frac{s}{{\bf k^2}}\right)\,(i_1 + i_2)\ , , \nonumber \end{aligned}\ ] ] with @xmath39 corresponding to the one rung gluon ladder and @xmath40 correspondent to the three gluons exchange graphs , whose order is also @xmath41 .
such structure is due to the fact that in the color singlet calculation there is no cancellation between graphs and one can not obtain an expression for the two - loop level which is proportional to the one loop amplitude @xcite .
we define @xmath40 through symmetry on the integration variables @xmath12 and @xmath13 ( see eqs .
( 1,3 ) ) and the factor @xmath42 collects the correspondent color factors and the remaining constants .
the explicit calculation of those integrals , yields @xmath43 @xmath44 some comments about the amplitude above are in order .
the scale of the factor @xmath34 should be at non - perturbative regime , i.e. @xmath45 gev@xmath46 . an interesting aspect is the behavior of the amplitude at the forward limit @xmath25 , where it becames very large .
this limit is a well known property of perturbative qcd calculations and there are several reasons to believe that the point @xmath25 plays a very special role , such that perturbation theory may even not be applicable . for the full bfkl series in the forward region there is still the diffusion on tranverse momenta , i.e. on @xmath47 , which extends into both the ultraviolet and the infrared regions @xcite .
nevertheless , the momentum scale @xmath3 supplies the control condition .
however , we will suppose that a smooth transition from a finite @xmath3 down to @xmath25 is possible and that the finite bfkl series gives the correct behavior on energy for the forward observables .
later we make use of this hypothesis to obtain the logarithmic slope @xmath48 and the differential elastic cross section . using quite general properties of the ifs ,
namely they vanish as transverse momenta go to zero , one can guess their behavior which is determined by the large scale nucleon dynamics .
regardless its exact shape , in general the proton if takes the form @xcite : @xmath49 where @xmath50 is a scale which is typical of the non - perturbative dynamics . as a consequence of this choice
the momentum transfer behavior is completely determined by the kernel .
the amplitude now reads : @xmath51 \nonumber \\ + \ , a^{\prime}\,\ln\left(\frac{s}{{\bf k}^2}\right)\,\left [ \frac{\ln ( \mu^2)}{(|t|- \mu^2 ) } + \frac{\ln(\mu^2)|t|}{(|t|-\mu^2)^2}\ln \left(\frac{\mu^2}{|t|}\right)\right ] \,.\nonumber \end{aligned}\ ] ] here we use the definition @xmath52 .
we observe again a divergent behavior at @xmath25 , nevertheless we claim that the forward amplitude is finite in this point and the dependence on energy is correctly described . despite obtaining an analytic expression to the elastic scattering amplitude ,
a direct comparison with the experimental data is known not to be reliable .
in fact , data on differential cross section at low @xmath3 are parameterized in the form @xmath53 , where @xmath54 is the forward slope @xcite . therefore
, we can obtain an expression for the differential cross section at small @xmath3 , using our previous results .
the usual relation to describe the cross section is : @xmath55_{t=0}\ , .
\end{aligned}\ ] ] in the regge framework the slope is obtained from the powerlike behavior of the scattering amplitude , dependent of the effective slope of the pomeron trajectory @xmath56 , namely @xmath57 .
the parameter @xmath58 comes from the slope of the @xmath59-@xmath59-@xmath6 vertex . in our case
we should obtain the slope from the non forward elastic scattering amplitudes @xmath60 obtained above . for the amplitude obtained employing the balitsky and kuchina impact factor it results the following slope @xmath61 where the first term does not contribute effectively at @xmath25 and we are left only with the second term . from simple inspection of the amplitude obtained with the usual impact factor ( see eq . ( 5 ) )
we also verify that one gets a similar expression to the correspondent slope . considering the specific form for the @xmath3-derivative of the amplitudes , their asymptotic values at @xmath25 depend only on the energy .
in fact , they take the form @xmath62 , where @xmath63 and @xmath64 are @xmath7-independent parameters . for our case , the amplitude is purely imaginary , then @xmath65 and @xmath66 .
putting all together , the corresponding slope is @xmath67\,\,,\end{aligned}\ ] ] and the elastic differential cross section is given by eq .
( 6 ) , where again @xmath68 . in order to obtain the parameters @xmath63 and @xmath64
, we use the slope experimental values for both low ( cern - isr ) and high energy ( cern - sps , tevatron ) points from @xmath19 reaction ( @xmath69gev@xmath70 gev ) @xcite . having the slope obtained from data ,
the elastic differential cross section is straighforwardly determined and a successful comparison with its experimental measurements at @xmath71 gev is shown in the fig .
( 2 ) . in summary
, we study the contribution of a truncated bfkl series to the hadronic process , specifically the @xmath8 collisions , considering two orders in perturbation theory corresponding to the bare two gluons exchange and the one rung gluon ladder . despite the restrictions imposed by the use of a perturbative approach for soft observables , a good description of the total cross sections
was obtained motivating an analysis of the elastic differential cross section .
although the qcd perturbation theory is in principle not reliable at the forward direction ( @xmath25 ) , nevertheless we suppose that perturbation theory gives the behavior on energy even in this region . the next step is to consider @xmath3 different from zero , where the momentum transfer furnishes a scale to perform suitable calculations . in order to proceed this
, we calculate the non forward amplitude introducing two distinct anstze for the proton impact factor , namely a factorizable @xmath3-dependent proposed recently by balitsky and kuchina and the usual non - perturbative impact factor . in order to describe data we used a small momentum transfer approximation and obtained an expression to the elastic slope @xmath72 and the correspondent parameters .
the elastic differential cross section is obtained straightforwardly , describing with good agreement the experimental data at both low and high energy values .
mbdg thanks the organizers of this lively meeting and the useful discussions with p. landshoff , a. kaidalov , e. levin , v. petrov and m. ryskin
. 9 e.a .
kuraev , l.n .
lipatov and v.s .
fadin , _ phys . lett _ * b60 * ( 1975 ) 50 ; _ sov
* 44 * ( 1976 ) 443 ; _ sov .
* 45 * ( 1977 ) 199 ; ya .
balitsky and l.n .
lipatov , _ sov .
_ * 28 * ( 1978 ) 822 . | a truncated bfkl series is studied and applied to hadronic processes . the @xmath0
are described with good agreement with data and in a way consistent with the unitarity bound .
the elastic scattering amplitude is calculated at @xmath1 , introducing two distinct anstze for the proton impact factor .
the @xmath2 is obtained at small @xmath3 approximation and compared with the data . |
the perturbative approach to quantum chromodynamics ( qcd ) is believed to solve all problems within its own limitations provided the correct choices of the expansion variable and the variable(s ) to be fixed are used . however , perturbative qcd ( pqcd ) currently has the status of being a large collection of seemingly independent approaches , since a single unified approach valid for all processes is not known .
this is problematic when one wants to use a range , qualitatively speaking , of different processes to constrain the same parameters , _
e.g. _ in global fits .
what is needed is a single formalism valid over the union of all ranges that the various pqcd approaches allow .
this unification must be consistent , _
i.e. _ it must agree with each approach in the set , when the expansion of that approach is used up to the order being considered . in this presentation , we review recent progress in establishing a generalized formalism for the evolution of fragmentation functions ( ffs ) from the smallest @xmath0 values probed so far way up to unity @xcite . after explaining how to treat soft - gluon and hadron - mass effects in secs . [
sec : two ] and [ sec : three ] , respectively , we study their phenomenological implications in sec .
[ sec : four ] .
our conclusions are contained in sec .
[ sec : five ] .
the current optimum description of single - hadron inclusive production is provided by the qcd parton model , which requires knowledge of the ffs @xmath1 , each corresponding at leading order ( lo ) to the probability for the parton @xmath2 produced at short distance @xmath3 to form a jet that includes the hadron @xmath4 carrying a fraction @xmath0 of the longitudinal momentum of @xmath2 . from now on we omit the label @xmath4 and put the ffs for all @xmath2 into the vector @xmath5 .
the evolution of the ffs in the factorization scale @xmath6 is also required , which at large and intermediate @xmath0 is well described @xcite by the dokshitzer - gribov - lipatov - altarelli - parisi ( dglap ) equation @xcite , given at lo by @xmath7 where @xmath8 are the lo splitting functions calculated from fixed - order ( fo ) pqcd .
we define @xmath9 , which at lo obeys @xmath10@xmath11 $ ] , where @xmath12 is the first coefficient of the beta function and @xmath13 is the asymptotic scale parameter of qcd . for the color gauge group su(3 ) ,
the color factors appearing in this presentation are @xmath14 , @xmath15 , and @xmath16 ; @xmath17 is the number of active quark flavors . on the other hand , at small @xmath0 , the fo approximation fails due to unresummed _ soft - gluon logarithms _ ( sgls ) .
the most singular sgls are the _
double logarithms _ ( dls ) , which give a @xmath18 singularity in the lo splitting function @xmath19 of the form @xmath20 .
these dls occur at all orders in the fo splitting function , being generally of the form @xmath21 for @xmath22 , implying that , as @xmath0 decreases , eq .
( [ dglapx ] ) becomes a poor approximation once @xmath23 .
an improvement of the small-@xmath0 description should be obtained by accounting for all dls , which is provided by the double logarithmic approximation ( dla ) @xcite , @xmath24 , \label{dlax}\end{aligned}\ ] ] where @xmath25 for valence - quark or non - singlet ffs , while @xmath26 when @xmath27 , where @xmath28 is the singlet ff . the modified leading logarithm approximation ( mlla ) @xcite improves the description over the dla only by including a part of the fo contribution that is known to be important at small @xmath0 . with certain qualifications @xcite
, the mlla leads to a good description of all data down to the smallest @xmath0 values .
however , what has been lacking is a single approach which can describe data from the largest to the smallest @xmath0 values .
such an approach can be simply and consistently constructed from the knowledge provided by the dla and the dglap equation .
the result reproduces the fo result when expanded in @xmath29 , while all small-@xmath30 singularities are resummed into a non - singular expression in mellin space ( see later ) . as shown in detail in ref .
@xcite , this approach is unique up to higher - order terms in the fo series and up to less singular sgls , _
e.g. _ the _ single logarithms _ ( sls ) .
such a formalism may be constructed simply by evolving according to eq .
( [ dglapx ] ) , but with the splitting function modified as @xmath31 where @xmath32 contains the complete resummed contribution to the splitting function to all orders from the dls , while @xmath33 is the remaining fo contribution at lo .
it is obtained by subtracting the lo dls , already accounted for in @xmath34 , from @xmath19 to prevent double counting .
we work to lo , since the less singular sgls that occur at next - to - leading order ( nlo ) are not known to all orders .
however , we leave the sl at lo unresummed , since it is not singular as @xmath18 .
we now use eq .
( [ dlax ] ) to gain some understanding of @xmath34 . for this
, we transform to mellin space , defined by @xmath35 upon mellin transformation , eq .
( [ dlax ] ) becomes @xmath36 substituting eq .
( [ replace ] ) with the fo term @xmath33 neglected in eq .
( [ dglapx ] ) , taking the mellin transform , @xmath37 and inserting this in eq .
( [ drappre ] ) , we obtain @xmath38 we choose the solution @xmath39 since its expansion in @xmath29 yields at lo the result @xmath40 which agrees with the lo dls from the literature @xcite .
equation ( [ dlresummedinp ] ) contains all terms in the splitting function of the form @xmath41 , being the dls in mellin space , and agrees with the results of refs .
we now return to @xmath0 space , where eq .
( [ dlresummedinp ] ) reads @xmath42 with @xmath43 being the bessel function of the first kind . to summarize our approach , we evolve the ffs according to eq .
( [ dglapx ] ) , but with the replacement of eq .
( [ replace ] ) , where @xmath32 is given by eq .
( [ alldlinzindelpclosed ] ) , and @xmath33 is given by @xmath44 after the terms proportional to @xmath20 have been subtracted .
to extend nlo calculations to small @xmath45 , the complete resummed dl contribution given by eq .
( [ alldlinzindelpclosed ] ) must be added to the nlo splitting functions .
these contain sgls belonging to the classes @xmath46 , which must be subtracted .
note that the nlo @xmath47 term is accounted for by the resummed dl contribution .
the @xmath48 term is a type @xmath49 term , and hence does not need to be subtracted . before we outline the phenomenological investigation of our approach , we note that it is more complete than the mlla , which can be shown as follows . with @xmath50
accounted for , eq .
( [ drappre ] ) is modified to @xmath51 , \label{drap}\end{aligned}\ ] ] up to terms which are being neglected in this presentation and which are neglected in the mlla .
if we approximate @xmath52 by its sls , defined at lo to be the coefficients of @xmath53 , equal to those in @xmath54 , @xmath55 and apply the approximate result that follows from the dla at large @xmath56 , @xmath57 which can be derived , _
e.g. _ , from eq .
( [ alldlinzindelpclosed ] ) , the gluon component of eq .
( [ drap ] ) becomes precisely the mlla differential equation .
therefore , we conclude that , since we do not use these two approximations , our approach is more complete and accurate than the mlla .
we now incorporate hadron mass effects into our calculations , using a specific choice of scaling variable . for this purpose ,
it is helpful to work with light cone coordinates , in which any four - vector @xmath58 is written in the form @xmath59 with @xmath60 and @xmath61 . in the center - of - mass ( cm ) frame , the momentum of the electroweak boson takes the form @xmath62 in the absence of hadron mass , @xmath45 ( whose definition @xmath63 applies only in the cm frame ) is identical to the light cone scaling variable @xmath64
. however , the definition @xmath63 applies only in the cm frame .
so , @xmath65 is a more convenient scaling variable for studying hadron mass effects , since it is invariant with respect to boosts along the direction of the hadron s spatial momentum .
taking this direction to be the three - axis , and introducing a mass @xmath66 for the hadron , the momentum of the hadron in the cm frame reads @xmath67 therefore , the relation between the two scaling variables in the presence of hadron mass is @xmath68 note that these two variables are approximately equal when @xmath69 , _
i.e. _ hadron mass effects can not be neglected when @xmath45 ( or @xmath65 ) are too small . in the leading - twist component of the cross section after factorization , the hadron @xmath4
is produced by fragmentation from a real , massless parton of momentum @xmath70 the @xmath71 component of everything other than this parton and of everything produced by the parton other than the observed hadron @xmath4 must be positive , implying that @xmath72 and @xmath73 , respectively . as a generalization of the massless case
, we assume the cross section we have been calculating is @xmath74 , _
@xmath75 which is related to the measured observable @xmath76 via @xmath77}\ , \frac{d\sigma}{d\eta}(\eta(x_p),s).\ ] ] note that the effect of hadron mass is to reduce the size of the cross section at small @xmath45 ( or @xmath65 ) .
the data we are studying are described by the sum of the production cross sections for each light charged hadron species , being the charged pion , the charged kaon , and the ( anti)proton , whose masses ( 140 , 494 , and 938 mev , respectively ) are substantially different .
therefore , separate ffs and hadron masses for each of the three species are needed .
clearly , so many free parameters would not be constrained by these data . however , since most of the produced particles are pions , it is reasonable to take the final - state hadrons to have the same mass , so that in our approach it is reasonable to use a single hadron mass parameter @xmath66 , whose fitted result should be closer to the pion mass than the proton mass , _
i.e. _ around 300 mev .
any significant deviation from this value would imply that some physics has not been accounted for .
in a first test of our approach , we compare its effects on fits of quark and gluon ffs to @xmath78 data with the standard fo dglap evolution .
we use normalized differential - cross - section data for the process @xmath79 , where @xmath4 is a light charged hadron and @xmath80 is anything else , from tasso at @xmath81 , 35 , 44 @xcite , and 22 gev @xcite , mark ii @xcite and tpc @xcite at 29 gev , topaz at 58 gev @xcite , aleph @xcite , delphi @xcite , l3 @xcite , opal @xcite , and slc @xcite at 91 gev , aleph @xcite and opal @xcite at 133 gev , delphi at 161 gev @xcite , and opal at 172 , 183 , 189 @xcite , and 202 gev @xcite .
we place a small-@xmath0 cut @xcite on our data of @xmath82 where @xmath83 is a mass scale of @xmath84 gev@xmath85 .
we fit the gluon ff , @xmath86 , as well as the quark ffs , @xmath87 , \nonumber\\ d_{uc}(x , q_0
^ 2)&=&\frac{1}{2}\left[d_u(x , q_0 ^ 2)+d_c(x , q_0 ^ 2)\right],\end{aligned}\ ] ] where @xmath88 gev . since the hadron charge is summed over , we set @xmath89 .
for each of these three ffs , we choose the parameterization @xmath90 which , at small @xmath0 , is a gaussian in @xmath91 for @xmath92 with center positive in @xmath91 for @xmath93 as is found to be the case , while it reproduces the standard parameterization ( _ i.e. _ that without the @xmath94 factor ) used in global fits at intermediate and large @xmath0 .
we use eq .
( [ dlarelfordquarkanddg ] ) to motivate the simplification @xmath95 to our parameterization .
we also fit @xmath13 , so that there is a total of 9 free parameters . since we only use data for which @xmath96 , where @xmath97 gev is the mass of the bottom quark , and since @xmath98 , we take @xmath99 . while the precise choice for @xmath17 does not matter in the dla , calculations in the fo approach strongly depend on it .
since the theoretical error in our lo approach is larger than that at nlo , some of our results should be interpreted somewhat qualitatively . in particular
, the result for @xmath13 has a theoretical error of a factor of @xmath100 . without the @xmath101 factors , eq .
( [ dlarelfordquarkanddg ] ) implies that @xmath102 since the @xmath103 factors are important at large @xmath0 , where eq .
( [ dlarelfordquarkanddg ] ) is no longer valid , we do not impose eq .
( [ constraintsonac ] ) , but rather test its validity over the whole @xmath0 range .
.[tab1]parameter values for the ffs at @xmath88 gev parameterized as in eq .
( [ genparam ] ) from a fit , with @xmath104 , to all data listed in the text using dglap evolution in the fo approach to lo .
the fit also yields @xmath105 mev . [ cols="^,^,^,^,^",options="header " , ] fit to data as described in table [ tab3].,scaledwidth=48.0% ] we now perform the last fit again , but with @xmath66 included in the list of free parameters . we obtain the results in table [ tab3 ] and fig . [ fig4 ] , and the values @xmath106 mev and @xmath107 mev .
treatment of hadron mass effects renders the value of @xmath13 obtained in the fit with dl resummation more reasonable and slightly improves the latter , yielding @xmath108 .
we conclude , therefore , that to improve the large-@xmath91 description and to achieve a reasonable value for @xmath13 , both dl resummation and treatment of hadron mass effects are required .
in conclusion , we have proposed a single unified scheme which can describe a larger range in @xmath0 than either fo dglap evolution or the dla , by implementing sgl resummation in the standard dglap formalism in a minimal and unambiguous way through any order of perturbation theory .
our present limitation to lo is caused by the lack of knowledge on sgl resummation necessary to match nlo dglap evolution .
the actual inclusion of higher orders , although conceptually straightforward in our scheme , requires a separate analysis , which lies beyond the scope of ref .
@xcite and is left for future work .
further improvement in the small-@xmath0 region can be expected from the inclusion of resummed sls .
our scheme allows for a determination of quark and gluon ffs over a wider range of data than previously achieved , and should be incorporated into global fits of ffs such as that in ref .
@xcite , since the current range of @xmath109 is very limited .
since ffs and their dglap evolution are universal , our approach should be expected to also improve the description of inclusive hadron production from reactions other than @xmath78 annihilation , such as those involving at least one proton in the initial state .
the author is grateful to the organizers of the ipm school and conference on lepton and hadron physics for the kind invitation and for creating such a stimulating atmosphere , and to simon albino , gustav kramer , and wolfgang ochs for their collaboration on the work presented here .
this work was supported in part by the deutsche forschungsgemeinschaft through grant no .
kn 365/3 - 1 and by the bundesministerium fr bildung und forschung through grant no
. 05 ht4gua/4 .
j. binnewies , b. a. kniehl , and g. kramer , z. phys .
c * 65 * , 471 ( 1995 ) ; phys .
d * 52 * , 4947 ( 1995 ) ; * 53 * , 3573 ( 1996 ) ; b. a. kniehl , g. kramer , and b. ptter , phys . rev .
* 85 * , 5288 ( 2000 ) ; nucl .
b * 582 * , 514 ( 2000 ) ; * 597 * , 337 ( 2001 ) ; s. kretzer , phys . rev .
d * 62 * , 054001 ( 2000 ) ; l. bourhis , m. fontannaz , j. ph .
guillet , and m. werlen , eur .
j. c * 19 * , 89 ( 2001 ) .
v. n. gribov and l. n. lipatov , yad .
fiz . * 15 * , 781 ( 1972 ) [ sov .
j. nucl .
phys . * 15 * , 438 ( 1972 ) ] ; l. n. lipatov , yad .
fiz . * 20 * , 181 ( 1974 ) [ sov .
j. nucl .
* 20 * , 94 ( 1975 ) ] ; g. altarelli and g. parisi , nucl .
b * 126 * , 298 ( 1977 ) ; yu .
l. dokshitzer , zh . eksp .
* 73 * , 1216 ( 1977 ) [ sov .
jetp * 46 * , 641 ( 1977 ) ] . | we review recent progress in the development of an approach valid to any order which unifies the fixed - order dglap evolution of fragmentation functions at large @xmath0 with soft - gluon logarithmic resummation at small @xmath0 . at leading order ,
this approach , implemented with the double logarithmic approximation , reproduces exactly the modified leading logarithm approximation , but it is more complete due to the degrees of freedom given to the quark sector and the inclusion of the fixed - order terms .
we find that data from the largest @xmath0 values to the peak region can be better fitted than with other approaches .
in addition , we develop a treatment of hadron mass effects that leads to additional improvements at small @xmath0 . |
hawking s vacuum conservation theorem @xcite states that in a spacetime which admits a time function and which satisfies the dominant energy condition , the stress - energy tensor vanishes on a compact domain delimited by two locally achronal hypersurfaces @xmath0 and @xmath1 provided it vanishes on @xmath0 ( or @xmath1 ) . roughly speaking ,
if a space region is empty at one time it will remain so at later times provided no matter - energy flows in from outside .
a more detailed proof can be found in ( * ? ? ?
4.3 ) and a simplified argument was obtained by carter in @xcite . in this work
we wish to remove the assumption on the existence of a time function .
this is desirable since the domain under consideration is often of the form @xmath2 , with @xmath0 partial cauchy hypersurface , where the horizon @xmath3 might form precisely due to the presence of closed timelike curves behind it , where a time function can not exist .
hawking was aware of this limitation in his theorem and in fact on @xcite he sketched a possible more involved proof which however demanded a slightly stronger energy condition ( * ? ? ?
* point ( 5 ) , p. 293 ) .
furthermore , in the new argument hawking still assumes that a time function of the globally hyperbolic set @xmath4 extends up to the boundary @xmath5 remaining @xmath6 , which should not be expected in general due to the lack of regularity of @xmath5 . in this work
we shall retain carter s simplification and we shall reach the desired conclusion without invoking the existence of a time function . also , we shall not need to strengthen the energy condition .
we recall that a _ spacetime _ @xmath7 is a paracompact , connected , time oriented lorentzian manifold of dimension @xmath8 .
the metric has signature @xmath9 . in our terminology
the lightlike , timelike or causal vectors are non - zero . a null ( resp .
non - spacelike ) vector is lightlike ( resp .
causal ) or zero .
we assume that the metric satisfies the einstein equations with cosmological constant , and that the stress - energy tensor satisfies : the _ dominant energy condition_. on every tangent space @xmath10 the linear endomorphism @xmath11 sends the future - directed non - spacelike cone into itself .
when @xmath12 is timelike and normalized @xmath13 represents the energy - mumentum flow which the condition demands to be causal , roughly speaking energy can not travel faster than light .
physically , the dominant energy condition is nothing more than that since the constraint for @xmath14 null follows by continuity from that for @xmath14 timelike .
let us recall some basic consequences of the dominant energy condition .
as observed by carter a non - trivial endomorphism ( [ llp ] ) can not send a future - directed timelike vector @xmath14 into the zero vector .
indeed , let us consider a small perturbation @xmath15 observe that since the timelike cone is open @xmath16 is timelike for sufficiently small @xmath17 . clearly ,
if the rank of @xmath18 is non - zero then we can find a @xmath19 with non - vanishing image , and altering its sign if necessary , we obtain an image vector which is not a future - directed non - spacelike vector .
the contradiction proves that only null vectors can be sent to the zero vector . since the scalar product of two future - directed causal vectors is non - positive and vanishes if and only if they are proportional and lightlike we see that , for any two future - directed causal vectors @xmath14 and @xmath20 @xmath21 equality holding only if at least one of the vectors is lightlike . in particular , the dominant energy condition ( dec ) implies the _ weak energy condition _ ( wec ) which demands that for every timelike vector @xmath14 ( and hence by continuity , for every null @xmath14 ) @xmath22 , and the wec implies the _ null energy condition _
( nec ) also called _ null convergence condition _ according to which @xmath23 for every null vector @xmath14 . we need to find some inequalities implied by the dominant energy condition @xcite .
observe that if @xmath24 is future - directed timelike and normalized and @xmath25 is orthogonal to it then , with @xmath26 , the vector @xmath27 is lightlike thus @xmath28 which can be rewritten @xmath29 similarly , if @xmath30 is another vector orthogonal to @xmath24,@xmath31 is lightlike thus @xmath32 @xmath33 we are going to obtain an inequality involving just @xmath34 , thus we can redefine @xmath35 if necessary to make @xmath36 . replacing @xmath37 and @xmath38 in eq.([aao ] ) gives @xmath39 from this inequality we obtain the two inequalities @xmath40 summing ( and considering the equation so obtained also for @xmath41 ) @xmath42 any tensor @xmath43 whose contractions with @xmath44 vanishes can be written as the sum of @xmath45 terms of the form @xmath46 thus @xmath47 equations ( [ one ] ) , ( [ two ] ) , prove that in any orthonormal frame in which @xmath48 is timelike @xmath49
we are ready to prove a first vacuum conservation result .
[ kaq ] let @xmath7 be a spacetime which satisfies the dominant energy condition .
let us consider an open relatively compact region @xmath50 of spacetime delimited by two locally achronal hypersurfaces @xmath0 and @xmath1 , possibly with edge , such that @xmath51 is the union of integral connected segments ( possibly degenerate to a point ) of a @xmath6 future - directed timelike vector field @xmath52 which have past endpoint in @xmath0 and future endpoint in @xmath1 .
then the stress energy tensor vanishes on @xmath0 iff it vanishes on @xmath1 iff it vanishes on @xmath53 .
let us first give the proof for @xmath0 and @xmath1 continuously differentiable .
let us normalize @xmath52 , @xmath54 .
let @xmath55 be the map which assigns to @xmath56 the endpoint of the curve @xmath57 \to m$ ] , @xmath58 , @xmath59 , wherever it exists .
by well known results on the dependence on initial conditions and external parameters of ordinary differential equations @xcite this flow map @xmath60 is @xmath6 on an open subset of @xmath61 where the map @xmath55 makes sense . for every @xmath62
the map @xmath63 is actually a local diffeomorphism wherever it is defined ( * ? ? ?
* theor .
2.9 ) . by assumption every point of @xmath51 can be reached from @xmath0 following an integral line of @xmath52 .
let @xmath64 be the function defined through : @xmath65 is such that @xmath66 for some @xmath67 .
by the implicit function theorem @xmath62 has the same regularity of @xmath0 thus @xmath6 , and by construction @xmath68=1 $ ] .
we stress that we are not claiming that @xmath62 is a time function , just that it is well defined over @xmath51 . for a real number @xmath69 to be specified later on , let us introduce the vector field @xmath70 which is future - directed timelike wherever @xmath71 or zero .
let us denote with @xmath72 the usual volume form .
let us consider the non - negative integrals @xmath73 if @xmath74 at a point @xmath75 , @xmath76 , then @xmath77 is timelike and hence transverse to @xmath78 , thus the integral is strictly positive . as a consequence , for any @xmath76 , @xmath79 iff @xmath80 on @xmath78 iff @xmath81 on @xmath78 .
recalling the identity @xmath82 we obtain from stokes theorem @xmath83 using @xmath84 @xmath85 .\ ] ] since @xmath86 v^\mu v^\nu =- c , \ ] ] the matrix in square brackets can be written @xmath87 for some continuous tensor fields @xmath88 such that their contractions with @xmath52 vanish .
this fact can be most easily understood passing to an orthonormal basis with @xmath89 . in the same way
one easily sees that @xmath90 and @xmath91 . on every relatively compact open set @xmath92 which admits a coordinate system the contraction @xmath93 $ ] can be made non - positive by choosing a sufficiently large value of @xmath69 , and non - negative by choosing a sufficiently large value of @xmath94 . indeed , by the dominant energy condition and by using the continuity of @xmath95 on the compact set @xmath96 we have that , due to eqs .
( [ one ] ) and ( [ three ] ) , there is a constant @xmath97 such that ( @xmath98 is the spacetime dimension ) @xmath99 where an equation analogous to the first one holds also for @xmath100 .
the statement at the beginning of the paragraph follows plugging ( [ dec ] ) into eq .
( [ div ] ) . since @xmath51 can be covered by a finite number of relatively compact coordinated neighborhoods we have on @xmath51 , @xmath101 .
observe that actually by choosing @xmath102 sufficiently large the equality case is excluded unless @xmath81 at the point under consideration .
thus for @xmath103 large enough @xmath104 where the second equality holds only if @xmath81 on @xmath50 and hence @xmath51 .
thus if we take @xmath69 large enough we have that @xmath81 on @xmath0 implies @xmath81 on @xmath53 and hence in @xmath1 .
analogously , for @xmath105 large enough @xmath106 where the second equality holds only if @xmath81 on @xmath50 and hence @xmath51 .
thus if we take @xmath94 large enough we have that @xmath81 on @xmath1 implies @xmath81 on @xmath53 and hence in @xmath0 .
the proof for the lipschitz regularity of @xmath0 and @xmath1 is as above , there are just two critical steps .
the first is the construction of function @xmath62 .
we can not define it starting the flow from @xmath0 since , as this hypersurface is only lipschitz , @xmath62 would not be @xmath6 .
however , in a neighborhood of @xmath107 one can introduce coordinates @xmath108 such that @xmath109 , and @xmath110 are induced from the @xmath52-projection on a smooth spacelike hypersurface @xmath111 transverse to @xmath52 .
then @xmath0 can be locally expressed as a lipschitz graph @xmath112 which can be approximated by a smooth graph by whitney approximation theorem ( * ? ? ?
* theor . 6.21 ) .
through a partition of unity it is possible to patch together these local graphs into a smooth hypersurface @xmath113 with the property that every @xmath52-integral curve which intersect @xmath0 intersects @xmath113 and conversely .
then the definition of @xmath62 is as above but starting the flow from @xmath113 .
the other delicate step concerns the application of the gauss - green ( divergence ) theorem which , fortunately , has indeed been generalized to lipschitz domains and even lipschitz fields ( * ? ? ?
5.8 ) @xcite . in this theorem
the boundary term @xmath114 , @xmath115 , can be calculated expressing @xmath78 as a local graph and extending the expression of the integral in terms of a @xmath6 function @xmath116 to the lipschitz case .
more in detail , if @xmath117 @xmath118 observe that @xmath119 is @xmath120 by the lipschitzness of @xmath116 , thus the integral @xmath114 is well defined .
@xmath121 the physical meaning of the integral @xmath122 can be understood , at least in the continuously differentiable case , as follows .
let us denote with @xmath123 , @xmath76 , the open subset of @xmath78 in which @xmath77 is timelike .
let us endow @xmath123 with the following volume form @xmath124 where @xmath125 is a @xmath126 future - directed causal field normal to @xmath78 ( possibly lightlike ) . here
@xmath127 is evaluated just on the tangent space to @xmath128 .
this choice of volume is independent of the transverse field @xmath77 but it depends on the scale of @xmath125 ( this multiplying factor could be naturally fixed through normalization if @xmath125 were timelike ) .
we have @xmath129 \nu= \int_{\hat s_{k } } e^{-ct } t(v , n ) \nu . \label{kos}\end{aligned}\ ] ] one should be very careful in using the last expression when @xmath78 is partly lightlike for one can easily miss the difference between @xmath123 and @xmath78 .
[ odp ] let @xmath130 be a locally achronal compact topological hypersurface possibly with edge and let @xmath131 be compact . under the dominant energy condition the stress energy tensor vanishes on @xmath130 iff it vanishes on @xmath132 iff it vanishes on @xmath131 .
let @xmath52 be any smooth normalized future - directed timelike field , and let @xmath133 and @xmath134 .
the assumption of theorem [ kaq ] are satisfied thus the desired conclusion follows .
@xmath121 the usual vacuum conservation theorem states that if the stress - energy tensor vanishes on @xmath130 then it vanishes on @xmath135 and on the horizon @xmath132 .
actually , the previous result establishes a converse implication which is of particular interest , namely that if @xmath71 somewhere on @xmath130 then @xmath71 at some point @xmath136 .
suppose that @xmath132 is @xmath6 at @xmath137 and that @xmath125 is a future directed lightlike vector orthogonal to @xmath138 .
observe that we can not conclude that @xmath139 at @xmath137 , for it can be that @xmath140 as for some stress - energy tensor of pure aligned radiation ( type ii ( * ? ? ?
* sect . 4.3 ) ) .
let us introduce two modifications of the dominant energy condition .
the _ stable dominant energy condition _ @xcite . on every tangent space @xmath10 the linear endomorphism ( [ llp ]
) sends the future - directed casual cone into the future - directed timelike cone .
+ the _ weakened stable dominant energy condition_. at every point @xmath81 or the stable dominant energy condition holds , which is equivalent to : the linear endomorphism ( [ llp ] ) sends the future - directed casual cone into the future - directed timelike cone plus the zero vector . or if no causal vector is sent to zero we have finished .
thus assume @xmath71 and there is a f.d .
causal vector @xmath20 which has zero image . since @xmath71 there is a f.d .
causal vector @xmath14 such that @xmath141 , necessarily f.d .
timelike by assumption , then @xmath142 which is a contradiction since it must be positive . ] the stable dominant energy condition is preserved under small perturbations of the endomorphism , which means that the source content of spacetime is not on the verge of violating the dominant energy condition .
unfortunately , it might seem a too strong condition as it excludes the vacuum case , that is , under this condition @xmath71 at every point .
observe that we might want to impose the stable dominant energy condition only where a source is really present .
this leads us to the weakened stable dominant energy condition which contemplates the vacuum case . in the four dimensional spacetime case and in the classification of (
4.3 ) , this condition allows only type i stress - energy tensors ( namely the diagonal ones ) where the energy density is larger than the absolute values of the principal pressures , @xmath143 or @xmath144 ( observe that types iii and iv are excluded by the null energy condition ) .
let @xmath136 be a differentiability point of the horizon , and let @xmath125 be a future directed lightlike vector tangent to the horizon at @xmath137 .
if the weakened stable dominant energy condition holds and @xmath71 at @xmath137 then @xmath139 at @xmath137 , thus , where @xmath125 is a lightlike vector , and then he applies a version of the conservation theorem which he has not really proved .
probably he used eq .
( [ kos ] ) missing the difference between @xmath145 and @xmath130 , and that between @xmath52 and @xmath125 . ] under the assumptions of corollary [ odp ] , and under the weakened stable dominant energy condition @xmath71 somewhere on @xmath130 implies @xmath139 somewhere on the horizon @xmath132 ( namely at some differentiability point of the horizon where a lightlike tangent @xmath125 to the horizon exists ) . we know that @xmath71 at some point @xmath146 , thus in a neighborhood @xmath50 of @xmath147 . as
the horizon is almost everywhere differentiable and @xmath18 is continuous we can find @xmath148 where the horizon is differentiable and @xmath71 .
@xmath121 this result will allow us to show that under the weakened stable dominant energy condition compact cauchy horizons can not form @xcite , a result which seems significative for a final resolution of the weak / strong cosmic censorship conjecture . physically speaking , in absence of quantum effects the weakened stable dominant energy condition seems reasonable in all those phenomena in which it is known that the source of spacetime is not just coherent radiation , particularly in the study of the development of future singularities .
less convincing is its validity at the beginning of the universe since the higgs field may still stay in its non - degenerate minimum so that , as all the particles would be massless , the only source content would be radiation . in this case
if @xmath130 is a present time partial cauchy hypersurface the radiation would emerge from the boundary @xmath149 which would play the role of big bang null hypersurface . for more on this picture of the beginning of the universe and its advantages
the reader is referred to @xcite .
we have proved a version of the vacuum conservation theorem which has some good features : ( a ) it involves causal rather than just spacelike hypersurfaces , ( b ) it holds for lipschitz hypersurfaces , ( c ) it does not assume the existence of a time function , ( d ) it includes a converse preservation claim according to which if the stress - energy tensor does not vanish on @xmath130 then it does not vanish on @xmath150 .
the application of this theorem to the study of compact cauchy horizons has been briefly commented .
carter , b. : energy dominance and the hawking - ellis vacuum conservation theorem . in : the future of the theoretical physics and cosmology ( cambridge , 2002 ) ,
cambridge univ . press ,
cambridge ( 2003 ) minguzzi , e. : causality and entropic arguments pointing to a null big bang hypersurface .
j. of phys .
: conf . ser . *
314 * , 012,098 ( 2011 ) .
ontribution to the proceedings of the conference xxxiii spanish relativity meeting , gravity as a crossroad in physics , granada ( spain ) , september 6 - 10 , 2010 | a version of the vacuum conservation theorem is proved which does not assume the existence of a time function nor demands stronger properties than the dominant energy condition .
however , it is shown that a stronger stable version plays a role in the study of compact cauchy horizons . |
observation of the neutrinoless ( 0@xmath6 ) double beta ( @xmath7 ) decay and/or dark matter particles would be extraordinary events for the modern physics .
it is because both the phenomena are related to the new physical effects beyond the standard model of the particle theory .
the great interest to the @xmath8 decay searches @xcite have arisen from the recent evidence of neutrino oscillations , strongly suggesting that neutrinos have nonzero mass .
while oscillation experiments are sensitive to the neutrinos mass difference , only the measured @xmath8 decay rate could establish the majorana nature of the neutrino and give the absolute scale of the effective neutrino mass .
several sensitivity studies of the @xmath7 decay processes were performed using low background crystal scintillators , which contain various candidate nuclei : @xmath9ca @xcite , @xmath10ca @xcite , @xmath11cd @xcite , @xmath12cd , @xmath13cd and @xmath14cd @xcite , @xmath15ce and @xmath16ce @xcite , @xmath17gd @xcite , @xmath18w and @xmath19w @xcite .
for instance , in the experiment @xcite with the cadmium tungstate ( cdwo@xmath1 ) scintillators enriched in @xmath14cd the very low counting rate of 0.04 counts/(yr@xmath20kev@xmath20 kg ) was reached in the energy window @xmath21 mev .
the half - life limit on @xmath8 decay of @xmath14cd was set as @xmath22 yr at 90% cl , which leads to one of the strongest restriction on the effective majorana neutrino mass @xmath23 ev .
there are strong evidence for a large amount of dark matter in the universe ( matter which reveals itself only through gravitational interaction ) . weakly interacting massive particles ( wimp ) in particular neutralino , predicted by the minimal supersymmetric extensions of the standard model
are considered as one of the most acceptable component of the dark matter @xcite .
these particles can be detected due to their scattering on nuclei producing low energy nuclear recoils .
extremely low counting rate and small energy of recoil nuclei are expected in experiments to search for the wimp .
therefore , the experiments for direct wimp search require extremely low level of background ( less than several events per kilogram per day per kev ) with very low energy threshold ( less than tens kev ) . direct methods of wimp detection are based on registration of ionization or / and excitation of recoil nucleus in the material of the detector itself . at present , most sensitive experiments apply different detectors for wimp search : ge semiconductor detectors @xcite , bolometers @xcite , scintillation detectors @xcite . an interesting possibility to reject background caused by electrons provides cryogenic technique , which uses simultaneous registration of heat and charge @xcite or heat and light signals @xcite . in ref .
@xcite calcium tungstate crystals ( cawo@xmath1 ) are discussed as promising material for such a kind of dark matter detector . however , radioactive contamination of cawo@xmath1 crystals used in @xcite , as well as investigated in @xcite , is too high for low counting measurements .
the purpose of our work was investigation of scintillation properties and radioactive contamination of zinc tungstate ( znwo@xmath1 ) crystals as possible detectors for the double beta decay and dark matter experiments .
three clear , slightly colored znwo@xmath1 crystals ( @xmath24 mm , @xmath25 mm , and @xmath26 mm ) were used in our studies .
the crystals were fabricated from monocrystals grown by the czochralski method .
the main properties of znwo@xmath1 scintillators are presented in table 1 , where characteristics of calcium and cadmium tungstates are given for comparison .
cdwo@xmath1 + density ( g/@xmath27 ) & 7.8 & 6.1 & 8.0 + melting point ( @xmath28c ) & 1200 & @xmath29 & 1325 + structural type & wolframite & sheelite & wolframite + cleavage plane & marked ( 010 ) & weak ( 101 ) & marked ( 010 ) + hardness ( mohs ) & @xmath30 & @xmath31 & @xmath30 + wavelength of emission maximum ( nm ) & 480 & @xmath32 & 480 + refractive index & @xmath33 & 1.94 & @xmath34 + effective average decay time@xmath35 ( @xmath36s ) & 24 & 8 & 13 + photoelectron yield [ % of nai(tl)]@xmath35 & 13% & 18% & 20% + + scintillation characteristics of zinc tungstate crystals have been studied in @xcite .
light yield of znwo@xmath1 was measured in @xcite with the help of silicon photodiodes as 9300 photons / mev ( which is @xmath37% of nai(tl ) light yield ) .
we have estimated the value of the photoelectron yield of the znwo@xmath1 in measurements with two scintillators : znwo@xmath1 @xmath25 mm and standard nai(tl ) @xmath38 mm .
the crystals were coupled to the pmt ( emi d724kflb ) with the bialkali rbcs photocathode .
the znwo@xmath1 detector was wrapped by ptfe reflector tape .
the relative photoelectron yield of znwo@xmath1 was determined as 13% of that of nai(tl ) .
the energy resolution of znwo@xmath1 crystals for 662 kev @xmath39 rays ( @xmath40cs ) has been reported as fwhm@xmath41 @xcite and fwhm@xmath42 @xcite . in the present work the energy resolution fwhm@xmath43 for 662 kev @xmath39 line of @xmath40cs was measured with the znwo@xmath1 @xmath25 mm crystal .
the crystal was wrapped by ptfe reflector tape and optically coupled to 3 " photomultiplier ( pmt ) philips xp2412 .
taking into account slow decay of znwo@xmath1 scintillation , the 24 @xmath36s shaping time of spectroscopy amplifier was set .
a substantial improvement of light collection ( @xmath44% ) and energy resolution was achieved by placing the crystal in liquid ( silicone oil with index of refraction @xmath45 ) .
the crystal was fixed in center of the teflon container @xmath46 mm and viewed by two pmts xp2412 .
1(a ) demonstrates the energy spectra of @xmath40cs and @xmath47co obtained in such a way with the @xmath487 mm znwo@xmath1 scintillation crystal . for the first time
the energy resolution 9.1% and 6% was measured with znwo@xmath1 crystal scintillator for 662 and 1333 kev @xmath39 lines , respectively .
furthermore , the resolution 11.3% for the 662 kev @xmath39 line was obtained with the larger crystal @xmath4928 mm in the same measurement conditions .
1(b ) shows the energy spectra of @xmath50am and @xmath51fe low energy gamma and x - ray lines measured with the znwo@xmath1 scintillator @xmath52 mm ( the crystal was wrapped by ptfe reflector tape and optically coupled to the pmt xp2412 ) .
the 6 kev peak of @xmath51fe source is still resolved from the pmt noise .
the @xmath53 ratio ratio is defined as ratio of @xmath54 peak position in the energy scale measured with @xmath55 sources to the energy of @xmath54 particles . because @xmath39 quanta interact with detector by @xmath56 particles , we use more convenient term
@xmath53 ratio .
] was measured ( in the energy range of @xmath57 mev ) with the help of the collimated beam of @xmath58 particles from an @xmath50am source and using various sets of thin mylar films as absorbers .
the energies of @xmath58 particles were determined with the help of a surface - barrier detector ( see for details @xcite ) .
the znwo@xmath59 crystal was irradiated in the directions perpendicular to ( 010 ) , ( 001 ) , and ( 100 ) crystal planes .
the dependence of the @xmath60 ratio on energy and direction of @xmath58 beam relatively to the crystal planes ( see fig .
2 ) is similar to that observed with cdwo@xmath1 scintillators @xcite .
as the quenching of the scintillation light caused by @xmath58 particles ( in comparison with electrons ) is due to the higher ionization density of @xmath58 particles , such a behaviour of the @xmath60 ratio can be explained by the energy dependence of ionization density of @xmath58 particles @xcite .
the time characteristics of znwo@xmath1 scintillators were studied as described in @xcite with the help of a transient digitizer based on the 12 bit adc ( ad9022 ) operated at the sample rate of 20 ms / s .
the scintillator was irradiated by @xmath61 mev @xmath58 particles in the direction perpendicular to the ( 010 ) crystal plane .
shapes of the light pulses produced by @xmath58 particles and @xmath39 rays are depicted in fig .
fit of the pulses by the function : @xmath62 , where @xmath63 are intensities ( in percentage of the total pulse intensity ) , @xmath64 are decay constants for different light emission components , @xmath65 is integration constant of electronics ( @xmath66s ) , gives three decay components @xmath67 , @xmath68 and @xmath69 @xmath36s with different amplitudes for @xmath39 rays and @xmath58 particles ( see table 2 ) .
the value of the slow decay component is in agreement with the result of 25 @xmath36s obtained in @xcite , while 0.7 and 7 @xmath36s decay time constants were identified at the first time .
we were not able to measure the 0.1 @xmath36s decay component reported in @xcite , because of the not enough fast digitizer used in our measurements .
.decay times of znwo@xmath1 scintillators for @xmath39 quanta and @xmath58 particles .
the decay constants and their intensities ( in percentage of the total intensity ) are denoted as @xmath70 and a@xmath71 , respectively . [ cols= "
< , < , < , < , < " , ] in addition , the observed with znwo@xmath1 scintillators dependence of pulse shape on the direction of @xmath58 particles irradiation relative to the crystal axes ( supposing that such a dependence would remain valid for nuclear recoils ) could be used to detect a diurnal asymmetry of wimp direction @xcite .
scintillation properties of znwo@xmath1 crystals were studied . the energy resolution 9.1% ( 662 kev @xmath40cs @xmath39 line )
was obtained with znwo@xmath1 crystal scintillator placed in liquid and viewed by two pmts .
shapes of scintillation signals was investigated , and clear pulse - shape discrimination for @xmath39(@xmath56 ) and @xmath58 events was achieved .
dependences of the @xmath0 ratio and scintillation pulse shape on the energy of alpha particles were measured .
it was also found that the @xmath0 ratio and pulse shape show the dependence on direction of @xmath58 irradiation relatively to the main crystal planes of znwo@xmath1 crystal .
radioactive contamination of znwo@xmath1 crystal was measured in the solotvina underground laboratory .
no radioactive pollution in znwo@xmath1 scintillator was observed .
limits on radioactive contamination in znwo@xmath1 crystal are on the level from tenth to a few mbk / kg which is one two order of magnitude lower than those for cawo@xmath1 . two potentially @xmath7 active zinc and two tungsten isotopes ( @xmath3zn , @xmath72zn , @xmath18w , and @xmath19w ) were studied with the help of znwo@xmath1 crystal .
new improved half - life limits on double beta decay in zinc isotopes were established . in particular , for the @xmath73 decay of @xmath3zn the half - life limit was set as @xmath74 yr at 68% cl . due to a low level of radioactive contamination ,
znwo@xmath1 crystals seems to be encouraging material for the dark matter experiments .
znwo@xmath1 crystals could provide a possibility to search for spin - dependent inelastic scattering of wimp with excitation of low energy nuclear levels .
most interesting targets are @xmath75zn and @xmath76w . a strong signature can be used to search for spin - dependent inelastic scattering of wimp on nuclei with nonzero spin ( which are present in znwo@xmath1 crystals ) , via registration of nuclear recoils and delayed @xmath39 quanta . due to dependence of znwo@xmath1 scintillation pulse shape on direction of irradiation by high ionizing particles
, there is @xmath77 a possibility to search for diurnal modulation of wimp direction .
it could be a strong signature for wimp evidence . c. bacci et al .
, astropart .
phys . 2 ( 1994 ) 13 ; r. bernabei et al .
b 389 ( 1996 ) 757 ; 424 ( 1998 ) 195 ; 450 ( 1999 ) 448 ; 480 ( 2000 ) 23 ; 509 ( 2001 ) 197 ; eur .
j. c 23 ( 2002 ) 61 ; riv .
26 ( 2003 ) 1 . | energy resolution , @xmath0 ratio , and the pulse shape discrimination ability of the znwo@xmath1 crystal scintillators were studied .
the radioactive contamination of a znwo@xmath1 crystal was investigated in the solotvina underground laboratory .
possibilities to apply znwo@xmath1 crystals for the dark matter and double beta decay searches are discussed .
new improved half - life limits on double beta decay in zinc isotopes were established , in particular , for @xmath2 decay of @xmath3zn as : @xmath4 yr and @xmath5 yr , both at 68% cl
. -0.5 cm -0.5 cm 0.2 cm f.a .
danevich , v.v .
kobychev , s.s .
nagorny , d.v .
poda , v.i .
tretyak , s.s .
yurchenko , yu.g .
zdesenko _ institute for nuclear research , msp 03680 kiev , ukraine
_ 0.4 cm 0.4 cm pacs numbers : 29.40.mc ; 23.40.-s ; 95.35.+d 0.4 cm keywords : znwo@xmath1 crystal scintillators ; double beta decay ; dark matter 1.0 cm |
the presence of strong x - ray and nonthermal radio emission in late - type stars is a well - known indication of enhanced coronal activity ( drake et al .
1992 ; gdel 2002 ) .
x - ray emission from coronae ( @xmath9k ) and chromospheric emission ( @xmath10k ) are also closely correlated , and thus stars with intense coronae will have strong chromospheric emission , as evidenced by their uv and ca ii h and k line emission ( e.g. ayres et al .
1995 ) . while many active stars have been identified as such through their x - ray and radio emission , it is only through detailed optical photometric and spectroscopic studies that their activity can be classified into known types .
rs cvn binaries are perhaps the most common type of active star , at least for stars having spectral types in the range from late f through k. in an rs cvn binary , the two stars are usually tidally locked so that the rotational period of each star is approximately the same as the orbital period ; however , the two stars are not undergoing mass transfer ( i.e. they form a detached system ) .
one star is generally a spectral type f to g dwarf or a subgiant that is @xmath11 1000 k hotter than its companion , which is usually a g to k type giant or sub - giant . in the more recent definition of the rs cvn class ,
there is no restriction on the spectral type of the secondary star or on the orbital period , except for the evolutionary constraint ( fekel et al .
rs cvns have been further subdivided into three groups according to their orbital period : the short period ( @xmath12 d ) , classical ( @xmath13 d ) , and long period ( @xmath14 d ) groups . in the soft ( 0.1 - 2.0 kev )
x - ray band , the luminosities of rs cvn binaries typically lie in the range of @xmath15 erg s@xmath16 ( drake et al .
1989 ) , compared to @xmath17 erg s@xmath16 for normal late - type stars ( schmitt & liefke 2004 ) .
vla and atca observations at cm wavelengths have shown that the radio luminosities of rs cvn systems are in the range from @xmath18 to @xmath19 erg s@xmath16 hz@xmath16 ( drake & linsky 1986 , drake et al.1989 ) , which is also enhanced compared to that found for other late - type stars , typically @xmath20 erg s@xmath16 hz@xmath16 ( gdel 2002 ) . the k0 iv star hd 81032 ( houk & smith - moore 1988 ) was first noticed as an x - ray active star when it was identified as the likely optical counterpart of a soft x - ray ( 0.2 - 4.0 kev ) source in the einstein slew survey , 1es 0920 - 13.6 , with an x - ray flux of @xmath21 erg @xmath22 s@xmath16 ( elvis et al .
1992 , schachter et al .
an x - ray source at this position was later detected in the rosat all - sky - survey ( rass ) at a weaker ( but statistically more significant ) level of @xmath23 erg @xmath22 s@xmath16 , named 1rxs j092253.7 - 134919 in the rass bright source catalogue ( voges et al .
this star was detected on february 12 , 1993 as a radio source , with a 3.6 cm flux of @xmath24 mjy ( drake , s. a. , private communication ) .
given its spectral type and luminosity class , and its v - magnitude of 8.91 ( wright et al .
2003 ) , the spectroscopic distance of hd 81032 is @xmath25 pc , implying an x - ray luminosity of @xmath26 erg s@xmath16 ( based on the rass count rate and an assumed conversion factor of @xmath27 erg @xmath22 pspc cts@xmath16 : a self - consistent x - ray luminosity based on the rass data is calculated in 8) , and a radio luminosity of @xmath28 erg s@xmath16 hz@xmath16 , where the distance uncertainty is the dominant contributor to the large errors in the luminosities .
these values clearly show that the star hd 81032 has an active corona . in this paper
, we present extensive optical photometric and spectroscopic observations , as well as an analysis of archival x - ray data , of the star hd 81032 .
this is the first detailed optical photometric and spectroscopic study of this star , although a brief discussion of some of the optical photometry of this star was given in pandey et al .
( 2002 ) . based on our long - term optical study ,
the star hd 81032 is shown to be a new member of the ( long - period ) rs cvn class .
the organization of this paper is as follows .
the observations and data reduction are presented in [ obsdat.sec ] . in
[ ligper.sec ] , we analyse the data for the periodicity . in
[ phopha.sec ] , we discuss the light curve and phase of minimum light .
chromospheric emission features are described in
[ halcai.sec ] . in
[ physical.sec ] and [ sed.sec ] , the spectral type and ir excess of hd 81032 are discussed , while the x - ray spectra are discussed in [ xspec.sec ] .
finally , [ con.sec ] summarizes the main results of this paper .
hd 81032 was observed in the johnson b , v and cousins r filter for 122 nights during four observing runs - year 2000 - 2001 ( 19 nights ) , year 2001 - 2002 ( 54 nights ) , year 2002 - 2003 ( 24 nights ) and year 2003 - 2004 ( 25 nights ) - at the aryabhatta research institute of observational sciences ( aries ) .
the observations were made with the 104-cm sampurnanand telescope and using a @xmath29 ccd camera in years 2000 - 2003 , and a @xmath30 ccd camera during the years 2003 - 2004 .
a few ccd frames were taken in the b , v and r filters on every night , with exposure times ranging from 2 to 60 secs , depending upon the seeing conditions and the filter used .
several bias and twilight flat frames were also taken during each observing run .
bias subtraction , flat fielding and aperture photometry were performed using iraf . for the star hd 81032
the comparison and check stars were tyc 5471 1345 1 and usno - a2.0 0750 - 06845737 , respectively .
differential photometry , in the sense of variable minus the comparison star , was done , since program , comparison and check stars were all in the same ccd frame .
ubvri observations of hd 81032 along with some of the landolt ( 1992 ) standard region sa 98 were obtained on 22 february 2004 for photometric calibration .
the average magnitudes of hd 81032 during the observing years 2003 - 2004 in the u , b , v , r and i filters were @xmath31 and @xmath32 , respectively .
spectroscopic observations were carried out during 2003 january 20 to 24 at the vainu bappu observatory , kavalur with the omr spectrograph fed by the 234-cm vainu bappu telescope .
the data were acquired with a @xmath33 ccd camera of @xmath34 square pixel size covering a range of @xmath35 and having a dispersion of @xmath36 .
the star was observed in the wavelength ranges of 3500 - 4700 and 5700 - 6900 .
four spectra of hd 81032 in the h@xmath2 region and three in the caii h and k region were obtained .
a signal - to - noise of ratio between 20 to 40 was achieved in these spectra .
hd 71952 , a k0iv type star , was also observed as a reference star for hd 81032 .
the spectra were extracted using the standard reduction procedures in the iraf packages ( bias subtraction , flat fielding , extraction of the spectrum and wavelength calibration using arc lamps ) .
the spectral resolution was determined by using emission lines of arc lamps taken on the same nights .
a spectral resolution ( @xmath37 ) of @xmath38 at @xmath39 and @xmath40 at @xmath41 was achieved .
all the spectra were normalized to the continuum and equivalent widths for the emission lines were computed using the iraf task _
splot_. the errors in the measurement of the equivalent widths of the emission lines were determined by measuring the equivalent widths of some moderate absorption lines for each spectra .
we computed the standard deviation for the equivalent width of each feature , and finally determined the mean standard deviation to be @xmath42 .
the star hd 81032 was observed and detected by the rosat pspc detector during the rosat all - sky - survey ( rass ) phase over a 2-day period from 1990 november 10 to 12 .
the exposure time was 501 s , and was accumulated in 26 separate short scans of this region of sky .
the pspc had an energy range from 0.1 - 2.4 kev with a ( low ) spectral resolution ( @xmath43 at 1 kev ) .
a full description of the x - ray telescope and detectors can be found in trmper ( 1983 ) and in pfeffermann et al.(1987 ) .
the rosat x - ray data for hd 81032 were obtained from the public archives , the relevant rass dataset being rs932025n00 .
source spectra for hd 81032 were accumulated from on - source counts obtained from a circular region on the sky centered on the x - ray peak and having a radius of 3.85 arcmin .
the background was accumulated from several neighboring regions at nearly the same offset from the source .
photometric light curves corresponding to the four observing runs were taken .
figure [ ligper.fig ] shows the v band differential light curves of the star hd 81032 at different epochs .
we did not find any significant variations in the comparison star ( see below ) .
the yearly mean of the standard deviation(@xmath44 ) between the different measures of comparison and check stars in the b , v and r filters was found to be 0.011 , 0.01 and 0.01 , respectively . each light curve shown in fig .
[ ligper.fig ] was analysed for periodicity . to find a period from unequally spaced data
, we used the clean algorithm ( roberts et al .
1987 ) in starlink s period software .
the power spectrum obtained using this method is shown in the inset of each panel of fig .
[ ligper.fig ] along with the period determined .
the cleaned power spectra presented were obtained after 100 iterations with a loop gain of 0.1 .
the period is found to be constant within error for each epoch . to improve the period determination of the star hd 81032 ,
the entire data from 2000 - 2004 were analysed using the same algorithm .
figure [ power.fig ] shows the cleaned power spectrum from the entire dataset .
the highest peak in the cleaned power spectrum corresponds to a period of @xmath45 .
the 18.802 d period is much more plausible than the 57 d period reported earlier ( pandey et al .
the previous determination was mainly due to observational limitations as our early data were too sparse and highly uneven . besides a subgiant like this one is unlikely to be synchronized in a binary of 57-days period .
the julian days of the observations were converted to the phases using the ephemeris : @xmath46 where the initial epoch corresponds to the conjunction with the first minimum observed .
figure [ fold.fig ] shows the differential b , v , r and v@xmath47 , light curves and , ( b - v ) and ( v - r ) colour curves of the star hd 81032 . here
v@xmath47 stands for the differential v band light curve of the comparison and the check stars .
each point in the light curves is mean of 3 - 4 independent observations taken over a night .
the light curve during the observing years 2001 - 2002 has a dense temporal coverage .
we , therefore , divided this light curve into two different epochs to see any variation in the @xmath48 and the amplitude .
the mean epoch of the light curves , the observed maximum ( @xmath49 ) and minimum ( @xmath50 ) in the v band , peak to peak amplitude ( @xmath51 ) , and phase of minima(@xmath48 ) are listed in table [ ligper.tab ] .
the value of @xmath49 was constant during each epoch indicating that the brightness of unspotted photosphere was constant from epoch to epoch .
however , the value of @xmath50 was reduced by 0.14 mag from epoch a to epoch b , and remained constant during the epoch c , d and e. the rs cvn systems usually show one or two well defined minima , thereby indicating that the rotational modulations caused by one or two prominent spots or groups of spots .
additional spots may be present at other longitudes , or in the circumpolar regions but contribution to the overall rotational modulation may not be appreciable .
the phase of the light minimum ( @xmath48 ) directly indicates the mean longitude of the dominant groups of spots .
a sharp minimum was observed during the epoch a ( see fig . [ fold.fig]a ) . at the same time
the amplitude of the v band light curve was found to be 0.288 mag , which was maximum during our observations .
the sharpness of the minimum indicate that it is the latitudinal extent of the groups of spots that may be responsible rather than the longitudinal extent .
broad minima during the epochs b , c and d indicate that the spots were spread over an appreciable longitudinal range ( fig .
[ fold.fig ] b to d ) .
it is interesting to see the light curve of the star hd 81032 during the epoch e. here a single large spot , characterized by a broad minimum during the epochs b , c and d separated into two groups of spots .
this can be easily seen by two well separated minima ( see fig .
[ fold.fig ] e ) .
significant change in @xmath48 ( see table [ ligper.tab ] ) is probably associated with a change in the spot configuration on the surface of the star .
variation in the colour of the star hd 81032 is correlated with its magnitude ( see top two panels of fig .
[ fold.fig ] a to e ) i.e. the star becomes redder when fainter , and bluer when brighter , supporting the starspot hypothesis .
the significance of the correlation has been calculated by determining the linear correlation coefficient , r , between the magnitude and the colours .
the value of r between v and ( b - v ) , v and ( v - r ) and ( b - v ) and ( v - r ) was found to be 0.20 , 0.55 and 0.40 , with the corresponding probability of no correlation being 0.0253 , @xmath52 and @xmath53 , respectively .
the conventional starspot model assumes that the spots are cooler than the surrounding photosphere , and hence one would expect the star to be the reddest at the light minimum , although flaring activity can lead an opposite effect being observed as seen in ux ari ( ulvs & henry 2003 ; padmakar & pandey 1999 ; ravindran & mohin 1995 ) .
figure [ phmin.fig ] shows the plot of the mean epoch versus phase minima ( @xmath48 ) of the v band light curves .
the following equation was fitted to the data by the method of linear least squares to determine the rate of the phase shift ( or migration period ) .
@xmath54 here @xmath55 is the rate of phase shift ( degrees / day ) , p is the corresponding period in days , t is time in days , @xmath56 is reference time , and @xmath57 is the reference spot longitude in degree .
after the epoch jd 2452400 ( or after the epoch d ) , the evolution of different spots seems to be more responsible than the migration of single group of spots , so , we have used only first three points of the fig .
[ phmin.fig ] ( i.e. the phase minima of epoch a , b and c ) in the linear least square fit of equation ( 1 ) .
the solid line in fig .
[ phmin.fig ] shows the linear least square fit of the equation ( 1 ) to the data .
the rate of phase shift for the group of spots was determined as @xmath58 degree / day , which corresponds to a migration period of @xmath1 years .
the smaller group of spots during the epoch e appears to rotate with the similar angular velocity with that of a group of spots during the epoch a , b and c ( see the dotted line in fig . [ phmin.fig ] ) .
however , long - term studies of spot groups evolution and migration in the star hd 81032 require much more frequent sampling in the time domain .
the arrangement of groups of spots in one permanent strip can be easily interpreted as a long lived active group rotating with the same velocity .
the open circles in the fig .
[ phmin.fig ] are not following that strip , and may be due to the formation of a new group of spots on the surface of the star .
the h@xmath2 and caii h & k emission lines are important indicators of chromospheric activity . in very active stars like
ii peg , v711 tau , ux ari the h@xmath2 emission is seen clearly above the continuum , whereas in less active stars only a filled in absorption line is observed .
figure [ halpha.fig ] shows spectra of hd 81032 in the h@xmath2 region .
spectrum of the star hd 71952 is shown for comparison .
h@xmath2 emission line can be seen clearly in the star hd 81032 ( fig .
[ halpha.fig ] ) .
it has been suggested that in the rs cvn systems at least one star must also show the intense emission in the h and k lines of caii ( fekel et al .
as shown in figure [ caiihk.fig ] , caii h and k lines are above the continuum at all the phases observed .
the measured equivalent widths ( ews ) of the h@xmath2 , caii h and k emission feature , the corresponding jd and phase are listed in table [ ew.tab ] .
the photometric observations during the epoch d were close to the spectroscopic observations .
figure [ splc.fig ] shows the plot of ews of ca ii h , k and h@xmath2 against phase , along with the v band light curve of the epoch d. it appears that h@xmath2 , ca ii h , and k emission features are variable and anti - correlated with the photometric phase i.e. the maximum at photometric minimum and minimum at photometric maximum .
this could be due to the presence of active cool spots on the surface of the star .
the value of the total galactic reddening e(b - v ) in the direction of hd 81032 is estimated to be 0.04 mag from the reddening map given by schlegel et al .
however , the star hd 81032 suffers negligible reddening at a distance of 140 pc , and we assume e(b - v ) = 0 for the remaining discussion . using the tycho parallax ( hog 1997 ) and the value of @xmath59 mag from the present photometry
the absolute magnitude @xmath60 of hd 81032 is 3.04 mag .
the value of @xmath60 is consistent with the luminosity class iv for this star .
the @xmath61 colour of hd 81032 derived from our photometry is @xmath62 consistent with the relatively precision to the value of @xmath63 given in the tycho catalogue ( wright et al .
the ( b - v ) colour is best matched with the spectral class k0iv , and is consistent with the spectral class identified by wright et al .
( 2003 ) . applying the bolometric correction -0.4 ( schmidt & kaller 1982 ) for k0iv star ,
the bolometric magnitude of hd 81032 is 2.65 mag .
we have determined the sed of hd 81032 using broad band ubvri ( present photometry ) , 2mass jhk ( cutri et al . 2003 ) and 12 , 25 , 60 and 100 @xmath64 iras ( moshir 1989 ) fluxes .
however , only upper limits are available at 25 , 60 and 100 @xmath65 . the observed sed of hd 81032 along with the synthetic sed is shown in figure [ sed.fig ] .
the synthetic sed is expected from the intrinsic properties of the star ( kurucz 1993 ) .
the model sed shown in fig .
[ sed.fig ] has been adjusted to coincide with the observed sed of the star hd 81032 at the v - band wavelength of 0.55 @xmath64 .
we have overplotted the synthetic sed for the different @xmath66 and @xmath67 combinations .
the values of @xmath66 and @xmath68 which best match the observed sed are @xmath69 k and @xmath70 , respectively , and are consistent within one subclass of the inferred k0iv spectral type of the star hd 81032 ( see [ physical.sec ] ) .
the continuum of hd 81032 is in good agreement with the normal values for the spectral type of k0iv up to jhk band but deviates in all bands longer than @xmath71 ( see fig .
[ sed.fig ] ) .
the difference between the observed and the expected ( j - k ) and ( h - k ) colours are @xmath72 and @xmath73 mag , respectively .
this indicates that there is no significant colour excess in j , h and k band .
the expected ( j - k ) and ( h - k ) colours were taken from the spectral type of the star ( koorneef , 1983 ) .
the 12 @xmath65 magnitude ( @xmath74 ) of the star hd 81032 was found to be @xmath75 mag using the relation @xmath76 ( mitrou et al .
the intrinsic @xmath77 $ ] colour of the star hd 81032 is determined to be @xmath78 mag .
the expected @xmath77 $ ] colour of 0.14 mag for the k0iv type star ( verma et al .
1987 ) gives the colour excess of 0.17 mag , which is consistent with the ideal value of @xmath79 mag ( mitrou et al .
the light curve for the source and background were extracted using the xselect package for the pspc 0.1 - 2.4 kev energy band which contains all the x - ray photons .
the background - subtracted x - ray light curve of hd 81032 is shown in figure [ xlc.fig ] plotted with a time bin size of 64 s. it appears from the light curve that a moderate flare occurred during the rass observations , with a peak of about 0.6 ct s@xmath16 ( compared to a pre - flare level of 0.2 - 0.3 ct s@xmath16 ) at approximately jd = 2448206.95 ( 1990/11/11 10:48:0.0 ) and a half - decay time of @xmath80 s. the x - ray spectrum of hd 81032 as observed with the rosat pspc is shown in figure [ xspec.fig ] .
response matrices based on the available off - axis calibration of the pspc and using the appropriate ancillary response files were created and data were fitted using the _ xspec _ ( version 11.3.1 ) spectral analysis package .
spectral models for thermal equilibrium plasma known as the mewe - kaastra - liedahl or mekal model ( liedahl , ostrecheld & goldstein 1995 ; mewe , kaastra & liedahl 1995 ) were used .
the background - subtracted x - ray spectra were fitted with 1-temperature ( 1 t ) and 2-temperature ( 2 t ) plasma models , either assuming solar photospheric abundances as given by anders & grevess ( 1989 ) or allowing the abundance of every element other than h to vary by a common factor relative to the solar ( photospheric ) values . in each of the above models
the interstellar absorption was assumed to follow the absorption cross - sections given by morrison & mccammon ( 1983 ) , and the total intervening hydrogen column density @xmath81 was allowed to vary freely .
the results of different model fits are summarized in table [ xdata.tab ] .
single - temperature mekal models with abundances fixed to the solar values gave unacceptably high values for @xmath82 , and thus can be rejected .
however , single - temperature plasma models with abundances of @xmath83 times solar and plasma temperature of @xmath84 were found acceptable .
alternatively , two - temperature mekal models with fixed solar abundances were also found to be acceptable ; the two temperatures being @xmath85 kev and @xmath86 kev . multi - temperature variable abundance models would also produce acceptable fits to the pspc spectrum , but the small number of accumulated counts does not warrant such complex models .
the pspc spectrum and the best - fit two - temperature plasma model with solar abundances are shown in fig .
[ xspec.fig ] along with the significance of the residuals in terms of their @xmath87 .
based on the best fit 2 t mekal model a source flux of @xmath88 erg @xmath22 s@xmath16 was obtained . at a distance of 140 pc the x - ray luminosity of hd 81032 is calculated to be @xmath89 , similar to that of known subgiants rs cvn binaries ( drake et al .
1989 , singh et al . 1995,1996 ; dempsey et al .
1997 ; padmakar et al .
2000 ; pandey et al .
the derived x - ray luminosity is about a factor of 2 bigger than the earlier estimate given in 1 .
this may be due to the non - trivial column density of @xmath90 @xmath91 inferred : objects with a high column density will typically have a higher count rate to flux conversion factor than the standard one which assumes a low column density .
the values of @xmath92 and @xmath93 of hd 81032 ( see table [ xdata.tab ] ) are consistent with those of other rs cvn systems studied by dempsey et al .
( 1997 ) , who found an average value of @xmath94 kev and @xmath95 kev for a sample of 28 rs cvn systems .
the volume emission measures em1 and em2 of hd 81032 are also found to be consistent with those of the other similar systems , where the average volume emission measures @xmath96 and @xmath97 are @xmath98 and @xmath99 @xmath100 , respectively ( dempsey et al .
1997 ) .
all the errors quoted in this section are internal statistical errors , and that the distance - dependent errors of 45% in luminosity which were mentioned in 1 have not been included in , for example , the em errors quoted in table 3 or the errors in the x - ray luminosity given the above paragraph of this section .
a photometric period of @xmath101d has been discovered in the star hd 81032 . the amplitude and the phase of the photometric light curves of hd 81032
are observed to be changing from one epoch to another .
the amplitude of variation is found to be maximum ( 0.29 mag ) during the observing year 2000 - 2001 .
the change in the amplitude is mainly due to a change in the minimum of the light curve , and this may be due to a change in the spot coverage on the surface of the star .
a single group of spots is found to migrate , and a migration period of @xmath102 years is determined from the first three years of the data . the star is seen to be redder at the light minimum and we interpret this is due to the relatively cooler temperature of the darker regions present in the visible hemisphere . a spectral type of k0iv is determined for hd 81032 .
the equivalent widths of h@xmath2 , ca ii h and k emission lines in the spectra of hd 81032 appear more intense close to the photometric light minimum .
this could be attributed to the presence of cool spots on the surface of the stars .
the observed x - ray spectrum and the inferred coronal plasma parameters for hd 81032 are typical of those seen in active stars such as rs cvn binaries .
all of the optical and x - ray properties found for hd 81032 are most consistent with it being an evolved rs cvn binary of the long - period type .
although there are a handful of apparently single evolved stars , e.g. , fk com , which can exhibit similar properties , identification of hd 81032 as such a rare object seems rather improbable . clearly , high - resolution optical spectroscopy of hd 81032 in order to confirm the presence of radial velocity variability and determine a spectroscopic binary orbit would be definitive evidence for its proposed binary nature .
we are grateful to the anonymous referee for valuable comments and suggestions .
we are thankful to the time allocation committee for giving time at 234-cm vbt .
this research has made use of data obtained from the high energy astrophysics science archive research center ( heasarc ) , provided by nasa s godard space flight center .
starlink is funded by pparc and based at the rutherford appleton laboratory , which is part of council for the central laboratory of the research councils , uk .
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 .
99 anders , e. , & grevesse , n. 1989 , geochim .
acta , 53 , 197 .
ayres , t. r. , et al . , 1995 , apjs , 96 , 223 cutri , r. m. , et al . , 2001 , 2mass all - sky catalog of point sources , vizier on - line data catalog ii/246 dempsey , r. c. , linsky , j. l. , fleming , t. a. , & schmitt , j. h. m. m. , 1997 , aj , 478 , 358 drake , s. a. private communication drake , s. a. , simon , j. , & linsky , j. l. , 1992 , apjs , 82 , 31 drake , s. a. , simon , j. , & linsky , j. l. , 1989 , apjs , 71 , 905 drake , s. a. , & linsky , j. l. , 1986 , aj , 91 , 1229 elvis , e. , plummer , d. , schachter , j. & fabbiano , g. , 1992 apjs , 80 , 257 fekel , f. c. , moffet , t. j. , & hennry g. w. , 1986 , apjs , 60 , 551 .
gdel , m. , 2002 , araa , 40 , 217 .
hog , e. , et al .
, 1997 , a&a 323 , l57 .
houk , n. & smith - moore , m. 1988 , michigan catalogue of two - dimensional spectral types for the hd stars , vol . 4 : -26 to -12 degrees .
koorneef , j. , 1983 , a&a , 128 , 84 .
kurucz , r. l. , 1993 , private communication landolt , a. u. , 1992 , aj , 104 , 340 liedahl , d. a. , osterheld , a. l. , & goldstein , w. h. , 1995 , apj , 438 , l115 mewe , r. , kaastra , j. s. , & liedahl , d. a. , 1995 , legacy , 6 , 16 mitrou , c. k. , doyle , j. g. , mathioudakis , m. & antonopolou , 1996 , a&as , 115 , 61 .
morrison , r. , & mccammon , d. , 1983 , apj , 270 , 119 moshir , m. , 1989 , explanatory supplement to iras faint source survey padmakar , p. s. , singh , k. p. , drake , s. a. , & pandey , s. k. , 2000 , mnras , 314 , 733 padmakar , p. s. , & pandey , s. k. , 1999 , a&as , 138 , 203 .
pandey , j. c. , singh , k. p. , drake , s. a. & sagar , r. , 2005 , aj , 130 , 1231 pandey , j. c. , singh , k. p. , sagar , r. & drake , s. a. , 2002 , japa , 23 , 9 .
pfeffermann , e. , et al .
, in soft x - ray optics and technology , proc .
spie , 733,519 raveendran , a. v. , & mohin , s. , 1995 , a&a , 301 , 788 .
roberts , d. h. , lehar , j. & dreher , j. w. , 1987 , aj , 93 , 968 p.493 schachter , j. f. , et al .
1996 , apj .
, 463 , 747 schlegel , d. j. , finkbeiner , d. p. , & davis , m. , 1998 , apj , 500 , 525 schmidt - kaler th .
, 1982 , in : landolt - bornstein , numerical data and functional relationship in science and technology , new series , group vi , vol . 2b , scaifers k. & voigt h. h. ( eds . )
springer - verlog , berlin schmitt , j. h. m. m. & liefke , c. 2004 , a&a , 417 , 651 .
singh , k. p. , drake , s. a. , & white , n. e. , 1995 , apj , 445 , 840 singh , k. p. , drake , s. a. , white , n. e. & simon , t. , 1995 , aj , 112 , 221 trmper , r. a. 1983 , adv .
space res , 2 , 241 .
ulvs v. aarum & henry , g. w. , 2003 , a&a , 402 , 1033 .
verma , r. p. , iyenger , k. v. k. , & rengarajan , t. n. , 1987 , a&a , 177 , 346 voges , w. et al . , 1999 ,
a&a , 349 , 389 .
wright , c. o. , egan , m. p. , kraemer , k. e. , & price , s. d. , 2003 , aj 125 , 359 .
a&a , 172 , 225 . | bvr photometric and quasi - simultaneous optical spectroscopic observations of the star hd 81032 have been carried out during the years 2000 - 2004 .
a photometric period of @xmath0 d has been detected for this star .
a large group of spots with a migration period of @xmath1 years is inferred from the first three years of the data .
h@xmath2 and ca ii h and k emissions from the star indicate high chromospheric activity . the available photometry in the bvrijhk bands is consistent with spectral type of k0 iv previously found for this star .
we have also examined the spectral energy distribution of hd 81032 for the presence of an infrared colour excess using the 2mass jhk and iras photometry , but found no significant excess in any band above the normal values expected for a star with this spectral type .
we have also analyzed the x - ray emission properties of this star using data obtained by the rosat x - ray observatory during its all - sky survey phase .
an x - ray flare of about 12 hours duration was detected during the two days of x - ray coverage obtained for this star .
its x - ray spectrum , while only containing 345 counts , is inconsistent with a single - temperature component solar - abundance coronal plasma model , but implies either the presence of two or more plasma components , non - solar abundances , or a combination of both of these properties .
all of the above properties of hd 81032 suggest that it is a newly identified , evolved rs cvn binary .
-0.4 cm 0.3 cm * unravelling the nature of hd 81032 - a new rs cvn binary * j. c. pandey@xmath3 , k. p. singh@xmath4 , s. a. drake@xmath5 , and r. sagar@xmath6 @xmath3 aryabhatta research institute of observational sciences , naini tal - 263 129 , india @xmath7 tata institute of fundamental research , mumbai - 400 005 , india @xmath8 usra & code 662 , nasa / goddard space flight center , greenbelt , md 20771 , usa _ keyword : _ star - x - ray ; star - variable ; star - late - type ; star - hd 81032 |
let @xmath7 denote the random @xmath3-uniform hypergraph with vertex set @xmath9:=\{1,2, ...
,n\}$ ] , where each of the @xmath10 potential ( hyper)edges of cardinality @xmath3 is present with probability @xmath11 , independently of all other potential edges .
likewise , let @xmath6 be the random @xmath3-uniform hypergraph on @xmath9 $ ] , where @xmath12 edges are chosen uniformly at random among all sets of @xmath12 potential edges .
the model @xmath6 can be gainfully viewed as a snapshot of the random hypergraph process @xmath13 where @xmath14 is obtained from @xmath15 by inserting an extra edge chosen uniformly at random among all @xmath16 remaining potential edges . for @xmath17 ,
these models are the typical random graph models , @xmath18 , @xmath19 and @xmath20 as customary , we say that for a given @xmath21 ( @xmath11 resp . )
some graph property @xmath22 holds _ with high probability _ , denoted _
_ , if the probability that @xmath23 ( @xmath24 resp . )
has property @xmath22 tends to 1 as @xmath25 .
further , @xmath26 is the _ sharp threshold _ for @xmath22 if for each @xmath27 ( fixed ) , w.h.p .
@xmath28 does not have @xmath22 and w.h.p @xmath29 does have @xmath22 .
for the random hypergraph process , the _ stopping time _ of @xmath22 , denoted
@xmath30 is the first moment that the process has this property @xmath22 ; we denote the hypergraph process stopped at this time by @xmath31 . in one of the first papers on random graphs , erds and rnyi @xcite showed that @xmath32 is a sharp threshold for connectivity in @xmath33 .
later , stepanov @xcite established the sharp threshold of connectivity for @xmath34 among other results .
more recently , bollobs and thomason @xcite proved the stronger result for the random graph process that w.h.p .
the moment the graph process loses its last isolated vertex is also the moment that the process becomes connected ; in other words , w.h.p .
@xmath35 . we prove the analogous result for the the random @xmath3-uniform hypergraph process ; a consequence of this result is that @xmath36 is a sharp threshold of connectivity for @xmath23 .
there are various measures for the _ strength _ of connectedness of a connected graph , but here we will focus on @xmath2-(vertex-)connectivity . for @xmath37 , a hypergraph with more than @xmath2 vertices
is @xmath2-_connected _ if whenever @xmath38 vertices are deleted , along with their incident edges , the remaining hypergraph is connected .
note that the definition of 1-connectedness coincides with connectedness .
necessarily , for a hypergraph to be @xmath2-connected , each vertex must have degree at least @xmath2 , because if a vertex @xmath39 has degree less than @xmath2 , then we can delete a neighbor from each incident edge to isolate @xmath39 .
however , as commonly seen in these types of results , the main barrier to @xmath2-connectivity in these random graph models arises from such vertices that can be separated from the rest of the graph by the deletion of their neighbors ( see for instance erds - rnyi @xcite , ivchenko @xcite , bollobs @xcite,@xcite ) . here
, we extend this idea to random @xmath3-uniform hypergraphs ; in particular , we find that if @xmath40 and @xmath41 , @xmath42 , then w.h.p .
@xmath43 is not @xmath2-connected and w.h.p .
@xmath44 is @xmath2-connected ; also we find an analogous threshold value for @xmath45 . a stronger result concerns the random graph process where edges are added one after another .
let @xmath46 and @xmath47 ; note that @xmath48 .
in @xcite , bollobs and thomason showed that for @xmath17 ( the graph case ) and any @xmath0 $ ] , @xmath49 .
we extend this result for @xmath3-uniform random hypergraphs albeit for _ fixed _ @xmath3 and @xmath2 . [
theorem : 1.25 ] w.h.p , at the moment the @xmath3-uniform hypergraph process loses its last vertex with degree less than @xmath2 , this process becomes @xmath2-connected .
formally , for @xmath50 and @xmath51 ( both fixed ) , @xmath52 . to prove this result
, we begin by determining the likely range of @xmath53 , and further that just prior to this window , at some @xmath54 edges , w.h.p .
there are not many vertices of degree less than @xmath2 .
then , we prove that w.h.p .
@xmath55 is almost @xmath2-connected in the sense that whenever @xmath38 vertices are deleted , there is a massive component using almost all leftover vertices .
third , we show that w.h.p . to isolate a vertex of @xmath56 you would have to delete at least @xmath2 of its neighbors ( this is trivially true for graphs , but not so for @xmath50 ) .
in particular , we show that w.h.p .
edges incident to degree @xmath38 vertices have trivial intersection ( just the vertex itself ) .
finally , we show that the probability that @xmath57 , but these three previous likely events also hold tends to zero , which completes the proof of the theorem .
the following corollary is nearly immediate in light of the theorem .
[ cor : 1.25 ] * ( i ) * let @xmath58 , where @xmath59 . w.h.p
. @xmath6 is @xmath60-connected , but not @xmath61-connected .
further , the probability that @xmath6 is @xmath2-connected tends to @xmath62 .
+ * ( ii ) * let @xmath63 , where @xmath59 .
@xmath7 is @xmath60-connected , but not @xmath61-connected .
further , the probability that @xmath7 is @xmath2-connected tends to @xmath62 . for the remainder of this paper ,
let @xmath50 and @xmath51 be fixed numbers .
[ lemma : 6.5.1 ] let @xmath64 , but @xmath65 , @xmath66 and @xmath67
. then w.h.p . , *
( i ) * the minimum degree of @xmath55 is @xmath38 and the number of vertices with degree @xmath38 is in the interval @xmath68.\ ] ] * ( ii ) * there are no vertices of degree @xmath38 in @xmath44 . consequently , w.h.p
. @xmath69 $ ] .
we prove that the number of vertices with degree @xmath38 , denoted by @xmath70 , is in the interval by chebyshev s inequality .
note that a given vertex can be in @xmath71 possible edges , so @xmath72 = n p[\text{deg}(1 ) = k-1 ] = n \frac { { { n-1 \choose d-1 } \choose k-1 } { { n \choose d}- { n-1 \choose d-1 } \choose m_0 - k + 1 } } { { { n \choose d } \choose m_0}}.\ ] ] here and elsewhere in this paper , we use the identity @xmath73 , where @xmath74 and later , we use the inequality @xmath75 .
now @xmath72 = ( 1+o(1/n ) ) \frac{n\cdot n^{(d-1)(k-1)}}{(k-1 ) ! ( ( d-1)!)^{k-1 } } \left ( \frac{d ! \ , m_0}{n^d } \right)^{k-1 } \frac { { { n \choose d } - { n-1 \choose d-1 } \choose m_0}}{{{n \choose d } \choose m_0}}.\ ] ] this latter fraction can be sharply approximated .
@xmath76 \right ) \nonumber \\ & = \left ( 1 + o\left(\frac{\ln n}{n } \right ) \right ) \exp \left ( - \frac{d m_0}{n } \right ) = \left ( 1 + o\left(\frac{\ln n}{n } \right)\right ) \frac{e^\omega}{n ( \ln n)^{k-1}}.\end{aligned}\ ] ] hence @xmath77 & = \left ( 1 + o\left ( \frac{\ln n}{n } \right)\right ) \frac{e^\omega}{(k-1 ) ! } \left ( \frac { d m_0/n } { \ln n } \right)^{k-1 } \\ & = \left ( 1 + o\left ( \frac{\ln \ln n}{\ln n } \right)\right ) \frac{e^\omega}{(k-1)!}.\end{aligned}\ ] ] for the second factorial moment , we have that @xmath78 = n ( n-1 ) p(\text{deg}(1)=\text{deg}(2 ) = k-1).\ ] ] we break this latter probability over @xmath79 , the number of edges that include both vertices @xmath80 and @xmath81 . in particular ,
vertex 1 is in @xmath82 edges that do not contain vertex @xmath81 and vice versa ; further there are @xmath83 edges that include neither vertex 1 or 2 .
since there are @xmath84 potential hyperedges containing both vertices and @xmath85 potential hyperedges containing one vertex but not the other , we have that @xmath86 just as in , we can estimate this latter fraction @xmath87 using these asymptotics , one can show that the @xmath79th term in is on the order of @xmath88 .
in particular , the sum of the terms over @xmath89 $ ] , is @xmath90 .
therefore @xmath91 whence @xmath78 = \left ( 1 + o\left ( \frac{\ln \ln n}{\ln n}\right)\right ) e[x]^2.\ ] ] consequently , @xmath92 = e[x ] + o\left ( ( e[x])^2 \frac{\ln \ln n}{\ln n } \right).\ ] ] by chebyshev s inequality , @xmath70 is concentrated around its mean and in particular , w.h.p .
@xmath70 is in the interval .
to finish the proof of part * ( i ) * , it remains to show that w.h.p .
there are no vertices of degree less than @xmath38 , which can be done by a first moment argument using similar techniques to the asymptotics of @xmath93 $ ] .
similarly , for part * ( ii ) * , one can easily show that the expected number of vertices of degree @xmath38 in @xmath44 tends to zero as well .
now we will establish that w.h.p .
@xmath55 is almost @xmath2-connected in the sense that if @xmath38 vertices are deleted , then there remains a massive component containing almost all left - over vertices . to this end , we prove an analogous statement for the random bernoulli hypergraph @xmath7 and use a standard conversion lemma to obtain the desired result for @xmath55 . in this next lemma , we pick a specific version of @xmath54 , one where @xmath94 .
[ lemma : 7.28.1 ] let @xmath95 and @xmath96 . with high probability , * ( i ) * @xmath7 has the property whichever @xmath38 vertices are deleted , there remains a giant component which includes all but up to @xmath97 leftover vertices . " * ( ii ) * @xmath98 has the property
whichever @xmath38 vertices are deleted , there remains a giant component which includes all but up to @xmath97 leftover vertices . " * ( i ) * given a set of @xmath38 vertices , @xmath99 , let @xmath100 be the event that if the vertices @xmath101 are deleted from @xmath7 along with their incident edges , then there remains no components of size at least @xmath102 . in particular , we wish to show that w.h.p .
@xmath7 is not in @xmath100 for any @xmath101 . using the union bound over all @xmath38 element sets of @xmath9 $ ] as well as symmetry , we find that @xmath103 where @xmath104 .
note that the remaining hypergraph left after deleting @xmath105 from @xmath7 is distributed as @xmath106 ( this is the primary reason that we consider the bernoulli hypergraph @xmath7 rather than @xmath6 ) .
therefore @xmath107 is precisely the probability that @xmath108 does not have a component of size at least @xmath109 . to bound @xmath107
, we note that any hypergraph on @xmath110 vertices without a component of size at least @xmath111 has a set of vertices @xmath112 such that there are no edges between @xmath112 and @xmath113 \setminus s$ ] where @xmath114.$ ] to see this fact , consider a hypergraph @xmath115 on @xmath110 vertices without such a large component and let @xmath116 be the vertex sets of the components of @xmath115 in increasing order by their cardinalities .
then there is some minimal @xmath117 so that @xmath118 further , @xmath119 is not empty and since @xmath120 , we have that @xmath121 and @xmath122 clearly , there are no edges including a vertex of @xmath123 and @xmath113 \setminus s$ ] . therefore , @xmath124 , |s|=s , \text { no edge between } s\text { and } [ n']\setminus s),\ ] ] and by symmetry over all such vertex sets @xmath112 , @xmath125\text { and } [ n ' ] \setminus [ s]).\ ] ] further , this latter probability is symmetric about @xmath126 ( i.e. the probabilities corresponding to @xmath127 and @xmath128 are equal ) . hence @xmath129\text { and } [ n ' ] \setminus
[ s]).\ ] ] the number of potential edges that contain at least one vertex from @xmath130 $ ] and at least one vertex from @xmath113\setminus[s]$ ] is @xmath131 .
hence @xmath132 where @xmath133 and @xmath134 are the sums over @xmath135 $ ] and @xmath136 $ ] respectively .
we begin with analyzing @xmath134 since these bounds will be cruder and simpler .
trivially , @xmath137 now let s take on these binomial coefficient terms .
trivially @xmath138 and so @xmath139 further , the function @xmath140 is decreasing for @xmath141 , so we have that @xmath142 now our bound in becomes @xmath143 in the previous expression , the leading order terms in the first and third terms will cancel and the middle term is absorbed in the error .
namely , we have that @xmath144 or equivalently @xmath145 now let s take on the sum @xmath133 .
we begin with taking the leading order terms in the exponent . @xmath146
uniformly over @xmath147 , we have that @xmath148 consequently , the exponent in is @xmath149 whence there is some ( fixed ) @xmath150 such that for all @xmath147 , @xmath151 using the bound @xmath152 as well as , becomes @xmath153 however , for @xmath147 , we have that @xmath154 therefore @xmath155 this sum is dominated by the first term ( @xmath156 ) because ratios of consecutive terms uniformly tend to zero .
consequently , we have that @xmath157.\ ] ] summing our bounds for @xmath133 and @xmath134 ( and , respectively ) , we have that @xmath158 which most definitely is @xmath159 and so by the bound , part * ( i ) * of the lemma is proved .
part * ( ii ) * is established by using a standard conversion technique between @xmath6 and @xmath7 .
for any hypergraph property @xmath160 , we have that @xmath161 where @xmath162 is the number of edges of @xmath115 .
therefore , for any ( possible ) @xmath12 , @xmath163 whence @xmath164 for @xmath165 and @xmath166 , one can show that @xmath167 hence in our case , @xmath168 in the proof of part * ( i ) * , we found that this latter probability tends to zero superpolynomially fast .
for the random graph process ( @xmath17 ) , it was found that the main barrier to @xmath2-connectivity is the presence of vertices of degree less than @xmath2 , which could be isolated with the deletion of their neighbors ( see erds - rnyi @xcite , ivchenko @xcite , bollobs @xcite,@xcite ) .
we will find a similar situation for the random hypergraph process .
however , we run into an additional issue here for hypergraphs .
even if the degree of a vertex @xmath39 is @xmath2 , we could isolate @xmath39 with the deletion of less than @xmath2 vertices .
for instance , if all of @xmath39 s edges also include vertex @xmath169 , then the deletion of just @xmath169 from the hypergraph ( along with its incident edges ) will isolate @xmath39 from the rest of the hypergraph .
our ultimate goal in this section is to show that w.h.p .
each vertex of @xmath170 has at least @xmath2 edges whose pairwise intersections are precisely @xmath171 ; in this case , for any vertex , you would need to delete at least @xmath2 of its neighbors to isolate it . to this end
, we first prove that w.h.p .
@xmath55 has this property for vertices with degree at least @xmath2 and as nearly as could be expected for vertices with degree @xmath38 .
a set of edges @xmath172 incident to vertex @xmath39 is _ quasi - disjoint _ if all pairwise intersections of these edges are @xmath173 formally , if @xmath174 , then @xmath175 [ lemma : 6.5.2 ] let @xmath176 , where @xmath177 but @xmath178 .
, @xmath55 is such that * ( i ) * the incident edges of a degree @xmath38 vertex form a quasi - disjoint set , * ( ii ) * vertices with degree at least @xmath2 have a quasi - disjoint set of incident edges with size at least @xmath2 . note that both parts of this lemma are trivially true for @xmath179 and part * ( i ) * is also trivially true for @xmath180 . let @xmath181 be the number of vertices whose maximum quasi - disjoint set has size @xmath117 and whose degree is @xmath182 . to prove this lemma
, it suffices to show that w.h.p . for @xmath183 and @xmath184 , we have that @xmath185 , which is shown by a first moment argument .
now @xmath186 = n p(j , \ell),\ ] ] where @xmath187 is the probability that a generic vertex @xmath39 has a maximum quasi - disjoint set of size @xmath117 and whose degree is @xmath188 . to bound this probability , note that @xmath39 has a set of @xmath117 quasi - disjoint edges and each of the remaining @xmath189 edges must have at least one vertex from the @xmath190 neighbors from the quasi - disjoint edges ; further the remaining @xmath191 edges do not include @xmath39 . hence @xmath192 we gave sharp asymptotics for the last fraction in . here and throughout the rest of the paper , we will use @xmath193 for @xmath194 when the formula for @xmath195 becomes too bulky .
therefore @xmath196 whence @xmath197 for @xmath198 ( independent of @xmath183 and @xmath184 ) .
thus @xmath199 \leq_b e^{\omega } \sum_{\ell \geq 1 } \left ( c \ , \frac{\ln n}{n } \right)^{\ell } \leq_b e^{\omega } \frac{\ln n}{n } \to 0,\ ] ] which completes the proof of the lemma .
[ cor : 7.28 ] w.h.p .
each vertex of @xmath170 has a quasi - disjoint set of incident edges with size at least @xmath2 .
this lemma is trivially true for @xmath179 .
suppose that @xmath200 .
let @xmath201 be the event that @xmath170 has a vertex that does _ not _ have a quasi - disjoint set of edges with size at least @xmath2 ; we wish to show that @xmath202 .
let @xmath203 be as defined in lemma [ lemma : 6.5.1 ] .
we have proved that w.h.p .
@xmath69 $ ] and that @xmath43 does not have vertices of degree less than @xmath38 .
further , w.h.p .
the number of degree @xmath38 vertices in @xmath55 is less than @xmath204 ( lemma [ lemma : 6.5.1 ] ) .
in addition , w.h.p .
@xmath55 has the two properties of the previous lemma ( lemma [ lemma : 6.5.2 ] ) .
let @xmath205 be the intersection of these four likely events . to prove the lemma , it suffices to show that @xmath206 .
let @xmath207 be the vertex set of vertices of degree @xmath38 in @xmath55 .
note that @xmath208 , \ , |v_0| \leq 3 e^{\omega}/(2(k-1 ) ! ) } p(a_n\cap b_n \cap \{\tilde{v}_0 = v_0\}).\end{aligned}\ ] ] on the event that @xmath201 and @xmath205 occur and @xmath209 , necessarily some edge @xmath210 is added in the hypergraph process at some step @xmath211 $ ] such that @xmath210 includes both a vertex @xmath212 and one of @xmath213s @xmath214 neighbors in @xmath55 .
thus @xmath215 therefore @xmath216 , \ , |v_0| \leq 3 e^{\omega}/(2(k-1 ) ! ) } p(\tilde{v}_0=v_0 ) \leq \frac{\omega e^{\omega}}{n } \to 0.\ ] ]
now that we have sufficient knowledge about the structure of @xmath55 and low - degree vertices in @xmath170 , we can prove our main theorem .
@xmath170 is @xmath2-connected . in short ,
w.h.p . @xmath217 .
let @xmath218 , for @xmath219 by lemma [ lemma : 6.5.1 ] , we have shown that w.h.p .
@xmath220 $ ] and by lemma [ cor : 7.28 ] , each vertex of @xmath170 has a quasi - disjoint edge set of size at least @xmath2 .
further , the property whichever @xmath38 vertices are deleted , there remains a giant component which includes all but up to @xmath97 leftover vertices , " denoted @xmath22 , is an increasing property ( closed under the addition of edges ) .
therefore , by lemma [ lemma : 7.28.1 ] , @xmath170 has property @xmath22 as well . to prove this theorem
, it suffices to show that the probability that these three likely events hold yet @xmath57 tends to zero . to this end , for @xmath221 $ ] , let @xmath222 be the event that @xmath6 is not @xmath2-connected , but each vertex has a quasi - disjoint edge set of size at least @xmath2 and @xmath6 has property @xmath22 . to prove this theorem , it suffices to prove that @xmath223 .
we will in fact show that @xmath224 uniformly over @xmath221 $ ] . in this case , @xmath225 as desired .
all that remains is to prove the bound . on the event @xmath222
, there are @xmath38 vertices , @xmath226 such that upon their deletion , there is a component of size @xmath128 for some @xmath227 .
in fact , since each remaining vertex must have at least one incident edge , we must have that @xmath228 .
let @xmath112 be the set of vertices not in this large component . by the union bound over all @xmath38 element sets of @xmath9 $ ] and sets @xmath112 , @xmath229 , as well as symmetry
, we have that @xmath230 where @xmath231 is the probability that each vertex of @xmath23 has a quasi - disjoint edge set of size at least @xmath2 and that after the deletion of @xmath232 from @xmath6 , the vertices @xmath233 $ ] form a component ; in this case , @xmath234 .
we now turn to showing that @xmath231 tends to zero sufficiently fast .
suppose that @xmath115 is some hypergraph in the event corresponding to @xmath231 .
after the deletion of the @xmath38 vertices , we know that vertex @xmath235 from @xmath112 has at least one incident edge , which necessarily must reside completely within @xmath112 .
further , before deletion , any incident edge to @xmath169 must be completely contained within @xmath112 or this edge must contain one of the @xmath38 to - be - deleted vertices
. moreover , there are at least @xmath2 edges incident to @xmath169 before the deletion .
therefore @xmath236 where @xmath237 is the corresponding probability to when there are ( at least ) @xmath79 incident edges to @xmath169 contained within @xmath112 and ( at least ) @xmath238 incident edges to @xmath169 that contain at least one of the to - be - deleted vertices . to bound the number of hypergraphs contributing to @xmath237 , we choose @xmath79 potential edges within @xmath112 containing @xmath169 , @xmath238 potential edges that include @xmath169 and at least one to - be - deleted vertex
; then we choose the remaining @xmath239 edges among all potential edges except those that include a vertex of @xmath112 and @xmath240 vertices of @xmath233 $ ] ( which necessarily can not be present ) .
note that these last chosen edges can include @xmath169 as well .
therefore @xmath241 first , we use trivial bounds on the first two binomial terms .
then we use the inequality @xmath242 .
namely , note that @xmath243 further , for @xmath244 , we have that @xmath245 hence @xmath246 which no longer depends on @xmath79 .
therefore @xmath247 and @xmath248 now taking on this sum , we find that @xmath249 and we find that @xmath250 , as desired .
as a consequence of theorem [ theorem : 1.25 ] , for any @xmath12 , we have that @xmath251 * ( i ) * let @xmath252 , where @xmath59 .
@xmath6 is @xmath60-connected , but not @xmath61-connected .
further the probability that @xmath6 is @xmath2-connected tends to @xmath62 .
+ * ( ii ) * let @xmath253 , where @xmath59 . w.h.p .
@xmath7 is @xmath60-connected , but not @xmath61-connected .
further the probability that @xmath7 is @xmath2-connected tends to @xmath62 . *
( i ) * first , note that w.h.p .
@xmath254 and @xmath255 by lemma [ lemma : 6.5.1 ] . therefore , by theorem [ theorem : 1.25 ] , w.h.p .
@xmath6 is @xmath60-connected , but not @xmath61-connected . in the lemma following this proof , we show that @xmath70 , the number of vertices of degree @xmath38 in @xmath6 is asymptotically poisson with parameter @xmath256 .
thus @xmath257 using the equation finishes off the proof . *
( ii ) * this part will be proved from * ( i ) * using a standard conversion technique similar to the one used in lemma [ lemma : 7.28.1 ] . since the number of edges in @xmath258 denoted @xmath259 ,
is binomially distributed on @xmath260 trials with success probability @xmath11 , we have that @xmath261 therefore , if @xmath262 where @xmath263 , then w.h.p .
using this fact , becomes @xmath265 notice that @xmath266 . by part *
( i ) * , if @xmath160 is @xmath267 or @xmath268 , then @xmath269 ; also , if @xmath270 , then @xmath271 .
to finish off the proof that @xmath272 has the same limits , we will use the fact that these properties are monotone . now an _ increasing _
( _ decreasing _ )
property is a property that is closed under the addition ( deletion resp . ) of edges . for any increasing property @xmath160
, we have that @xmath273 for any @xmath274 .
this fact is obvious when you consider @xmath23 and @xmath275 to be snapshots of the random hypergraph process @xmath276 .
moreover , if @xmath160 is an increasing property , then @xmath277 further , if @xmath160 is decreasing , then the inequalities above are reversed .
consequently , for a monotone property @xmath160 such that both @xmath278 tend to the same number , then @xmath272 does as well .
we prove this lemma using the method of moments ( see @xcite for a description of this method ) . in order to prove the lemma , it suffices to show that for each @xmath280 ( fixed ) , @xmath281 = \left ( \frac{e^{-c}}{(k-1)!}\right)^r.\ ] ] to compute the @xmath282th factorial moment , note that @xmath283 = e_0 + e_1 + \ldots + e_{(k-1)r},\ ] ] where @xmath284 is the expected number of ordered @xmath282-tuples of vertices of degree at most @xmath38 such that there are exactly @xmath285 edges containing at least one of these vertices .
we will see that the terms other than @xmath286 are negligible .
let s first consider @xmath286 .
since the number of edges is @xmath287 , each of these @xmath282 vertices have degree @xmath38 and must necessarily not be adjacent ; so @xmath288 taking on this last factor , we have that @xmath289 by sharply approximating this last fraction , we have that @xmath290 note that @xmath291 we also have @xmath50 so that @xmath292 similarly , we have that @xmath293 further @xmath294 and @xmath295 thus @xmath296 now we turn to @xmath284 , for @xmath297 . note that @xmath284 is less than the expected number of such @xmath282-tuples where these @xmath282 vertices have exactly @xmath298 adjacent edges ( we dropped the degree condition ) .
then , for @xmath297 , we have that @xmath299 in particular , we have that @xmath300 using the fact that @xmath301 along with the bound , we have that @xmath302 } \left ( \frac{\ln n}{n^{d-1 } } \right)^{(k-1)r - j } \left ( \frac{1}{n ( \ln n)^{k-1}}\right)^r = \frac{1}{(\ln n)^j},\ ] ] which completes the proof of the lemma . | bollobs and thomason ( 1985 ) proved that for each @xmath0 $ ] , with high probability , the random graph process , where edges are added to vertex set @xmath1 $ ] uniformly at random one after another , is such that the stopping time of having minimal degree @xmath2 is equal to the stopping time of becoming @xmath2-(vertex-)connected .
we extend this result to the @xmath3-uniform random hypergraph process , where @xmath2 and @xmath3 are fixed .
consequently , for @xmath4 and @xmath5 , the probability that the random hypergraph models @xmath6 and @xmath7 are @xmath2-connected tends to @xmath8 * keywords : * random hypergraph ; vertex connectivity |
a surface - plasmon - polariton ( spp ) is a coupled state between electromagnetic ( em ) field and electron plasma oscillations at the interface between a metal and a dielectric , for which the em field could be confined in subwavelength dimensions normal to the surface of metal .
consequently , metallic cavities supporting spp modes have been used to realize spp lasers ( also known as plasmonic lasers or spasers ) with subwavelength dimensions @xcite .
the energy in a spaser can remain confined as coherent spps or it can be made to leak out from the spaser as radiation . in many targeted applications in integrated optics and nanophotonics , spasers are developed as nanoscale sources of coherent electromagnetic ( em ) radiation and show interesting properties such as ultrafast dynamics for applications in high - speed optical communications .
parallel - plate metallic cavities supporting spp modes are also utilized for terahertz quantum - cascade lasers ( qcls ) @xcite to achieve low - threshold and high - temperature performance @xcite owing to the low - loss of spp modes at terahertz frequencies that are much smaller than the plasma frequency in metal .
the most common type of plasmonic lasers with long - range spps , which include terahertz qcls , utilize fabry - prot type cavities in which at least one dimension is longer than the wavelength inside the dielectric @xcite .
one of the most important challenges for such plasmonic lasers is their poor coupling to free - space radiation modes owing to the subwavelength mode confinement in the cavity , which leads to small radiative efficiency and highly divergent radiation patterns .
this problem is also severe for terahertz qcls based on metallic cavities and leads to low - output power and undesirable omnidirectional radiation patterns from fabry - prot cavities @xcite .
a possible solution to achieve directionality of far - field emission from spasers is to utilize periodic structures with broad - area emission , which has been used for both short - wavelength spasers @xcite as well as terahertz qcls @xcite . on chip phased - locked arrays @xcite or metasurface reflectors composed of multiple cavities @xcite
have also been utilized for directional emission in terahertz qcls . however , edge - emitting fabry - prot cavity structures with narrow cavity widths are more desirable , especially for electrically pumped spasers to achieve small operating electrical power and better heat removal from the cavity ( along the width of the cavity in lateral dimension , through the substrate ) for continuous - wave ( cw ) operation . in this paper , we theoretically and experimentally demonstrate a new technique for implementing distributed - feedback ( dfb ) in plasmonic lasers with fabry - prot cavities , which is termed as an _ antenna - feedback _ scheme .
this dfb scheme has no resemblance to the the multitude of dfb methods that have been conventionally utilized for semiconductor lasers .
the key concept is to couple the guided spp mode in a spaser s cavity to a single - sided spp mode that can exist in its surrounding medium by periodic perturbation of the metallic cladding in the cavity .
such a coupling is possible by choosing a bragg grating of appropriate periodicity in the metallic film .
this leads to excitation of coherent single - sided spps on the metallic cladding of the spaser , which couple to a narrow - beam in the far - field .
the narrow beam emission is due in part to the cavity acting like an end - fire phased - array antenna at the microwave frequencies , as well as due to the large spatial extent of a coherent single - sided spp mode that is generated on the metal film as a result of the feedback scheme .
experimentally , the antenna - feedback method is implemented for terahertz qcls for which the method is shown to be an improvement over the recently developed third - order dfb scheme for producing directional beams @xcite since it does not require any specific design considerations for phase - matching @xcite .
the emitted beam is more directional and the output power is also increased due to increased radiative field by virtue of this specific scheme .
single - mode operation in spasers with fabry - prot cavities could be implemented in a straightforward manner by periodically perturbing the metallic film that supports the resonant spp modes .
the schematic in fig .
[ fig1](a ) shows an example of a periodic grating in the top metal cladding for a parallel - plate metallic cavity that could be utilized to implement conventional @xmath1-th order dfb by choosing the appropriate periodicity .
since the spp mode has maximum amplitude at the interface of metal and dielectric active medium , a periodic perturbation in the metal film could provide strong bragg diffraction up to high - orders for the counter - propagating spp waves inside the active medium with incident and diffracted wavevectors @xmath2 and @xmath3 respectively , such that @xmath4 where @xmath5 is the grating period , @xmath6 is the grating wavevector , and @xmath1 is an integer ( @xmath7 ) that specifies the diffraction order . for plane - wave like modes at frequencies far away from the plasma resonance in metal , @xmath8 , where @xmath9 is the free - space wavelength corresponding to the spp mode and @xmath10 is the effective propagation index in active medium ( approximately the same as refractive index of the medium ) , the so - called bragg mode with @xmath11 is resonantly excited because it is , by design , the lowest - loss mode in the dfb cavity within the gain spectrum of the active medium . for terahertz qcls with metallic cavities , first - order @xcite and second - order @xcite dfbs have been implemented to achieve robust single - mode operation .
however , these conventional dfb techniques do not achieve directionality of far - field radiation in both directions .
there is phase mismatch for spp waves on either side of metal claddings and destructive interference between successive apertures , as shown in fig .
[ fig1](b ) for propagating spp waves .
therefore , no coherent single - sided spp waves can be established on the metallic cladding in the surrounding medium as demonstrated in fig .
[ fig1](c ) .
consequently , 2d photonic - crystal dfb structures have been utilized for broad - area ( surface ) single - mode emission @xcite for which diffraction - limited beams could be achieved at the expense of large cavity dimensions .
in contrast to conventional dfb methods in which periodic gratings couple forward and backward propagating waves inside the active medium itself , the antenna - feedback scheme couples a single - sided spp wave that travels in the surrounding medium with the spp wave traveling inside the active medium as illustrated in fig . [ fig1](d ) .
the spp wave inside the active medium with incident wavevector @xmath8 is diffracted in the opposite direction in the surrounding medium with wavevector @xmath12 .
for the first - order diffraction grating ( @xmath13 ) , equation results in @xmath14 that leads to excitation of a dfb mode with @xmath15 , which is different from any of the @xmath1-th order dfb modes that occur at @xmath11 .
hence , the antenna - feedback mode could always be excited by just selecting the appropriate grating period @xmath5 such that the wavelength occurs close to the peak - gain wavelength in the active - medium . for gaas / algaas based terahertz qcls , @xmath16 , @xmath17 , and hence for a chosen grating - period @xmath5
, the first - order dfb , antenna - feedback , and second - order dfb modes occur at wavelengths @xmath18 , @xmath19 , and @xmath20 respectively .
the typical gain - bandwidth of terahertz qcls is less than @xmath21 of the peak - gain wavelength , which suggests that the grating has to be designed specifically to excite the antenna - feedback mode .
the antenna - feedback scheme leads to excitation of a coherent single - sided spp standing - wave on the metallic cladding of the spaser , which is phase - locked to the resonant - cavity spp mode inside the active medium as shown in fig .
[ fig1](f ) .. both waves maintain exact same phase relation at each aperture location , where they exchange electromagnetic ( em ) energy with each other due to diffraction as illustrated in fig . [ fig1](e ) .
the spp wave in the surrounding medium is excited due to scattering of em field at apertures that generates a combination of propagating quasi - cylindrical waves and spps @xcite that propagate along the surface of the metal - film .
the scattered waves thus generated at each aperture superimpose constructively in only the end - fire ( @xmath22 ) direction owing to the phase - condition thus established at each aperture .
for coupling to far - field radiation , the radiation is therefore analogous to that from an end - fire phased array antenna that produces a narrow beam in both @xmath23 and @xmath24 directions .
thick and based on a @xmath25 gaas/@xmath26 qcl design ( details in supplementary material , section s1 ) .
a scanning - electron microscope image of the fabricated qcls is shown on the right .
( b ) experimental light - current - voltage characteristics of a representative qcl with antenna - feedback of dimensions @xmath27 at different heat - sink temperatures .
the qcl is biased with low duty - cycle current pulses of @xmath28ns duration and @xmath29khz repetition rate .
inset shows lasing spectra for different bias where the spectral linewidth is limited by instrument s resolution .
the emitted optical power is measured without any cone collecting optic inside the cryostat .
( c ) measured spectra for four different antenna - feedback qcls with varying grating periods @xmath5 , but similar overall cavity dimensions .
the qcls are biased at a current - density @xmath30 at @xmath31 .
, width=316 ] a third - order dfb technique was recently shown to achieve emission in a narrow beam for terahertz qcls with fabry - prot cavities @xcite .
it can achieve high directionality for the radiated beam in both directions perpendicular to propagation , so long as the effective propagation index of the spp wave inside the active medium could be made @xmath32 by complex deep dry etching in the slits @xcite or lateral corrugated geometry @xcite .
the so - called phase matching condition is possible for gaas - based qcls by cavity engineering @xcite since the @xmath33 is close to @xmath34 .
the antenna - feedback technique in this work offers similar outcome as a perfectly matched third - order dfb with improved directionality as well as a novel outcoupling mechanism of the radiated beam from terahertz qcls .
it is to be noted that the antenna - feedback scheme is automatically phase - matched and hence it could be utilized for any type of spaser without any restrictions on the required index in the active medium .
figure [ fig2 ] shows comparison of the eigenmode spectrum of a terahertz qcl cavity with conventional dfb , taking third - order dfb as an example versus antenna - feedback gratings computed using a finite - element solver @xcite .
the occurrence of band - gaps in the spectra is indicative of the dfb effect . in both cases ,
the lower - frequency band - edge mode is the lowest - loss mode by way of dfb action , since the dfb modes result in a standing - wave being established along the length of the cavity with an envelope shape that vanishes close to the longitudinal boundaries ( end - facets ) . for antenna - feedback grating ,
a standing - wave for the single - sided spp wave is additionally established in air , as can be seen from the field - plot of the band - edge mode in fig .
[ fig2](b ) .
in contrast , third - order dfb leads to negligible amplitude of the single - sided spp wave in air , as mentioned previously and illustrated schematically in fig .
[ fig1](c ) . for the dominant tm polarized ( ey ) electric field of antenna - feedback ,
the hybrid spps mode bound to the top metal layer consists of both quasi - cylindrical waves and spps , which are evanescent - field with a free - space propagation constant .
particularly at long wavelengths , such as mid infrared and thz region , spps and quasi - cylindrical waves complexly mix with each other @xcite , contributing to the large spatial extent of spp mode in surrounding medium as shown in fig . [ fig2](b ) , which does not exist for any conventional dfb , such as first - order , second - order and third - order dfb .
in addition , hybrid spps on top of metallic grating and standing - wave inside laser cavity show clearly different periodicity , with the ratio of free space wavelength over guided wavelength , which further confirms that the excitation of the coherent spp wave on both side of top metallic surface contribute to feedback and coupling mechanism with antenna - feedback scheme .
the absorbing boundaries at the longitudinal ends of the cavity @xcite increase the relative loss of the modes that are further away from the band - edge mode , which helps in mode discrimination and will lead to excitation of the desired band - edge mode for single - mode operation of the spaser .
the active region and metal layers are modeled as lossless since the exact loss contribution due to each is not clear in literature for terahertz qcls at cryogenic temperatures . if lossy metal is used , the relative loss of various resonant modes for the dfb cavities are not impacted and neither are the mode - shapes and the corresponding resonant frequencies . for the band - edge modes in fig .
[ fig2](a ) and fig . [ fig2](b ) , a loss of @xmath35 was estimated as a contribution from the absorbing boundaries .
consequently , the radiative ( outcoupling ) loss of of the third - order dfb is @xmath36 as compared to @xmath37 for the antenna - feedback .
the radiative loss of the third - order dfb is smaller since the band - edge mode has zeros of the radiative field ( @xmath38 ) being located at each aperture , since the grating period @xmath5 is integer multiple of half - wavelengths in the gaas / algaas active medium ( @xmath39 , where @xmath40 ) .
such a low - outcoupling efficiency is also existent in surface - emitting terahertz qcls with second - order dfb @xcite . for the cavity with antenna - feedback ,
the radiative loss is higher because the grating period is not an integer multiple of half - wavelengths inside the active medium ( @xmath41 ) that leads to large amplitudes of the radiative - field ( @xmath38 ) in alternating apertures as shown in the figure . as a consequence ,
the output power from terahertz qcls with antenna - feedback should be greater than that with conventional dfb gratings , which is an additional advantage of the antenna - feedback scheme for terahertz qcls .
this was also verified experimentally from the measured output power .
the recently developed second - order dfb qcls with graded periodicity @xcite achieve high - power emission for the same reason , non - zero radiative field under the metallic apertures .
figure [ fig3 ] shows experimental results from terahertz qcls implemented with antenna - feedback gratings .
details about fabrication and measurement methods are presented in supplementary material ( section s1 ) .
[ fig3](b ) shows representative @xmath42-@xmath43 curves versus heat - sink temperature for a qcl with @xmath44 .
the qcl operated up to a temperature of @xmath45 . in comparison , multi - mode fabry - prot cavity qcls on the same chip that did not include longitudinal or lateral absorbing boundaries operated up to @xmath46 .
light - current characteristics and spectra at different bias with fabry - prot cavity are shown in supplementary material ( section s3 ) .
the temperature degradation due to absorbing boundaries is relatively small and similar to previous reports of dfb terahertz qcls @xcite .
the inset shows measured spectra at different bias at @xmath31 .
most qcls tested with different grating periods showed robust single - mode operation except close to peak - bias when a second - mode was excited for some devices at a shorter - wavelength , which suggests it is likely due to a higher - order lateral mode being excited due to spatial - hole burning in the cavity .
peak - power output of @xmath47 was detected from the antenna - feedback qcl measured directly at the detector without using any collecting optics . for comparison ,
a terahertz qcl with third - order dfb ( without phase matching ) and similar dimensions was also fabricated on the same chip , which operated up to a similar temperature of @xmath48 and emitted peak - power output of @xmath49 ( see supplementary material , section s3 ) .
the antenna - feedback gratings lead to greater radiative outcoupling compared to conventional dfb schemes for terahertz qcls , as discussed in the previous section .
[ fig3](c ) shows spectra measured from four different terahertz qcls with antenna - feedback gratings of different grating periods @xmath5 .
the single - mode spectra scales linearly with @xmath5 , which is the clearest proof that the feedback mechanism works as expected and the lower band - edge mode is selectively excited in each case . using @xmath50 from equation , the effective propagation index of the spp mode in the active medium @xmath10 is calculated as @xmath51 , @xmath52 , @xmath53 , and @xmath54 for qcls with @xmath5 of @xmath55 , @xmath56 , @xmath57 , and @xmath58 respectively .
the effective mode - index @xmath10 decreases because a larger @xmath5 introduces larger sized apertures in the metal film since the grating duty - cycle was kept same for all devices .
consequently , a greater amount of field couples to the single - sided spp mode in air for increasing @xmath5 , thereby reducing the modal confinement in the active medium that reduces the propagation index of the guided spp mode further . in pulsed mode and biased at @xmath30 while lasing in single - mode .
the plots are for qcls with @xmath59 mm long cavities and ( b ) @xmath60 width and @xmath61 grating emitting at @xmath62 , ( c ) @xmath63 width and @xmath61 grating emitting at @xmath62 , and ( d ) @xmath63 width and @xmath44 grating emitting at @xmath64 respectively .
, width=316 ] experimental far - field beam patterns for antenna - feedback qcls with varying designed parameters are shown in fig .
single - lobed beams in both lateral ( @xmath23 ) and vertical ( @xmath24 ) directions were measured for all qcls . as shown in fig .
[ fig4](b ) , the full - width half - maximum ( fwhm ) for the qcl with @xmath60 width , @xmath61 is @xmath65 , which is the narrowest reported beam - profile from any terahertz qcl to - date .
in contrast , previous schemes for emission in a narrow - beam have resulted in divergence angles of @xmath66 using very long ( @xmath67 ) cavities and a phased - matched third - order dfb scheme @xcite and @xmath68 using broad - area devices with 2d photonic - crystals @xcite for single - mode terahertz qcls , and @xmath69 @xcite and @xmath70 @xcite for multi - mode qcls using metamaterial collimators .
[ fig4](c ) and [ fig4](d ) show representative beam patterns from qcls with wider cavities of @xmath63 width , for the smallest and largest @xmath5 in the range of fabricated devices respectively .
the beam divergence is relatively independent of @xmath5 as expected .
more importantly , the measurements show that the wider cavities result in a slightly broader beam .
such a result is counter - intuitive because typically a laser emits in a narrower beam as its cavity s dimensions are increased due to an increase in the size of emitting aperture .
such a behavior is unique for a spaser with antenna - feedback , and is discussed along with full - wave 3d fem simulation of the beam pattern in supplementary material ( section s2 ) .
it can be argued that the size of the beam could be further narrowed by utilizing narrower cavities for terahertz qcls , which will be extremely beneficial to develop cw sources of narrow - beam coherent terahertz radiation .
in this article , we have presented a novel antenna - feedback scheme to achieve single - mode operation and a highly directional far - field radiation pattern from plasmonic lasers with subwavelength apertures and fabry - prot type cavities .
it is conceptually different from any other previously utilized dfb schemes for solid - state lasers , and is based on phase - locking of a single - sided surface - plasmon - polariton ( spp ) mode on ( one of ) the metal film(s ) in the spaser s cavity with the guided spp mode inside the spaser s active medium .
the phase - locking is established due to strong bragg diffraction of the spp modes by periodically perforating the metal film in the form of a grating of holes or slits .
the uniqueness of the method lies in the specific value of the grating s period , which leads to the spaser s cavity radiating like an end - fire phased - array antenna for the excited dfb mode .
additionally , coherent single - sided spps are also generated on the metal film that have a large spatial extent in the surrounding medium of the laser s cavity , which could have important implications for applications in integrated plasmonics .
coherent spps with large spatial extent could make it easy to couple spp waves from the plasmonic lasers to other photonic components , and could also potentially be utilized for plasmonic sensing .
experimentally , the scheme is implemented in terahertz qcls with subwavelength metallic cavities .
a beam - divergence angle as small as @xmath0 is achieved for single - mode qcls , which is narrower than that achieved with any other previously reported schemes for terahertz qcls with periodic photonic structures .
compared with the third - order dfb method , the new antenna - feedback scheme is easier to implement for fabrication by standard lithography techniques without any other complex fabrication technique to precisely match a well defined effective mode index , and achieves a superior radiative outcoupling owing to the fact that the grating period is not an integer multiple of half - wavelengths of the standing spp - wave inside the active medium .
terahertz qcls with antenna - feedback could lead to development of new modalities for terahertz spectroscopic sensing and wavelength tunability due to access of a coherent terahertz spp wave on top of the qcl s cavity , possibilities of sensing and imaging at standoff distances of few tens of meters , and development of integrated terahertz laser arrays with a broad spectral coverage for applications in terahertz absorption spectroscopy .
see supplementary materials section for additional supporting content
. 10 d. j. bergman and m. i. stockman , `` surface plasmon amplification by stimulated emission of radiation : quantum generation of coherent surface plasmons in nanosystems , '' phys .
lett . * 90 * , 027402 ( 2003 ) .
m. t. hill and m. c. gather , `` advances in small lasers , '' nat .
* 8 * , 908 ( 2014 ) .
r. f. oulton , `` surface plasmon lasers : sources of nanoscopic light , '' mater . today * 15 * , 26 ( 2012 ) .
p. berini and i. d. leon , `` surface plasmon - polariton amplifiers and lasers , '' nat .
* 6 * , 16 ( 2012 ) .
m. a. noginov , g. zhu , a. m. belgrave , r. bakker , v. m. shalaev , e. e. narimanov , s. stout , e. herz , t. suteewong , and u. wiesner , `` demonstration of a spaser - based nanolaser , '' nature * 460 * , 1110 ( 2009 ) .
r. khler , a. tredicucci , f. beltram , h. e. beere , e. h. linfield , a. g. davies , d. a. ritchie , r. c. iotti , and f. rossi , `` terahertz semiconductor - heterostructure laser , '' nature * 417 * , 156159 ( 2002 ) .
b. s. williams , `` terahertz quantum - cascade lasers , '' nat .
* 1 * , 517525 ( 2007 )
. m. t. hill , m. marell , e. s. p. leong , b. smalbrugge , y. zhu , m. sun , p. j. van veldhoven , e. j. geluk , f. karouta , y .- s .
oei , r. ntzel , c .- z .
ning , and m. k. smit , `` lasing in metal - insulator - metal sub - wavelength plasmonic waveguides , '' opt .
express * 17 * , 11107 ( 2009 ) .
r. f. oulton , v. j. sorger , t. zentgraf , r .- m .
ma , c. gladden , l. dai , g. bartal , and x. zhang , `` plasmon lasers at deep subwavelength scale , '' nature * 461 * , 629 ( 2009 )
. y .- j .
lu , j. kim , h .- y .
chen , c. wu , n. dabidian , c. e. sanders , c .- y .
wang , m .- y .
lu , b .- h .
li , x. qiu , w .- h .
chang , l .- j .
chen , g. shvets , c .- k .
shih , and s. gwo , `` plasmonic nanolaser using epitaxially grown silver film , '' science * 337 * , 450 ( 2012 ) .
b. s. williams , s. kumar , h. callebaut , q. hu , and j. l. reno , `` terahertz quantum - cascade laser at @xmath71 m using metal waveguide for mode confinement , '' appl .
lett . * 83 * , 21242126 ( 2003 ) .
a. j. l. adam , i. kaalynas , j. n. hovenier , t. o. klaassen , j. r. gao , e. e. orlova , b. s. williams , s. kumar , q. hu , and j. l. reno , `` beam patterns of terahertz quantum cascade lasers with subwavelength cavity dimensions , '' appl .
. lett . * 88 * , 151105 ( 2006 ) .
e. e. orlova , j. n. hovenier , t. o. klaassen , i. kaalynas , a. j. l. adam , j. r. gao , t. m. klapwijk , b. s. williams , s. kumar , q. hu , and j. l. reno , `` antenna model for wire lasers , '' phys . rev .
* 96 * , 173904 ( 2006 ) .
w. zhou , m. dridi , j. y. suh , c. h. kim , d. t. co , m. r. wasielewski , g. c. schatz , and t. w. odom , `` lasing action in strongly coupled plasmonic nanocavity arrays , '' nat .
* 8 * , 506 ( 2013 ) .
f. van beijnum , p. j. van veldhoven , e. j. geluk , m. j. a. de dood , g. w. t hooft , and m. p. van exter , `` surface plasmon lasing observed in metal hole arrays , '' phys .
lett . * 110 * , 206802 ( 2013 ) .
x. meng , j. liu , a. v. kildishev , and v. m. shalaev , `` highly directional spaser array for the red wavelength region , '' laser & photon .
* 8 * , 896 ( 2014 ) .
a. yang , t. b. hoang , m. dridi , c. deeb , m. h. mikkelsen , g. c. schatz , and t. w. odom , `` real - time tunable lasing from plasmonic nanocavity arrays , '' nature comm .
* 6 * , 6939 ( 2015 ) .
a. h. schokker and a. f. koenderink , `` lasing at the band edges of plasmonic lattices , '' phys .
b * 90 * , 155452 ( 2014 ) .
a. v. dorofeenko , a. a. zyablovsky , a. p. vinogradov , e. s. andrianov , a. a. pukhov , and a. a. lisyansky , `` steady state superradiance of a 2d - spaser array , '' opt .
express * 21 * , 1453914547 ( 2013 ) .
l. mahler and a. tredicucci , `` photonic engineering of surface - emitting terahertz quantum cascade lasers , '' laser & photon . rev . * 5 * , 647658 ( 2011 ) . c. sirtori , s. barbieri , and r. collombelli ,
`` wave engineering with thz quantum cascade lasers , '' nat
* 7 * , 691 ( 2013 ) .
kao , q. hu , and j. l. reno , `` phase - locked arrays of surface - emitting terahertz quantum - cascade lasers , '' appl .
* 96 * , 101106 ( 2010 ) .
y. halioua , g. xu , s. moumdji , l. li , j. zhu , e. h. linfield , a. g. davies , h. e. beere , d. a. ritchie , and r. colombelli , `` phase - locked arrays of surface - emitting graded - photonic - heterostructure terahertz semiconductor lasers , '' opt .
express * 23 * , 6915 ( 2015 ) .
l. xu , c. a. curwen , p. w. hon , q .- s .
chen , t. itoh , and b. s. williams , `` metasurface external cavity laser , '' appl .
. lett . * 107 * , 221105 ( 2015 ) .
m. i. amanti , m. fischer , g. scalari , m. beck , and j. faist , `` low - divergence single - mode terahertz quantum cascade laser , '' nat . photon . * 3 * , 586590 ( 2009 ) .
t. y. kao , q. hu , and j. l. reno , `` perfectly phase - matched third - order dfb thz quantum - cascade lasers , '' opt . lett .
* 37 * , 2070 ( 2012 ) .
b. s. williams , s. kumar , q. hu , and j. l. reno , `` distributed - feedback terahertz quantum - cascade lasers with laterally corrugated metal waveguides , '' opt .
lett . * 30 * , 2909 ( 2005 ) .
j. a. fan , m. a. belkin , f. capasso , s. khanna , m. lachab , a. g. davies , and e. h. linfield , `` surface emitting terahertz quantum cascade laser with a double - metal waveguide , '' opt .
express * 14 * , 11672 ( 2007 ) .
s. kumar , b. s. williams , q. qin , a. w. m. lee , q. hu , and j. l. reno , `` surface - emitting distributed feedback terahertz quantum - cascade lasers in metal - metal waveguides , '' opt .
express * 15 * , 113 ( 2007 ) .
y. chassagneux , r. colombelli , w. maineult , s. barbieri , h. e. beere , d. a. ritchie , s. p. khanna , e. h. linfield , and a. g. davies , `` electrically pumped photonic - crystal terahertz lasers controlled by boundary conditions , '' nature * 457 * , 174 ( 2009 ) .
y. halioua , g. xu , s. moumdji , l. h. li , a. g. davies , e. h. linfield , and r. colombelli , `` thz quantum cascade lasers operating on the radiative modes of a 2d photonic crystal , '' opt . lett . * 39 * , 3962 ( 2014 ) . p. lalanne , j. hugonin , h. liu , and b. wang , `` a microscopic view of the electromagnetic properties of sub-@xmath9 metallic surfaces , '' surf .
* 64 * , 453 ( 2009 ) .
a. babuty , a. bousseksou , j .-
tetienne , i. m. doyen , c. sirtori , g. beaudoin , i. sagnes , y. d. wilde , and r. colombelli , `` semiconductor surface plasmon sources , '' phys .
lett . * 104 * , 226806 ( 2010 ) .
m. i. amanti , g. scalari , f. castellano , m. beck , and j. faist , `` low divergence terahertz photonic - wire laser , '' opt .
express * 18 * , 6390 ( 2010 ) . , a finite - element partial differential equation solver from comsol inc .
o. demichel , l. mahler , t. losco , c. mauro , r. green , a. tredicucci , j. xu , f. beltram , h. e. beere , d. a. ritchie , and v. tamoinuas , `` surface plasmon photonic structures in terahertz quantum cascade lasers , '' opt .
express * 14 * , 5335 ( 2006 ) .
g. xu , r. colombelli , s. p. khanna , a. belarouci , x. letartre , l. li , e. h. linfield , a. g. davies , h. e. beere , and d. a. ritchie , `` efficient power extraction in surface - emitting semiconductor lasers using graded photonic heterostructures , '' nature comm .
* 3 * , 952 ( 2012 ) .
g. liang , e. dupont , s. fathololoumi , z. r. wasilewski , d. ban , h. k. liang , y. zhang , s. f. yu , l. h. li , a. g. davies , e. h. linfield , h. c. liu , and q. j. wang , `` planar integrated metasurfaces for highly - collimated terahertz quantum cascade lasers , '' sci . rep .
* 4 * , 7083 ( 2014 ) .
n. yu , q. j. wang , m. a. kats , j. a. fan , s. p. khanna , l. li , a. g. davies , e. h. linfield , and f. capasso , `` designer spoof surface plasmon structures collimate terahertz laser beams , '' nat . materials * 9 * , 730 ( 2010 ) . s. khanal , l. zhao , j. l. reno , and s. kumar , `` temperature performance of terahertz quantum - cascade lasers with resonant phonon active - regions , '' j. opt * 16 * , 094001 ( 2014 ) .
c. a. balanis , _ antenna theory : analysis and design _ ( wiley - interscience , 2005 ) , 3rd ed . c. boyle , c. sigler , j. kirch , d. lindberg iii , t. earles , d. botez , and l. mawst , `` high - power , surface - emitting quantum cascade laser operating in a symmetric grating mode , '' appl . phys . lett . * 108 * , 121107 ( 2016 ) .
national science foundation ( nsf ) ( eccs 1128562 , eccs 1351142 , cmmi 1437168 ) .
this work is performed , in part , at the center for integrated nanotechnologies , a u.s .
department of energy ( doe ) , office of basic energy sciences user facility .
sandia national laboratories is a multiprogram laboratory managed and operated by sandia corporation , a wholly owned subsidiary of lockheed martin corporation , for the u.s .
doe s national nuclear security administration under contract de - ac04 - 94al85000 .
a united states patent application ( pending ) for the described technology has been filed through lehigh university ( application number 14/984,652 , filed on dec .
30 , 2015 ) .
@subsection terahertz plasmonic laser radiating in an ultra - narrow beam +
the active - medium of the qcls is based on a three - well resonant - phonon design with gaas/@xmath26 superlattice ( design rtrp3w197 , wafer number vb0464 ) , which is described in ref . , and was grown by molecular - beam epitaxy .
the qcl superlattice is @xmath72 thick with an average @xmath73-doping of @xmath74 , and surrounded by @xmath75 and @xmath76 thick highly - doped gaas contact layers at @xmath77 on either side of the superlattice .
fabrication of qcls with parallel - plate metallic cavities followed a cu - cu thermocompression wafer bonding technique as in ref . with standard optical contact lithography . following wafer - bonding and substrate removal ,
positive - resist lithography was used to selectively etch away the @xmath76 thick highly - doped gaas layer from all locations where top - metal cladding would exist on individual cavities except in @xmath78 wide regions at the outer rectangular boundaries of the top - metal layer due to overlapping mask layers .
the removal of this layer beneath the top - metal does not impact electrical transport significantly except adding a small voltage drop at the top contact during qcl operation .
importantly , this lithography step allows radiative outcoupling from the apertures in the finally fabricated qcl cavities ; and simultaneously serves to implement longitudinal and lateral absorbing boundaries as illustrated in fig .
3(a ) by leaving the highly - doped gaas layer exposed at both longitudinal and lateral edges of the qcl cavity .
the absorbing boundaries result in a highly lossy propagation of the spp modes in those regions @xcite .
while the longitudinal boundaries help in dfb mode discrimination , the lateral boundaries are useful to eliminate higher - order lateral guided modes by making them more lossy in comparison to the fundamental mode @xcite .
ti / au metal layers of thickness @xmath79 nm were used as the top metal cladding , using an image - reversal lithography mask for implementing gratings in the metal layers .
another positive - resist lithography step was used to cover the grating - metal with photoresist to be used as a mask for wet - etching of ridges in a h@xmath80so@xmath81:h@xmath80o@xmath80:h@xmath80o @xmath82:@xmath83:@xmath84 solution for @xmath85minutes .
after etching the @xmath72 thick superlattice active region when the bottom metal layer is exposed , an over - etch of @xmath86minute was followed up to reduce the slope on the sidewalls of the qcl s cavities to allow for a more uniform current - density distribution through the height of the cavity ( in @xmath24 direction ) .
a ti / au contact was used as the backside - metal contact for the finally fabricated qcl chips to assist in soldering . before deposition of backside - metal of the wafer
, the substrate was mechanically polished down to a thickness of @xmath87 to improve heat - sinking . for measurements ,
a cleaved chip consisting of qcls with antenna - feedback gratings of different periods along with some fabry - protridge qcls and third - order dfb qcls was in - soldered on a cu block , the qcl to be tested was wire - bonded for electrical biasing , and the cu block was mounted on the cold - stage of a liquid nitrogen vacuum cryostat for measurements .
initial measurements of the fabricated qcl cavities did not yield lasing devices at @xmath31 .
it was then realized that the @xmath88 wide lateral absorbing boundaries on each side were perhaps introducing high optical losses in the cavities .
the finally fabricated wafers were then etched again ( post - fabrication ) in a h@xmath80so@xmath81:h@xmath80o@xmath80:h@xmath80o @xmath82:@xmath83:@xmath84 solution for @xmath89seconds , which is likely to etch away most of the exposed highly - doped gaas layer . however , @xmath78 wide highly - doped gaas layer remains unetched underneath the top - metal cladding on all four outer boundaries of the cladding , which serves to perform the role of an absorbing region .
the post - etched devices lased upon re - testing , and their robust single - mode operation suggests that both lateral and longitudinal absorbing boundaries were able to suppress undesired resonant - modes in the cavity as desired .
the emitted optical power was measured with a deuterated triglycine sulfate pyroelectric detector ( dtgs ) pyroelectric detector and calibrated with a terahertz thermopile powermeter ( scientech model ac2500h ) without any optical component or cone collecting optic inside the cryostat to improve collection of radiated power in the case of dfb qcls .
( a winston cone was used to collect power , but only for the fabry - prot - cavity qcl reported in fig .
[ figs3 ] . )
the spectra were measured at @xmath31 in linear - scan mode with a resolution of @xmath90 ( @xmath91 ghz ) using a bruker fourier - transform infrared spectrometer ( ftir ) under vacuum , with a room - temperature dtgs detector placed inside the ftir .
far - field radiation patterns were measured with a dtgs pyroelectric cell detector mounted on a computer - controlled two - axis ( @xmath92 plane ) moving stage in the end - fire ( @xmath22 ) direction at a distance of @xmath93 mm from the qcls end - facets .
the detector s size leads to spatial averaging of the measured beam with a step - size of @xmath94 , which has a negligible impact on the measured beam - width .
finite - element ( fem ) simulations were performed using a commercially available software @xcite .
the eigenmodes of the dfb cavities are computed using an eigenfrequency solver .
the loss in absorbing boundaries of the cavity is implemented using a complex dielectric constant computed using drude - model @xcite .
the imaginary part of the computed eigenfrequencies is used to compute the propagation loss of the resonant - cavity modes in the units of inverse of length .
the active region is modeled as lossless and the metal is modeled to be perfect electrical conductors , in which case , the computed loss is the sum of loss at absorbing boundaries as well as that due to radiation ( outcoupling ) .
the thickness of the metallic layer is @xmath95 nm in simulation .
the relative losses of various resonant modes for the dfb cavities are not changed and neither are the mode - shapes and the corresponding resonant frequencies when using lossy metal , or when the thickness of the metallic film are altered ( as long as it is kept significantly thinner than that of the active region ) . for the 2d simulations ( cavities of infinite - width )
the far - field boundaries are modeled as absorbing layers as described in ref .
for 3d simulations , the in - built perfectly matched layers in the fem solver were utilized to model the far - field boundaries .
the narrow - beam emission from plasmonic lasers with antenna - feedback is due to a combination of two factors that are related to the radiative behavior of phased - array antennas .
in addition to the _ array - factor _ that leads to narrower beams for more number of elements in a phased - array ( in this case , a longer length of the spaser s cavity ) , the far - field radiation pattern is additionally narrowed owing to an _ element - factor _ , which is defined as the far - field radiation pattern due to an individual emitter of the phased - array @xcite . the large spatial extent of the single - sided spp wave in the surrounding medium results in a narrower element - factor compared to an omnidirectional point - source emitter .
as measured experimentally and shown in fig . 4 , the antenna - feedback terahertz qcl cavities with a narrower width of @xmath60 resulted in a beam with smaller divergence as compared to the @xmath63 wide cavity .
this seemingly unique behavior from antenna - feedback qcls serves to further validate the concept of the specific feedback scheme , in which the cavity radiates like a phased - array antenna .
a narrower cavity causes greater lateral spread of the single - sided spp mode ( in the @xmath23 dimension ) in the surrounding medium , especially when the cavity s width is sub - wavelength .
a single - sided spp mode with a broader cross - section in the @xmath92 plane will lead to a more directional far - field radiation pattern owing to a narrower element - factor for the phased - array antenna structure . to further validate the experimental results of fig . 4 ,
full - wave 3d fem simulations were carried out for terahertz qcl cavities with antenna - feedback .
figure [ figs1 ] shows the computed far - field radiation patterns for band - edge dfb mode for terahertz qcl cavities implemented with antenna - feedback gratings in the top metal cladding .
the in - built perfectly - matched layers ( pmls ) in the fem solver @xcite serve as an effective absorbing boundary for computation of the radiation pattern .
the pmls parameters are coarsely adjusted to achieve minimum reflection at terahertz frequencies and transform the propagating waves to exponentially decaying waves .
the pmls are placed far enough to avoid interaction with the single - side spp standing - wave in the near - field of the cavity . considering the computer s memory limitations , laterally absorbing
boundaries are not modeled in the cavity , which does not impact the simulation s results since the dfb mode could always be found , even though it is not the lowest loss mode in the eigenmode spectrum of the cavity due to other higher order lateral modes that can have even lower propagation loss .
all metal layers are modeled as perfectly conducting surfaces to limit the mesh size in the geometry . in order to set up a far - field calculation , a far - field domain node
is implemented that is a single closed surface surrounding all radiating apertures in the cavity .
the distribution of far em field is based on the fourier transform of the near - field as implemented in the fem solver .
simulated 3d beam patterns demonstrate single - lobed beams with narrow divergence for the band - edge mode with antenna - feedback as shown in the figure .
the shape and fwhm of the computed radiation - pattern is in close agreement with the experimentally measured beams that are shown in fig .
the fwhm of simulated beam patterns is slightly larger than that of the measured results .
the difference are likely due to the relative simplification of implemented 3d model , which does not account for the sloped sidewall profile of the ridges and also the fact the ground plane around the qcl s cavity is different for the measured chip as compared to that in the simulated geometry .
( @xmath96 ) with grating period @xmath97 .
, width=316 ] . . ,
figure [ figs2 ] shows @xmath42-@xmath43 characteristics of a representative terahertz qcl with third - order dfb without well defined phase - matching condition that was fabricated on the same chip by standard lithography to pattern grating on top metallic layer , and with similar dimensions as that of the antenna - feedback qcls whose data is reported in fig .
3 . using the phase condition @xmath11 as determined from equation 1 for @xmath98 ( third - order dfb ) , an effective - mode index of @xmath99
is estimated for the resonant - cavity spp mode in the active medium of the cavity . for ideal phase - matching ,
a mode - index value of @xmath34 is required , hence the phased - matching condition is not satisfied for this qcl . consequently , the measured fwhm of the far - field radiation pattern @xmath100 for this qcl is not as good as that for phased - matched third - order dfb qcls with reduced effective mode index @xcite .
elongated and multi - lobed beam pattern for third - order dfb when refractive index is close to @xmath101 is observed in ref .
the peak output power for the third - order dfb qcl without phase matching is @xmath49 at @xmath31 , which is about three times less than that obtained for the antenna - feedback qcl with similar dimensions lasing at similar frequency .
the output power slope efficiencies at @xmath31 are @xmath102mw / a and @xmath103mw / a respectively for the third - order dfb qcl and antenna - feedback qcl respectively .
the optical power was measured without cone collecting optics .
the relative differences in the measured optical power are in good agreement with the radiative losses estimated via 2d fem simulations for similar cavity dimensions as in fig . 2 .
the maximum operating temperature of this qcl is almost similar to that of the antenna - feedback qcl , which suggests that radiative loss is relatively a small fraction of the overall waveguide loss in both types of dfb qcl cavities .
this also suggests that there is a scope for enhancing radiative efficiency to increase optical power output without causing a significant degradation in temperature performance of terahertz qcls with antenna - feedback .
calculation of surface - outcoupling efficiency as in ref .
@xcite is outside the scope of this work due to large uncertainty in the absolute loss contribution of metal layers , absorbing boundaries , as well as the active region itself .
the optimized phase - matched third - order dfb @xcite with bound - to - continuum design shows lower current - density and higher efficiency without absorbing regions .
since the feedback of laser mode is achieved by spp mode propagating on the surface for antenna - feedback , technique of covering waveguide sidewall with metal @xcite or lateral corrugated grating @xcite , which does not absorb effective outcoupled laser radiation from antenna - feedback , can replace lateral absorber layer here and will further enhance the outcoupling efficiency .
l - i characteristics at different heat - sink temperature of a representative terahertz qcl with fabry - prot ( fp ) cavity is shown in fig .
[ figs3 ] . exposed highly - doped lossy contact layer serving as absorber introduces @xmath104 temperature degradation for dfb cavity . because sub - wavelength mode confinement in fp cavities results in highly divergent output beams of thz qcls with parallel - plate metallic cavities , output power is measured with a winston cone and the detector was placed adjacent to the cryostat window .
diameter of the circular opening of cone is @xmath105 .
frequency of gain medium covers @xmath106 to @xmath107 at @xmath108 .
direction ) for the band - edge dfb modes for ( a ) third - order dfb cavity , and ( b ) cavity with antenna - feedback respectively . , width=307 ] figure [ figs4 ] shows energy - density profile for the resonant band - edge modes of the cavities with third - order dfb and antenna - feedback schemes respectively , along the entire length of the cavity ( @xmath59 mm ) .
energy - density is calculated and summed up along the cavity s height ( @xmath72 ) from 2d simulations ( that effectively model cavities of infinite width ) .
the eigenmode spectra and electric - field distributions for the cavities are shown in fig . 2 . both third - order dfb and antenna - feedback schemes show non - uniform envelope shapes , which gradually decrease from the center to the end - facets of cavity and provide indication of typical dfb action due to coupling of propagating waves along the length of cavity as per the corresponding schematics in fig . 1
. long cavity and dominant electric field @xmath109 of a fabry - prot cavity type mode that exists sufficiently away from the photonic bandgap .
( c ) energy density profile and electric - field profile along the length of the cavity for the `` desired '' antenna - feedback mode that exists at the edge of the photonic bandgap .
the @xmath22 direction is along the length of laser ridge , as marked in fig .
[ figs4 ] .
, width=336 ] a thin highly - doped gaas layer below the top - metal cladding was left unetched at the regions close to longitudinal facets , serving as the longitudinal absorbing boundary to ensure the excitation of the desired antenna - feedback mode as the lowest - loss lasing mode .
firstly , such absorbing boundaries reduce the end - facet reflectivities that strengthens the distributed - feedback coupling for the antenna - feedback scheme .
secondly , the longitudinal absorber boundaries play an important role in providing the necessary mechanism for mode - discrimination , to selectively excite the desired band - edge antenna - feedback mode for lasing .
figure [ figs5](a ) shows that eigenmode spectrum of antenna - feedback scheme without longitudinal absorbing boundaries .
there exist the fabry - perot type modes with uniform energy density profile along the whole length of laser cavity and large electric field distribution close to end facets of the cavity , as shown in fig .
[ figs5](b ) . however , the loss of such modes is small due to highly reflective end - facets .
the desired antenna - feedback mode , located at the lower bandedge , shows non - uniform energy - density profile along the length of cavities as an indicator of dfb action and strong coupling between plasmonic mode on top of metallic layer and guided mode in the active core , as seen from fig .
[ figs5](c ) .
however , its radiative loss is higher due to the improved and large radiative field of antenna - feedback mode and hence , it can not be excited for lasing in the absence of longitudinal absorbing boundaries .
when longitudinal absorbing boundaries are introduced , the relative loss of the modes with large intensity closer to longitudinal ends of cavity , which are further away from the photonic bandgap , will be selectively and significantly increased .
therefore , the antenna - feedback mode is excited for lasing as the robust lowest - loss mode as seen from the eigenmode spectrum in fig .
2(b ) .
width , @xmath72 thickness , and @xmath110 length is simulated as in fig .
[ figs1 ] .
the frequency of the band - edge resonant - cavity mode is @xmath111 , which is the desired `` antenna - feedback '' mode .
inset : mode shape along lateral @xmath23 dimension of cavity at a distance of @xmath72 above the top metal cladding ; the fwhm of the lateral electric - field profile is @xmath112 .
, width=316 ] figure [ figs6 ] shows lateral mode shape of the antenna - feedback mode at approximately the longitudinal center of the cavity from 3d simulation .
the spp mode in surrounding medium extends in both lateral and vertical directions significantly , and the extent of the mode in lateral ( @xmath23 ) dimension is greater than the width of the cavity itself .
the fhwm of the mode shape along lateral ( @xmath23 ) dimension of cavity at a distance of @xmath113 above top gratings ( picked arbitrarily to keep it slightly away from location where finite - element mesh in simulation changes abruptly ) is @xmath112 .
this large effective wavefront created on top of the waveguide leads to a narrow far - field , i.e. a narrow `` element - factor '' in terminology of the radiation from a phased - array antenna . while the narrowing of `` array - factor '' due to constructive interference of radiation from all apertures in end - fire direction is also at play
; it is not the dominant effect to cause narrow beaming for terahertz qcls in this work for short - length cavities ( @xmath114 ) .
in contrast , for third - order dfb qcls with phase - matching , beam narrowing is primarily due to the narrow array - factor only , which is why very long cavities ( @xmath115 ) are needed to achieve beam divergence values of less than @xmath116 @xcite .
profile in conventional dfb ( third - order dfb as an example ) and antenna - feedback scheme . * ( a ) @xmath117 profile for third - order dfb cavity plotted along perpendicular @xmath24 direction ( cavity s height direction ) .
( b ) @xmath117 profile for antenna - feedback scheme with an expanded view close to top metal cladding .
, width=316 ] figure [ figs7 ] shows the electric - field squared modulus @xmath117 profile inside the cavity and above the metal cladding for both conventional dfb and antenna - feedback scheme from 2d simulations .
layer sequence follows active region with @xmath72 thickness , top metal layer with @xmath95 nm thickness ( lossless metal ) and air region . at the location above metal cladding
, the mode - intensity in conventional dfb immediately decays to zero .
in contrast , for antenna - feedback scheme , the sub - wavelength plasmonic resonant - optical mode inside the cavity is phased - locked with a spp mode that travels on top of the metal cladding outside the cavity .
the spp wave is relatively tightly bound to the top metal cladding and decays in the length of the order of the wavelength in vacuum .
the decay length is independent of the loss in metal or thickness in metal or various duty cycles of grating ( which was verified via finite - element simulations ) .
the establishment of a hybrid spp mode on top of the cavity in vacuum is unique in antenna - feedback scheme . in a third - order dfb terahertz
qcl , even with effective mode index getting close to @xmath34 , the feedback mechanism remains same as that in conventional dfb lasers , i.e. distributed - feedback couples propagating modes inside the cavity without involving the propagating spp waves in the surrounding medium .
correspondingly , no standing spp wave ( with large amplitude ) is established in the surrounding medium in a third - order dfb qcl s cavity as it does in antenna - feedback scheme .
this could be verified by simulating a terahertz qcl cavity structure with third - order dfb gratings , but with the refractive index of the active region arbitrarily set to a value of @xmath32 to satisfy the so - called `` phase - matching '' condition in simulation . | plasmonic lasers ( spasers ) generate coherent surface - plasmon - polaritons ( spps ) and could be realized at subwavelength dimensions in metallic cavities for applications in nanoscale optics .
plasmonic cavities are also utilized for terahertz quantum - cascade lasers ( qcls ) , which are the brightest available solid - state sources of terahertz radiation .
a long standing challenge for spasers that are utilized as nanoscale sources of radiation , is their poor coupling to the far - field radiation . unlike conventional lasers that could produce directional beams ,
spasers have highly divergent radiation patterns due to their subwavelength apertures . here
, we theoretically and experimentally demonstrate a new technique for implementing distributed - feedback ( dfb ) that is distinct from any other previously utilized dfb schemes for semiconductor lasers . the so - termed _ antenna - feedback _ scheme leads to single - mode operation in plasmonic lasers , couples the resonant spp mode to a highly directional far - field radiation pattern , and integrates hybrid spps in surrounding medium into the operation of the dfb lasers .
experimentally , the antenna - feedback method , which does not require the phase matching to a well - defined effective index , is implemented for terahertz qcls , and single - mode terahertz qcls with beam divergence as small as @xmath0 are demonstrated , which is the narrowest beam reported for any terahertz qcl to - date .
moreover , in contrast to negligible radiative - field in conventional photonic band - edge lasers , in which the periodicity follows the integer multiple of half - wavelength inside active medium , antenna - feedback breaks this integer - limit for the first time and enhances the radiative - field of lasing mode .
terahertz lasers with narrow - beam emission will find applications for integrated as well as standoff terahertz spectroscopy and sensing .
the antenna - feedback scheme is generally applicable to any plasmonic laser with a fabry - prot cavity irrespective of its operating wavelength , and could bring plasmonic lasers closer to practical applications . |
supersymmetric grand unified theories ( guts ) provide a very attractive framework for representing particles and forces of nature as they solve the gauge hierarchy problem , unify three forces of nature , and also explain tiny neutrino masses through seesaw paradigm @xcite while providing possible cold dark matter candidates of the universe
. an evidence of supersymmetry at the large hadron collider ( lhc ) would be a land - mark discovery which would certainly change the future course of physics .
but , in the absence of any evidence of supersymmetry so far , it is worth while to explore new physics prospects of nonsupersymmetric ( non - susy ) guts @xcite and , particularly , those based upon @xmath0 which has grown in popularity as it unifies all fermions of one generation including the right - handed ( rh ) neutrino into a single spinorial representation .
it provides spontaneous origins of p ( @xmath10 parity ) and cp violations @xcite .
most interestingly , in addition to predicting the right order of tiny neutrino masses through mechanisms called the canonical ( @xmath11 type - i ) @xcite seesaw and type - ii @xcite seesaw , it has high potentiality to explain all fermion masses @xcite including large mixings in the neutrino sector @xcite with type - ii seesaw dominance @xcite .
in fact neither seesaw mechanism , nor grand unification require supersymmetry per se .
although gauge couplings automatically unify in the minimal supersymmetric standard model @xcite , and they fail to unify through the minimal particle content of the standard model ( sm ) in one - step breaking of non - susy su(5 ) or so(10 ) , they do unify once intermediate symmetries are included to populate the grand desert in case of non - susy so(10 ) @xcite .
in addition , with intermediate gauge symmetries @xmath0 also predicts signals of new physics which can be probed at low or accelerator energies . +
a hallmark of @xmath0 grand unification is its underlying quark - lepton symmetry @xcite because of which the canonical seesaw scale is pushed closer to the gut scale making it naturally inaccessible to direct tests by low - energy experiments or collider searches .
the energy scale of type - ii seesaw mechanism in @xmath0 is also too high for direct experimental tests .
in contrast to these high scale seesaw mechanisms , an experimentally verifiable and attractive mechanism that has been recently introduced into @xmath0 @xcite is the radiative seesaw @xcite where the quark - lepton unification has no role to play and additional suppression to light neutrino mass prediction occurs by loop mediation proportional to a small higgs quartic coupling that naturally emerges from a plank - scale induced term in the gut lagrangian .
the model predicts a rich structure of prospective dark matter candidates also verifiable by ongoing search experiments .
it has been further noted that this embedding of the radiative seesaw in @xmath0 may have a promising prospect for representing all fermion masses . a number of other interesting neutrino mass generation mechanisms including type - iii seesaw , double seesaw , linear seesaw , scalar- triplet seesaw have been suggested and some of them are also experimentally verifiable @xcite . in the context of non - susy @xmath0 in this work our purpose is to explore the prospects of another neutrino mass generation mechanism called the inverse seesaw @xcite which is also different from canonical or type - ii seesaw mechanism and has the potentiality to be experimentally verifiable because of the low scale at which it can operate although higher scale inverse seesaw models have been suggested @xcite . in a large class of models @xcite , the implementation requires the introduction of fermionic singlets under the gauge group of the model .
likewise , its implementation in @xmath0 introduces a new mass scale @xmath12 into the lagrangian corresponding to the mass matrix of the additional singlet fermions of three generations and the tev - scale seesaw requires this parameter to be small .
there is an interesting naturalness argument in favor of its smallness based upon exact lepton number conservation symmetry @xcite . below the tev scale , in the limit @xmath13
, the corresponding lagrangian has a leptonic global @xmath14 symmetry which guarantees left - handed neutrinos to remain massless .
the small value of @xmath15 essentially needed to match the neutrino oscillation data with the tev - scale inverse seesaw formula may be taken as a consequence of very mild breaking of the exact global symmetry .
thus the small value of the parameter protected by the exact lepton number conservation is natural in the t hooft sense @xcite .
however , inspite of such interesting naturalness argument , there has been no dynamical understanding of its origin so far although an interpretation using higgs mechanism has been given in the context of a model with extended gauge , fermion and higgs sectors @xcite . in a different class of @xmath0 models ,
the small singlet fermion mass parameter has been generated radiatively @xcite where more nonstandard fermions have been found to be necessary . in susy @xmath0 singlet fermions have also been used to derive new forms of fermion mass matrix while predicting standard fermion mass ratios @xcite and to obtain new seesaw formula for neutrinos while explaining baryon asymmetry of the universe through leptogenesis @xcite . while most of the inverse seesaw models need gauge singlet fermions under the sm gauge group or its extensions @xcite and the use of @xmath0-singlet fermions may point to the disadvantages of the corresponding gut - based models , extended electroweak theory based upon @xmath16 gauge symmetry @xcite contains such singlets in its fundamental representations . to give some examples of guts , one @xmath0-singlet fermion per generation
is automatically contained in the @xmath17 dimensional fermion representation of @xmath18 @xcite where @xmath19 under @xmath0 but @xmath20 under @xmath21 and , in the latter case , an additional discrete @xmath22 symmetry is needed to qualify it as a trinification gut model @xcite .
interesting properties of @xmath23 gauge theory including experimentally verifiable predictions at accelerator energies have been discussed @xcite .
gauge boson mediated proton decays are suppressed in @xmath23 type of models .
in addition to the rh neutrino and the other singlet fermion needed for inverse seesaw , these models ( @xmath23 and @xmath24 ) also contain @xmath25 nonstandard fermions per generation and no experimental data are yet available on their masses at low energies so as to persue the present bottom - up approach to derive the dirac neutrino mass matrix from fermion mass fits at the gut scale .
the same argument holds against any other model that may contain additional nonstandard fermions beyond the rh neutrino and the singlet - fermion needed for inverse seesaw . regarding the potentiality of so(10 ) motivated inverse seesaw in the visible sector , the same quark - lepton symmetry that forces canonical seesaw scale to be far beyond the experimentally accessible range , makes the tev - scale inverse seesaw predict observable nonunitarity effects as new physics signals verifiable at low and accelerator energies and at neutrino factories @xcite .
recently in a series of interesting investigations , using inverse seesaw mechanism , bhupal dev and mohapatra @xcite have shown that susy @xmath0 , besides admitting a low spontaneous breaking scale of @xmath26 gauge symmetry with right handed gauge bosons @xmath1 and @xmath2 accessible to lhc , is also capable of fitting all fermion masses and mixings at the gut scale while predicting observable nonunitarity effects .
the model has been also shown to account for the observed baryon asymmetry of the universe through leptogenesis caused due to the decay of tev scale masses of the pseudo dirac rh neutrinos @xcite .
currently considerable attention has been devoted to propose models with an extra neutral @xmath2 gauge boson which may also emerge from pati - salam or left - right gauge theories , or @xmath0 and @xmath24 grand unified theories , or also from string inspired models @xcite .
different from the supersymmetric @xmath0 model of ref.@xcite , here we adopt the view that there may not be any manifestation of supersymmetry at accelerator energies and that the actual parity restoration scale may be high . instead of both the @xmath1 and the @xmath2 boson masses being low , there may be only some remnants of high scale left - right symmetry or quark lepton symmetry manifesting at low and accelerator energies as smoking gun signatures such as the @xmath2 @xcite gauge boson and the associated nonunitarity effects of the tev - scale inverse seesaw . with this point of view in this work
we show that a non - susy @xmath0 with @xmath27 gauge symmetry at the tev scale and left - right gauge theory at higher intermediate scale , with or without d - parity , achieves precision gauge coupling unification , and predicts a low mass @xmath2 making them suitable for implementation of tev - scale inverse seesaw mechanism .
the model can also be verified or falsified through its predictions on observable nonunitarity effects and additional contributions to lepton flavor violations .
another testing ground for the model could be through the @xmath0 prediction on gauge boson mediated proton decay on which dedicated search experiments are ongoing .
we derive renormalization group equations in the presence of two intermediate gauge symmetries for running fermion masses and mixings , and determine the dirac neutrino mass at the tev scale from successful fits to the fermion masses at the gut scale . in this approach
we find a simple relation between the rh neutrino masses in the model .
we also point out a different type of relation in the partial degenerate case that permits much lower values of rh neutrino masses resulting in a cp - violating phase increased by @xmath28 orders larger than the degenerate case . some of our predictions include branching ratios for @xmath29 enhanced by @xmath30 orders . out of the two minimal models , while the intermediate scale d - parity conserving model is ruled out by proton decay constraint , the proton lifetime for @xmath6 in the intermediate scale d - parity nonconserving model is predicted to be well within the accessible range of ongoing search experiments .
we have also discussed simple extensions of the two models with longer proton lifetime predictions .
this method can also be implemented using pati - salam model or left - right models @xcite .
this paper is organized in the following manner . in sec.2
we briefly discuss the model and carry out gauge coupling unification and proton lifetime predictions in sec.3 . with a brief explanation of inverse seesaw mechanism in sec.4 we summarize relevant formulas encoding nonunitarity effects and lepton flavor violations . in sec.5
we discuss renormalization group evolution of fermion masses and mixings to the gut scale in the presence of nonsupersymmetric gauge theories @xmath31 and @xmath32 at intermediate scales . in this section
we also show how fermion masses are fitted at the gut scale and information on the dirac neutrino mass matrix is obtained .
nonunitarity effects are discussed in sec.6 with predictions on the moduli of relevant matrix elements . in sec.7
we give predictions on cp - violating parameters and lepton flavor violation where we also discuss possible limitations of the present models .
in sec.8 we provide a brief summary and discussion along with conclusion . in the appendix
a we provide beta function coefficients for gauge coupling unification while in appendix b we summarize derivations of renormalization group equations ( rges ) for fermion masses and mixings .
there has been extensive investigation on physically appealing intermediate scale models @xcite in non - susy @xmath0 .
although in the minimal two step - breaking of non - susy @xmath0 models @xcite we found no suitable chain with a sufficiently low scale to implement the inverse seesaw , the following chain with two intermediate gauge symmetries appears to be quite suitable , so(10 ) & & g_i + & & su(2)_l u(1)_r u(1)_b - lsu(3)_c [ g_2113 ] + & & su(2)_l u(1)_ysu(3)_c [ sm ] + & & su(3)_c u(1)_q , [ chain ] where we will consider two possibilities for @xmath33 . + as model - i @xmath34(g_{2l } \neq g_{2r})$ ] is realized by breaking the gut - symmetry and by assigning vacuum expectation value ( vev ) to the d - parity odd singlet in @xmath35 @xcite .
as the left - right discrete symmetry is spontaneously broken at the gut scale , the higgs sector becomes asymmetric below @xmath36 causing inequality between the gauge couplings @xmath37 and @xmath38 .
this model does not have the cosmological domain wall problem .
the second step of symmetry breaking takes place by the right - handed ( rh ) higgs triplet @xmath39 whereas the third step of breaking to sm takes place by the @xmath31-submultiplet @xmath40 contained in the rh doublet of @xmath41 .
it is well known that sm breaks to low energy symmetry by the sm higgs doublet contained in the bidoublet @xmath42 under @xmath32 which originates from @xmath43 of @xmath0 .
this is the minimal particle content for the model to carry out the spontaneous braking of gut symmetry to low - energy theory .
but a major objective of the present work is to explore the possibility of observable nonunitarity effects for which it is required to extract information on the dirac neutrino mass matrix ( @xmath44 ) from a fit to the fermion masses at the gut scale and this is possible by including two higgs doublets instead of one @xcite .
we assume these doublets to originate from two separate bidoublets contained in @xmath45 .
implementation of inverse seesaw also requires the minimal extension by adding three @xmath0-singlet fermions @xmath46 , one for each generation @xcite . as model - ii
, we treat the gut symmetry to be broken by the vev of the @xmath32-singlet @xmath47 which is even under d - parity @xcite .
this causes the higgs sector below gut - scale to be left - right symmetric resulting in equal gauge couplings in @xmath48 . for the sake of simplicity we treat the rest of the symmetry breaking patterns of model - ii similar to model - i and we assume the presence of three singlet fermions .
we call these two models , model - i and model - ii , as minimal models with two low scale higgs doublets in each .
we now examine precision gauge coupling unification for these two models .
in this section we examine gauge coupling unification in the minimal model - i and and the minimal model - ii and make predictions on proton lifetimes while we also predict the corresponding quantities in their simple extensions .
* a. unification in minimal models * it was shown in @xcite that with @xmath31 gauge symmetry at the lowest intermediate scale in @xmath0 there is substantial impact of two - loop effects on mass scale predictions in a number of cases . the one - loop and the two - loop beta - function coefficients for the evolution of gauge couplings @xcite for model - i and model - ii with two higgs doublets for each case
are given in appendix a. we have also included small mixing effects @xcite due to two abelian gauge factors @xmath49 in both the models below the @xmath50 scale . using @xmath51 and @xmath52 @xcite we find that with @xmath53 tev precision unification of gauge couplings occurs for the following values of masses at one - loop and two - loop levels for the model - i , m_u^ol&=&10 ^ 15.978 gev , m^ol_r^+=10 ^ 10.787 gev , _
g^ol=0.02253 , ( one - loop ) , + m_u&=&10 ^ 15.530 gev , m_r^+=10 ^ 11.15 gev , _
g=0.02290 ( two - loop).[solmod1 ] the rg evolution of gauge couplings at two - loop level is shown in fig .
[ fig1 ] exhibiting precision unification at @xmath54 gev . in model - ii coupling unification occurs with similar precision but at @xmath55 gev .
the decay width of the proton for @xmath6 is @xcite ( pe^+^0 ) & = & ( ) |a_l|^2||_h|^2(1+d+f)^2r .
[ width ] where @xmath56 $ ] for @xmath0 , @xmath57 the @xmath58 element of @xmath59 for quark mixings , @xmath60 is the short - distance renormalization factor in the left ( right ) sectors and @xmath61 long distance renormalization factor . @xmath62 degenerate mass of @xmath63 superheavy gauge bosons in @xmath0 , @xmath64 hadronic matrix element , @xmath65 proton mass @xmath66 mev , @xmath67 pion decay constant @xmath68 mev , and the chiral lagrangian parameters are @xmath69 and @xmath70 . with
@xmath71 gev@xmath72 estimated from lattice gauge theory computations , we obtain @xmath73 for model - i .
the expression for the inverse decay rates for both the minimal models is expressed as , ^-1(pe^+^0 ) & = & ( 1.0110 ^ 34 yrs)(0.012 gev^3_h)^2 ( 2.726 a_r)^2(1/43.6_g)^2 + & & ( 7.6f_q ) ( m_u ) ^4 , [ lifetime ] where the factor @xmath74 for @xmath0 . now using the estimated values of the model parameters in each case the predictions on proton lifetimes for both models
are given in table [ tab1 ] where the uncertainties in unification scale and proton lifetime have been estimated by enhancing the error in @xmath75 to @xmath76 level .
it is clear that with maximal value @xmath77 yrs . , model - i predicts the proton lifetime closer to the current experimental lower bound @xmath78 yrs . @xcite which is accessible to ongoing proton decay searches in near future @xcite . on the other hand model - ii
is ruled out at two - loop level as it predicts lifetime nearly two orders smaller .
the reduction of lifetime by nearly two - orders compared to one - loop predictions in both cases is due to the corresponding reduction in the unification scale by a factor of @xmath79 . [ cols="^,^,^,^,^,^,^,^,^ " , ]
each of the two @xmath0 models we have considered for inverse seesaw has two types of nonstandard gauge symmetries , @xmath32 or @xmath80 and @xmath31 . here
we derive rges for running yukawa and fermion mass matrices from which , following the earlier approach @xcite , we derive rges for the mass eigenvalues and mixing angles .
we define the rescaled @xmath81- functions 16 ^ 2= _ f_i.[b1 ] with @xmath31 symmetry the scalar field @xmath82 through its vev @xmath83 gives masses to down quarks and charged leptons while @xmath84 through its vev @xmath85 gives dirac masses to up quarks and neutrinos .
these fields are embedded into separate bi - doublets in the presence of @xmath32 and their vacuum structure has been specified in sec.4 .
we have derived the beta functions for rg evolution of yukawa matrices ( @xmath86 ) , fermion mass matrices ( @xmath87 ) , and the vacuum expectation values ( @xmath88 ) .
the rescaled beta functions are given below in both cases .
+ * @xmath31 symmetry : * _ y_u&=&[32y_uy_u^+ 12y_dy_d^+t_u-_i c_i^qg_i^2]y_u , + _ y_d&=&[32y_dy_d^+12y_uy_u^+t_d- _ i c_i^qg_i^2]y_d , + _ y_&=&[32y_y_^+12
y_ey_e^+t_u-_i c_i^lg_i^2]y _ , + _
y_e&=&[32y_ey_e^+12 y_y_^+t_d-_i c_i^lg_i^2]y_e , + _ m_u & = & [ 32y_uy_u^+12y_dy_d^-_ic^q_ig_i^2]m_u , + _ m_d & = & [ 32y_dy_d^+ 12 y_uy_u^-_ic^q_ig_i^2]m_d , + _ m_d&=&[32y_uy_^+12y_ey_e^-_ic^l_ig_i^2]m_d , + _
m_e&=&[32 y_euy_e^+12 y_y_^-_i c^l_ig_i^2]m_e , + where the beta - functions for vevs are _
-t_u]v_u , + _
v_d&=&[_ic^v_ig_i^2 -t_d]v_d,[matrge ] with t_u = tr(3y_u^y_u+y_^y _ ) , t_d = tr(3y_d^y_d+ y_e^y_e).[tutd ] the parameters occurring in these equations , and also in eq.([eigenbeta ] ) and eq .
( [ ckmgen ] ) given below are a & = & 32 , b=12 , a^=b^=0 , + c^q_i&=&(9/4,3/4,1/4,8 ) , c^l_i=(9/4,3/4,9/4,0 ) , + c^q_i&=&(0,0,1/4,8),c^l_i = ( 0,0,9/4,0 ) , c^v_i=(9/4,3/4,0,0 ) , ( i=2l,1r , bl,3c).[paramat1 ] * @xmath32 symmetry : * + following definitions of sec.4 in the presence of left - right symmetry .
the rescaled beta functions for rges of the yukawa and fermion mass matrices are _
y_u&=&(y_uy_u^+y_dy_d^)y_u+y_u(y_u^y_u+y_d^y_d ) + t_uy_u+ t_1y_d- _
i c_i^qg_i^2y_u , + _
y_d&=&(y_dy_d^+ y_uy_u^)y_d+y_d(y_d^y_d+ y_u^y_u)+t_dy_d+t_2y_u -_i c_i^qg_i^2y_d , + _ y_&= & ( y_y_^+y_ey_e^)y_+ y_(y_^y_+y_e^y_e)+t_uy_+t_1y_e -_i c_i^lg_i^2y _ , + _ y_e&=&(y_ey_e^+y_y_^)y_e+ y_e(y_e^y_e+y_^y_)+t_dy_e+ t_2y_-_i c_i^lg_i^2y_e , _ m_u & = & ( y_uy_u^+y_dy_d^)m_u+m_u(y_u^y_u+y_d^y_d ) -_ic^q_ig_i^2m_u+t_1m_d , + _ m_d & = & ( y_dy_d^+y_uy_u^)m_d+m_d(y_d^y_d+y_u^y_u ) -_ic^q_ig_i^2]m_d+t_2 m_u , + _ m_d&=&(y_uy_^+ y_ey_e^)m_d+m_d ( y_^y_+y_e^y_e ) -_ic^l_ig_i^2m_d+t_1m_e , + _ m_e&=&(y_ey_e^+y_y_^)m_e+ m_e(y_e^y_e+y_^y _ ) -_ic^l_ig_i^2m_e + t_2m_d,[g2213 ] where the rescaled beta functions for vevs @xmath89 are the same as in eq.([matrge ] ) with different coefficients @xmath90 defined below and functions @xmath91 and @xmath92 are the same as in eq.([tutd ] ) . other two traces entering in this case are _ 1&= & tr(3y_d^y_u+y_e^y _ ) , + t_2&=&tr(3y_u^y_d+y_^y_e).[hat12 ] the parameters occurring in these equations and also in eq.([eigenbeta ] ) and eq.([ckmgen ] ) given below are a&=&b=2 , a^=b^=1 , + c^q_i&=&(9/4,9/4,1/4,8 ) , c^l_i=(9/4,9/4,9/4,0 ) ,
c^q_i=(0,0,1/4,8 ) , + c^l_i&=&(0,0,9/4,0 ) , c_i^v=(9/4,9/4,0,0 ) , ( i=2l,2r , bl,3c ) . [ paramat2 ] then following the procedure described in @xcite , and using the definition of parameters in the two different mass ranges , given above we obtain rges for mass eigenvalues and elements of ckm mixing matrix @xmath93 which can be expressed in the generalized form for both cases , + * mass eigenvalues : * + _ m_i&=&m_i , i = u , c , t , + _ m_i&=&m_i , i = d , s , b , + _
m_i&=&m_i , i = e , , , + _
m_i&=&m_i , i = n_1,n_2,n_3 .[eigenbeta ] * ckm matrix elements : * + _ v _ & = & _ = u , c , t ; v _ + & & -_=d , s , b ; v_.[ckmgen ] then using third generation dominance , the beta functions for all the 9 elements are easily obtained for respective mass ranges where in addition to the parameters in the respective cases in eq.([paramat1 ] ) and eq.([paramat2 ] ) , @xmath94 in the mass range @xmath95 with @xmath31 symmetry , but @xmath96 in the mass range @xmath97 with @xmath32 or @xmath80 symmetry and , in the latter case , the nonvanishing traces @xmath98 are easily evaluated in the mass basis . 99 for reviews on different neutrino mass generation mechanisms with or without guts , supersymmetry , or flavor symmetry
see , for instance , g. altarelli , arxiv : 0905.2350[hep - ph ] ; g. altarelli , arxiv:1011.5342 ; g. altarelli , arxiv:1107.1980[hep - ph ] ; e. ma , proc .
9th workshop in high energy physics phenomenology ( whepp9 ) , 3 - 14 jan .
, 2006 , bhubaneswar , india ; pramana * j. phys .
67 * , 803 ( 2006 ) ; r. n. mohapatra _ et al _ , rept .
. phys . * 70 * , 1757 ( 2007 ) ; r. n. mohapatra , prog .
nucl . phys .
* 64 * , 307 ( 2010 ) ; s. raby _
et al _ arxiv : 0810.4551 [ hep - ph ] ; e. ma , arxiv:0905.0221 ; e. ma , neutrino mass , aip conf . proc .
* 116 * , 239 ( 2009);arxiv:0810.5574[hep - ph ] ; h. hettamansperger , m. lindner , and w. rodejohann , j. high energy phys . *
1104 * ( 2011 ) 123 ; g. senjanovic , riv .
nuovo cim . * 034 * , 1 ( 2011 ) .
j. c. pati and a. salam , phys . rev .
* d 8 * , 1240 ( 1973 ) ; ibid .
* d 10 * , 275 ( 1974 ) .
h. georgi and s. l. glashow , phys .
rev . lett . * 32 * , 438 ( 1974 ) .
h. georgi , particles and fields , _ proceedings of aps division of particles and fields , _ ed c. carlson , ( aip , new york , 1975 ) , p.575 ; h. fritzsch , p. minkowski , ann .
( berlin ) * 93 * , 193 ( 1975 ) .
d. chang , r. n. mohapatra and m. k. parida , phys .
* 52 * , 1072 ( 1984 ) ; d. chang , r. n. mohapatra and m. k. parida , phys . rev . * d 30 * , 1052 ( 1984 ) .
d. chang , r. n. mohapatra , j. gipson , r. e. marshak , and m. k. parida , phys .
d 81 * , 1718 ( 1985 ) .
p. minkowski , _ phys .
* b 67 * , 421 ( 1977 ) ; t. yanagida in _ workshop on unified theories , kek report 79 - 18 _ , p. 95
, 1979 ; m. gell - mann , p. ramond and r. slansky , _ supergravity _ , p. 315 .
amsterdam : north holland , 1979 ; s. l. glashow , _ 1979 cargese summer institute on quarks and leptons _ , p. 687
, new york : plenum , 1980 ; r. n. mohapatra and g. senjanovic , _ phys .
lett . _ * 44 * , 912 ( 1980 ) .
m. magg and c. wetterich , phys .
* b 94 * , 61 ( 1980 ) ; g. lazaridis , q. shafi and c. wetterich , nucl . phys . * b 181 * , 287 ( 1981 ) ; j. schechter and j. w. f. valle , phys . rev . *
d 22 * ( 1980 ) 2227 ; r. n. mohapatra and g. senjanovic , phys . rev . *
d 23 * , 165 ( 1981 ) .
marciano and g. senjanovic , phys . rev .
* d 25 * ( 1982 ) 3092 ; u. amaldi , w. de boer and h. furstenau , phys .
* b 260 * ( 1991 ) 447 ; p. langacker and m. luo , phys . rev .
* d 44 * , 817 ( 1991 ) .
n. g. deshpande , e. keith , and p. pal , phys .
d 46 * , 2261 ( 1992 ) .
r. n. mohapatra and m. k. parida , phys . rev .
* d 47 * , 264 ( 1993 ) .
d. g. lee , r. n. mohapatra , m. k. parida and m. rani , phys . rev . * d 51 * , 229 ( 1995 )
. m. k. parida , phys . lett .
* b 704 * , 206 ( 2011 ) .
e. ma , phys .
d 73 , 077301 ( 2006 ) .
s. k. kang and c. s. kim , phys . lett .
* b 646 * , 248 ( 2007 ) ; j. ellis , d. v. nanopoulos and k. olive , phys .
* b 300 * , 121 ( 1993 ) ; s. k. majee , m. k. parida , and a. raychaudhuri , phys . lett . * 668 * , 299 ( 2008 ) ; m. k. parida and a. raychaudhuri , phys . rev .
* d 82 * , 093017 ( 2010 ) ; s. k. majee , m. k. parida , a. raychaudhuri , and u. sarkar , phys . rev .
* d 75 * , 075003 ( 2007 ) .
see , for instance , f. deppisch , t. s. kosmas , and j. w. f. valle , nucl .
* b 752 * , 80 ( 2006 ) ; f. deppisch and j. w. f. valle , phys . rev . * d 72 * , 036001 ( 2005 ) ; j. garoya , m. c. gonjalez - gartia , and n. rius , j. high energy phys .
* 02 * ( 2007 ) 021 ; c. arina , f. bazzochi , n. forengo , j. c. romao , and j. w. f. valle , phys .
* 101 * , 161802 ( 2008 ) ; f. bazzochi , d. g. cerdeno , c. munoz and j. w. f. valle , phys . rev . * d 81 * , 051701 ( 2010 ) ; m. b. gavela , t. hambye , d. hernandez , and p. hernandez , j. high energy phys .
* 09 * ( 2009 ) 038 ; m. malinsky , t. ohlsson , z. -z .
xing , and h. zhang , phys .
* b 679 * , 242 ( 2009)l ; m. hirsch , t. kernreiter , j. c. romao , and a. villanova del moral , j. high energy phys . * 1001 * ( 2010 ) 103 ; arxiv:0910.2435[hep - ph ] ; f. bazzocchi , phys . rev . *
d 83 * , 093009 ( 2011 ) ; m. e. catano , r. martinez and r. ochoa , arxiv:1206.1966 ; m. hirsch , w. porod and f. staub , arxiv:1206.3516 ; x. g. he and e. ma , phys
* b 683 * , 178 ( 2010 ) .
a. abada , d. das and c. weiland , arxiv:1111.5836[hep - ph ] , a. abada , d. das , a. vicente , and c. weiland , arxiv:1206.6497[hep - ph ] ; p. s. b. dev , s. mondal , b. mukhopadhyaya and s. roy , arxiv : 1207.6542 .
s. khalil , phys .
* d 82 * , 07702 ( 2010 ) ; s. khalil , h. okada and t. toma , j. high energy phys .
* 1107 * ( 2011 ) 026 .
gu and u. sarkar , phys .
lett . * b694 * , 226 ( 2010 ) .
p. s. bhupal dev and r. n. mohapatra , phys .
d 81 * , 013001 ( 2010 ) ; p. s. bhupal dev and r. n. mohapatra , phys . rev . *
d 82 * , 035014 ( 2010 ) .
g. t hooft , in _ proceedings of the 1979 cargese summer institute on recent developments in gauge theories _ , edited by g. t hooft _ et al .
_ ( plenum press , new york , 1980 ) .
e. ma , mod .
. lett . * a 24 * , 2491 ( 2009 ) ; d. ibanez , s. morisi and j. w. f. valle , phys . rev . * d 80 * , 053015 ( 2009 ) .
e. ma , phys . rev .
* d 80 * , 013013 ( 2009 ) . c. h. albright and s. m. barr , phys .
lett . * 85 * , 244 ( 2000 ) ; ibid .
phys . rev . *
d 62 * , 093008 ( 2000 ) . for
a review see s. m. barr , int .
a18 * , 3971 ( 2003 ) .
s. m. barr , phys .
lett . * 92 * , 101601 ( 2004 ) ; c. h. albright and s. m. barr , phys . rev .
* d 69 * , 073010 ( 2004 ) .
a. g. dias , c. a. de s.pires , p. s. rodrigues da silva and a. sampieri , arxiv:1206.2590[hep - ph ] .
f. gursey , p. ramond and p. sikivie , phys . lett . *
b60 * , 177 ( 1976 ) ; e. ma , phys . rev .
d 36 * , 274 ( 1987 ) . s. l. glashow and k. kang , _ in grand unification .
proceedings , 5th workshop , providence _
, usa , april 12 - 14 , 1984 , edited by k. kang ( world scientific , singapore , 1985 ) .
c. cauet , h. pas and s. weisenfeldt , phys . rev .
* d 83 * , 093009 ( 2011 ) .
s. heinemeyer and e. ma , fortsch .
phys . * 58 * , 729 ( 2010 ) ; s. heinemeyer , e. ma , m. mondragon , and g. zoupanos , aip conf . proc . * 1200 * , 568 ( 2010 ) ; arxiv:0910.0501[hep - ph ] ; e. ma , j. phys .
* 315 * ( 2011)012006 .
s. blanchet , p. s. bhupal dev and r. n. mohapatra , phys .
d 82 * , 115025 ( 2010 ) .
p. langacker , rev .
bf 81 , 1199 ( 2009 ) ; p. langacker , arxiv:0909.3260 . for earlier work on @xmath2 boson with two - step breaking of left - right gauge symmetry and in guts
see m. k. parida and a. raychaudhuri , phys . rev .
* d 26 * , 2364 ( 1982 ) ; m. k. parida and c. c. hazra , phys .
* b 121 * , 355 ( 1983 ) ; m. k. parida and c. c. hazra , phys . rev .
d 40 * , 3074 ( 1989 ) . g. j. p. eboli , j. gonzalez - fraile and m. gonzalez - gartia , arxiv:1112.0316[hep - ph ] ; a. falkowski , c.grojean , a. kaminska , s. pokorski , and a. weiler , j. high energy phys .
* 1111 * ( 2011 ) 028 ; arxiv:1108.1183 [ hep - ph ] .
s. bertolini , l. di luzio and m. malinsky , phys .
* d80 * , 015013 ( 2009 ) ; s. bertolini , l. di luzio and m. malinsky , phys . rev . *
d 85 * , 095014 ( 2012 ) .
h. georgi , h. r. quinn , and s. weinberg , phys .
lett . * 33 * , 451 ( 454 ) .
d. r. t. jones , phys . rev .
* d 25 * , 581 ( 1982 ) .
l. lavoura , phys . rev .
* d 48 * , 2356 ( 1993 ) ; f. del aguila , g. d. coughlan and m. quiros , nucl . phys .
* b 307 * , 633 ( 1988 ) ; f. del aguila , g. m. masip and m. perez - victoria , nucl . phys . * b 456 * , 531 ( 1995 ) .
k. nakamura _
( particle data group ) , j. phys.*g 37 * , 075021 ( 2010 ) ; c. amsler _ et al .
_ ( particle data group ) , phys . lett . * b 667 * , 1 ( 2008 ) .
p. nath and p. filviez perez , phys .
rep . * 441 * , 191 ( 2007 ) ; b. bajc , i. dorsner and m. nemevsek , j. high energy phys .
* 0811 * ( 2008 ) 007 ; p. langacker , phys .
rept.*72 * , 185 ( 1981 ) .
h. nishino _
et al . _ ,
lett . * 102 * , 141801 ( 2009 ) ; h. nishino _ et al .
_ , phys . rev . *
d 85 * , 112001 ( 2012 ) .
m. shiozawa,neutrino.kek.jp / jhfnu / workshop2/ohp / shiozawa.pdf ; k. abe _ et al .
_ , arxiv:1109.3262[hep - ex ] .
b. a. dobrescu , k. kong and r. mahbubani , j. high energy phys .
* 07 * ( 2007 ) 006 ; y. bai and b. a. dobrescu , j. high energy phys .
* 1107 * ( 2011 ) 100 ; m. k. parida , p. k. sahu , and k. bora , phys . rev .
* d 83 * , 093004 ( 2011 ) ; m. k. parida , ref.@xcite ; r. n. mohapatra and m. k. parida , ref .
@xcite .
s. antusch , j. p. baumann , and e. fernandez - martinez , nucl . phys . *
b810 * , 369 ( 2009 ) ; s. antusch , m. blennow , e. fernandez - martinez , and j. lopez - pavon , phys . rev . * d 80 * , 033002 ( 2009 ) ; s. antush , c. biggio , e. fernandez - martinez , m. belen gavela , and j. lopez - pavon , j. high energy phys . * 10 * ( 2006 ) 084 ; d.v .
forero , s. morisi , m. tartola and j. w. f. valle , j. high energy phys . * 09 * ( 2011 ) 142 .
e. fernandez - martinez , m. b. gavela , j. lopez - pavon and o. yasuda , phys .
* b 649 * , 427 ( 2007 ) .
k. kanaya , prog .
phys.,*64*,2278 ( 1980 ) .
j. kersten and a. y. smirnov , phys . rev .
* d 76 * , 073005 ( 2007 ) . m. malinsky , t. ohlsson , h. zhang , phys . rev . * d 79 * , 073009 ( 2009 ) .
g. altarelli and d. meloni , nucl . phys .
* b 809 * , 158 ( 2009 ) .
f. del aguila and j. a. aguilar - saavedra , phys . lett .
* b 672 * , 158 ( 2009 ) ; f. del aguila and j. a. aguilar - saavedra and j. de blas , acta phys . polon .
* b 40 * , 2901 ( 2009 ) ; arxiv:0910.2720 [ hep - ph ] .
a. van der schaaf , j. phys .
* g 29 * , 2755 ( 2003 ) ; y. kuno , nucl . phys . *
b * , proc .
149 * , 376 ( 2005 ) .
et al . _ , [ t2k collaboration ] , phys .
lett . * 107 * , 041801 ( 2011 ) ; f. p. an _ et al._[daya - bay collaboration ] , phys . rev . lett . * 108 * , 171803 ( 2012 ) ; p. adamson _
et al . _ , [ minos collaboration ] , phys .
lett . * 107 * , 181802 ( 2011 ) ; j. k. ahn _
[ reno collaboration ] , phys .
* 108 * , 191802 ( 2012 ) .
a. ilakovac and a. pilaftsis , nucl
* b 437 * , 491 ( 1995 ) .
e. akhmedov , m. lindner , e. schnapka , and j. w. f. valle , phys . rev . *
d 53 * , 2752 ( 1996 ) ; arxiv : hep - ph/9509255 .
c. r. das and m. k. parida , eur .
j. * c 20 * , 121 ( 2001 ) .
conrad sanderson , _ armadillo : an open source c++ linear library for fast prototyping and computationally intensive experiments _ , technical report , nicta , 2010 .
g. l. fogli , e. lisi , a. marrone , a. palazzo , and a. m. rotunno , arxiv:1205.5254[hep - ph ] ; t. schwetz , m. tartola and j. w. f. valle , new j. phys .
* 13 * , 063004 ( 2011 ) ; d. v. forero , m. tartola and j. w. f. valle , arxiv:1205.4018[hep - ph ] j. adam _ et al . _
( meg collaboration ) , phys .
rev . lett.*107 * , 171801 ( 2011 ) ; arxiv : 1107.5547[hep - ex ] . k. hayasaka _
et al , _ ( belle collaboration ) , phys .
* b 666 * , 16 ( 2008 ) , arxiv:0705.0650[hep - ex ] . for
a review see f. r. joaquim , a. rossi , nucl .
* b 765 * , 71 ( 2007 ) . y. kuno ( prime working group ) , nucl .
b. * proc . suppl . *
149 * , 376 ( 2005 ) . v.
v. dixit and m. sher , phys . rev . *
d 40 * , 3765 ( 1989 ) . m. k. parida and p. k. patra , phys .
lett . * 66 * , 858 ( 1991 ) ; m. k. parida and p. k. patra , phys .
68 * , 754 ( 1992 ) ; m. k. parida , phys . rev . * d 57 * , 2736 ( 1998 ) .
h. an , p. s. bhupal dev , y. cai , and r. n. mohapatra , phys . rev . lett . * 108 * , 081806 ( 2012 ) . t. w. b. kibble , g. lazaridis , and q. shafi , phys . rev . * d 26 * , 435 ( 1982 ) . | recently realization of tev scale inverse seesaw mechanism in supersymmetric @xmath0 framework has led to a number of experimentally verifiable predictions including low - mass @xmath1 and @xmath2 gauge bosons and nonunitarity effects .
using nonsupersymmetric @xmath0 grand unified theory , we show how a tev scale inverse seesaw mechanism for neutrino masses is implemented with a low - mass @xmath2 boson accessible to large hadron collider .
we derive renormalization group equations for fermion masses and mixings in the presence of the intermediate symmetries of the model and extract the dirac neutrino mass matrix at the tev scale from successful gut - scale parameterization of fermion masses .
we estimate leptonic nonunitarity effects measurable at neutrino factories and lepton flavor violating decays expected to be probed in near future . while our prediction on the nonunitarity matrix element @xmath3 for degenerate right - handed neutrinos is similar to the supersymmetric @xmath0 case
, we find new predictions with significantly enhanced value of its phase @xmath4 when partial degeneracy among these neutrino masses is adequately taken into account by a constraint relation that emerges naturally in this approach .
other predictions on branching ratios and cp - violating parameters are discussed .
an important distinguishing characteristic as another test of the best identified minimal model is that the threshold corrected two - loop prediction on proton lifetime with maximum value @xmath5 yrs .
is accessible to ongoing search experiments for the decay @xmath6 .
simple model extensions with longer proton lifetime predictions are also discussed . * inverse seesaw mechanism in nonsupersymmetric @xmath0 , proton lifetime , nonunitarity effects , and a low - mass @xmath2 boson * .15 in * ram lal awasthi @xmath7 and mina k. parida @xmath8 * + _ @xmath9_centre of excellence in theoretical and mathematical sciences , + soa university , khandagiri square , bhubaneswar 751030 , india _ _ + _ @xmath7_harish - chandra research institute , chhatnag road , jhusi , + allahabad 211019 , india _ + _ |
one dimensional quantum wires provide an interesting opportunity to study mesoscopic physics in an interacting system . on the one hand , it is well established theoretically that interactions give rise to unique electronic properties in one dimension , described by the so called tomonaga
luttinger liquid ( tll ) model.@xcite a detailed comprehension of the remarkable interplay between interactions and disorder has been achieved.@xcite on the other hand , little attention has been paid to the role of the contacts which are known to influence strongly the transport properties of mesoscopic structures and quantum wires . in this respect ,
two simplified models were proposed recently .
either one connects a finite wire to the reservoirs by tunneling barriers,@xcite leading to the suppression of the ballistic conductance at low temperature due to the dramatic effect of the interactions .
alternatively , in the opposite situation , the wire is perfectly connected to one
dimensional leads,@xcite yielding a perfect conductance independent on the interactions of any range less than the wire length.@xcite the latter result contradicts the conductance reduction by the interactions predicted in a wire without contacts,@xcite and is in agreement with recent experiments on micron length quantum wires.@xcite the perfect conductance can be explained through an extension of landauer s approach to the interacting wire.@xcite the reservoirs are taken into account by the flux they inject which acts as an initial condition for the equation of motion of the density.@xcite the incident flux is perfectly transmitted for any range of interactions .
but reservoirs inject electrons which are not the proper modes of the wire .
this leads us to introduce intermediate noninteracting leads so that we can properly identify the injected and transmitted electronic flux . under theses circumstances an incident electron undergoes multiple internal reflections at the contacts due to change in interactions , leading to a perfect transmission into a series of spatially separated charges.@xcite while the ballistic conductance of a quantum wire can not reveal the tll character of the wire , the natural question one asks is if the other spectacular manifestations of the tll model , that are the signature of its non fermi liquid behavior , can be still observed .
it is the purpose of this paper to study these features .
some have already been addressed in the literature,@xcite but new interesting effects of the leads will be seen to appear .
we will also discuss recent experiments on quantum wires.@xcite beyond these points , in the present paper we develop a formal framework to deal with the properties of any inhomogeneous tomonaga luttinger liquid ( it ll ) .
an it ll can occur in circumstances more general than contact effects .
for example , a spatially varying effective interaction can be due to a varying width of the wire or to a nearby gate with a peculiar geometry .
one can also think of another ideal system , the edge states in the fractional quantum hall effect ( fqhe).@xcite the it ll model might be relevant to describe transitions between edges at different filling , or an edge state connected to a fermi liquid.@xcite this motivates us to investigate not only electrons with spin but also spinless electrons with external interacting leads .
let us now summarize our main results .
a typical feature of a tll is the separation of the charge and spin dynamics .
imagine that one injects a spin polarized flux of electrons into one external lead , and detects transmitted spin and charge on the second external lead .
since the latter is noninteracting , one might suspect that charge and spin recombine .
but we show that this is not the case : the charge and spin parts are separated even in the noninteracting leads .
such process has been first studied in refs . for spinless electrons .
the propagation is defined in terms of a quasiparticle current ( corresponding to laughlin quasiparticle in the fqhe ) , and to a superposition of electron - hole excitations in a quantum wire .
the basic buildingblock in the description of this effect is the scattering matrix of one quasiparticle at the contact between a tll of parameter @xmath12 and a second one of parameter @xmath13.@xcite the reflection coefficient is @xmath14 , thus is negative when @xmath15 .
such exotic result indicates an analogy with andreev@xcite reflection even for repulsive interactions . in connection with this andreev reflection
, we are here led to show a second aspect , the manifestation of proximity effects.@xcite it is known that a typical feature of the tll is the nonuniversal algebraic decay of correlations of different types , reinforcing the tendency towards a cdw or superconducting order which can not be of long range due to the importance of quantum fluctuations in one dimension .
it is important to know the effect of the finite size and of the leads on the fluctuations .
we will show here that the dominant tendency in the wire extends towards the external leads .
a new aspect of the `` mutual proximity effect '' manifests itself when the wire has a tendency towards a wigner crystal ( @xmath16 cdw ) .
this happens usually for coulomb interactions@xcite but also for very repulsive short range interactions . for electrons with spin with charge interaction parameter @xmath3 ,
the wigner crystal dominates the @xmath1 cdw at any temperature in an infinite wire , while the inverse holds at @xmath7 . with external leads connected to a wire with parameter @xmath3
, we show the existence of a crossover temperature depending on position @xmath17 below which the @xmath1 cdw dominates the wigner crystal .
@xmath17 has a non trivial powerlaw dependence on both the distance to the contacts and the wire length .
the behavior of density correlations can be probed through coupling to a backscattering potential .
the effect of weak impurities on the conductance of the wire with leads was found to be qualitatively similar to that without leads,@xcite and this seems in agreement with experiments.@xcite but compared to the usual tll , the conductance is affected in a nontrivial way by the contacts , especially when isolated barriers are considered.@xcite we show here that for electrons with spin this dependence becomes more crucial for very repulsive interactions because , as mentioned above , the external leads obscure the tendency towards the formation of a wigner crystal at low temperature .
apart from the specific model we consider , we improve the study of weak impurities .
usually , the renormalization equations in the presence of one barrier@xcite are derived for both the thermal length , @xmath18 , and the wire length , @xmath19 , infinite : their finite values are accounted for semi empirically by introducing @xmath20 as a cutoff . here , the renormalization equations are explicitly derived for finite lengths @xmath19 and @xmath21 . besides
, they are extended to a backscattering potential whose total spatial extension is less than @xmath21 and @xmath19 .
another important question concerns the fluctuations of the conductance in the presence of a random potential distribution .
this question can arise in experiments even if the mean free path is much larger than the wire length due to the imperfection and roughness on the boundaries . here
we show that the conductance is self - averaging at high temperature , more precisely @xmath8 is of the order @xmath22.@xcite but this ratio saturates at @xmath10 at temperatures low compared to @xmath23 .
this has to be taken seriously into account when one tries to infer precise powerlaws from the experimental value of @xmath24 .
let us now give the plan of our paper which does not follow the above sequence of results .
instead , we separate the paper in two parts : the first deals with spinless electrons which are also relevant for the fqhe , while we restore spin in the second part . for spinless electrons ,
we begin by describing the formalism and the inhomogeneous luttinger liquid model .
we discuss the discontinuous interaction case which is suitable to understand qualitatively the results ; it might be relevant for edge states in fqhe , but is has to be taken with some caution for quantum wires .
we then study the density
density correlation functions for two reasons .
first , in order to show how the dominant tendency in the wire extends towards the external leads , in analogy with proximity effects .
next , to investigate the role of backscattering . in particular
, the nonlocal correlation functions at finite temperature and at low frequencies are computed .
any extended non
random weak potential is then considered , for which the leading renormalization equation is derived explicitly at finite temperature .
the conductance is expressed perturbatively in the backscattering potential .
we discuss the random impurity case where the most important new result concerns the conductance fluctuations .
in the second part dealing with electrons with spin , most of the spinless treatment can be extended except for two important points .
first , the transmission process is studied explicitly .
secondly , we show in which way leads affect the tendency towards the wigner crystal that exists for very repulsive short range interactions .
this induces different regimes for the correction to the conductance in the presence of a backscattering potential .
finally , we discuss the experimental results obtained in quantum wires .
in a strictly one dimensional system with short range interactions between spinless electrons the low energy properties can be fully parameterized by two parameters @xmath25 and @xmath4 . without interactions @xmath26 and @xmath27 .
the hamiltonian can then be expressed as @xmath28,\ ] ] where @xmath29 is related to the long - wavelength component of the electron density through @xmath30 , and @xmath31 is the canonical momentum conjugate to @xmath29 , @xmath32=i\pi\delta(x - y)$ ] . taking into account the discreteness of the electrons , haldane derived the representation of the total electron density in terms of @xmath29:@xcite @xmath33 where @xmath34 .
the coefficients @xmath35 obey @xmath36 .
for the noninteracting case only @xmath37 and @xmath38 are nonzero , however in the presence of interactions higher coefficients appear .
the @xmath35 then can be calculated perturbatively , as done implicitly in ref .. consider now an interacting wire delimited by @xmath39 $ ] whose length will be denoted by @xmath40 connected perfectly to non interacting leads .
the global system is described by the hamiltonian ( [ h ] ) with spatially varying parameters @xmath41 , and @xmath42.@xcite we require @xmath25 and @xmath4 to be uniform on the external leads , i.e. for @xmath43 , taking values @xmath44 and @xmath45 .
even though the most relevant situation of noninteracting leads corresponds to @xmath26 and @xmath46 , it is interesting to keep @xmath45 for other possible applications , for instance the fqhe.@xcite we note that the absence of translational invariance in the quantum wire gives rise to an inhomogeneous chemical potential @xmath47 that can be incorporated by a translation in @xmath29.@xcite then one has to replace , in eq.([representation ] ) , @xmath48 let us discuss the microscopic arguments for the it ll model .
if one assumes that the screening does not take place in the same way in the two dimensional gas into which the wire opens and inside the wire we can argue that interactions are described by a function @xmath49 which is not translationally invariant .
but then the electronic momentum is not conserved anymore in the vicinity of the contacts .
indeed , this appears when one bosonizes the interaction hamiltonian @xmath50 , using @xmath51 from eq.([representation ] ) : the exponential terms in @xmath29 can not be ignored in general , and those terms violate momentum conservation .
but if @xmath52 where @xmath53 varies slowly on scales @xmath54 , one can reduce the hamiltonian to a quadratic form.@xcite strictly speaking , this condition is required to find a perfect conductance , as discussed in ref ..
if this condition is not met , and for not too strong interactions , the perfect conductance is recovered in the high temperature limit , but is reduced due to the abrupt change in interactions at low temperature .
for instance , when @xmath55 and @xmath56 are constants for @xmath57 , exponential terms in @xmath58 can not be ignored , and give rise to effective barriers determined by the jump of the interactions . nevertheless , due to the symmetry of the structure , and since one needs @xmath59 to be able to use the tll model , one can easily achieve resonances that suppress the role of these terms . for quantum wires opening adiabatically into a two dimensional gas ,
the situation of abrupt variations is not realistic ; it is adopted for mathematical convenience , and the results can be be trusted only far from the contacts compared to @xmath54 where the behavior should not depend on the profile of variations at the contacts .
we shall however look at the behavior at the contact because it is very similar to that in the strong tunneling limit of two coupled luttinger liquids:@xcite the effective parameter at the contact , yielding the local conductance @xcite and controlling the correlation functions@xcite is exactly the same as the parameter for the equivalent tll liquid in ref .. the situation for edge states in the fractional quantum hall effect ( fqhe ) might be different .
quadratic terms arise from tunneling between edge states , which is more difficult to achieve due to their spatial separation .
recently , the model with discontinuous parameters was shown to be suited to describe transitions between edges with different filling,@xcite or the connection between an edge state and a fermi liquid.@xcite in any case , most of the results we derive here can be extended to any profile of variations of the parameters .
it is well known that symmetry breaking phase transitions can not occur in one dimension .
but the interactions in an infinite wire enhance charge density or superconducting fluctuations depending on whether they are repulsive or attractive .
the natural question is to know whether such tendencies persist in the presence of leads .
in the present subsection , we would like to compute the correlation functions not only to answer this question , but also to explain the proximity effect,@xcite and to give the necessary tools for studying the role of backscattering in the next section , closely related to the tendency towards formation of a charge density wave .
let us write the density density correlation function , using eq.([representation ] ) : @xmath60}\right],\end{aligned}\ ] ] where @xmath61 is given in terms of the fundamental bosonic green function@xcite @xmath62 as @xmath63 @xmath64 is a cutoff different but in general of the order of @xmath65 .
we impose on the imaginary time @xmath66 , so that @xmath67 can be computed at finite temperature , without dependence on @xmath64.@xcitethe consequence of this cutoff procedure on analytic continuation is elucidated in appendix [ correlation ] . in eq.([rhorho ] ) @xmath68 is given by eq.([phi0 ] ) , and @xmath69 note that the first term ( in @xmath61 ) in eq.([rhorho ] ) can be obtained from the limit @xmath70 of @xmath71
. instead of writing out explicitly @xmath71 the imaginary part of its fourier transform in the low frequency limit @xmath72 will be of more use : @xmath73 where @xmath74 is a dimensionless function .
the function @xmath75 can be obtained from @xmath76 by multiplying @xmath42 by @xmath77 , thus we will often study @xmath76 , denoted @xmath78 for simplicity .
the time dependence of @xmath61 being due to the nonlocal part @xmath79 in eq.([utau ] ) , we can write ( for details see appendix [ correlation ] ) @xmath80 where @xmath81 and @xmath82 is a dimensionless time integral , eq.([gint ] )
. we can find the properties of @xmath83 for general smooth variations of @xmath4 and @xmath25 , and we have computed it explicitly for any @xmath84 in the discontinuous interaction case and at finite temperature . by deforming the contour of integration in the complex plane , the computation is made much easier , especially in the low temperature limit . in the high temperature limit ,
more steps are needed .
the results are given explicitly in appendix [ correlation ] .
in an infinite wire with uniform parameter @xmath4 , @xmath85 is a function of @xmath86 , and we denote it by @xmath87 .
it decreases exponentially at separations larger than @xmath88 , @xmath89 , while its local value has the typical powerlaw in temperature @xmath90 that diverges in the zero temperature limit when @xmath91 , indicating an enhancement of the density fluctuations compared to the noninteracting case . for interaction parameters
@xmath92 constant in the bulk of the wire and reaching their asymptotic values @xmath93 on the external leads , four general properties of @xmath78 can be shown .
the first and second one hold at high temperature , where comparison can be made with @xmath94 . 1 . at @xmath95
, we have @xmath96 where @xmath87 is the value of @xmath78 in an infinite wire with parameter @xmath4 .
thus we recover the translationally invariant behavior for any @xmath84 far from the contacts compared to @xmath21 .
2 . again , at @xmath95 , but for any location of @xmath84 , we have @xmath97 for @xmath98 , while the inverse inequality holds for @xmath99 .
3 . at any temperature , and for small separations compared to @xmath21 , @xmath78 factorizes : @xmath100 this is due to the fact that @xmath101 [ eq.([chic ] ) ] , up to corrections of order @xmath102 .
the local value of @xmath78 is almost constant on segments much less than the wire length and the thermal length .
if we denote @xmath103 then @xmath104 these equalities hold up to corrections of the order of @xmath105 .
let us now specialize to the case of discontinuous parameters .
first we introduce the quasiparticle density @xmath106 for @xmath107 which are the right and left going proper modes in a uniform tll .
they correspond to laughlin quasiparticle currents in edge states of the fqhe .
as we said before , the basic quantity for understanding the physics is the scattering matrix@xcite of quasiparticles at the contact of two perfectly connected tll with parameters @xmath4 and @xmath45 .
it acts in the space @xmath108 and is given by @xmath109 where @xmath110 is the reflection coefficient @xmath111 the matrix @xmath112 relates the outgoing flux to the injected flux , defined not in term of wavefunction amplitudes as in usual scattering approaches , but directly in terms of the quasiparticle current @xmath113 .
the local effective parameter ( yielding a local kubo conductance@xcite ) is given by @xmath12 multiplied by the transmission coefficient,@xcite @xmath114 let us first discuss the high temperature limit .
it is worth noting that a lower ( upper ) bound to @xmath85 , completing eq.([majoration ] ) , is given by @xmath115 for @xmath116 ( @xmath117 ) . in eq.([chic ] ) , the nonlocal part @xmath83 is given by eq.([cf1 ] ) in the limit @xmath95 , while it simplifies to a constant [ eq.([clow ] ) ] in the limit @xmath118 , in accordance with property ( [ decouple ] ) .
the green function @xmath67 is given by eq.([apgz ] ) .
instead of writing explicitly the factor @xmath119 that enters in eq.([chic ] ) , we use the local value of @xmath78 .
this is equivalent as far as the dominant behavior is concerned because @xmath120 is a slowly varying function of @xmath121 ( cf .
appendix [ correlation ] ) .
thus we drop it from @xmath122 in the high temperature limit @xmath123 ) , the multiple reflections caused by the change in interactions do nt affect the correlation functions due to the lack of thermal coherence along the wire .
only one reflection is felt within a thermal length @xmath124 from the contacts .
we now define @xmath125 the time it takes for a quasiparticle emanating at @xmath121 to reach the closest contact , measured in units of @xmath64 . for @xmath126 [ eq.([tx ] ) ] , we find @xmath127 again , for points far from the contacts compared to @xmath21 , this expression coincides with that obtained in an infinite tll with parameter @xmath4 , @xmath128 .
at the contacts , we find @xmath129 [ cf eq.([chia ] ) ] ( this amounts to setting @xmath130 in eq.([vrenorm ] ) , using @xmath131 ) . in the limit of temperatures low compared to @xmath132 , the multiple reflections affect the correlation functions , thus the external leads with parameter @xmath45 determine the temperature dependence .
but there is a nontrivial dependence on the wire length as well as on the distance of the barrier from the contacts : @xmath133 this gives the behavior for @xmath126 .
@xmath134 can be obtained by setting @xmath130 in this expression .
remember that one obtains @xmath135 from @xmath136 ( in both temperature limits ) by multiplying both @xmath4 and @xmath45 by @xmath77 , which leaves @xmath110 [ eq.([gamma ] ) ] unchanged . clearly , the @xmath137 cdw is less important than the @xmath1 one .
further , if electrons in the external leads are non interacting , @xmath46 , the @xmath137 cdw are suppressed in the zero - temperature limit because @xmath138 [ eq.([integration ] ) ] .
let us first recall an exotic phenomena emanating from this model , the analogy with andreev reflection.@xcite [ [ andreev - reflection ] ] andreev reflection + + + + + + + + + + + + + + + + + + the reflection coefficient of a quasiparticle from a tll with parameter @xmath12 incident on another one with parameter @xmath13 is given by @xmath110 , eq.([gamma ] ) , and is _ negative _ for @xmath15 , i.e. a partial quasi - hole is reflected back .
this is similar to andreev reflection of an incident electron on a normal superconductor interface .
the similarity is the closest for @xmath139 and @xmath140 , because the wire has a tendency towards superconductivity , but it holds more generally even for repulsive interactions , for instance for a quasiparticle in the wire with @xmath91 incident on the noninteracting lead .
it is worth noting that in the case of two semi
infinite wires with @xmath12 and @xmath13 as parameters , and @xmath15 , the local kubo conductance at the interface is given by the transmission coefficient from @xmath141 to @xmath142 multiplied by @xmath12 , i.e. by @xmath143 .
this verifies @xmath144.@xcite this is very similar to the inequalities for an n - s interface , if one interprets @xmath12 as the conductance ( kubo ) of the normal side , @xmath145.@xcite finally , let us comment on a recent paper@xcite based on the model in ref . of two tlls connected by a tunneling term .
the bosonized theory was associated with conformal field theories to describe quasiparticles in term of soliton states .
interestingly , in the strong tunneling limit , the scattering matrix for these states at the interface turns out to be the same as eq.([scattering ] ) .
this gives a stronger foundation of the quasiparticle scattering scheme we introduced previously.@xcite [ [ proximity - effect-1 ] ] proximity effect + + + + + + + + + + + + + + + + the reflection at the contact gives rise to a proximity effect for both superconducting or cdw correlations.@xcite this can be seen computing the correlation functions on the external leads .
in particular ,
when @xmath95 , there is no coherence between the endpoints of the wire , and it is sufficient to consider half the system .
the wire and one lead have a symmetric role ; it is sufficient to permute @xmath4 and @xmath45 in eq.([vrenorm ] ) in order to find @xmath78 for @xmath146 , and to take @xmath147 . by the way ,
this situation is relevant for the fqhe , if two edge states with different filling are connected.@xcite at @xmath148 , instead of eq.([integration ] ) inside the wire , outside the wire one has @xmath149 up to @xmath150 . beyond a distance of the order of @xmath19 in one external lead ,
we recover the simple law @xmath151 . for clarity
, we discuss the consequence for noninteracting leads , @xmath46 .
then if @xmath91 , the density density correlations in the bulk are similar to those in an infinite wire , but are reduced when one goes to the contacts because @xmath152 decreases for @xmath153 .
besides , this enhancement extends in the external leads up to a distance @xmath154 [ eq.([lmin])].@xcite for @xmath155 , the pairing correlation function , which can be obtained from @xmath78 if one replaces @xmath4 by @xmath156 and @xmath45 by @xmath157 , is enhanced up to a distance @xmath158 .
this increases with @xmath4 , i.e. where interactions are more attractive .
this is reminiscent of the proximity effect.@xcite in general , we can show that the proximity effect extends up to the minimum length scale at hand .
we must note that this holds even for a smooth profile of @xmath25 and @xmath4 , as for instance the properties of @xmath78 exposed before are general .
we study now the role of impurities in the wire .
the conductance with impurities was shown to depend on the interactions and is generally affected by the external leads.@xcite the main goal of this part is to give a general scheme to treat the effect of a weak backscattering potential @xmath159 in a non translational invariant system and at finite temperature .
the coupling of the conduction electrons to impurities with potential @xmath159 is @xmath160 , where @xmath51 is given by eq.([representation ] ) .
it is then convenient to do an integration by parts giving @xcite @xmath161 the forward scattering term @xmath162 is absorbed by a translation of @xmath29 by @xmath163 which has to be included in the scalar function @xmath68 , eq.([phi0 ] ) .
note that the impurity hamiltonian ( [ nouvelle ] ) is similar to that considered usually if we replace our @xmath164 by @xmath165 .
for instance , if @xmath166 has the form of a kink , it corresponds to what is commonly treated as one barrier . in order to derive the renormalization equations , the exact partition function @xmath167 at finite temperature
is expanded in terms of @xmath168 : @xmath169 where @xmath170 is given by eqs.([gtau],[utau ] ) , and the integration runs over @xmath171 and @xmath172 .
@xmath167 describes a neutral gas of integer charges restricted to a cylinder whose perimeter is @xmath173 , and whose height is determined by the spatial extension of @xmath168 , reducing to a circle when @xmath168 is local .
the charges interact via @xmath174 , eq.([utau ] ) .
this is not the coulomb interaction but an infinite series of logarithmic terms related to the transmission process discussed earlier ( see eq.([apgz])).@xcite the renormalization procedure is implemented by increasing the cutoff to @xmath175 , where @xmath64 is the bare cutoff @xmath176 , and modifying the parameters in order to keep @xmath167 invariant .
it is immediate to derive the leading order equation for @xmath166 . according to eq.([utau ] )
, the cutoff appears in the local part of @xmath61 .
the change of @xmath177 due to a change of the cutoff by @xmath178 is @xmath179d\tau_0(l),\ ] ] the exponential of such a term for a pair @xmath180 factorizes and , once the global neutrality is used , can be absorbed separately into @xmath181 and @xmath182 .
this gives the leading - order flow equations for @xmath159 _ explicitly at finite temperature _
@xmath183 note that in the limit of zero temperature , an infinite uniform wire with parameter @xmath4 , and a purely local @xmath166 , the fourier transform of this equation yields the well - known flow equation for @xmath184 , @xmath185,@xcite provided @xmath186 is independent on the cutoff as we required .
in general , one expects @xmath159 to renormalize the interactions . in the extreme case of a local barrier ,
it was shown by integrating out degrees of freedom away from the barrier that the interactions are not renormalized.@xcite we can both recover and generalize this result in a different way , by using the expansion above at finite temperature and in the finite wire : new interaction terms @xmath187 between the charges are generated by the renormalization , but they decay faster than @xmath188 at long time scales . for instance , in the zero - temperature limit , and for a barrier in the center of the wire , @xmath189 is corrected by @xmath190 , where @xmath191 is renormalized by @xmath166 . at finite temperature , we have @xmath192 , i.e. @xmath187 is still decaying faster than @xmath61.@xcite let us now discuss a more extended potential than one barrier , but whose total extension obeys @xmath193 [ eq.([lmin ] ) ] .
it turns out that when one goes to a cutoff such that @xmath194 , the partition function ( [ zprem ] ) becomes identical to that of a local barrier.@xcite this is related to the property ( [ chiloc ] ) .
then it is tempting to assume that the effect of such a potential would be similar to that of a local barrier at length scales larger than the potential extension @xmath195 .
but one has to check that the renormalization from the bare cutoff up to @xmath195 does not modify the local interaction parameter , i.e. @xmath42 at points where @xmath196 .
this would induce non translational invariant effective interactions even if the bare @xmath4 is uniform .
we can for instance show that @xmath68 [ eq.([phi0 ] ) ] is renormalized by a complicated complex function , whose equation is rather lengthy to write .
thus we think this point needs a more thorough study .
finally , it is worth writing the next order corrections to eq.([dvmsurdlx ] ) in the case where the partition function becomes equivalent to that of a barrier and therefore the integrations over @xmath171 can be done explicitly in eq.([zprem ] ) . for given @xmath197 ,
one has to add to this equation @xmath198 up to nonuniversal prefactors depending on @xmath199.@xcite if we ignore renormalizations of the interaction , as is surely correct for a local barrier or for a weak enough extended potential , the equation ( [ renormalisationspin ] ) can be integrated straightforwardly . in the absence of an external energy scale , the unique limitation on increasing the cutoff comes naturally from the fact that one has to put at least two charges on the same cylinder of radius @xmath200 , thus we stop at @xmath201 .
then the renormalized potential is simply @xmath202 where @xmath136 , eq.([chig ] ) , has been studied in the previous subsection .
one can also consider the fourier transform @xmath203 of the potential .
if the extension of @xmath166 is less than @xmath204 , we can use eq.([chiloc ] ) to factor out @xmath78 from the integral , thus allowing us to recover the dominant term of the renormalized fourier component of @xmath166 : @xmath205 by simple multiplication by @xmath206 . note that whenever @xmath68 [ eq.([phi0 ] ) ] can be replaced by @xmath207 , @xmath208 .
however the nonlocality of @xmath75 has to be taken into account when one looks to the next leading term.@xcite the derivation above is valid for any profile of parameters . in case @xmath4
varies from its external value @xmath45 towards a plateaus at @xmath4 inside the wire , we can show that the dependence of @xmath136 on @xmath209 is monotonic .
its monotony is determined for any @xmath121 by the sign of @xmath210 at @xmath211 , thus is similar to that in an infinite wire .
the renormalized @xmath168 , eq.([vrenormalized ] ) obtained at a given temperature goes up ( down ) when temperature is lowered for @xmath91 ( @xmath155 ) . at @xmath148 ,
the @xmath209 dependence is controlled by the sign of @xmath212 , in particular @xmath168 saturates for @xmath46 at @xmath148 . in the case of discontinuous parameters ,
the explicit value of @xmath78 is given by eqs.([vrenorm ] ) and ( [ integration ] ) for the high and low temperature regime .
the renormalized @xmath166 , depending on the location of the barriers , is shown in fig.[f : allcourbes ] for the case @xmath46 , @xmath91 .
we point out that these curves are not inferred from cutting the scaling at @xmath209 or at @xmath132 , but result from the renormalization at finite temperature and for the finite wire .
note that the saturation at @xmath148 occurs only for @xmath46 and for @xmath213.@xcite let us first recall that the conductance of a wire without impurities @xmath214 is independent of interactions if the external leads are noninteracting .
the conductance can be related to the transmission , which turns out to be perfect , thus generalizing landauer s approach to an interacting wire.@xcite one can also use the kubo formula,@xcite provided the external leads are noninteracting , and one is limited to the linear and stationary regime .
we have to stress an important point : the external field has to be used when interactions are included exactly in the hamiltonian , and not the internal electric field modified by the interactions .
this is clear for instance through the equation of motion method we have developed in ref . to study transport , for which the _ external field _
forms a source term . the same conditions to justify the kubo formula hold with impurities ,
if one assumes that the reservoirs impose an external bias @xmath215 ( that can be maintained in the noninteracting leads ) independently of the current . then the stationary current is given by @xmath216 where @xmath217 is the nonlocal conductivity , related to the fourier transform of @xmath67 [ eq.([gtau ] ) ] computed now with the total hamiltonian @xmath218 , and continued to real frequency , @xmath219 the uniformity of the current as well as time reversal symmetry require the zero frequency limit of @xmath217 to be uniform.@xcite thus @xmath220 .
if the external leads are interacting , one can not impose an external value of the potential , rather the potential is now renormalized by the interactions.@xcite nevertheless , the spatial separation of the two edges in a hall bar can lead to different couplings with reservoirs as discussed in , and it is also possible to measure locally the potential on each edge .
thus it is still of interest to let the external parameter @xmath45 arbitrary . in appendix [ dyson ] , we derive a novel exact dyson equation for the conductivity in a non translationally invariant system .
this is very suitable to write the perturbative expression for the conductance@xcite @xmath221 where @xmath222 is given by @xmath223 and @xmath224 is the contribution from backscattering of @xmath197 electrons , @xmath225,\end{aligned}\ ] ] where @xmath75 is defined by eqs .
( [ utau],[gtau],[imaginary],[xm ] ) .
remember that the forward scattering contribution is included in @xmath68 [ eq.([translation ] ) ] .
we have already obtained general properties of @xmath85 .
if parameters change abruptly , one can inject the explicit form ( appendix [ correlation ] ) to express the conductance at any temperature , for sufficient weak potential _ with any extension_. if the potential extension is much less than @xmath226 [ eq.([lmin ] ) ] , we can use eq.([chiloc ] ) to show easily that the dominant term of equation ( [ eqrmnew ] ) reduces to the square of the renormalized potential @xmath227 where @xmath228 is given by eq.([vm ] ) .
as long as @xmath229 ( i.e. out of resonance ) , the dominant term in @xmath222 is given by @xmath230 .
on resonance , the dominant term comes still from @xmath213 or from @xmath231;@xcite this depends on the values of @xmath4 and @xmath45 , as well as the extension and strength of the potential .
it appears that at low temperature , because the @xmath16 cdw correlation function acquires a factor @xmath232 [ eq.([integration ] ) with @xmath4 and @xmath45 multiplied by @xmath233 , one must have @xmath234 for @xmath222 to vanish at zero temperature ( for a short enough wire ) .
remarkably , for @xmath46 , the low temperature conductance is still @xmath235 on resonance , _ even for @xmath236_. we will not discuss the resonance in further detail here , even though one has a variety of behaviors , and we refer to the explicit expression of @xmath222 , eqs.([rsommerm][eqrmnew ] ) , and to the computation of appendix [ correlation ] . out of resonance , and for discontinuous parameters , @xmath46 , @xmath222
is given by the same curves as those yielding the renormalized @xmath166 in fig.[f : allcourbes ] .
the wire has to be short enough so that @xmath222 stays weak enough when reaching the plateaus below @xmath132 , otherwise the perturbative computation breaks down at a temperature above @xmath132 .
we refer to ref . for the strong barrier case .
now we consider random impurities distributed all over the wire . for simplicity ,
we limit ourselves to the case of a gaussian distribution , @xmath237 , and @xmath238 by averaging eq.([eqrmnew ] ) over disorder , one obtains the leading term of @xmath222 [ eq.([rsommerm ] ) ] coming from the @xmath213 contribution @xmath239 where @xmath240 ^ 2/(u k_f)^2 $ ] .
the forward scattering part in @xmath241 is @xmath242 [ eq.([translation ] ) ] , and thus can be dropped when averaging .
thus the dependence of @xmath243 on @xmath121 is due to the inhomogeneity of the interactions .
it could also be induced by a space dependent disorder strength @xmath244 .
nevertheless , we will assume @xmath245 for simplicity , in which case @xmath243 is uniform and is equal to @xmath246 .
this holds for instance in the bulk of the wire , and thus this is a plausible approximation whenever the impurities in the bulk dominate .
note that we have obtained eq.([rdes ] ) by performing first a double integration by parts in eq.([eqrmnew ] ) and then retaining only the term with no derivative of @xmath78 .
the explicit integration of the function @xmath78 can be done but is tedious .
however , the results can be understood easily.@xcite there are two main contributions : one from the bulk that behaves , for @xmath211 , as @xmath247 , the other from the contacts behaving as @xmath248 . for @xmath148 , up to nonuniversal constants , these contributions become@xcite @xmath249\ ] ] where @xmath250 is given by eq.([ka ] ) .
let us discuss the case of noninteracting leads , @xmath46 . in this case
, the contact contribution dominates for very attractive interactions , @xmath251 , and for not too high temperatures ( or not too long wires ) . inspecting eq.([lowmathr ] )
, we see that @xmath252 saturates at @xmath148 at a value @xmath253 for @xmath254 , but that it increases with wire length for @xmath255 . in the latter case ,
the perturbative computation is valid only for @xmath256 where @xmath257 this coincides with the localization length inferred by scaling arguments in .
let us now compute the variance of the conductance @xmath258 .
clearly , it is also equal to the variance of @xmath222 which can be expressed perturbatively , using eq.([eqrmnew ] ) , as @xmath259 \chi^2(x , y).\ ] ] here we have used eq.([gauss ] ) to perform the average of @xmath260 , have dropped the product of two correlation functions for different pairs of integers , and neglected the role of the derivative in eq.([nouvelle ] ) .
it is interesting to consider first the low temperature limit @xmath148 , where a general inequality , valid for arbitrary profile @xmath42 , can be shown . using the factorization property ( [ decouple ] ) which holds now for any @xmath84 in the wire because @xmath261 , we can write eq.([variancer ] ) as @xmath262 the term for @xmath263 on the right hand side is less than that for @xmath264 , this latter is nothing but @xmath265 [ eq.([rdes ] ) ] .
thus we have the interesting inequality @xmath266 in particular , this shows that @xmath267 if @xmath268 is simply @xmath207 , this is a very good estimate of @xmath6 because the integral for @xmath263 in eq.([variance ] ) contains a rapidly oscillating function , @xmath269 , while @xmath78 varies slowly compared to @xmath270 . besides , @xmath271 in order to validate the bosonization , so that the integral corresponding to @xmath263 in eq.([variance ] ) can be neglected .
this justifies eq.([ratiolow ] ) , showing that the fluctuations of @xmath222 are of the same order as its average value . in the special case of short range interactions with discontinuous parameters ,
@xmath6 is obtained as the square of eq.([lowmathr ] ) .
we now consider the high temperature limit , @xmath211 . to simplify the discussion , we restrict ourselves to the case @xmath91 , which is probably the case in a quantum wire
. then the bulk contribution dominates over the contact effects , so that we can restrict the integral in eq.([variancer ] ) to @xmath272 , and therefore @xmath273 . over this segment
, @xmath78 depends only on @xmath86 ( see eq.([chibulk ] ) ) , and we can write , after a change of variables @xmath274 , @xmath275 where @xmath276 ^ 2 dv,\ ] ] and @xmath277 is given by eq.([ckv ] ) . since @xmath278
, we have extended the integral approximatively to infinity .
but since we are restricted to temperatures much less than the fermi energy , we have @xmath279 , thus the oscillatory part of @xmath280 can be neglected and @xmath280 tends to a constant .
the ratio @xmath281 is now much less than unity , contrary to the low temperature limit , thus the relative fluctuations of @xmath11 are small .
it is worth noting that one can extrapolate the ratio ( [ ratiohight ] ) obtained for @xmath278 to @xmath282 , to recover the same order of magnitude as eq.([ratiolow ] ) at which @xmath283 saturates at low temperature , i.e. for @xmath284 .
let us take into account the spin of the electrons , which is necessary when dealing with quantum wires , and to confront the theory with experiments .
this case might eventually also be relevant for two - channel edge states in fqhe .
the typical feature for interacting electrons with spin is the separation of the charge and spin degrees of freedom at low energy .
the density for each spin component @xmath285 ( @xmath286 denoting the spin up or down ) has a long - wavelength part that is related to a boson field @xmath287 by @xmath288 .
the total density of electrons with spin @xmath289 can be written as eq.([representation ] ) where @xmath29 has now the subscript @xmath289 .
the charge and spin fields are defined by @xmath290 the hamiltonian describing the low - energy properties can be decoupled in a charge and spin parts , @xmath291 , where @xmath292\ ] ] for @xmath293 .
the boson fields are related to the charge and spin density ( @xmath51 and @xmath294 ) through @xmath295 and @xmath296 .
@xmath297 is the momentum density conjugate to @xmath298 . in the absence of interactions , @xmath299 .
recall that an additional term has normally to be added to @xmath300 [ eq.([hnuspin ] ) ] corresponding to the backscattering of electrons of opposite spin , @xmath301 . for spin - invariant interactions ,
@xmath302 renormalizes to zero if it is initially positive , while a spin gap is opened if it is negative .
then @xmath303 scales respectively to @xmath141 or zero which are the only values consistent with su@xmath304 symmetry . the inhomogeneous tll model is now characterized by @xmath121 dependent functions @xmath305 and @xmath306 . for simplicity
, we take noninteracting leads , otherwise the parameters would be too numerous ; indeed this is the most relevant situation for quantum wires . nevertheless , the treatment of @xmath302 is not trivial for non translationally invariant interactions because @xmath307 now is space dependent .
we have derived the renormalization equations for this case , @xcite nevertheless their integration is quite involved .
we can however draw qualitative but not firm conclusions , restricting ourselves to spin isotropic interactions .
the most difficult case corresponds to attractive interactions . as in the uniform case
, we expect a spin gap to develop , but it is not clear how it extends spatially and how is it affected by the leads in the low temperature regime .
the way @xmath303 would vary from @xmath141 on the external leads to @xmath308 inside the wire is not clear neither ; this is an interesting problem to solve . for repulsive interactions ,
the inhomogeneous @xmath303 is expected to renormalize towards a value respecting the su(2 ) symmetry @xmath309 . loosely speaking ,
this is easier to study because the leads have @xmath310 , so that they do nt prevent this scaling and rather favor it .
this seems the most relevant situation for quantum wires , even though we can sometimes allow for @xmath303 to take any value .
the first issue we consider is the transmission process of an incident electron .
the case of electrons without spin was discussed in refs . , and this can be extended easily to electrons injected from reservoirs .
let us imagine that we inject an electron of definite spin in the left lead .
this corresponds to create a kink in both @xmath311 and @xmath312 whose time evolution has to be solved given the initial conditions .
the equations of motion for these fields are decoupled , and require their continuity at the contacts as well as that of @xmath313 . in particular , both the charge and spin current , @xmath314 are conserved .
this is because we neglect interactions that violate the conservation of @xmath315 and @xmath316 , corresponding respectively to the umklapp process and the backscattering of electrons with opposite spin ( @xmath302 is irrelevant for repulsive interactions , see ref . for the it ll ) . in view of the above continuity requirements , the charge and spin of the incident electron
are reflected at the first contact with two different coefficients similar to eq.([gamma ] ) with a subscript @xmath317 , @xmath318 the transmitted charge and spin propagate at different velocities @xmath319 and thus reach the second contact where they get partially transmitted at different times @xmath320 . since the transmitted charge and spin propagate at the same fermi velocity in the right non interacting lead , they will stay _ spatially separated _ , at a distance @xmath321 ( fig.[f : refspin ] ) .
due to subsequent internal reflections , we end up with a series of spatially separated partial charge and spin spikes on the two non interacting leads .
the transmitted spikes correspond to a complicated superposition of electron - hole excitations .
note that in the relevant case of spin - invariant repulsive interactions , @xmath310 , the spin part is not reflected but gets directly transmitted to the right lead , while the charge undergoes the multiple reflection process . at times very long compared to @xmath322 ( @xmath323 ) , the series of transmitted charge ( spin ) spikes sum up to unity .
thus the transmission of an incident spin polarized flux is perfect . in an infinite wire , interacting electrons with spin
have a quite involved `` phase diagram '' ; here we do not intend to study it entirely when the wire is connected to leads .
we give some qualitative results and then discuss thoroughly the tendency towards the formation of a wigner crystal . as for spinless electrons , we expect the dominant tendency to extend towards the leads in analogy with the proximity effect , but the leads can also intervene in the competition as will be shown later , especially at low enough temperature .
let us first discuss the case where superconducting order can take place , as is the case in an infinite wire with @xmath324 .
one has to distinguish singlet ss , corresponding to the spin gap case , @xmath308 and triplet ss , @xmath309 .
if we naively take these values for @xmath303 inside the wire , then we can show that : for @xmath308 inside the wire in order to take into account the spin gap , we can show that the tendency towards singlet super - conductivity holds for @xmath324 , but the proximity effect towards the external leads shows up only for @xmath325 .
remarkably , for @xmath326 , an incident electron with spin up is reflected back into one hole with spin down , and two quasi - particles of opposite spin , of charge unity and moving at velocity @xmath327 are transmitted , recalling andreev reflection .
@xcite for @xmath324 , and @xmath328 , we can show that triplet superconductivity develops and extends towards the external leads , but the andreev reflection is more subtle to interpret.@xcite in the following , we focus on the situation @xmath329 , @xmath309 , as it should be for quantum wires , and determine the dominant cdw in the density
density correlation function .
let us compute the density - density correlation function .
in order to express the density , we can superpose eq .
( [ representation ] ) for the spin up or down , thus with the additional index @xmath330 or @xmath294 , then express @xmath298 as function of @xmath331 .
this yields @xmath332,\ ] ] where the sum runs over integers @xmath333 and @xmath334 of the same parity,@xcite and @xmath335 .
normally only @xmath336 are allowed .
but the other harmonics with @xmath337 are generated by renormalization in the induced density that develops as a response to an external perturbation.@xcite the fermionic correlation functions can be expressed through the bosonic correlation function @xmath338 ( and its dual ) for @xmath317 as in eq.([utau ] ) with @xmath41 and @xmath42 indexed by @xmath339 . then the density
density correlation function is similar to eq.([rhorho ] ) if one adds the subscript @xmath51 to @xmath61 in the first line , and @xmath71 is replaced by @xmath340 obtained from eq.([xm ] ) by the following substitution @xmath341 thus @xmath342 where @xmath343 is the analogous of @xmath71 without spin with @xmath25 and @xmath4 indexed by @xmath339 .
the imaginary part of the fourier transform of @xmath340 at low frequencies , @xmath344 is defined as in eq.([imaginary ] ) .
it can not be factored unless the spin and charge velocities are equal . for studying the dominant tendency , it is sufficient to consider its local values .
up to a slowly varying function of @xmath121 , we have @xmath345 where @xmath346 is exactly that computed without spin , with the additional index @xmath339 , in particular there are two thermal lengths associated with the different charge and spin velocity @xmath347 , two temperatures @xmath348 , and two times ( in units of @xmath64 ) @xmath349 for a spin ( @xmath350 ) or charge ( @xmath330 ) excitation to go from @xmath121 to the closest contact .
the calculation can be carried out for any parameters for charge and spin , on the leads or inside the wire , but let us specify eq.([factorspin ] ) for non interacting leads , and for @xmath310 . in the high temperature limit , using eq.([vrenorm ] ) , @xmath351 in particular , at a distance much greater than @xmath352 from the contacts , @xmath353 is the same as that in an infinite wire @xmath354 for @xmath355 , the @xmath356 component of the density
density correlation function is enhanced compared to the noninteracting case . in the low temperature limit , using eq.([integration ] ) @xmath357 note that the superconducting correlation functions can be obtained in a similar fashion , one has to distinguish the triplet from the singlet superconducting tendency , and we refer to for more details . in an infinite wire , the @xmath16 cdw , corresponding to wigner crystal , dominates the @xmath1 cdw for very repulsive interactions , more precisely for @xmath358 and @xmath310 as one can inspect using eq.([chispinbulk ] ) .
we show here that the leads can induce also a proximity effect in the opposite sense , by suppressing the wigner crystal at low temperature .
this tendency is even more important for long range coulomb interactions to which case one could extend qualitatively the results obtained here .
our study without spin has shown that the enhancement of a @xmath1 cdw persists with leads .
we have also observed that the @xmath137 cdw are suppressed at low temperature due to the @xmath209 dependent term @xmath359 in @xmath75 ; but this effect intervenes only when the backscattering potential is studied on resonance , because the @xmath1 cdw dominates @xmath137 for @xmath360 . with spin , one has to compare the @xmath1 to the @xmath16 cdw , thus the correlation functions @xmath344 corresponding respectively to @xmath361 and @xmath362 .
the competition is much less trivial than without leads and is @xcite discussed in detail in appendix [ appwigner ] .
here we summarize the results , illustrated by figs.[fg : klesst ] and [ fg : kinterm ] .
we will focus on the @xmath358 case , since for @xmath363 the @xmath1 cdw is dominant at any temperature .
then contrary to the infinite wire where the @xmath16 dominates , we show here that there is a crossover temperature @xmath17 from the @xmath364 cdw to the @xmath1 cdw as @xmath209 is lowered below @xmath17 .
@xmath17 decreases monotonically when @xmath121 varies from one contact to the center of the wire where it reaches its minimum value , interestingly given by the following power of @xmath132 : @xmath365 one can check that @xmath366 because @xmath367 .
indeed @xmath17 saturates at @xmath368 for all @xmath121 such that @xmath369 , i.e. @xmath150 . thus for any @xmath370 , the wigner crystal is suppressed all over the wire and the dominant tendency is towards the @xmath1 cdw . in order to discuss the case @xmath371 we shall investigate in detail the behavior close to the contacts .
there is a second particular value other than @xmath372 that arises : all the correlation functions at @xmath373 are analogous to those in the center of the wire with @xmath374 replaced by @xmath375 ( the analogue of eq.([ka ] ) with the supplementary index @xmath293 ) .
when @xmath310 , @xmath376 , the limiting value @xmath377 corresponds to @xmath378 . in particular ,
as long as @xmath379 one has @xmath380 , thus the @xmath1 cdw dominates the @xmath16 cdw at the contacts at any temperature .
this yields different expressions for @xmath17 depending on @xmath381 : * for @xmath382 , it is given by eq.([tckmore ] ) , and diverges at @xmath373 ; reassuringly , we find that @xmath17 is continuous at the point @xmath383 [ eq.([tx0 ] ) ] where it reaches @xmath132 from below or above ( fig.[fg : kinterm ] ) . * for @xmath384 , @xmath17 is given by eq.([tcl ] ) .
@xmath17 is below @xmath132 for any @xmath121 , and reaches its maximum value at @xmath385 ; @xmath386 is similar to eq.([tc0 ] ) where @xmath381 is replaced by @xmath387 .
( fig.[fg : klesst ] ) if we consider now any @xmath358 and a given temperature in the range @xmath388 $ ] ( @xmath386 is infinite for @xmath389 ) , we see that the dominant tendency switches from the @xmath16 to the @xmath390 cdw when one goes from the center to the contacts .
the point @xmath121 at which this transition occurs is a temperature dependent function @xmath391 that is the inverse of @xmath17 . for @xmath209
higher than @xmath386 ( lower than @xmath368 ) , the @xmath16 ( @xmath1 ) cdw dominates all over the wire . to conclude , the external leads _ suppress the importance of the @xmath16 cdw _ at low enough temperature or close to the contacts , thus preventing the tendency towards the formation of a wigner crystal .
the role of impurities when spin is taken into account can be treated in a similar fashion as without spin , by doubling the indices as above , leading to different powerlaws on temperature .
nevertheless , in view of our previous discussion , the main additional complication arises from the competition between the @xmath1 and @xmath16 backscattering : the leads intervene more strongly if @xmath358 .
we will first derive briefly the renormalization equations at finite temperature , then give results for the conductance .
the coupling to impurities is described in terms of @xmath392 , where @xmath51 is given by eq.([rhospin ] ) . for any potential ,
the partition function can be expanded in terms of @xmath168 in a similar way as for the spinless case [ eq.([zprem ] ) ] , but with @xmath393 replaced by a couple of integers @xmath394 .
then @xmath167 describes integer charges restricted to two neutral cylinders , corresponding to the charge ( @xmath395 ) and the spin ( @xmath396 ) degree of freedom .
the radius of each cylinder is given by @xmath397 , and its height determined by the spatial extension of the potential .
only charges on the same cylinder @xmath339 interact via @xmath398 , eq.([utau ] ) with the index @xmath339 on any function or parameter . following similar steps as in the spinless case
, we can infer the leading renormalization equation analogous to ( [ dvmsurdlx ] ) , where @xmath61 is replaced according to the recipe ( [ receipe ] ) , @xmath399v ' .\end{aligned}\ ] ] for simplicity we have omitted the arguments on the right hand side . in particular , through @xmath338 , @xmath166 now acquires an additional spatial dependence .
the bare value of @xmath400 is zero if these terms are not present in the initial density representation , eq.([rhospin ] ) . in order to illustrate how they are generated ,
we specialize to the case of one barrier .
then the cylinders shrink to circles , and the partition function reads @xmath401.\end{aligned}\ ] ] the sum runs over all @xmath402-tuples of integers @xmath403 and @xmath404 with total vanishing sum separately , and the bare transform of @xmath166 @xmath405 acquires a dependence on @xmath406 during renormalization .
we can show that the interactions are not renormalized as for the spinless case .
the next leading term to eq.([renormalisationspin ] ) is obtained by the contraction of one pair of charges on each circle , and is equal to@xcite @xmath407 up to non universal prefactors , the sum running over @xmath408 and @xmath409 . for a potential with extension @xmath195
much less than @xmath19 and @xmath21 , we can show that the partition function is analogous to eq.([zonebarrier ] ) when a cutoff larger than @xmath195 is reached .
again , at the location of @xmath166 the interaction parameters might be renormalized before reaching this stage .
if the renormalization of the local effective interactions , as well as next leading terms can be ignored , eq.([renormalisationspin ] ) can be integrated easily .
the maximum cutoff is half the radius of one cylinder , i.e. @xmath410 , where we get the renormalized parameters@xcite @xmath411 @xmath353 has been studied previously , and is given by eq.([chihighspin ] ) ( respectively ( [ chilowspin ] ) ) in the high ( low ) temperature limit . for @xmath310 , @xmath329 , and for any @xmath121 ,
@xmath168 is enhanced when the temperature is lowered , and is less enhanced when one gets closer to the contacts because the leads moderate the repulsive interactions . if the potential is weak enough or the wire is short enough , we can perform the renormalization at temperatures below @xmath132 .
the renormalized @xmath412 obtained at a given @xmath148 is the same for all @xmath148 , but this is not the case for all the other couples @xmath413 , for which @xmath414 goes down to zero at zero temperature .
for instance , we refer to our previous discussion of the competition between the @xmath1 and @xmath16 backscattering contribution .
now we give the perturbative expression for the conductance . for a general weak potential @xmath159
, the conductance can be expressed as in eq.([gimp ] ) .
@xmath222 is similar to eq.([rsommerm ] ) with @xmath197 standing for a couple of integers @xmath415 with the same parity .
this substitution applies also to eq.([eqrmnew ] ) , giving @xmath416}\chi_{m_\rho , m_\sigma}(x , y).\end{aligned}\ ] ] the nonlocal function @xmath417 is now more difficult to compute due to the different velocities of spin and charge .
but if we assume @xmath418 , it can be inferred from that without spin ( see appendix [ correlation ] ) by replacing @xmath419 when the potential has an extension less than @xmath19 and @xmath420 ( @xmath317 ) , only the local value is required as far as _ the dominant term _
is concerned.@xcite then the leading term is similar to eq.([rlocal ] ) with @xmath197 replaced by the pair @xmath356 , thus replacing @xmath228 by eq.([vmspin ] ) , and using eqs .
( [ chihighspin],[chilowspin ] ) . for @xmath358
, we refer to figures [ fg : klesst ] and [ fg : kinterm ] . above ( below ) the crossover temperature @xmath17 , where @xmath121 is the location of the barrier , the @xmath16 ( respectively @xmath1 ) backscattering
gives the dominant contribution , and can be inferred from eqs .
( [ chihighspin],[chilowspin ] ) by taking @xmath361 ( respectively @xmath421 ) depending on whether @xmath148 or @xmath211 . for a random potential verifying eq.([gauss ] ) , the leading term of @xmath222 is given explicitly in figure [ f : spindesordre ] .
again the case @xmath358 is more complicated : we have to distinguish three temperature regimes , separated by @xmath132 than @xmath368 , eq.([tc0 ] ) .
we can also compute the variance of @xmath222 , using eqs .
( [ eqrmnewspin],[gauss ] ) .
similar steps to that without spin can be carried on . while the powerlaw behavior of @xmath6 is different from that without spin , the relative fluctuations are the same , i.e. we have still eq.([ratiolow ] ) in the low temperature limit , and eq.([ratiohight ] ) ( with a different function @xmath280 that does not affect the order of magnitude ) in the high temperature limit . let us give the behavior of @xmath6
. in the low temperature limit @xmath148 , the inequality ( [ encadrereflection ] ) holds again , showing that @xmath6 is of the same order as @xmath260 , eq.([ratiolow ] ) , given in fig.[f : spindesordre ] .
in particular , in the case @xmath422 , @xmath423 for @xmath358 , this powerlaw holds at lower temperature @xmath370 , eq.([tc0 ] ) , but @xmath424 for @xmath425 . in the high temperature limit @xmath211 , restricting ourselves to @xmath329 , we can retain only the contribution of the bulk .
then we find @xmath426 up to a bounded function of both @xmath427 for @xmath317 .
again , we have ignored the terms due to first and second derivatives of @xmath78 , and kept only the pair @xmath356 that yields the dominant term .
this pair is equal to @xmath428 ( respectively to @xmath362 ) for @xmath363 ( respectively @xmath358 ) as is the case for the dominant contribution to @xmath429 .
remarkably , the ratio ( [ ratiohight ] ) is _ the same _ as that without spin up to the function @xmath280 , and is @xmath253 in _ any parameter range_.
different aspects of transport in an interacting wire with measuring leads have been treated in this paper . in this section , we will compare the effect of backscattering on the conductance to that obtained without leads , and then discuss the results that can be generalized to smooth variations of the interactions at the contacts . in an infinite wire with repulsive interactions , as is thought to be appropriate for a quantum wire , a unique local barrier is sufficient to suppress transport completely.@xcite the finite length is intuitively introduced as a cutoff , thus preventing the conductance to vanish for weak enough backscattering potential or short enough wire.@xcite with noninteracting leads and spinless electrons , we recover the results of refs .
over all the temperature range only for a barrier whose separation from the contacts is of the order of the wire length @xmath19 , and under the additional condition @xmath363 for electrons with spin .
if we consider extended disorder , the agreement holds for @xmath430 without spin , and for @xmath431 with spin ( see fig [ f : spindesordre ] ) . in this parameter range ,
the localization length we infer from the breakdown of our perturbative evaluation of @xmath222 coincides with that found in ref . in the weak disorder limit .
we now discuss the extension of our results to smooth variations . as we mentioned before
, the discontinuous parameter model can be relevant for edges states in the fqhe@xcite , but is not well justified microscopically for quantum wires .
thus it is important to know which results can be trusted for a more realistic interaction profile .
the detailed behavior of the transmission dynamics of an incident electron was given previously for abrupt jumps.@xcite by inspecting the boson green function , we can maintain the qualitative conclusion ; in particular , the transmitted charge and spin stay separated in the external leads .
the correlation functions are directly related to this process.@xcite all the properties of @xmath78 found in subsection [ properties ] still hold . to be specific ,
assume that the leads are non interacting , @xmath46 and that @xmath4 decreases slowly enough from @xmath141 to its bulk value @xmath4 inside the wire .
again , for @xmath91 , [ cf .
( [ chibulk ] ) and ( [ majoration ] ) ] the tendency towards the cdw holds inside the wire , but is reduced when one goes to the leads , until it reaches its noninteracting value in the leads at a distance @xmath432 where @xmath433 is the maximum energy at hand , recalling the proximity effect .
analogous conclusions hold for the superconducting fluctuations when interactions are attractive , the coherence length being now @xmath434 . in the presence of a non random backscattering potential ,
the renormalization equations ( [ renormalisationspin ] ) are general .
the role of backscattering is determined by the density density correlations .
consider for instance a barrier at a point @xmath121 .
when @xmath121 is at a distance of order @xmath19 of the contacts , the role of backscattering is the same as that in ref .. for @xmath91 , @xmath222 ( see eqs .
( [ majoration ] ) and ( [ rlocal ] ) ) decreases when @xmath121 gets closer to the contacts .
the noninteracting leads ( where a barrier is marginal ) tend to reduce the importance of the repulsive interactions when one gets closer to them .
using the boson green function properties on the external leads ( restricting ourselves to the case @xmath310 with spin ) , the temperature dependence of the correlation function corresponding to the @xmath137 cdw at @xmath148 has a term @xmath435 .
thus for @xmath46 , this saturates for @xmath213 but goes to zero at zero temperature for @xmath436 .
in particular , the usual scaling argument inferring the low temperature behavior @xmath148 from that at @xmath211 is justified only if the @xmath1 contribution intervenes at all temperatures , which is not the case in our model for spinful electrons with @xmath3 , or for any parameter range when the resonance in the presence of backscattering potential is achieved . for a random potential ,
all the results can be generalized for repulsive interactions .
but the inequality ( [ encadrereflection ] ) is more general than that , and holds even if @xmath4 varies in any way inside the wire .
when spin is taken into account , and for @xmath310 , the tendency towards a wigner crystal is expected to be suppressed at low temperature , more precisely the dominant tendency is that towards the @xmath1 cdw at any @xmath370 [ eq.([tc0 ] ) ] .
this is because the dependence on temperature of the @xmath16 cdw is controlled by @xmath46 , thus is in @xmath437 , vanishing in the zero temperature limit . in the presence of coulomb interactions ,
the ballistic conductance is still independent on the interactions.@xcite one can make an analogy with the case @xmath358 to expect the transmitted charge and spin to stay separated in the leads , and the tendency towards wigner crystal to be suppressed by the leads at low enough temperature .
we now comment on the experimental situation . in the high temperature limit ,
the ballistic value of the conductance of one - channel quantum wire @xmath438 was observed by tarucha _
et al._,@xcite as well as in the remarkable clean wires of yacoby _
et al_.@xcite but the low temperature behavior is more tricky .
on the one hand , tarucha _ et al . _
fitted the observed @xmath222 with the intuitive law @xmath439^{(k_\rho -1)/2}$ ] of ref . , and infer the parameter @xmath440 .
we however find a more complex dependence of @xmath222 on the temperature and the location of the barriers .
for instance , if the backscattering is merely due to the contacts , the exponent @xmath441 would correspond to the local parameter @xmath442 , thus to @xmath443 .
of course we do nt expect the same combination of parameters to appear in the case of smooth variations of @xmath4 , but this serves as an illustration ; as we discussed in section [ secgeneral ] , the effective parameter that controls the powerlaw increases when one gets closer to the wire .
probably the potential configuration depends on each wire , and more precise fit and evaluation of @xmath444 are required .
even more caution is needed if the potential is random . while the conductance is self averaging at high temperature ( eq.([ratiohight ] ) ) , its fluctuations become more important at low temperature , of the order of the reduction @xmath222 , eq.([ratiolow ] ) ; then one has not to measure simply the absolute value of @xmath222 , but also its fluctuations . on the other hand , yacoby _
et al._@xcite fabricated high quality quantum wires using a sophisticated technique .
the most striking result is that the conductance shows perfect plateaus as function of the gate voltage .
their value is @xmath445 per spin mode and channel where @xmath446 .
@xmath67 depends on the temperature and the wire length , but is reproducible for different wires made in wells of the same width .
this indicates that the reduction in the conductance is not due to potential fluctuations , but rather to the backscattering of electrons from the two dimensional gas when entering the wire . in the high temperature limit , the one - channel wire conductance is @xmath235 , ( @xmath447 ) in accordance with theoretical results .
but @xmath67 decreases when one lowers temperature , or increases the wire length , in a way similar qualitatively to the law @xmath448 or @xmath449 for an interacting wire with parameter @xmath4 .
the main objection of yacoby _ et al .
_ against this tll interpretation is that temperature is scaled by the fermi energy which varies with electron density , in contradiction with the observed plateaus which are density independent .
further , the parameter @xmath4 itself is expected to depend on the density .
the importance of these two effects is however not so obvious to assess .
first , energy scales different from the fermi energy might play the role of the cutoff .
second the way @xmath4 depends on the density ( i.e. the gate voltage ) is difficult to estimate : a fully microscopic calculation of @xmath4 would be necessary , including the capacitive effects of the gates .
we have however also to remark that our model is not the best suited to describe the complicated geometry of the experiments .
the transition between the two dimensional gas and the wire is not easy to describe , and is different from the usual adiabatic transition .
thus we think model closer to the actual experiment needs to be developed.@xcite
in this paper we have studied the properties of an interacting wire connected to measuring leads .
we have obtained a number of new results , mainly concerning the suppression of the wigner crystal and its effect on the conductance with impurities , and the conductance fluctuations .
we have also completed our study of the transmission process by taking spin into account .
we have shown here that the clean wire acts as a separator of the charge and spin of an incident electron , which do nt recombine in the noninteracting leads .
we can hope that experimental progress allows this feature to be observable by appropriate spin and charge detectors .
we have also given a systematic computation of the nonlocal correlation functions and their fourier transform at temperature @xmath450 .
this has allowed to investigate both the dominant tendency in and outside the wire and to study the role of impurities . as first mentioned in ref . , we have explained in more detail how the proximity effect shows up . for repulsive ( attractive ) interactions , the charge density ( pairing ) fluctuations
are enhanced in the bulk of the wire , and decrease , while staying enhanced compared to the noninteracting case , when one goes beyond the contacts .
but the novel effect is that the leads influence the competition between different tendencies in the wire at low enough temperature .
this was shown to be the case for very repulsive short range interactions @xmath3 .
normally , a tendency towards a wigner crystal shows up , corresponding to the @xmath16 cdw .
when connected to leads , such tendency gets suppressed below a space - dependent crossover temperature , and is replaced by the @xmath1 cdw . in particular , this allows for simultaneous @xmath1 and @xmath16 cdw in the wire in a certain temperature range .
@xmath451 is controlled by interactions , thus its measurement could provide a way to measure @xmath4 . the wigner crystal shows up more strongly for coulomb interactions ; we have not studied this case , but we expect the same qualitative conclusions to hold .
we have also studied the role of a backscattering potential .
we have developed a formal framework to study its role for any non translationally invariant interactions , and explicitly at finite temperature .
we have derived the renormalization equations for a non random potential @xmath166 with any extension , but we expect interactions to be renormalized by a nonlocal @xmath166 . if @xmath159 is weak enough for this effect to be neglected , the conductance @xmath24 was expressed in detail in the discontinuous parameter case .
@xmath24 is controlled by the interactions , but is influenced by the external leads , especially for very repulsive interactions because the leads suppress the tendency towards the wigner crystal below @xmath17 [ cf figs .
[ fg : klesst],[fg : kinterm ] ] .
we have computed the conductance fluctuations , whose behavior is nonuniversal .
but remarkably , the ratio @xmath8 stays of the same order , whether the spin is taken into account or not , or whether @xmath7 or not .
it is @xmath452 for @xmath148 , and @xmath453 for @xmath211 .
we can generalize the main results to a smooth interaction profile near the contacts .
one of us ( i. s. ) would like to thank b. douot , d. c. glattli , t. martin , d. l. maslov and a. yacoby for interesting discussions .
we derive the analogue of a dyson equation for the nonlocal conductivity .
we begin by spinless electrons , taking into account spin will the be straightforward .
we consider a general hamiltonian @xmath454 where @xmath455 is given by eq.([h ] ) , and @xmath456 depends on @xmath29 , but not on its conjugate canonical momentum @xmath457 .
let us first give the general equation of motion for @xmath29 .
using @xmath458=i\delta(x - y)$ ] , any functional @xmath459 of @xmath29 verifies @xmath460 = -i\frac{\partial { \mathcal{f}}}{\partial \phi}.\ ] ] this allows us to write the system @xmath461 where @xmath462 note that @xmath463 has the dimension of a force that intervenes in the time derivative of @xmath464 . eliminating @xmath465 in eq.([piphi ] ) , the equation of motion for @xmath466 reads @xmath467 where we denote @xmath468.\ ] ] for an arbitrary nonlocal interaction potential @xmath49 , @xmath469 contains also an integral operator,@xcite and all the remaining analysis can be generalized to that case .
note that without @xmath470 , we have @xmath471 , and thus recover the equation of motion @xmath472 . in the following ,
the corresponding correlation functions carry a subscript @xmath473 , in order to distinguish them from those computed with @xmath474 .
next we derive a dyson equation for the nonlocal conductivity @xmath475 , eq.([cali ] ) . for any @xmath476 depending on @xmath466 in a local or nonlocal way
, we can show , using eq.([relbase ] ) and eq.([moti ] ) , that @xmath477 \right\rangle \right\ } = i\theta(t)\left\langle \left [ { \cal i}(x , t),a\right ] \right\rangle + \delta ( t)\frac{\partial a}{\partial \phi ( x)}\label{eqor}\ ] ] where @xmath478 is the heaviside function .
the fourier transform with respect to time of eq.([eqor ] ) yields @xmath479 where we adopt the notation @xcite @xmath480 \right\rangle_{h_{tot}}.\ ] ] we point out that the correlation functions are defined here with the total hamiltonian @xmath481 , eq.([htot ] ) .
similarly , we can show that @xmath482 applying these two identities successively to the green function @xmath483 we get @xmath484 where @xmath53 denotes @xmath485 .\ ] ] in order to invert the relation above , and thus to obtain @xmath486 where @xmath487 are arbitrary , we multiply its two members by @xmath488 , integrate over @xmath489 , and use @xmath490 then we multiply both sides by @xmath491 and integrate over @xmath121 .
we get @xmath492 we can also express this equation in terms of the dimensionless nonlocal conductivity ( measured in units of @xmath235 ) , eq.([sigmag ] ) : @xmath493 the conductance measured in units of @xmath235 is given by eq.([gsigma ] ) , @xmath494 independent of its arguments , which ensures current conservation .
it is also real because it is a correlation function of two hermitian fields .
indeed , for two hermitian operators @xmath495 , we have @xmath496 \right\rangle = \left\langle \left\langle a;b\right\rangle \right\rangle _ 0 \;\;\;\mbox{real}.\ ] ] besides , @xmath497 \right\rangle dt \;\;\;\mbox{is purely imaginary}\ ] ] thus @xmath498 is real while its derivative at zero frequency is purely imaginary because @xmath475 is hermitian .
if one supposes @xmath499 finite , one has @xmath500
= 0.\ ] ] then the zero frequency limit of eq.([cond ] ) yields the exact conductance measured in units of @xmath235 @xmath501 .\ ] ] where @xmath502 is the zero frequency limit of eq.([eq : fnonlocal ] ) and @xmath503 is the dimensionless conductance in the absence of @xmath470 . if one takes spin into account , @xmath504 can depend on the spin field @xmath505 and its conjugate canonical momentum @xmath506 , both commuting with the field @xmath507 .
this is the case for instance for backscattering potential , where the contribution of the @xmath356 harmonics of the density ( [ rhospin ] ) to @xmath508 reads , after integration by parts : @xmath509 .
we will show how to compute the imaginary part of the fourier transform of the nonlocal correlation function ( [ xm ] ) at frequency @xmath510 , i.e. @xmath85 [ eq.([imaginary ] ) ] .
first we clarify the way analytic continuations from matsubara to real frequencies are done when we regularize by imposing the imaginary time to obey @xmath511 .
consider any general operators @xmath476 and @xmath512 .
using the equivalent of the notation ( [ notationomega ] ) for the matsubara frequency @xmath513 , we have @xmath514 where we have deformed the integration contour in the complex plane , and used @xmath515 let us now specialize to the case @xmath516 where @xmath197 is normally an integer but can be any real , and consider @xmath71 , eq.([xm ] ) @xmath517 then we have @xmath518 this is because the transformation : @xmath519 leaves @xmath455 invariant , but @xmath520 .
besides , @xmath71 is invariant when exchanging @xmath121 and @xmath489 because of time reversal symmetry . equation ( [ ab ] ) becomes , after the analytic continuation @xmath521 , @xmath522 the derivative with respect to @xmath433 yields , after changing @xmath523 in the second term above , @xmath524 where @xmath525 .
according to eq.([eq : dro ] ) , and after the change of variables @xmath526 , @xmath527 dz , \label{dchim}\ ] ] where @xmath528 , and @xmath61 is given by eq.([utau ] ) .
owing to the analytic properties of @xmath61 , we can move the integration contour in eq.([dchim ] ) as explained now .
let us write the integral in eq.([dchim ] ) as @xmath529 where the spatial arguments are implicit .
@xmath61 is analytic on the complex band @xmath530 , as well as @xmath531 .
let @xmath532 .
we then have @xmath533 .
we can show that for @xmath534 , @xmath535 then we can translate the contour of integration by @xmath536 thus making it along @xmath537 . on the other hand
, we can show that @xmath538 is an even function of @xmath539 , thus @xmath540 finally , the integral ( [ dchim ] ) becomes @xmath541 this form is much more convenient than ( [ dchim ] ) for three reasons : the integrand is real , the singularities are avoided , and approximations are easier to make .
thus the dimensionless function @xmath78 , eq.([imaginary ] ) , corresponding to @xmath213 can be expressed as @xmath542 this gives eq.([chic ] ) : @xmath543 where @xmath544 . \label{gint}\ ] ] all the @xmath75 can be obtained by multiplying @xmath67 by @xmath77 .
let us give the function @xmath67 in the case of discontinuous interaction parameters@xcite @xmath545\end{aligned}\ ] ] with @xmath546 , \ ] ] and @xmath547 @xmath110 is given by eq.([gamma ] ) .
it is worth noting that one can deduce @xmath67 from @xmath217 as explained in ref . , and that @xmath217 is related to the transmission process.@xcite let us first consider the limit @xmath95 .
we can show in detail that a good approximation of the integral @xmath83 , eq.([gint ] ) , can be obtained if we approximate @xmath67 by @xmath548\right\}. \label{gzfaible}\ ] ] there is no coherence between the two contacts separated by @xmath278 , so that only the role of one reflection matters .
the integral can be expressed as @xmath549 where @xmath550 is the hypergeometric function generalized to two variables,@xcite and @xmath250 given by eq.([ka ] ) . in the limit of an infinite wire with constant parameters , we can set @xmath551 , i.e. @xmath552 . then @xmath83 is a function of @xmath86 , or of @xmath553 [ eq.([lesv ] ) ] and becomes @xmath554 where @xmath512 is the beta function and @xmath555 is the hypergeometric function of one variable . at distances
far from the contacts compared to @xmath21 , thus for @xmath556 , we have @xmath557 in order to express @xmath78 , eq.([apchic ] ) , we need also to exponentiate eq.([gzfaible ] ) for @xmath558 , and @xmath559 .
thus one gets @xmath85 for any @xmath84 .
[ [ local - value ] ] local value + + + + + + + + + + + for @xmath560 , we can check that @xmath561 .
nevertheless , for @xmath562 to hold , we must have @xmath563 .
thus strictly speaking the property ( [ chiloc ] ) ( now @xmath564 since we are in the high temperature limit ) holds only for @xmath565 .
but this is related to the abrupt jump of parameters at the contacts , thus we do nt expect it to hold for smooth variations , where eq.([chiloc ] ) can be generalized .
besides bosonization describes large separations compared to @xmath566 , thus the variations of @xmath4 as well as the behavior for @xmath567 are not relevant .
recall that @xmath568 is the time it takes for a `` quasiparticle '' to go from @xmath121 to the closest contact , measured in units of the cutoff @xmath64 [ eq.([tx0 ] ) ] : @xmath569 let us write the local value of @xmath78 : @xmath570 ^{k_a - k}\\\nonumber&&{_2f_1}\left ( \frac 12,k_a - k , k_a+\frac 12,\tanh ^2(\pi \underline{t}t_x)\right).\end{aligned}\ ] ] the hypergeometric function @xmath555 is unity at the contact ( @xmath571 ) and stays bounded and nonvanishing all over the wire . at @xmath572 , @xmath573 at @xmath574 ,
an analogous expression holds with @xmath575 . besides
, we can show that:@xcite @xmath576 for @xmath577 , while the inverse holds for @xmath578 .
we note finally that the simplified expression ( [ vrenorm ] ) was obtained by ignoring the hypergeometric function in eq.([upsilonxxtext ] ) , as well as constant factors , and taking @xmath579 .
but we can check that @xmath134 , eq.([chia ] ) , can be recovered from eq.([vrenorm ] ) by letting @xmath130 .
[ [ expression - of - chiax ] ] expression of @xmath580 + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + another interesting expression is that of @xmath580 for @xmath581 , thus @xmath582 in eq.([cf1 ] ) : @xmath583 the computation of @xmath83 is now much easier . since in eq.([gint ] ) @xmath584 , @xmath585 .
then one can approximate , in eq.([apgz ] ) , @xmath586 , \label{g}\ ] ] thus eq.([gint ] ) reduces simply to @xmath587 independent of @xmath84 .
@xmath83 is the same as @xmath120 evaluated in an infinite wire with parameter @xmath45 .
but @xmath78 [ eq.([apchic ] ) ] has also parts depending on @xmath121 and @xmath489 due to @xmath67 . by letting @xmath558 in eq.([apgz ] ) we get @xmath588 up to a bounded and slowly varying function of @xmath568 .
inserting this expression for each of @xmath84 as well as eq.([clow ] ) in eq.([apchic ] ) , we obtain @xmath85 , factorizing as in eq.([decouple ] ) because @xmath261 . note
that taking @xmath589 in eq.([glow ] ) or ignoring the @xmath141 in front of @xmath590 then taking @xmath130 gives the same result for @xmath134 .
the behavior for @xmath591 is not very relevant since it depends closely on the unrealistic abrupt profile of @xmath92 , and also because the tll model is valid at distances much larger than @xmath566 , in particular @xmath126 .
here we discuss in detail the competition between the @xmath1 cdw and the @xmath16 cdw , separating the high and low temperature regimes . for simplicity , we restrict ourselves to the case @xmath310 , and we drop the index @xmath51 from all the parameters or length scales .
note that the effective local parameter at the contacts , eq.([ka ] ) , for the spin degrees of freedom is also equal to one : the correlation function @xmath592 is now equal to that in a noninteracting wire for any @xmath121 and @xmath489 .
we consider here @xmath593 .
recall that the correlation functions at distances far from the contacts compared to @xmath594 are identical to those in an infinite tll , eq.([chispinbulk ] ) . at the contacts ,
the equivalent tll has a parameter @xmath595 , eq.([ka ] ) . for any @xmath121 ,
let us use eqs.([factorspin],[integration ] ) to obtain the ratio between the @xmath1 and @xmath16 cdw correlation functions , @xmath596 where @xmath110 is given by eq.([gammanu ] ) , and @xmath597 is the time it takes for a charge excitation emanating at @xmath121 to reach the closest contact .
the @xmath16 cdw dominates the @xmath1 one wherever @xmath598 .
thus : * for @xmath599 , the @xmath16 cdw is dominant for any @xmath3 as in an infinite tll , because @xmath600 for @xmath601 . * for @xmath602 ,
i.e. at the contacts , @xmath603 : the @xmath16 dominate for @xmath604 , i.e. for @xmath605 .
* for any @xmath606 , we can replace @xmath607 if @xmath608 , and the ratio ( [ ratio_correlation_high ] ) becomes @xmath609this expression holds for @xmath610 , but the computation can be carried out for any @xmath121 ( see appendix [ correlation ] ) .
+ since @xmath611 , @xmath612 , @xmath613 increases when one gets closer to the contacts , being @xmath253 for @xmath579 , and of order @xmath141 at the contacts .
this means that the importance of the @xmath1 compared to the @xmath16 is enhanced as one gets closer to the leads .
we have to discuss whether @xmath605 or @xmath614 . in the former case , @xmath598 ( cf eq.([ratio_small_tx ] ) ) . in the latter case @xmath615
reaches unity for @xmath616 thus , for @xmath617 , the @xmath1 cdw dominates ( see fig.[fg : kinterm ] ) . to summarize the high temperature regime , the @xmath16 ( @xmath1 )
cdw is dominant all over the wire for @xmath605 ( @xmath7 ) . in the intermediate parameter region ,
i.e. for @xmath618 , the @xmath16 cdw dominates only for points far enough from the contacts , i.e. for @xmath619 , eq.([xch ] ) .
consider now the low temperature limit , @xmath620 .
let us write again eq.([chilowspin ] ) @xmath621 this holds for @xmath622 , and the value at the contact can be recovered by simple substitution @xmath623 in eq.([lowchispin ] ) as we can check from explicit computation for any @xmath121.@xcite the essential difference between @xmath624 and @xmath625 , due to the leads , is through the temperature dependent term @xmath626 in @xmath625 that vanishes in the zero - temperature limit , while @xmath624 is temperature - independent .
then the @xmath1 cdw dominates the @xmath16 cdw in the zero - temperature limit all over the wire for any parameter @xmath4 . for @xmath7 ,
the @xmath1 dominates everywhere at any temperature . for @xmath3 ,
the @xmath16 cdw dominates in the high temperature limit .
thus there is a crossover temperature we determine now .
let us write the ratio of the corresponding correlation functions at finite temperature , using eq.([lowchispin ] ) , @xmath627 [ transition ] again , we have to distinguish the cases @xmath605 and @xmath614 : * for @xmath605 , we can easily get the crossover temperature by letting @xmath628 , @xmath629 we can check that for any @xmath630 $ ] , @xmath631 , and @xmath451 is a decreasing function of @xmath568 ( fig.[fg : klesst ] ) .
* for @xmath618 , the same expression for the crossover temperature holds , eq.([tcl ] ) .
but it is inside the actual range of temperatures , i.e. @xmath632 only for points less than @xmath383 where @xmath383 verifies
@xmath633 , i.e. given by @xmath634 * for @xmath605 , the @xmath16 cdw dominates for @xmath635 , eq.([tcl ] ) , while the @xmath1 cdw dominates below ( fig.[fg : klesst ] ) .
in particular , for a given temperature in the range @xmath388 $ ] , there is a spatial crossover between the two tendencies occurring at @xmath636 , one can obtain by inverting eq.([tcl ] ) . * for @xmath618
, we can invert the relation ( [ xch ] ) to get the crossover temperature higher than @xmath132 for @xmath637 , and we can ( which is reassuring ) check that the point where it reaches @xmath132 is again given by @xmath383 , eq.([tx0 ] ) , showing that we have a continuous function @xmath17 given by @xmath638 note that the @xmath1 cdw dominates for any temperature at the contacts : @xmath639 , thus @xmath386 in the second line diverges ( fig.[fg : kinterm ] ) .
m. fabrizio and a. gogolin , phys .
b * 51 * , 17827 ( 1995 ) .
in the same spirit , but for the different problem of the fractional quantum hall effect , see m. r. geller _
et al . _ ,
77 * , 4612 ( 1997 ) .
the result @xmath438 was recovered without leads by a. kawabata , j. phys .
* 65 * , 30 ( 1996 ) , y. oreg and a. m. finkelstein , phys .
b * 54 * , r14265 ( 1996 ) and r. egger and h. grabert , phys . rev
77 * , 538 ( 1996 ) . but
this requires the current to be not renormalized by the interactions@xcite .
after the original version of this paper was submitted , appeared a paper by scot r. renn and daniel p. arovas , phys .
* 78 * , 4091 ( 1997 ) where the fluctuations of the conductance are computed in the infinite wire for spinless electrons , and coincide with our results only in the high temperature limit and for spinless electrons . | we investigate a one dimensional wire of interacting electrons connected to one dimensional noninteracting leads in the absence and in the presence of a backscattering potential .
the ballistic wire separates the charge and spin parts of an incident electron even in the noninteracting leads .
the fourier transform of nonlocal correlation functions are computed for @xmath0 . in particular , this allows us to study the proximity effect , related to the andreev reflection .
a new type of proximity effect emerges when the wire has normally a tendency towards wigner crystal formation .
the latter is suppressed by the leads below a space dependent crossover temperature ; it gets dominated everywhere by the @xmath1 cdw at @xmath2 for short range interactions with parameter @xmath3 .
the lowest order renormalization equations of a weak backscattering potential are derived explicitly at finite temperature . a perturbative expression for the conductance in the presence of a potential with arbitrary spatial extension is given .
it depends on the interactions , but is also affected by the noninteracting leads , especially for very repulsive interactions , @xmath3 .
this leads to various regimes , depending on temperature and on @xmath4 .
for randomly distributed weak impurities , we compute the conductance fluctuations , equal to that of @xmath5 . while the behavior of @xmath6 depends on the interaction parameters , and
is different for electrons with or without spin , and for @xmath3 or @xmath7 , the ratio @xmath8 stays always of the same order : it is equal to @xmath9 in the high temperature limit , then saturates at @xmath10 in the low temperature limit , indicating that the relative fluctuations of @xmath11 increase as one lowers the temperature .
# 1 |
.the five most energetic pulsars : from the left to the right columns , pulsar name ( psr ) , rotation period ( @xmath7 ) in units of second , spin down age @xmath8 ) in units of @xmath9 , the magnetic field strength at the light cylinder ( @xmath10 ) in units of @xmath11 g , distance to the source @xmath12 in units of kpc , x - ray efficiency and gamma - ray efficiency ( in @xmath13 radian ) .
@xmath14 ; the x - ray efficiency in 2 - 10kev energy bands ( mineo et al . 2004 ) .
@xmath15 ; the x - ray efficiency in 0.3 - 10kev and gamma - ray efficiency above 100mev ( abdo eta l. 2013 ) .
@xmath16 ; the x - ray efficiency calculated with `` absorbed '' fluxes in 2 - 10kev and 20 - 100kev ( campana et al .
2008 ) and gamma - ray efficiency above 100mev ( ackermann et al .
@xmath17 ; the efficiency in 20 - 100kev , including pulsar and pwn and upper limit of gamma - ray efficiency ( renaud et al .
2010 and reference therein ) . [ cols="^,^,^,^,^,^,^",options="header " , ] psrs j0537 - 6910 and j0540 - 6919 are energetic young pulsars in the large magellanic cloud ( hereafter lmc ) , and they were discovered by the x - ray observations ( seward et al .
1984 ; marshall et al .
the spin - down powers of psrs j0537 - 6910 and j0540 - 6919 are @xmath18 and @xmath19 , respectively , which are similar to @xmath20 of the crab pulsar . among currently known pulsars , these three
, psr j0537 - 6910 , crab and j0540 - 6919 , have the top three highest spin - down power ( see table 1 ) . in this paper , `` crab - like pulsars '' is used to refer to all three . the fermi large area telescope ( hereafter fermi - lat )
resolved the gamma - ray emissions from the two crab - like pulsars in the lmc , and furthermore detected the pulsed emissions from psr j0540 - 6919 ( ackermann et al .
2015 ) .
psr j0540 - 6919 ( spin period @xmath21s ) is known as the `` crab - twin '' , because not only the spin - down parameters but also the properties of the pulsed emissions in multi - wavelength bands are similar to those of the crab pulsar .
first , fermi - lat found that the ratio of x - ray luminosity and @xmath22gev gamma - ray luminosity is @xmath23 , which is similar to @xmath24 for the crab pulsar ( abdo et al .
this feature is clearly distinct from @xmath25 of the others ( abdo et al .
2013 for the fermi - lat pulsar catalog ) .
second , the pulse peaks in different wavelength bands are all in phase , just like the pulse profiles of the crab pulsar . furthermore , psr j0540 - 6919 emits the giant radio pulses that appear at the positions of the pulse peaks in higher - energy bands ( johnston et al 2004 ) .
this property is also the same as for the crab pulsar ( shearer et al .
it is likely that there three features represent the nature of the pulsars with @xmath26 . while the crab - like pulsars are similar in their spin - down properties , there are several remarkable differences in the observed radiation : ( 1 ) the pulse shape , and ( 2 ) the radiation efficiency , which is defined as the ratio of the radiation luminosity to the spin - down power @xmath27 .
psr j0540 - 6919 shows a broad pulse profile with a small dip at the center ( campana et al .
2008 for the x - ray pulse and gradari et al .
2011 for the optical pulse ) , while the crab pulsar shows a sharp double - peak structure with the phase separation of @xmath28 ( abdo et al .
the integrated luminosity of the pulsed x - ray / gamma - ray emissions from psr j0540 - 6919 is @xmath29erg / s ( in @xmath13 sr ) , which is about a factor of 3 larger than that of the crab pulsar , @xmath30erg / s . as a result ,
the radiation efficiency @xmath0 of psr j0540 - 6019 is a factor of ten larger than that of the crab pulsar , @xmath2 .
the fermi - lat confirmed that the luminosity of the non - thermal radiation from a pulsar tends to increase as @xmath31 ( abdo et al .
2013 ) , which yields @xmath32 .
this empirical relation can not explain the ratio @xmath33 . for psr j0537 - 6910
, the fermi - lat did not detect the pulsed emissions , and it measured the spectrum fitted by a power - law function , suggesting the emissions originate from the pulsar wind and/or a supernova remnant .
the pulsed x - ray emissions from psr j0537 - 6910 were observed to be @xmath34 in 2 - 10kev
( mineo et al . 2004 ) , indicating the efficiency is @xmath35 , which is much lower than those of the crab and psr j0540 - 6919 . since the distance to the crab ( @xmath36kpc ) , and the lmc ( @xmath37kpc ) are well determined , the uncertainty of the efficiency due to that is the distance should be small
. the observed emission properties of the three energetic pulsars pose a challenge theoretically to see whether a unique model can explain all these systems .
electron / positron acceleration and high - energy emission processes in the pulsar magnetosphere have recently been discussed within the framework of the slot gap model ( harding et al .
2008 ; harding & kalapotharakos 2015 ) , outer gap model ( cheng et al . 2000 , hirotani 2015 , takata et al .
2016 ) , current sheet of the force - free magnetosphere model ( spitkovsky 2006 ; bai & spitkovsky 2010 ) , and pulsar wind model ( aharonian et al .
2012 ) . in this paper
, we will discuss the high - energy emission process within the framework of the outer gap accelerator model .
for the crab - like pulsars , the outer gap model has predicted that most of @xmath38gev photons from the outer gap are converted into pairs by the pair - creation process and can not escape from the light cylinder ( see section [ model ] , cheng et al .
2000 ; takata & chang 2007 ; tang et al .
synchrotron radiation and the inverse - compton process of the secondary pairs can produce the observed emissions in the optical to tev energy bands .
the outer gap model predicts that the shape of the pulse profile is sensitive to the viewing angle and magnetic inclination angle measured from the spin axis . in the fermi - lat pulsar catalog ( abdo et al .
2013 ) , @xmath39 of the sources show a double - peak structure in the pulse profile and @xmath40 show a wide phase ( 0.4@xmath41 0.6 ) separation between the two peaks .
the outer gap model explains the widely separated two peaks by assuming a larger magnetic inclination angle and a larger earth viewing angle ( takata et al .
2011 ; watters & romani 2011 ) .
on the other hand , takata & chang ( 2007 ) explain the pulse profile of psr j0540 - 6919 by a smaller inclination angle @xmath42 and a larger viewing angle @xmath6 .
the observed geometry of the pulsar wind tori also suggests the viewing angle @xmath6 for psr j0540 - 6919 ( ng & romani 2004 , 2008 ) .
previous studies of psr j0540 - 6919 ( zhang & cheng 2000 ; takata & chang 2007 ) mainly discussed the optical / x - ray emissions , since only the upper limit of the gev flux had been reported before the launch of the fermi . hence , it is not obvious why the efficiencies of the observed radiations among the crab - like pulsars are so different . in this paper , therefore , we will revisit the non - thermal emission process of the crab - like pulsars with the outer gap model . in section 2
, we will describe our theoretical model for the crab - like pulsars . in section 3 ,
we present our result of the fitting spectrum for psr j0540 - 6919 and discuss the differences between the crab and this pulsar . in section 4 , we will discuss the emissions from psr j0537 - 6910 .
we apply the calculation method developed in takata et al .
( 2016 ) , which solves the outer gap structure in the three - dimensional space with a rotating dipole magnetic field .
they obtained the structure of the accelerating electric field and gap currents by solving the poisson equation , and the continuity equations for the electrons and positrons and the pair - creation process . for the outer gap model , some electrons / positrons , which migrate along the magnetic field lines , should enter the outer gap from the gap boundaries and they initiate the gamma - ray radiation and subsequent pair - creation cascade processes .
takata et al . ( 2016 ) solved the pair - creation cascade inside the outer gap by assuming the number of the electron / positron injections at the gap boundaries , which is the crucial factor for control of the outer gap structure .
we refer takata et al ( 2016 ) for detailed calculations .
takata et al .
( 2016 ) assumed that the outer gap structure is variable in time , rather than stationary , because of the time - dependent injection of the electrons / positrons at the gap boundaries .
the model argued that the observed gamma - ray spectrum is a superposition of the emissions from different stationary gap structures with different injection rates at the gap boundaries .
this dynamic model provides a better fit for the spectra of the fermi - lat pulsars . as we will argue later ( see section [ discuss ] )
, our model suggests that the observed emissions from the crab - like pulsars are not from primary electrons / positrons accelerated in the outer gap , but from the secondary pairs created outside the gap , and therefore the dynamic behavior of the outer gap may be less important in explaining the pulsar s observed gev spectra . in our calculations ,
the model parameters that determine the gap dynamics are the inclination angle of the magnetic axis , the surface temperature of the neutron star , and the number of the electrons and positrons that enter the outer gap from the gap boundaries .
we assume the magnetic inclination angle @xmath43 less than @xmath44 measured from the spin axis .
for the rotator with a @xmath3 , the positrons and electrons can enter the gap from outside along the magnetic field lines by crossing the inner boundary ( star side ) and outer boundary ( light cylinder side ) , respectively .
we assume that the rate of the particle injection is constant over the inner and the outer boundaries .
we parameterize the injection current in units of the goldreich - julian value and denote @xmath45 and @xmath46 as the normalized injection rates at the gap inner and outer boundaries , respectively .
our local model has to treat the injection currents ( @xmath47 ) as the model fitting parameters .
takata et al .
( 2016 ) discussed the origin of the electrons / positrons that enter the outer gap at the gap boundaries . in the outer gap , the positrons and electrons crossing
the gap boundaries initiate the gamma - ray emission and a subsequent pair - creation cascade .
the electrons / positrons are accelerated by the electric field parallel to the magnetic field and emit the gamma - rays via the curvature radiation process and/or the inverse - compton scattering process .
the emitted gamma - rays may be converted into pairs , by the pair - creation process , with the surface x - rays .
the new pairs created in the gap are accelerated by the electric field and emit the curvature photons .
the pairs created outside the gap lose their energy via the synchrotron radiation and the inverse - compton scattering process .
we denote as `` primary '' pairs as the electrons / positrons accelerated inside the gap , and as `` secondary '' pairs those produced outside the gap .
since no measurements on the surface temperature of the crab - like pulsars have been made , we assume @xmath48k as the temperature of the entire stellar surface . the main difference from the calculation in takata et al .
( 2016 ) is that the current model of the crab - like pulsars takes into account emission from the pairs created outside the outer gap .
the x - rays produced by the synchrotron radiation of the secondary pairs become the target soft - photon field for the photon - photon pair - creation process occurring outside the gap .
one important difference in the circumstellar conditions between the crab - like pulsars and other fermi - lat pulsars is the mean - free path of the pair - creation process between a @xmath49gev photon and a background soft photon produced by the secondary pairs .
the optical depth inside the light cylinder may be written down as @xmath50 where @xmath51 is the number density of the soft photons , @xmath52 is the cross - section , @xmath53 is the x - ray luminosity , @xmath54 is the energy of the soft photon , and @xmath55 is the light cylinder radius .
the optical depth of the crab - like pulsars is usually larger than unity with @xmath56 . for the crab - like pulsars ,
therefore , most of the primary gamma - rays with an energy @xmath49gev are absorbed by the pair - creation process and the secondary particles will emit the x - rays via synchrotron emission .
this explains the ratio of x - ray and gamma - ray luminosity @xmath57 for the crab - like pulsars , where @xmath58 is the _ apparent _ luminosity from the magnetosphere . for other fermi - lat pulsars ,
the mean - free path is of order @xmath59 with @xmath60 , and results in @xmath61 . to calculate the emission from the pairs produced outside the outer gap , we trace the propagation of the @xmath38gev photons and the pair - creation rates on the trajectory .
we calculate the pair - creation mean - free path by assuming the number density of the soft - photons inferred from the observations ; that is , @xmath62 where @xmath63 is the observed spectrum in the optical to hard x - ray energy bands and @xmath64kpc is the distance to the lmc .
the pitch angle , @xmath65 , of the newborn pairs produced outside the outer gap is calculated from @xmath66 where @xmath67 is the unit vector of the magnetic field , and @xmath68 is the propagation direction of the gamma - rays , and @xmath69 and @xmath70 represent the positions of the pair - creation and the radiation , respectively .
the emission direction is calculated from @xmath71 , where @xmath72 is the co - rotation velocity , and @xmath73 is calculated from @xmath74 . in the calculation , there is an uncertainty in the collision angle between the gamma - ray and magnetospheric soft photons .
since the latter are emitted by the secondary pairs , which has a pitch angle @xmath65 , we may assume a collision angle of @xmath75 .
grand based cherenkov telescopes have observed the pulsed emissions up to @xmath76 1tev from the crab pulsar ( abdo et al .
2010 ; aleksi@xmath77 et al . 2011 , 2012 , 2014 ;
aliu et al .
2008 , 2011 ) .
the emissions between 10gev and 1tev are well fitted by a single power - law function .
the standard curvature radiation process can not easily explain the emissions above 100gev from the crab pulsar , which suggests the inverse - compton scattering process inside the magnetosphere ( aleksi@xmath77 et al .
2011 ; harding and kalapotharakos 2015 ) or at the pulsar wind region ( aharonian et al . 2012 ) . within the framework of the outer gap scenario , the @xmath78gev emissions of the crab pulsar
are explained by the emission process of tev primary pairs and/or secondary pairs that were produced by the pair - creation process of the tev photons from the inverse - compton scattering process of the primary pairs .
if the infrared ( ir ) photons from the secondary pairs enter the outer gap , they are up - scattered by @xmath79tev electrons / positrons whose lorentz factor is @xmath80 , and become @xmath81tev gamma - rays .
most of the tev gamma - rays from the outer gap are absorbed by the soft photons outside the gap , and create @xmath79tev electrons / positrons .
the tev secondary pairs also emit photons via synchrotron radiation and inverse - compton scattering processes .
furthermore , the high - energy secondary photons also targets for the pair - creation . in this paper
, we also examine the pair - creation cascade outside the outer gap which is initiated by the tev gamma - rays from the outer gap , and we will discuss its contribution to the observed emissions of psr j0540 - 6919 . since the ir photons are produced above the outer gap , we assume that they irradiate the outer gap at around the upper boundary , say @xmath81% of the gap thickness , with the number density estimated from equation ( [ xnumber ] ) . since the emission direction of the ir will be related to the pitch angle in equation ( [ pitch ] ) , we may roughly estimate the collision angle of the inverse - compton scattering as @xmath82 at the emission point .
and viewing angle @xmath83 ( or @xmath84 ) and @xmath85 .
the dashed - double - dotted line is the expected sensitivity of the cherenkov telescope array ( acharya et al .
the observed x - ray data were taken from de plaa et al .
( 2013 ) . for fermi - lat data
, we read the data from ackermann et al .
( 2015 ) and multiplied 0.75 for estimating the pulsed fluxed , which is suggested in their paper .
, scaledwidth=90.0% ] [ fit ] in this section , we apply the model to psr j0540 - 6919 .
since there are two kinds of the secondary pairs in our calculation , we define the terminology `` low - energy secondary '' which represents the pairs created by the primary curvature photons , and high - energy secondary for those created by primary tev photons via the inverse - compton scattering process .
figure [ j0540spe ] shows the multi - wavelength spectrum of psr j0540 - 6919 with the model fitting curves . in the figure
, the dashed line shows the synchrotron and inverse - compton scattering processes of the low - energy secondary pairs , and dashed - dotted line is the emissions from the high - energy secondary pairs .
the results are for the inclination angle @xmath86 and the observer viewing angle @xmath83 ( or @xmath84 ) .
in addition , we assume that the injection rate at the inner and outer boundaries is 1% of the goldreich - julian value , @xmath85 .
figure [ intrinsic ] shows the intrinsic spectra for the curvature radiation ( solid line ) and inverse - compton scattering process ( dashed line ) inside the gap .
as we can see in figure [ j0540spe ] , the emissions ( dashed line ) from the low - energy secondary pairs explain the observed emissions in the 100ev-1gev energy bands .
however , the calculated spectrum above 1gev decays faster than the fermi - lat data . to reconcile with fermi - lat data above 1gev
, therefore , the present model predicts that the residual curvature emissions ( thin solid line ) and/or the emissions from high - energy secondary pairs ( dashed - dotted line ) contribute to the fermi - lat observations . in the current calculation , a fraction of high - energy photons ( @xmath3810gev ) emitted by the secondary pairs created near the light cylinder can escape from the pair - creation process . as we will argue in section [ discuss ] , the dynamic behavior of the outer gap discussed in takata et al .
( 2016 ) will not be the main reason to explain the observed spectrum above the cut - off energy of psr j0540 - 6919 . with a small inclination angle @xmath86 , the calculated gamma - ray light curve ( solid line in figure [ j0540light ] ) shows a broad pulse with two narrow peaks separated by @xmath87 , which is consistent with the observations .
the inclination angle is @xmath86 for solid line and @xmath88 for dashed line , respectively.,scaledwidth=90.0% ] .
the results are for @xmath85 . , scaledwidth=90.0% ]
the fermi - lat observations found that the efficiency , @xmath89 , of psr j0540 - 6919 is about a factor of ten larger than the crab pulsar , and this result is incompatible with the empirical relation @xmath90 of the fermi - lat pulsars ( abdo et al .
2013 ) . here
we suggest the smaller magnetic inclination of psr j0540 - 6919 causes the larger radiation efficiency than the crab pulsar , whose magnetic inclination angle will be relatively large .
figure [ lumi ] shows the calculated gamma - ray luminosity as a function of the inclination angle . in the figure
, the vertical axis is normalized by the calculated luminosity at @xmath86 .
we find in the figure that the calculation luminosity tends to decrease as the inclination angle increases . in the current model ,
this dependency was caused by the dependency on ( 1 ) the position of the null charge surface of the goldreich - julian charge density and on ( 2 ) the maximum gap current on the inclination angle .
the gap power depends on the thickness of the gap in the poloidal plane , and it decreases with decreasing of the thickness .
the electrodynamics of the conventional gap models expects the relation that @xmath91 , where @xmath92 is defined by the ratio between the size of the outer gap measured on the stellar surface and the polar cap size ( takata et al .
for the crab - like pulsars , the outer gap size may be controlled by the mean - free path of the pair - creation process between the curvature photons and soft x - rays from the neutron star surface ( wang et al .
since the null charge surface on the last - open field lines approaches to the stellar surface with the increase of the inclination angle , the location of the outer gap is closer to the stellar surface for larger inclination angle .
since the number density of x - rays from the stellar surface is inversely proportional to the square of the radial distance , the pair - creation mean - free path inside the gap is shorter for the outer gap closer to the stellar surface .
hence , the outer gap becomes thinner and as a result the gap radiation power decreases with the increasing of the inclination angle . in the current calculation , the pair - creation process inside the gap is occurring due to collision between gev gamma - rays and surface x - rays . in this case , the mean free path of the pair - creation process at around the light cylinder is estimated as @xmath93 . with this mean - free path and the injection rate @xmath94 ,
the gap thickness is determined so as to produce @xmath95 curvature photons inside the outer gap by one particle injected at the outer boundary , and thus to make @xmath96 of pairs by the pair - creation process inside the gap ( see section [ discuss ] ) . with a constant mean - free path @xmath97 , we find that the gev photons have to travel a distance of @xmath98 inside the gap to screen the gap .
the mean - free path actually depends on the position as @xmath99 .
for a smaller inclination angle , because the null charger surface is close to the light cylinder , a constant mean - free path with @xmath100 is a good approximation . for a larger inclination angle , on the other hand
, the null charge surface is closer to the stellar surface and the radial dependency of the mean - free path becomes more important , indicating the average mean - free path is shorter .
therefore , the required travel distance of the gamma - rays to create @xmath96 pairs becomes shorter than @xmath98 of the lower inclination case , and therefore the gap thickness reduces .
it has been suggested that the inner boundary of the middle part of the outer gap tends to be shifted toward the stellar surface as the gap current increases .
the model suggests that the inner boundary will touch on the stellar surface if the gap current is @xmath101 in units of the goldreich - julian value ( takata et al .
2004 ) , which decreases with increasing inclination angle .
this is because the charge density that is created by the gap current at the inner boundary should match with the local value of goldreich - julian charge density , which on the polar cap region is @xmath102 .
one may expect that if the inner boundary of the outer gap once touches on the stellar surface , the latter supplies copious particles to close the outer gap .
therefore , the gap luminosity decreases with increasing inclination angle .
as described in section [ fit ] , the smaller magnetic inclination angle of psr j0540 - 6919 preferentially explains the observed small separation of the two peaks in the pulse profile . for a larger magnetic inclination angle @xmath103 and a viewing angle @xmath6 , the phase - separation between two peaks is @xmath104 , as shown in figure [ j0540light ] , and this would be the case for the crab pulsar .
we emphasize , therefore , that the smaller inclination magnetic angle of psr j0540 - 6919 can explain both the higher radiation efficiency and the narrower phase separations of the two peaks than those of the crab pulsar .
fermi - lat resolved the gamma - ray emissions from the high spin - down powered pulsar , j0537 - 6910 , in the lmc with a flux level of @xmath105 .
however , the pulsed emissions in fermi - lat data have yet to be confirmed , and the observed spectrum fitted by a single power - low function indicates the emissions to be from the pulsar wind nebula and/or a supernova remnant ( ackermann et al .
2015 ) . since psr j0537 - 6910 has the largest spin - down power ( @xmath106 ) and the strongest magnetic field at the light cylinder ( @xmath107 g ) among the known pulsars ( see the atnf pulsar catalog , manchester et al .
2005 ) , it is likely that this pulsar produces gamma - rays in the magnetosphere , and but they are buried under the background emission , or the gamma - ray beam is out of the line of sight .
the observed emission properties of psr j0537 - 6910 are very different from those of the crab and j0537 - 6910 ; ( 1 ) the pulsed emissions have been discovered only in the x - ray bands , ( 2 ) the observed radiation efficiency in the x - rays is very low @xmath108 , and ( 3 ) the pulse width , @xmath109 , in the x - ray bands ( marshall et al .
1998 ) is narrower than those of the other pulsars .
no detection of the pulsed emissions by the fermi - lat makes if difficult for us to discuss the electromagnetic spectrum in the wide energy bands , and to constrain the magnetic inclination and the earth viewing angle .
however , we may expect that the radiation process of psr j0537 - 6910 is similar to those of the crab and j0540 - 6919 , and we may assume that the flux level of the pulsed gamma - rays measured on the earth is @xmath110 , which is the case for the crab and psr j0540 - 6919 .
under those assumptions , the observed radiation efficiency will be of the order of @xmath111 , which is about two orders of magnitude smaller than that of j0540 - 6919 .
as expected from figure [ lumi ] , we would say that it is difficult to explain @xmath112 by the effect of the inclination angle . if both psrs j0537 - 6910 and j0540 - 6919 have a viewing angle @xmath6 , it is also difficult to explain the difference in the pulse width with the difference in the inclination angle .
we suggest therefore that the earth viewing angle is very different between the two pulsars .
our model suggests that the earth viewing angle of psr j0540 - 6919 is close to @xmath6 measured from the spin axis , which is also suggested by a study of the pulsar wind ( ng & romani 2004,2008 ) . since most of the pairs inside the gap are created around the null charge surface , the outer gap emission is stronger for an earth viewing angle of @xmath6 .
as the viewing angle deviates from the @xmath6 , therefore , the observed gap emission rapidly decreases and hence the apparent radiation efficiency decreases ( see figures 3 and 4 in takata et al .
2011 ) ; at the same time , the pulse width becomes narrower . on these grounds ,
we speculate that the main reason for difference in the observed efficiencies and in the observed pulsed widths between psrs j0540 - 6919 and j0537 - 6910 is the difference in the earth viewing angle . in figure [ j0540spe ]
, we assumed the same particle injection rates at the inner and outer boundaries .
the assumption of equal injection rates at the gap boundaries is arbitrary , and it is not necessary for the real case . in the current
_ local _ model , however , it would not be possible to consistently solve the injection particles at the gap boundaries , for which we would have to solve the global structure including the polar cap activities , outer gap activities , and pulsar wind region . to see the dependency on the choice of the injection current
, we examined the case for @xmath113 and @xmath114 , that is , no particles enter into the gap from the inner boundary ( star side ) or outer boundary ( light cylinder side ) , respectively .
figure [ j0540spe2 ] summarizes the dependency of the emissions from the low - energy secondary pairs on the injection currents @xmath45 and @xmath46 ; the solid line , dashed line and dashed - dotted line are results for @xmath115 , @xmath116 and @xmath117 , respectively .
we find in the figure that the calculated spectra become harder for @xmath114 ( dashed line in figure [ j0540spe2 ] ) .
this is related to the fact that most of the pairs are created by the inwardly propagating gamma - rays .
collision with the x - rays from the surface is a head - on process for inwardly propagating gamma - rays , while it is tail - on process for outwardly propagating gamma - rays .
hence , the mean - free path of the former is shorter than that of latter , and most of the pairs are created by the inwardly propagating gamma - rays .
this indicates that the gap size is mainly controlled by the pair - creation process of the inwardly propagating gamma - rays . for @xmath114 ,
therefore , the outer gap has to be thick to create enough pairs , and as a result the calculated spectrum becomes harder . in our calculation , the gap structure is controlled by the magnitude of @xmath46 , except for the case @xmath119 .
we quantitatively discuss how the gap size depends on the injection rate @xmath46 . since the gap thickness of the crab - like pulsar is about 10% of the light cylinder radius , we can approximate that the propagation direction of the gamma - rays is the same as the direction of the particle s motion . under this approximation ,
the evolution of the number density of the inwardly moving particles ( electrons ) and gamma - rays in the pair - creation region is described by @xmath120 and @xmath121 respectively , where @xmath122 is the distance from the outer boundary , and @xmath123 and @xmath124 are number density and photon number density normalized by the goldreich - julian density , respectively .
we ignore the effect of the pair - creation by the gamma - rays propagating outward .
in addition , @xmath125 and @xmath126 are the mean free path of the photon - photon pair - creation process and the rate of the curvature radiation . in the present calculation ,
since we assume the surface temperature @xmath127k , the mean - free path inside the gap at the light cylinder is estimated as @xmath128 , where we used @xmath129 and @xmath130 with @xmath131 being stefan - boltzmann constant .
the rate of the curvature radiation is estimated as @xmath132 , where @xmath133 being the curvature radius , and @xmath134 the lorentz factor of the accelerated particles . to solve the equations ( [ dn ] ) and ( [ dg ] )
, we impose the boundary conditions as @xmath135 and @xmath136 , where @xmath137 represents the outer boundary .
we assume that the rate of the curvature radiation process , @xmath126 , is constant along the magnetic field line . by assuming that mean free path is constant along the distance @xmath122 from the outer boundary ( @xmath138 )
, we find the solution that @xmath139 where @xmath140 .
since we consider the surface x - ray emission as the soft - photon field for the photon - photon pair - creation process inside the outer gap , the mean - free path will decrease as @xmath141 . to take into account this effect
, we explore the solution with the mean - free path in the form @xmath142 .
the solution becomes @xmath143 , \label{sol2}\ ] ] where @xmath144 with @xmath145 .
figure [ pair ] shows the evolution of the ratio of the local number density and that at the outer boundary of the inwardly moving particles , @xmath146 , as a function of the distance from the outer boundary .
the solid and dashed lines are solutions given by equations ( [ sol2 ] ) and ( [ sol1 ] ) , respectively ; here we adopted @xmath147 and @xmath148 .
we can find in the figure that the multiplicity of the particles injected at the outer boundary becomes @xmath96 if the gamma - rays travel @xmath98 from the outer boundary .
this suggests that if the injection rate at the boundary is 1% of the goldreich - julian value ( that is , @xmath94 ) , the number density becomes the goldreich - julian value after the gamma - ray travels @xmath98 in the outer gap and the created pairs will significantly screen the accelerating electric field .
this is consistent with the gap structure solved in this paper .
figure [ pair ] also indicates that for a smaller injection rate , the gamma - rays have to travel a greater distance to achieve the goldreich - julian number density of the pairs by the pair - creation , and hence the gap size becomes larger .
as figure [ j0540spe ] shows , the current model with using constant injection rate @xmath149 predicts that the residual curvature radiation of the primary particles and/or the emissions from the high - energy secondary pairs can explain the observed emissions above 1gev .
takata et al .
( 2016 ) , on the other hand , argued that sub - exponential decays of the gev spectra of the fermi - lat pulsars reflect the time - dependent emission process of the outer gap .
they proposed that the injection rate at the gap boundaries is time - dependent variable and the observed gamma - ray spectrum is emitted from different gap structures with different injection rates . in the model , the observed spectrum was fitted better as the superposition of several power - law plus exponential cut - off functions with varying the cut - off energy , for which the different components are produced at the different injection rates at the gap boundaries .
we discuss the shape of the observed gev spectrum of psr j0540 - 6191 by the fermi - lat with the dynamics model in takata et al .
we find however that the calculated gev spectra do not greatly affect the assumed extent of injection rate at the gap boundaries .
this is because the gev gamma - rays observed on the earth do not come from the primary pairs in the gap , but the secondary pairs that are decelerated by the radiation process .
figure [ j0540spe1 ] shows the spectra of the emissions from the low - energy secondary pairs calculated with different injection rates ; @xmath150 ( solid line ) , @xmath151 ( dashed line ) and @xmath152 ( dotted line ) .
we can see that the hardness ( peak energy ) of the `` synchrotron emissions '' ( low - energy component ) increases with decreasing of the injection rate .
this is because the gap thickness increases and hence the electric field in the gap becomes stronger as the injection rate decreases . as a result ,
the energy distribution of the low - energy secondary pairs that emit synchrotron photons becomes harder for a gap with a smaller injection rate .
on the other hand , the energy peak of the spectra by `` inverse - compton scattering '' ( high - energy component ) does not greatly depend on the injection rate .
this is because the synchrotron cooling is more important for the particles with a lorentz factor @xmath153 than the radiation cooling , due to the inverse - compton scattering .
as figure [ j0540spe1 ] shows , therefore , the energy peak of the inverse - compton scattering of low - energy secondary pairs always appears around @xmath96mev , regardless of the injection rates . as a result , it is obvious from figure [ j0540spe1 ] that even if we superpose the emissions calculated with different injection rates , the combined spectrum decays faster and still has a large discrepancy with the fermi - lat spectrum above 1gev . on these grounds ,
we conclude that the residual curvature photons and/or the emissions from the high - energy secondary pairs contribute to the observed emissions above 1gev . in summary
, we discussed the gamma - ray emissions from the crab - like pulsars , psrs j0537 - 6910 and j0540 - 6919 , in the lmc .
the pulsed emissions from psr j0540 - 6919 is observed to have an efficiency that is a factor of ten larger than that of the crab pulsar . by solving the electrodynamics of the outer gap accelerator
, we concluded that the difference in the radiation efficiencies of psr j0540 - 6919 and the crab pulsar is caused by the difference in the inclination angle .
inferred from the observed x - ray emissions , the radiation efficiency of psr j0537 - 6910 is about two orders of magnitude smaller than that of psr j0540 - 6919 .
because of the very narrow x - ray pulse and low radiation efficiency of psr j0537 - 6910 , we suspect that the earth viewing angle of psr j0537 - 6910 greatly deviates from @xmath6 , which is the case for the crab and psr j0540 - 6919 .
the authors thank k. hirotani , the referee , for insightful comments and suggestions on the manuscript .
we al thank a. h. kong , c. y. hui , p. h. t. tam , m. ruderman , and s. shibata for the useful discussions .
jt is supported by the nsfc grants of china under 11573010 .
ksc is supported by a grf grant of the hong kong government under hku17300814p .
all calculations were done under the high performance computing cluster ( hyperion ) of the institute of particle physics and astrophysics , hust .
abdo , a. a. ; ackermann , m. ; ajello , m. ; atwood , w. b. ; axelsson , m. ; baldini , l. ; ballet , j. et al . , 2010 , apj , 708 , 1254 abdo , a. a. ; ajello , m. ; allafort , a. ; baldini , l. ; ballet , j. ; barbiellini , g. ; baring , m. g. ; bastieri , d. , et al .
, 2013 , apjs , 208 , 17 acharya , b. s. ; actis , m. ; aghajani , t. ; agnetta , g. ; aguilar , j. ; aharonian , f. ; ajello , m. ; akhperjanian , a. et al . , 2013 , aph , 43 , 3 ackermann , m. ; albert , a. ; baldini , l. ; ballet , j. ; barbiellini , g. ; barbieri , c. ; bastieri , d. ; bellazzini , r. , et al . , 2015 , sci , 350 , 801 aharonian , f. a. ; bogovalov , s. v. ; khangulyan , d. , 2012 , nature , 482 , 507 bai , x .-
n . ; spitkovsky , a. , 2010 , apj , 715 , 1282 aleksi@xmath77 et al . 2011 , apj , 742 , 43 aleksi@xmath77 et al .
2012 , a&a , 541 , 13 aleksi@xmath77 et al .
2014 , a&al , 565 , 12 aliu , e. , et al . , 2008 , sci , 322 , 1221 aliu , e. , et al . , 2011 ,
sci , 334 , 69 campana , r. ; mineo , t. ; de rosa , a. ; massaro , e. ; dean , a. j. ; bassani , l. , 2008 , mnras , 389 , 691 cheng , k. s. ; ruderman , m. ; zhang , l. , 2000 , apj , 537 , 964 de plaa , j. ; kuiper , l. ; hermsen , w. , 2003 , a&a , 400 , 1013 gradari , s. ; barbieri , m. ; barbieri , c. ; naletto , g. ; verroi , e. ; occhipinti , t. ; zoccarato , p. ; german@xmath157 , c. , 2011 , mnras , 412 , 2689 harding , a.k . ; kalapotharakos , c. , 2015 , apj , 811 , 63 johnston , simon ; romani , roger w. ; marshall , f. e. ; zhang , w. , 2004 , mnras , 355,31 manchester , r. n. ; hobbs , g. b. ; teoh , a.;hobbs , m. , astron .
j. , 129 , 1993 - 2006 ( 2005 ) ( astro - ph/0412641 ) marshall , f. e. ; gotthelf , e. v. ; zhang , w. ; middleditch , j. ; wang , q. d. , 1998 , apj , letter , 499 , 179 mineo , t. ; cusumano , g. ; massaro , e. 2004 , nuclear physics b proceedings supplements , volume 132 , p. 632 - 635
ng , c .- y . ; romani , roger w. , 2008 , apj , 673 , 411 ng , c .- y . ; romani , roger w. , 2004 , apj , 601 , 479 renaud , m. ; marandon , v. ; gotthelf , e. v. ; et al .
2010 , apj , 716 , 663 seward , f. d. ; harnden , f. r. , jr .
; helfand , d. j. , 1984 , apj , letter , 287 , 19 shearer , a. ; stappers , b. ; oconnor , p. ; golden , a. ; strom , r. ; redfern , m. ; ryan , o. , 2003 , sci , 301 , 4923 spitkovsky , a. , 2006 , apj , letter , 648 , 51 takata , j. ; ng , c. w. ; cheng , k. s. , 2016 , mnras , 455 , 4249 takata , j. ; wang , y. ; cheng , k. s. , 2011 , mnras , 415 , 1827 takata , j. ; wang , y. ; cheng , k. s. , 2010 , apj , 715 , 1318 takata , j. ; chang , h .- k . , 2007 , apj , 670 , 67 takata , j. ; shibata , s. ; hirotani , k. , 2004 , mnras , 348 , 241 tang , anisia p. s. ; takata , j. ; jia , j. j. ; cheng , k. s. , 2008 , apj , 676 , 562 wang , y. ; takata , j. ; cheng , k. s. , 2010 , apj , 720 , 178 watters , kyle p. ; romani , roger w. , 2011 , apj , 727 , 123 | we discuss x - ray and gamma - ray emissions from crab - like pulsars , psrs j0537 - 6910 and j0540 - 6919 , in large magellanic cloud .
fermi - lat observations have resolved the gamma - ray emissions from these two pulsars and found the pulsed emissions from psr j0540 - 6919 . the total pulsed radiation in the x - ray / gamma - ray energy bands of psr j0540 - 6919 is observed with the efficiency @xmath0 ( in 4@xmath1 sr ) , which is about a factor of ten larger than @xmath2 of the crab pulsar . although psr j0537 - 6910 has the highest spin - down power among currently known pulsars , the efficiency of the observed x - ray emissions is about two orders of magnitude smaller than that of psr j0540 - 6919 .
this paper mainly discusses what causes the difference in the radiation efficiencies of these three energetic crab - like pulsars .
we discuss electron / positron acceleration and high - energy emission processes within the outer gap model . by solving the outer gap structure with the dipole magnetic field ,
we show that the radiation efficiency decreases as the inclination angle between the magnetic axis and the rotation axis increases . to explain the difference in the pulse profile and in the radiation efficiency , our model suggests that psr j0540 - 6919 has an inclination angle much smaller than the that of crab pulsar ( here we assume the inclination angles of both pulsars are @xmath3 ) . on the other hand
, we speculate that the difference in the radiation efficiencies between psrs j0537 - 6910 and j0549 - 6919 is mainly caused by the difference in the earth viewing angle , and that we see psr j0537 - 6910 with an earth viewing angle @xmath4 ( or @xmath5 ) measured from the spin axis , while we see psr j0540 - 6919 with @xmath6 . |
throughout this paper we work in the piecewise linear category .
let @xmath4 be an embedding of a finite graph @xmath5 into the @xmath0-sphere .
then @xmath4 is called a _ spatial embedding _ of @xmath5 and @xmath6
is called a _ spatial graph_.
we denote the set of all spatial embeddings of @xmath5 by @xmath7 .
we call a subgraph @xmath8 of @xmath5 which is homeomorphic to the circle a _ cycle _ of @xmath5 . for a positive integer @xmath9
, @xmath10 denotes the set of all cycles of @xmath5 if @xmath11 and the set of all unions of mutually disjoint @xmath9 cycles of @xmath5 if @xmath12 .
in particular , we denote @xmath13 by @xmath14 simply . for an element @xmath15 in @xmath10 and a spatial embedding @xmath4 of @xmath5 , @xmath16 is none other than a knot if @xmath11 and an @xmath9-component link if @xmath12 .
a graph @xmath5 is said to be _ intrinsically linked _ ( il ) if for every spatial embedding @xmath4 of @xmath5 , @xmath6 contains a nonsplittable @xmath1-component link .
conway - gordon @xcite and sachs @xcite showed that @xmath17 is il , where @xmath18 denotes the _ complete graph _ on @xmath19 vertices .
moreover , il graphs have been completely characterized as follows . for a graph @xmath5 and an edge @xmath20 of @xmath5
, we denote the subgraph @xmath21 by @xmath22 .
let @xmath23 is an edge of @xmath5 which is not a loop .
we call the graph which is obtained from @xmath22 by identifying the end vertices @xmath24 and @xmath25 the _ edge contraction of @xmath5 along @xmath20 _ and denote it by @xmath26 . a graph @xmath27 is called a
_ minor _ of a graph @xmath5 if there exists a subgraph @xmath28 of @xmath5 and the edges @xmath29 of @xmath28 such that @xmath27 is obtained from @xmath28 by a sequence of edge contractions along @xmath29 . a minor @xmath27 of @xmath5
is called a _ proper minor _ if @xmath27 does not equal @xmath5 .
let @xmath30 be a property for graphs which is _ closed _ under minor reductions ; that is , for any graph @xmath5 which does not have @xmath30 , all minors of @xmath5 also do not have @xmath30 .
a graph @xmath5 is said to be _ minor - minimal _ with respect to @xmath30 if @xmath5 has @xmath30 but all proper minors of @xmath5 do not have @xmath30 .
note that @xmath5 has @xmath30 if and only if @xmath5 has a minor - minimal graph with respect to @xmath30 as a minor . by the famous theorem of robertson - seymour @xcite , there are finitely many minor - minimal graphs with respect to @xmath30 .
neetil - thomas @xcite showed that il is closed under minor reductions , and robertson - seymour - thomas @xcite showed that the set of all minor - minimal graphs with respect to il equals the _ petersen family _ which is the set of all graphs obtained from @xmath17 by a finite sequence of _ @xmath2-exchanges _ and _ @xmath3-exchanges_. here a @xmath2-exchange is an operation to obtain a new graph @xmath31 from a graph @xmath32 by removing all edges of a cycle @xmath33 of @xmath32 with exactly three edges @xmath34 and @xmath35 , and adding a new vertex @xmath36 and connecting it to each of the vertices @xmath37 and @xmath38 as illustrated in fig .
[ delta - y ] ( we often denote @xmath39 by @xmath40 ) .
a @xmath3-exchange is the reverse of this operation .
this family contains exactly seven graphs as illustrated in fig .
[ petersen ] , where @xmath41 means that @xmath28 can be obtained from @xmath5 by a single @xmath2-exchange .
note that @xmath42 is isomorphic to the _ petersen graph_. we remark here that if @xmath32 is il then @xmath31 is also il @xcite , and if @xmath31 is il then @xmath32 is also il @xcite .
namely @xmath2 and @xmath3-exchanges preserve il .
on the other hand , a graph @xmath5 is said to be _ intrinsically knotted _ ( ik ) if for every spatial embedding @xmath4 of @xmath5 , @xmath6 contains a nontrivial knot .
conway - gordon @xcite showed that @xmath43 is ik .
fellows and langston @xcite showed that ik is closed under minor reductions , and motwani - raghunathan - saran @xcite showed that @xmath43 is a minor - minimal ik graph . although additional minor - minimal ik graphs are known by kohara - suzuki @xcite and foisy @xcite , @xcite , ik graphs have not been completely characterized yet .
we remark here that if @xmath32 is ik then @xmath31 is also ik @xcite , but if @xmath31 is ik then @xmath32 may not always be ik .
namely the @xmath3-exchange does not preserve ik in general .
actually flapan - naimi @xcite exhibited that there exists a graph @xmath44 which is obtained from @xmath43 by five times of @xmath2-exchanges and twice @xmath3-exchanges such that it is not ik .
we call the set of all graphs obtained from @xmath43 by a finite sequence of @xmath2 and @xmath3-exchanges the _ heawood family_. or @xmath45 by a finite sequence of @xmath2 and @xmath3-exchanges the heawood family , where @xmath45 is the complete @xmath46-partite graph on @xmath47 vertices . ]
this family contains exactly twenty graphs as illustrated in fig .
[ heawood ] , where @xmath41 means that @xmath28 can be obtained from @xmath5 by a single @xmath2-exchange .
note that @xmath48 is isomorphic to the _ heawood graph _
, see remark [ hea ] .
kohara - suzuki @xcite showed that a graph @xmath5 in the heawood family is a minor - minimal ik graph if @xmath5 is obtained from @xmath43 by a finite sequence of @xmath2-exchanges , namely @xmath5 is one of fourteen graphs @xmath43 , @xmath49 , @xmath50 , @xmath51 and @xmath52 . in (
5 ) is wanting .
] on the other hand , @xmath53 is isomorphic to @xmath44 , namely @xmath53 is not ik .
our first purpose in this paper is to determine completely when a graph in the heawood family is ik as follows .
[ main1 ] let @xmath5 be a graph in the heawood family
. then the following are equivalent : 1 .
@xmath5 is ik , 2 .
@xmath5 is obtained from @xmath43 by a finite sequence of @xmath2-exchanges , 3 .
@xmath54 is the empty set .
namely , each of the graphs @xmath55 and @xmath56 in the heawood family is not ik , and only these graphs in the heawood family contain a union of mutually disjoint three cycles .
our second purpose in this paper is to show that any of the graphs in the heawood family is a minor - minimal graph with respect to a certain kind of intrinsic nontriviality even if it is not ik .
we say that a graph @xmath5 is _ intrinsically knotted or completely @xmath0-linked _
( i(k or c3l ) ) if for every spatial embedding @xmath4 of @xmath5 , @xmath6 contains a nontrivial knot or a @xmath0-component link any of whose @xmath1-component sublink is nonsplittable .
note that an ik graph is i(k or c3l ) .
as we will show in proposition [ closed ] , i(k or c3l ) is closed under minor reductions .
then we have the the following . [ main2 ] all of the graphs in the heawood family are minor - minimal i(k or c3l ) graphs . actually , each of the graphs @xmath55 and @xmath56 in the heawood family is not ik but i(k or c3l ) , and they are minor - minimal with respect to i(k or c3l ) . [ rem ] 1 .
a graph @xmath5 is said to be _ intrinsically @xmath9-linked _
( i@xmath9l ) if for every spatial embedding @xmath4 of @xmath5 , @xmath6 contains a nonsplittable @xmath9-component link @xcite @xcite .
note that i@xmath1l coincides with il .
let @xmath5 be a graph in the heawood family which is not ik .
then we will show in example [ not3l ] that there exists a spatial embedding @xmath4 of @xmath5 such that @xmath6 does not contain a nonsplittable @xmath0-component link .
namely @xmath5 is neither ik nor i@xmath0l .
a graph @xmath5 is said to be _ intrinsically knotted or @xmath0-linked _
( i(k or 3l ) ) if for every spatial embedding @xmath4 of @xmath5 , @xmath6 contains a nontrivial knot or a nonsplittable @xmath0-component link @xcite .
it is clear that i(k or c3l ) implies i(k or 3l ) , but the converse is not true .
actually in @xcite , although foisy discovered an i(k or 3l ) graph @xmath5 and exhibit a spatial embedding @xmath4 of @xmath5 such that @xmath6 contains a nonsplittable @xmath0-component link but does not contain a nontrivial knot , each of the nonsplittable @xmath0-component links in @xmath6 contains a split @xmath1-component sublink .
the rest of this paper is organized as follows . in the next section ,
we show the general results about graph minors , @xmath2-exchanges and spatial graphs .
we prove theorems [ main1 ] and [ main2 ] in sections @xmath0 and @xmath46 , respectively .
let @xmath27 be a minor of a graph @xmath5 . then there exists a natural injection @xmath57 for any positive integer @xmath9 . in particular ,
we denote @xmath58 by @xmath59 simply .
let @xmath4 be a spatial embedding of @xmath5 and @xmath20 an edge of @xmath5 which is not a loop .
then by contracting @xmath60 into one point , we obtain a spatial embedding @xmath61 of @xmath26 .
similarly we also can obtain a spatial embedding @xmath61 of @xmath27 from @xmath4 .
thus we obtain a map @xmath62 then we immediately have the following . [ map2 ] for a spatial embedding @xmath4 of @xmath5 and an element @xmath15 in @xmath63 , @xmath64 is ambient isotopic to @xmath65 .
@xmath66 now we show the following .
[ closed ] i(k or c3l ) is closed under minor reductions .
let @xmath5 be a graph which is not i(k or c3l ) and @xmath27 a minor of @xmath5 .
let @xmath4 be a spatial embedding of @xmath5 which contains neither a nontrivial knot nor a @xmath0-component link any of whose @xmath1-component sublink is nonsplittable .
then by proposition [ map2 ] , @xmath61 also contains neither a nontrivial knot nor a @xmath0-component link any of whose @xmath1-component sublink is nonsplittable .
this implies that @xmath27 is not i(k or c3l ) .
[ closed2 ] proposition [ map2 ] also implies that ik , i@xmath9l and i(k or 3l ) are closed under minor reductions .
let @xmath32 and @xmath31 be two graphs such that @xmath31 is obtained from @xmath32 by a single @xmath2-exchange as we said in the previous section .
let @xmath15 be an element in @xmath67 which does not contain @xmath33 .
then there exists an element @xmath68 in @xmath69 such that @xmath70 .
thus we obtain a map @xmath71 for any positive integer @xmath9 .
in particular , we denote @xmath72 by @xmath73 simply .
note that @xmath74 is surjective and the inverse image of @xmath15 by @xmath74 contains at most two elements in @xmath67 for any element @xmath15 in @xmath69 .
note also that the surjectivity of @xmath74 implies the following .
[ main1lem2 ] for @xmath12 , if @xmath67 is the empty set , then @xmath69 is also the empty set .
@xmath66 let @xmath4 be a spatial embedding of @xmath31 and @xmath75 a @xmath1-disk in the @xmath0-sphere such that @xmath76 and @xmath77 .
let @xmath78 a spatial embedding of @xmath32 such that @xmath79 for @xmath80 and @xmath81 .
thus we obtain a map @xmath82 then we immediately have the following .
[ map ] for a spatial embedding @xmath4 of @xmath31 and an element @xmath15 in @xmath69 , @xmath16 is ambient isotopic to @xmath83 for each element @xmath84 in the inverse image of @xmath15 by @xmath74 .
@xmath66 now we show the following lemmas .
[ dy ] if @xmath32 is i(k or c3l ) , then @xmath31 is also i(k or c3l ) .
assume that @xmath31 is not i(k or c3l ) , namely there exists a spatial embedding @xmath4 of @xmath31 which contains neither a nontrivial knot nor a @xmath0-component link any of whose @xmath1-component sublink is nonsplittable . in the following we show that @xmath85 also contains neither a nontrivial knot nor a @xmath0-component link any of whose @xmath1-component sublink is nonsplittable .
let @xmath8 be an element in @xmath86 .
if @xmath8 is not @xmath33 , then @xmath87 is ambient isotopic to @xmath88 by proposition [ map ] and @xmath88 is a trivial knot by the assumption . since @xmath89 is also a trivial knot , it follows that @xmath85 does not contain a nontrivial knot .
let @xmath15 be an element in @xmath90 .
if @xmath15 does not contain @xmath33 , then @xmath91 is ambient isotopic to @xmath92 by proposition [ map ] and @xmath92 is a @xmath0-component link which contains a split @xmath1-component sublink by the assumption . if @xmath15 contains @xmath33 , then @xmath91 is a split @xmath0-component link .
thus we see that @xmath85 does not contain a @xmath0-component link any of whose @xmath1-component sublink is nonsplittable .
[ minor - min - lem ] if @xmath31 is minor - minimal for i(k or c3l ) , then @xmath32 is also minor - minimal for i(k or c3l ) . in the following
we show that for any edge @xmath20 of @xmath32 which is not a loop , there exists a spatial embedding @xmath4 of @xmath93 and a spatial embedding @xmath94 of @xmath95 such that each of @xmath96 and @xmath97 contains neither a nontrivial knot nor a @xmath0-component link any of whose @xmath1-component sublink is nonsplittable . if @xmath20 is not
@xmath98 , @xmath99 or @xmath35 , then there exists a spatial embedding @xmath100 of @xmath101 and a spatial embedding @xmath102 of @xmath103 such that both @xmath104 and @xmath105 contains neither a nontrivial knot nor a @xmath0-component link any of whose @xmath1-component sublink is nonsplittable .
note that @xmath101 ( resp .
@xmath103 ) is obtained from @xmath93 ( resp .
@xmath95 ) by a single @xmath2-exchange at the same @xmath33
. then we see that each of @xmath106 and @xmath107 contains neither a nontrivial knot nor a @xmath0-component link any of whose @xmath1-component sublink is nonsplittable in the similar way as the proof of lemma [ dy ] .
if @xmath20 is one of @xmath98 , @xmath99 and @xmath35 , we may assume that @xmath23 without loss of generality .
now there exists a spatial embedding @xmath100 of @xmath108 such that @xmath109 contains neither a nontrivial knot nor a @xmath0-component link any of whose @xmath1-component sublink is nonsplittable
. then we can see that @xmath110 . on the other hand
, there exists a spatial embedding @xmath102 of @xmath111 such that @xmath112 contains neither a nontrivial knot nor a @xmath0-component link any of whose @xmath1-component sublink is nonsplittable .
take a @xmath1-disk @xmath113 in the @xmath0-sphere such that @xmath114 and @xmath115 .
then @xmath116 may be regarded as the image of a spatial embedding of @xmath117 , which is denoted by @xmath94 .
it is clear that @xmath118 contains neither a nontrivial knot nor a @xmath0-component link any of whose @xmath1-component sublink is nonsplittable .
[ ot_rem ] lemma [ minor - min - lem ] has already been proven by ozawa - tsutsumi in a more general form ( * ? ? ?
* lemma 3.1 , exercise 3.2 ) .
but we give a proof of lemma [ minor - min - lem ] as described above for the reader s convenience .
[ main1lem ] each of the graphs @xmath55 and @xmath56 in the heawood family is not ik .
since the case of @xmath53 has been already shown by flapan - naimi @xcite , we show that each of the graphs @xmath119 and @xmath56 is not ik .
let @xmath120 be the spatial embedding of @xmath121 as illustrated in fig .
[ no_knots ] .
then it can be checked directly that @xmath122 does not contain a nontrivial knot .
thus @xmath121 is not ik .
let @xmath123 be the spatial embedding of @xmath124 as illustrated in fig .
[ no_knots ] .
let @xmath125 be the map from @xmath126 to @xmath127 induced by the @xmath3-exchange from @xmath124 to @xmath121 at the marked @xmath40 as illustrated in fig .
[ no_knots ] .
then it is clear that @xmath128 .
since @xmath122 does not contain a nontrivial knot , by proposition [ map ] it follows that @xmath129 also does not contain a nontrivial knot .
namely @xmath124 is not ik . by repeating this argument
, we can see that each of the graphs @xmath130 and @xmath56 is also not ik , see fig . [
no_knots ] .
first we show that ( 1 ) and ( 2 ) are equivalent .
since we have already known that ( 2 ) implies ( 1 ) , we show that ( 1 ) implies ( 2 ) . if @xmath5 is ik , then by lemma [ main1lem ] we see that @xmath5 is not one of @xmath55 or @xmath56 . namely @xmath5 is obtained from @xmath43 by a finite sequence of @xmath2-exchanges .
next we show that ( 2 ) and ( 3 ) are equivalent .
assume that @xmath5 is obtained from @xmath43 by a finite sequence of @xmath2-exchanges .
note that @xmath131 is the empty set .
thus by proposition [ main1lem2 ] we see that @xmath54 is the empty set .
conversely , if @xmath5 is one of @xmath55 , and @xmath56 , then @xmath54 is not the empty set .
this completes the proof .
[ fn_revisited ] let @xmath132 be the spatial embedding of @xmath133 as illustrated in fig .
[ no_knots ] and @xmath134 the spatial embedding of @xmath53 as illustrated in fig .
let @xmath135 be the map from @xmath136 to @xmath137 induced by the @xmath3-exchange from @xmath133 to @xmath53 at the double - marked @xmath40 as illustrated in fig .
then it is clear that @xmath138 .
moreover we can see that @xmath134 coincides with flapan - naimi s example of a spatial embedding of @xmath53 whose image does not contain a nontrivial knot as illustrated in the left side of fig .
[ fn ] @xcite .
[ cgty ] ( conway - gordon @xcite , taniyama - yasuhara @xcite ) let @xmath5 be a graph in the petersen family and @xmath4 a spatial embedding of @xmath5 .
then there exists an element @xmath15 in @xmath139 such that @xmath140 is odd , where @xmath141 denotes the _ linking number _ in the @xmath0-sphere .
let @xmath142 be the graph as illustrated in fig .
we denote the set of all cycles with exactly four edges of @xmath142 by @xmath143 . for a spatial embedding @xmath4 of @xmath142 , we define @xmath144 where @xmath145 denotes the second coefficient of the _ conway polynomial_. note that @xmath146 of a knot @xmath147 is congruent to the _ arf invariant _ modulo two @xcite .
then the following is known .
[ d4type ] let @xmath5 be a graph which contains @xmath142 as a minor and @xmath4 a spatial embedding of @xmath5 . for two elements @xmath153 and
@xmath154 in @xmath155 , if both @xmath156 and @xmath157 are odd , then @xmath151 . for two elements @xmath15 and @xmath84 in @xmath149 , we see that both @xmath158 and @xmath159 are odd by the assumption
. then by proposition [ map2 ] , it follows that @xmath160 and @xmath161 are also odd .
thus by lemma [ tya2 ] , we have that @xmath162 [ n9fn ] let @xmath5 be @xmath121 or @xmath53 . for every spatial embedding @xmath4 of @xmath5 , there exists an element @xmath8 in @xmath14 such that @xmath163 is odd , or there exists an element @xmath15 in @xmath54 such that each @xmath1-component sublink of @xmath16 has an odd linking number .
first we show in the case of @xmath167 .
let @xmath4 be a spatial embedding of @xmath121 .
note that @xmath121 contains @xmath17 as the proper minor @xmath168 thus by lemma [ cgty ] , there exists an element @xmath169 in @xmath170 such that @xmath171 is odd .
hence by proposition [ map2 ] , there exists an element @xmath153 in @xmath172 such that @xmath156 is odd .
note that @xmath172 consists of ten elements , and by the symmetry of @xmath121 , we may assume that @xmath173\cup [ 2\ 6\ 5\ 8]$ ] or @xmath174\cup [ 4\ 5\ 6]$ ] without loss of generality .
note that @xmath121 contains @xmath175 as the proper minor @xmath176 thus by lemma [ cgty ] , there exists an element @xmath177 in @xmath178 such that @xmath179 is odd .
hence by proposition [ map2 ] , there exists an element @xmath154 in @xmath180 such that @xmath157 is odd .
note that @xmath180 consists of nine elements @xmath181\cup [ 1\ 2\ 8\ 7],\ \mu'_{2}= [ 1\ 5\ 4\ 7]\cup [ 2\ 3\
9\ 8],\ \mu'_{3}= [ 2\ 8\ 5\ 4]\cup [ 3\ 1\ 7\ 9],\\ & & \mu'_{4}= [ 1\ 2\ 4\ 7]\cup [ 3\ 5\ 8\ 9],\ \mu'_{5}= [ 1\ 2\ 3]\cup [ 4\ 7\ 8\ 5],\ \mu'_{6}= [ 1\ 2\ 8\ 5]\cup [ 3\ 4\ 7\ 9],\\ & & \mu'_{7}= [ 2\ 3\ 4]\cup [ 1\ 5\ 8\ 7],\ \mu'_{8}= [ 7\ 8\ 9]\cup [ 1\ 2\ 4\ 5],\ \mu'_{9}= [ 1\ 5\ 3]\cup [ 2\ 8\ 7\ 4 ] .
\end{aligned}\ ] ] for @xmath182 , let @xmath183 be the subgraph of @xmath121 which is @xmath184 if @xmath185 and @xmath186 if @xmath187 . assume that @xmath188 is odd for some @xmath189 .
then it can be easily seen that @xmath183 contains a graph @xmath190 as a minor so that @xmath190 is isomorphic to @xmath142 and @xmath191 .
since both @xmath156 and @xmath188 are odd , by lemma [ d4type ] there exists an element @xmath8 in @xmath192 such that @xmath163 is odd .
next assume that @xmath193 is odd .
we denote two elements @xmath194\cup [ 1\ 2\ 6\ 5]$ ] and @xmath194\cup [ 4\ 2\ 6\ 5]$ ] in @xmath195 by @xmath196 and @xmath197 , respectively .
we denote the subgraph @xmath198 of @xmath199 by @xmath200 .
then it can be easily seen that @xmath201 contains a graph @xmath202 as a minor so that @xmath202 is isomorphic to @xmath142 and @xmath203 .
note that @xmath204 = [ 1\ 2\ 6\ 5 ] + [ 4\ 2\ 6\ 5]$ ] in @xmath205 , where @xmath206 denotes the homology group with @xmath207-coefficients .
then , by the homological property of the linking number , we have that @xmath208 thus we see that @xmath209 is odd or @xmath210 is odd . in either case , by lemma [ d4type ] there exists an element @xmath8 in @xmath211 such that @xmath163 is odd .
note that @xmath121 contains @xmath213 as the proper minor @xmath214 thus by lemma [ cgty ] , there exists an element @xmath177 in @xmath215 such that @xmath216 is odd .
hence by proposition [ map2 ] , there exists an element @xmath154 in @xmath217 such that @xmath157 is odd .
note that @xmath217 consists of seven elements , and by the symmetry of @xmath121 , we may assume that @xmath218\cup [ 2\ 6\ 9\ 3\ 4]$ ] or @xmath194\cup [ 1\ 5\ 3\ 4\ 2\ 6]$ ] without loss of generality .
we denote the subgraph @xmath219 of @xmath121 by @xmath220 .
assume that @xmath218\cup [ 2\ 6\ 9\ 3\ 4]$ ] .
we denote two elements @xmath221\cup [ 4\ 3\ 2]$ ] and @xmath221\cup [ 6\ 9\ 3\ 2]$ ] in @xmath222 by @xmath223 and @xmath224 , respectively .
we denote the subgraph @xmath186 of @xmath220 by @xmath225 .
then @xmath183 contains a graph @xmath190 as a minor so that @xmath190 is isomorphic to @xmath142 and @xmath226 . since @xmath227 = [ 4\ 3\ 2 ] + [ 6\ 9\ 3\ 2]$ ] in @xmath228
, it follows that @xmath229 this implies that @xmath230 is odd or @xmath231 is odd . in either case , by lemma [ d4type ] there exists an element @xmath8 in @xmath192 such that @xmath163 is odd .
next assume that @xmath232\cup [ 1\ 5\ 3\ 4\ 2\ 6]$ ] .
we denote four elements @xmath194\cup [ 3\ 4\ 5]$ ] , @xmath194\cup [ 4\ 5\ 6]$ ] , @xmath194\cup [ 1\ 5\ 6]$ ] and @xmath194\cup [ 2\ 4\ 6]$ ] in @xmath222 by @xmath233 and @xmath234 , respectively .
since @xmath235= [ 3\ 4\ 5 ] + [ 4\ 5\ 6 ] + [ 1\ 5\ 6 ] + [ 2\ 4\ 6]$ ] in @xmath228 , it follows that @xmath236 this implies that @xmath237 is odd for some @xmath238 or @xmath46 .
moreover , by the symmetry of @xmath220 , we may assume that @xmath239 is odd or @xmath240 is odd without loss of generality .
assume that @xmath239 is odd .
we denote the subgraph @xmath241 of @xmath121 by @xmath242 .
then @xmath242 contains a graph @xmath243 as a minor so that @xmath243 is isomorphic to @xmath142 and @xmath244 .
since both @xmath156 and @xmath230 are odd , by lemma [ d4type ] there exists an element @xmath8 in @xmath245 such that @xmath163 is odd .
next assume that @xmath240 is odd .
we denote four elements @xmath194\cup [ 1\ 2\ 6]$ ] , @xmath194\cup [ 1\ 2\ 3]$ ] , @xmath194\cup [ 2\ 3\ 4]$ ] and @xmath194\cup [ 1\ 3\ 5]$ ] in @xmath222 by @xmath246 and @xmath247 , respectively . since @xmath235=
[ 1\ 2\ 6 ] + [ 1\ 2\ 3 ] + [ 2\
3\ 4 ] + [ 1\ 3\ 5]$ ] in @xmath228 , we have that @xmath248 thus we see that @xmath237 is odd for some @xmath249 or @xmath250 .
moreover , by the symmetry of @xmath220 , we may assume that @xmath251 is odd or @xmath252 is odd without loss of generality .
assume that @xmath251 is odd .
we denote the subgraph @xmath253 of @xmath121 by @xmath254 .
then @xmath254 contains a graph @xmath255 as a minor so that @xmath255 is isomorphic to @xmath142 and @xmath256 .
since both @xmath156 and @xmath257 are odd , by lemma [ d4type ] there exists an element @xmath8 in @xmath258 such that @xmath163 is odd .
finally , assume that @xmath252 is odd .
let us consider the 3-component link @xmath259\cup [ 4\ 5\ 6]\cup [ 7\ 8\ 9])$ ] .
since all @xmath1-component sublinks of @xmath260 are @xmath261 and @xmath262 , each of the @xmath1-component sublinks of @xmath260 has an odd linking number .
next we show in the case of @xmath263 .
let @xmath4 be a spatial embedding of @xmath53 .
note that @xmath53 contains @xmath175 as the proper minor @xmath264 thus by lemma [ cgty ] , there exists an element @xmath169 in @xmath178 such that @xmath265 is odd .
hence by proposition [ map2 ] , there exists an element @xmath153 in @xmath266 such that @xmath156 is odd . note that @xmath266 consists of nine elements , and by the symmetry of @xmath53
, we may assume that @xmath267\cup [ 2\ 10\ 3\ 9\ 6]$ ] , @xmath268\cup [ 1\ 10\ 3\
9\ 6]$ ] , @xmath269\cup [ 1\ 6\ 2\ 4\ 7]$ ] , @xmath270\cup [ 1\ 10\ 2\ 6]$ ] or @xmath271\cup [ 1\ 6\ 9\ 3\ 4\ 7]$ ] without loss of generality .
note that @xmath53 contains @xmath213 as the proper minor @xmath272 thus by lemma [ cgty ] , there exists an element @xmath177 in @xmath215 such that @xmath273 is odd .
hence by proposition [ map2 ] , there exists an element @xmath154 in @xmath274 such that @xmath157 is odd .
note that @xmath274 consists of seven elements @xmath275\cup [ 1\ 6\ 2\ 4\ 7],\ \mu'_{2}= [ 1\ 7\ 8\ 10]\cup [ 2\ 4\ 3\ 9\ 6],\\ & & \mu'_{3}= [ 1\ 10\ 2\ 6]\cup
[ 3\ 4\ 7\ 8\ 9],\ \mu'_{4}= [ 2\ 4\ 3\ 10]\cup [ 1\ 7\ 8\ 9\ 6],\\ & & \mu'_{5}= [ 2\ 4\ 7\ 8]\cup [ 1\ 10\ 3\
9\ 6],\ \mu'_{6}= [ 2\ 8\ 9\ 6]\cup [ 1\ 10\ 3\ 4\ 7],\\ & & \mu'_{7}= [ 2\ 8\ 10]\cup [ 1\ 6\ 9\ 3\ 4\ 7 ] .
\end{aligned}\ ] ] for @xmath276 , let @xmath183 be the subgraph of @xmath53 which is @xmath277 if @xmath278 and @xmath186 if @xmath279 .
assume that @xmath188 is odd for some @xmath280 .
then @xmath183 contains a graph @xmath190 as a minor so that @xmath190 is isomorphic to @xmath142 and @xmath191 .
since both @xmath156 and @xmath188 are odd , by lemma [ d4type ] there exists an element @xmath8 in @xmath192 such that @xmath163 is odd .
note that @xmath53 contains another @xmath213 as the proper minor @xmath282 thus by lemma [ cgty ] , there exists an element @xmath177 in @xmath215 such that @xmath273 is odd .
hence by proposition [ map2 ] , there exists an element @xmath154 in @xmath274 such that @xmath157 is odd .
note that @xmath274 consists of seven elements @xmath283\cup [ 2\ 4\ 5\
3\ 10],\ \mu'_{2}= [ 1\ 7\ 4\ 5]\cup [ 2\ 10\ 3\ 9\ 6],\\ & & \mu'_{3}= [ 3\ 5\ 6\ 9]\cup [ 1\ 10\ 2\
4\ 7],\ \mu'_{4}= [ 1\ 5\ 3\ 10]\cup [ 2\ 4\ 7\ 9\ 6],\\ & & \mu'_{5}= [ 1\ 10\ 2\ 6]\cup [ 3\ 9\ 7\ 4\ 5],\ \mu'_{6}= [ 1\ 5\ 6]\cup [ 2\ 4\ 7\ 9\ 3\ 10],\\ & & \mu'_{7}= [ 2\ 4\ 5\ 6]\cup [ 1\ 10\ 3\ 9\ 7 ] .
\end{aligned}\ ] ] for @xmath276 , let @xmath183 be the subgraph of @xmath53 which is @xmath284 if @xmath285 and @xmath186 if @xmath286 .
assume that @xmath188 is odd for some @xmath280 .
then @xmath183 contains a graph @xmath190 as a minor so that @xmath190 is isomorphic to @xmath142 and @xmath191 .
since both @xmath156 and @xmath188 are odd , by lemma [ d4type ] there exists an element @xmath8 in @xmath192 such that @xmath163 is odd .
let @xmath213 be the proper minor of @xmath53 and @xmath288 the element in @xmath289 as in case 2 .
for @xmath276 , let @xmath183 be the subgraph of @xmath53 which is @xmath290 if @xmath291 and @xmath186 if @xmath292 .
assume that @xmath188 is odd for some @xmath280 .
then @xmath183 contains a graph @xmath190 as a minor so that @xmath190 is isomorphic to @xmath142 and @xmath191 .
since both @xmath156 and @xmath188 are odd , by lemma [ d4type ] there exists an element @xmath8 in @xmath192 such that @xmath163 is odd .
note that @xmath53 contains another @xmath175 as the proper minor @xmath294 thus by lemma [ cgty ] , there exists an element @xmath177 in @xmath178 such that @xmath295 is odd .
hence by proposition [ map2 ] , there exists an element @xmath154 in @xmath266 such that @xmath157 is odd .
note that @xmath266 consists of nine elements @xmath296\cup [ 1\ 10\ 2\ 4\ 7],\ \mu'_{2}= [ 3\ 10\ 8\ 9]\cup [ 1\ 6\ 2\ 4\ 7],\\ & & \mu'_{3}= [ 1\ 5\ 8\ 10]\cup [ 2\ 4\ 7\ 9\ 6],\ \mu'_{4}= [ 7\ 8\ 9]\cup [ 1\ 10\ 2\ 6],\\ & & \mu'_{5}= [ 2\ 8\ 10]\cup
[ 1\ 6\ 9\ 7],\ \mu'_{6}= [ 2\ 8\ 5\ 6]\cup [ 1\ 10\ 3\ 9\ 7],\\ & & \mu'_{7}= [ 1\ 7\ 8\ 5]\cup [ 2\ 10\ 3\ 9\ 6],\ \mu'_{8}= [ 1\ 5\ 6]\cup [ 2\ 4\ 7\ 9\ 3\ 10],\\ & & \mu'_{9}= [ 2\ 4\ 7\ 8]\cup [ 1\ 10\ 3\ 9\ 6 ] .
\end{aligned}\ ] ] for @xmath182 , let @xmath183 be the subgraph of @xmath53 which is @xmath297 if @xmath298 and @xmath186 if @xmath299
. assume that @xmath188 is odd for some @xmath300 .
then @xmath183 contains a graph @xmath190 as a minor so that @xmath190 is isomorphic to @xmath142 and @xmath191 .
since both @xmath156 and @xmath188 are odd , by lemma [ d4type ] there exists an element @xmath8 in @xmath192 such that @xmath163 is odd .
next assume that @xmath193 is odd .
we denote two elements @xmath301\cup [ 2\ 4\ 3\ 10]$ ] and @xmath301\cup [ 3\ 4\ 7\ 9]$ ] in @xmath195 by @xmath196 and @xmath197 , respectively .
we denote the subgraph @xmath302 of @xmath199 by @xmath303 and the subgraph @xmath304 of @xmath53 by @xmath305 .
then @xmath201 contains a graph @xmath202 as a minor so that @xmath202 is isomorphic to @xmath142 and @xmath203 . since @xmath306 = [ 2\ 4\ 3\ 10 ] + [ 3\ 4\ 7\ 9]$ ] in @xmath205
, it follows that @xmath208 this implies that @xmath209 is odd or @xmath210 is odd . in either case , by lemma [ d4type ] there exists an element @xmath8 in @xmath211 such that @xmath163 is odd .
finally assume that @xmath307 is odd .
note that @xmath53 contains another @xmath213 as the proper minor @xmath308 thus by lemma [ cgty ] , there exists an element @xmath309 in @xmath215 such that @xmath310 is odd .
hence by proposition [ map2 ] , there exists an element @xmath311 in @xmath274 such that @xmath312 is odd .
note that @xmath274 consists of seven elements @xmath313\cup [ 1\ 10\ 3\ 4\ 7],\ \mu''_{2}= [ 4\ 5\ 8\ 7]\cup [ 1\ 10\ 3\ 9\ 6],\\ & & \mu''_{3}= [ 1\ 7\ 8\ 10]\cup [ 3\ 4\ 5\ 6\ 9],\ \mu''_{4}= [ 3\ 10\ 8\ 9]\cup [ 1\ 7\ 4\ 5\ 6],\\ & & \mu''_{5}= [ 1\ 6\ 9\ 7]\cup [ 3\ 4\ 5\ 8\ 10],\ \mu''_{6}= [ 3\ 9\ 7\ 4]\cup [ 1\ 10\ 8\ 5\ 6],\\ & & \mu''_{7}= [ 7\ 8\ 9]\cup [ 1\ 10\ 3\ 4\ 5\ 6 ] .
\end{aligned}\ ] ] for @xmath314 , let @xmath315 be the subgraph of @xmath53 which is @xmath316 if @xmath317 and @xmath318 if @xmath319 .
assume that @xmath320 is odd for some @xmath321 .
then @xmath315 contains a graph @xmath322 as a minor so that @xmath322 is isomorphic to @xmath142 and @xmath323 .
since both @xmath307 and @xmath320 are odd , by lemma [ d4type ] there exists an element @xmath8 in @xmath324 such that @xmath163 is odd .
next assume that @xmath325 is odd .
we denote three elements @xmath194\cup [ 1\ 5\ 3\ 10]$ ] , @xmath194\cup [ 1\ 5\ 6]$ ] and @xmath194\cup [ 3\ 4\ 5]$ ] in @xmath326 by @xmath327 , @xmath328 and @xmath329 , respectively .
we denote the subgraph @xmath330 of @xmath53 by @xmath331 .
then @xmath332 contains a graph @xmath333 as a minor so that @xmath333 is isomorphic to @xmath142 and @xmath334 . since @xmath335 = [ 1\ 5\ 3\ 10 ] + [ 1\ 5\ 6 ] + [ 3\ 4\ 5]$ ] in @xmath336
, it follows that @xmath337 this implies that @xmath338 is odd for some @xmath164 .
if @xmath339 is odd or @xmath340 is odd , then by lemma [ d4type ] there exists an element @xmath8 in @xmath341 such that @xmath163 is odd .
if @xmath342 is odd , let us consider the 3-component link @xmath343\cup [ 7\ 8\ 9]\cup [ 1\ 10\ 2\ 6])$ ] . since all @xmath1-component sublinks of @xmath260 are @xmath344 and @xmath345 , each of the @xmath1-component sublinks of @xmath260 has an odd linking number .
we denote two elements @xmath271\cup [ 1\ 6\ 9\ 7]$ ] and @xmath271\cup [ 3\ 9\ 7\ 4]$ ] in @xmath326 by @xmath347 and @xmath348 , respectively .
since @xmath349 = [ 1\ 6\ 9\ 7 ] + [ 3\ 9\ 7\ 4]$ ] in @xmath336 , it follows that @xmath350 this implies that @xmath351 is odd or @xmath352 is odd . by the symmetry of @xmath53
, we may assume that @xmath351 is odd .
note that @xmath53 contains another @xmath175 as the proper minor @xmath353 thus by lemma [ cgty ] , there exists an element @xmath177 in @xmath178 such that @xmath295 is odd .
hence by proposition [ map2 ] , there exists an element @xmath154 in @xmath266 such that @xmath157 is odd .
note that @xmath266 consists of nine elements @xmath354\cup [ 1\ 6\ 2\ 4\ 7],\ \mu'_{2}= [ 1\ 7\ 8\ 5]\cup [ 2\ 4\ 3\ 9\ 6],\\ & & \mu'_{3}= [ 1\ 5\ 6]\cup [ 3\ 9\ 7\ 4],\ \mu'_{4}= [ 3\ 4\ 5]\cup [
1\ 6\ 9\ 7],\\ & & \mu'_{5}= [ 5\ 6\ 9\ 8]\cup [ 1\ 10\ 3\ 4\ 7],\ \mu'_{6}= [ 4\ 5\ 8\ 7]\cup [ 1\ 10\ 3\ 9\ 6],\\ & & \mu'_{7}= [ 1\ 5\ 3\ 10]\cup [ 2\ 4\ 7\ 9\ 6],\ \mu'_{8}= [ 2\ 4\ 5\
6]\cup [ 1\ 10\ 3\ 9\ 7],\\ & & \mu'_{9}= [ 7\ 8\ 9]\cup [ 1\ 10\ 3\ 4\ 2\ 6 ] .
\end{aligned}\ ] ] for @xmath182 , let @xmath183 be the subgraph of @xmath53 which is @xmath355 if @xmath356 and @xmath357 if @xmath358 .
assume that @xmath188 is odd for some @xmath359 .
then @xmath183 contains a graph @xmath190 as a minor so that @xmath190 is isomorphic to @xmath142 and @xmath360 .
since both @xmath351 and @xmath188 are odd , by lemma [ d4type ] there exists an element @xmath8 in @xmath192 such that @xmath163 is odd .
next assume that @xmath361 is odd .
we denote two elements @xmath194\cup [ 1\ 6\ 2\ 10]$ ] and @xmath194\cup [ 2\ 4\ 3\ 10]$ ] in @xmath362 by @xmath363 and @xmath364 , respectively .
we denote the subgraph @xmath365 of @xmath366 by @xmath367 and the subgraph @xmath368 of @xmath53 by @xmath369 .
then @xmath370 contains a graph @xmath371 as a minor so that @xmath371 is isomorphic to @xmath142 and @xmath372 . since @xmath373 = [ 1\ 6\ 2\ 10 ] + [ 2\ 4\ 3\ 10]$ ] in @xmath374
, it follows that @xmath375 this implies that @xmath376 is odd or @xmath377 is odd . in either case , by lemma [ d4type ] there exists an element @xmath8 in @xmath378 such that @xmath163 is odd .
finally assume that @xmath307 is odd .
note that @xmath53 contains another @xmath213 as the proper minor @xmath379 thus by lemma [ cgty ] , there exists an element @xmath309 in @xmath215 such that @xmath310 is odd . hence by proposition [ map2
] , there exists an element @xmath311 in @xmath274 such that @xmath312 is odd .
note that @xmath274 consists of seven elements @xmath380\cup [ 1\ 10\ 2\ 4\ 7],\ \mu''_{2}= [ 3\ 9\ 7\ 4]\cup [ 1\ 5\ 8\ 2\ 10],\\ & & \mu''_{3}= [ 1\ 7\ 4\ 5]\cup [ 2\ 8\ 9\ 3\ 10],\ \mu''_{4}= [ 2\ 4\ 5\ 8]\cup [ 1\ 10\ 3\ 9\ 7],\\ & & \mu''_{5}= [ 2\ 4\ 3\ 10]\cup [ 1\ 5\ 8\ 9\ 7],\ \mu''_{6}= [ 1\ 5\ 3\ 10]\cup [ 2\ 4\ 7\ 9\ 8],\\ & & \mu''_{7}= [ 3\ 4\ 5]\cup [ 1\ 10\ 2\ 8\ 9\ 7 ] .
\end{aligned}\ ] ] for @xmath314 , let @xmath315 be the subgraph of @xmath53 which is @xmath381 if @xmath382 and @xmath318 if @xmath383 .
assume that @xmath320 is odd for some @xmath321 .
then @xmath315 contains a graph @xmath322 as a minor so that @xmath322 is isomorphic to @xmath142 and @xmath323 .
since both @xmath307 and @xmath320 are odd , by lemma [ d4type ] there exists an element @xmath8 in @xmath324 such that @xmath163 is odd .
next assume that @xmath325 is odd .
we denote two elements @xmath270\cup [ 1\ 10\ 8\ 9\ 7]$ ] and @xmath270\cup [ 2\ 8\ 10]$ ] in @xmath326 by @xmath327 and @xmath328 , respectively .
we denote the subgraph @xmath384 of @xmath53 by @xmath385 .
then @xmath385 contains a graph @xmath386 as a minor so that @xmath386 is isomorphic to @xmath142 and @xmath387 . since @xmath388 = [ 1\ 10\ 8\ 9\ 7 ] + [ 2\ 8\ 10]$ ] in @xmath336
, it follows that @xmath389 this implies that @xmath339 is odd or @xmath340 is odd .
if @xmath339 is odd , then by lemma [ d4type ] there exists an element @xmath8 in @xmath390 such that @xmath163 is odd .
if @xmath340 is odd , let us consider the 3-component link @xmath343\cup [ 2\ 8\ 10]\cup [ 1\ 6\ 9\ 7])$ ] . since all @xmath1-component sublinks of @xmath260 are @xmath391 and @xmath392 , each of the @xmath1-component sublinks of @xmath260 has an odd linking number .
this completes the proof .
note that a graph in the heawood family is obtained from one of @xmath43 , @xmath121 and @xmath53 by a finite sequence of @xmath2-exchanges .
thus by lemma [ dy ] , theorem [ n9fn ] and the fact that @xmath43 is ik , namely i(k or c3l ) , it follows that every graph in the heawood family is i(k or c3l ) . on the other hand
, a graph in the heawood family is obtained from one of @xmath393 and @xmath48 by a finite sequence of @xmath3-exchanges . since
each of @xmath393 and @xmath48 is a minor - minimal ik graph and @xmath394 and @xmath395 are the empty sets , it follows that @xmath393 and @xmath48 are minor - minimal i(k or c3l ) graphs .
thus by lemma [ minor - min - lem ] , we have the desired conslusion . [ apex ]
a graph is said to be _
@xmath1-apex _ if it can be embedded in the @xmath1-sphere after the deletion of at most two vertices and all of their incidental edges .
it is not hard to see that any @xmath1-apex graph may have a spatial embedding whose image contains neither a nontrivial knot nor a @xmath0-component link any of whose @xmath1-component sublink is nonsplittable .
thus any @xmath1-apex graph is not i(k or c3l ) .
it is known that every graph of at most twenty edges is 2-apex @xcite ( see also @xcite ) . since the number of all edges of every graph in the heawood family is twenty one ,
we see that any proper minor of a graph in the heawood family is @xmath1-apex , namely not i(k or c3l ) .
this also implies that any graph in the heawood family is minor - minimal for i(k or c3l ) .
[ not3l ] let @xmath396 be the spatial embedding of @xmath121 and @xmath397 the spatial embedding of @xmath53 as illustrated in fig .
[ no_3links ]
. then it can be checked directly that both @xmath398 and @xmath399 do not contain a nonsplittable @xmath0-component link .
thus neither @xmath121 nor @xmath53 is i3l .
moreover , we can see that @xmath400 and @xmath56 are not i3l in a similar way as the proof of lemma [ main1lem ] , see fig .
[ no_3links ] . [ hea ]
we remark that the heawood graph is ik .
the heawood graph is the dual graph of @xmath43 which is embedded in a torus .
it is known that there exists a unique graph @xmath48 obtained from @xmath43 by seven times applications of @xmath2-exchanges @xcite .
the seven triangles corresponds to the black triangles of black - and - white coloring of the torus by @xmath43 .
then @xmath48 and @xmath27 are mapped to each other by parallel transformation of the torus , see fig .
[ torus ] .
thus they are isomorphic . since @xmath48 is ik
, we have the result . [ k3311 ]
it is known that all twenty six graphs obtained from the complete four - partite graph @xmath45 by a finite sequence of @xmath2-exchanges are minor - minimal ik graphs @xcite , @xcite .
there exist thirty two graphs which are obtained from @xmath45 by a finite sequence of @xmath2 and @xmath3-exchanges but can not be obtained from @xmath45 by a finite sequence of @xmath2-exchanges .
recently , goldberg - mattman - naimi announces that these thirty two graphs are also minor - minimal ik graphs @xcite .
h. sachs , on spatial representations of finite graphs , _ finite and infinite sets , vol .
i , ii ( eger , 1981 ) , _ 649662 , colloq .
janos bolyai , * 37 * , _ north - holland , amsterdam , _ 1984 . k. taniyama and a. yasuhara , realization of knots and links in a spatial graph , _ topology appl . _
* 112 * ( 2001 ) , 87109 . | we say that a graph is intrinsically knotted or completely @xmath0-linked if every embedding of the graph into the @xmath0-sphere contains a nontrivial knot or a @xmath0-component link any of whose @xmath1-component sublink is nonsplittable .
we show that a graph obtained from the complete graph on seven vertices by a finite sequence of @xmath2-exchanges and @xmath3-exchanges is a minor - minimal intrinsically knotted or completely 3-linked graph . |
measurements of the hubble constant are a unique data set for statistical analysis for two reasons .
first , huchra s compilationhuchra/. ] with over 400 measurements is one of the largest collection of measurements of a single quantity .
second , the hubble constant is now one of the more precisely determined cosmological parameters ( see , e.g. , freedman et al . 2001 ; bennett et al .
2003 ) .
it is also of great interest to understand how well the hubble constant has been measured , both because it is an important cosmological parameter and because of the role it plays in various cosmological tests , most importantly the expansion time test ( see , e.g. , peebles & ratra 2003 for a review )
. assuming a value for the hubble constant in the body of this paper we work with @xmath0 = 71 km s@xmath1 mpc@xmath1 , the central value from the combined wmap and other data analysis of spergel et al .
( 2003 ) = 67 km s@xmath1 mpc@xmath1 from the gott et al .
( 2001 ) median statistics analysis of a major , earlier subset of the @xmath0 measurements considered here , showing that our conclusions are not sensitive to the precise central value of @xmath0 assumed in the estimated 10 % ( two standard deviation ) range now under discussion ( see , e.g. , gott et al .
one may use huchra s compilation of @xmath5 ( where @xmath6 and @xmath7 are the upper and lower one standard deviation error bars ) to construct the distribution of errors of the hubble constant measurements .
this is a plot of the number of measurements as a function of the number of standard deviations ( @xmath8 ) the measurement deviates from the actual value @xmath0 .
here @xmath9 when @xmath10 and @xmath11 when @xmath12 . in our analysis here we use measurements from huchra s compilation up to and including measurement 2003.239 . deleting the four entries from 1924 and 1925 that lack actual estimates of @xmath0 , we use
461 published estimates of @xmath0 in our analysis here , 40 % more than the 331 used in the analysis of gott et al .
observers often note that there could be unknown systematic errors , however authors quoted errors have been used to evaluate the accuracy of @xmath0 estimates and so it is important to understand the quoted error distribution . in @xmath13 2
we describe our analysis of this collection of 461 measurements , assuming that @xmath0 = 71 km s@xmath1 mpc@xmath1 , the central value from the combined wmap and other data analysis of bennett et al .
( 2003 ) and spergel et al .
( 2003 ) . for comparison , to show that the results are robust to small changes in the true value of @xmath0 , summary results from an analysis based on the central value of @xmath0 = 67 km s@xmath1 mpc@xmath1 from the gott et al .
( 2001 ) median statistics study are presented in the appendix .
we conclude in @xmath13 3 .
figure 1 shows the distribution of deviations of the 461 measurements from the central wmap value of @xmath0 = 71 km s@xmath1 mpc@xmath1 , in units of the quoted standard deviation of the measurement .
this is the error distribution of the @xmath0 measurements ; the left panel shows the signed error distribution and the right panel shows the absolute magnitude of the errors ( the distribution in the right panel is symmetric about @xmath14 ) .
these error distributions have significant tails : there are numerous measurements 5 and even 10 standard deviations away .
more precisely , in the signed error distribution of fig .
1@xmath15 68.3 % and 95.4 % of the probability lies in the range @xmath16 and @xmath17 , respectively , and for the absolute magnitude error distribution of fig .
1@xmath18 the corresponding limits are @xmath19 and @xmath20 , respectively .
an alternative characterization of the tails of this distribution is provided by the fraction of data within the @xmath21 = 1 and 2 ranges , which for the distribution shown in fig .
1@xmath18 is 48 % and 69 % , respectively .
these are impressively high ( nearly half the observed values are within one standard deviation of 71 km s@xmath1 mpc@xmath1 ) but still clearly at odds with what is expected for a gaussian distribution .
it is of interest to quantify how well the data of fig .
1 are fit by various simple distribution functions , and to determine the parameters of these functions that result in the best fit to the data . to do this
we proceed as follows . for our purposes
it is useful to maximize the number of data points in each bin as well as the number of bins .
this is perhaps best accomplished by using 21 bins ( close to the square root of 461 ) , labelled by integer @xmath22 that runs from 1 to 21 , and adjusting the widths of the bins , @xmath23 , to ensure equal expected probability ( for the assumed distribution function ) in each bin .
thus , for an assumed distribution ( such as a gaussian ) we construct 21 bins such that the expected number of data points in each bin would be 21.95 .
then we compare with the number of data points observed in each of the 21 bins and do a @xmath24 analysis , as discussed in the next paragraph .
( since the number expected in each bin is large compared to unity , a @xmath24 analysis is justified . ) with this prescription , the data binning depends on the assumed probability distribution function , @xmath25 ( in this paper we present results only from the fit to the symmetric absolute error distribution , e.g. , that in fig .
1@xmath18 ) . to estimate goodness of fit we use the assumed probability distribution function to compute the expected number of measurements in each bin @xmath22 , @xmath26 , where @xmath27 is the total number of measurements .
since there are a finite number of measurements in each bin , they should be poisson distributed with mean value @xmath26 for the @xmath28 bin . for the poisson distribution
the variance @xmath29 is equal to the mean hence the total @xmath24 is @xmath30 ^ 2 } \over n p(|n_\sigma|_j ) } , \ ] ] where @xmath31 is the observed number of measurements in each bin .
we shall tabulate the reduced @xmath24 , @xmath32 , where @xmath33 is the number of degrees of freedom , i.e. , the number of bins ( 21 ) less the number of constraints and fitting parameters .
given @xmath34 and @xmath33 one may compute the probability that the assumed distribution well describes the spread of the measurements . in the computation of this probability
we assume that the bins are uncorrelated , which is not necessarily true ( since lower rungs of the distance ladder introduce correlations in subsets of the measurements ) .
it is therefore wise to place quantitative emphasis on just the @xmath34 values and use the corresponding probabilities as simply a qualitative indicator of goodness of fit .
we consider four probability distribution functions and as mentioned above focus on the absolute magnitude error distribution , as in fig .
1@xmath18 , so all distributions we consider will be centered at @xmath14 .
one constraint that must be satisfied is that the total number of measurements must sum to 461 .
since we consider 21 bins and normalize to fit the total number of measurements , a probability distribution function with no free parameters will have @xmath35 degrees of freedom . even though we have noted the existence of extended tails in the error distributions of fig .
1 , it is natural perhaps pavlovian to first consider the gaussian distribution , initially with width chosen so that @xmath36 corresponds to one standard deviation , and then with a scale factor to vary the width of the distribution .
that is , we take as probability distribution function the gaussian expression @xmath37 , \ ] ] for the case where @xmath36 is equivalent to one standard deviation , and then consider the function @xmath38 where @xmath39 is a scale factor that is adjusted to minimize @xmath24 .
( we allow @xmath39 to vary over the range 0.5 to 3 in steps of 0.1 when computing @xmath24 . ) in the first case there are no additional free parameters so @xmath35 ; the scale factor @xmath39 is an additional free parameter in the second case so here we have @xmath40 degrees of freedom .
figure 2 shows the measurement error histograms and the best - fit gaussians , both normalized to unit area .
numerical values are listed in table 1 .
these show that if @xmath0 = 71 km s@xmath1 mpc@xmath1 then the measurement error distributions are extremely poorly fit by a gaussian , even if the gaussian width is allowed to be a free parameter . interestingly , if the width is allowed to float while minimizing @xmath24 it favors 1.8 , i.e. , the assumed distribution favors identifying @xmath41 with one standard deviation , almost double the value one would naively infer from the measurement errors , thus perhaps indicating that in this gaussian case it might not be unreasonable to roughly double the quoted error bars , consistent with our earlier discussion of extended tails .
this situation might profitably be contrasted with what happened in the early days of cosmic microwave background spatial anisotropy measurements , where a number of models fit the measurements extremely well , perhaps indicating that the error bars had been over estimated ( ganga , ratra , & sugiyama 1996 ) . in any case , it is very unlikely that the @xmath0 measurement errors are described by a gaussian distribution .
( note that the probability is a little higher for the @xmath0 = 67 km s@xmath1 mpc@xmath1 case , but even here a gaussian distribution is a very poor fit . )
the fact that the error distribution of hubble constant measurements is non - gaussian does not necessarily imply an underlying non - gaussianity in the measurement errors .
rather , the distribution tells us something about the observers ability to correctly estimate systematic and statistical uncertainties .
figure 2 indicates that the distribution of hubble constant measurement errors has a more extended tail than is predicted by a gaussian probability distribution .
perhaps the most well - known distribution with an extended tail is the cauchy , or lorentzian , or breit - wigner distribution , @xmath42 we also consider the case @xmath38 where the scale factor @xmath39 is allowed to vary while @xmath24 is minimized .
figure 3 shows the data and best - fit cauchy distributions , and numerical values are listed in table 1 .
unlike the gaussian case , the cauchy distribution can not be rejected ; it is acceptable at 9.9 % or 8.7 % depending on whether @xmath39 is fixed to unity or allowed to vary ( and it does significantly better at @xmath0 = 67 km s@xmath1 mpc@xmath1 ) .
however , it is clear from fig .
3 that the cauchy distribution has greater probability in the extended tails than does the hubble constant measurements error distribution .
the cauchy distribution has a similar central peak , with a 50 % chance that @xmath43 , but a 95.4 % chance that @xmath44 instead of 7.0 as observed . it would therefore be beneficial to search for a distribution that has broader tails than the gaussian one but narrower than the cauchy case .
a cauchy distribution with @xmath45 would result if the errors were gaussian distributed , observers took measurements free of systematic errors , divide their data into two parts , used each half to produce two independent estimates of the hubble constant , @xmath46 and @xmath47 , and produced a mean estimate @xmath48 with an error estimate ( standard deviation of the mean ) of @xmath49 .
if @xmath46 and @xmath47 are drawn from an underlying gaussian distribution centered on the true value @xmath50 then @xmath51 is distributed like a cauchy distribution with @xmath52 .
that gives a 50 % chance that @xmath53 is within 1 @xmath54 of the true value and a 95 % chance that @xmath53 is within 12.7 @xmath54 of the true value .
the large tails result because in a gaussian distribution there is an appreciable chance that @xmath55 will be significantly less than the true sigma for the distribution . in this scenario the observer is really using the self - consistency of her observations to set the error bars .
if one measures the distance to two galaxies using cepheids , and gets two values for the hubble constant that are close to each other , one may well be tempted to think that one s method has the high degree of accuracy implied by the observed value of @xmath49 .
indeed , if one estimated the errors by other means ( estimated uncertainties in measuring the observed quantities required to measure the hubble constant , along with standard propagation of errors ) and one got an error significantly larger than @xmath55 then one might be suspicious that one should be so lucky as to obtain such a small value of @xmath55 . yet rarely , such lucky coincidences do occur and it is precisely these cases that cause the large tails in the cauchy distribution .
the cauchy distribution with @xmath45 is acceptable at 9.9 % , but not a good fit and there are other distributions that are better fits .
as we have noted , a cauchy distribution with @xmath45 would result from a true gaussian distribution if the observer divided his data into two parts , used the data itself to set error bars , and made the mistake of assuming the errors should be distributed according to a gaussian distribution rather than the student @xmath4 distribution ( which for the case of two data points is the @xmath56 student @xmath4 distribution , or the cauchy distribution ) .
this prompts us to investigate the general student @xmath4 distributions .
student s @xmath4 distribution is @xmath57 \over\sqrt{\pi n}\ , \gamma(n/2 ) } { 1 \over \left(1 + |n_\sigma|^2/n\right)^{(n+1)/2 } } , \ ] ] where @xmath58 is positive and @xmath59 the gamma function .
we also consider the distribution @xmath60 where the scale factor @xmath39 is allowed to vary while @xmath24 is minimized . when @xmath61 student s @xmath4 distribution becomes the gaussian distribution and for @xmath56 it is the cauchy distribution . for @xmath62 student
s @xmath4 distribution has narrower tails than the cauchy case but broader ones than the gaussian distribution , just as wanted .
we have fit student s @xmath4 distribution to the @xmath0 measurement errors data while allowing @xmath58 to take on integer values between 2 and 6 ( and sometimes going up to 30 ) , so in this case we have one additional parameter and hence one less degree of freedom .
we find @xmath63 always minimizes the value of @xmath24 and so show this case in fig . 4 and table 1 . from table 1
we see that if the scale factor @xmath39 is held at unity student s @xmath4 distribution is an unlikely fit to the data , especially if @xmath0 = 71 km s@xmath1 mpc@xmath1 .
however , if @xmath39 is allowed to vary as @xmath24 is minimized , student s @xmath4 distribution with @xmath63 is an excellent fit to the @xmath0 measurements error distribution , and fig .
4@xmath18 shows that there is very good agreement between the expected and measured counts in the last bin .
the final probability density distribution we consider is the double exponential or laplace distribution , @xmath64 this falls off less rapidly than the gaussian distribution but faster than the cauchy distribution .
the sample median is the best estimator for the mean of this distribution ( eadie et al .
the results of the fit are shown in fig . 5 and listed in table 1 . as in the case for student s @xmath4 distribution with @xmath63 ,
when @xmath39 is held fixed at unity the double exponential is an unacceptable fit to the @xmath0 measurements error distribution but when @xmath39 is allowed to vary it is an excellent fit to the data . in the first paragraph of this section we noted that in fig . 1@xmath18 68.3 % and 95.4 % of the probability lies in the range @xmath19 and @xmath20 , respectively , and the @xmath65 and @xmath66
ranges include 48 % and 69 % of the data points , respectively .
( see the appendix for the corresponding numbers for the @xmath0 = 67 km s@xmath1 mpc@xmath1 case . )
tables 2 and 3 show the related limits for the various probability density distributions we consider in this paper .
these numerical values provide another indication of the non - gaussianity of the hubble constant measurement error distribution .
our analysis of a perhaps unique ( because of its size ) data set , the measurement errors of all available estimates of the hubble constant , makes for some interesting conclusions . if all observers have done perfect jobs at estimating their errors and the true errors were gaussian , as might be expected , then the distributions in fig
. 1 should be gaussian with standard deviation of unity .
first , and perhaps not totally unexpectedly , the errors in the hubble constant are not gaussianly distributed , even if the scale factor @xmath39 is allowed to vary when minimizing @xmath24 . at the minimum value of @xmath24 , @xmath67 , suggesting it might be reasonable to roughly double the magnitude of @xmath0 measurement error bars .
early observers using inferior equipment or techniques would have larger errors , but knowing that their methods were uncertain should have established larger error bars .
as methods improved the measurements become more accurate but the stated error bars become smaller .
early or late observers are at no relative disadvantage relative to others .
indeed , each observer has freedom to state her error bars and has a priori an equal chance of having the true value occur within one standard deviation of their result . over - optimism
would produce error bars that were too small while over - conservatism would produce error bars that were too large . which occurs in practice ?
the results here suggest that astronomers were over - optimistic by almost a factor of 2 .
in some case there were systematic errors of which the observers were simply unaware ( such as mistaking hii regions for bright stars ) .
in other cases , standard candles were not as standard as imagined , leaving some steps in the distance ladder wrong by more than people thought .
also , using self consistency in the data as a check on the errors can lead to large tails because it occasionally induces one to be over - optimistic ( the student @xmath4 effect ) . and the real data may have non - gaussian tails ( say in the luminosity of standard candles ) . in general over - conservatism ( the urge to be right ) always competes with over - optimism ( the urge to have more interesting limits ) . in the case of the hubble constant astronomers
were over - optimistic . in a history - of - science context , it might be of interest to more closely examine the most deviant measurements of fig . 1 , those that have @xmath21 larger than say 7 , to understand why these are so deviant , but this is not our purpose here .
the hubble constant measurement history suggests that to be really sure ( 95.4 % ) you have to go to 7 @xmath2 .
this may explain why some people are cautious upon hearing of a three standard deviation result .
it s not that they believe the errors but want to be more sure than 99.7 % .
it s that they suspect there is a large chance ( @xmath68 50 % ) that the error bars may have been underestimated by a factor of 2 or 3 and the chance it is really correct is consequently really significantly less than 99.7 % .
second , an @xmath63 student s @xmath4 distribution , with @xmath69 , or a double exponential distribution , with @xmath70 , are excellent fits to the @xmath0 measurement errors distribution , with @xmath0 = 67 km s@xmath1 mpc@xmath1 having a somewhat higher probability than @xmath0 = 71 km s@xmath1 mpc@xmath1 .
the hubble constant measurement history gives an interesting example where we can access how trustworthy quoted errors might be in fundamental measurements .
it would be interesting to study comparative examples from other fields .
in particular , it would be interesting to know whether the @xmath63 student @xmath4 distribution or the double exponential distribution also provides a good description of the measurement errors of other quantities .
we are grateful to j. huchra for the compilation of @xmath0 measurements and acknowledge useful discussions with a. kosowsky .
gc and br acknowledge support from nsf career grant ast-9875031 and doe epscor grant de - fg02 - 00er45824 .
jrg acknowledges support from nsf grant ast-9900772 .
in the main body of the paper we assumed @xmath0 = 71 km s@xmath1 mpc@xmath1 , the central value from the wmap analysis . the wmap @xmath0 error bars are @xmath71 km s@xmath1 mpc@xmath1 ( bennett et al .
@xmath0 is pinned down to only about 10 % at two standard deviations ( gott et al .
2001 ) so it is reasonable to find out how our conclusions depend on the value of @xmath0 .
in this appendix we use @xmath0 = 67 km s@xmath1 mpc@xmath1 , the central value from the gott et al .
( 2001 ) median statistics analysis of a subset ( 331 measurements prior to mid 1999 ) of the 461 measurements used here .
= 68 km s@xmath1 mpc@xmath1 ( using all 461 measurements ) ; the small shift in the median after a 40 % increase in the number of measurements considered is great tribute to its robustness .
see gott et al .
( 2001 ) , podariu et al .
( 2001 ) , avelino , martins , & pinto ( 2002 ) , and chen & ratra ( 2003 ) for other cosmological applications of median statistics . ]
figure 6 shows the @xmath0 measurements error distribution for the case when @xmath0 = 67 km s@xmath1 mpc@xmath1 .
this distribution has a somewhat less prominent central peak than the @xmath0 = 71 km s@xmath1 mpc@xmath1 case . in the signed error distribution of fig .
6@xmath15 68.3 % and 95.4 % of the probability lies in the range @xmath72 and @xmath73 , respectively , while for the absolute magnitude error distribution of fig .
6@xmath18 the corresponding limits are @xmath74 and @xmath75 , respectively . in fig .
6@xmath18 the @xmath76 and @xmath66 ranges include 51 % and 72 % of the data points , respectively .
again , these are at odds with what is expected for a gaussian distribution .
table 1 also lists the numerical fitting results for the @xmath0 = 67 km s@xmath1 mpc@xmath1 case . as in the case when @xmath0 = 71 km s@xmath1 mpc@xmath1 , here again the @xmath63 student s @xmath4 distribution and the double exponential distribution provide excellent fits to the @xmath0 error histogram when the scale factor @xmath39 is allowed to vary when minimizing @xmath24 .
these two distributions are shown in figs .
. it might be significant that the @xmath0 = 67 km s@xmath1 mpc@xmath1 case always has a lower @xmath24 than the @xmath0 = 71 km s@xmath1 mpc@xmath1 case , indicating perhaps that the median statistics value determined from a large fraction of the data is more robust time will tell . in any case , a comparison of the entries in table 1 shows that the results presented here are robust with respect to small changes in the value of @xmath0 .
lccccccccccccc gaussian & & 1 & 19.8 & 20 & @xmath77 & & 1 & 15.0 & 20 & @xmath77 + gaussian & & 1.8 & 2.63 & 19 & @xmath77 & & 1.7 & 1.92 & 19 & 0.94 + cauchy & & 1 & 1.42 & 20 & 9.9 & & 1 & 1.10 & 20 & 35 + cauchy & & 1.1 & 1.46 & 19 & 8.7 & & 1.0 & 1.15 & 19 & 29 + @xmath63 student s @xmath4 & & 1 & 2.58 & 19 & @xmath77 & & 1 & 1.56 & 19 & 5.7 + @xmath63 student s @xmath4 & & 1.3 & 0.717 & 18 & 80 & & 1.2 & 0.326 & 18 & 99.7 + double exponential & & 1 & 7.12 & 20 & @xmath77 & & 1 & 5.11 & 20 & @xmath77 + double exponential & & 1.5 & 0.501 & 19 & 96 & & 1.6 & 0.325 & 19 & 99.7 lcccccccc gaussian & & 1 & 1.0 & 2.0 & & 1 & 1.0 & 2.0 + gaussian & & 1.8 & 1.8 & 3.6 & & 1.7 & 1.7 & 3.4 + cauchy & & 1 & 1.8 & 14 & & 1 & 1.8 & 14 + cauchy & & 1.1 & 2.0 & 15 & & 1.0 & 1.8 & 14 + @xmath63 student s @xmath4 & & 1 & 1.3 & 4.5 & & 1 & 1.3 & 4.5 + @xmath63 student s @xmath4 & & 1.3 & 1.7 & 5.9 & & 1.2 & 1.6 & 5.4 + double exponential & & 1 & 1.2 & 3.1 & & 1 & 1.2 & 3.1 + double exponential & & 1.5 & 1.7 & 4.6 & & 1.6 & 1.8 & 4.9 + observed & & & 1.9 & 7.0 & & & 1.7 & 7.5 lcccccccc gaussian & & 1 & 0.68 & 0.95 & & 1 & 0.68 & 0.95 + gaussian & & 1.8 & 0.42 & 0.73 & & 1.7 & 0.44 & 0.76 + cauchy & & 1 & 0.50 & 0.71 & & 1 & 0.50 & 0.71 + cauchy & & 1.1 & 0.47 & 0.68 & & 1.0 & 0.50 & 0.71 + @xmath63 student s @xmath4 & & 1 & 0.58 & 0.82 & & 1 & 0.58 & 0.82 + @xmath63 student s
@xmath4 & & 1.3 & 0.48 & 0.74 & & 1.2 & 0.51 & 0.76 + double exponential & & 1 & 0.63 & 0.87 & & 1 & 0.63 & 0.87 + double exponential & & 1.5 & 0.49 & 0.74 & & 1.6 & 0.47 & 0.71 + observed & & & 0.48 & 0.69 & & & 0.51 & 0.72 | we construct the error distribution of hubble constant ( @xmath0 ) measurements from huchra s compilation of 461 measurements of @xmath0 and the wmap experiment central value @xmath0 = 71 km s@xmath1 mpc@xmath1 .
this error distribution is non - gaussian , with significantly larger probability in the tails of the distribution than predicted by a gaussian distribution .
the 95.4 % confidence limits are 7.0 @xmath2 in terms of the quoted errors .
it is remarkably well described by either a widened @xmath3 student s @xmath4 distribution or a widened double exponential distribution .
these conclusions are unchanged if we use instead the central value @xmath0 = 67 km s@xmath1 mpc@xmath1 found from a median statistics analysis of a major subset of @xmath0 measurements used here . |
since the original prediction by anderson that the disorder could lead to the absence of electron diffusion in imperfect crystals,@xcite anderson localization has always been one of the most fascinating phenomena in condensed - matter physics and much progress has been achieved in this research field.@xcite in fact , anderson localization is a ubiquitous quantum phenomenon and has been directly observed in a variety of disordered systems.@xcite on the other hand , the complementary subject to explore metallic states in low - dimensional disordered systems arises and has attracted extensive attention among the physics community , in the context of electron transport and direct application of realistic materials which are always disordered . for instance , the extended states have been reported in one - dimensional ( 1d ) disordered systems when the short- or long - range correlations are incorporated.@xcite another celebrated example is the aubry - andr ( aa ) model@xcite @xmath0 \varphi_n= t ( \varphi_{n-1 } + \varphi _ { n+1 } ) , \label{eq1}\end{aligned}\ ] ] where @xmath1 is the fermi energy , @xmath2 is the strength of the on - site potential , @xmath3 is an irrational number and is incommensurate with the lattice , @xmath4 is the nearest - neighbor hopping integral , and @xmath5 is the amplitude of the wavefunction at the @xmath6th site . because @xmath3 is irrational , the on - site potential displays quasiperiodicity and the aa model can be regarded as a quasiperiodic ( qp ) system .
let us consider the commensurate case of @xmath7 with @xmath8 and @xmath9 being co - prime integers .
it is obvious that the resultant energy spectrum of eq .
( [ eq1 ] ) consists of @xmath9 subbands and all electronic states are described by bloch s theorem . since an irrational number can be viewed as the ratio of two infinite integers , the energy spectrum of the aa model splits into infinite subbands , leading to a highly fractal band structure .
besides , it has been claimed that all electronic states of the aa model are delocalized when @xmath10 and are exponentially localized with lyapunov exponent @xmath11 when @xmath12.@xcite this metal - insulator transition ( mit ) in parameter space has been verified experimentally in both bose - einstein condensates and photonic lattices .
_ have demonstrated this phase transition for noninteracting ultracold atoms in optical lattices by studying the transport properties and both spatial and momentum distributions.@xcite lahini _ et al . _
have directly measured the spreading of initially narrow wave packets in photonic lattices and observed a localization phase transition at @xmath13.@xcite these recover the interest of the electronic properties of the aa model .
et al . _ have shown that a ladder network , consisting of two identical aa chains , exhibits an mit at multiple fermi energies in the presence of next - nearest - neighbor ( nnn ) hopping.@xcite biddle _ et al . _
have studied the localization properties of the aa chain by considering non - nearest - neighbor hopping and found the mobility edges.@xcite ( color online ) schematic view of the qp system ( white region with open symbols ) connected by two leads ( gray region with filled circles ) .
the black , blue , and red lines denote the intrachain , interchain , and nnn hopping integrals , respectively . here
, the width of the system is @xmath14.,scaledwidth=40.0% ] since the realistic materials usually have finite width , there is a growing interest in the localization properties of quasi-1d system@xcite and of two - dimensional ( 2d ) system in recent years.@xcite in this perspective , it would be significant to investigate the localization transition and the scaling properties of both quasi-1d and 2d qp systems , which are made up from coupling several aa chains along the transverse direction , and there has already been one work along this research field.@xcite one wonders whether there still exists an mit in these systems in parameter space and whether the scaling theory of localization remains valid in the 2d qp system . in this paper , we will answer the above two questions and fig .
[ fig : one ] plots the schematic illustration of the system contacted by two leads . by combining the landauer - bttiker formula with recursive green s function technique
, the conductance of the qp system is calculated in the presence of nnn hopping .
our results reveal that an mit could occur in the quasi-1d qp system by tuning not only the nnn hopping integral but also the number of the aa chains .
all these results are general and still hold by coupling distinct aa chains with different strengths of the on - site potentials and other model parameters .
in addition , we show from finite - size scaling that the transport properties of the 2d qp system can be described by a single parameter and the scaling function can reach the value 1 , contrary to the scaling theory of localization .
finally , the mit phase diagram is obtained and the statistical properties of the conductance are performed for the 2d qp system .
the rest of the paper is constructed as follows . in sec .
[ sec2 ] , the theoretical model and the method are presented . in sec .
[ sec3 ] , the conductance and the localization length are shown for different situations , and the scaling and statistical properties are also studied for the 2d qp system .
finally , the results are summarized in sec .
within the framework of the tight - binding approximation , the hamiltonian of the qp system connected by two leads can be written as @xmath15 , \label{eq2}\end{aligned}\ ] ] where @xmath16 ( @xmath17 ) creates ( annihilates ) an electron at lattice site @xmath18 of the left lead ( @xmath19 ) , of the central region ( @xmath20 , white region in fig .
[ fig : one ] ) , and of the right lead ( @xmath21 ) . here , @xmath22 $ ] is the chain index , @xmath23 is the width , and @xmath24 is the length of the central region .
the on - site potential @xmath25 at the central region is set as @xmath26 with @xmath27 the strength of the on - site potential of the @xmath28th chain and @xmath3 the irrational number .
@xmath4 is the intrachain hopping integral and is chosen as the energy unit . @xmath29 and @xmath30 are , respectively , the interchain and nnn hopping integrals between two neighboring chains ( see fig . [ fig : one ] ) .
the schrdinger equation @xmath31 in the site representation can be expressed as @xmath32 here , @xmath33 is the @xmath34 identity matrix , @xmath35 is the sub - hamiltonian matrix of the @xmath6th layer @xmath36 @xmath37 with @xmath38 the amplitude of the wavefunction at lattice site @xmath18 and @xmath39 the transpose , and @xmath40 is the sub - hamiltonian matrix connecting two successive layers @xmath41 at zero temperature , the landauer conductance @xmath42 can be obtained from the landauer - bttiker formula:@xcite @xmath43 \label{eq7},\end{aligned}\ ] ] with @xmath44 the quantum conductance .
@xmath45 is the linewidth function , @xmath46 is the green s function , @xmath47 is the retarded self - energy due to the coupling to the left / right lead with the on - site potential being @xmath48,@xcite and @xmath49 is the hamiltonian of the central region . in the numerical calculation ,
the conductance can be evaluated by using recursive green s function technique for very large system size . besides
, the lyapunov exponents @xmath50 s are also calculated by the transfer - matrix method.@xcite for the @xmath51 system , there are @xmath23 propagating channels along the longitudinal direction and each @xmath52 reflects the exponential decay of the corresponding channel . the smallest @xmath50 is the most physically significant quantity and its inverse is the localization length @xmath53 .
we emphasize that the propagating channel is distinct from the chain in eq .
( [ eq2 ] ) because of the interchain hopping integral @xmath29 and the nnn one @xmath30 , and these two concepts are equivalent to each other when @xmath54 .
the fixed boundary condition is imposed in the transverse direction . in this situation ,
the system is closer to the realistic materials , of which their edges usually do not connect to each other ; and corrections to scaling disappear faster with increasing system size.@xcite in the following , the irrational number is taken as @xmath55 and the results are usually averaged over @xmath56 qp samples to reach convergence , except for specific indication in the figure caption .
each qp sample is uniformly selected from the infinite system [ see eq .
( [ eq3 ] ) where the integer @xmath6 can be @xmath57 theoretically ] by using the sliding window strategy.@xcite for instance , the @xmath58th sample can be expressed as @xmath59= w_j \cos [ 2\pi \alpha n + \varphi_{i , m}]$ ] with @xmath60 and @xmath58 being positive integers . here
, the integer @xmath60 is chosen as @xmath61 and the results are independent on @xmath60 when the number of samples is sufficiently large .
we first consider the case of all @xmath62 and prove analytically the nnn hopping - induced mit in the qp system which is composed of several identical aa chains . in this regard , all @xmath35 s and @xmath40 can be diagonalized by applying a single unitary matrix @xmath63 through @xmath64 and @xmath65.@xcite then , eq .
( [ eq4 ] ) can be decoupled into following schrdinger equations : @xmath66 with @xmath67 $ ] , @xmath68 , and @xmath69 and @xmath70 the @xmath71th eigenvalue of @xmath35 and @xmath72 , respectively . @xmath73 and @xmath74 accordingly
, the qp system can be transformed into @xmath23 isolated aa chains with identical strength @xmath2 of the on - site potential [ eq . ( [ eq9 ] ) ] and different hopping integrals [ eq . ( [ eq10 ] ) ] .
these decoupled chains are exactly the propagating channels as mentioned above .
one can see that the interchain hopping integral only determines the positions of the energy bands of the decoupled chains and the energy bands will be farther away even separated from each other for greater @xmath29 .
in contrast , the nnn hopping integral only contributes to the hopping integrals of the decoupled chains , and some of the decoupled chains will be more delocalized with increasing @xmath30 and contrarily the others become more localized . besides
, we can also demonstrate the mobility edges from eqs .
( [ eq9 ] ) and ( [ eq10 ] ) for the specific case @xmath75.@xcite it is clear that when anyone of the decoupled chains has extended states , the qp system shows metallic behavior .
since the most conducting decoupled chain is the first one with @xmath76 , the critical condition to observe the mit in the qp system is@xcite @xmath77 when @xmath78 there will be at least one delocalized decoupled chain and the system exhibits metallic behavior , whereas @xmath79 all of the decoupled chains are localized and the system is insulating . although the analytical argument is performed on the special case of all @xmath62 , we believe that the nnn hopping - induced mit is general for other quasi-1d qp systems by coupling distinct aa chains .
( color online ) energy - dependent localization length @xmath53 of the two - leg ladder for ( a ) @xmath80 ( solid line ) , ( b ) @xmath81 , and ( c ) @xmath82 with @xmath83 .
the insets show the corresponding conductance @xmath42 and the dotted line in ( a ) denotes the dos .
the above results are performed for a single qp sample .
( d ) length - dependent averaged conductance @xmath84 for typical values of @xmath1 with @xmath82 , which is obtained from @xmath56 qp samples .
other parameters are @xmath85 , @xmath86 , @xmath87 , and @xmath88.,scaledwidth=45.0% ] to explore the generality of the nnn hopping - induced mit in the qp system , we study the localization properties of the two - leg ladder model as an example , by coupling two distinct aa chains with different @xmath27 s . in this situation ,
it is difficult to provide analytical results and the numerical calculations are performed instead . figures [ fig : two](a ) , [ fig : two](b ) , and [ fig : two](c ) show the localization length @xmath53 for @xmath80 , @xmath89 , and @xmath90 , respectively , as a function of energy @xmath1 with @xmath87 and @xmath91 , while the insets display the corresponding landauer conductance @xmath42 . here , @xmath53 is averaged within a very small energy interval of @xmath92 to avoid numerical fluctuations . although the localization properties of the two - leg ladder strongly depend upon the nnn hopping integral @xmath30 ( discussed later )
, one can see several general features that are irrespective of @xmath30 .
firstly , all the energy spectra consist of two isolated subbands , because of the repulsion effects between the two chains driven by the large interchain hopping integral @xmath93 [ eq . ( [ eq9 ] ) ]
. and these two subbands are exactly the propagating channels of the two - leg ladder . secondly
, there exist many sharp peaks in the curve @xmath53-@xmath1 due to the highly fractal energy band of the aa model .
this feature can be further identified in the density of states ( dos ) [ see the dotted line in fig . [ fig : two](a ) ] which is obtained by diagonalizing the hamiltonian of the central region , and the quantized energy levels coincide with the peaks in the curve @xmath53-@xmath1 [ fig . [
fig : two](a ) ] . here , we emphasize that the dos profile does not change with the system length @xmath24 when the two - leg ladder is sufficiently long .
the origin of the extremely fractal energy spectrum can be also understood as follows : the aa model can be projected from the 2d square lattice which develops degenerate landau energy levels in the presence of a perpendicular magnetic field.@xcite in the absence of @xmath30 , the two subbands are almost symmetric with respect to @xmath94 and the two - leg ladder exhibits insulating behavior with @xmath95 and @xmath96 [ see the solid line and the inset of fig .
[ fig : two](a ) ] , although all electronic states of the first aa chain are delocalized ( @xmath87 ) , consistent with previous study that the system composed of one disordered chain and another completely ordered chain is localized.@xcite this is attributed to the fact that when an electron is injected into the first chain , it will be continuously scattered from larger ( localized ) potentials of the second chain , owing to the interchain hopping integral .
the longer the system is , the more the scattering the electron suffers .
accordingly , @xmath42 will be exponentially declined to zero with increasing @xmath24 , a typical feature of anderson localization . while in the presence of @xmath30 that could lead to electron - hole asymmetry , the two subbands become asymmetric and opposite behaviors
could be observed with increasing @xmath30 . for the left subband ,
its bandwidth decreases monotonically from @xmath97 at @xmath80 to @xmath98 at @xmath82 and @xmath53 is gradually declined by increasing @xmath30 , and hence the electronic states become more localized . in contrast
, the bandwidth of the right subband is enhanced from @xmath97 at @xmath80 to @xmath99 at @xmath82 and @xmath53 increases with @xmath30 , and nonzero conductance could emerge in the region of large @xmath30 [ see the insets of figs . [
fig : two](b ) and [ fig : two](c ) ] .
this phenomenon can be explained as follows .
in fact , the nnn hopping integral could produce additional two propagating channels with opposite sign of hopping integrals [ eq . ( [ eq10 ] ) ] and thus affects the transport properties of the two - leg ladder by mixing these channels with the original ones at @xmath80 .
when the hopping integrals of the additional and original propagating channels are of different signs , the effective hopping integral will be reduced and leads to the shrink of both the bandwidth and the localization length .
contrarily , when they have identical sign , the effective hopping integral can be enhanced and both the bandwidth and the localization length can be increased , leading to a possible mit in the two - leg ladder .
actually , the electronic states of finite conductance observed in the insets of figs . [ fig : two](b ) and [ fig : two](c ) are delocalized even in the thermodynamic limit , because the conductance of the localized states presents exponential dependence on the system length and should be zero when @xmath83 .
this statement can be further verified in fig .
[ fig : two](d ) , where the averaged conductance @xmath84 is plotted for several typical energies as a function of @xmath24 with other parameters identical to those in fig .
[ fig : two](c ) .
it clearly appears that @xmath84 fluctuates around a certain value with increasing @xmath24 and does not tend to zero .
therefore , we conclude that these electronic states are truly delocalized and an mit could be driven in the two - leg ladder by increasing @xmath30 .
( color online ) energy - dependent @xmath42 of the two - leg ladder by coupling two distinct aa chains with @xmath80 ( solid line ) and @xmath82 ( dotted line ) . ( a ) @xmath100 , @xmath101 , and @xmath102 .
( b ) @xmath87 , @xmath91 , and @xmath103 .
other parameters are @xmath85 , @xmath83 , and @xmath104 .
the results are calculated from a single qp sample.,scaledwidth=45.0% ] to further substantiate the nnn hopping - induced mit and the stability of the extended states in the region of large @xmath30 , we then investigate the localization properties of the two - leg ladder with various model parameters , as illustrated in figs .
[ fig : three ] and [ fig : four ] .
firstly , we consider other values of @xmath27 s , small @xmath29 , and large @xmath3 .
figure [ fig : three](a ) shows the conductance @xmath42 versus @xmath1 by coupling two localized aa chains with @xmath100 and @xmath101 , while fig . [
fig : three](b ) plots @xmath42 versus @xmath1 for large irrational number @xmath103 , with the small interchain hopping integral @xmath105 .
it can be seen that all electronic states are localized with @xmath96 in the absence of @xmath30 [ see the solid lines in figs .
[ fig : three](a ) and [ fig : three](b ) ] and the extended states of nonzero conductance could emerge when @xmath30 becomes large [ see the dotted lines in figs . [
fig : three](a ) and [ fig : three](b ) ] .
since the band structure of the two - leg ladder is sensitive to both @xmath3 and @xmath27 s , the position of the extended states is distinct from each other when @xmath3 or @xmath27 is different .
( color online ) ( a ) eigenvalue spectrum of the two - leg ladder by varying the irrational number @xmath3 within @xmath106 .
( b ) 2d plot of @xmath107 versus the eigenvalue number @xmath71 [ see the horizontal coordinate @xmath108 in ( a ) ] and @xmath3 .
larger @xmath71 refers to higher @xmath108 . here , @xmath109 and other parameters are the same as those in fig . [
fig : two](c ) .
all these results are performed for a single qp sample , and are similar when other qp samples are considered or by coupling two identical aa chains with @xmath110.,scaledwidth=45.0% ] secondly , we consider the irrational number @xmath3 within a wide range .
figure [ fig : four](a ) plots the eigenvalue spectrum of the two - leg ladder by changing @xmath3 from @xmath48 to @xmath111 .
it clearly appears that there exist two hofstadter - like butterflies in the energy spectrum , although the system is composed of two distinct aa chains with different @xmath27 s .
the left subband exhibits the compressed configuration of the hofstadter s butterfly due to the small width @xmath112 ( the original hofstadter s butterfly was reported at @xmath13 with width @xmath113),@xcite and the right subband shows the expanded version with width @xmath114 . to illustrate the localization properties at different @xmath3 , it is convenient to calculate the participation ratio @xmath107:@xcite @xmath115 which measures the extension of the wavefunction . figure [ fig : four](b ) shows the corresponding 2d plot of @xmath107 versus the eigenvalue number @xmath71 and @xmath3 .
it can be seen that @xmath107 is very small in the left subband ( @xmath116 ) and the wavefunctions are confined within a small region .
our further calculations reveal that @xmath107 decreases with increasing @xmath24 , indicating that all electronic states are localized in the left subband.@xcite in contrast , @xmath107 is usually very large in the right subband ( @xmath117 ) and the wavefunctions can spread over the entire system for all investigated values of @xmath3 , except for some electronic states which locate at the band edge and the corresponding @xmath107 remains quite small [ see the four corners of the right subband in fig . [
fig : four](b ) ] .
the small @xmath107 in these corners may originate from the fact that : ( 1 ) the qp character of the system is not obvious for short system length @xmath109 when @xmath3 is very close to an integer ; ( 2 ) the electronic states in the band edge are isolated from others and can not form a band , and thus become fragile .
however , we find that the large @xmath107 is independent on @xmath24 ( data not shown ) , implying that the electronic states are delocalized in the right subband . besides
, one can see that @xmath107 exhibits clustering patterns in the right subband and is extremely large within these regions , owing to the highly degenerate energy levels .
therefore , our results demonstrate that the extended states still exist in the case of large @xmath30 for all investigated values of @xmath27 , @xmath29 , and @xmath3 , and thus the nnn hopping - induced mit is general . since the localization properties of the system does not depend upon @xmath29 , in the following we fix the interchain hopping integral as @xmath118 without loss of generality .
( color online ) @xmath42 versus @xmath1 of several quasi-1d qp systems with different width @xmath23 . ( a )
@xmath119 for all @xmath28 s .
( b ) @xmath120 when @xmath28 is odd and @xmath121 when @xmath28 is even .
other parameters are @xmath83 , @xmath104 , and @xmath122 . the results are performed for a single qp sample.,scaledwidth=45.0% ] besides the nnn hopping , it can be seen from eq .
( [ eq10 ] ) that the hopping integrals of some decoupled chains can be enhanced with increasing the system width @xmath23 in the presence of @xmath30 and the extended states can emerge in the case of large @xmath23 when @xmath78 .
figure [ fig : five](a ) displays the conductance @xmath42 versus @xmath1 for several quasi-1d qp systems which are composed of identical aa chains with all @xmath119 and @xmath83 .
one notices that @xmath42 is zero within the numerical accuracy for @xmath85 [ see the solid line of fig .
[ fig : five](a ) ] and is progressively increased with increasing @xmath23 as expected [ see the dotted and dash - dotted lines of fig . [
fig : five](a ) ] , implying the occurrence of mit in the quasi-1d qp systems .
the underlying physics for the extended states observed in large @xmath23 can be ascribed to the dimensionality effects together with the nnn hopping .
just as the two - leg ladder model , the nnn hopping integral could also generate additional propagating channels with various hopping integrals in the quasi-1d systems [ eq . ( [ eq10 ] ) ] .
these additional propagating channels will interact with the original ones at @xmath80 and change the transport properties of the quasi-1d system with both localized and delocalized channels . since both the number of the delocalized channels and the conductance of the original ones increase with @xmath23 , the system
s conductance can become larger with increasing @xmath23 .
this dimensionality effects - induced mit is a general phenomenon and can still take place in the qp systems by coupling distinct aa chains of different @xmath27 s , as illustrated in fig .
[ fig : five](b ) , where the conductance @xmath42 is shown for several quasi-1d qp systems by coupling two kinds of aa chains . besides , the conductance profiles exhibit several gaps for small @xmath23 [ see the dotted lines of figs . [
fig : five](a ) and [ fig : five](b ) ] and these gaps tend to disappear for large @xmath23 , because the energy spectra of the delocalized channels are different and can distribute over a wide energy range with increasing @xmath23 .
therefore , we conclude that an mit could be also driven in the quasi-1d qp systems by increasing the system width .
( color online ) scaling properties of @xmath123 for the 2d qp system with @xmath124 and all @xmath62 .
( a ) @xmath123 versus @xmath1 for different @xmath24 and ( b ) @xmath125 versus @xmath24 for several values of @xmath1 with @xmath126 and @xmath122 .
( c ) @xmath123 versus @xmath30 for distinct @xmath24 , ( d ) @xmath123 versus @xmath24 for various @xmath30 which is very close to the critical value @xmath127 ( see text ) , and ( e ) @xmath123 versus @xmath24 for several @xmath30 which is distant from @xmath127 , by fixing @xmath128 and @xmath129 . in fig .
[ fig : six](d ) , the squares , circles , down triangles , diamonds , and left triangles denote @xmath130 , @xmath131 , @xmath132 , @xmath133 , and @xmath134 , respectively . in fig .
[ fig : six](e ) , the circles and down triangles represent , respectively , @xmath81 and @xmath135 with their vertical axis on the left side , while the squares correspond to @xmath136 with its vertical axis on the right side ( as indicated by the arrow ) .
the solid lines in ( b ) and ( e ) are the linear fitting curves @xmath137 with @xmath138 being about @xmath111 .
the results are averaged over @xmath56 qp samples.,scaledwidth=47.0% ] we then study the localization phase transition and the scaling properties of the 2d qp system with @xmath139 .
for the purpose of scaling , the contact resistance is subtracted from @xmath42 to yield the thouless conductance @xmath140 , which is expressed as@xcite @xmath141 here , n is the number of propagating channels in the leads at fermi energy @xmath1 with @xmath142 the contact resistance .
then , the scaling function@xcite @xmath143 is numerically evaluated and is used to determine the localization and scaling properties . @xmath144 and
@xmath145 correspond to the insulator and the metal , respectively .
figure [ fig : six](a ) shows the energy - dependent @xmath123 for the 2d qp system with several values of @xmath24 .
it can be seen that @xmath123 increases with @xmath24 in a wide energy range and exhibits strong fluctuations in the band edge , owing to the fractal energy spectrum of the aa model . the scaling behavior of @xmath123 on @xmath24 is plotted in fig . [ fig : six](b ) for several typical energies .
it clearly appears that @xmath123 increases monotonically with @xmath24 and the dependence of @xmath123 on @xmath24 can be well fitted by a simple function @xmath137 , with the exponent @xmath146 being about 1 .
this implies that the 2d qp system presents truly metallic behavior in the thermodynamic limit .
figure [ fig : six](c ) displays @xmath123 versus @xmath30 for several @xmath24 .
one can see that @xmath123 increases with @xmath30 , due to the emergence of additional delocalized channels and the enhancement of the conductance of the original ones in the 2d system .
besides , the curves @xmath123-@xmath30 for different @xmath24 intersect at a critical value @xmath127 of the nnn hopping integral . for @xmath147 , @xmath123 decreases with @xmath24 and corresponds to the localized regime ; whereas for @xmath148 , @xmath123 increases with @xmath24 and the system is in the metallic regime .
this feature can be further observed in figs .
[ fig : six](d ) and [ fig : six](e ) , where @xmath123 is shown as a function of @xmath24 with @xmath30 close to and distant from the critical value @xmath127 , respectively .
it can be seen that there exist fluctuations in the curve @xmath123-@xmath24 in the region of small conductance [ see the squares , circles , and down triangles in fig .
[ fig : six](d ) ] .
the curve @xmath123-@xmath24 is smooth in the region of large conductance [ see the left triangles in fig . [
fig : six](d ) ] and can be also fitted well by the function @xmath137 with @xmath149 when @xmath123 becomes larger [ see the circles and down triangles in fig .
[ fig : six](e ) ] .
( color online ) main frame : scaling function @xmath150 versus @xmath123 for the 2d qp system .
@xmath150 is calculated from a wide range of @xmath2 , @xmath30 , and @xmath1 . here
, the length is taken as @xmath151 to capture the qp character of the system , the step of the length is chosen as @xmath152 , and @xmath56 qp samples are considered .
the oblique line is the fitting curve of @xmath153 for @xmath154 and the horizontal line is @xmath155 .
the top inset displays the mit phase diagram in the `` @xmath2-@xmath30 '' space ( black symbols ) , which is extracted from fig .
[ fig : six](c ) , and the solid line denotes the curve @xmath156 .
the bottom inset shows @xmath123 as a function of the sample number @xmath157 with @xmath129 , @xmath158 , and other parameters being identical to those in fig .
[ fig : six](a).,scaledwidth=45.0% ] figure [ fig : seven ] plots the scaling function @xmath150 versus @xmath123 and the bottom inset shows the dependence of @xmath123 on the sample number @xmath157 for typical parameters .
it clearly appears that the oscillating amplitude of @xmath123 declines quickly with increasing @xmath157 and @xmath123 saturates when @xmath159 ( see the bottom inset of fig .
[ fig : seven ] ) .
for instance , the ratio of the difference between @xmath123 at @xmath160 and its saturated value to the saturated value is 0.08% , where the saturated value of @xmath123 is prescribed at @xmath161 . here ,
@xmath150 is calculated from @xmath56 samples to avoid spurious effects and the length is taken as @xmath151 to capture the qp character of the system .
it can be seen from fig .
[ fig : seven ] that all data points tend to construct a single curve of @xmath162 . for small conductance
, the scaling function can be linearly fitted by a simple function @xmath153 ( see the oblique line ) , in good agreement with the scaling theory of localization,@xcite implying that the conductance falls off exponentially with increasing @xmath24 in the localized regime . for large conductance ,
the scaling function is approximately @xmath111 ( see the horizontal line of @xmath155 ) , indicating that the conductance increases linearly with @xmath24 .
this point is contrary to the scaling theory of localization that the scaling function should be less than @xmath48 for the 2d disordered systems.@xcite the underlying physics can be attributed to the fact that the system we studied is qp but not disordered . in this situation , the delocalized channels as well as their conductances will be increased by increasing @xmath24
. however , there exists fluctuation in the curve @xmath162 , especially in the regime of small conductance .
this phenomenon is mainly due to the highly fractal energy spectrum of the qp system .
one can see from fig .
[ fig : six](a ) that @xmath123 fluctuates very strongly for small @xmath123 , in sharp contrast to the situation that it is smooth for large @xmath123 , for whatever the system size is .
these fluctuations can be also identified in fig .
[ fig : six](d ) as discussed above .
the top inset of fig .
[ fig : seven ] displays the mit phase diagram in the `` w - t@xmath163 '' space ( see the black symbols ) , which is subtracted from fig .
[ fig : six](c ) .
it can be seen that the critical nnn hopping integral @xmath30 to observe metallic states increases monotonically with @xmath2 and the curve of @xmath30-@xmath2 can be fitted by @xmath164 .
( color online ) statistical properties of the thouless conductance for the 2d qp system .
( a ) distribution of @xmath140 and ( b ) energy - dependent standard deviation @xmath165 .
the results are obtained from @xmath166 qp samples and other parameters are identical to those in fig . [
fig : six](a ) .
, scaledwidth=37.0% ] finally , we investigate the statistical properties of the thouless conductance of the 2d qp system .
figures [ fig : eight](a ) and [ fig : eight](b ) show , respectively , the distribution of @xmath140 for typical parameters and the energy - dependent standard deviation @xmath165 for three system sizes with @xmath167 .
it can be seen from fig .
[ fig : eight](a ) that the distribution of @xmath140 is extremely complicated and is not normal , which is different from previous studies in disordered systems.@xcite because of the cosine modulation of the on - site energies and the quasiperiodicity , we find that the conductance presents oscillating behavior with various amplitudes when the sample is successively chosen from an infinite aa chain by using the sliding window strategy ( data not shown ) .
besides , we can see that the probability is quite large or small when the conductance is around specific values . as a result
, the distribution of @xmath140 will deviate from the normal one . by inspecting fig .
[ fig : eight](b ) , one notices the following features : ( 1 ) the magnitude of @xmath168 is comparable to @xmath169 [ see fig .
[ fig : six](a ) and @xmath170 can be viewed as approximation of @xmath169 ] , where @xmath168 is large around the band center and becomes very small within the band edge .
it can be also observed that @xmath168 becomes greater when @xmath169 is larger , indicating that the fluctuation of the conductance will be stronger for larger conductance .
( 2 ) @xmath168 is enhanced by increasing @xmath24 around the band center , where the conductance increases with @xmath24 [ see fig . [
fig : six](a ) ] and the system presents metallic behavior .
this is different from the universal conductance fluctuation theory that @xmath168 should be independent of the system size .
( 3 ) there are several dips in the profile of @xmath172 , similar to that observed in fig .
[ fig : six](a ) .
in summary , we investigate the localization phase transition and the scaling properties of both quasi - one - dimensional and two - dimensional quasiperiodic systems , which are constructed by coupling several aubry - andr chains along the transverse direction . by employing the landauer - bttiker formula and recursive green s function method ,
the two - terminal conductance is calculated for the quasiperiodic systems with next - nearest - neighbor hopping .
the numerical results indicate that a metal - insulator transition could be produced in these quasiperiodic systems in parameter space by adjusting the next - nearest - neighbor hopping integral and the number of chains .
these results are general and hold by coupling distinct aa chains with different strengths of the on - site potentials and other model parameters .
we find that the energy spectrum of the two - leg ladder model exhibits two hofstadter - like butterflies .
besides , we show from the finite - size scaling that the transport properties of the two - dimensional quasiperiodic system can be described by the single thouless conductance and the scaling function can reach the value 1 , contrary to the scaling theory of localization .
we thank yan - yang zhang for useful discussions .
this work was financially supported by nbrp of china ( 2012cb921303 and 2012cb821402 ) , nsf - china under grant nos .
11274364 and 91221302 , and pdsf - china under grant no .
2013m540153 .
p. w. anderson , phys . rev . * 109 * , 1492 ( 1958 ) .
e. abrahams , p. w. anderson , d. c. licciardello , and t. v. ramakrishnan , phys .
rev . lett . * 42 * , 673 ( 1979 ) .
f. evers and a. d. mirlin , rev .
phys . * 80 * , 1355 ( 2008 ) . ,
edited by e. abrahams ( world scientific , singapore , 2010 ) . s. johri and r. n. bhatt , phys .
* 109 * , 076402 ( 2012 ) .
t. schwartz , g. bartal , s. fishman , and m. segev , nature ( london ) * 446 * , 52 ( 2007 ) .
y. lahini , a. avidan , f. pozzi , m. sorel , r. morandotti , d. n. christodoulides , and y. silberberg , phys .
lett . * 100 * , 013906 ( 2008 ) .
h. hu , a. strybulevych , j. h. page , s. e. skipetrov , and b. a. van tiggelen , nature phys .
* 4 * , 945 ( 2008 ) . j. billy , v. josse , z. zuo , a. bernard , b. hambrecht , p. lugan , d. clment , l. sanchez - palencia , p. bouyer , and a. aspect , nature ( london ) * 453 * , 891 ( 2008 ) . g. roati , c. derrico , l. fallani , m. fattori , c. fort , m. zaccanti , g. modugno , m. modugno , and m. inguscio , nature ( london ) * 453 * , 895 ( 2008 ) .
s. faez , a. strybulevych , j. h. page , a. lagendijk , and b. a. van tiggelen , phys .
lett . * 103 * , 155703 ( 2009 ) .
y. lahini , r. pugatch , f. pozzi , m. sorel , r. morandotti , n. davidson , and y. silberberg , phys .
lett . * 103 * , 013901 ( 2009 ) .
s. s. kondov , w. r. mcgehee , j. j. zirbel , and b. demarco , science * 334 * , 66 ( 2011 ) .
f. jendrzejewski , a. bernard , k. mller , p. cheinet , v. josse , m. piraud , l. pezz , l. sanchez - palencia , a. aspect , and p. bouyer , nature phys . * 8 * , 398 ( 2012 ) . f. a. b. f. de moura and m. l. lyra , phys .
lett . * 81 * , 3735 ( 1998 ) . v. bellani , e. diez , r. hey , l. toni , l. tarricone , g. b. parravicini , f. domnguez - adame , and r. gmez - alcal , phys . rev
. lett . * 82 * , 2159 ( 1999 ) .
f. m. izrailev and a. a. krokhin , phys . rev
. lett . * 82 * , 4062 ( 1999 ) .
u. kuhl , f. m. izrailev , a. a. krokhin , and h .- j .
stckmann , appl .
. lett . * 77 * , 633 ( 2000 ) .
f. m. izrailev , a. a. krokhin , and n. m. makarov , phys . rep . * 512 * , 125 ( 2012 ) .
s. aubry and g. andr , ann .
. soc . * 3 * , 133 ( 1980 ) .
s. sil , s. k. maiti , and a. chakrabarti , phys .
* 101 * , 076803 ( 2008 ) .
j. biddle and s. das sarma , phys .
lett . * 104 * , 070601 ( 2010 ) .
j. biddle , d. j. priour jr .
, b. wang , and s. das sarma , phys .
b * 83 * , 075105 ( 2011 )
. f. a. b. f. de moura , r. a. caetano , and m. l. lyra , phys .
b * 81 * , 125104 ( 2010 ) .
w. zhang , r. yang , y. zhao , s. duan , p. zhang , and s. e. ulloa , phys .
b * 81 * , 214202 ( 2010 ) . a .-
guo and s .- j .
xiong , phys .
b * 83 * , 245108 ( 2011 ) .
xie , v. e. kravtsov and m. mller , phys .
b * 86 * , 014205 ( 2012 ) . j. j. krich and a. aspuru - guzik , phys . rev
. lett . * 106 * , 156405 ( 2011 ) .
t. a. sedrakyan , j. p. kestner , and s. das sarma , phys .
a * 84 * , 053621 ( 2011 ) .
a. rodriguez , a. chakrabarti , and r. a. rmer , phys .
b * 86 * , 085119 ( 2012 ) .
m. hilke , phys .
* 91 * , 226403 ( 2003 ) .
a. punnoose and a. m. finkelstein , science * 310 * , 289 ( 2005 ) .
m. onoda , y. avishai , and n. nagaosa , phys .
lett . * 98 * , 076802 ( 2007 ) .
j. h. bardarson , j. tworzydo , p. w. brouwer , and c. w. j. beenakker , phys .
* 99 * , 106801 ( 2007 ) .
k. nomura , m. koshino , and s. ryu , phys .
lett . * 99 * , 146806 ( 2007 ) .
zhang , j. hu , b. a. bernevig , x. r. wang , x. c. xie , and w. m. liu , phys .
lett . * 102 * , 106401 ( 2009 ) .
z. xu , l sheng , r shen , b. wang and d y xing , j. phys . : condens .
matter * 25 * , 065501 ( 2013 ) .
, edited by s. datta ( cambridge university press , cambridge , england , 1995 ) .
d. h. lee and j. d. joannopoulos , phys . rev .
b * 23 * , 4997 ( 1981 ) . k. slevin , t. ohtsuki , and t. kawarabayashi , phys .
lett . * 84 * , 3915 ( 2000 ) .
k. slevin , p. marko , and t. ohtsuki , phys .
rev . lett .
* 86 * , 3594 ( 2001 ) .
guo and s .- j .
xiong , phys .
e * 79 * , 041924 ( 2009 ) . a .-
guo , s .- j .
xiong , x. c. xie , and q .- f .
sun , j. phys . : condens .
matter * 25 * , 415501 ( 2013 ) .
d. r. hofstadter , phys .
b * 14 * , 2239 ( 1976 ) .
k. slevin , y. asada , and l. i. deych , phys .
b * 70 * , 054201 ( 2004 ) . | in this paper , we explore the localization transition and the scaling properties of both quasi - one - dimensional and two - dimensional quasiperiodic systems , which are constituted from coupling several aubry - andr ( aa ) chains along the transverse direction , in the presence of next - nearest - neighbor ( nnn ) hopping .
the localization length , two - terminal conductance , and participation ratio are calculated within the tight - binding hamiltonian .
our results reveal that a metal - insulator transition could be driven in these systems not only by changing the nnn hopping integral but also by the dimensionality effects .
these results are general and hold by coupling distinct aa chains with various model parameters .
furthermore , we show from finite - size scaling that the transport properties of the two - dimensional quasiperiodic system can be described by a single parameter and the scaling function can reach the value 1 , contrary to the scaling theory of localization of disordered systems . the underlying physical mechanism is discussed . |
granular superconductors ( sc ) are composed of a very large number of small ( micron - size ) superconductive grains which are coupled together due to the josephson tunnelling ( or , in some cases , due to the proximity effect ) .
these systems are inherently disordered due to randomness in the sizes of grains and in their mutual distances .
usually the strength of josephson coupling between grains is rather weak , so the maximum josephson energy of the contact between two grains is much below the intragrain superconductive condensation energy .
therefore granular sc can be considered as systems with a two - level organization : their short - scale properties are determined by the superconductivity of individual grains , whereas the macroscopic sc behaviour is governed by the weak intergrain couplings . in the treatment of the latter
, one can neglect any internal structure of sc grains and describe them just by the phases @xmath5 of their superconductive order parameters @xmath6 where @xmath7 is the phase difference induced by the electromagnetic vector potential @xmath8 and @xmath9 , whereas the coupling strengths @xmath10 are proportional to the maximum josephson currents : @xmath11 .
the vector potential @xmath8 in eq.([ham ] ) is the sum of the vector potential @xmath12 of the external magnetic field @xmath13 and of the josephson currents - induced vector potential @xmath14 . in the absence of external magnetic field ,
the lowest - energy state for the `` hamiltonian '' ( [ ham ] ) is , clearly , a macroscopically superconductive state with all phases @xmath5 equal to each other .
thus that granular sc system looks similar to the random xy ferromagnet with randomness in the values of the coupling strengths @xmath10 s ( apart from the possible role of the induced vector potential @xmath14 which will be dicussed later ) ; within this analogy the role of xy `` spin components '' is taken by @xmath15 , @xmath16 . the situation becomes a lot more complicated in the presence of non - zero external magnetic field @xmath13 , which makes the system _ randomly frustrated _
( since magnetic fluxes penetrating plaquettes between neighbouring grains are random fractional parts of @xmath17 ) . when the external field is sufficiently strong , @xmath18 ( here @xmath19 is the characteristic intergrain distance ) , the random phases @xmath20 become of the order of @xmath21 or larger , which means complete frustration of the intergrain couplings
i.e. the system is then expected to resemble the xy _
spin glass_. actually the random josephson network in a magnetic field is not exactly identical to the xy spin glass due to the following reasons @xcite : i ) .
the effective couplings @xmath22 between `` spins '' @xmath23 of the frustrated sc network are _ random complex _
numbers whereas in the xy spin glass model , they are _ real random _ numbers .
ii ) . generally the phases @xmath20 depend on the total magnetic induction @xmath24 , i.e. the effective couplings @xmath25 depend on the phase variables @xmath5 determining the intergrain currents @xmath26 . in some cases
the effects produced by the self - induced magnetic field @xmath27 are weak and can be neglected ( the quantitative criterion will be discussed later on ) , so that phases @xmath20 can be considered as being fixed by the external field . the model described by the hamiltonian ( [ ham ] ) with fixed @xmath20 s and @xmath28
is usually called `` gauge glass '' model .
it is expected on the basis of the analytical @xcite as well as numerical @xcite results that the gauge glass model in 3d space exhibits a true phase transition into a low - temperature glassy superconductive ( nonergodic ) state .
the mean - field theory of such a low - temperature state shows @xcite that it is characterized by the presense of a finite effective penetration depth for the _ variation _ of an external field , nonzero macroscopic critical current , and the absense of a macroscopic meissner effect . the full model ( [ ham ] ) with @xmath29 s containing contribution from @xmath27
is sometimes called `` gauge glass with screening '' @xcite .
the effect of screening on the presence and properties of the phase transition into a glassy state is not completely clear ; some numerical results @xcite indicate the absence of a true phase transtion in a 3d model with screening .
quantitatively , the strengh of screening is determined by the ratio @xmath30 where @xmath31 is the characteristic inductance of an elementary intergrain current loop @xcite . in the ceramics with @xmath32 ,
screening effects become important on a long - distance scale @xmath33
only ( i.e. they are similar to the strongly type - ii superconductors with disorder ) . apart from its relevance for the description of granular superconductors , the gauge glass model with screening is rather often considered ( e.g. ref .
@xcite ) as a simplified model describing the large - scale behaviour of disordered bulk type - ii superconductors in the mixed state ( so - called vortex glass problem ) .
actually it is unclear a priori how these two problems are related ; an obvious difference between them is that the basic ingredient of the latter is the vortex lattice which is clearly an anisotropic object , whereas the former does not contain any prescribed direction in the 3d space . on the other hand , the granular superconductor in a moderate magnetic field @xmath34
may be considered as a kind of disordered type - ii superconductor , where the notion of a hypervortex ( which is the macroscopic analogue of the abrikosov vortex ) can be introduced @xcite .
therefore , the macroscopic properties of a granular network at @xmath35 may resemble those of the vortex glass ; in such a scenario a phase transtion between vortex glass and gauge glass phases would be expected in a granular superconductive network at @xmath36 ( cf .
ref.@xcite for a more detailed discussion of this subject ) .
recently , it was noted that granular superconductors may become glassy even in the absence of external magnetic field , if a large enough part of josephson junctions are anomalous , i.e. their minimum josephson coupling energy corresponds to a phase difference @xmath37 instead of @xmath38 ( so - called @xmath21-junctions ) .
two completely different origins of @xmath21-junctions were proposed : mesoscopic fluctuations in dirty superconductors @xcite and the pairing with non - zero momentum @xcite .
recent experiments revealing the @xmath4-wave nature of pairing in high - temperature superconductors @xcite indicate the possibility of observing glassy superconductive behaviour in htsc ceramics in virtually zero magnetic field .
note that ceramics with equal concentrations of usual and @xmath21-junctions are completely equivalent ( if screening effects can be neglected ) to the @xmath39 spin glass .
contrary to the 3d gauge glass model , the xy spin glass in 3d is expected to have no true thermodynamic phase transition at finite temperature @xcite ; recently , it has been suggested that the xy spin - glass and d - wave ceramic superconductor might have a new equilibrium ordered phase , the so - called chiral - glass phase @xcite .
however , these issues are hardly relevant for the measurable response at temperatures much below `` bare '' glass transition temperature @xmath40 , which we consider in this paper .
experimental studies of granular superconductors reveal @xcite an appearance of magnetic irreversibility ( a difference between meissner and shielding magnetizations or , in other terms , between field cooled ( f.c . ) and zero field cooled ( z.f.c . ) magnetizations ) below some temperature @xmath40 , which is lower than the sc transition temperature @xmath41 of the grains .
however , detailed analysis of the magnetic response in such systems is usually complicated by the mixing of contributions from individual grains and from the intergrain currents .
the goal of this paper is to develop a method which makes it possible to extract from the raw data on d.c .
magnetic response the intergrain contribution and to compare its behaviour with existing theoretical predictions .
the compound la@xmath1sr@xmath2cuo@xmath3 was chosen in this study for experimental convenience , since its critical temperature ( @xmath42 ) is within the optimal temperature range of our noise and a.c .
susceptibility measurements setup .
the sample was fabricated by standard solid state reaction of la@xmath43o@xmath44 , srco@xmath44 and cuo @xcite .
mixed powder was pressed into pellets which were sintered in air at @xmath45c for 12 hours . the material was then submitted to three cycles of regrinding , sifting to @xmath46 , pressing and sintering again at @xmath47c for 12 hours .
samples prepared in two successive runs were used in this study . in the first one ( sample a )
, pellets @xmath48 thick and @xmath49 diameter were obtained , with a density about @xmath50 of the theoretical bulk value . in the second one ( sample b ) ,
cylinders of diameter @xmath51 and length 5 to @xmath51 were prepared with a density ratio about @xmath52 . in both preparations ,
grains sizes were in the range 1 10 @xmath53 .
room temperature x - ray powder diffraction patterns showed the presence of a small amount ( @xmath54 ) of the non superconductive compound la@xmath55sr@xmath56cu@xmath43o@xmath57 .
the rest of the paper is organized as follows . in section
ii the general analysis of the magnetic response data obtained on two different samples ( a and b ) of la@xmath1sr@xmath2cuo@xmath3 ceramics is presented and the intergrain ( josephson ) contribution to the overall response is extracted .
section iii is devoted to the detailed study of the magnetic response of josephson intergrain network in the low - field range .
it is found that the macroscopic critical current is suppressed considerably ( by a factor 2 ) , in a magnetic field of only about @xmath58 .
the lower - field d.c.response to field variations of order @xmath59 was analyzed for the f.c .
states obtained at @xmath60 and two temperatures , @xmath61 and @xmath62 .
the data at @xmath63 and @xmath64 and @xmath65 are shown to be compatible with the bean critical - state picture @xcite and the low - field critical current value is identified .
the rest of the data are in a sharp contrast with bean - model predictions : the screening current _ grows sublinearly _ ( approximately as a square root ) with increasing exitation field .
very low field , low frequency a.c .
measurements are presented , which reveal the strongly irreversible nature of that anomalous response .
a new phenomenological model is proposed for the treatment of these data .
its first predictions are found to be in a reasonable agreement with the data . in section iv
, the theoretical analysis of our experimental results is given in terms of the existing theories of `` gauge - glass '' state .
it is shown that the observed transition temperature to the low - temperature state of the network and the magnitude of the ( low-@xmath66 ) critical current are in sharp contradiction with the ( usual ) assumption that the zero - field granular network is unfrustrated .
on the contrary , under the assumption of a strongly frustrated network at @xmath67 , all basic measured parameters of the ceramic network are in mutual agreement .
we believe that these estimates indicate the existence of a large proportion of @xmath21-junctions in the la@xmath1sr@xmath2cuo@xmath3 ceramics , possibly due to the @xmath4-wave nature of superconductivity in cuprates . the section v is devoted to the development of a new model of diamagnetic response in glassy superconductors , which is necessary for the description of the anomalous data described at the end of section iii .
this new model ( in some sense , intermediate between the bean @xcite and the campbell @xcite ones ) is based on two ideas : i ) the existence of two characteristic `` critical '' currents ( @xmath68 and @xmath69 ) , and ii ) the fractal nature of free energy valleys in the ceramic network .
our conclusions are presented in section vi , whereas some technical calculations can be found in the appendix . for convenience
, the e.m.u system of units will be used for experimental data , and gaussian units for the theoretical discussions .
the d.c . magnetization was measured by the classical extraction method .
two squid magnetometers were used : one a home made apparatus used in several previous spin - glass studies @xcite , the other a commercial system ( cryogenics s500 ) . in this section ,
we describe successively the static magnetic response of samples a and b and present a preliminary treatment of these data , in order to distinguish between the magnetic response of individual grains and intergrain currents @xcite ( a detailed study of the latter is the subject of the next section ) .
firstly , we present results obtained after cooling the samples in various d.c . fields and applying small field increases .
secondly , we will derive from the results the response of the josephson currents as a function of field and temperature .
finally , we will show that the behavior of the field cooled ( f.c . ) susceptibility can be satisfactorily accounted for if the system of josephson currents does not carry meissner magnetization .
it will be shown that the same interpretation accounts fairly well for the f.c .
results which , at first sight , are rather different for the samples a and b. sample a is a @xmath70 thick pellet with an approximately ellipsoidal shape of @xmath71 .
its calculated volume is @xmath72 and the demagnetizing field coefficient for the field parallel to the longitudinal axis is @xmath73 @xcite .
+ = fig .
[ ahyst ] displays the magnetic dipole moment of the sample cooled to @xmath74 in zero field and submitted to cycles @xmath75 for several values of @xmath76 up to @xmath77 . at the lowest increasing fields , the moment increases initially with a slope @xmath78 . above @xmath79
, it approaches a slope @xmath80 .
the remanent positive moment saturates for @xmath81 .
the calculated moment of the sample for perfect volume shielding in an homogeneous field is ( e.m.u .
system ) : @xmath82 owing to the error in the evaluation of the volume , this value is determined with an accuracy of only @xmath83 . nevertheless , it is in fair agreement with the slope @xmath84 in fig .
[ ahyst ] . on the other hand ,
the slope @xmath80 is about @xmath85 , a rather small value since the density ratio of the sample is about @xmath50 . at such low temperatures ( in comparison with @xmath86 ) , where the lower critical field of the grain s material is above @xmath87
, one would expect expulsion of the field by the grains with a penetration depth @xmath88 .
the expected value for the magnetization @xmath89 of the system of uncoupled grains system can be calculated as @xcite : @xmath90 where @xmath91 is the volume fraction of the superconductive material and @xmath92 is the demagnetizing field for the grains . for an estimate , we assume grains to be spherical ( @xmath93 ) and , using @xmath94 and @xmath95 , we find @xmath96 .
this value is considerably below the volume fraction of the sample filled by grains ( @xmath97 ) ; we assume that the difference is due to the intergrain penetration depth @xmath88 being comparable to the grain size @xmath98 and estimate an effective value of @xmath88 as @xmath99 taking an average size of @xmath100 for the grains , we obtain @xmath101 .
values reported for the mean penetration depth in la@xmath1sr@xmath2cuo@xmath3 are about @xmath102 @xcite .
the value found here is larger than the expected mean value for the homogeneous material , indicating that the grains are not monocrystalline .
this will be confirmed below by the results of field - cooling experiments .
= the shielding susceptibility is plotted in fig .
[ amdhvst ] , as a function of temperature and for several values of the ambient f.c .
the measurements were performed according to the following procedure : the sample was cooled in a field @xmath103 down to the working temperature and the moment was measured after waiting 300 sec ; then the field was increased by a small amount @xmath104 and the moment was measured again after waiting 300 sec .
the figure displays the experimental shielding susceptibility normalized to the value for total shielding , i.e. : @xmath105 the curves show the double step usually ascribed to the action of both intragrain currents and josephson intergrain currents @xcite . at high temperature , the onset of grains diamagnetism occurs at about @xmath106 . above @xmath107
, the response corresponds to the diamagnetism of the grains . at a fixed temperature ,
it is @xmath103 independent for @xmath108 , and decreases for increasing @xmath109 .
below @xmath107 , the onset of josephson currents manifests as a second step of the diamagnetic response .
this second step appears at a decreasing temperature as @xmath103 increases . at the lowest temperatures ,
the diamagnetic moment amounts to about @xmath110 of flux expulsion at @xmath111 and decreases with increasing @xmath103 . at @xmath112 ,
the flux expulsion saturates at a value slightly above @xmath113 which corresponds roughly to the level of @xmath85 determined above for the grains response .
the susceptibility in fig .
[ amdhvst ] contains the contributions of grains and josephson currents .
the contributions can be separated on the line of the work by dersh and blatter @xcite .
the induction in the sample is given by @xmath114 where @xmath115 and @xmath116 stand respectively for the magnetization of grains and of the josephson currents .
it should be noted that the magnetization due to macroscopic circulating currents in a superconductor is sample - size dependent , i.e. the corresponding susceptibility is not a local quantity . at the macroscopic scale of the circulating currents , the magnetization @xmath115
can always be written as @xmath117 , where @xmath118 is homogeneous over the sample . in
what follows , we consider quantities averaged over the volume of the sample : in that case , @xmath116 is the averaged moment per volume unit due to the currents .
the demagnetizing field effect will be neglected in the calculations .
we have verified that , owing to the small value of the demagnetizing factor , this does not modify the essential features of the result while allowing a simpler derivation ( the effect of demagnetizing factor will be taken into account when analyzing the data from the sample b ) .
we get @xmath119 then @xmath120 ( with @xmath121 ) , and @xmath122 eq .
( [ eq - m_j ] ) must be considered with care since @xmath123 is history and field dependent .
in fact it is well - adapted to the description of the result of zero ( or small ) field cooling experiments .
more generally , we must consider the response to field increments @xmath124 to obtain @xmath125 then , the polarizability @xcite @xmath126 of the josephson network reads : @xmath127 note that we can equivalently consider the response of the currents system in an homogeneous medium with permeability @xmath128 .
if the applied field is varied by @xmath124 , the josephson network sees a variation of internal field @xmath129 and develops a polarization @xmath130 .
then , we recover eq .
( [ eq - chi_j ] ) . the value of @xmath123 could be determined in principle if we were able to obtain a packing of disconnected grains equivalent to the packing of the sintered sample . in practice
this was not possible .
indeed , mechanical grinding resulted in breaking a large part of the grains and thus modifying the characteristics of the material . nevertheless , it is possible to extract @xmath123 , at least approximately , from the data of fig .
[ amdhvst ] . at high temperature , above the onset of intergrain currents at @xmath131 ,
the shielding susceptibility @xmath132 is due to the grains alone , independent on @xmath103 below @xmath133 . at low temperatures , for @xmath103 above @xmath133
, the @xmath132 curves superpose and there is no manifestation of the onset of intergrain currents .
thus , here also , @xmath132 represents the response of the grains alone .
hence , the response @xmath123 of the grains can be reasonnably approximated by an interpolation between these two limits .
the interpolation curve , obtained by a smoothing procedure between both curves at @xmath134 and @xmath135 is displayed on fig .
[ amdhvst ] ( dotted curve ) .
the values of @xmath126 derived from eq .
( [ eq - chi_j ] ) are plotted versus temperature in fig .
[ akhijdc ] , for @xmath136 .
= note that the dependence of @xmath126 on @xmath137 seen in the figure is supposed to reflect the behavior of the initial shielding properties of the josephson network with the increase of @xmath103 .
nevertheless , non linearity of the response due to the correlative increase of the value of @xmath138 ( @xmath139 ) can not be excluded : this aspect will be studied in detail in sample b. finally , one can note the similarity of our data with the results of earlier numeric simulations on a gauge glass system @xcite .
above we have dicussed the system s responses to the variation of magnetic field at fixed temperature ( i.e. shielding responses ) and extracted from these data the polarizability @xmath126 of the intergrain system .
now we turn to the description of the results of the field cooling ( f.c . )
( meissner ) magnetization was measured by the standard procedure between @xmath140 and @xmath141 for fields from @xmath142 to @xmath143 .
the results are reported in fig .
[ amfcvst ] versus temperature and fig .
[ amfcvsh ] versus applied field .
data are normalized to the value of the moment for @xmath110 shielding .
= = even at the smallest field , the flux expulsion rate is no more than @xmath144 , less than the @xmath85 shielding by the grains . at low fields , below @xmath145 , there is an approximate affinity between the curves of @xmath146 versus @xmath147 .
@xmath146 can be extrapolated linearly to @xmath148 .
the result is plotted in fig .
[ amdhvst ] ( solid circles ) : one can see that the extrapolated f.c .
susceptibility superposes exactly with the low d.c .
field shielding susceptibility above @xmath107 .
therefore , at low d.c .
field above @xmath107 , the response of the grains system is reversible and it is well described by the low d.c .
field shielding curves ; this justifies the hypothesis used above for the calculation of @xmath149 . on the other hand ( see fig . [ amfcvsh ] ) , the behavior of the f.c .
susceptibility as a function of @xmath150 is not trivial .
@xmath146 decreases with increasing field and reaches a stable level ( about @xmath151 at the lowest temperatures ) at roughly @xmath145 .
whatever the temperature , this decrease is centered at a constant value of the field , about @xmath152 . above @xmath153 ,
@xmath146 decreases once more with increasing field .
note an essential difference between the f.c .
results presented on fig .
[ amfcvst ] and the shielding results above ( fig .
[ amdhvst ] ) : the f.c .
curves do not show any increase of the response @xmath146 with the temperature decrease below 20k , where the intergrain coupling grows considerably ( as it is seen from fig .
[ amdhvst ] ) .
this means that the network of intergrain currents does not produce meissner ( f.c . )
magnetization , whereas it does produce _ shielding _ magnetization . the behavior of the f.c .
susceptibility @xmath154 as a function of the applied field @xmath150 depicted in fig .
[ amfcvsh ] shows two nontrivial features : i ) crossover between two plateaus ( at low and moderate values of @xmath150 ) , which takes place between @xmath65 and @xmath155 independent of temperature , and ii ) the value of the low - field @xmath156 is noticeably lower than the meissner response of uncoupled grains ( @xmath144 versus @xmath85 ) .
these features can be understood in terms of ( i ) , a polycrystalline structure of the grains , which can be suspected from the large values of the penetration depth obtained from the results of fig .
[ ahyst ] , and ( ii ) , self - shielding ( pinning of the magnetic flux ) by the josephson currents when lowering the temperature in an applied field .
we start from the feature i ) ; the curves of f.c .
magnetization in fig .
[ amfcvsh ] are rather similar to those which were measured by ruppel et al .
@xcite in ybacuo ceramics .
the authors interpreted their results on the basis of a theory of the flux expulsion by strongly anisotopic randomly oriented crystallites as derived by wohllebeen et al .
we stress that the model is not based on any activated flux creep mechanism .
it is thus well - adapted to the analysis of our results : indeed , flux creep effects can hardly be invoked here since the temperature has no apparent effect of on the characteristic field related to the decrease of magnetization .
the starting point of the model is that , provided the size @xmath157 of the crystallites is such that @xmath158 , the longitudinal magnetization of a crystallite whose c - axis makes an angle @xmath29 with the field is given by @xmath159 , where @xmath160 is a factor close to 1 , depending on the ratio @xmath161 .
after averaging over @xmath29 , one obtains @xmath162 .
it must be stressed that the system of intragrain crystallites is a strongly - coupled system , contrary to the system of grains which composes the ceramic .
therefore , a grain consists of an ensemble of interconnected josephson loops surrounding crystallites whose planes are nearly along the field and are thus transparent to the field . at low fields ,
this system will expel the flux with a penetration depth depending on the junction coupling energy . nevertheless , when the field is such that a loop sees a flux larger than @xmath163 , the macroscopic magnetization of the josephson currents vanishes and the system reacts as an ensemble of disconnected crystallites @xcite .
the characteristic field of this crossover is such that @xcite : @xmath164 recently determined values for the penetration depth in la@xmath1sr@xmath2cuo@xmath3 @xcite are @xmath165 and @xmath166 .
older measurements indicate a higher anisotropy , up to a factor 14 @xcite .
we can thus reasonnably consider that the model can be applied in our case .
taking @xmath167 , we obtain @xmath168 . with @xmath169
this leads to a mean diameter @xmath170 for the crystallites . above @xmath171 ,
the system acts as an ensemble of crystallites whose average susceptibility is @xmath172 . with the density ratio @xmath173 , taking @xmath174 and supposing spherical crystallites we obtain from eq .
( [ eq - yaron ] ) @xmath175 which is above the experimental value ( the latter being about @xmath176 ) .
nevertheless , it must be noted that we have neglected here the effect of the factor @xmath160 and used a rather unrealistic spherical approximation for the shape of crystallites .
finally , it has been seen that above @xmath177 , the f.c .
magnetization starts to decrease once more with increasing field although @xmath178 is larger than @xmath179 in la@xmath1sr@xmath2cuo@xmath3 .
this can be due to intrinsic pinning inside the crystallites themselves when the applied field is such that the flux in the cross section of one crystallite is larger than @xmath180 . with a mean radius of @xmath181 for the crystallites
, this crossover occurs at about @xmath182 .
now we turn to the discussion of the feature ii ) mentioned above . at temperatures below @xmath183 ,
the josephson currents become active .
their effect is that , at @xmath184 , the shielding response of the system of grains amounts at about @xmath85 , while the f.c .
susceptibility saturates at about @xmath144 .
this difference is enough to be significant and can be interpreted as the result of pinning by the josephson network .
in fact , this pinning can be understood as a back shielding effect of the josephson currents against the decrease of local internal field , due to the temperature dependence of the grain s system permeability @xmath128 .
we have seen above that the response of the system consists of the two parts : ( i ) for an applied field @xmath150 , the internal field due to the grains seen by the intergrain currents is @xmath185 , and ( ii ) the intergrain currents system reacts to all variation of @xmath186 with a polarizability @xmath126 and generates a magnetization @xmath187 .
thus , when the temperature is decreased by @xmath188 , the internal field decreases by @xmath189 and the josephson network tends to screen this variation . since the intergrain currents give no meissner effect , we consider their response as totally irreversible .
thus for a variation @xmath188 of the temperature , in a field @xmath190 , the variation of induction is : @xmath191 on the other hand , @xmath192 . with @xmath193
, we finally obtain : @xmath194 @xmath195 is the magnetization produced by the josephson currents due to variation of @xmath128 with decreasing temperature . as @xmath123 is known only in the limit @xmath196 , eq .
( [ eq - chi_fc ] ) has been used to calculate @xmath156 versus @xmath147 in the limit of low field . in order to do it , we started from the values of @xmath197 as derived above ; for @xmath126
, we have used the values given in fig .
[ akhijdc ] for the smallest field @xmath198 .
the result is plotted on fig .
[ afccalc ] .
the agreement of calculated values with experimental data is rather satisfactory , although not perfect .
= this discrepancy is emphasized if eq .
( [ eq - chi_fc ] ) is reversed in order to calculate @xmath126 as a function of @xmath123 and @xmath199 .
the reason is that we have used here the simplest linear model of back shielding .
in fact , as we will see later , the response of the currents system is strongly non - linear , with the susceptibility @xmath126 decreasing with increasing @xmath200 , and this effect becomes stronger as the temperature increases .
the result is that the calculated efficiency of back shielding is underestimated , since the value of the experimental susceptibility is determined by applying finite increments @xmath200 .
sample b was machined from one of the original cylinders , in form of a parallelepiped of dimensions approximately @xmath201 .
its calculated volume is @xmath202 and its demagnetizing field coefficient for a longitudinal field is @xmath203 . in a longitudinal field ,
its calculated moment for perfect flux expulsion is given by @xmath204 measurements of the initial magnetization at @xmath74 are in fair agreement with this value .
for @xmath103 above @xmath205 and up to @xmath206 @xmath207 reaches a stable level about @xmath208 which corresponds to the response of the grains alone . with the density ratio of @xmath52 for this sample and using eq .
( [ eq - yaron ] ) one finds @xmath209 , yielding @xmath210 , i.e. the same value as derived for sample a. the shielding susceptibility was measured in this sample by using a more sophisticated method , in order to reduce the effect of non linearity . after cooling the sample at the working temperature in the d.c .
field , the field was increased by 5 successive steps @xmath200 , and @xmath211 was measured . at the lowest fields , @xmath212 and ( to keep a good signal / noise ratio ) @xmath213 at the highest ones .
then , the value of @xmath214 was extrapolated to @xmath215 by least square fit .
like in the case of sample a , all curves at @xmath216 merge at high temperatures to a common curve which corresponds to the flux expulsion by the grains .
the main difference with the sample a is that in the sample b the onset of josephson currents shielding occurs at higher temperatures .
this is consistent with the fact that sample b is more dense , resulting in a better coupling between grains ; moreover , its size is larger , which also increases the total shielding magnetization . at low temperature ,
the magnetization curve at @xmath135 reaches a level slightly above @xmath217 , which corresponds to the low temperature level for the grains .
the shielding response of the josephson currents is obtained with the procedure already used for the sample a. here the demagnetizing factor can not be neglected ( @xmath203 ) .
two kind of quantities are to be considered : ( i ) the responses @xmath123 and @xmath218 of an equivalent sample without demagnetizing field ( e.g. an infinitely long cylinder with the same cross - section for instance ) ; here @xmath123 is the response of the system of grains alone , without intergrain currents , and @xmath218 is the total response of the system of intragrain plus intergrain currents , and ( ii ) , the measured responses @xmath219 and @xmath220 ; they correspond to the measured moment for each case , normalized to the moment for total flux expulsion in the volume of the sample .
the relation between both kinds of quantities is given by @xmath221 .
a relation of the same kind holds for @xmath123 and @xmath222 . with the use of eq .
( [ eq - chi_j ] ) , we finally obtain : @xmath223 where @xmath224 , @xmath225 .
similar to the case of the sample a , an approximate curve has been determined for @xmath219 by interpolating between the small @xmath103 curves at high temperatures , and the curve at @xmath135 at low temperatures .
then the values of @xmath126 have been derived from eq .
( [ eq - chi_j2 ] ) and plotted on fig .
[ bkhijic ] .
the set of curves is similar to the set for sample a , except for the higher onset temperature of the intergrains currents .
= field cooled magnetization data , normalized to the value of the moment for full flux expulsion , are reported in fig .
[ bmeissvsh ] as a function of fields up to @xmath206 .
= at the lowest field and temperature , the f.c .
magnetization does not exceed @xmath226 of its value for full flux expulsion .
furthermore , at low temperatures the curves versus field present a second maximum at about @xmath177 . we expect that this complicated behavior is due to the back shielding effect of the intergrain currents , as discussed for sample a. to take them into account , a relation similar to eq .
( [ eq - chi_fc ] ) ( but with the demagnetizing effect taken into account ) should be derived .
the internal field is given as usual by @xmath227 , and the value of the local field seen by the currents is @xmath228 .
thus , under a temperature variation @xmath229 , @xmath230 with @xmath231 , and using the relation @xmath232 one obtains after integration : @xmath233 here @xmath234 , with @xmath126 reported on fig
. [ bkhijic ] , whereas the value of @xmath128 was obtained using the relation @xmath235 from the value of @xmath219 as derived above .
the values of @xmath236 for @xmath237 have been calculated using the values of @xmath219 as determined above , and the values of @xmath126 at @xmath111 .
the calculated value of @xmath236 was found to be about @xmath238 at @xmath63 , whereas its measured value was about @xmath239 .
the discrepancy between measured and calculated values is larger here than in corresponding results for sample a. we believe that the origin of this discrepancy is the same as in the case of sample a , i.e. it stems from the nonlinear response effect .
this effect is numerically larger in sample b since here the onset of josephson currents occurs in a range of temperature where @xmath123 still varies strongly , contrary to the case of sample a. the above analysis shows ( irrespectively to the above - mentioned discrepancy ) that the back shielding effect leads to a strong reduction of the field cooled susceptibility as compared with the susceptibility of the grains alone .
it is then easy to understand the complex behavior of @xmath236 as a funtion of field : at @xmath74 for instance , the onset of back shielding occurs at about @xmath240 , and its amplitude increases with decreasing field due to the increase of @xmath241 .
starting from the two - step behavior of @xmath123 expected from the theory of wohllebeen et al .
@xcite ( and seen in the data of sample a , where back shielding is less important ) , back shielding results on the double maximum shape of the measured curves .
in this subsection we will present the procedure we used to extract the value of the macroscopic critical current in our sample b. this procedure is not quite trivial since we are interested in the dependence of the critical current on the background d.c .
field in the sample , so we need to analyse the magnetization curves which depend both on the cooling field @xmath242 and on the field variation @xmath124 .
magnetization has been recorded at 10 and @xmath243 as a function of increasing @xmath138 with the smallest possible field steps ( @xmath244 ) , and starting from several f.c .
states . from the @xmath245 data , it is possible to derive the value of the current response @xmath246 as a function of @xmath138 .
for this , we use eq .
( [ eq - chi_j2 ] ) which can be written as : @xmath247 where @xmath248 is the magnetization of the grains alone ; @xmath249 and @xmath250 are defined in section ii - b .
the value of the grains system response is approximately derived in the same section : @xmath251 at @xmath74 and @xmath252 at @xmath243 . calculated values of @xmath246 at @xmath74 are plotted in fig . [ minij_10 ] .
note the analogy of these results to the magnetization curves of classical type ii superconductors with strong pinning ( the difference is that here @xmath138 plays the role of @xmath150 ) .
= after cooling the sample at zero d.c .
field , its response is obviously symmetric with respect to @xmath138 . when it is cooled in a finite d.c .
field , it is not the case anymore , as was explained in the previous section .
the magnetic moment of the sample just after cooling is @xmath253 where @xmath254 is the positive moment due to the back shielding by the josephson currents which have been developed during the cooling process ( see eq .
( [ eq - chi_j ] and [ eq - chi_j2 ] ) ) .
so , the total moment produced by the intergrain currents after increasing the field by @xmath138 is @xmath255 .
it is this moment which vanishes when @xmath256 ( at large enough @xmath138 ) , and thus @xmath246 approaches @xmath257 . in fig .
[ mpm2l ] we show the data recorded at @xmath258 and @xmath259 .
curves recorded at positive and negative @xmath138 both converge to the value corresponding to @xmath257 : at @xmath184 , @xmath257 is about @xmath260 .
= when @xmath261 , it is natural to expect that the response of the josephson network does not depend on the initial state .
a simple illustration can be given by analogy with bean - like pinning in type ii superconductors @xcite . at large @xmath200 , when the induction profile has penetrated up to the center of the sample , the magnetization no longer depends on @xmath138 but only on @xmath262 . if , as it is the case in real materials , @xmath262 varies with the induction in the sample , the magnetization depends on the total @xmath190 , whatever the value of @xmath103 in which the sample was cooled .
actually , when plotted as a function of the total field @xmath263 , the curves giving the total moment of network currents @xmath264 merge in their `` large '' field part ( i.e. above their maximum ) .
the values have been calculated , with @xmath265 and @xmath266 for @xmath267 and @xmath268 respectively . in order to obtain an optimal overlap between the curves ,
the following values have been used for @xmath269 : @xmath270 at @xmath271 , @xmath272 at @xmath267 and @xmath268 .
indeed , the calculated values for @xmath246 at large @xmath138 are extremely sensitive to those for @xmath269 .
this allows us to refine the determination of @xmath269 .
note that the values quoted above do not differ by more than @xmath273 , which is compatible with experimental accuracy and the possible variations of grains response with @xmath274 . finally , from the knowledge of the true josephson shielding response , in `` large '' fields we can now derive a rough evaluation of the critical current .
namely , above the maximum of @xmath275 , we calculate the value @xmath276 of the average critical current which would give the value of the measured moment by use of the bean formula @xcite in a cylindrical geometry . for strong penetration ,
the magnetization is given in e.m.u .
by @xmath277 . with @xmath278 and
the values of the moment measured at @xmath74 and @xmath243 with @xmath271 , we obtain the data displayed in fig .
[ jcavg ] .
note that the data are limited to fields such that @xmath279 below which the above approximate evaluation is no longer relevant .
= we can now concentrate on the behavior of the josephson currents moment at small @xmath200 .
for this discussion , the currents susceptibility @xmath280 is plotted versus @xmath138 at @xmath74 and @xmath243 in fig .
[ xinis_10 ] and [ xinis_20 ] , respectively .
= = at @xmath184 , after zero field cooling or cooling in a small field @xmath281 , the response varies linearly with @xmath138 for small values of @xmath138 up to about @xmath282 .
this linear slope of @xmath280 is considered as the result of classical bean - like pinning with critical current density @xmath283 , where @xmath284 is the initial slope of the curve @xcite .
this initial slope is reported on the figure as the short dashed line which corresponds to @xmath285 , leading to @xmath286 . at larger
@xmath200 , the behaviour of currents susceptibility @xmath280 deviates from linear , which is the result of both the magnetic - field dependence of the critical current @xmath287 , an intrinsic effect , and of the increasing degree of flux penetration into the sample , a purely size - dependent effect .
usually one uses the bean model ( generally with some @xmath288-dependent critical current ) in an appropriate geometry in order to deconvolute these two effects .
however , one should keep in mind that the bean model is a severe simplification of the problem of constant pinning force , corresponding to the limit @xmath289 ( i.e. the london penetration depth is supposed to be negligible with respect to the bean penetration length ) . for the simplest sample shapes ( thin slab or cylinder )
it means that the condition @xmath290 should be fulfilled , which is usually the case .
however the situation is more complicated for samples of square cross - section ( like our one ) , where the effect of corners may become important even at @xmath290 .
for such a geometry , the use of bean model leads to exactly the same relation between critical current , external field and measured magnetization as for the cylindrical ones , whereas one expects some difference if finite-@xmath88 corrections are taken into account . at the present stage
, we are not able to evaluate these corrections and therefore the values of the magnetization corresponding to our experimental geometry with non negligible @xmath291 .
nevertheless , we expect that it lies between the curves for two extreme limits .
the upper one corresponds to the @xmath292 limit , where the magnetization is given simply by the bean s formula for the cylinder : @xmath293 for @xmath294 and @xmath295 for @xmath296 .
a lower limit ( thought rather artificial ) consists of the `` double slab '' case in which the variation of magnetization is counted twice ( once for each pair of edges ) : @xmath297 for @xmath298 and @xmath299 for @xmath300 .
both curves are plotted in the fig .
[ xinis_10 ] ( dot dashed and long dashed curves respectively ) for @xmath301 and @xmath302 at @xmath303 .
let us now discuss the data starting from those obtained for low d.c .
fields , @xmath304 .
one can see that , after the initial linear part , the absolute value of the measured susceptibility is always smaller than the calculated ones .
this corresponds to the decrease of @xmath262 with increasing induction , as it is classically expected in granular materials , due to the suppression of intergrain critical currents by magnetic field penetration into the josephson junctions@xcite .
this `` classical '' behavior for granular superconductors is usually analyzed by considering the volume - averaged josephson medium as a kind of type ii superconductor in the dirty limit , provided its macroscopic penetration depth @xmath305 is large as compared with the grains size @xcite . at @xmath306
the behavior of @xmath280 is quite different : there is no initial linear slope , but a monotonic curvature is present down to the smallest @xmath200 .
it is no longer possible to adjust a bean like curve to the data .
for instance , the bean curve plotted on the lowest @xmath138 data for @xmath267 is reported on the fig .
[ xinis_10 ] as a dashed line .
it corresponds to a very small critical current of order @xmath307 , and it is evident that the effective screening current becomes much larger with increasing @xmath200 . here
, contrary to the case of @xmath308 , the absolute value of the measured susceptibility is always _ larger _ than the calculated one for a constant shielding current corresponding to the limit @xmath309 .
this means that , whereas at @xmath111 the effective screening current density stays constant and then slowly decreases with increasing @xmath200 ( which corresponds to classical josephson pinning ) , at @xmath306 it _ increases _ with @xmath200 _ sublinearly _ ( since a linear increase would correspond to a susceptibility independent of @xmath138 ) .
such behaviour is quite unusual within the commonly accepted picture of screening in superconductors ; indeed , we know that , for vanishing field excitations , the screening current may be either i ) _ linear _ in @xmath138 and _ reversible _ , as in the london ( or campbell @xcite ) shielding regime , or ii ) _ constant _ ( equal to the initial critical current @xmath262 ) and _ irreversible _ as in the case of the bean - type critical state ( or of any other known critical model , e.g kim model @xcite , exponential model @xcite , etc ) .
the above anomalous screening behaviour is even more pronounced at @xmath243 where , even after zero field cooling , no initial linear slope of @xmath280 can be seen in the data .
all curves show the same anomalous behavior as the data at @xmath74 in fields from @xmath310 .
this specific behavior is emphasized by plotting the difference between the measured susceptibility @xmath280 and its value for total flux expulsion @xmath311 , versus @xmath138 on a log log scale .
in such a plot , at least in the regime of weak penetration , i.e. where @xmath280 is larger than @xmath312 , sublinear variation of the shielding current density results in a logarithmic slope smaller than 1 for the curves of @xmath313 ( for @xmath280 smaller than @xmath312 , we are in a regime of strong penetration where it is no longer possible to relate simply the variations of the moment response to those of the shielding current ) . in fig .
[ dx2_1020 ] , we have reported the three curves for which data are found in the range above @xmath312 , i.e. at @xmath314 and @xmath310 , and @xmath315 . = at @xmath74 and @xmath316 , the logarithmic slope is about 1 as expected , although at the smallest fields the curve crosses over to a smaller logarithmic slope closer to @xmath317 . at @xmath243 and @xmath316
the logarithmic slope is about @xmath318 at the lowest @xmath200 .
approximately the same value of the slope characterizes the data obtained at @xmath74 and @xmath319 , although the dispersion of data points at lowest @xmath200 makes its accurate determination difficult .
the above anomalous behaviour makes it tempting to try a simple _ ansatz _ for the behavior of the response current density of the system versus induction variations .
let us suppose that @xmath320 with @xmath29 between @xmath38 and @xmath321 .
the case with @xmath322 corresponds simply to classical screening with penetration length @xmath88 ( since @xmath323 ) ; the case with @xmath324 corresponds to constant @xmath325 , i.e. the classical bean case .
anomalous response arises for non integer @xmath326 .
for very small excitation @xmath200 , the length of induction penetration is small as compared with the size of the sample and we need to consider the effect of the excitation in the lowest order in @xmath327 only . for the purpose of illustration
we consider the simplest slab geometry .
then the induction profile is determined by the maxwell equation @xmath328 where @xmath329 is the coordinate perpendicular to the edge of the sample .
for an external field @xmath200 , the induction in the sample is given by @xmath330 where @xmath331 is the coordinate of penetration and @xmath332 and @xmath333 are normalizing factors ; for @xmath334 , @xmath335 , i.e. @xmath336 .
then , integrating the field profile ( [ dbx ] ) over @xmath329 , we get @xmath337 where @xmath338 is the mean magnetization variation due to the field variation @xmath138 . if we now compare the result ( [ dmdh ] ) with the data shown in fig .
[ dx2_1020 ] , we find values of @xmath29 in the range @xmath339 at both @xmath74 and @xmath243 .
thus a simple choice for the relation between the screening current @xmath262 and the induction variation @xmath327 allows us to imitate the experimental results for the simplest protocol of a weak monotonic @xmath138 variation on top of a homogeneous state of the network .
nevertheless , it is evident that @xmath327 has no clear meaning if the variation of @xmath150 is non - monotonic or if the initial state is obtained by non - zero field cooling .
indeed , in the later case , induction in the sample varies during cooling due to the variation of @xmath128 with @xmath340 , giving the response @xmath341 as seen before .
furthermore , we will see below that the response is irreversible even for extremely low exitation fields .
problems of sensitivity limit the range of small excitations which can be used in d.c . experiments .
the preceding results clearly show the sublinear nature of the low field response , but they do not allow its precise determination . in order to extend by several orders of magnitude the range of our lower excitations investigation , we have been led to perform a.c .
susceptibility measurements .
the use of a.c .
response measurements is always questionnable when equilibrium ( or quasi - equilibrium ) properties are under investigation , since the results can be affected by the time - dependent part of the response function .
it has been shown that the latter is the response of a very good conductor with complex conductivity @xcite .
hence , it is necessary to work at low frequency , in a range where the susceptibility is roughly frequency independent .
we present here preliminary results obtained on a long cylinder obtained by stacking several of the original sample b cylinders .
measurements were done at @xmath243 , at a working frequency of @xmath342 in the equipment used for noise experiments @xcite .
the sample was simply shifted into the upper half of the third order gradiometer . at this temperature and frequency
, we have verified that the in - phase susceptibility is almost frequency independent , which ensures that the results are mainly dependent on the ( quasi ) static part of the response .
the susceptibility was recorded using classical method of squid magnetometry .
we used a.c .
excitation fields in the range @xmath343@xmath344 and the sample was cooled in d.c .
fields from @xmath38 to @xmath345 . from the data ,
the values of the josephson network susceptibility was extacted using the method developed in section ii , with the susceptibilities in eq .
( [ eq - chi_j2 ] ) being complex quantities .
the susceptibility measured at @xmath346 at the lowest a.c .
amplitude was taken as the level for perfect diamagnetism .
[ xsjvslh ] displays a log - log plot of the out - of - phase susceptibility @xmath347 versus the amplitude of the a.c .
field , and for several values of the f.c .
static field .
the response is irreversible down to the lowest a.c amplitudes , and the irreversibility increases with the superimposed d.c
. field .
= = all curves follow a power law , with the same exponent close to @xmath348 .
going towards the smallest a.c .
excitations , they show some downward bend which could be related with the approach to a linear regime ( with @xmath349 ) below @xmath350 , although the dispersion of the data is too high to conclude . the in - phase susceptibility @xmath351 is plotted as a function of @xmath352 in fig .
[ xpjvsla ] . here
as well , the anomalous nature of the response is clearly seen .
@xmath353 behaves like @xmath354 where both the constant @xmath355 and the slope @xmath160 increase with increasing superimposed static field @xmath274 .
the dependence of the harmonic susceptibility on the a.c .
field amplitude is a genuine proof of the existence of static irreversibility in the response .
this is not astonishing by itself , but these results stress the anomalous aspect of this irreversibility .
for instance , in the classical bean case with a weak penetration , it is known that @xmath356 and @xmath347 are proportional to @xmath357 whereas fig .
[ xsjvslh ] and [ xpjvsla ] clearly show the proportionality to @xmath352 . a further evidence is provided by plotting @xmath347 versus @xmath356 as displayed in figure [ dxpvsxs ] .
it can be shown that if the a.c .
response is driven only by static irreversibility , both are proportional . in the bean case ,
the coefficient of proportionality is @xmath358 . in the figure [ dxpvsxs ] ,
the part of data which lie in the range of @xmath359 variation of @xmath351 ( where the relations for slab geometry are approximately valid ) show that @xmath347 is indeed proportional to @xmath356 , but with a bit smaller coefficient @xmath360 . = in order to understand the meaning of the above results , we generalize the crude _ ad hoc _ model of section iii - b to the irreversible case . in order to do it
, we generalize the protocol of the bean model .
namely , in the bean model , the current is given by a step function of the variation of induction , @xmath361 according to the sign of @xmath327 , as long as the induction variation is monotonous .
if the sign of variation of @xmath288 is reversed , @xmath325 also changes sign , which can be written in terms of the _ variation _ of the current density ( with respect to the initial current distribution obtained after monotonous variation of the field , @xmath362 , @xmath363 where @xmath364 and @xmath365 .
such a representation ( which is not needed in the analysis of the bean model itself ) will allow us to construct the necessary generalization of the relation between current and variation of the field used in eq .
( [ dbdx ] ) .
actually our goal here is rather limited : we are going to find a consistent description of the simplest hysteresis cycle which consists of the initial increase of @xmath327 from zero to @xmath366 , then reversing the sign of the field variation until the value of @xmath367 is reached , and then reversing @xmath368 once more and finishing at @xmath369 .
the description of this cycle will be consistent if we find that the value of the current density at the end - point , @xmath370 , coincides with the one after the original increase of the field @xmath366 , @xmath362 .
this simply means that the hysteresis loop is closed .
it is easy to check that the above condition will be fulfilled by the following choice of the @xmath371 dependence : @xmath372 where @xmath332 and @xmath333 have the same meaning as in eq.([dbdx ] ) .
actually the only difference between the eq .
( [ jbnew ] ) and the original used in the eq.([dbdx ] ) is the coefficient @xmath373 .
the bean model limit then corresponds to @xmath374 , so the above coefficient approaches 2 as it should .
then instead of eq .
( [ dbdx ] ) we obtain : @xmath375 where @xmath376 .
the induction profile , induced magnetization and harmonic response are calculated in the appendix .
the main conclusions are that the fundamental components @xmath356 and @xmath347 are both proportional to @xmath377 , and that their ratio @xmath378 decreases from @xmath379 to @xmath38 when @xmath29 goes from 0 to 1 . for @xmath380 ,
we get ( cf . fig . [ pentxsxp ] ) @xmath381 , a value which is in good agreement with the data presented on fig .
[ dxpvsxs ] .
note that the degree of irreversibility ( measured by this ratio ) is similar ( although a bit lower ) to the one of the bean model .
it should be emphasized that the numerical coefficient in eq .
( [ jbnew ] ) was `` fitted '' in order to obtain consistent ( i.e. closed ) hysteresis loop ; one can expect that an analogous equation describing current variation after some more complicated history of the field variations will contain another ( history - dependent ) numerical coefficient instead of @xmath373 .
it can be seen from figs .
[ xsjvslh ] and [ xpjvsla ] that @xmath382 increases with increasing ambient d.c . field .
it is natural to expect a decrease of @xmath332 with increasing @xmath274 .
the increase of @xmath383 with @xmath103 means that @xmath384 decreases more quickly than @xmath332 when @xmath103 increases .
the presence and behavior of the constant @xmath355 can not be predicted on the basis of the above simple model .
in fact , the latter neglects the possibility of elastic displacement of flux lines under the action of the external applied field .
such an effect would result in a response analogous to the campbell response due to the elastic displacement of vortices in their pinning potential in type ii superconductors @xcite .
campbell s response is linear , so it would result in an a.c .
field independent positive contribution to @xmath385 , whose amplitude would be inversely proportional to the strength of the restoring force of the pinning potential wells .
it is natural to expect that the pinning force decreases with increasing ambient d.c .
field in our granular system , due to the reduction of the junctions critical currents .
hence , such an effect would give a positive contribution to @xmath385 , which would increase with increasing d.c .
this corresponds rather well to the behavior of the offset @xmath355 seen in the data .
in this section we compare the experimental results described above with the theoretical results available for the randomly frustrated josephson networks .
we start from a simple estimate for the mean energy @xmath387 using the experimental value of the low - temperature , z.f.c .
( @xmath388 ) critical current density @xmath286 . using the estimate @xmath389 for the mean size of the grains , one could naively obtain @xmath390 and the corresponding low - temperature josephson energy @xmath391 ( this value was derived from @xmath262 measured at @xmath63 , but we do not expect much difference in the intrinsic josephson energies at @xmath392 and at @xmath393 since the bulk transition temperature in la@xmath1sr@xmath2cuo@xmath3 is @xmath394 ) .
however such an estimate is in contradiction with the measured value of the glass transition temperature @xmath395 .
indeed , let us assume that the mean coordination number ( number of `` interacting neigbours '' ) @xmath396 in the ceramics is around 6 , as for a simple cubic lattice .
then for the estimate of the relation between @xmath397 and @xmath40 one can use the simulation data @xcite which give @xmath398 , where we took into account the linear dependence of @xmath397 on @xmath399 close to the bulk transition temperature . as a result
, one gets @xmath400 i.e. a factor 30 larger than the naive estimate above
. however we will show now that this discrepancy may be resolved if we assume that the current network producing the measured critical current density @xmath262 was actually _ strongly frustrated _ in spite of the absence of background d.c .
field in this measurement .
the macroscopic critical current density @xmath262 for a strongly frustrated josephson network was calculated in ref .
@xcite within the mean - field approach ( we are not aware of any calculations of this kind beyond the scope of the mean - field theory ) . it was shown that frustration strongly reduces @xmath262 as compared to its value @xmath401 for an unfrustrated system , @xmath402 , where the factor @xmath403 was obtained by numerical solution of the _ slow cooling _
equations @xcite describing the evolution of the glassy state under slow variations of temperature and magnetic field . in the low - temperature limit
, this relation amounts to a factor 25 reduction of the @xmath262 value with respect to @xmath401 .
correspondingly , the characteristic value of the critical current for an individual junction will be obtained as @xmath404 and results in a josephson coupling energy @xmath405 , in a fairly good agreement with the above estimate ( [ e_j ] ) .
the above estimates show that the network of josephson junctions in la@xmath1sr@xmath2cuo@xmath3 is frustrated even in the absence of an external magnetic field .
a careful reader could question this conclusion since we have used some results from the mean - field theory which may be a poor approximation for a 3d gauge - glass .
we believe , however that the qualitative result of the above estimates is sufficiently robust because a strong reduction of @xmath262 with respect to @xmath401 should be a general feature of a glassy network , so that unaccuracy due to mean - field approximation can not compensate for a huge discrepancy obtained between @xmath406 and the estimate ( [ e_j ] ) .
additional evidence in favor of the glassy nature of our system is provided by the similarity of the low-@xmath138 diamagnetic response at @xmath407 with zero as well as non - zero @xmath103 , as described in section iii above , as well as the low - frequency noise data obtained in ref .
@xcite on the same type of ceramics . what could be the origin of that frustration ?
we believe that most probably it is the result of the d - wave nature of superconductivity in cuprates @xcite and randomness of the crystalline orientations in ceramics @xcite .
it was shown there that the form of the effective phase - dependent hamiltonian for such ceramics is of the same form as in ( [ ham ] ) except for the fact that the random phases @xmath20 at @xmath67 are just @xmath38 or @xmath21 depending on the mutual orientation of grains @xmath408 and @xmath409 . therefore such a system at @xmath67 is equivalent to the @xmath39 spin glass , with the low - temperature state characterized by a completely random orientation of phases @xmath410 , as in the gauge - glass model with uniformly random distribution of @xmath20 s .
therefore the low - temperature state is characterized by the presence of randomly distributed intergrain currents and , therefore , of the magnetic field generated by these currents .
it means that the actual phases @xmath20 will contain contributions due to the self - induced magnetic field .
its relative importance is characterized by the ratio of the corresponding magnetic flux penetrating elementary loops of the ceramics @xmath411 to the flux quantum @xmath17 , i.e. just by the parameter @xmath412 where @xmath31 is the characteristic inductance of an elementary loop @xcite . estimating the elementary inductance as @xmath413 and using eq.([e_j ] ) to estimate @xmath414
, we obtain @xmath415 so the self - field ( screening ) effects are relatively weak , though perhaps not always negligible .
it is also of interest to estimate the effective penetration depth @xmath416 of a very weak magnetic field perturbation @xmath124 into the ceramics .
roughly , the value of @xmath416 can be estimated as @xmath417 .
another ( hopefully more accurate ) estimate can be obtained using mean - field results @xcite which allow one to express @xmath416 via the critical current density @xmath262 : @xmath418 where we inserted ( as compared with ref .
@xcite ) @xmath419 and approximated the random nearest - neigbour network by a cubic lattice with coordination number @xmath420 , which amounts to the relation @xmath421 between the effective interaction range @xmath422 and the intergrain distance @xmath19 .
the characteristic magnetic field variation producing the critical current density @xmath262 at the boundary can be estimated as @xmath423 , whereas the numerical solution @xcite gives @xmath424 within the theoretical approach of ref .
@xcite , @xmath425 marks a crossover between reversible ( although still non - linear at @xmath426 ) and irreversible penetration of the magnetic field into the intergrain network .
the value of @xmath425 obtained in eq.([h_c ] ) is on the lower border of the range of the field variations used to measure our d.c .
magnetization curves , so we could just conclude that we always have @xmath427 and thus are producing the bean - like critical state . indeed , the data at @xmath111 , @xmath428 look compatible with such an interpretation ( cf .
fig.([dx2_1020 ] ) , where some deviations from the logarithmic slope 1 ( which is the characteristic of a bean state ) are seen at lowest @xmath429 ) .
however , as far as the data obtained at @xmath74 with d.c .
fields @xmath430 , or all data at higher temperature ( @xmath407 ) , including d.c . and a.c
. results at zero-@xmath103 , are concerned ( cf .
figs.([xinis_10]-[xpjvsla ] ) ) , the low - field magnetization response is drastically different from bean - type predictions , as explained at the end of section iii .
qualitatively , the most surprising feature of these data is the existence of a very broad range of @xmath138 within which the response is non - linear but still not like the critical - state one .
we are not aware of any microscopic theory which predicts fractional - power behaviour of the shielding susceptibility over such a broad range of @xmath431 variations .
it can not be excluded _ a priori _ that such a behaviour is related to a very wide range of intergrain critical currents , which might exist in ceramics ( till now we have neglected inhomogenity of intergrain coupling strengths in our theoretical discussion ) .
moreover , we may expect that the relative importance of such inhomogenities increases with the field and/or temperature ( cf .
@xcite ) in section v , we will try to formulate a new phenomenological model appropriate for the understanding of our data ( leaving its theoretical justification for a future study ) ; this model will be seen to be an interpolation between campbell s and bean s regimes of flux penetration into hard superconductors .
we showed at the end of section iii - c that a simple generalization , eq.([dbdx2 ] ) , of bean s relation between variation of the applied magnetic induction @xmath327 and current @xmath325 results in reasonably good agreement with our data .
contrary to the original bean relation , the new one was not based on any physical picture ; it was just a convenient description of the data . in this section
we propose a phenomenological model which provides a qualitative understanding of the irreversible diamagnetic behaviour mimicked by eq.([dbdx2 ] ) .
we start from the picture of non - linear response of the current @xmath432 to a variation of the vector potential @xmath433 derived in ref .
@xcite within the mean - field approximation and presented in fig . 2 of that paper .
here the current induced by a variation of @xmath433 is linear at very small @xmath433 , then grows sublinearly , and finally reaches its maximum value @xmath262 at the critical @xmath434 such that the differential response @xmath435 . at @xmath436 the numerical instability of the _ slow cooling _
equations was detected and interpreted as an indication of the absence of any solution which would interpolate smoothly between zero and large ( i.e. @xmath437 ) values of @xmath438 .
in other terms , some kind of `` phase slip '' was expected to happen in the model @xcite , leading to a new metastable state , which would have lower ( free ) energy at the new value of the vector potential @xmath439 ( in other terms , a state similar to the one obtained by the f.c .
procedure at constant @xmath440 , which does not carry macroscopic current )
. further increase of @xmath441 again induces a macroscopic current until it reaches the maximum value @xmath262 at @xmath442 , and so on .
thus the whole @xmath443 dependence emerging from the mean - field solution @xcite is periodic ; it leads immediately to the irreversibility of the response , since the inverse function @xmath444 is multivalued : different vector potential values may correspond to the same value of current .
of course , such a periodic @xmath443 dependence does not correspond to the usual campbell - bean picture , which would better be represented by @xmath445 it is important to note that the @xmath443 dependence ref .
@xcite was obtained from the _ space - independent _
solution for the glassy correlation function @xmath446 ; such an approximation , being reasonable for the description of smooth `` adiabatic '' transformations in a system with long - range interactions , will probably break down when the jump from one metastable state to another happens .
in other terms , the above - mentioned `` phase slip '' should have something to do with spatially inhomogeneous processes like vortex penetration in hard type - ii superconductors .
the problem of the solution of the general history- and space - dependent system of integral equations ( which may be derived following the method of ref .
@xcite ) is formidable and the method to solve it is still unknown .
therefore we can only speculate on possible properties of its solution .
the simplest idea would be that the macroscopic @xmath443 response becomes ( after averaging over inhomogenities of the space - dependent solution ) similar to the campbell - bean type of the response ( [ jbean ] ) .
indeed , our analysis of the low - field diamagnetic response at @xmath63 and @xmath447 ( section iii - b ) developed in section iv on the basis of such an assumption , is in reasonable agreement with the data .
however other sets of data ( for higher temperature and/or lower field ) are described by completely different _ ansatz _ ( [ jbnew ] )
. we will now propose a ( phenomenological ) generalization of the @xmath443 relation compatible with eq.([jbnew ] ) .
the relation we are looking for should be an _
( i.e. independent on the sample geometry ) and _ general _ ( i.e. usable for an arbitrary magnetic history of the sample ) relation between the current and variation of the vector potential .
remember that eq .
( [ jbnew ] ) was written for the simplest nonmonotonic variation of @xmath327 , and that it relates the true vector @xmath448 and the pseudovector @xmath449 .
so , in writing this equation , some additional information on the geometry of the sample has been used ( we use the simplest slab geometry ) .
thus a natural basic equation should relate the current density @xmath448 and the variation of the vector potential @xmath450 . in a generalized model ,
the diamagnetic current response should possess two major properties : + i ) it must scale as some fractional power @xmath451 with the amplitude of exitation field @xmath452 , and + ii ) it must be strongly irreversible ( as it follows from the analysis of the ratio @xmath453 ) .
we consider these two conditions in sequence .
the condition i ) is rather easy to fulfill : it is enough to suppose that the differential response of the current to the variation of the vector potential @xmath438 is given by a non - linear generalization of the london relation @xmath454 where the current - dependent `` effective penetration depth '' is given by @xmath455 in the case of a monotonic field variation applied to an initially uniform induction distribution , the eqs.([londgen ] , [ lambdaef ] ) lead to the simple relation @xmath320 with @xmath456 . indeed , with @xmath457 and approximating @xmath458 by @xmath459 , one obtains @xmath460 .
thus we need to choose @xmath461 in order to reproduce the observed scaling with @xmath462 .
however , the set of equations ( [ londgen ] , [ lambdaef ] ) does not fulfill the second condition ii ) above : the corresponding solutions are _ reversible _ , as it follows from the existence of a single - valued function @xmath463 which follows from eqs.([londgen ] , [ lambdaef ] ) . in other words , the system described by eqs.([londgen ] , [ lambdaef ] ) would exhibit nonlinearity and harmonics generation , but would not show finite @xmath464 in the @xmath465 limit . in order to avoid this inconsistency , we need to formulate a model with the same kind of scaling between @xmath438 and @xmath325 as in eqs.([londgen ] , [ lambdaef ] ) , but with a nonmonotonic @xmath443 dependence allowing for the irreversible behaviour .
a model obeying very similar properties was formulated and studied in ref .
@xcite in a different physical context ( one - dimensional spin glass ) .
the low - energy spin configurations in this model are described by a phase variable @xmath466 such that two such configurations ( local energy minima ) which differ by a phase shift @xmath467 in a region around some point @xmath468 , have a characteristic energy difference @xmath469 and a characteristic spatial extent of the phase deformation @xmath470 .
this scaling holds for the intermediate range of phase deformations @xmath471 ; at smaller @xmath472 the energy cost of deformation is @xmath473 , whereas at @xmath474 the energy growth obviously saturates due to @xmath475 periodicity .
the above @xmath476 scaling leads to a _ sublinear growth _ of the characteristic `` force ''
@xmath477 with @xmath478 in the same intermediate range .
the main contribution to the second derivative @xmath479 ( curvature of the energy valleys ) comes from the smallest scale @xmath480 , i.e. from the curvature of individual local minima .
it was explained in ref .
@xcite that such a scaling means a fractal organization of the energy minima as a function of @xmath481 with fractal dimensionality @xmath482 .
it means that the number of energy minima discernable on a scale @xmath481 grows as @xmath483 at finer scales ; new minima appear primarily due to the splitting the older ( broader ) ones .
this picture emerged in ref .
@xcite from the microscopic analysis of the original hamiltonian for a one - dimensional spin - glass model formulated in ref
. @xcite .
we can borrow the qualitative features of this construction for our present purpose ( leaving for future studies the problem of its microscopic justification for the case of superconductive glasses ) .
suppose that the free energy @xmath484 of the josephson network behaves ( as a function of vector potential variations with respect to a `` virgin '' state with a homogeneous induction ) in a way similar to @xmath485 at @xmath486 .
namely , suppose that the free energy is parabolic , @xmath487 , at very small variations of vector potential @xmath488 , but on a larger scale , @xmath489 , it contains many local minima whose characteristic free energies scale ( with respect to the lowest state with @xmath490 ) as @xmath491 with the exponent @xmath492 ( see the definition of @xmath493 below ) .
then the characteristic value of the current @xmath494 scales as @xmath495 in the same interval of @xmath438 . at large @xmath496 variations ,
the growth of the induced current should saturate at the true critical current value @xmath262 , so we can estimate @xmath497 on the other hand , weak @xmath498 leads to the usual linear london ( or campbell ) response with an effective penetration depth @xmath499 ; matching at @xmath500 leads to the following estimate : @xmath501 the estimate ( [ j_a ] ) looks very much like the previous version defined by ( [ londgen ] , [ lambdaef ] ) , so one can find the relation between the exponents : @xmath502 = however the whole picture is substantially altered : the current is now supposed to be an ( irregularly ) oscillating function of @xmath438 ( see fig .
[ simulj ] ) , thus only its envelope @xmath503 defined on a scale @xmath438 follows the scaling relation ( [ j_a ] ) . as a result , the inverse function @xmath504 is multivalued and the irreversibility of the response is ensured .
similar to the spin - glass model of ref .
@xcite , the fractal dimensionality @xmath505 of the low - energy valleys can be defined ; it is given now by @xmath506 .
the proposed picture is based on the existence of two substantially different scales of currents , @xmath68 and @xmath262 , and corresponding vector potential variations @xmath507 and @xmath434 ; thus it can be compared with the usual campbell - bean picture of critical currents in the same way as the thermodynamics of type - ii superconductors is compared with that of the type - i ones . in order to describe quantitatively the diamagnetic response in the `` fractal '' range ( [ f_a ] ) we need to determine the _ distribution function _ @xmath508 $ ] ( which would lead , in particular , to the estimate ( [ j_a ] ) for @xmath503 ) . moreover , in general , a relation of the type of ( [ j_a ] ) could be nonlocal ( i.e. the current depends on the @xmath509 distribution in some region of space , whose size may depend on @xmath438 itself ( see again ref .
we leave this complicated problem for future studies , and just note here that merely the existence of relation ( [ j_a ] ) is sufficient for the existence of some `` natural '' properties of the response ( like the presence of a closed hysteresis loop , as it was assumed in section iii - c ) .
in this paper , we have presented experimental results on the low temperature ( @xmath74 and @xmath243 ) response of the granular ht@xmath0 superconductor la@xmath1sr@xmath2cuo@xmath3 to small field excitations .
the general properties of the magnetic response were investigated in two samples ( a and b ) differing by the strength of the coupling between grains . by cooling the samples in various d.c .
fields up to @xmath143 and applying small field increases , we were able to measure the shielding response of the material and to derive a method , inspired by the work of dersh and blatter @xcite , to extract from the data the polarizability of the intergrain currents system .
the field cooled ( f.c . ) magnetization was measured in fields up to @xmath240 .
analysis of the results led to the conclusion that i ) the structure of the grains is polycrystalline , resulting in a step decrease of the f.c .
magnetization with increasing field , which can be interpreted on the basis of the model by wohllebeen et al .
@xcite ; ii ) self shielding ( pinning ) by the intergrain currents when lowering the temperature strongly reduces the value of the f.c .
magnetization ; iii ) there is no macroscopic meissner magnetization due to the system of intergrain currents .
further detailed study of the response of the josephson network was performed in sample b. it was shown that the response is asymmetric with respect to the sign of variation of the applied field after field cooling ; this is due to the shielding currents pinned during cooling .
the macroscopic critical current is found to be strongly reduced by moderate values of the external d.c .
field , about @xmath310 .
very low field magnetization measurements were performed by applying field steps of @xmath510 or low frequency a.c .
fields in the range @xmath511 to @xmath344 , after cooling in d.c .
fields up to @xmath512 .
the results show that the response is strongly non linear , the shielding current growing _
sublinearly _ with increasing applied field .
furthermore , the a.c .
results show that it is strongly _
irreversible _ down to the smallest excitations used .
it is shown that a non - linear relation between the shielding current and the induction , @xmath513 with @xmath462 , together with a natural assumption about the existence of a closed hysteresis loop , give predictions in a reasonable agreement with the data .
theoretical analysis of our experimental results was developed on the frame of the existing gauge - glass " theories .
it was show that the extremely low value of the low - temperature , zero - field critical current density ( @xmath514 at @xmath74 ) together with the rather high temperature of the transition to the low - temperature glassy state , can be coherently interpreted only under the assumption that the josephson network is strongly frustrated even at zero applied field .
this contradicts the usual assumption that frustration in the interactions arises only due to the local magnetic induction , but supports the hypothesis of the existence of a large proportiont of @xmath21-junctions in the granular system .
these @xmath21-junctions are possibly due to the d - wave nature of the pairing , combined with the randomness of grain orientations in la@xmath1sr@xmath2cuo@xmath3 ceramics . finally , a new model of diamagnetic response in the glassy state of granular superconductors
was developed in order to describe the anomalous ( fractional - power ) behavior of the shielding current response .
this model , based on the idea of a fractal organization of the free energy landscape in the granular network , can provide a qualitative account for the main features of the anomalous response .
its further development will be the subject of future studies .
we are grateful to l. b. ioffe for many important discussions which helped to clarify a number of issues considered in this paper .
research of m.v.f . was supported by the dga grant # 94 - 1189 , by the joint grant m6m300 from the international science foundation and the russian goverment , and by the grant # 95 - 02 - 05720 from the russian foundation for fundamental research .
the hysteretic behavior of the current as a function of the induction variations is represented by the relation : @xmath515 @xmath516 when starting from zero induction state , and 1 otherwise .
@xmath517 and @xmath518 where @xmath519 and @xmath520 are the ( old ) values just before the last reversal of the sign of variation of @xmath521 .
the _ ansatz _
[ djmod ] ensures that we have a stable closed hysteresis loop , and that there is no hysteresis for @xmath522 which describes the london case .
figure [ jdeb ] displays the shape of the hysteresis loops for three characteristic values of @xmath326 .
the induction profile is determined by the maxwell equation which leads , for the case of weak penetration , to @xmath523 where @xmath524 ; @xmath329 is the distance from the edge of the sample . after increasing applied field from 0 to @xmath525 , starting from zero induction state , the induction profile is given by @xmath526 , leading to : @xmath527 where @xmath528 field penetrates till @xmath529 .
when @xmath530 decreases from @xmath525 , we get @xmath531 . hence : @xmath532 modification of induction relative to @xmath520 extends up to @xmath533 . for @xmath534 , @xmath535 where @xmath520 is given by eq .
( [ initho ] ) . when @xmath536 is reached , eq .
( [ reversh ] ) gives simply @xmath537 .
after reversing the sign of variation of @xmath530 once more , the profiles are simply symmetrical of those given by eq .
( [ reversh ] ) .
examples of induction profiles are given on fig .
[ bprofil ] for @xmath538 .
99 for a review , see g. blatter , m. v. feigelman , v. b. geshkenbein , a. i. larkin and v. m. vinokur , rev .
phys . * 66 * , 1125 ( 1994 ) s. john and t. c. lubensky , phys . rev . *
b34 * , 4815 ( 1986 ) .
v. m. vinokur , l. b. ioffe , a. i. larkin and m. v. feigelman , sov .
jetp * 6 * , 1986 ( 1987 ) .
v. s. dotsenko , m. v. feigelman and l.b .
ioffe in spin glasses and related problems , soviet scientific review /section a , chapter iv . m. v. feigelman and l. b. ioffe , phys . rev . lett . * 74 * , 3447 ( 1995 ) .
j. d. reger , t. a. tokuyasu , a. p.
young and m. p. a. fisher phys.rev .
* b44 * , 7147 ( 1991 ) .
d. huse and h. seung , phys.rev .
b 42 , 1059 ( 1990 ) . h. s. bokil and a. p. young phys .
lett . * 74 * , 3021 ( 1995 ) .
m. borik , m. chernikov , v. velasago and v. stepankin , j. low temp . phys . *
85 * , 283 ( 1991 ) .
d. s. fisher , m. p. a. fisher and d. a. huse , phys . rev .
* b43 * , 130 ( 1991 ) .
e. b. sonin , jetp lett .
* 47 * , 496 ( 1988 ) ; e. b. sonin and a. k. tagancev , sov.phys.jetp * 68 * , 572 ( 1989 ) .
b. z. spivak and a. yu .
zyuzin jetp lett .
* 47 * , 267 ( 1988 ) ; b. z. spivak and s. kivelson phys . rev . * b43 * , 3740 ( 1991 ) . v. b. geshkenbein , a. i. larkin and a. barone , phys . rev .
* b36 * , 235 ( 1987 ) .
m. sigrist and t. m. rice , reviews of modern physics * 67 * , 503 ( 1995 ) .
j. r. kirtley , c. c. tsuei , j. z. sun , c. c. chi , l. s. yu - jahnes , a. gupta , m. rupp and m. b. ketchen , nature * 373 * , 225 ( 1995 ) .
h. kawamura , phys .
rev . * b51 * , 12398 ( 1995 ) and j. phys .
japan * 64 * , 711 ( 1995 ) .
i. morgenstern , k. a. mller and j. g. bednorz , z. phys .
* b69 * , 33 ( 1987 ) .
p. imbert and g. jehanno , hyperfine interactions * 4748 * , 233 ( 1989 ) . c. p. bean , rev . mod .
* 36 * , 31 ( 1964 ) .
a. m. campbell , j.phys .
* c2 * , 3186 ( 1969 ) .
m. ocio , m. alba and j. hammann , j. physique lett .
* 46 * , 1101 ( 1985 ) .
h. dersh and g. blatter , phys . rev .
* b38*,11391 ( 1988 ) .
j. a. osborn , phys .
67 * , 351 ( 1945 ) .
u. yaron , y. kornyushin and i. felner , phys . rev .
* b46 * , 14823 ( 1992 ) . c. p. poole , t. datte and h. a. farach , copper oxide superconductors , john wiley & sons , new - york 1988 .
r. renker , i. apfelstedt , h. kpfer , c. politis , h. rietschel , w. shauer , w. wll , u. gottwick , h. kneissel , u. rauchschwalbe , h. spille and f. steglich , z. phys . *
b67 * , 1 ( 1987 ) .
the susceptibility , which in the usual sense is a local quantity representing @xmath552 , has no meaning in the case of circulating currents in a conductor .
we speak rather of a polarizability @xmath126 which represents the average value @xmath553 , and describes the global effect of the currents over the whole volume of the sample .
s. ruppel , g. michels , h. geus , j. kalenborn , w. schlabitz , b. roden and d. wohllebeen , physica * c174 * , 233 ( 1991 ) .
d. wohllebeen , g. michels and s. ruppel , physica * c174 * , 242 ( 1991 ) . c. ebner and d. stroud , phys .
rev * b31 * , 165 ( 1985 ) .
t. shibauchi , h. kitano , k. uchinokura , a. maeda , t. kimura and k. kishio , phys .
* 72 * , 2263 ( 1994 ) .
m. suzuki , physica * c185189 * , 2243 ( 1991 ) .
y. b. kim , c. f. hempstead and a. r. strnad , phys .
* 9 * , 306 ( 1962 ) .
s. senoussi , m. oussna , g. collin and i. a. campbell , phys . rev . * b37 * , 9772 ( 1988 ) .
j. r. clem , physica * c153155 * , 50 ( 1988 ) .
l. leylekian , m. ocio and j. hammann , physica * b194196 * , 1865 ( 1994 ) . l. leylekian , m . ocio and j. hammann ,
physica * c185189 * , 2243 ( 1991 ) . l. b. ioffe , phys.rev . *
b38 * , 5181 ( 1988 ) .
see the review ref.@xcite , p.183 .
l. glazman , a. koshelev and a. lebed , zhetf * 94 * , 259 ( 1988 ) ; sov.phys.jetp * 67 * , 1235 ( 1988 ) .
see the review ref.@xcite , section 3.2 .
m. v. feigelman and l. b. ioffe , z.phys.*b 51 * , 237 ( 1983 ) . | the properties of the low excitation field magnetic response of the granular high temperature ( ht@xmath0 ) superconductor la@xmath1sr@xmath2cuo@xmath3 have been analyzed at low temperatures .
the response of the josephson currents has been extracted from the data .
it is shown that intergrain current response is fully irreversible , producing shielding response , but do not carry meissner magnetization .
analysis of the data shows that the system of josephson currents freezes into a glassy state even in the absense of external magnetic field , which is argued to be a consequence of the @xmath4-wave nature of superconductivity in la@xmath1sr@xmath2cuo@xmath3 . the macroscopic diamagnetic response to very weak variations of the magnetic field is shown to be strongly irreversible but still qualitatively different from any previously known kind of the critical - state behaviour in superconductors . a phenomenological description of these data is given in terms of a newly proposed `` fractal '' model of irreversibility in superconductors .
pacs numbers 74.50.+r , 74.60.ge , 74.80.bj = 22 cm = 18 cm = .8 cm = -1.5 cm = -1.5 cm epsf |
the sample preparation as well as the measurements were conducted in ultra high vacuum ( uhv ) with a base pressure of @xmath56 mbar .
the n - doped si(111 ) substrate was annealed at 1100@xmath57c by direct current heating for 10 minutes and cooled down slowly to 800@xmath1c over a time interval of 10 minutes until a sharp ( 7@xmath587 ) low energy electron diffraction ( leed ) pattern was observed .
one monolayer ( ml ) of bismuth was deposited at a substrate temperature of 470@xmath1c using a commercial electron beam evaporator to form an adlayer of trimers .
the substrate temperature was measured with an optical pyrometer .
after bi - deposition leed measurements showed a ( @xmath59)r30@xmath1 reconstruction .
as both the monomer phase at 1/3 ml bi coverage and the trimer phase at 1 ml bi coverage show the same ( @xmath0)r30@xmath1 reconstruction we used quantitative leed measurements to distinguish between the two phases .
we measured the integrated intensity of the ( 10 ) and ( 01 ) spots as a function of electron energy and compared them to calculations done by wan _ et al .
the measured data was averaged over three equivalent spots and smoothed .
the result is shown in figure [ fig : leed ] .
the agreement between measured and calculated spectra is quite convincing , allowing an unambiguous identification and preparation of the trimer phase . in figure
[ fig : stm ] two ( 10@xmath5810)nm@xmath60 topographic stm images are shown . in ( a )
the monomer phase is of the bi / si(111 ) structure is shown at a bias voltage of @xmath211.55v and a tunneling current of 0.2na is shown . in ( b ) the trimer phase is shown at a bias voltage of @xmath210.95v .
the tunneling current was 0.2na as well .
angle - resolved photoemission spectroscopy ( arpes ) measurements were done with a hemispherical specs hsa3500 electron analyzer with an energy resolution of @xmath6110mev .
the images were recorded with a step size of 1@xmath1 in angular direction .
we used monochromatized he i radiation with an energy of 21.2ev for the arpes measurements .
the sample was kept at 90k during the measurements .
we measured the band structure along the two high symmetry directions @xmath6 and @xmath7 of the surface brillouin zone ( sbz ) associated with the ( @xmath62)r30@xmath1 reconstruction ( see fig.[fig : sbz ] ) .
the zero for the initial state energy in the measured data was set to the fermi level .
we chose an n - doped si substrate because the position of the fermi level is close to the conduction band .
this allows states within the bulk band gap to be easily accessible by photoemission spectroscopy . from the recorded data we subtracted a shirley background and normalized every energy distribution curve ( edc ) of the image to the same integrated intensity .
spin - degeneracy is a consequence of both time reversal and spatial inversion symmetry .
if the latter is broken spin - degeneracy can be lifted by the spin - orbit interaction . for a free electron gas in two dimensions the spin - orbit coupling hamiltonian in the rashba - bychkov ( rb ) model is given by the following equation @xcite : @xmath63 where the coupling constant @xmath16 is called rashba parameter , @xmath64 are the pauli matrices , @xmath65 is the in - plane momentum and @xmath66 is a unit vector normal to the plane of the two dimensional electron gas ( 2deg ) .
the resulting energy dispersion is : @xmath67 with the effective mass @xmath68 , the offset in parallel momentum @xmath15 and the band extremum at @xmath69 .
the rb hamiltonian lifts the spin - degeneracy and introduces an offset in parallel momentum of the nearly free electron parabola . at high symmetry
points the bands are degenerate due to time inversion symmetry . as a result the band dispersion of a rb split band has a very characteristic shape .
the rashba parameter @xmath16 in the most straight forward formulation of the rb hamiltonian is proportional to the potential gradient perpendicular to the plane of the 2deg .
an additional contribution from an in - plane gradient can strongly enhance the spin - splitting . from the photoemission data the rashba energy @xmath17 ( the energy difference between the crossing point of the two parabolas and the band extremum ) , the offset of the two parabolas in parallel momentum @xmath15 and the effective mass @xmath68 can be determined .
these parameters are related to an effective rashba - parameter @xmath16 via @xmath70 this effective rashba parameter @xmath16 contains all the different contributions to the spin - splitting , i. e. the potential gradient in the plane of as well as perpendicular to the surface and the atomic spin - orbit coupling from the atoms involved .
the rashba parameter @xmath16 determined from equation ( [ equation ] ) should be understood as a characteristic parameter for comparison with other systems . | we demonstrate the induction of a giant rashba - type spin - splitting on a semiconducting substrate by means of a bi trimer adlayer on a si(111 ) wafer .
the in - plane inversion symmetry is broken so that the in - plane potential gradient induces a giant spin - splitting with a rashba energy of about 140mev , which is more than an order of magnitude larger than what has previously been reported for any semiconductor heterostructure .
the separation of the electronic states is larger than their lifetime broadening , which has been directly observed with angular resolved photoemission spectroscopy .
the experimental results are confirmed by relativistic first - principles calculations .
we envision important implications for basic phenomena as well as for the semiconductor based technology . exploiting
the electron spin for information processing is one of the leading goals in the rapidly growing field of spintronics . at its heart
lies the rashba - bychkov ( rb ) type spin - splitting , where the spin - orbit interaction lifts the spin degeneracy in a symmetry broken environment @xcite .
many device proposals make use of this concept @xcite with some interesting proofs of principle @xcite .
the materials of choice are semiconductor heterostructures , albeit the size of the spin - splitting is typically very small .
a large spin - splitting is desirable as it would , for example , decrease the precession time of the spin in a spin transistor @xcite so that it is smaller than the spin relaxation time .
furthermore , a separation of the spin - split states beyond their lifetime broadening is an important criterion for distinguishing between the intrinsic and extrinsic spin hall effect @xcite .
the different influences on the intrinsic spin hall conductivity , such as disorder and elastic / inelastic lifetime , are still under debate @xcite .
recently , a giant spin - splitting has been demonstrated for noble metal based surface alloys @xcite , where heavy elements with a strong atomic spin - orbit coupling are incorporated into the surface .
these systems , however , are not suitable for the field of spintronics because of the presence of spin - degenerate bands at the fermi level originating from the metallic substrate .
one possible alternative is to grow thin films with spin - split bands onto a semiconducting substrate @xcite .
however , due to confinement effects a multitude of quantum well states arise , which potentially influence the transport properties of the system .
it is , therefore , desirable to transfer the concept of the giant spin - splitting directly onto a semiconductor surface .
here we show that a monolayer of bi trimers on a si(111 ) surface forms a two - dimensional ( 2d ) electronic structure with a giant spin - splitting much larger than what has been observed so far at the interfaces of semiconductor heterostructures .
the effect can be traced to a strong contribution of an in - plane potential gradient due to an inherent structural inversion asymmetry ( rb model ) .
while the structure of this system has been studied both theoretically as well as experimentally @xcite , the electronic structure , in particular a possible spin - splitting of the electronic states , has remained a controversial issue @xcite .
we demonstrate unequivocally that bi induces a giant spin - splitting at the silicon surface .
furthermore , the spin - splitting is observed to be larger than the lifetime broadening , so that the bi / si(111 ) system is a prime candidate for spintronics applications or studying the intrinsic spin hall effect . in addition , the silicon substrate allows for excellent compatibility with existing silicon - based semiconductor electronics .
a single layer of bi on si(111 ) grows in a monomer as well as a trimer configuration , both of which show a ( @xmath0)r30@xmath1 reconstruction @xcite . a structural model
is shown in fig .
[ fig : structure ] for the monomer phase ( a ) and the trimer phase ( b ) .
both the monomers and the trimers are centered on top of second layer si atoms ( @xmath2 lattice sites ) forming a symmorphic space group based on the point group @xmath3 .
the si substrate breaks the in - plane inversion symmetry for both the monomer and the trimer phase . looking at the isolated bi adlayer alone
, the trimer formation introduces a reduction of the symmetry because the mirror plane @xmath4 is missing .
the mirror plane @xmath5 holds for both the monomer and the trimer phase as well as for the combination of adlayer and si substrate . from these simple symmetry considerations we conclude that the bi trimer phase is the least symmetric structure and , hence , should lead to the bigger spin - splitting
we , therefore , only consider the trimer phase .
its preparation was verified with quantitative low - energy electron diffraction measurements @xcite .
the experimental band structure measured with angular resolved photoemission spectroscopy ( arpes ) along the two high symmetry directions of the surface brillouin zone ( sbz ) @xmath6 and @xmath7 is displayed in fig .
[ fig : experiment ] ( a ) and ( b ) , respectively @xcite .
the intense feature near @xmath8 at an energy of about @xmath9ev can be attributed to the silicon bulk .
the other features ( s1 , s2 , s3 ) in fig .
[ fig : experiment](a ) originate from the 2d electronic structure of the surface .
s1 is most intense at the @xmath10-point at an initial state energy of about @xmath11ev .
this band splits in two components when moving away from the high symmetry point @xmath10 , which is a strong indication of a rb - type spin - splitting .
s2 is located at about @xmath9ev at the second @xmath8-point and disperses upwards towards the @xmath10-points .
the third state s3 shows the highest intensity at the second @xmath10-point at an energy of about @xmath12ev .
this band moves downwards in energy towards the second @xmath8-point .
the bandwidth of s3 is smaller than the one for s2 .
these three 2d states are also visible along the @xmath7-direction as shown in fig .
[ fig : experiment ] ( b ) .
s1 appears as a parabolic band with negative effective mass with a band maximum located at about @xmath11ev at the @xmath10-point .
along the @xmath7 direction no splitting of this band has been observed .
s2 is located around @xmath13ev at @xmath10 , but only with a very weak intensity .
the most intense feature along the @xmath7-direction is the s3 2d state with a band minimum at about @xmath12ev at the @xmath10-point and an upwards dispersion towards the neighboring @xmath14-points .
a possible spin - splitting in the bi / si(111 ) system is an unresolved issue in the literature . while kinoshita _
et al .
_ @xcite consider a splitting in the three 2d states related to a strong spin - orbit interaction of the bi atoms , it has been dismissed by kim _
et al .
_ @xcite . in the following
, we will show from the experimental data as well as spin - resolved band structure calculations that the band structure shows a giant spin - splitting of the electronic states due to the rb effect .
spin - degeneracy is a consequence of both time reversal and spatial inversion symmetry .
if the latter is broken spin - degeneracy can be lifted by the spin - orbit interaction ( rb model ) @xcite .
in addition , if the inversion symmetry is also broken in the plane of the two - dimensional electron gas , the contribution from an in - plane gradient can strongly enhance the spin - splitting @xcite .
the characteristic parameters quantifying the strength of the splitting are the momentum offset @xmath15 , the coupling constant in the hamiltonian @xmath16 ( rashba parameter ) , as well as the rashba energy @xmath17 .
they are related by @xmath18 and @xmath19 . here
@xmath20 is the effective mass .
a close up of the band structure near the @xmath10-point is shown in fig .
[ fig : closeup ] .
the bands along @xmath6 ( fig.[fig : closeup](a ) ) near @xmath211.2ev clearly show the characteristic dispersion of a rb type spin - splitting with the band crossing at the @xmath10-point and the shift of the maxima away from it . from the data we extract the momentum offset @xmath22@xmath23 , an effective mass of @xmath24 ( @xmath25 free electron mass ) as well as the rashba energy @xmath26mev . from these values
we can calculate the rashba parameter @xmath16=1.37ev .
the spin - splitting is well resolved in the data .
the average line width for the spin - split states at the band maximum ( @xmath27@xmath23 ) is 195mev , which accounts for intrinsic lifetime as well as interactions and scattering .
the separation of the states is about 220mev .
the spin - splitting at the @xmath10-point in fig.[fig : closeup](a ) is strongest along the @xmath6-direction . along the @xmath28-direction in fig .
[ fig : closeup](b ) the spin - splitting at the @xmath10-point is much weaker and can not be resolved in the experiment .
this peculiar band topology can be related to the symmetry properties of the @xmath10-point . as the @xmath10-point is located on the border of the first sbz it has a reduced symmetry as compared to the @xmath8-point . despite the symmetry breaking of the bi trimers , the mirror symmetry @xmath5 ( see fig.[fig : structure ] ) holds so that for the dispersion along the @xmath28-direction the spin - splitting is greatly reduced , i. e.
it can not be observed in the data . to support our interpretation of the observed spin - splitting , we conducted spin - resolved first principles band structure calculations , which were performed in close analogy to our previous calculations on the rb effect @xcite .
the surface geometry of the trimer structure is determined from first principles using the vienna ab - initio simulation package ( vasp ) which provides precise total energies and forces @xcite .
the bi trimers ( milkstool structure ) are relaxed outward by @xmath29{\%}$ ] from the ideal positions ( 100% corresponds to the si bulk interlayer distances , lattice constant @xmath30{\aa}$ ] ) .
the subsurface relaxations are small ( @xmath31{\%}$ ] ) and neglected in the korringa - kohn - rostoker ( kkr ) calculations .
the in - plane displacement of the bi trimer atoms @xmath32 is @xmath33 , with @xmath34 indicating bi on - top of first layer si atoms and @xmath35 coinciding bi - trimer atoms on @xmath2 sites .
the subsequent kkr and relativistic layer - kkr calculations use the structural data from vasp as input . the spectral density @xmath36 is obtained from the imaginary part of the site - dependent green function .
resolved with respect to spin orientation ( index @xmath37 ) and angular momentum , it allows a detailed analysis of the electronic structure .
the difference @xmath38 reveals the characteristic spin splitting of rb - split bands .
the results of the band structure calculations are shown in fig.[fig : calculation ] for the @xmath6-direction in ( a ) and for the @xmath39-direction in ( b ) .
the intensity scale shows the total spectral density ( @xmath40 ) of the states multiplied by the sign of the spin - polarization sgn(@xmath41 ) perpendicular to the high symmetry line , i. e. blue and red colors correspond to opposite spin - polarizations .
the calculations reproduce all the main features of the measured band structure . in particular , the splitting of the s1 band around the @xmath10-point along the @xmath6-direction is well documented .
as can be seen in figure [ fig : calculation ] the two branches of the split s1 band clearly show opposite spin - polarization , i. e. a giant spin - splitting in the electronic structure of bi / si(111 ) .
the spin - splitting is strongly anisotropic around @xmath42 .
the peculiar band topology , which was observed in the experiment is clearly reproduced in the calculations .
this can again be attributed to the lower symmetry of wave vectors @xmath43 within ( @xmath44 ) or perpendicular ( @xmath45 ) to a mirror plane of the system .
it is conceived that this feature results from the ` trimerization ' of the three bi sites in the 2d unit cell ; calculations with a reduced @xmath32 ( i. e. larger distance between bi trimer atoms ) indicate an even smaller splitting along @xmath45 .
furthermore , the calculations show that about 83% of the spin - split states at the @xmath10-point are localized in the bi adlayer and about 16% in the first si layer .
one can thus speculate that the spin - splitting is particularly influenced by the bi adlayer and that trimerization symmetry breaking increases the effect of the in - plane potential gradient .
the giant spin - splitting in the bi / si(111 ) trimer system has a similar origin as in the bi / ag(111 ) surface alloy : an inversion symmetry breaking in the plane of the surface leads to a strong contribution from an in - plane potential gradient , which substantially enhances the spin - splitting . in both systems
the threefold symmetry of the underlying substrate breaks the in - plane inversion symmetry .
however , considering only the topmost layer , the trimer formation in bi / si(111 ) also leads to a breaking of the in - plane inversion symmetry ( see fig .
[ fig : structure ] ) , which is not the case for the bi / ag(111 ) surface alloy . comparing the spin - splitting of the bi / si(111 ) electronic structure to semiconductor heterostructures
, we find that in the latter the spin - splitting is substantially smaller .
for example , for an inverted ingaas / inalas heterostructure a rashba constant of @xmath46ev has been measured @xcite . with an effective mass of @xmath47 ,
a rashba energy of @xmath48ev can be calculated . for hgte quantum wells
a rashba constant @xmath49ev has been found @xcite .
however , here the spin - splitting has been identified to be proportional to @xmath50 instead of a linear dependence @xcite . for the bi / si(111 ) system ,
the rashba energy @xmath26mev as well as the rashba parameter @xmath51ev are much bigger . from the momentum
offset @xmath52@xmath23 we can calculate that a phase shift of the spin precession angle @xmath53 can be obtained after a length @xmath54 of only @xmath55 nm . in the ingaas / inalas heterostructure a length of 400 nm has been estimated . while these figures show the excellent potential of the bi / si(111 ) system , additional measurements giving insight into the transport properties , such as shubnikov - de haas oscillations , are necessary to further elaborate the suitability of this system for spintronics applications .
corresponding experiments are in progress .
we have shown that the trimer phase of bi on si(111 ) shows a giant spin - splitting .
the experimental results reveal the characteristic band dispersion of a rb - type spin - splitting with a peculiar band topology at the @xmath10 point .
they are confirmed by first principles band structure calculations .
the splitting is caused by the spin - orbit interaction induced rb effect in combination with a strong contribution from the in - plane gradient due to the reduced symmetry of the trimer structure and the substrate .
furthermore , this spin - splitting is of the same order of magnitude as the one reported for bi / ag(111 ) and orders of magnitude lager than a typical spin - splitting reported for semiconductor heterostructures . in this way
, we have transferred the concept of giant spin - splitting onto a semiconducting substrate .
this gives excellent perspectives for the use of this concept in the field of spintronics .
in particular , the silicon substrate makes this system compatible with existing semiconductor technology . on the fundamental side
such systems are interesting for , e. g. , the spin hall effect .
since the energy separation of the spin - split states ( 220mev ) is larger than the lifetime broadening ( 195mev ) , it may be easier to distinguish the extrinsic and intrinsic spin hall effects .
12 y. a. bychkov and e. i. rashba _ sov .
phys .
jetp lett .
_ * 39 * , 78 ( 1984 ) j. sinova , d. culcer , q. niu , n. a. sinitsyn , t. jungwirth and a. h. macdonald _ phys
. rev .
lett . _ * 92 * , 126603 ( 2004 ) s. datta and b. das _
appl .
phys .
lett . _ * 56 * , 665 ( 1990 ) j .- i .
ohe , m. yamamoto , t. ohtsuki and j. nitta _ phys .
rev .
b _ * 72 * , 041308(r ) ( 2005 ) t. koga , j. nitta , h. takayanagi , and s. datta _ phys
. rev .
lett . _ * 88 * , 126601 ( 2002 ) y. kato , r. c. myers , a. c. gossard , and d. d. awschalom , _ nature _ * 427 * , 50 ( 2004 ) s. a. wolf , d. d. awschalom , r. a. buhrman , j. m. daughton , s. von molnr , m. l. roukes et al .
_ science _ * 294 * , 1488 ( 2001 ) j. wunderlich , b. kaestner , j. sinova , and t. jungwirth , _ phys . rev .
lett . _ * 94 * , 047204 ( 2005 ) .
b. a. bernevig , s .- c .
zhang , _ phys .
rev .
lett . _ * 95 * , 016801 ( 2005 ) .
p. wang and y. q. li , _
j. phys . : condens
. matter _ * 28 * , 215206 ( 2008 ) and references therein . w. yang , k. chang , and s. c. zhang , _ phys .
rev .
lett . _ * 100 * , 056602 ( 2008 ) and references therein . c. r. ast , d. pacil , m. falub , l. moreschini , m. papagno , g. wittich , p. wahl , r. vogelgesang , m. grioni , k. kern , cond - mat/0509509 ( 2005 ) . c. r. ast , j. henk , a. ernst , l. moreschini , m. c. falub , d. pacil , p. bruno , k. kern and m. grioni _ phys .
rev .
lett . _ * 98 * , 186807 ( 2007 ) c. r. ast , g. wittich , p. wahl , r. vogelgesang , d. pacil , m. c. falub , l. moreschini , m. papagno , m. grioni and k. kern _ phys .
rev .
b _ * 75 * , 201401(r ) ( 2007 ) d. pacil , c. r. ast , m. papagno , c. da silva , l. moreschini , m. falub , a. p. seitsonen , and m. grioni , _ phys . rev .
b _ * 73 * , 245429 ( 2006 ) .
e. frantzeskakis , s. pons , h. mirhosseini , j. henk , c. r. ast , and m. grioni , _ phys .
rev .
lett . _ * 101 * , 196805 ( 2008 ) . k. he , t. hirahara , t. okuda , s. hasegawa , a. kakizaki , and i. matsuda , _ phys
. rev .
lett . _ * 101 * , 107604 ( 2008 ) .
t. hirahara , t. nagao , i. matsuda , g. bihlmayer , e. v. chulkov , yu . m. koroteev , p. m. echenique , m. saito , and s. hasegawa , _ phys .
rev .
lett . _ * 97 * , 146803 ( 2006 ) .
r. h. miwa , t. m. schmidt and g. p. srivastava _ j. phys .
: condens . matter _ * 15 * , 2441 ( 2003 ) k. j. wan , t. guo , w. k. ford and j. c. hermanson _ surf .
sci . _ * 261 * , 69 ( 1992 ) r. shioda , a. kawazu , a. a. baski , c. f. quate , and j. nogami , _ phys .
rev .
b _ * 48 * , 4895 ( 1993 ) see epaps supporting information t. kinoshita , s. kono and h. nagayoshi _
j. phys .
soc .
jpn . _ * 56 * , 2511 ( 1987 ) y. kim , j. s. kim , c. c. hwang , s. p. shrestha , k. s. an , and c. y. park _ j. kor .
phys .
soc . _ * 39 * , 1032 ( 2001 ) g. kresse and j. furthmller , _ phys .
rev .
b _ * 54 * , 11169 ( 1996 ) j. nitta , t. akazaki , h. takayanagi , and t. enoki , _ phys .
rev .
lett . _ * 78 * , 1335 ( 1997 ) m. schultz , f. heinrichs , u. merkt , t. colin , t. skauli , and s. lvold , _ semicond .
sci .
technol . _ * 11 * , 1168 ( 1996 ) x. c. zhang , a. pfeuffer - jeschke , k. ortner , v. hock , h. buhmann , c. r. becker , and g. landwehr , _ phys .
rev .
b _ * 63 * , 245305 ( 2001 ) |
cosmic shear , the weak gravitational lensing effect by the large - scale structure of the universe , will be one of the primary probes in most of the forthcoming cosmological galaxy surveys , including kids , des , hsc , lsst , and _ _ euclid _ _ @xcite .
the correlations of the distortions of distant galaxy images induced by the continuous deflection of their light by the matter distribution along the line of sight are sensitive to both the evolution of structure and the geometry of the cosmos , so that cosmic shear can provide excellent constraints on dark matter , dark energy , and potential deviations from general relativity @xcite .
precision measurements of cosmological parameters need to be matched by an equally superb control of systematic errors .
in particular , weak lensing surveys will have to undertake major efforts to control their dominant astrophysical contamination , the intrinsic alignment of galaxy shapes , which mimics the correlations induced by gravitational lensing . to date ,
the knowledge about intrinsic alignments is limited , and virtually non - existent at the typical redshifts of galaxy samples used for weak lensing analyses around @xmath5 , and at small , non - linear scales from where the majority of cosmological constraints originates ( e.g. * ? ? ?
hence several methods have been designed to remove or calibrate the intrinsic alignment signal with minimal assumptions about the form of the intrinsic correlations ( see e.g. * ? ? ?
* ; * ? ? ?
? * ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
? * ; * ? ? ?
while most of these approaches constitute valuable consistency checks , it is likely that none is capable of controlling the intrinsic alignment contamination without a significant loss of statistical power of the survey , unless prior constraints on the systematic are available .
therefore a robust and comprehensive model for intrinsic galaxy alignments is paramount to allow for the optimal exploitation of the future rich weak lensing data sets .
even the design of surveys would profit from an improved understanding of intrinsic alignments which , for instance , are the main driver for the requirements on the accuracy of galaxy redshift measurements @xcite .
moreover , intrinsic alignments potentially constitute an interesting astrophysical signal that probes the dependence of galaxy formation and evolution on environment , and that is complementary to established techniques .
this calls for an intrinsic alignment model that can explain e.g. their dependence on redshift , environment , merger history , and physical properties of galaxies .
the classical picture of how galaxies acquire certain shapes assumes a dichotomy between spiral and elliptical galaxies .
the shapes of the former are determined by their disc whose rotation axis is aligned with the halo angular momentum , which in turn was created via tidal torquing during the assembly of the galaxy @xcite . as the resulting angular momentum scales with the square of the quadrupole of the gravitational potential , one expects the correlations between the angular momenta and hence the projected shapes of spiral galaxies to vanish to first order @xcite .
the shapes of elliptical galaxies are assumed to follow the generally triaxial shape of their haloes which are subjected to tidal stretching by the surrounding large - scale matter distribution .
the projected ellipticity of the galaxy is then proportional to the quadrupole of the gravitational potential , leading to the linear alignment model for early - type galaxies @xcite .
even if these simple prescriptions prove to be valid for correlations at large galaxy separation , one expects modifications on small scales where non - linear structure evolution , baryonic physics , and dynamical processes in high - density regions become relevant .
one way forward is to incorporate the alignments of satellite galaxies in a halo model @xcite , which predicts similar correlations as an earlier empirical ansatz by @xcite to employ the full non - linear power spectrum in the linear alignment model .
the predictions of intrinsic alignments by the tidal torque and tidal stretching paradigms are consistent with current observations , which however are still fraught with relatively large uncertainties .
measurements among late - type galaxies in the sdss main samples @xcite and the overlap region between the sdss and wigglez surveys @xcite have not yielded any significant detection , except at very low redshift @xcite . in contrast , correlations between intrinsic shapes as well as the alignment of intrinsic shapes towards overdensities of early - type galaxies have been robustly detected in several data sets out to redshift 0.7 ( @xcite ; see also @xcite for a detection in cosmic shear data ) .
the observed correlation functions are in agreement with the galaxy separation and redshift dependencies predicted by the linear alignment model and its non - linear modifications ( @xcite ; see also @xcite ) , featuring in addition a clear increase in amplitude with galaxy luminosity @xcite . concerning small scales , observations of satellite galaxy shape alignment towards the centre of clusters reach back to @xcite .
recent results range between weakly significant radial alignment of satellites @xcite and non - detection of anisotropy in the orientations @xcite , with uncertain levels of selection effects and systematic errors .
a successful model of intrinsic galaxy shapes has to fit large - scale correlations of galaxy ellipticities , small - scale satellite alignment , as well as the distribution of galaxy ellipticities simultaneously . in @xcite , paper i
hereafter , dark matter halo properties from the millennium simulation were combined with semi - analytic models of galaxy evolution and analytic models for the link between the shape of galaxies and their underlying halo .
the resulting shape models were confronted with one - point statistics and distributions of galaxy ellipticities measured in the cosmos survey . in this work
we continue to follow this ansatz of semi - analytic galaxy shape modelling , now investigating the two - point statistics of ellipticities .
the millennium simulation @xcite is well suited to build our galaxy shape models because it has a sufficiently large volume to allow for measurements of correlations among widely separated galaxies , combined with good mass resolution for accurate halo shape and angular momentum measurements ( cf .
@xcite who use a similar approach but rely on simulations that have 20 times higher particle mass ) .
numerous publications ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) have investigated the correlations of dark matter halo ellipticities and angular momenta with each other , and with the large - scale matter distribution , in n - body simulations . to enable an accurate selection of galaxy samples for comparison with observations , we supplement this information with multi - band photometry and galaxy type classifications from the semi - analytic models of galaxy formation and evolution by @xcite . to incorporate the critical link between the shape of the visible , baryonic matter and the structure of the underlying dark matter distribution , we make use of the statistical properties of this relation extracted from a large number of small - scale , high - resolution hydrodynamic simulations by @xcite . furthermore , as the millennium data includes the positions of satellite galaxies , but not the shapes of the corresponding subhaloes , we resort to simple models of satellite shapes and alignments , based on the high - resolution simulations by @xcite whose findings are qualitatively in agreement with similar investigations by @xcite .
this article is structured as follows . in section@xmath6[sec : sims ] we briefly summarise the main aspects of the underlying simulations , before providing in section@xmath6[sec : galaxymodels ] a synopsis of our models of galaxy shapes .
section@xmath6[sec : simresults ] highlights the dependence of the shape correlations resulting from the simulations on redshift , mass , luminosity , and environment . in section@xmath6[sec : iaobs ] the intrinsic alignments in wigglez and sdss galaxy samples are compared to the simulation - based alignment signals .
the impact of the best - matching intrinsic alignment model on weak lensing surveys is studied in section@xmath6[sec : impact ] . in section@xmath6[sec : conclusions ] we summarise and conclude on our findings . unless stated otherwise , rest - frame magnitudes are @xmath7-corrected to @xmath0 and computed assuming the cosmology of the millennium simulation ( see below ) except a hubble constant @xmath8 with @xmath9 .
magnitudes extracted from the millennium database are given in the vega system , while all observations use the ab system .
if direct comparison is necessary , we resort to the conversion tables of @xcite . when multiple data sets with error bars are plotted , points are slightly offset horizontally for clarity throughout .
in this section we briefly summarise the key aspects of the simulations that we use .
more details are provided in section@xmath62 of paper i. the basis of our galaxy shape models is the millennium simulation @xcite which provides us both with the volume ( comoving box size @xmath10 ) to measure correlations on cosmological scales and with the mass resolution ( particle mass @xmath11 ) needed to determine the properties of galaxy - sized dark matter haloes accurately .
we will work with the 32 snapshot outputs from @xmath12 to @xmath0 .
the cosmology of the millennium simulation follows a spatially flat @xmath13cdm model with matter density parameter @xmath14 .
the baryon density parameter is @xmath15 , the hubble parameter @xmath16 , the power - law index of the initial power spectrum @xmath17 , and the normalisation of the power of matter fluctuations @xmath18 .
this latter value is high compared to recent observational constraints suggesting @xmath19 ( e.g. * ? ? ?
* ; * ? ? ?
* ) , but the impact of @xmath20 on the strength of intrinsic galaxy correlations is unknown at present ( see paper i for a discussion on the expected effects of high @xmath20 as well as of baryons on intrinsic alignments ) .
we will compare the simulation - based intrinsic alignment signals directly to observations , thereby assuming that the high value of @xmath20 does not affect our results significantly .
note however that , when calculating the importance of intrinsic alignments relative to cosmic shear correlations , we will use @xmath18 also for the latter to be consistent .
models of galaxies were placed into the millennium haloes via the semi - analytic galaxy evolution model ` galform ` in the version of @xcite . from these models we extract apparent and rest - frame magnitudes in various bands as well as the distinction between central and satellite galaxies , where
the former are defined as the galaxy in the most massive substructure of a halo at any given time .
moreover , we adopt the classification of galaxy morphologies via the bulge - to - total ratio of @xmath21-band luminosity , @xmath22 , as proposed by @xcite .
we define our early - type galaxy sample via @xmath23 and the late - type sample accordingly by @xmath24 .
the latter comprises spiral and lenticular galaxies , the cut at @xmath25 being motivated by the strikingly different merger histories of these populations ( see * ? ? ?
ray - tracing is employed to compute the gravitational shear at the positions of dark matter haloes , generated by the light deflection of the matter distribution along the line of sight .
the method is detailed in @xcite and provides us with 64 light cones with an area of @xmath26 each , yielding a mock survey of @xmath27 out to a redshift of @xmath12 .
special care has been taken to avoid that parts of dark matter haloes end up on different lens planes which in the context of this paper could e.g. create artificial alignments of close halo / galaxy pairs .
all vectorial quantities such as satellite positions , shape ellipsoid major axes , or angular momenta are transformed into a coordinate system with the line of sight as one basis vector , which facilitates projection onto the plane of the sky .
in the following we provide a synopsis of the various shape models which we investigate . here
we will focus on the alignment properties , while the assignment of galaxy shapes has been laid out in detail in section@xmath63 of paper i. generally , we divide the galaxy sample into central and satellite galaxies , as identified by the semi - analytic models , and further distinguish between early- and late - type galaxies based on the bulge - to - total luminosity , @xmath28 .
an overview on the model variants used and their naming conventions is given in table [ tab : galaxymodels ] . [ cols="<,<,<,<",options="header " , ] [ tab : galaxymodels ] we base our models of galaxy shapes on the morphologies and angular momenta of the underlying dark matter haloes .
the same approach as in @xcite is used to identify bound structures in the matter distribution of the simulation and to compute the shapes and angular momenta of haloes .
a friends - of - friends algorithm is employed to construct groups of simulation particles within which self - bound structures are detected with the ` subfind ` code @xcite .
a halo is then defined as a collection of self - bound sub - haloes .
a halo shape is computed via the quadrupole tensor of the mass distribution @xmath29 with components @xmath30 where @xmath31 is the number of all particles that belong to the halo , and where @xmath32 denotes the position vector of particle @xmath33 with respect to the halo centre ( defined as the location of the gravitational potential minimum ) .
the eigenvalues and eigenvectors of @xmath29 define an ellipsoid , with the eigenvalues per unit mass giving the square of the semi - axis lengths , and the corresponding eigenvectors specifying the axis orientations .
we interpret this ellipsoid as an approximation to the shape of the halo .
our input catalogues contain all haloes in the millennium database that host a galaxy , with particle numbers down to 20 . in paper
i it was demonstrated that a minimum particle number of @xmath34 is required to measure halo shapes and orientations accurately , i.e. to have less than @xmath35 deviation of the measured axis ratios from the truth , and less than @xmath36degrees uncertainty in the direction of the largest eigenvector ( see also * ? ? ?
sparsely sampled haloes tend to produce smaller axis ratios ( and hence larger ellipticity on average ) as well as rapidly increasing uncertainty in the halo orientation , leading to an underestimation of ellipticity correlations @xcite .
the minimum requirement can be relaxed to @xmath37 when accepting @xmath38 deviation of the measured axis ratios and @xmath39degrees deviation of the measured orientation of the largest eigenvector .
since @xmath40 is a restrictive condition that would discard more than half of the haloes identified in the millennium simulation , we will adopt this less stringent limit when comparing the simulation intrinsic alignment signals to data and when predicting ii and gi signals , but keep to a minimum particle number of 300 for the investigation of dependencies in the simulation of intrinsic alignment correlations in section@xmath6[sec : simresults ] .
furthermore we make use of the @xcite calculations of the specific angular momentum of haloes within the virial radius , restricted to @xmath40 .
we find that shape models based on the halo angular momentum have negligible correlations for halo masses slightly above @xmath41 ( see section@xmath6[sec : simresults ] ) , so that the absence of angular momentum information for low - mass haloes should not affect our results .
we test whether our results depend on the radius within which the angular momentum is calculated . to this end
, we calculate correlations between the normalised angular momenta of massive haloes with at least 300 particles within @xmath42
. however , the correlations in orientation resulting from computing @xmath43 within @xmath42 and within @xmath44 do not show significant deviations beyond the error bars computed from field - to - field variance .
all central galaxies with @xmath45 and in haloes with more than 100 particles are assumed to have the same three - dimensional shape as their host haloes .
the complex galaxy ellipticity ( in projection ) is computed from the symmetric tensor @xmath46 as defined in section 3.1 of paper i via , as opposed to the polarisation @xmath47 which was employed in paper i. the two quantities are related via @xmath48 @xcite . ]
@xmath49 a frequently used alternative is to base the shape of elliptical galaxies on the reduced inertia tensor of the halo , arguably producing a better approximation of the shape of the galaxy residing close to the halo centre as it gives more weight to small radii . to model the impact of switching from the simple inertia tensor of equation ( [ eq : massquadrupole ] ) to the reduced one , we devise models for which we increase the halo semi - axis ratios by about @xmath50 , thereby approximating the results of @xcite .
early - type models are denoted by ` est ` if based on the simple inertia tensor , and by ` ert ` if based on the reduced inertia tensor .
additionally we construct a model named ` ema ` which is based on the simple inertia tensor and includes a random misalignment of halo major axes .
the misalignment angles are drawn from a gaussian distribution with a scatter of @xmath51 , as proposed by @xcite who obtained this scatter by matching the intrinsic ellipticity correlations of an sdss lrg sample to halo shape correlations extracted from n - body simulations . since ellipticity correlations among early - type galaxies , whose haloes have @xmath52 and thus no measured halo shape , are expected to be small ( see also section@xmath6[sec : simresults ] ) , it is safe to assume that these galaxies have random orientations .
we make the further assumption that the statistical halo shape properties of galaxies with @xmath52 are the same as those of more massive galaxies . in each redshift slice
we construct two - dimensional histograms of halo axis ratios and assign randomly sampled shapes from these histograms to low - mass galaxies at the same redshift .
all central galaxies with @xmath40 and @xmath53 are modelled as circular thick discs whose orientation is determined by the angular momentum vector @xmath54 of the underlying halo .
the ellipticity of the galaxy image is determined analogously to the procedure outlined in paper i , except that , due to the different definition of ellipticity , paper i , equation ( 6 ) is modified to @xmath55 where @xmath56 is the polar angle of the image ellipse .
the axis ratio of the ellipse reads @xmath57 where @xmath58 is the ratio of disc thickness to disc diameter , i.e. approximately the axis ratio for a galaxy viewed edge - on .
we explore two values of this parameter : @xmath59 , similar to what is expected for the disc of e.g. the milky way ( models denoted by ` sth ` ) , and @xmath60 , designed to include the effect of bulges ( models denoted by ` sal ` ) and motivated by the observations of @xcite . like in the case of the early - type galaxy shape model
, the projection implicitly assumes that the three - dimensional light distribution is uniform with a sharp cut - off at the perimeter .
we refrain from using more complicated schemes involving a realistic radial light distribution as this could imply variable ellipticity as a function of radius .
moreover we neglect any small deviations of the image from an elliptical shape in the projection of discs .
re - simulations of dark matter haloes with high resolution and including baryonic particles modelled in different ways indicate that galaxy discs are not perfectly aligned with the halo angular momentum .
@xcite collected the misalignment results from over 500 haloes simulated with baryons and galaxy formation physics ( from @xcite and @xcite , using the simulations of @xcite and @xcite respectively ) , as well as 95 haloes without baryons ( also from @xcite ) .
@xcite found that the polar misalignment angles @xmath61 between angular momentum vector and disc rotation axis can be jointly described by the distribution @xmath62 with @xmath63 .
the cumulative distribution function of equation ( [ eq : lmisalignment1 ] ) is conveniently calculated and inverted analytically , so that we can compute misalignment angles via @xmath64 } \right\}}\;,\ ] ] where @xmath65 is a uniform random number in @xmath66}$ ] .
we determine a new vector @xmath67 by rotating @xmath43 by an amount @xmath61 around an arbitrary orthogonal axis , followed by another rotation around @xmath43 by an angle @xmath68 uniformly sampled in @xmath69}$ ] .
the galaxy disc is then constructed perpendicular to @xmath67 .
note that @xcite used a similar prescription for late - type galaxy models .
disc thickness is accounted for by rescaling ellipticities as @xmath70 , which in the edge - on limit corresponds to an axis ratio of 0.16 , hence lying in - between the models for @xmath58 that we consider .
misalignment angles between angular momentum vector and rotation axis were extracted from the simulations by @xcite .
the set of simulations studied in @xcite is superior in number of galaxy / halo systems and resolution , although both agree on a median misalignment of the order @xmath71degrees .
the fit to the distribution of @xmath61 suggested by @xcite appears to be more accurate in particular for large misalignments than the truncated gaussian proposed by @xcite .
this misalignment prescription is implemented for disc models with @xmath60 , denoted as ` sma ` .
analogous to low - mass early - type galaxies , late - type galaxies with @xmath72 that have no angular momentum information are modelled as randomly oriented thick discs , where @xmath58 has the same value as the model used for the corresponding model of central late - type galaxies with @xmath40 . for galaxies residing in the substructures of haloes we do not have information about the properties of their dark matter distribution .
therefore we have to rely on external data to model both the shapes and orientations of satellite galaxies . for early - type satellites
we proceed in analogy to the central galaxies with @xmath73 and sample the axis ratios of three - dimensional ellipsoids from the histograms obtained for massive haloes with shape information at each redshift slice ( model ` est ` ) .
optionally these axis ratios are rescaled to mimic the use of the reduced inertia tensor ( model ` ert ` ) .
the ellipsoids are then oriented to point their major axis towards the central galaxy of the halo and subsequently projected along the line of sight using equation ( [ eq : ellipseprojection3 ] ) to yield image ellipticities .
the radial alignment of satellite galaxies towards the centre of the halo is strongly supported by simulations ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) and has also been detected in galaxy clusters ( e.g. * ? ? ?
* ) , although the statistical significance is still low and detections are plagued by systematic effects ( see * ? ? ?
simulation results indicate that tidal torquing is the cause for satellite alignment @xcite .
the perfect radial alignment that we assume by default is to be understood as an upper limit for the satellite ellipticity correlation signal . as an alternative model
we implement the modifications of shapes and orientations of subhaloes found by @xcite in high - resolution dark matter - only simulations ( knebe08 model hereafter ; identifier ` ekn ` ) .
the corresponding modifications to halo shapes are detailed in paper i , section 3.3 .
the same publication also confirmed that satellite galaxies are radially aligned towards their host , albeit with substantial scatter in the angle @xmath74 between the satellite major axis and the direction to the halo centre .
the probability distribution of @xmath75}$ ] is given by ( @xcite ; typo corrected ) @xmath76 where the constants @xmath77 and @xmath78 are independent of the halo mass . after changing the satellite major axis orientation by an angle @xmath74 sampled from equation ( [ eq : knebedistribution ] ) , we allow for a rotation by @xmath68 , uniformly sampled in @xmath69}$ ] , around the connecting line to the halo centre , before projecting the satellite shape along the line of sight .
similar to their central counterparts , late - type satellite galaxies are assumed to be thick circular discs perpendicular to the angular momentum vector of the halo , where we place the angular momentum vector at an arbitrary pointing perpendicular to the line connecting the position of the satellite with the centre of the halo .
this means that a satellite disc galaxy viewed edge - on points towards the centre of a halo .
otherwise we explore the same model variants as for central late - type galaxies , i.e. the two values of @xmath58 ( models ` sth ` and ` sal ` for @xmath59 and @xmath60 respectively ) and the optional misalignment of the angular momentum vector from its default direction according to @xcite , which in this case is equivalent to the misalignment of the position vector of the satellite with the major axis of the projected image of an inclined disc ( model ` sma ` ) .
we begin our analysis by studying galaxy shape correlations and their trends with various galaxy properties , separated into early types ( i.e. based on halo shapes ) and late types ( i.e. based on halo angular momenta ) . here
we focus on correlations of the projected shapes of two galaxies as a function of their three - dimensional , comoving separation @xmath79 , computed via ( e.g. * ? ? ?
* ) @xmath80 where @xmath81 denotes the average over all positions @xmath82 .
the tangential(+ ) and cross(@xmath83 ) components of the galaxy ellipticity are defined as @xcite @xmath84 where @xmath85 denotes the polar angle of the line connecting the galaxy pair in a reference coordinate system .
the statistic @xmath86 allows for a compromise between the proximity to observable quantities ( which are projected along the line of sight ) and the readiness with which the signal can be physically interpreted .
we perform the measurement on the simulation boxes , choosing the line of sight parallel to the edges of the cube .
error bars are determined from the variance between eight equal - sized cubic sub - volumes of the simulation box .
correlations between measurements of @xmath86 at different separation @xmath79 are generally weak , the correlation coefficient exceeding 0.5 only occassionally .
unless explicitly stated otherwise , we will limit ourselves in this section to galaxies for which we have reliable measurements of halo properties , i.e. central galaxies with halo particle number exceeding 300 .
early - type models are based on the simple inertia tensor , and for late - type models @xmath60 and zero misalignment between disc rotation axis and angular momentum vector is assumed . in section@xmath6[sec : modeleffect ] we will highlight the impact of different assumptions for our central galaxy models and the effect of including satellite galaxies . as a function of comoving galaxy pair separation @xmath79 for late - type ( top panel ) and early - type ( bottom panel ) galaxies .
black squares show the results for galaxies at redshift @xmath0 , red circles for galaxies at @xmath87 , and blue triangles for galaxies at @xmath88 .
the sample is restricted to the mass range @xmath89 .
note the different scaling of the ordinate axes . ]
computed for every redshift output in a single radial bin covering @xmath90 .
results for early - type ( late - type ) galaxies are shown as black dots ( red diamonds ) .
negative points are plotted with their absolute values as open symbols . since
the signal is computed from the same simulation volume at different redshifts , points are strongly correlated . ]
the dependence of the galaxy ellipticity correlation function on redshift is illustrated in fig.@xmath91[fig:3d_z ] , where @xmath86 is plotted for three redshift outputs , at @xmath0 , @xmath87 , and @xmath88 , respectively .
the samples are restricted to the halo mass range @xmath92 to avoid confusion with a potential mass dependence of the correlations .
late - type galaxies generally do not show significant correlations of their projected ellipticities .
the two high - redshift samples have marginally positive correlations below @xmath93 , while at @xmath0 @xmath86 is consistent with a constant small positive signal on all scales . to achieve higher signal - to - noise
, we compute @xmath86 in a single radial bin covering @xmath90 , plotted as a function of redshift in fig.@xmath91[fig:3d_z_analysis ] . for late - type galaxies the resulting correlations are marginally positive at most output redshifts , increasing slightly in amplitude by a factor of a few between @xmath0 and @xmath1 .
ellipticity correlations for early - type galaxies are statistically significant out to @xmath94 for all redshift bins , increasing by about three orders of magnitude from @xmath95 to @xmath96 ( see fig.@xmath91[fig:3d_z ] ) . the overall amplitude of @xmath86 decreases towards lower redshifts although this decrease is weaker on non - linear scales , i.e. for @xmath97 at @xmath87 and @xmath98 at @xmath0 .
this may be caused by an increase of alignments in regimes of strongly non - linear structure growth .
the decline of correlation strength is also evident from fig.@xmath91[fig:3d_z_analysis ] , and is , to good approximation , exponential as a function of look - back time ( and hence spacing of output redshifts ) .
the only exception to this trend is a mild excess signal at @xmath99 which we relate to the larger impact of non - linear scales in the range @xmath90 at these redshifts .
while the mass cuts remove any redshift dependence of the mean halo mass for the late - type sample , there is still a moderate increase by @xmath100 from @xmath88 to @xmath0 for the early - type sample .
since correlations are expected to become stronger with increasing halo mass ( see the following section ) , this might lead to a small under - estimation of the redshift dependence .
note that , since we fix the halo mass , we are not tracking the same sample of haloes as a function of redshift .
for instance , the high - redshift haloes with strong alignments correspond to objects that formed early and are likely to be among the most massive objects at @xmath0 .
the results of fig.@xmath91[fig:3d_z_analysis ] thus suggest a link between an early formation time of the halo and strong intrinsic alignments . to gain more insight into the evolution of galaxy ellipticity correlations with redshift , it is useful to consider the alignment between the major axes ( normalised to unit - length , denoted by @xmath101 ) for the same galaxy samples , as this measure employs a three - dimensional quantity and thus avoids projection effects .
we compute the correlation function @xmath102 where 0.5 is subtracted to ensure that random orientations of the major axes imply a null signal .
the statistic @xmath103 features the same scaling with redshift as shown for @xmath86 in the bottom panel of fig.@xmath91[fig:3d_z ] ( hence not shown ) .
similarly , @xcite observed an increase in the correlation of halo major axes at high redshift in the millennium simulation , although they used different halo definitions and selection criteria .
strong intrinsic alignment signals at high redshift pose a challenge to deep weak lensing surveys , in particular if a similar increase happens for correlations between galaxy ellipticity and the mass distribution ( and hence the gravitational shear ) .
that such correlations are indeed significant is suggested by the results of @xcite who found an increase of halo major axis - halo number density correlations with redshift . as a function of comoving galaxy pair separation @xmath79 , split into three magnitude bins ( rest frame , @xmath79-band )
shown are correlations between early - type galaxies at @xmath87 in the halo mass range @xmath104 for @xmath105 ( @xmath106 ; @xmath107 ) as black squares ( red dots ; blue triangles ) .
_ bottom panel _ : same as above , but split into three halo mass bins .
correlations between early - type galaxies at @xmath87 in the mass range @xmath108 ( @xmath109 ; @xmath110 ) are shown as black squares ( red dots ; blue triangles ) . ] at @xmath87 in a single radial bin covering @xmath90 as a function of halo mass @xmath111 ( bottom panel ) and rest - frame magnitude @xmath112 at fixed mass ( top panel ) .
the signal for early - type ( late - type ) galaxies is shown as black squares ( red diamonds ) .
negative points are plotted with their absolute values as open symbols . ]
an increase of galaxy number density - ellipticity correlations with luminosity for early types was firmly established in sdss lrg samples by @xcite and confirmed with a substantially extended data set by @xcite .
it is reasonable to assume that this result implies stronger intrinsic correlations and stronger alignments with the large - scale structure of more massive dark matter haloes .
we test this scenario with our mock sample by calculating @xmath86 at @xmath87 for three mass and rest - frame magnitude bins , respectively , as shown in fig.@xmath91[fig:3d_mass ] . to cleanly isolate the mass and luminosity dependencies , we restrict the analysis of the luminosity dependence to galaxies in the mass range @xmath89 .
note that the results for late - type galaxies are noisy and without clear signal detections , so that we only show early - type correlations in this plot .
we observe a clear increase in the overall amplitude of @xmath86 for the higher mass samples ( bottom panel of fig.@xmath91[fig:3d_mass ] ) , without significant changes in the radial dependence .
the top panel of fig.@xmath91[fig:3d_mass ] shows no clear dependence of intrinsic correlations on luminosity for fixed mass range .
we re - compute @xmath86 with narrower bins in @xmath111 and @xmath112 , using only a single radial bin in the range @xmath90 .
moreover , to avoid any residual effect of the pronounced mass dependence , we choose the mass ranges for each @xmath112 bin to cover a decade in @xmath111 and have an average of @xmath113 .
figure [ fig:3d_mass_analysis ] , top panel , reveals that there is no clear luminosity dependence of @xmath86 if the halo mass is kept fixed . the amplitude of the early - type correlations as a function of @xmath111 follows a power - law up to @xmath114 . only the point at the highest masses , which includes the haloes of groups and clusters , drops below this relation
this could possibly indicate that the high merger activity expected for these objects reduces the halo ellipticity on average and/or partially destroys halo alignments .
@xcite reported a similarly strong mass dependence of major axis correlations using two mass bins over a similar mass range as in fig.@xmath91[fig:3d_mass_analysis ] .
late - type galaxies generally show a very weak signal , with no clear trends seen in the mass and luminosity dependence .
the results of fig.@xmath91[fig:3d_mass_analysis ] demonstrate that mass rather than luminosity is the dominant parameter governing the intrinsic alignment strength . in flux - limited samples ,
on which weak lensing studies are usually based , the combined redshift and mass dependencies lead to an even more pronounced increase of the alignment signal with redshift at a given physical galaxy separation .
however , as a function of angular scale , this scaling is counteracted by the drop in intrinsic alignment strength for pairs of galaxies with larger physical separation . at @xmath87 and for fixed halo mass in a single radial bin covering @xmath90 as a function of local galaxy number density @xmath115 .
the signal for early - type ( late - type ) galaxies is shown as black dots ( red diamonds ) .
negative points are plotted with their absolute values as open symbols . ]
an estimate of the local density around a given galaxy is determined by computing the comoving distance @xmath116 to the fourth - nearest neighbour , where satellite galaxies are included , and deriving @xmath117 $ ] .
we refrain from calculating mass densities because we only have halo masses for central galaxies , and besides , number densities are more readily constrained observationally .
as we do not find significant differences in amplitude or radial dependence between subsamples binned according to @xmath115 , we again compute @xmath86 for a single radial bin in the range @xmath90 , devising four logarithmically spaced bins in number density .
like in the preceding section , we minimise the impact of a possible residual mass dependence by setting the mass range of haloes included in each @xmath115 bin to span a decade and have a mean of @xmath113 . as shown in fig.@xmath91[fig:3d_ndensdep ] , the amplitude of early - type correlations is constant across more than five orders of magnitude in @xmath115 .
the signal for late - type galaxies is again noisy , with a marginal preference for stronger alignments in the lowest density bin .
we deem it physically plausible that it is easier for isolated spiral galaxies to remain aligned after formation with the surrounding large - scale structure , and thus neighbouring galaxies in a similar environment , than for disc - dominated galaxies in high - density regions which frequently and/or strongly interact .
the spheroidal galaxies in our early - type sample have all undergone at least one major merger , so that in this case a different , apparently density - independent alignment mechanism , which is preserved or possibly generated by strong gravitational interactions , must be at work .
a dependence of galaxy shape correlations on environment could complicate attempts to model intrinsic alignments and cause subtle biases when trying to infer the properties of the large - scale structure via weak lensing observations , making an observational test of these results highly desirable .
the only attempt hitherto to constrain environment effects on intrinsic alignment signals was undertaken by @xcite who isolated the brightest galaxies of groups and clusters in the sdss lrg sample .
however , this selection more closely resembles a split into central and satellite galaxy populations and is therefore not directly compatible to our results . as a function of comoving galaxy pair separation @xmath79 for different modelling assumptions about shapes of central galaxies , in the redshift range @xmath118 . _
top panel _ : results for late - type galaxies modelled as perfectly aligned with the angular momentum of their host halo ( black squares ) , modelled as having a random misalignment between disc rotation axis and halo angular momentum vector ( red circles ) , and modelled as in the foregoing case but with a thinner disc of @xmath59 instead of @xmath60 ( blue triangles ) . _ bottom panel _ : results for early - type galaxies with shapes based on measurements of the simple inertia tensor ( black squares ) , the reduced inertia tensor ( red circles ) , or including random misalignments ( blue triangles ) . note the different scaling of the ordinate axes . ]
so far we have only used central haloes with @xmath119 for which the galaxy models are derived directly from the halo properties measured in the simulation .
furthermore we have assumed perfect alignment between disc rotation axes and angular momentum vectors , as well as halo shapes and galaxy shapes , to obtain the clearest correlation signals . in fig.@xmath91[fig:3d_models ] we show how the different assumptions made about the galaxy models change the results presented in section@xmath6[sec : galaxymodels ] .
note that in this section we measure the correlations in the light cones ( which include satellite galaxies ) rather than the simulation boxes , limiting the redshift range to @xmath118 , and considering haloes with @xmath119 only .
error bars are now and henceforth determined from the field - to - field variance in the 64 light cones .
if galaxy shapes are assigned according to the reduced inertia tensor measurements ( model ` ert ` ) , @xmath86 is suppressed by a factor of 2.8 , equally across all scales , as all projected galaxy images are closer to circular .
the interiors of haloes , where the reduced inertia tensor has higher weight , are not only closer to spherical but are also less aligned with the large - scale structure ( e.g. * ? ? ? * ) and hence with each other across large distances .
this radial dependence of halo alignment would further reduce the amplitude of correlations , manifesting the strong dependence of shape correlations on the details of the method with which simulated dark matter halo shapes are measured .
note that , nonetheless , correlations for early - type galaxies are significant out to around @xmath120 .
random misalignments ( model ` ema ` ) lead to a suppression of the amplitude of @xmath86 by about a factor of 4.5 , in good agreement with @xcite .
the rescaling is slightly dependent on physical separation , with stronger suppression below @xmath121 . as a function of comoving galaxy pair separation @xmath79 for different contributions of satellite galaxies .
galaxies are selected to lie in the range @xmath118 and be brighter than @xmath122 . _
top panel _ : results for late - type galaxies , where black squares correspond to the signal without satellite galaxies , red circles ( blue upward triangles ) to modelling satellites as misaligned discs with @xmath59 ( @xmath60 ) , green downward triangles to satellites which are perfectly aligned with the direction towards the central galaxy , and brown diamonds to the signal for staellites only .
negative points are shown as open symbols . _
bottom panel _ : results for early - type galaxies , where black squares correspond to the signal without satellite galaxies , red circles ( blue upward triangles ) to modelling satellite shapes based on the reduced ( simple ) inertia tensor , green downward triangles to satellites modelled according to the knebe08 model , and brown diamonds to satellites only .
the inset shows the fraction of central - central ( cen - cen ) , satellite - central ( sat - cen ) , and satellite - satellite ( sat - sat ) galaxy pairs contributing to the correlations at different @xmath79 .
note that differences in pair composition between early- and late - type samples are small . ]
all our late - type models are consistent with zero over the range shown in fig.@xmath91[fig:3d_models ] , where the models ` sma ` and ` sth ` have a @xmath123-value that is about twice as high as for model ` sal ` .
it is expected that random misalignments of the angular momentum vector ( model ` sma ` ) further dilute any signal . using a thinner disc
while retaining the misalignments ( model ` sth ` ) moves data points marginally away from zero but also increases the scatter and hence the error bars . returning to the default models for central galaxies but now including satellites in @xmath86
, we obtain the correlation functions shown in fig.@xmath91[fig:3d_sat ] .
satellite galaxies modify the correlation signal most significantly on scales below about @xmath124 where they clearly dominate the galaxy pairs available ( see the inset ) .
the models ` sth ` , ` sma ` , ` ert ` , and ` est ` have similar correlation signals as the correponding central galaxies ( cf . the respective plots for satellite galaxies only in fig.@xmath91[fig:3d_sat ] ) and thus cause little change in the total @xmath86
. as expected , satellite models with strong alignment like ` sal ` boost the signal on small scales while models with less elliptical shapes and weaker alignments ( ` ekn ` ) yield the strongest suppression of @xmath86 .
we now turn to the comparison of two - point statistics of galaxy shapes from the millennium simulation with the results from observational data sets . to test our models of early - type galaxy shapes , we will use the most comprehensive analysis to date by @xcite , based on several early - type sdss samples . relying on the shape measurements by @xcite , @xcite jointly analysed early - type galaxy samples constructed from the sdss main and luminous red galaxy ( lrg ) spectroscopic samples , and the megaz - lrg sample @xcite which features photometric redshifts based on sdss @xmath125 photometry .
the sdss main and megaz - lrg samples were each split into two redshift bins chracterised by a substantially different mean luminosity .
the lrg sample was divided into three luminosity bins plus a further split in redshift , so that in total 10 data sets were analysed , with quite uniform coverage of redshifts out to @xmath126 and rest - frame magnitudes - corrections of @xcite , but using the same values of the relevant cosmological parameters as those of the millennium simulation . ] over the range @xmath127 .
high - quality shape measurements are available for about @xmath128 galaxies with spectroscopic redshifts and more than @xmath129 galaxies with photometric redshifts .
the selection criteria of the different samples are summarised in @xcite .
if we applied these directly to the simulated galaxy catalogues , the analysis would be subjected to the potentially unreliable colours produced by semi - analytic galaxy evolution models ; see e.g. @xcite who found that the red sequence extracted from the millennium simulation has a tilt and larger spread compared to observations ( note that they did not use a ` galform`-based model though ) .
hence we refrain from applying survey selection criteria that involve colours directly to the simulation - based catalogues .
moreover the luminous galaxies in the sdss samples have low number densities and cover about one order of magnitude larger survey area than the combined millennium catalogues , so that the selected mock samples would be undesirably small .
hence we chose a different route : after homogenising the colours of the samples , @xcite jointly fit a three - parameter model , finding excellent agreement between every data set and the model .
it is based on the linear alignment model @xcite , heuristically extended into the non - linear regime @xcite , and fitted with extra redshift and luminosity dependencies , which results in the matter - intrinsic power spectrum ( for a formal definition see the appendix of @xcite ) @xmath130 with the matter power spectrum @xmath131 .
the best - fit parameters of this model are @xmath132 , @xmath133 , and @xmath134 ( all marginalised and @xmath135 ) .
we adopt the pivot redshift @xmath136 and luminosity @xmath137 , corresponding to an absolute magnitude of @xmath138 in the @xmath79-band , from that work .
the constant @xmath139 is set for convenience to quantify the amplitude dependence in terms of the dimensionless parameter @xmath140 , where @xmath141 corresponds to the normalisation determined by @xcite using supercosmos data @xcite .
the minus sign in equation ( [ eq : iamodel ] ) accounts for the fact that the major axis of a galaxy is aligned with the stretching direction of tidal forces ( see * ? ? ?
the growth factor of matter fluctuations @xmath142 and the matter power spectrum are evaluated for the millennium cosmology , using the @xcite transfer function and non - linear corrections by @xcite .
the latter has been shown to fit the simulation matter power spectrum reasonably well on the relatively large scales we are interested in ( e.g. * ? ? ?
* ) , but note that potential amplitude mismatches would be calibrated via the galaxy bias measurements described below .
we now define galaxy samples from the simulation catalogues by selecting all galaxies with @xmath143 and @xmath144 by default , so that their redshifts and luminosities are covered by the observational data sets . from a rest - frame colour - magnitude diagram of these galaxies , shown in the right panel of fig.@xmath91[fig : cmd ] , right panel , we determine a cut along the line @xmath145 which isolates the ( bright part of the ) red sequence and thus predominantly elliptical ( and lenticular ) galaxies .
the sample consists of @xmath146 central early - type galaxies , @xmath147 central late - type galaxies ( including lenticulars ) , as well as @xmath148 early - type and @xmath149 late - type satellites .
we compute the redshift and luminosity distributions of these sets of simulated galaxies and feed them into equation ( [ eq : iamodel ] ) , from which a model intrinsic alignment power spectrum is computed . the projected correlation function between the matter distribution and the radial intrinsic shear
is then determined via @xcite @xmath150 \label{eq : wdeltaispec } & & \end{aligned}\ ] ] where the kernel @xmath151 denotes the bessel function of the first kind of order @xmath152 . here
, @xmath153 is a weighting with the redshift distribution that takes the effects in a magnitude - limited sample into account ( see the appendix of @xcite for the explicit expression ) .
in addition , we incorporate into @xmath154 a linear weighting with the number of available galaxy pairs in a given transverse separation bin , which affects @xmath155 at the largest scales because pairs with large @xmath156 can only be found at higher redshift , due to the small mock survey size . rcc sample & @xmath157 & @xmath158 + + full sample & @xmath159 & 1.82 + @xmath160 & @xmath161 & 0.52 + @xmath162 & @xmath163 & 1.17 + @xmath164 & @xmath165 & 2.00 + @xmath166 & @xmath167 & 1.97 + @xmath168 & @xmath169 & 0.80 + @xmath170 & @xmath171 & 1.05 + @xmath172 & @xmath173 & 0.56 + @xmath174 & @xmath175 & 1.12 + [ tab : galaxybias ] to obtain the corresponding signal from the simulation - based models , we calculate the correlation function @xmath176 as a function of comoving transverse galaxy separation @xmath156 and line - of - sight separation @xmath177 . here , @xmath178 is the number of galaxies in the sample and @xmath179 the galaxy ellipticity component of galaxy @xmath33 tangential to the line connecting galaxies @xmath33 and @xmath180 .
the function @xmath181 is unity if galaxies @xmath33 and @xmath180 have separations that lie within the bins defined by @xmath156 and @xmath177 , and zero otherwise .
the minus sign in equation ( [ eq : corr3d ] ) implies that radial alignment induces a positive correlation , as is customary in intrinsic alignment measurements .
we achieve higher signal - to - noise ( s / n ) by summing the correlation function along the line of sight , @xmath182 where the summation runs over @xmath183 bins of width @xmath184 .
we place @xmath185 bins in the range @xmath186 . to convert @xmath187 to @xmath188 , we proceed analogously to the observational analyses , i.e. we assume @xmath189 ( linear and deterministic galaxy bias ) and measure the clustering signal of the same galaxy sample in the linear and quasi - linear regime to determine @xmath157 . for convenience
the galaxy clustering correlation function @xmath190 is measured as a function of angular separation @xmath56 rather than
@xmath156 with the publicly available tree code ` athena ` , employing the estimator by @xcite .
the clustering correlation function is modelled as @xcite @xmath191}\ ; p_\delta ( k_\perp , z ) + c\ ; , \end{aligned}\ ] ] where @xmath192 denotes comoving distance .
the same matter power spectrum as used for evaluating equation ( [ eq : iamodel ] ) is employed .
the free parameters in equation ( [ eq : wgg ] ) are the linear bias @xmath157 and the offset @xmath193 ( to account for the undetermined integral constraint ; see e.g. @xcite for the analogous procedure ) , which are jointly fit to @xmath190 as determined from the simulation . to avoid the deeply non - linear clustering regime and effects of the survey aperture on large scales , the fit range is limited to @xmath194 , using 7 logarithmically spaced bins .
table [ tab : galaxybias ] summarises the resulting values of the galaxy bias in the different simulation - based samples that we consider .
we reproduce the general trend of increased galaxy bias with higher redshift and higher luminosity ( the latter only weakly though ) of the galaxy sample ( cf . * ? ? ?
the measured @xmath187 are divided by @xmath157 , with the uncertainty on the bias included in the error bars . as a function of transverse separation @xmath156 .
the black solid line is the j11 model obtained from the best - fit parameters in @xcite , i.e. the model that jointly fits the signal from the sdss early - type galaxy samples considered in that work .
the dark grey band is an estimate of the @xmath135 uncertainty of this model .
points with error bars correspond to the signals of the simulation - based galaxy models .
red circles ( orange squares ; green downward triangles ; blue upward triangles ; brown pentagons ) correspond to the ` est - sma - est - sma ` ( ` est - sal - est - sal ` ; ` ert - sma - ert - sma ` ; ` ert - sma - ekn - sma ` ; ` ema - sma - ekn - sma ` ) model .
the light grey region indicates the regime with non - linear bias which hinders the conversion from @xmath195 to @xmath188 .
note that the ordinate axis has been rescaled by @xmath196^{0.8}$ ] .
] in fig.@xmath91[fig : megaz_amplitude ] the resulting observation - based model ( j11 model henceforth ) for @xmath155 is plotted as the solid curve . on scales
@xmath197 the assumption of linear galaxy bias employed in the transition from galaxy - shape correlation to matter - shape correlation of the sdss samples breaks down , so that the model is not reliable in this regime .
the uncertainty of the model is estimated by propagating the @xmath135 errors on the fit parameters @xmath140 , @xmath198 , and @xmath68 , assuming that they are uncorrelated , which is a fair ( see fig.@xmath9114 of @xcite ) and conservative assumption .
the resulting correlation functions @xmath188 for several galaxy shape models are also shown in fig.@xmath91[fig : megaz_amplitude ] .
both the j11 model and the simulation signals agree fairly well in the dependence on @xmath156 , but not in amplitude in all cases . for a more quantitative interpretation of fig.@xmath91[fig : megaz_amplitude ] we model the simulation - based signals @xmath199 as a rescaled version of the fiducial j11 model , i.e. @xmath200 , and fit the amplitude @xmath201 in the range @xmath202 ( including the full covariance between @xmath156 bins and the contributions to the diagonal due to the errors on @xmath157 ) .
results for the four models under consideration are shown in table [ tab : megazamp ] . as the reduced @xmath203 around unity of the fits suggest , the @xmath156 dependence of the simulation - based signals is indeed in good agreement with the j11 model , and therefore with the prediction of the linear alignment model , as well as the observations of the sdss early - type samples . if we include the points at @xmath204 in the fit , the goodness of fit degrades substantially in all cases .
provided that our assumptions about linear biasing ( as well as about the simple amplitude rescaling due to different values of @xmath20 ) still hold in this regime , the discrepancy could point at deficiencies in the modelling of satellites and small - scale effects . only the model ` ema - sma - ekn - sma ` is consistent at @xmath135 with the fiducial j11 model , i.e. @xmath205 . all other combinations that we explore have amplitudes that are more than @xmath206 above the j11 model , where the most discrepant ones are based on early - type modelling with the simple inertia tensor . switching to the reduced inertia tensor
lowers the correlation function amplitude by @xmath4 ( compare ` est - sma - est - sma ` to ` ert - sma - ert - sma ` ) , whereas the introduction of misaligned halo major axes leads to a factor 2 reduction ( compare ` est - sma - est - sma ` to ` ema - sma - ekn - sma ` ) , in good agreement with @xcite .
the shape modelling of late types has significant impact on the analysis of these early - type samples , with a @xmath4 effect on overall amplitude ( compare ` est - sal - est - sal ` to ` est - sma - est - sma ` ) .
the inclusion of satellites is equally important , as can be seen from the @xmath50 amplitude reduction between models ` ert ` and ` ekn ` .
rcc model & @xmath201 & @xmath158 + + ` est - sal - est - sal ` & @xmath207 & 0.52 + ` est - sma - est - sma ` & @xmath208 & 1.38 + ` ert - sma - ert - sma ` & @xmath209 & 1.45 + ` ert - sma - ekn - sma ` & @xmath210 & 1.00 + ` ema - sma - ekn - sma ` & @xmath211 & 0.85 + [ tab : megazamp ] to compare the redshift and luminosity dependencies of @xmath188 for the different models , we split the default mock galaxy sample ( @xmath143 ; @xmath212 ) at @xmath87 into two redshift bins with a similar number of galaxies .
then we extend the sample by two extra redshift bins with boundaries at @xmath213 and @xmath214 , keeping the constraint @xmath144 .
likewise , we divide the original full sample into two magnitude bins at @xmath215 and add two more magnitude bins with boundaries at @xmath216 and @xmath217 , retaining the original redshift cut @xmath143 . note that the additional bins reach substantially beyond the ranges probed by the observations in @xcite , so that the observation - based model has to be extrapolated with increasing uncertainty . for each redshift and magnitude bin
we proceed as outlined in section@xmath6[sec : megazmethod ] to calculate @xmath188 in a single broad @xmath156 bin that covers the ( quasi-)linear range between @xmath218 and @xmath219 .
the resulting signals are given in fig.@xmath91[fig : megaz_lzdep ] , again including the error on @xmath157 into the error bars of the simulation - based models , and propagating the error of the fit parameters in equation ( [ eq : iamodel ] ) into the uncertainty on the j11 model . .
transverse separation scales have been averaged over in the ( quasi-)linear regime @xmath220<35 $ ] .
black diamonds correspond to the j11 model obtained from the best - fit parameters in @xcite ; error bars are propagated from the errors on these parameters .
symbols and colours of the simulation - based models are the same as in fig.@xmath91[fig : megaz_amplitude ] .
note that the two low - redshift bins cover the range for which observational data was available to determine the j11 model .
_ bottom panel _ : same as above , but showing the rest - frame absolute @xmath79-band magnitude @xmath112 dependence of @xmath221 .
note that the two brightest magnitude bins cover the range for which observational data was available to determine the j11 model . ]
all signals agree well in an overall weak decrease of the amplitude of @xmath188 with redshift , including the redshift range @xmath222 in which the j11 model is not directly constrained by observations . in stark contrast , the j11 and simulation - based models are clearly discrepant in the dependence on absolute magnitude .
observations , in particular of the sdss lrg samples , clearly favour a significant increase of intrinsic alignments with higher galaxy luminosity , which the j11 model extrapolates to @xmath223 .
the simulation - based models predict a shallower slope at the bright end and weak or no dependence on absolute magnitude for fainter galaxies .
this breakdown could either indicate that the roughly linear luminosity dependence observed for the sdss samples does not extend to fainter galaxies , or that our shape models for faint galaxies , mostly satellites ( the satellite fractions in the samples corresponding to the four bins shown in fig.@xmath91[fig : megaz_lzdep ] , bottom panel , are @xmath224 , @xmath225 , @xmath226 , and @xmath227 , from bright to faint ) , produce too strong alignment .
both scenarios are plausible and can be tested with forthcoming improved observations and simulations .
this latter result limits the validity of predictions of intrinsic alignment contamination made with the simulation - based models .
however , note that that a typical galaxy in a _ euclid_-like survey has an absolute rest - frame magnitude of @xmath228 around the median redshift , i.e. it lies in a range where the simulation performs reasonably well .
the intrinsic alignment signal of faint galaxies at low redshift is likely to be over - predicted though , in particular on small scales where satellite galaxies contribute most .
we also attempt a comparison with the intrinsic alignment measurement by @xcite , the only analysis to date that includes a late - type galaxy sample covering a substantial range in redshift .
it is based on the wigglez dark energy survey @xcite which is a spectroscopic redshift survey of bright emission - line galaxies that obtained about @xmath229 redshifts over a total of about @xmath230 of equatorial sky out to @xmath231 .
the northern parts of the survey overlap with sdss imaging from which @xcite determined more than @xmath232 shapes of galaxies ( see that work for details about the shape measurements ) .
the overlap region contains @xmath233 galaxies with reliable redshifts from wigglez of which about a third have high - quality shape measurements . the sample is additionally split into two redshift bins at @xmath234 .
the criteria for the wigglez target selection rely on uv photometry which we do not have at our disposal in the millennium galaxy catalogues . besides
, this would again involve the direct application of colour cuts .
instead , we impose the constraint @xmath235 , where the faint limit originates from shape measurement catalogue while the bright limit is an original wigglez cut to remove contamination by low - redshift objects , to construct another colour - magnitude diagram ( see fig.@xmath91[fig : cmd ] , left panel ) . as fig.@xmath913 in @xcite shows ,
the wigglez colour cut @xmath236 selects roughly the bluest third of the blue cloud .
we reproduce this selection approximately by introducing a cut @xmath237 .
this is supplemented by the criterion @xmath238 , which avoids a spurious contamination with very blue , very bright galaxies classified as early types ; see fig.@xmath91[fig : cmd ] .
it is known that the @xcite models overpredict the abundance of luminous blue galaxies , and that the bulge - to - total luminosity ratios of the full galaxy sample tend to large values for @xmath239 @xcite .
therefore we conclude that these galaxies are not representative of galaxies in wigglez and exclude them from further analysis .
we obtain @xmath187 as well as the correlation function @xmath240 which is calculated by analogy to equations ( [ eq : corr3d ] ) and ( [ eq : corrprojected ] ) , except that in this case the tangential ellipticities of galaxy pairs are correlated . as
many wigglez galaxies are at redshifts @xmath241 , this quantity could be contaminated by correlations due to gravitational shear .
however , we find that including the contribution to galaxy ellipticity by shear leaves @xmath242 practically unchanged , in line with the conclusions of @xcite .
we find that all simulation - based correlations are consistent with zero over the same scales as shown in fig.@xmath913 of @xcite , with very little dependence on the model assumptions .
the associated @xmath123-values ( the probability of obtaining a @xmath203 value at least as extreme as the one observed assuming that the hypothesis of a zero signal is true ) generally exceed 0.5 ; only for @xmath187 in the high - redshift bin @xmath243 .
this is in line with the null detection of intrinsic alignments in the wigglez data , but note that error bars are more than an order of magnitude larger over all scales in the observational data sets , hence there is little discriminatory power .
a caveat in this direct comparison of @xmath187 measurements in wigglez and the millennium simulation is that we have assumed that our models do not only provide a fair description of the correlation between galaxy shape and the matter distribution but also of the link between matter and the galaxy distribution . in other words
, we have implicitly assumed that the simulated galaxies have the same galaxy bias as the observed wigglez galaxies , which is not guaranteed because of the different sample selection criteria and the high @xmath20 value of the simulation .
however , the modelled and observed results would still be broadly consistent even with a difference in galaxy bias by a factor of two or more .
the extremely blue wigglez galaxies are rare objects , which leads to the large error bars on correlation functions due to low number densities , and exhibit very little intrinsic alignment , as confirmed by our results .
this makes them unfavourable for intrinsic alignment investigations .
a large - area survey to similar depths of more generic late - type galaxies , which also make up the bulk of samples used for weak lensing studies , would be desirable but is not available to date . at lower redshift the blue subsamples of sdss spectroscopic galaxies studied by @xcite
could potentially provide another test of our models , but applying the corresponding selection criteria to the simulated catalogues produces only a few hundred galaxies per line of sight , so that any results would be inconclusive due to large error bars .
a selection based on a colour - magnitude diagram as for the wigglez samples is problematic as @xcite split the population by observed rather than rest - frame magnitude which leads to an uncertain yet significant contamination of the blue sample by early - type galaxies ( see the discussions in @xcite and @xcite ) .
hence we refrain from a detailed analysis .
one of the main motivations for studying intrinsic shapes and alignments of galaxies , beside the insights into galaxy formation and evolution , is to assess the contamination of weak gravitational lensing measurements .
the ability to extract information on the growth of structure and the geometry of the universe from the two - point function of gravitational shear ( gg henceforth ) may be severely limited by correlations between the intrinsic shapes of galaxies ( ii henceforth ) as well as correlations between intrinsic shapes of foreground galaxies and the shear acting on background galaxy images ( gi henceforth ) . in the following we will assess the impact of the galaxy shape model that fared best in the comparison to observational data sets on a cosmic shear measurement with typical parameters of forthcoming surveys . combining the results of section@xmath6[sec : iaobs ] with those of paper i , the model ` ert - sma - ekn - sma ` provides overall fair agreement with observational constraints and will hence be employed for the intrinsic alignment predictions of this section .
the constraints on late - type intrinsic alignments and shape distributions are still weak or inconclusive .
we take the observed suppression of the simulated correlations for the red sdss early - type galaxies due to the choice of late - type model ( see figs.@xmath91[fig : megaz_amplitude ] and [ fig : megaz_lzdep ] ) as an indication in favour of the ` sma;sma ` models which are expected to yield the smallest signal ( see also section@xmath6[sec : modeleffect ] ) . the model combination ` ert ; ekn ` produces the best agreement in amplitude and redshift dependence with sdss early - type observations .
it also fares well at describing the intrinsic ellipticity distribution of a typical early - type galaxy sample in cosmos ( see figs.@xmath915 and 7 in paper i ) .
moreover , the choice ` ert - sma - ekn - sma ` should be slightly more conservative in terms of the expected weak lensing contamination than ` ema - sma - ekn - sma ` , the only other model combination tested that provides a good match in amplitude , as the resulting intrinsic alignment signal is larger at all redshifts ( see fig.@xmath91[fig : megaz_lzdep ] , top panel ) .
lccc survey & band & mag .
limit & @xmath244 + + _ euclid_-like & @xmath245 & 24.5 & 0.9 + kids - like & @xmath79 & 24.4 & 0.7 + kids - like & @xmath246 & 24.6 & 0.7 + shallow & @xmath79 & 23.9 & 0.6 + [ tab : surveydepths ] by default we work with a galaxy sample similar to the one anticipated for the _ euclid _ wide imaging survey , by requiring a limiting ab magnitude of 24.5 ( for @xmath247 extended sources ) in the broad @xmath245 _ euclid _ passband @xcite . as magnitudes in this filter are not available from the semi - analytic models , we approximate @xmath245 fluxes , which cover the wavelength range @xmath248 nm , by adding the fluxes of the sdss @xmath79 and @xmath33 filters , as well as half of the @xmath249 filter ( whose throughput lies between @xmath250 nm ) .
these assumptions result in a total sample with a median redshift of @xmath251 and a mean galaxy number density of @xmath252 , which is in good agreement with the numbers quoted in @xcite , @xmath253 and @xmath254 .
we will also consider a survey design closer to forthcoming ground - based surveys such as kids @xcite .
adopting the magnitude limits given in @xcite , we obtain a mock sample with a depth of @xmath255 , in good agreement with other predictions for weak lensing science .
we construct two kids - like samples selected in the @xmath246 and @xmath79 bands , respectively , as well as a shallower survey ( 0.5 magnitudes brighter limit , @xmath256 ) for comparison .
the corresponding properties of the galaxy samples are summarised in table [ tab : surveydepths ] .
all current and future cosmic shear surveys will make use of photometric redshift information to study the large - scale structure in tomographic slices .
we do not take into account the scatter along the line of sight introduced by photometric redshift estimation , but defer analysing the impact of photometric redshift scatter on the intrinsic alignment signals and their control to future work .
this implies in particular that the ii signal is restricted to redshift auto - correlations in our study , and that these auto - correlations are expected to be weaker in reality due to photometric redshift scatter .
the correlations induced by weak lensing are generated along the line of sight and thus best measured as a function of angular separation , in contrast to intrinsic alignments which depend on the physical separation of the galaxies correlated .
therefore we now use the standard correlation functions ( e.g. * ? ? ?
* ) @xmath257 where @xmath258 denotes the average over all angular separation vectors @xmath61 , the correlation function @xmath86 ( see equation [ eq:3dcorr ] ) which depends on physical separation @xmath79 but is otherwise identical , and the correlation functions @xmath240 and @xmath259 ( as employed in section@xmath6[sec : iaobs ] ) which depend on transverse separation @xmath156 and additionally stack the signal from several line - of - sight bins . ] .
note that in this and the following equations we suppress indices indicating the correlation of redshift bin subsamples for notational convenience .
the correlation functions @xmath260 are the measures of choice to be applied to galaxy catalogues derived from real data as they are not affected by masking or a complex survey geometry . as gravitational lensing creates only curl - free shear patterns to first order ,
it is desirable to separate the two - point statistics of the shear field into a curl - free e - mode and a divergence free b - mode . this can be achieved by means of the aperture masses @xcite @xmath261 where @xmath262 is an arbitrary axi - symmetric weight function .
the corresponding two - point statistics , the aperture mass dispersions , can be determined from the correlation functions via ) ; see e.g. @xcite .
as we are interested in the relevance of intrinsic alignments for a realistic mock survey and data analysis , we opt for an approach that could readily be applied to real data . ]
\label{eq : map } & & \end{aligned}\ ] ] where @xmath264 are weight functions whose form is explicitly given in @xcite for the simplest polynomial weight function @xmath262 introduced by @xcite , which we also employ .
in practice the integration of equation ( [ eq : map ] ) is executed as a riemann sum over very finely binned correlation functions ( @xmath265 logarithmically spaced angular separation bins ) .
errors on @xmath260 are propagated into @xmath266 by the formalism described in @xcite and @xcite .
the smallest @xmath56 bin entering the riemann sum is given by the minimum separation for which @xmath260 is measured .
as this is always larger than zero , some leakage from e - modes to b - modes and vice versa is expected .
consulting the results of @xcite , we choose the minimum angular separation such that any leakage has a negligible effect on our e - mode signals ( @xmath267 ) .
the aperture - mass dispersions are calculated for 12 logarithmically spaced angular bins in the range @xmath268 .
it is convenient to express the total intrinsic alignment contribution relative to the weak lensing signal at a given scale via @xmath269 where , by using the absolute value in the numerator , we have accounted for the fact that the gi signal is typically negative ( e.g. * ? ? ?
errors on @xmath270 are propagated from the full covariances of @xmath271 , @xmath272 , and @xmath273 , neglecting any cross - variances between the three statistics . to further condense this information
, we seek to devise a single ratio @xmath274 of intrinsic alignment over lensing signal for a given redshift bin combination .
since a logarithmic angular binning for the aperture - mass dispersion is used , we deem a simple average over angular bins appropriate , if statistical error and cross - correlation between these bins are taken into account . hence , introducing the shorthand notation @xmath275 and @xmath276 , we define the covariance - weighted average @xmath277 here , @xmath278 denotes the standard deviation of @xmath274 .
we fix @xmath279 , i.e. at the highest separation bin , and set @xmath280 , corresponding to an angular range @xmath281 included in @xmath274
. the lower limit of @xmath282 is motivated by the lack of observational constraints on the two - point statistics on smaller scales ( at the median redshift of _ euclid _ a @xmath282 angular separation corresponds to @xmath283 ; compare this to fig.@xmath91[fig : megaz_amplitude ] ) .
moreover , the small - scale intrinsic alignment signal is strongly dominated by satellite galaxies for which our shape models are most uncertain . for gravitational lensing ( gg ) and intrinsic alignment signals ( gi and ii ) for three redshift bin combinations in a _ euclid_-like survey ( @xmath284 ; @xmath285 ) .
the black squares correspond to the gg signal , the blue circles to the ii signal , and red triangles to the negative gi signal ( open symbols show the signal where positive ) . in the small panels
the ratio of total intrinsic alignment over lensing signal @xmath286 is plotted .
grey dotted lines are used to indicate an @xmath270 threshold of @xmath38 , as well as @xmath287 , which serves as a rough separator between the non - linear and ( quasi-)linear regimes . _
top panels _ : auto - correlations in a redshift bin defined by @xmath288 .
_ centre panels _ :
cross - correlations between two redshift bins with @xmath288 and @xmath289 , respectively . _ bottom panels _ : auto - correlations in a redshift bin defined by @xmath289 . ] in fig.@xmath91[fig : angular_example ] the e - mode aperture mass dispersion of the gg , gi , and ii signals is shown as a function of filter scale @xmath56 .
we use a galaxy sample representative of a _ euclid_-like survey , presenting auto - correlations in a low - redshift bin ( @xmath289 ) and a high - redshift bin ( @xmath288 ) , as well as cross - correlations between the two samples . note that in particular the ii signal strength depends on the redshift bin width .
we have chosen the widths to be close to the minimum that is realistically possible for a _ euclid_-like survey , so that the ii correlations shown represent an upper limit .
the cosmic shear signal from the auto - correlations is in good agreement with the results by @xcite who computed @xmath290 directly from the shear field at @xmath291 .
the ii signal is clearly detected on all scales considered in the auto - correlations at low and high redshift , increasing strongly for smaller angular separations .
the gi signal is negative where significant , i.e. foreground ellipticities and the gravitational shear on the images of galaxies in the background are preferentially anti - correlated .
it remains subdominant in amplitude to ii in the auto - correlations ( because the redshift bins are relatively narrow ) , but has a strong signal with a steep dependence on @xmath56 in the cross - correlation .
figure [ fig : angular_example ] also provides the ratio of the total intrinsic alignment correlations over the lensing signal , calculated via equation ( [ eq : riatheta ] ) .
somewhat arbitrarily , we use @xmath292 as a threshold above which the intrinsic alignment contamination is considered severe . future surveys like _ euclid _ strive to measure cosmic shear two - point statistics to sub - percent accuracy , so that any intrinsic alignment effects at the per cent level and higher need to be modelled or removed .
any uncertainty in the knowledge about the intrinsic alignment signals propagates through to increased errors on cosmology ; the stronger the signal the larger the impact on constraints .
if these effects exceed @xmath293 , the requirements on the accuracy of the intrinsic alignment models and on the knowledge of their free parameters will become increasingly strict and hence challenging to meet .
the _ euclid _ baseline restricts the cosmological weak lensing analysis to redshifts larger than 0.5 , mainly because the lensing signal at low redshift is small while intrinsic alignments are comparatively strong . for the auto - correlation of the low - redshift bin , which is located at this transitional redshift , @xmath294 on scales @xmath295 , and @xmath296 below @xmath297 .
thus , our models suggest that for @xmath298 intrinsic alignments become too strong to allow for a precise cosmological measurement via weak lensing , whereas at higher redshifts this is feasible if the intrinsic alignment signal is well understood and the analysis is not extended too far into the non - linear regime . on these small scales
the exploitation of cosmic shear measurements remains limited because of the difficulties of modelling the non - linear contributions to the matter power spectrum , the coupling of modes due to non - linear evolution , and the impact of baryonic physics ( e.g. * ? ? ?
* ; * ? ? ?
* ) . at @xmath299
the intrinsic alignment contamination is a few per cent on all relevant scales , which is still too high to be left unaccounted for .
this is worrisome due to the absence of deep , large - area spectroscopic surveys for representative galaxy samples needed to study intrinsic alignments at these redshifts .
however , we caution again that in practice some dilution of the ii signals is expected because of photometric redshift scatter
. this does not apply to the gi signal in the cross - correlation which yields @xmath300 in the well constrained regime above @xmath301 .
since the model - independent removal of the gi signal severely reduces the cosmological power of the weak lensing survey @xcite , this result implies tight requirements on the quality of intrinsic alignment models over a wide range in redshift . as a function of redshift for a _ euclid_-like survey ( @xmath284 ; @xmath285 ) , averaged over angular scales @xmath302 .
auto - correlations of redshift bins as a function of the mean redshift in a bin are shown as black squares .
cross - correlations between a foreground redshift bin whose mean redshift is given on the abscissa and the background bin at the highest redshift ( @xmath303 ) are shown as red triangles .
the grey dotted line indicates an @xmath274 of @xmath38 . ] to illustrate the redshift dependence of the intrinsic alignment contamination , we define six redshift bins containing roughly equal numbers of galaxies for a _ euclid_-like survey , using the bin boundaries @xmath304 ; @xmath305 ; @xmath306 ; @xmath307 ; @xmath308 ; @xmath309 . for each bin
we calculate the averaged ratio @xmath274 ( see equation [ eq : ria ] ) for the auto - correlation , as well as the cross - correlation between each bin and the highest redshift bin . as the gi signal increases slightly stronger with redshift than the lensing signal for a given foreground redshift bin ( e.g. * ? ? ?
* ) , due to the different weights in the respective limber equations , @xmath274 of the cross - correlation with the highest redshift bin in the background is largest , i.e. we consider the worst case . as is shown in fig.@xmath91[fig : angular_redshift ] , auto- and cross - correlations behave similarly in that @xmath274 decreases considerably with redshift , from close to 0.5 at @xmath310 to a few per cent or less at @xmath311 . at lower redshift
there is less structure along the line of sight that has contributed to the gravitational shear , causing a lower gg signal , whereas a given angular range probes smaller physical galaxy separations at which intrinsic alignment is stronger .
this redshift range , below @xmath312 , is hence well suited to provide strong constraints on intrinsic alignments in a joint tomographic weak lensing analysis ( see * ? ? ?
* ) . at
@xmath313 the shape model predicts an intrinsic alignment contamination below @xmath38 on average , so that the sensitivity of a cosmological weak lensing analysis to intrinsic alignments is somewhat reduced ( cf . * ? ? ?
* ; * ? ? ?
the ratio @xmath274 for the gi signal remains significant at a few per cent at @xmath314 where direct observational constraints on intrinsic alignments will be difficult to obtain , limiting the fidelity of any modelling attempt .
a reliable intrinsic alignment model that captures the dependence on galaxy luminosity , colour , and redshift accurately can in principle be used to optimise a weak lensing survey towards minimum contamination .
for instance , one naively expects that a survey with a brighter apparent magnitude limit has higher intrinsic alignment contamination because at any given redshift galaxies will on average be more luminous and thus more strongly aligned .
, averaged over angular scales @xmath302 , for four different combinations of survey parameters ( passband and limiting magnitude ; see table [ tab : surveydepths ] ) .
black squares ( red dots ; blue triangles ) correspond to auto - correlations in the redshift bin @xmath289 ( auto - correlations in the redshift bin @xmath288 ; cross - correlations between the two bins ) .
the grey dotted lines indicates @xmath315 . ]
generally , one seeks to optimise the statistical power of a survey by keeping the total exposure time fixed and trading off survey area and depth . considering the weak lensing signal in isolation
, this leads to maximising the area once a certain medium depth is reached @xcite .
when including intrinsic alignments , this conclusion is reversed for two reasons : a ) the relative strength of intrinsic alignments with respect to the cosmological signal becomes weaker if less luminous galaxies become part of the sample , and b ) a deeper survey allows for a longer baseline in redshift , hence facilitating the calibration of intrinsic alignments with a minimum of assumptions about models @xcite . as a precursory study
, we investigate the dependence of @xmath274 on the four different survey setups listed in table [ tab : surveydepths ] , varying the limiting magnitude and the passband . in fig.@xmath91[fig :
angular_depth ] we have plotted @xmath274 for the two redshift auto - correlations and the cross - correlation of the same low- and high - redshift bins used in section@xmath6[sec : angdep ] . at low redshift brighter magnitude limits affect the galaxy population only marginally by removing the faintest objects , hence the small differences for the surveys observing in red filters . in the @xmath246-band survey @xmath274
is reduced ( although the significance is low due to large error bars ) , which suggests that preferentially selecting blue , i.e. late - type , galaxies could be effective at reducing the intrinsic alignment contamination .
alternatively , a joint weak lensing analysis of two samples selected in different passbands could help to discern the lensing and intrinsic alignment signals by taking advantage of the achromatism of the lensing effect .
the intrinsic alignment contamination in the redshift cross - correlation is essentially unaffected by modifications of the survey parameters . as this signal
is dominated by gi , this implies that the alignment of the different foreground samples with the surrounding large - scale structure that lenses the background sources is on average unchanged . at high
redshift the impact of the magnitude limit is expected to be stronger .
indeed , @xmath274 for the _ euclid_-like survey is lowest , but the significance is low , and switching between @xmath316 and @xmath317 does not show the same trend .
similar to the low - redshift auto - correlation , selecting galaxies in the @xmath246-band rather than @xmath79 slightly reduces @xmath274 . for the cross - correlation between redshifts in the range @xmath289 and @xmath288 .
black squares correspond to @xmath286 without any cuts ( for reference ; same as in fig.@xmath91[fig : angular_example ] ) , red dots ( blue triangles ) to the signal when @xmath38 ( @xmath4 ) of the reddest galaxies in the foreground sample have been removed using the cut @xmath318 ( @xmath319 ) . the ratio @xmath286 is always computed using the lensing signal for the full sample to keep error bars small . ]
the ii signal is relatively straightforward to suppress by removing physically close galaxy pairs , as demonstrated by @xcite .
this procedure is not effective on the more worrisome gi contamination in cross - correlations of galaxy samples . using the same setup as in section@xmath6[sec : angdep ] ,
we test how well a simple cut in observed @xmath320 colour to remove the reddest galaxies in the low - redshift sample is capable of reducing the gi contamination .
we consider two cases , cutting @xmath38 and @xmath4 of the sample via @xmath318 and @xmath319 , respectively .
figure@xmath6[fig : angular_ccut ] shows that increasingly strict colour cuts can suppress the gi signal on all scales . for the @xmath4 cut
we find a reduction in signal by almost a factor of two around @xmath321 and for @xmath322 , and by about @xmath50 around @xmath323 .
this cut would imply an increase of the error bars on the cosmic shear signal by at most @xmath324 ( on small scales where sample variance is negligible ) .
cuts in rest - frame colour and absolute magnitude may prove to be even more effective in reducing intrinsic alignment contamination , but their performance depends strongly on the accuracy and precision of photometric redshift estimates .
the same applies to the close - pair removal tailored towards the suppression of the ii signal .
these mitigation techniques will be assessed in detail in forthcoming work by combining the semi - analytic models presented here with mock photometric redshift catalogues .
the gravitational deflection of light does not create b - modes in the resulting shear field to first order , so that two - point statistics of b - modes are routinely used to test for residual systematic effects .
@xcite detected b - modes generated by gi correlations in their simulations and proposed to use these as a diagnostic for intrinsic alignments .
we calculate the b - mode aperture mass dispersion @xmath325 for similar redshift bin combinations as used in section@xmath6[sec : angdep ] , again assuming a _ euclid_-like survey .
both lensing and intrinsic alignment b - mode signals are generally much smaller than the e - modes ( @xmath326 on scales @xmath327 , except for ii at low redshift which attains values similar to the e - mode around @xmath323 ) , but in some cases we find @xmath328 for ii and gg at statistically significant levels .
this is unphysical as the aperture mass is given by an integral over a non - negative power spectrum with a non - negative kernel @xcite .
increasing the minimum angular separation entering the integral of equation ( [ eq : map ] ) to @xmath329 removes the negative b - modes , at the price that only scales of @xmath323 and above are immune to leakage from the e - mode . even after these modifications
, we see a small but significant positive b - mode lensing signal out to large scales .
this is unexpected as @xcite showed that multiple light deflections and lens - lens coupling generate b - modes which are less than a per mil of the e - mode signal .
contributions by source redshift clustering @xcite should be small , or zero in the case of the cross - correlations ( because the redshift distributions do not overlap ) .
the limited resolution of the n - body simulation or the shear maps can not be the cause of @xmath328 because in that case smaller scales and lower redshifts should be affected more strongly , in contrast to what we observe .
a potential explanation is simulation particle shot noise which affects @xmath330 only at zero lag but contributes to @xmath331 at @xmath332 .
excess signal in @xmath333 leads to more negative b - modes than expected from a cosmological signal ; see equation ( [ eq : map ] ) .
we can not exclude that the residual gg signal is spurious , and that the ii and gi b - mode signals are affected at a similar level . therefore we defer a detailed investigation of the significance of b - modes generated by intrinsic alignments to future analysis .
note that these systematics affect the e - mode signals at the per cent level and are hence negligible in the preceding analysis .
in this work we investigated the correlations induced by the intrinsic alignment of galaxies using semi - analytical models of galaxy shapes . these models are based on the dark matter halo properties extracted from the millennium simulation , supplemented by semi - analytic models of galaxy formation and information on the link between dark matter and the luminous part of galaxies from various simulations with and without baryons . in a first step
the resulting intrinsic ellipticity correlations were analysed for central galaxies whose halo properties were measured in the simulation , differentiating between early and late types as identified by the semi - analytic models .
late - type galaxies are only weakly aligned , with small positive correlations on scales below @xmath334 which increase marginally with redshift but do not show any strong scaling with mass , luminosity , or environment , although statistical errors remain large for this sample .
early - type galaxies are strongly aligned with a significant signal beyond physical separations of @xmath335 which increases towards smaller separations ( three orders of magnitude over two decades in @xmath79 ) and towards higher redshifts ( one order of magnitude from @xmath0 to @xmath1 ) .
we found that the alignment strength increases with halo mass up to @xmath336 beyond which it remains constant , while it does not depend on luminosity or galaxy number density in the environment if the halo mass remains fixed .
we compared our galaxy shape models to a suite of recent observations of intrinsic alignments in both blue- and red - selected galaxy samples .
taking into account a realistic composition of the galaxy sample ( see also * ? ? ?
* ) and the effects of satellite galaxies , we obtained combinations of simple shape models that are in good agreement with early - type observations .
models that are based on reduced inertia tensor measurements of dark matter halo shapes ( which tend to be closer to spherical ) , and that include random misalignments of satellite galaxies ( taken from * ? ? ? * ) , yield better fits .
a model including random misalignments of central early - type galaxies as proposed by @xcite is also in fair agreement with observations , but such a misalignment is not required to match the signals from observations and simulations
. reduced inertia tensor measurements arguably yield a better approximation to the shape of a galaxy as they put more weight towards the centre of a halo .
ideally one would obtain the shape estimate only from the inner parts of the halo , but this increases the requirements on mass resolution as several hundred simulation particles now have to be within the chosen sub - region alone .
dark matter haloes become more spherical towards their centre @xcite , a trend that becomes more pronounced if the impact of baryons is taken into account @xcite .
we demonstrated that , depending on how the halo shape is measured , the prediction of intrinsic alignment amplitudes can vary by factors of a few . while the comparison with the measurement of late - type galaxy alignments by @xcite remains inconclusive due to the large statistical uncertainty of the observational data , good agreement of the radial and redshift dependencies between the observations and model predictions for early types
was found .
none of our models was capable of reproducing the observed luminosity , largely over - predicting alignments for faint objects .
this discrepancy could be due to a failure of the model for satellite alignments or indicate that the luminosity scaling found observationally for lrg - like galaxies does not extend to @xmath337 .
both hypotheses can be tested with forthcoming simulations and observations . using the galaxy shape model that best reproduced the joint observations of two - point statistics and the shape distributions analysed in paper i
, we predicted the intrinsic alignment contamination of a _ euclid_-like survey , measured in terms of the ratio of total e - mode intrinsic alignment ( ii+gi ) over lensing signal , @xmath274 .
we obtained @xmath274 below @xmath38 for @xmath338 and on the angular scales relevant for weak lensing analyses , i.e. above a few arcminutes , with higher contamination on smaller scales and at lower redshift .
these results are of the same order as earlier estimates @xcite and confirm that intrinsic alignments will be one of the major systematics for forthcoming cosmic shear surveys , but not significantly degrade constraints on cosmology if moderate priors on the intrinsic alignment signal can be imposed ( @xcite ; see also @xcite for the impact on cosmological parameter estimation by several intrinsic alignment models ) .
we investigated the effect of the galaxy sample selection on the strength of intrinsic alignments , selecting survey parameters similar to kids and _ euclid _ , and applying cuts in observed colour to remove the reddest galaxies .
variations in limiting magnitude and the filter in which the selection is made can change alignment signals , in particular the ii term , by factors of a few . the gi signal can almost be halved by reducing the sample size in the foreground by @xmath4 , corresponding to an increase in statistical error on the lensing signal by less than @xmath324 .
of course , one can undertake a much more sophisticated sample selection , e.g. by using cuts in rest - frame magnitude and colour , which can be approximately calculated by means of photometric redshifts .
this should be more effective at selecting galaxies with particularly low ( or high ) @xmath274 and will be investigated in future work . note that here we only suggested to make the sample selection in different filters . when also performing the shape measurement at different wavelengths , different structures within the galaxies dominate the light distribution , and thus the alignment properties may change as well .
we can not investigate this effect with our simple analytic models , but the forthcoming multi - band surveys will shed more light on its significance . ultimately , the intrinsic alignment contamination may become a major reason to perform weak lensing studies in a completely different part of the electromagnetic spectrum ( see e.g. * ? ? ?
b - modes generated by intrinsic ellipticity correlations seem to be affected by spurious contributions , so that we can only set an upper limit of @xmath326 . a potential exception is the ii b - mode signal at @xmath339 , which attains similar amplitudes as the corresponding e - mode .
such signals could serve as an independent complement to e - modes to constrain intrinsic alignment models .
future analysis of highly accurate simulations with very good mass and spatial resolution is needed to firmly establish whether intrinsic alignments leave a clear signature in the b - mode , especially at high redshift .
a semi - analytic approach to galaxy shape modelling as proposed here , or alternatively an advanced halo model ( see * ? ? ?
* ; * ? ? ?
* ) , has the major advantage of jointly predicting various shape and clustering statistics for a wide range of galaxy samples .
such models readily combine constraints from large and small scales and include realistic galaxy bias , allowing access to the deeply non - linear regime and thus enabling two - point analyses of small - scale surveys such as cosmos . although the shape models employed in this work are still simplistic and involve strong assumptions , the actual major concern is the lack of observational constraints . to date , there are essentially no informative intrinsic alignment measurements for blue , under - luminous galaxies which , however , are the main constituents of galaxy samples used for weak lensing .
furthermore the amplitude of intrinsic alignments at redshifts of order unity and beyond is currently unknown although the majority of galaxies in forthcoming weak lensing surveys will populate this regime .
future observations need to improve statistical constraints and extend them closer to the depths of weak lensing studies .
this is a challenging task as both accurate galaxy shape measurements and redshift estimates have to be available , as e.g. in the overlap regions of imaging and spectroscopic redshift surveys . reducing shot noise and
sample variance sets tight requirements on number density , area coverage , and survey geometry for the galaxy sample to be studied . since a critical ingredient is a redshift estimate for every galaxy which is accurate to roughly the per - cent level , multi - band photometric surveys like pau , or low - resolution spectroscopy as performed by primus @xcite , targeted at weak lensing surveys , are a promising strategy .
the _ euclid _ mission not only sets the highest requirements on the control of intrinsic alignments , but also provides data ideally suited for their calibration , with unprecedented constraints expected on emission - line galaxies from the spectroscopic survey , and the opportunity of directly measuring the intrinsic alignment signal on the weak lensing sample thanks to the highly accurate photometric redshifts supplied by the deep survey @xcite .
next steps in the simulation - based modelling of galaxy shapes include the usage of a hierarchical suite of n - body simulations which will provide the cosmological volume to improve on statistical power , as well as the mass resolution to determine robust halo properties of low - mass and satellite galaxies .
semi - analytical models that simultaneously determine the photometry and the morphology , based on halo properties and merger history , will establish a more realistic link between the shape of a galaxy and its luminosity and colours , thereby allowing us to examine the efficiency of suppressing the contamination by intrinsic alignments via colour cuts on the galaxy sample .
the results of the hydrodynamic simulations performed by @xcite challenge the paradigm that the different evolution of disc and spheroidal galaxies is caused by the details of the merger history .
they found that the difference between late and early galaxy types can rather be associated with the infall directions and thermo - dynamic state of the accreted gas .
similarly , @xcite concluded from their suite of hydrodynamic simulations that the angular momentum of galactic gas is larger than that of the dark matter halo , and closely linked to the surrounding large - scale structure .
these recent findings question the choice of analytic models of galaxy shapes that are directly derived from halo properties , as taken in this paper .
however , the proposed scenarios are likely to produce distinctively different predictions for e.g. the radial dependence ( due to a stronger link of angular momentum with filamentary structures ) and the evolution ( due to a stronger correlation with gas infall events than with mergers ) of intrinsic alignments , and hence can be put to the test by comparison to intrinsic alignment observations in analogy to our approach . once hydro - dynamic simulations can be run on cosmologically relevant scales with sufficient resolution
, one can by - pass the critical assumptions about the link between the shapes of dark and luminous matter and directly create mock galaxy images on which a standard shape measurement code can then be run .
there would still be critical assumptions involved , most prominently about the sub - grid physics of baryons in the simulation .
for instance , a strong dependence on feedback models , as seen for weak lensing power spectra @xcite , is also expected for intrinsic alignment signals .
however , these issues would be shared with the models of galaxy evolution and of cosmological probes , and hence allow for intrinsic alignments to be integrated smoothly into a wider cosmological context .
we are grateful to rachel mandelbaum for providing us with the wigglez correlation function data , and to catherine heymans for providing us with code to compute aperture mass statistics .
moreover we would like to thank them , as well as chris blake , tim schrabback , andy taylor , and ismael tereno for helpful discussions .
we are grateful to our referee for constructive remarks .
bj acknowledges support by an stfc ernest rutherford fellowship , grant reference st / j004421/1 .
es and hh acknowledge the support of the netherlands organization for scientific research ( nwo ) through a vidi grant , as well as support from the erc under fp7 grant number 279396 .
peb acknowledges support by the deutsche forschungsgemeinschaft ( dfg ) under the project schn 342/71 in the framework of the priority programme spp-1177 , and the initiative and networking fund of the helmholtz association , contract ha-101 ( ` physics at the terascale ' ) .
jh , sh , and ps acknowledge support by the dfg through the priority programme 1177 ` galaxy evolution ' ( schn 342/6 and wh 6/3 ) and the transregional collaborative research centre trr 33 ` the dark universe '
. sh also acknowledges support by the national science foundation ( nsf ) grant number ast-0807458 - 002 .
the simulations used in this paper were carried out as part of the programme of the virgo consortium on the regatta supercomputer of the computing centre of the max - planck - society in garching , and the cosmology machine supercomputer at the institute for computational cosmology , durham .
the cosmology machine is part of the dirac facility jointly funded by stfc , the large facilities capital fund of bis , and durham university . | intrinsic galaxy alignments constitute the major astrophysical systematic of forthcoming weak gravitational lensing surveys but also yield unique insights into galaxy formation and evolution .
we build analytic models for the distribution of galaxy shapes based on halo properties extracted from the millennium simulation , differentiating between early- and late - type galaxies as well as central galaxies and satellites .
the resulting ellipticity correlations are investigated for their physical properties and compared to a suite of current observations . the best - faring model is then used to predict the intrinsic alignment contamination of planned weak lensing surveys .
we find that late - type galaxy models generally have weak intrinsic ellipticity correlations , marginally increasing towards smaller galaxy separation and higher redshift .
the signal for early - type models at fixed halo mass strongly increases by three orders of magnitude over two decades in galaxy separation , and by one order of magnitude from @xmath0 to @xmath1 .
the intrinsic alignment strength also depends strongly on halo mass , but not on galaxy luminosity at fixed mass , or galaxy number density in the environment .
we identify models that are in good agreement with all observational data , except that all models over - predict alignments of faint early - type galaxies .
the best model yields an intrinsic alignment contamination of a _ euclid_-like survey between @xmath2 at @xmath3 and on angular scales larger than a few arcminutes . cutting @xmath4 of red foreground galaxies using observer - frame colours can suppress this contamination by up to a factor of two .
[ firstpage ] methods : statistical methods : numerical cosmology : observations galaxies : evolution gravitational lensing : weak large - scale structure of universe |
understanding the epoch of cosmological reionization has been a key challenge in modern observational cosmology . the gunn - peterson test with spectra of high - redshift ( high-@xmath12 ) quasars
is an observational tool for detecting the reionization epoch , and suggests that the intergalactic medium ( igm ) was reionized at @xmath13 @xcite .
there is also observational evidence that neutral hydrogen still remains in the igm of the universe at @xmath14 , e.g. , studies with quasars @xcite and with gamma - ray bursts ( grbs ) @xcite .
recent analysis of the cosmic microwave background ( cmb ) indicated the redshift of the cosmic reionization of @xmath15 by planck @xcite .
galaxies associated with ly@xmath2 emission , i.e. , lyman-@xmath2 emitters ( laes ) are also valuable probes to infer the neutral hydrogen ( h i ) fraction @xmath16 of the igm .
@xcite suggested the effectiveness of this concept from a theoretical viewpoint .
observational studies indicated that the luminosity functions ( lfs ) of the lyman - break galaxies ( lbgs ) show a monotonic decline in number density at @xmath17 to @xmath18 @xcite .
in contrast , the lfs of laes show little change at @xmath19 , implying an increasing fraction of laes relative to lbgs with increasing redshift ; in turn , they show a significant decline at @xmath13 in the number density , e.g. , @xcite who presented a clear decrease in the ly@xmath2 lf at @xmath20 from the @xmath21 . this rapid decline in the ly@xmath2 lf
can be attributed to a change in @xmath16 at @xmath13 .
the neutral hydrogen in the igm could decrease observed ly@xmath2 photons from star - forming galaxies by resonant scattering@xcite .
@xcite suggested an increase in the h i fraction from @xmath22 to @xmath23 and the estimated @xmath16 is 0.22 at @xmath23 .
they suggested possible cosmic variance in the number density of laes across a @xmath24 sq.degree area by a factor of 2 to 10 depending on their nb921 magnitudes . @xcite
inferred @xmath25 at @xmath23 from a different data set , suggesting existence of the cosmic variance in @xmath16 with a possible patchy reionization process in the universe .
a caveat in investigating reionization with the lfs of galaxies is that the lfs are likely to be influenced by galaxy evolution .
in contrast , the ly@xmath2 fraction , i.e. , the fraction of galaxies exhibiting strong ly@xmath2 emission of the total lbgs , has an advantage in that it should be less affected by galaxy evolution , especially in terms of the number density , than the ly@xmath2 lfs , which require comparison with the uv lfs @xcite .
in addition , star formation activity and dust / metal content of galaxies which contribute to the equivalent width ( ew ) of ly@xmath2 , do not change rapidly even at @xmath26 compared to those at lower redshifts @xcite .
the ly@xmath2 fraction at each redshift has been studied intensively to yield a complementary information to the cosmic reionization process @xcite .
the studies by @xcite and @xcite suggested that the ly@xmath2 fraction of lbgs drops significantly from @xmath27 to @xmath18 in contrast to its monotonic increase at @xmath28 .
their faint sample showed a more rapid decline than the bright sample .
@xcite also studied the ly@xmath2 fraction at @xmath29 and discussed the decline in ly@xmath2 fraction at @xmath30 with the possibility of witnessing the ongoing cosmic reionization process at @xmath31 .
a recent study with hubble space telescope ( hst ) slitless spectroscopy targeting gravitationally lensed clusters of galaxies ( glass ) also indicated that the number of laes at @xmath11 with respect to that of the lbgs is consistent with a lower probability of ly@xmath2 emission at @xmath29 than at @xmath14 @xcite . while the relative number of laes is an effective tracer of the cosmic reionization state , a fair application would require consideration of several factors that may affect the observable ly@xmath2 emission from high-@xmath12 galaxies .
the strength of ly@xmath2 emitted from galaxies may be affected by a combination of physical properties of galaxies and the transmission of the igm along the lines of sight .
@xcite discussed the importance of the coupled effects by internal dust extinction and the geometry and kinematics of h i gas surrounding a galaxy , to determine the net ly@xmath2 photons that can escape from the galaxy .
lbgs with low luminosities tend to have low metallicity or dust extinction , leading to a bluer uv color @xcite and likely association of strong ly@xmath2 emission .
this may introduce a possible luminosity dependence of the ly@xmath2 fraction , as suggested in previous studies @xcite .
@xcite discussed a similar trend in the stellar mass dependence on the strong ly@xmath2 emission .
observational studies suggested that lbgs with high uv luminosities tend to show a deficit in strong ly@xmath2 emission @xcite .
based on spectroscopic observations of lbgs , @xcite reported that the ly@xmath2 fraction of lbgs with low uv luminosities is higher than that of those with high luminosities at @xmath19 .
in addition , the ly@xmath2 fraction ( @xmath32 ) rises as redshift increases over @xmath28 .
@xcite supported this trend at @xmath33 based on a spectroscopic survey of galaxies in the vimos - vlt deep survey ( vvds ) sample .
@xcite reported the same increase in ly@xmath2 fraction ( @xmath32 ) from @xmath34 to @xmath35 at bright magnitudes ( @xmath36 ) .
recent spectroscopic studies have also reported detection of ly@xmath2 in a few sources in a similar luminosity range @xcite .
the redshift above @xmath37 is an important epoch to understand the progression of reionization . in particular , it is of interest to investigate even brighter objects ( @xmath38 ) , i.e. , in environments in very massive dark matter halos .
such massive halos are thought to evolve and form galaxies with strong clustering in an earlier epoch , and so are likely to host larger amounts of ionizing sources than less massive halos .
therefore , it is expected that reionization would progress earlier in the environments around the massive dark matter halos than in other regions . however , despite intensive observational studies of the ly@xmath2 fraction at @xmath13 , only small numbers of the lbgs at @xmath29 are available . primarily due to limitations in the surveyed area and as brighter galaxies are rarer ,
very little is known about the lbgs with very bright magnitudes .
previous studies on the ly@xmath2 fraction were conducted only at uv magnitudes fainter than @xmath39 .
thus , new deep and wide imaging and spectroscopic observations are indispensable to extend our knowledge of the bright magnitude regime .
focusing on such uv - bright lbgs is also an effective approach that facilitates follow - up spectroscopy and allows determination of a reliable ly@xmath2 fraction at high redshifts .
we have undertaken a new survey program to cover an unprecedentedly wide area by a deep imaging survey to detect uv - bright lbgs , followed by spectroscopic observation of some of the high-@xmath12 candidate galaxies . in this study , with the new spectroscopic sample from our survey program , we examine the ly@xmath2 fraction of lbgs at @xmath18 for very bright magnitudes .
we also discuss the cosmic reionization at redshift 7 , providing a new constraint on the cosmic neutral h i fraction . in section 2 , we summarize the observations and data reduction in our @xmath4-band imaging survey program , and explain the method of target selection for lbg candidates at @xmath18 . section 3 describes the follow - up spectroscopy and properties of the spectroscopic sample .
the ly@xmath2 ew of our uv - bright galaxy sample is described in section 4 .
we discuss the ly@xmath2 fraction of lbgs in section 5 , and its implications and interpretation regarding the state of the cosmic reionization in section 6 .
finally , conclusions are presented in section 7 . throughout the paper ,
we use the ab magnitude system @xcite , and assume a flat universe with @xmath40 .
this paper aims to study the ly@xmath2 fraction based on lbg candidates at @xmath18 in the two common fields , cosmos / ultravista and sxds / ukidss - uds ( hereafter , ultravista and uds , respectively ) .
these two survey fields represent a unique and powerful combination for studies on high-@xmath12 galaxies , where extensive multi - wavelength surveys at various facilities have been performed . in particular , the two ground - based deep near - infrared ( nir ) surveys ultravista @xcite and ukidss - uds @xcite provide an unprecedentedly wide and deep data set over an area of 1.65 sq.degree , making the two fields the most suitable for searches of bright lbgs at @xmath18 . initially , however , only relatively shallow optical data sets were available in these fields , especially in the @xmath4 band , which is crucial for detecting the dropout feature of the lyman - break of the @xmath18 galaxies , e.g. , cosmos imaging @xcite by the subaru telescope covering the ultravista field in the @xmath41 and @xmath4 bands , with @xmath42 ( 5 @xmath7 , 3-arcsec aperture ) , and the subaru / xmm - newton deep survey ( sxds ) @xcite located at the uds field in the @xmath43 , and @xmath4 bands to a depth of @xmath44 ( 5 @xmath7 , 2-arcsec aperture )
. this situation motivated us to undertake a new observational campaign of ultra - deep imaging in the @xmath4 band in the two fields .
we use a combined sample of the lbgs presented previously by @xcite , and a small addition to this work in the uds field as described below . in the following sections , we first present an overview of the entire @xmath4-band imaging campaign and data reduction . then , construction of the sample of lbgs in each field is explained .
the deep imaging observations in the @xmath4 band were conducted with suprime - cam @xcite using the 8.2 m subaru telescope , which uses fully depleted ccds that were introduced in 2008 august .
our survey plan was designed to cover the two target fields ultravista and uds ( @xmath45 sq.degree ) with eight pointings of suprime - cam , the field of view of which is @xmath46 sq.arcmin .
figure [ fig : pointing ] summarizes this pointing strategy .
each pointing was planned to be integrated by 20 hours , divided into shorter exposures .
the observing times were awarded as an open - use subaru intensive program ( pi : furusawa ; s08b-051 ) . originally , 20 nights were allocated over three semesters from 2008 october to 2009 november . to compensate for low completion rate due to poor weather conditions , etc .
, another 16 nights were added by a regular open - use program and the observatory s discretionary time .
final exposure times in each pointing , which are merged into the stacked images are 19.6 , 18.3 , 18.7 , and 20.8 hours for the ultravista field , and 11.7 , 16.2 , 11.7 , and 15.5 hours for the uds field .
the @xmath4-band data obtained in this campaign are summarized in table [ tab : zdata ] .
data reduction of the @xmath4-band data performed to generate mosaic - stacked images from raw ccd images , following the procedure described in @xcite .
the sdfred2 software @xcite provided by the observatory was used .
small fluctuations due to cross - talk between amplifiers of the ccd were corrected @xcite for the uds field data . the full width half - maximums ( fwhms ) of the point spread function ( psf ) of the stacked images before any convolution or transformation are 0.76 , 0.84 , 0.70 , and 0.82 arcsec for the ultravista field , and 0.75 , 0.81 , 0.78 , and 0.81 arcsec for the uds field .
the celestial coordinates of the stacked images were calibrated with the usno - b1.0 catalog @xcite , using sextractor and scamp .
the resultant residuals of the determined coordinates from the usno - b1.0 catalog coordinates are on the order of 0.5 to 0.8 arcsec rms .
this accuracy is good enough to perform image warping in section [ sec : sample_selection_uds ] .
magnitude zero points of the stacked images were determined with a standard star gd71 ( @xmath47 ) .
the estimated depths in the @xmath4 band are 26.6 , 26.5 , 26.5 , and 26.8 ( 5 @xmath7 , 2-arcsec aperture ) for the ultravista field , and 26.3 , 26.6 , 26,6 , and 26.6 for the uds field .
this data set is the deepest ever achieved in the @xmath4 band covering the entire two deep fields ( @xmath45 sq.degree ) .
sample selection of @xmath18 galaxies in the ultravista field ( ra @xmath48 , dec .
@xmath49 , j2000 ) was performed by @xcite , in an update to their initial work @xcite . in the field ,
the ultravista dr2 provides nir photometry in the @xmath50 , and @xmath51 bands ( @xmath52 in ultradeep strips ; 5 @xmath7 , 1.8-arcsec aperture ) over a field of 0.91 sq.degree e.g. , figure 1 in @xcite .
they combined the ultravista - dr2 data with the multi - waveband data including the cfhtls data in the @xmath53 bands ( @xmath54 ; 5@xmath7 , 2-arcsec aperture ) , and the final @xmath4-band data set ( table [ tab : zdata ] ) provided by our imaging campaign .
the high-@xmath12 candidates were selected based on a photometric redshift ( @xmath55 ) analysis applied to the sources detected in the @xmath56 or @xmath57 band under conditions with no detection in the @xmath58 or bluer bands .
the resultant @xmath18 lbg sample in the ultravista field contained 19 candidates with @xmath59 , including 13 sources at @xmath60 .
the uds - dr10 data ( omar almaini , private communication ) covers an area of 0.74 sq.degree centered on the uds field ( ra @xmath61 , dec .
@xmath62 , j2000 ) to depths of @xmath63 , and @xmath64 ( 5 @xmath7 , 2-arcsec aperture ) .
@xcite performed selection of @xmath18 lbgs in the uds field with photometric redshift analysis using the multi - waveband data set including the nir uds ( @xmath65 , and @xmath66 ) data and the optical sxds ( @xmath43 , and @xmath4 ) data .
the data set also included the @xmath4-band data in our imaging campaign .
however , the @xmath4-band data were of interim status , which consisted of 11.1 , 16.2 , 11.2 , and 15.5 hours slightly shallower in the uds1 and uds3 than the final data set listed in table [ tab : zdata ] .
additional @xmath4-band data have been obtained since their study .
their lbg sample in the uds field included only two candidates with @xmath59 in contrast to the ultravista field .
this relatively small number of high-@xmath12 candidates in the uds field may have been due to their requirement of significant detection in the @xmath67 band to avoid contamination by severe cross - talk from the ukirt / wfcam , which is only applied to the uds field . to add as many @xmath18 candidates to their lbg sample in this field as possible , in this study we used the combined data of the uds - dr10 , the sxds , and the updated @xmath4-band data set with the final depths ( table [ tab : zdata ] ) .
moreover , we applied a different sample selection to the data from that adopted in @xcite .
our additional sample selection procedure is described below .
we transformed the @xmath4-band stacked images in the uds field onto the same pixel coordinates as those of the @xmath56-band image .
the images in the @xmath4 to @xmath66 bands were convolved with a single gaussian smoothing kernel to have the same fwhms ( 0.83 arcsec ) of the psf as the @xmath56-band images . on the @xmath68 images , we ran sextractor ( ver .
2.19.5 ) in double - image mode , with detection on the @xmath56-band image , to generate @xmath56-band detected multi - waveband catalogs ( @xmath69 ; 5 @xmath7 , 2-arcsec aperture ) , separately for each of the four pointings .
we also removed sources located at pixels that were affected by low signal - to - noise ratio ( s / n ) around the image edges , strong blooming , and strong halos of bright stars .
we applied the following condition to the magnitude - limited sample of galaxies : ( 1 ) @xmath70 , ( 2 ) @xmath71 , and ( 3 ) complete dropout in all of the @xmath72 and @xmath73 bands .
the first condition involves selection of the lyman - break , which together with the second condition minimizes contamination by galactic lt - dwarf stars .
the third condition is to reduce contamination by lower - redshift galaxies .
the resultant intermediate catalogs that satisfied the above conditions included 659 , 757 , 725 , and 700 sources in each pointing , respectively . to remove false sources due to severe cross - talk on the @xmath74 images , we visually inspected every candidate source in the intermediate catalogs by eye , instead of requiring detection in the @xmath67 band . the false signals due to cross - talk appeared at positions of 191 pixels and/or its multiples from bright root signals in either the x- or y - axis , in each of the @xmath74 bands .
these locations of cross - talk signals are inherited from the characteristics of the read - out electronics of ukirt / wfcam .
if the candidate is an artifact of cross - talk , there must be a root signal that causes a series of cross - talk located at one of the positions of the multiples of 191 pixels from the candidate .
importantly , a series of false signals decays monotonically with distance from the root signal , and also does not go beyond the quadrant area where the root signal is located .
therefore , for every candidate source , we inspected 24 locations with distances of 191 and 282 pixels from the candidate on the x- and y - axes and grid points according to their combinations . in cases where there is any possible false signal at these locations
, we excluded the candidate source as it is likely to be a false signal .
we finally selected three new candidates of galaxies at @xmath11 in the uds field , which are listed in table [ tab : uds_cand ] .
figure [ fig : stamps ] shows postage stamp images of the three sources .
the three candidates , fh2 - 22303 , fh2 - 48620 , and
fh4 - 42903 , have bright @xmath56 magnitudes ( 24.3 , 25.4 , 25.3 ; 2-arcsec aperture ) , or rest - frame uv magnitudes of @xmath75 , @xmath76 , and @xmath77 , respectively .
photometric redshifts were derived using data from all of the broad bands from @xmath78 to @xmath66 bands based on the same method employed in @xcite .
the resultant @xmath55 are @xmath79 , and @xmath80 for fh2 - 22303 , fh2 - 48620 , and fh2 - 42903 , respectively , supporting the high redshifts @xmath11 of the candidate galaxies .
the associated errors correspond to their @xmath81 confidence levels .
the best - fit spectral energy distributions ( seds ) for the three sources are shown in figure [ fig : photoz_fit ] .
the first two sources ( fh2 - 22303 and fh2 - 48620 ) are brighter than all @xmath18 galaxies studied to date by spectroscopy for the ly@xmath2 fraction @xcite , if confirmed at @xmath18 .
we confirmed that the number density of the combined data set ( @xmath82 ) of the sample by @xcite and the new 3 sources @xmath83 , is consistent with the uv lf presented by @xcite .
we selected a total of 24 photometric sample of galaxies at @xmath84 .
the sample consists of 19 galaxies in the ultravista field , 2 in the uds field presented by @xcite , and the 3 new galaxies in the uds field added in this study . of the 24 lbg candidates at @xmath18 , 18 galaxies have very bright uv magnitudes of @xmath38 ( 14 galaxies in the ultravista , and 4 in the uds fields ) . as discussed in section [ sec : spec ] , nine galaxies are spectroscopically followed up in this study , and seven of these belong to the very bright magnitude range .
we conducted spectroscopic observations of nine sources of the 24 @xmath18 candidate galaxies selected in the previous section .
these targets were chosen to include the four high - priority sources in the ultravista field , which were qualified as ` robust ' in @xcite , and all the five sources in the uds field .
the ` robust ' targets were categorized based on results of @xmath55 analysis that did not accept either galactic stars or low-@xmath12 dusty galaxies in @xcite .
a summary of the target sources is shown in table [ tab : spec_fluxlimit ] .
the ids of all the sources presented in @xcite are taken from table 2 of @xcite with the prefix ` b14- ' , containing four sources in the ultravista field and two in the uds field .
the three sources with the prefix fh are the sources in the uds field added in this study .
the observations were performed on the four nights of 2013 march 5 and 6 and 2014 october 24 and 25 , in multi - object spectroscopy ( mos ) mode with subaru / focas @xcite .
the vph900 grism combined with the oh58 order - sort filter was employed , covering a wavelength range of 7,500 to 10,450 giving a spectral resolution of @xmath85 ( 0.74 pixel@xmath86 ) with a slit width of 0.8 arcsec .
this combination has the highest throughput at the target wavelengths among the 8-m class telescopes , especially at around @xmath87 , and is therefore highly suitable for the follow - up spectroscopy of ly@xmath2 emission at @xmath18 .
each mos mask has a single target source at @xmath88 .
the integrated times of 1.3 to 6 hours are devoted to each source ( table [ tab : spec_fluxlimit ] ) , being split into individual exposures of 1,200 s , with a dithering width of 1.0 arcsec in the spatial direction between exposures .
data analysis was performed in a regular manner using the standard pipeline focasred and mosred , implemented based on iraf and provided by the observatory .
the raw spectrum data were first bias - subtracted and flatfielded by domeflat data .
the pixel areas of target spectra were extracted by applying distortion correction based on a predefined pattern .
then , wavelengths were calibrated based on night sky oh emission lines embedded in the 2d spectra , and sky background was subtracted from the 2d spectra .
the reduced 2d spectra of individual exposures were shifted in the spatial direction by the dithering width and combined by taking median with 3-@xmath7 clipping of outlying pixels .
flux calibration was performed on the stacked 2d spectra using spectrophotometric standard stars gd153 , g191b2b , and feige110 , with iraf tasks ` standard ' , ` sensfunc ' , ` fluxcalib ' , and ` extinction ' . with these tasks ,
the atmospheric extinction was corrected assuming the atmospheric attenuation as a function of wavelength at maunakea , which was measured by cfht .
this procedure also removed features of the flatfield as a function of wavelength .
the overall fitting error was @xmath89 rms in this calibration , which is good enough for the following discussion .
we determined a spatial range of the source on the calibrated 2d spectrum using the iraf task ` prows ' , and extracted a 1d spectrum by summing fluxes within a spatial range of 6 pixels or 1.2 arcsec centered at the flux peak .
the effects of slit - loss on the measured fluxes of each source were corrected by estimating the missing fluxes based on the seeing size during the spectroscopic observations and the intrinsic sizes of the sources on the @xmath56-band images .
the estimated slit - loss fractions ranged from @xmath90 to @xmath91 . only a source b14 - 065666 displays a possible ly@xmath2 emission at @xmath11 , while the other eight sources showed no convincing emission line features to the s / n level of 3 , i.e. , no detection of ly@xmath2 .
figure [ fig : b14 - 065666spec ] shows the 2d and 1d spectra of the b14 - 065666 at around wavelengths including the emission line profile .
it was confirmed that the line is detected in each of 2d spectra generated separately for different dither positions , and spatial positions of the line are consistent with the dithering pattern .
only a single line feature is detected in the source b14 - 065666 .
this suggests that the line is unlikely to be h@xmath2 @xmath92 , as [ nii]@xmath936548,6584(s / n@xmath94 ) lines are not visible within the covered wavelength range on the spectrum .
in addition , the red color of @xmath95 can not be explained by ordinary galaxy seds .
this line is unlikely to be the [ oiii ] , because the doublet lines [ oiii]@xmath934959,5007 @xmath96 apart could be resolved assuming line ratio@xmath97 in the areas that are free from the oh emission , although the s / n ( @xmath98 ) is not sufficiently large .
in addition , the red @xmath99 color can not be explained due to lack of any break feature around the wavelength range .
the [ oii ] line in this case was located out of the wavelength coverage and could not be detected .
the likelihood of the [ oii ] is also small , as the double lines [ oii]@xmath933727,3729 ( 5.3 apart in the observer s frame ) should be resolved , although the s / n may not be sufficiently large assuming line ratio @xmath24 near an oh emission band . nonetheless , the photometric redshift of @xmath100@xcite and the red @xmath99 color of the source could exclude being a lower-@xmath12 galaxy at @xmath101 .
the property of the possible ly@xmath2 emission was measured on the 1d spectrum of b14 - 065666 . by profile fitting with the gaussian function by the iraf task ` splot ' , the line profile is centered at @xmath102 , which corresponds to a redshift of 7.168 for ly@xmath2 .
this spectroscopic redshift agrees with the @xmath55 estimated by @xcite .
the measured ly@xmath2 flux density is @xmath103 erg s@xmath104 @xmath105 ( s / n of 5.5 ) , and the ly@xmath2 luminosity was @xmath106 erg s@xmath104 ( table [ tab : spec_fluxlimit ] ) . here
, the flux error was estimated by fluctuation of sky background counts over the wavelength range of the 2d spectra where the line profile is located , avoiding the severe oh skylines .
this luminosity is fainter than @xmath107 of the ly@xmath2 lf in the redshift range by a factor of @xmath45 @xcite to @xmath108 @xcite . as the other eight sources did not show any possible emission lines , we studied noise statistics of their 2d spectra to estimate the observational upper limits of their ly@xmath2 fluxes . to measure the noise fluctuation , we sampled about 500 locations on the 2d spectra in the wavelength range corresponding to ly@xmath2 at @xmath109 to @xmath110 for five sources , including both regions free from and affected by the oh skylines .
each sampling area covered a rectangular area of 1 arcsec in the spatial direction and a width of 25 in the wavelength direction , as a typical ly@xmath2 observed line width .
for the other three sources ( b14 - 169850 , b14 - 035314 , b14 - 118717 ) , @xmath55 of which are below @xmath111 @xcite , we chose the sampling wavelength range to cover the redshifts spanning @xmath112 .
this range corresponds to the maximum 1-@xmath113 error of all the photometric sample galaxies as a conservative choice .
the histogram of the measured counts was fitted by the gaussian function and the 1-@xmath7 error was estimated for each candidate source in the same manner as adopted in @xcite .
we calculated the 3-@xmath7 upper limits of the ly@xmath2 flux density , which are also shown in table [ tab : spec_fluxlimit ] .
we derived the upper limit of the ly@xmath2 flux ew for our candidate sources based on the uv luminosities and the 3-@xmath7 upper limit of the ly@xmath2 flux determined in the previous section .
the ew upper limit in the observer s frame ( @xmath114 ) is converted to that in the rest - frame ( @xmath115 ) on the assumption of @xmath18 , except the source of b14 - 065666 ( @xmath5 ) .
the estimated upper limits of @xmath115 range from @xmath116 to @xmath117 , which are shown in table [ tab : spec_fluxlimit ] .
the uv magnitudes were calculated from the @xmath56-band magnitudes derived from mag_auto and assuming a flat spectrum of galaxies located at @xmath18 . for the only source b14 -
065666 exhibiting a possible ly@xmath2 emission line , the line flux was converted to @xmath115 using a uv magnitude of @xmath118 as a continuum flux , assuming redshift of @xmath119 .
the resultant @xmath115 is @xmath120 , which is consistent with the tendency for uv - bright galaxies to have small @xmath115 even in this very bright magnitude range ( @xmath121 ) .
the relations between ly@xmath2 @xmath115 and uv magnitudes of galaxies in different redshift bins are shown in figure [ fig : ew0_z_scatter ] .
our data are compared with previous studies , which provide the following results .
first , in panels ( a)(c ) for @xmath122 , the galaxy sample by @xcite includes galaxies with very bright magnitudes ( @xmath38 ) .
however , only a few galaxies show very small @xmath115 of ly@xmath2 emission in this magnitude range .
second , in panels ( c ) and ( d ) , our data are shown with the upper limits to the @xmath115(ly@xmath2 ) , which provide a new galaxy sample to the very bright magnitude range at @xmath11 .
these galaxies are associated with very small @xmath115 of @xmath123 .
our data suggest the existence of a considerable fraction of galaxies with low @xmath115(ly@xmath2 ) at @xmath11 at the brightest uv magnitudes ( @xmath121 ) studied to date ; only a source with @xmath124 at @xmath18 in this magnitude range was reported @xcite prior to this study . in panel ( c ) @xmath125 ,
the data points by @xcite seem to be located at systematically higher @xmath115 than our data . however , their data points at @xmath38 in panel ( c ) should be at @xmath126 based on their ew ( 540 to 1250 ) in the observer s frame .
in addition , with an independent sample based on photometric redshifts , @xcite indicated that @xmath127 of bright galaxies ( @xmath128 ) at @xmath129 have small @xmath115 of @xmath130 , including galaxies with very small @xmath115 ( @xmath123 ) .
therefore , there should be no marked difference in the relation of @xmath115 and uv magnitudes at @xmath13 between the previous studies @xcite and this study .
finally , panel ( e ) compares our results for the lbg candidates with those on the narrow - band selected laes in a previous study that surveyed a @xmath131 sq.degree field @xcite .
our uv - selected galaxies show lower @xmath115 than the laes in the same redshift range .
we aim to obtain a new constraint on the ly@xmath2 fraction for the very bright magnitude range at @xmath18 , which is studied here for the first time . the new result should be complementary to previous results at fainter magnitudes to understand a luminosity dependence of the ly@xmath2 fraction . to derive our results according to the same formalism as adopted in previous studies @xcite
, we use the fraction of galaxies with a rest - frame ly@xmath2 @xmath115 larger than 50 , @xmath132 .
previous studies discussed @xmath132 in the two magnitude ranges ( bright : @xmath133 , and faint : @xmath134 ) . as seven sources of our nine sample galaxies
are brighter ( table [ tab : spec_fluxlimit ] ) than the above bright magnitude range , we introduce a new magnitude range of ` very bright ' ( @xmath0 ) .
we chose the threshold @xmath115 of 50 , as the @xmath132 in the bright magnitudes ( @xmath135 ) at @xmath136 presented by @xcite is only the previous result that can be directly compared with our data at @xmath18 .
it is essential to compare our ly@xmath2 fraction at @xmath18 with those at lower redshifts to discuss a change in the ly@xmath2 fraction and progress in the reionization at these epochs .
we estimated the @xmath132 as follows .
first , the seven sources with uv magnitudes spanning from @xmath137 to @xmath76 ( @xmath121 ) were selected .
then , weights for each source were derived by the inverse of detection completeness of each source as a function of brightness .
the derived weighted numbers were summed , yielding a weighted total number of sample galaxies of @xmath138 for all seven sources . finally , as none of the sample galaxies show ly@xmath2 emission with @xmath8 , we estimated upper limit of the ly@xmath2 fraction . here , as an upper limit of the number of sources , we adopted a @xmath139-@xmath7 upper confidence level ( 1.84 ) for the observed count zero based on the poisson distribution @xcite .
the @xmath132 were estimated by dividing this upper limit @xmath140 by the weighted total number of sample galaxies @xmath138 . in this procedure ,
the completeness values for the four sources in the ultravista field and one in the uds field were taken from @xcite .
we conducted a monte carlo simulation to determine the completeness of the other sources in the uds fields ( section [ sec : uds_selection ] ) .
each simulation was performed by adding 1,000 artificial point sources with known magnitudes at random positions on the @xmath56-band image , and detecting them again .
the completeness was calculated as the fraction of recovered sources among the input sources .
the simulations were repeated 20 times and the completeness of each simulation was averaged to determine the effective completeness .
we confirmed that the completeness values from @xcite and this study are consistent and the possible uncertainty if any should not change the following discussion .
we did not correct for an effect of the oh skylines on detection of the ly@xmath2 emission assuming the finding by @xcite .
their simulation showed that over @xmath141 of simulated ly@xmath2 lines were recovered and the effect could be ignored at the same wavelengths as those in this study .
moreover , even if b14 - 065666 were not a true ly@xmath2 detection , the following results on ly@xmath2 fraction would not be changed .
figure [ fig : xlya50_z ] shows changes in the ly@xmath2 fraction ( @xmath142 ) over @xmath143 in the three different uv magnitude ranges : ( a ) faint ( @xmath134 ) , ( b ) bright ( @xmath133 ) , and ( c ) very bright ( @xmath144 ) .
our data provide upper limits to the @xmath132 at @xmath18 in the very bright range for the first time .
the following findings are obtained .
first , in the ( c ) very bright magnitude range , our data point at @xmath18 possibly implies that the ly@xmath2 fraction at @xmath18 remains at a similar fraction or lower than that at @xmath27 , although the data points have large errors . in the same panel , the data points derived from @xcite may support a rapid increase in @xmath132 from @xmath145 to @xmath27 .
@xcite suggested that the ly@xmath2 fraction shows a significant decline at @xmath18 from @xmath27 , which is also seen in panels ( a ) and ( b ) at the fainter magnitudes .
our results are consistent with this finding even in the very bright magnitude range .
second , in comparison between panels ( b ) and ( c ) , our upper limit may indicate a weak trend of the ly@xmath2 fraction in the very bright magnitude range ( @xmath144 ) at @xmath18 consistently being at a similar level to that at fainter magnitudes .
the possibility of a slightly higher ly@xmath2 fraction at @xmath18 in the ( c ) very bright range than in the ( b ) bright range also can not be excluded .
this result seems to extend the relation at @xmath146 @xcite where the ly@xmath2 fraction at very bright magnitudes ( @xmath147 ) is at the same level or slightly higher than that of fainter magnitudes ( @xmath148 ) .
a similar trend was suggested by @xcite in their bright magnitude range ( @xmath36 ) at @xmath14 .
they suggested that the ly@xmath2 fraction for the bright sample may be significantly higher ( by @xmath149 or a factor of @xmath45 ) than that of the fainter sample @xcite .
our finding may present some hints of a different trend from that suggested by @xcite , where the ly@xmath2 fraction for the faint ( ( a ) @xmath134 ) sample is consistently higher than that of the bright ( ( b ) @xmath133 ) sample at @xmath150 .
finally , our upper limit may imply that the ly@xmath2 fraction in the ( c ) very bright magnitudes at @xmath18 is comparable to or lower than that in the ( a ) faint magnitude range at the same redshifts .
the results in the previous section could provide only weak implications for the ly@xmath2 fraction of bright galaxies due to the current large errors .
nevertheless , the trend seen in the ly@xmath2 fraction may possibly imply the existence of physical mechanisms that cause the lbgs with large luminosities to have an ly@xmath2 escape fraction similar to or higher than that of fainter lbgs ( @xmath151 ) .
for example , @xcite discussed a possible excess at the bright end of the uv lf at @xmath18 by extrapolation of the conventional schechter function .
the similar excess of the ly@xmath2 lfs has also been studied by @xcite .
the excess in the very bright galaxies would require consideration of some special physical processes of galaxy formation , such as lack of an efficient mass - quenching mechanism @xcite in this epoch .
@xcite implied that the feedback from a central active galactic nucleus ( agn ) of galaxies , where star formation rate is suppressed by outflows from a luminous agn , could be a mass - quenching mechanism .
if the agn feedback is weak in the bright galaxies at @xmath18 , destruction of the h ii regions associated with star formation would be reduced , which supports higher luminosities of ly@xmath2 emission . at the lower redshifts , where the agn feedback is thought to work more efficiently , star formation in bright galaxies
is more suppressed than in faint galaxies .
the bright galaxies would have higher metallicity in the gas , and possibly more dust than in the counterparts in @xmath29 ; hence , the ly@xmath2 emission would be less efficient at the lower redshifts .
in addition , the outflow from a galaxy , if any , may help an increase in the ly@xmath2 escape fraction , due to scattering of the ly@xmath2 lines , e.g. , @xcite . moreover , the very bright lbgs may be located in massive dark matter halos where the reionization progresses in an earlier epoch . this trend may lead to a higher transmission of the igm than in less massive halos .
our result seems to be consistent with this trend .
it is difficult to clearly interpret the observation that the ly@xmath2 fraction in the very bright range at @xmath129 is still higher than that in the bright range at the same redshifts
. this could be in part accounted for by magnification of the sample galaxies due to gravitational lensing by their foreground galaxies @xcite , although the result can not be fully explained .
we note that the above discussion depends on the uncertainty of measurements of the ly@xmath2 fraction at each redshift .
as the error bars for the very bright sample galaxies at both lower redshifts reported by @xcite and at @xmath18 examined in this study are still large , we can not draw strong conclusions regarding the trend based only on the existing data , and further spectroscopic samples are necessary to obtain more reliable constraints . now that we have derived the ly@xmath2 fraction in the unprecedentedly bright uv magnitude range , it is interesting to derive a new constraint on the neutral h i fraction ( @xmath16 ) in the cosmic reionization epoch . the cumulative probability distribution function ( pdf ) of ly@xmath2 @xmath115 , which is the probability of galaxies having a rest - frame @xmath115 larger than a given value , has been used to discuss the neutral h i fraction with faint spectroscopic samples of galaxies ( @xmath134 ) at @xmath18 in previous studies @xcite .
@xcite reported that the neutral h i fraction may be 0.6 to 0.9 by comparing their ew - pdf with the model of @xcite , which is consistent with those suggested by @xcite and @xcite .
@xcite obtained @xmath152 at @xmath18 with 22 relatively faint @xmath12-dropout galaxies ( @xmath153 ) , followed up by fors2 of the keck telescope .
the slight difference in resultant neutral h i fraction from that of @xcite may have been due to difference in the magnitude ranges studied .
figure [ fig : pdf_ew0 ] shows the resultant ew - pdf of our new spectroscopic sample in the very bright uv magnitude range ( @xmath144 ) , compared with the results reported in the literature at the same redshift @xmath18 .
we estimated the upper limit of ew - pdf at @xmath154 as follows .
of our seven spectroscopic sample galaxies in the very bright magnitude range ( i.e. , without b14 - 118717 and fh4 - 42903 ) , the largest value of the observational upper limit of the @xmath115 was @xmath155 ( fh2 - 48620 ; table [ tab : spec_fluxlimit ] ) .
hence , we assume that the ly@xmath2 emission with @xmath156 , if any , would be detected in the seven spectroscopic data , but none of the seven sources shows such an emission line .
therefore , the ew - pdf upper limit was derived by dividing the @xmath139-@xmath7 upper limit @xmath140 by the weighted total number of sample galaxies @xmath138 , to ew - pdf(@xmath157)@xmath158 ( filled square with downward arrow ) . here , we assume the non - detection of ew to 3-@xmath7 significance based on the noise statistics averaged over the studied wavelength range , which includes both areas with and without oh emission
. this assumption may be slightly optimistic for this small ew range , as there is a chance that possible lines , if any , would be located in the relatively noisier oh emission bands .
nevertheless , the essence of the following discussion is not changed by the uncertainty .
although we assume @xmath18 for all the sample galaxies except @xmath5 of b14 - 065666 , our discussions in this study would also not be changed by making use of @xmath55 in estimation of @xmath115 .
the typical difference in the upper limits determined by @xmath18 and @xmath55 is sufficiently small ( @xmath159 ) .
we compare our data with the model prediction of ew - pdf presented by @xcite .
their model was generated by computing pdf of a fraction of ly@xmath2 photons transmitted through the igm , combining models of galactic shell - wind outflows with large - scale semi - numeric simulations of reionization .
they determined the reference model ew - pdf at @xmath27 ( @xmath160 ) to coincide with the median ly@xmath2 fraction ( @xmath161 ) at @xmath27 by @xcite .
the data point at @xmath14 by @xcite in the very bright magnitude bin at @xmath162 ( open circle ) is also plotted in figure [ fig : pdf_ew0 ] , which is still consistent with the reference ew - pdf within the error bars .
we estimated the ew - pdf models for the various neutral h i fractions at @xmath18 by rescaling the reference ew - pdf ( see caption of figure [ fig : pdf_ew0 ] ) .
it is noted that the model of @xcite was calculated for the uv magnitude range @xmath133 , which is fainter than that of our data , due to the dark matter halo mass range considered in their calculation . nevertheless , inclusion of more massive halos , which is naively interpreted as the inclusion of uv - brighter galaxies , should produce little if any change in the result ( mark dijkstra , private communication ) .
hence , we simply compare their model with our data .
our upper limit would not contradict the model curves of @xmath163 .
although this is a speculative comparison due to the large error and the small sample size , our data may possibly imply @xmath16 around @xmath164 to @xmath165 with the assumed escape fraction of @xmath166 .
this @xmath16 is consistent with the report of @xcite in the fainter magnitude range .
this result in the new very bright magnitude range may give further support to the findings discussed in previous studies that the reionization of the universe progresses rapidly from @xmath18 to @xmath167 .
we examined the same comparison with the models at fainter magnitudes ( @xmath168 ) using the data points in @xcite , and found an inferred @xmath16 of @xmath169 .
the combination of the results in the very bright and faint magnitudes may favor @xmath170 on the assumption of the escape fraction @xmath166 . based on our data together with the previous results of @xcite and @xcite at the same redshift
, the possibility of lower @xmath171 can not be excluded . as shown by the long - dashed curve ( magenta ) , evolution of the escape fraction of ionizing photons from galaxies
may contribute to the net change in ew - pdf .
the possible increase in escape fraction may significantly reduce the ly@xmath2 fraction under the same igm transmission , which would lower the @xmath16 required to explain the observed ew - pdf at @xmath18 .
moreover , @xcite mentioned the effect on their model of the possible existence of lbgs with ly@xmath2 absorption ( @xmath172 ) at high redshifts @xmath13 , which was reported by @xcite for @xmath17 lbgs .
they suggested that the ew - pdf model would be shifted by @xmath173 by considering lbgs with @xmath172 .
such a shift may explain the ly@xmath2 fraction ( @xmath174 , @xmath175 ) at @xmath176 of @xmath177 presented by @xcite .
if this is the case at @xmath18 in this study , the possible @xmath16 range would be further decreased to @xmath178 or lower . however ,
if we rescale the @xmath27 reference ew - pdf model to exactly match the data point of @xcite ( open circle ) , even higher h i fractions than the @xmath170 would be favored . together with the above discussion
, the lack of any strong ly@xmath2 , such as @xmath179 , in our spectroscopic sample may imply a hint of the unlikely very low @xmath16 at @xmath18 .
thus , further investigations to gain a better understanding of the physical properties and evolution of galaxies at high redshifts are necessary to provide stringent constraints on @xmath16 .
in addition , the number of spectroscopic samples of galaxies is still limited , and therefore future efforts to increase the size of the sample from low to high luminosity ranges , for reliably determining the ly@xmath2 fraction will be indispensable .
we conducted a search for uv - bright lbgs in two legacy survey fields , ultravista and ukidss - uds , which cover an area of 1.65 sq.degree ( @xmath180 ) .
very deep imaging observations on the two survey fields have been undertaken in the @xmath4 band ( @xmath181 ; 5@xmath7 , 2-arcsec aperture ) with subaru / suprime - cam .
we performed selection of @xmath18 candidates in the uds field from the multi - waveband catalogs ( @xmath182 ; 5@xmath7 ) with the updated @xmath4-band data , based on the red color in @xmath99 , combined with relatively mild to red color in @xmath183 , and dropout in all optical bands . in the uds field , we chose three candidates of possible uv - bright galaxies at @xmath18 .
the 19 candidate sources in the ultravista field and two sources in the uds field at @xmath84 presented by @xcite were combined with the three new sources in the uds field , yielding 24 targets in total .
the spectroscopic observations were obtained on nine sources of the @xmath18 candidates with subaru / focas .
only a single source b14 - 065666 shows possible ly@xmath2 emission at @xmath5 , while the other eight sources show no emission line features . by measurements of noise fluctuations on the 2d spectra
, the observational upper limits of the 3-@xmath7 upper limits of the ly@xmath2 fluxes and @xmath115 were estimated based on these eight sources .
the upper limits of ly@xmath2 @xmath115 span @xmath116 to @xmath117 for these sources .
the only source showing a possible ly@xmath2 emission , b14 - 065666 , has an estimated @xmath115 of @xmath120 with a uv magnitude of @xmath118 .
this result supports a tendency for uv - bright galaxies to likely have small @xmath115 even at very bright magnitudes ( @xmath121 ) .
the upper limits of ly@xmath2 ew at @xmath18 were compared with those of previous studies at @xmath184 .
our new data provide clear upper limits to the @xmath115(ly@xmath2 ) at @xmath11 in the very bright uv magnitude @xmath121 , implying that a considerable fraction of galaxies with very bright magnitudes ( @xmath121 ) have low ew at @xmath11 .
based on the upper limits of the ly@xmath2 ew , the 1-@xmath7 upper limits of ly@xmath2 fraction with thresholding @xmath185 ( @xmath132 ) at @xmath18 was derived for the brightest magnitude range studied to date ( @xmath0 ) .
while the ly@xmath2 fraction may support a rapid increase from @xmath145 to @xmath27 even in this bright magnitude range , it may possibly imply leveling off , remaining at the same level or below at @xmath18 .
this result may support the findings at fainter magnitudes reported by @xcite .
our data may also indicate a weak trend whereby the ly@xmath2 fraction at @xmath144 may be similar to that of the fainter magnitudes at @xmath186 , possibly up to @xmath18 .
a slightly higher fraction can not be ruled out . with a large uncertainty
, this result may witness the earlier progression of reionization in more massive dark matter halos , as well as a possible physical mechanism for providing a higher ly@xmath2 escape fraction in the very uv - bright galaxies at redshift 7 . finally , we discussed the ly@xmath2 ew - pdf derived from the upper limits of the ly@xmath2 fraction at @xmath18 , to provide a new constraint on the neutral h i fraction ( @xmath16 ) .
our resultant ew - pdf at @xmath187 at the very bright magnitudes was combined with previous studies at fainter magnitudes by @xcite and @xcite .
we performed a speculative comparison of the observed ew - pdf with the model of @xcite .
although the constraint is not strong given the large error bars , we derived the neutral h i fraction of @xmath188 favored by our data , leaving open the possibility of lower @xmath16 depending on evolution of the physical properties of galaxies , such as the escape fraction . the same comparison with ew - pdf for the fainter counterparts ( @xmath168 ) may support @xmath189 , which is consistent with or below the value for the brighter galaxies .
however , further compilation of spectroscopic observations is necessary to determine more reliable constraints on the ly@xmath2 fraction and reionization at @xmath29 based on a firm understanding of the evolution of galaxies .
we thank the anonymous referee for helpful comments that have improved the manuscript .
we are grateful to rebecca bowler for her valuable contribution to the proposals for spectroscopic observations .
we express our gratitude to yutaka ihara , yuko ideue , nana morimoto for their contributions to imaging observations , data reduction , and data evaluation , and kimihiko nakajima for data evaluation .
we are grateful to takashi hattori for supporting observations and data reduction .
we thank omar almaini for preparing the uds data , and mark dijkstra for providing his latest theoretical model of the ly@xmath2 ew - pdf .
the subaru imaging observations were assisted by yoshiyuki inoue , misae kitamura , kohki konishi , mariko kubo , yu niino , and takahiro ohno .
we thank satoshi miyazaki and suprime - cam team for commissioning suprime - cam fdccds , which has realized this work .
this work is based in part on data collected at subaru telescope , which is operated by the national astronomical observatory of japan .
ultravista is based on data products from observations made with eso telescopes at the la silla paranal observatory under eso programme i d 179.a-2005 and on data products produced by terapix and the cambridge astronomy survey unit on behalf of the ultravista consortium .
data analysis were in part carried out on common use data analysis computer system at the astronomy data center , adc , of the national astronomical observatory of japan .
ukidss uses the ukirt / wfcam ( casali et al .
2007 ) and a photometric system described in hewett et al .
( 2006 ) , and the calibration is described in hodgkin et al .
the pipeline processing and science archive are described in irwin et al .
( 2016 , in prep ) and hambly et al .
jsd acknowledges the support of the european research council via the award of an advanced grant .
this work is partly supported by jsps kakenhi grant number 23740159 ( hf ) and 15h03645 ( nk ) .
ando , m. , ohta , k. , iwata , i. et al .
, 2006 , , 654 , l9 becker , g. d. , bolton , j. s. , madau , p. et al .
2015 , , 447 , 3402 bongiorno , a. , schulze , a. , & merloni , a. , et al .
2016 , , 588 , a78 bouwens , r. j. , illingworth , g. d. , oesch , p. a. et al .
2014 , , 793 , 115 bouwens , r. j. , illingworth , g. d. , oesch , p. a. et al .
2015 , , 803 , 34 bowler , r. a. a. , dunlop , j. s. , mclure , r. j. et al .
2012 , , 426 , 2772 bowler , r. a. a. , dunlop , j. s. , mclure , r. j. et al .
2014 , , 440 , 2810 bowler , r. a. a. , dunlop , j. s. , mclure , r. j. et al .
2015 , , 452 , 1817 capak , p. , aussel , h. , ajiki , m. et al .
2007 , , 172 , 99 casali , m. , adamson , a. , & alves de oliveira , c. , et al .
2007 , , 467 , 777 cassata , p. , le fvre , o. , garilli , b. et al .
2011 , , 525 , a143 cassata , p. , tasca , l. a. m. , le fvre , o. et al .
2015 , , 573 , a24 caruana , j. , bunker , a. j. , wilkins , s. m. et al .
2014 , , 443 , 2831 curtis - lake , e. , mclure , r. j. , pearce , h. j. et al . 2012 , , 422 , 1425 dijkstra , m. , mesinger , a. , & wyithe , j. s. b. et al . 2011 , , 414 , 2139 dijkstra , m. , wyithe , s. , haiman , z. et al .
2014a , , 440 , 3309 dijkstra , m. 2014b , pasa , 31 , e040 dow - hygelund , c. c. , holden , b. p. , bouwens , r. j. et al .
2007 , , 660 , 47 dunlop , j. , rogers , a. b. , mclure , r. l. et al .
2013 , , 432 , 3520 fan , x. , strauss , m. a. , becker , r. h. et al .
2006 , , 132 , 117 furusawa , h. , kosugi , g. , akiyama , m. et al .
2008 , , 176 , 1 furusawa , j. , sekiguchi , k. , takata , t. et al .
2011 , , 727 , 111 fontana , a. , vanzella , e. , pentericci , l. et al .
2010 , , 725 , l205 gehrels , n. 1986 , , 303 , 336 hambly , n. c. , collins , r. s. , & cross , n. j. g. , et al .
2008 , , 384 , 637 hartoog , o. e. , malesani , d. , fynbo , j. p. u. et al .
2015 , , 580 , a139 hewett , p. c. , warren , s. j. , & leggett , s. k. , et al .
2006 , , 367 , 454 hodgkin , s. t. , irwin , m. j. , & hewett , p. c. , et al .
2009 , , 394 , 675 kashikawa , n. , aoki , k. , asai , r. et al .
2002 , , 54 , 819 kashikawa , n. , shimasaku , k. , matsuda , y. et al .
2011 , , 734 , 119 kobayashi , m. a. r. , totani , t. , & nagashima , m. 2010 , , 708 , 1119 konno , a. , ouchi , m. , ono , y. et al .
2014 , , 797 , 16 lawrence , a. , warren , s. j. , almaini , o. et al .
2007 , , 379 , 1599 le fvre , o. , cassata , p. , cucciati , o et al .
2013 , , 559 , 14 matthee , j. , sobral , d. , santos , s. et al .
2015 , , 451 , 400 mccracken . h. j. , milvang - jensen , b. , dunlop , j. s. et al .
2012 , , 544 , a156 mclure , r. j. , dunlop , j. s , bowler , r. a. a. et al .
2013 , , 432 , 2696 mcquinn , m. , hernquist , l. , zaldarriaga , m. et al .
2007 , , 381 , 75 miyazaki , s. , komiyama , y. , sekiguchi , m. et al .
2002 , , 54 , 833 monet , d. g. , levine , s. e. , canzian , b. et al .
2003 , , 125 , 984 oesch , p. a. , van dokkum , p. g. , illingworth , g. d. , et al .
2015 , , 804 , l30 oke , j. b. & gunn , j. e. 1983 , , 266 , 713 ono , y. , ouchi , m. , mobasher , b. et al .
2012 , , 744 , 83 ouchi , m. , shimasaku , k. , okamura , s. et al . , 2004 , apj , 611 , 660 ouchi , m. , shimasaku , k. , furusawa , h. et al .
2010 , , 723 , 869 peng , y. , lilly , s. j. , kova , k. et al . 2010 , , 721 , 193 pentericci , l. , fontana , a. , vanzella , e. et al .
2011 , , 743 , 132 planck collaboration 2015 , arxiv:1502.01589 roberts - borsani , g. w. , bouwens , r. j. , oesch , p. a. , et al .
2016 , , in press ( arxiv:1506.00854 ) schenker , m. a. , stark , d. p. , ellis , r. s. et al .
2012 , , 744 , 179 schmidt , k. b. , treu , t. , brada , m. , et al .
2016 , , 818 , 38 shapley , a. , e. , steidel , c. c. , pettini , m. et al .
2003 , , 588 , 65 shimasaku , k. , kashikawa , n. , doi , m. et al .
2006 , , 58 , 313 stanway , e. r. , bunker , a. j. , glazebrook , k. , et al .
2007 , , 376 , 727 stark , d. p. , ellis , r. , chiu , k. et al .
2010 , , 408 , 1628 stark , d. p. , ellis , r. , & ouchi , m. 2011 , , 728 , l2 steidel , c. c. , erb , d. k. , shapley , a. e. et al .
2010 , , 717 , 289 tilvi , v. , papovich , c. , finkelstein , s. l. et al . , 2014 , , 795 , 5 totani , t. , aoki , k. , hattori , t. et al . , 2014 , , 66 , 63 treu , t. , schmidt , k. b. , trenti , m. et al . , 2013 , , 775 , l29 van der burg , r. f. j. , hildebrandt , h. , erben , t. et al . 2010 , , 523 , a74 vanzella , e. , pentericci , l. , fontana , a. , et al . 2011 , , 730 , l35 yagi , m. , kashikawa , n. , sekiguchi , m. et al . 2002 , , 123 , 66 yagi , m. 2012 , pasp , 124 , 1347 zitrin , a. , labb , i. , belli , s. et al .
2015 , , 810 , l12 lrrr ultravista1 & 1173 & 26.63 & 0.76 + ultravista2 & 1096 & 26.46 & 0.84 + ultravista3 & 1123 & 26.47 & 0.70 + ultravista4 & 1250 & 26.75 & 0.82 + uds1 & 701 & 26.30 & 0.75 + uds2 & 970 & 26.57 & 0.81 + uds3 & 703 & 26.60 & 0.78 + uds4 & 928 & 26.58 & 0.81 + lccrrrrrrrr fh2 - 22303 & 02:16:25.092 & @xmath19004:57:38.50 & @xmath19128.7 & @xmath19128.2 & @xmath19128.0 & @xmath19128.0 & @xmath19127.6 & @xmath192 & @xmath193 & @xmath194 + fh2 - 48620 & 02:17:39.083 & @xmath19004:42:48.71 & @xmath19128.8 & @xmath19128.1 & @xmath19128.2 & @xmath19128.1 & @xmath19127.6 & @xmath195 & @xmath196 & @xmath197 + fh4 - 42903 & 02:17:26.306 & @xmath19005:10:16.27 & @xmath19128.5 & @xmath19128.2 & @xmath19128.0 & @xmath19128.1 & @xmath19127.6 & @xmath198 & @xmath195 & @xmath195 + lrrrrr b14 - 169850 & 4800 & @xmath199 & @xmath200 & @xmath201 & 4.7 + b14 - 065666 & 14400 & @xmath103 & @xmath202 & @xmath203 & @xmath204 + b14 - 304416 & 7200 & @xmath205 & @xmath206 & @xmath207 & 2.6 + b14 - 238225 & 6000 & @xmath208 & @xmath209 & @xmath210 & 6.5 + b14 - 035314 & 7200 & @xmath211 & @xmath212 & @xmath213 & 4.7 + b14 - 118717 & 6740 & @xmath205 & @xmath206 & @xmath214 & 7.1 + fh2 - 22303 & 19200 & @xmath215 & @xmath216 & @xmath217 & 1.8 + fh2 - 48620 & 4800 & @xmath218 & @xmath219 & @xmath213 & 7.1 + fh4 - 42903 & 7200 & @xmath220 & @xmath221 & @xmath222 & 10.7 + | we study the extent to which very bright ( @xmath0 ) lyman - break selected galaxies at redshifts @xmath1 display detectable ly@xmath2 emission . to explore this issue ,
we have obtained follow - up optical spectroscopy of 9 @xmath1 galaxies from a parent sample of 24 @xmath1 galaxy candidates selected from the 1.65 deg@xmath3 cosmos - ultravista and sxds - uds survey fields using the latest near - infrared public survey data , and new ultra - deep subaru @xmath4-band imaging ( which we also present and describe in this paper ) .
our spectroscopy has yielded only one possible detection of ly@xmath2 at @xmath5 with a rest - frame equivalent width @xmath6 .
the relative weakness of this line , combined with our failure to detect ly@xmath2 emission from the other spectroscopic targets allows us to place a new upper limit on the prevalence of strong ly@xmath2 emission at these redshifts . for conservative calculation and to facilitate comparison with previous studies at lower redshifts , we derive a 1-@xmath7 upper limit on the fraction of uv - bright galaxies at @xmath1 that display @xmath8 , which we estimate to be @xmath9 .
this result may indicate a weak trend where the fraction of strong ly@xmath2 emitters ceases to rise , and possibly falls between @xmath10 and @xmath1 .
our results also leave open the possibility that strong ly@xmath2 may still be more prevalent in the brightest galaxies in the reionization era than their fainter counterparts
. a larger spectroscopic sample of galaxies is required to derive a more reliable constraint on the neutral hydrogen fraction at @xmath11 based on the ly@xmath2 fraction in the bright galaxies . |
as has been discussed in the course of this workshop , the classification of stars is not as simple and straightforward as for other groups of chemically peculiar stars .
when only low resolution classification spectra are used , it is likely that stars with different astrophysical properties but similar characteristics in the narrow spectral range are included .
therefore a detailed spectroscopic investigation for a large number of confirmed stars ( e.g. paunzen et al .
1997 ) covering the whole range of atmospheric parameters is required to establish a common abundance pattern for these stars .
up to now we have analyzed the chemical composition of six stars where two of them turned out to be spectroscopic binaries ( paunzen et al .
1998 ) . in order to put our results on more solid
ground we investigated the impact of using different programmes for the calculation of the model atmospheres on the abundances of three stars with different temperatures , gravities and metallicities ( heiter et al .
the two issues we examined concerned the inclusion of individual line opacities and the calculation of the convective flux in the model atmospheres .
high resolution spectra with high signal - to - noise ratios have been obtained at various sites as listed in table [ obs_tab ] .
they have been reduced with noao iraf ( willmarth & barnes 1994 ) , and atmospheric parameters and abundances have been derived as described by gelbmann et al .
( 1997 , 1998 ) .
the resulting atmospheric parameters are presented in table [ par_tab ] .
the abundances of each element for which at least one `` unblended '' line could be detected in the observed spectra of the programme stars are plotted in figs .
[ abu_fig1 ] and [ abu_fig2 ] .
`` unblended '' means that no other line with a depth greater than 30 % of that of the examined line is found within a wavelength window whose width depends mainly on the @xmath0 of the respective star .
.observations used for this analysis . [ cols="<,<,<,<,<,<",options="header " , ]
the abundances of the three stars from fig .
[ abu_fig1 ] were used to calculate individual opacity distribution functions ( odfs ) for these stars with a programme written by n. piskunov ( piskunov & kupka 1998 ) .
the subsequent inclusion of these odfs in the model atmosphere calculations instead of the tabulated opacities from kurucz ( 1993 ) showed that the @xmath1-relations and therefore the synthetic spectra of all three stars do not change significantly . in the next step we compared the results of the original atlas9 programme ( kurucz 1993 , mlt ) with three versions where the treatment of convection ( or overshooting ) had been changed : 1 ) deletion of the part of the programme concerning the overshooting ( no ) , 2 ) implementation of the convection model by canuto & mazzitelli ( 1991 , cm ) , 3 ) variation of the treatment of overshooting ( castelli 1996 , ovok ) .
the differences can be summarized as follows : using the no version results in similar changes compared to mlt as for the cm model , whereas ovok yields practically the same @xmath1-relation as mlt . for hd84123
, the largest differences occur in the visible spectral range ( see fig .
[ con_fig1 ] ) , where the abundances of all elements would have to be decreased by 0.1 dex to match the mlt calculations . for hd192640 ,
significant differences are only seen in the uv spectral range ( fig .
[ con_fig2 ] ) .
the cm flux seems to represent the observations better than that of mlt .
note the absorption feature at 1600 , which can not be reproduced by either of the models .
for the hottest of the stars , hd183324 , the differences are negligible . this investigation was carried out within the working group _ asteroseismology ams _ with funding from the fonds zur frderung der wissenschaft - lichen forschung ( project _ s7303-ast _ ) .
we thank dr .
e. solano for providing the iue spectra .
use was made of the simbad database , operated at cds , strasbourg , france .
castelli , f. : 1996 , in _ model atmospheres and spectrum synthesis _ , eds . : s.j
. adelman , f. kupka and w.w .
weiss , asp conf .
108 , p. 85
gelbmann , m. : 1998 , _ this workshop _
kurucz , r.l . : 1993 , _ cd - rom 1 - 23 _ , smithsonian astrophysical observatory willmarth , d. , barnes , j. : 1994 , _ a user s guide to reducing echelle spectra with iraf _ | the results of the abundance analyses of six stars are presented . for three stars ,
the impact of individual odfs and various treatments of convection in the calculactions of the model atmospheres are investigated . |
quasars and other types of agn activity are believed to be powered by supermassive black holes .
the agn activity may produce highly collimated outflows from the immediate vicinity of the black hole ( e.g. rees 1984 ; blandford 1990 ) , which can power a large - scale radio source . two defining properties of astrophysical black holes that can be measured in principle are the black hole mass and spin . a significant amount of progress has been made toward measuring the masses of supermassive black holes ( e.g. kormendy & richstone 1995 ; magorrian et al .
1998 ; ferrarese & merritt 2000 ; gebhardt et al .
2000 ; ferrarese & ford 2005 ) , though measuring the spin has been more challenging .
general theoretical studies suggest that the merger and accretion history of a supermassive black hole is encoded in the spin of the hole .
for example , successive mergers are likely to produce black holes that are spinning at a moderate rate ( @xmath1 ) , while powerful accretion events are likely to produce rapidly rotating black holes ( berti & volonteri 2008 ) .
alternatively , most black holes may grow due to many short - lived , uncorrelated accretion episodes , which would lead to lower spin values ( king & pringle 2006 , 2007 ) .
numerous uncorrelated accretion episodes tend to cause the spin of the hole to decrease over time with substantial fluctuations in spin caused by each episode ( king , pringle , & hofmann 2008 ) . at present , only a few observations allow black hole spins to be studied directly
. observations of seyfert galaxies suggest rapidly rotating black holes in these systems ( wilms et al .
2001 ; fabian et al .
a large spin is indicated by observations of the galactic center black hole ( genzel et al .
2003 ; aschenbach et al .
2004 ) . and
, x - ray observations of active galaxies suggest rapidly rotating black holes in these systems ( crummy et al .
outflows from agn that produce extended radio sources allow a lower bound to be placed on the spin of the black hole that powers the outflow . assuming only that the outflow is powered by the spin energy of the black hole
, daly ( 2009 ) showed that the outflow energy and the black hole mass may be combined to obtain a lower limit on the spin of the black hole . for a sample of 19 very powerful classical double sources , the lower bound was about the same for each source in the sample , and indicated @xmath2 . here
, beam powers and black hole masses of radio sources are combined to study black hole spins in the contexts of the models of blandford & znajek ( 1977 ) , meier ( 1999 ) , and models with similar functional forms .
it is shown that the spin and magnetic field strength of a supermassive black hole can be rather tightly constrained in the contexts of these models .
the method is described in section 2 .
it is applied to two samples of sources , and the results are presented in section 3 , discussed in section 4 , and summarized in section 5 . a standard cosmological model with @xmath3 km / s / mpc , @xmath4 , @xmath5 , and zero space curvature is assumed throughout .
the well - known model of blandford & znajek ( 1977 ) , referred to as the bz model , and other models to power highly collimated outflows from agn are considered here . in the bz model , the beam power @xmath6 that will be produced is related to the dimensionless black hole spin @xmath7 , the black hole horizon @xmath8 , and the poloidal magnetic field strength @xmath9 at the horizon , where @xmath10 is the black hole spin angular momentum , m is the black hole mass , g is newton s constant , and c is the speed of light ( e.g. macdonald & thorne 1982 ; thorne , price , & macdonald 1986 ) @xmath11 here @xmath12 is the value of @xmath9 in units of @xmath13 g , @xmath14 is the black hole mass in units of @xmath15 , and the factor @xmath16 depends on the angular velocity of the field lines @xmath17 relative to that of the hole @xmath18 and is taken to have its maximum value of @xmath19 ( e.g. blandford 1990 ) .
the magnetic field strength is not expected to exceed the eddington magnetic field strength , given by @xmath20 g ( e.g. dermer , finke , & menon 2008 ) . rewriting equation ( 1 ) in terms of the black hole spin @xmath21 and normalized magnetic field strength @xmath22
, we have @xmath23 where @xmath24 is the eddington magnetic field strength in units of @xmath13 g , @xmath25 , @xmath26 is the beam power in units of @xmath27 erg / s , and @xmath28 is the eddington luminosity , @xmath29 erg / s . equation ( 2 ) allows a determination of the quantity @xmath30 when observations allow determinations of the beam power of the source , @xmath6 , and the black hole mass , @xmath31 .
the parameter @xmath30 provides a lower bound on @xmath21 and a lower bound on @xmath32 in the context of the bz model since the maximum value of @xmath21 is unity and the maximum value of @xmath32 is unity .
when this parameter , referred to as the `` j - b '' parameter , is equal to one , it suggests that both @xmath21 and @xmath32 are close to one .
if the `` j - b '' parameter exceeds one , it suggests a problem with the underlying model ( in this case , the bz model ) .
thus , the `` j - b '' parameter indicated by equation ( 2 ) provides a lower bound on @xmath21 and a lower bound on @xmath32 in the context of the bz model , and provides a test of the bz model .
models may be characterized by comparison with bz model predictions by writing @xmath33 .
then @xmath34 of that model is @xmath35 . if @xmath30 exceeds unity , then models with values of @xmath36 with @xmath37 ^ 2 $ ] would be indicated .
consider , for example , the hybrid model proposed by meier ( 1999 ) .
this model includes characteristics of the blandford - payne ( blandford & payne 1982 ) and blandford - znajek ( blandford & znajek 1977 ) models .
the beam power produced in this model is @xmath38 erg / s ( meier 1999 ) , or @xmath39 , and @xmath40 . requiring that @xmath34 satisfy @xmath41 allows constraints to be placed on @xmath36 , and this indicates which models can account for the characteristics of these systems .
the method is applied to the samples of 19 very powerful frii radio galaxies ( referred to as friib sources ) and 29 central dominant galaxies ( cdgs ) studied by daly ( 2009 ) .
the friib sources have radio powers at least a factor of ten above the classical fri / frii transition ( fanaroff & riley 1974 ) .
black hole masses and beam powers are available for all of these sources ; the masses are listed in tables 1 and 2 of daly ( 2009 ) . beam powers for the powerful classical double radio galaxies are obtained from guerra et al .
( 2000 ) , wan , daly , & guerra ( 2000 ) , and odea et al .
( 2009 ) ; note that these beam powers are independent of offsets of the extended radio source from minimum energy conditions ( odea et al .
beam powers for the cdgs are obtained from rafferty et al .
( 2006 ) , and also are independent of offsets from minimum energy conditions .
almost all of the radio sources associated with cdgs have fri radio source structure or amorphous radio structure , with a few exceptions such as cygnus a ( birzan et al . 2008 ) .
the name , redshift , total beam power , black hole mass , and eddington magnetic field strength are listed for each source in tables 1 and 2 for the friib and cdg sources , respectively .
the beam powers and black hole masses were combined to solve for @xmath30 using equation ( 2 ) and are listed in tables 1 and 2 .
values of @xmath42 were obtained explicitly for the friib sources using @xmath43 so @xmath44 , and are included in table 1 .
the value of @xmath45 is listed in the final column of tables 1 and 2 , and is discussed below . for simplicity , the average values of asymmetric error bars was used .
the parameter @xmath30 is shown as a function of black hole mass in figure 1 . analyzing the two samples separately ,
there is no indication of a dependence of @xmath30 on black hole mass ( see figure 1 ) .
since @xmath30 , and @xmath42 , are proportional to @xmath46 ( see eq .
2 ) , this is equivalent to finding no dependence of @xmath47 on black hole mass .
the parameter @xmath30 is shown as a function of redshift in figure 2 .
each sample clearly exhibits a dependence of @xmath34 on redshift ( see figure 2 ) .
the sample of 19 powerful friib sources has the dependence @xmath48 .
given that these are the most powerful radio sources at each redshift and that they are drawn from a complete sample of sources , this represents the envelope of the distribution of @xmath34 values as a function of redshift .
thus , this redshift dependence can be interpreted as the evolution of the maximum value of @xmath34 as a function of redshift , and is obtained in the context of the bz model .
the normalization of @xmath34 decreases by a factor of about @xmath49 if the meier model is considered , but the redshift behavior is unaffected .
the beam power is shown as a function of black hole mass in figure 3 .
the cdgs have much lower beam powers than the friib sources , though the sources are powered by black holes with similar mass .
the value of @xmath21 and @xmath32 may be independent , or the value of @xmath32 could depend upon @xmath21 .
the results of daly et al .
( 2009 ) and daly & guerra ( 2002 ) suggest that @xmath50 , that is , @xmath51 . setting @xmath52 , and noting that @xmath53 and @xmath54 , the results for @xmath34 listed in table 1 indicate that @xmath55 for the bz model .
thus , @xmath56 is of order unity , or @xmath57 . in this case ,
@xmath58 , and the redshift evolution of the maximum value of @xmath21 varies as @xmath59 .
the values of @xmath21 obtained using @xmath60 are listed in the final column of tables 1 and 2 and are obtained in the context of the bz model .
all of the empirically determined @xmath30 parameters are consistent with @xmath61 at about the 1 sigma level , thus the current data do not require modifications to the bz model ( see tables 1 and 2 and figure 1 ) .
it is interesting that many sources have @xmath62 , indicating that both @xmath21 and @xmath32 are close to one if the bz model provides an accurate description of the physics of these systems .
almost all of the empirically determined @xmath63 parameters are less than one , indicating that either @xmath21 or @xmath32 is less than one if this model provides an accurate description of these systems .
as noted in section [ method ] , the parameter @xmath34 provides a lower bound on @xmath21 and a lower bound on @xmath32 since neither is expected to exceed unity . the lower bounds on @xmath21 obtained for friib sources in the bz and the meier models are about 0.4 and 0.2 , respectively ( see table 1 ) .
these are larger than the lower bound of @xmath64 obtained for these same sources by daly ( 2009 ) using a model - independent approach , assuming only that the large - scale outflow from the agn is powered by the spin energy of the black hole . combining the results obtained here with those obtained by daly ( 2009 )
indicates that only a small fraction of the spin energy per unit black hole mass , @xmath65 , of a system is extracted during one particular outflow event for the friib sources studied and most of the cdgs .
thus , a single outflow event does not significantly reduce the spin of the black hole for friib sources and most radio sources associated with cdgs .
the parameter @xmath30 is clearly increasing with redshift ( see figure 2 ) , and is roughly @xmath66 for the friib sources .
the fact that lower @xmath30 sources are missing at high redshift results from the radio power selection effect , since all of the higher redshift sources are from the 3crr catalog ( laing , riley , & longair 1983 ) ; obviously , this does not explain the lack of high @xmath34 sources at low redshift .
given that the sources with high radio power have high beam power , and , for a fixed black hole mass , sources with high beam power have high spin ( see eq .
1 ) , the friib sources studied here are likely to have spin values that are close to the maximal values for sources at that redshift .
thus , the evolution of spin of these systems should be considered to be the envelope of the distribution , presenting the maximal spin as a function of redshift .
that is , the black holes at the centers of massive elliptical galaxies that produce friib radio sources have maximal spin values that are slowly decreasing as the universe evolves .
the results of daly ( 2009 ) suggest that the energy extracted per unit black hole mass , @xmath65 , during each outflow event is independent of source redshift and black hole mass , and is roughly constant for all of the friib sources studied .
given the selection effects affecting the cdgs , the data were consistent with those sources having a value of @xmath67 similar to that of the friib sources .
the lower bounds on @xmath21 obtained here in the context of the bz model are consistent with @xmath67 of @xmath68 obtained for the friib sources by daly ( 2009 ) and with most of the cdgs at about 1 to 2 @xmath69 . in the case that @xmath70 , as discussed in section [ results ] and as indicated by detailed studies of friib sources ( e.g. daly et al .
2009 ) , the value of @xmath21 is obtained by taking the square root of the values of @xmath34 listed in tables 1 and 2 , and is listed in the final columns of each table . in this case , all of the friib and cdg sources have values of @xmath21 that are consistent with the `` model - independent '' minimum value @xmath64 obtained by daly ( 2009 ) . in addition , this implies that the maximum value of @xmath21 evolves as @xmath71 .
the evolution of the spins of supermassive black holes may be used to study the accretion and merger history of these sources , as discussed , for example , by king & pringle ( 2006 , 2007 ) , berti & volonteri ( 2008 ) , and king , pringle , and hofmann ( 2008 ) .
it is particularly intriguing that the redshift evolution obtained here for friib sources is very similar to that predicted by the detailed model of king , pringle , & hofmann ( 2008 ) for spin down by gas accretion through a series of repeated accretion episodes .
in comparing the results obtained here with theoretical predictions , it should be kept in mind that all of the cdgs are by definition associated with giant elliptical galaxies , and the friib sources are likely to evolve into cdgs ( lilly & longair 1984 ; best et al .
1998 ; mclure et al .
2004 ) . due to the way each sample
is selected , the cdgs allow a glimpse into the spin values of sources with low radio power and low beam power , while the friib sources are the most powerful extended radio sources at their respective redshifts .
the combination of the results obtained here with those obtained earlier suggest that each source may undergo multiple outflow events .
the energy extracted per outflow event is roughly constant , suggesting very similar physical conditions in the source when the outflow event is triggered , and the energy extracted per outflow event is a small fraction of the total available energy for most of the sources ( daly 2009 ) .
thus , each source may undergo numerous outflow events .
these outflow events could significantly affect the gaseous medium in the vicinity of the radio source .
this heating of the gaseous environment of the source could cause the radio source structure to shift from frii to fri structure .
this is consistent with the ideas that very powerful 3crr radio galaxies evolve into cdgs ( lilly & longair 1984 ; best et al .
1998 ; mclure et al .
2004 ) , that different radio structures are determined by the source environment ( e.g. hardee et al . 1992
; odonoghue , eilek , & owen 1993 ; burns et al .
1994 ; bicknell 1995 ; barai & wiita 2007 ) , and that the evolution of the gaseous environments could be due in part to the large - scale outflows ( e.g. silk & rees 1998 ; eilek & owen 2002 ) .
the cdgs do have much lower beam powers than the friib sources , which may also explain the differences in the radio source structures .
both the cdg and friib samples are illustrative of outflows from massive black holes associated with giant elliptical galaxies since friib sources are expected to evolve into cdgs ( lilly & longair 1984 ; best et al .
1998 ; mclure et al .
the cdgs allow a study of black hole spins and magnetic field strengths of lower power radio sources that have a range of beam power , black hole spin , and magnetic field strengths .
the friib sources allow a study of black hole spins and magnetic field strengths of the most powerful sources with large - scale outflows at their respective redshifts ; these are the sources that have the largest radio power , beam power , and black hole spin .
these sources were studied in the context of the bz and meier models , which allowed the beam power and black hole mass of each source to be combined to obtain lower bounds on the black hole spin and strength of the magnetic field close to the black hole .
the lower bounds on the black hole spin obtained for friib sources were significantly larger than those obtained by daly ( 2009 ) by studying the outflow energy and black hole mass of each system .
this implies that , if the black hole spin powers the large - scale outflows in these systems , then the energy extracted in each outflow event is a small fraction of the total spin energy available .
thus , the evolution of the spin with redshift is a reflection of the merger and accretion history of the source , and is not significantly affected by energy losses associated with the outflow . and , each source may undergo numerous outflow events .
independent empirical work has shown that the magnetic field strength is proportional to the black hole spin ( e.g. daly et al .
this means that the values of @xmath34 obtained here can be used to obtain estimates of the black hole spin @xmath21 ; the normalization between @xmath32 and @xmath21 is of order unity , but is not precisely known .
values of @xmath21 were obtained in the context of the bz model for the friib and cdgs .
the values obtained for friib sources range from about 0.7 to 1 , while those for cdgs range from about 0.14 to 0.8 .
if considered in the context of the meier ( 1999 ) model , these values decrease by a factor of about 1.5 .
all of the values are greater than or similar to the `` model - independent '' minimum value of @xmath72 obtained by daly ( 2009 ) .
losses due to outflows will affect the spin of the black hole when that spin is close to @xmath67 .
thus , this could affect the spins of black holes of cdgs with low values of @xmath21 , but is unlikely to affect the those associated with friib sources .
gebhardt , k . , kormendy , j. , ho , l. c. , bender , r. , bower , g. , dressler , a. , faber , s. m. , filippenko , a. v. , green , r. , grillmair , c. , lauer , t. r. , magorrian , j. , pinkney , j. , richstone , d. , & tremaine , s. 2000 , apj , 543 , l5 lccclllc 3c 405 & 0.056 & @xmath73&@xmath74&@xmath75&@xmath76&@xmath77&@xmath78 + 3c 244.1 & 0.43 & @xmath79&@xmath80&@xmath81&@xmath82&@xmath83&@xmath84 + 3c
172 & 0.519 & @xmath85&@xmath86&@xmath87&@xmath88&@xmath89&@xmath90 + 3c 330 & 0.549 & @xmath91&@xmath92&@xmath93&@xmath94&@xmath95&@xmath96 + 3c 427.1 & 0.572 & @xmath85&@xmath97&@xmath98&@xmath99&@xmath100&@xmath101 + 3c 337 & 0.63 & @xmath102&@xmath103&@xmath104&@xmath105&@xmath100&@xmath106 + 3c34 & 0.69 & @xmath107&@xmath108&@xmath109&@xmath88&@xmath89&@xmath110 + 3c441 & 0.707 & @xmath111&@xmath112&@xmath113&@xmath114&@xmath115&@xmath116 + 3c 55 & 0.72 & @xmath117&@xmath97&@xmath118&@xmath119&@xmath120&@xmath121 + 3c 247 & 0.749 & @xmath122&@xmath123&@xmath124&@xmath125&@xmath126&@xmath127 + 3c 289 & 0.967 & @xmath128&@xmath129&@xmath124&@xmath130&@xmath131&@xmath132 + 3c 280 & 0.996 & @xmath133&@xmath129&@xmath124&@xmath134&@xmath135&@xmath136 + 3c 356 & 1.079 & @xmath137&@xmath138&@xmath139&@xmath140&@xmath141&@xmath142 + 3c 267 & 1.144 & @xmath143&@xmath144&@xmath145&@xmath146&@xmath147&@xmath148 + 3c 324 & 1.21 & @xmath149&@xmath150&@xmath151&@xmath152&@xmath153&@xmath154 + 3c 437 & 1.48 & @xmath155&@xmath156&@xmath145&@xmath157&@xmath158&@xmath159 + 3c 68.2 & 1.575 & @xmath160&@xmath161&@xmath162&@xmath163&@xmath164&@xmath165 + 3c 322 & 1.681 & @xmath166&@xmath167&@xmath139&@xmath168&@xmath169&@xmath170 + 3c 239 & 1.79 & @xmath171&@xmath172&@xmath162&@xmath173&@xmath174&@xmath175
+ [ table1 ] m84 & 0.0035 & @xmath176&@xmath177&@xmath178&@xmath179&@xmath180 + m87 & 0.0042 & @xmath181&@xmath182&@xmath183&@xmath184&@xmath185 + centaurus & 0.011 & @xmath186&@xmath182&@xmath183&@xmath187&@xmath188 + hcg 62 & 0.014 & @xmath189&@xmath190&@xmath191&@xmath192&@xmath193 + a262 & 0.016 & @xmath194&@xmath182&@xmath183&@xmath195&@xmath196 + perseus & 0.018 & @xmath197&@xmath198&@xmath199&@xmath200&@xmath201 + pks 1404 - 267 & 0.022 & @xmath202&@xmath190&@xmath191&@xmath203&@xmath204 + a2199 & 0.03 & @xmath205&@xmath206&@xmath207&@xmath208&@xmath209 + a2052 & 0.035 & @xmath210&@xmath198&@xmath199&@xmath211&@xmath212 + 2a 0335 + 096 & 0.035 & @xmath213&@xmath214&@xmath215&@xmath216&@xmath217 + mkw 3s & 0.045 & @xmath218&@xmath182&@xmath183&@xmath219&@xmath220 + a4059 & 0.048 & @xmath221&@xmath222&@xmath223&@xmath224&@xmath225 + hydra a & 0.055 & @xmath226&@xmath227&@xmath228&@xmath229&@xmath230 + a85 & 0.055 & @xmath231&@xmath222&@xmath223&@xmath232&@xmath233 + cygnus a & 0.056 & @xmath234&@xmath222&@xmath223&@xmath235&@xmath236 + sersic 159/03 & 0.058 & @xmath237&@xmath238&@xmath239&@xmath240&@xmath241 + a133 & 0.06 & @xmath242&@xmath243&@xmath207&@xmath244&@xmath245 + a1795 & 0.063 & @xmath246&@xmath247&@xmath207&@xmath248&@xmath249 + a2029 & 0.077 & @xmath250&@xmath251&@xmath252&@xmath253&@xmath254 + a478 & 0.081 & @xmath255&@xmath256&@xmath257&@xmath258&@xmath259 + a2597 & 0.085 & @xmath260&@xmath182&@xmath183&@xmath261&@xmath262 + 3c 388 & 0.092 & @xmath263&@xmath198&@xmath199&@xmath264&@xmath265 + pks 0745 - 191 & 0.103 & @xmath266&@xmath267&@xmath223&@xmath268&@xmath269 + hercules a & 0.154 & @xmath270&@xmath271&@xmath272&@xmath273&@xmath274 + zw 2701 & 0.214 & @xmath275&@xmath238&@xmath239&@xmath276&@xmath277 + ms 0735.6 + 7421 & 0.216 & @xmath278&@xmath271&@xmath272&@xmath279&@xmath280 + 4c 55.16 & 0.242 & @xmath281&@xmath214&@xmath215&@xmath282&@xmath283 + a1835 & 0.253 & @xmath284&@xmath285&@xmath286&@xmath287&@xmath288 + zw 3146 & 0.291 & @xmath289&@xmath290&@xmath291&@xmath292&@xmath293 + [ table ] | beam powers and black hole masses of 48 extended radio sources are combined to obtain lower bounds on the spins and magnetic field strengths of supermassive black holes .
this is done in the context of the models of blandford & znajek ( 1977 ) ( the bz model ) and meier ( 1999 ) ; a parameterization for bounds in the context of other models is suggested .
the bounds obtained for very powerful classical double radio sources in the bz model are consistent with black hole spins of order unity for sources at high redshift .
the black hole spins are largest for the highest redshift sources and decrease for sources at lower redshift ; the sources studied have redshifts between zero and two .
lower power radio sources associated with central dominant galaxies may have black hole spins that are significantly less than one . combining
this analysis with other results suggests that the maximum values of black hole spin associated with powerful radio galaxies decline from values of order unity at a redshift of 2 to values of order 0.7 at a redshift of zero , falling roughly as @xmath0 , while lower power radio sources have spin values that range from about 0.1 to 0.8 .
these black hole spin values decrease if the data are considered in the context of the meier model rather than the bz model . |
harmonic generation of optical fields in nonlinear optical crystals can be modeled with classical coupled wave equations @xcite .
define the fundamental and second harmonic fields as @xmath9 and @xmath10 , respectively . in ideal second harmonic generation ,
@xmath11 . while this frequency ratio has been measured very precisely ( @xmath12 @xcite )
, we present a lower bound to the frequency noise , or the spectrum of the time dependent quantity @xmath13 . several experiments at the forefront of precision metrology and frequency standards use harmonic generation in their experiments .
these include iodine stabilized nd : yag lasers @xcite , optical frequency combs @xcite , measurement of optical frequency ratios @xcite , and precision atomic spectroscopy @xcite .
many of these experiments provide bounds to any excess frequency noise which might be found in the second harmonic generation ( shg ) process ( e.g. from thermodynamic fluctuations in the crystal temperature ) . as experiments like these push towards lower noise levels , the fundamental noise sources in second harmonic generation may become a relevant noise source .
another field which utilizes precision shg is interferometric detection of gravitational waves .
these interferometers require measurements of mirror displacements at the level of 10@xmath14 @xcite . to achieve such high precision , these interferometers use seismically isolated suspended mirrors which make up multiple coupled optical cavities with narrow linewidths .
it has proven challenging to control these cavities and bring them simultaneously into resonance . in order to simplify this problem ,
auxiliary fields will be generated by shg and used to form independent signals for two of the cavities .
this technique is being developed using some small scale prototypes @xcite .
to understand the limits of such a scheme , we need to bound the frequency noise introduced in the second harmonic generation process . in our case , we use a 1064 nm carrier ( nd : yag ) , and generate the 532 nm doubled light with ppktp crystals . in order to meet the stability requirements for this technique
, the excess frequency noise must be less than 70 mhz rms from 10 mhz to 30 hz @xcite . in the future , optical configurations using more than one wavelength inside the interferometer to beat standard quantum noise limits of the detection @xcite may be used .
fundamental noise limits in harmonic generation may be relevant there .
figure [ fig : diagram ] shows the experimental setup used to measure the uncorrelated frequency noise between the fundamental and the second harmonic .
the beam from a 2w non - planar ring laser at 1064 nm is passed through a spatial filter cavity @xcite .
the spatially filtered beam then enters the mach - zehnder through a hole in an acoustic enclosure .
ppktp crystals ( raicol crystals ) inside temperature stabilized ovens are placed in each arm , with lenses added to mode match to the each crystal .
two additional dichroic mirrors ( m3 , m4 ) are placed in one arm for alignment .
the beams recombine on a dichroic 50:50 beamsplitter and low loss hr mirrors ( t @xmath15 10 ppm at 1064 nm ) are used to separate the fundamental from the second harmonic .
commercial dichroic mirrors ( hr532 , ar1064 ) were used to further separate the 532 nm from the 1064 nm light . for the 532 nm and 1064 nm detection , si ( hamamatsu 1223 ) and ingaas ( gpd 2000 ) photodiodes were used , respectively .
the mach - zehnder arm lengths were adjusted to be mid fringe for both 532 and 1064 nm simultaneously .
the interferometer is locked to this point by applying feedback from the 1064 nm pd difference signal to the pzt on m1 ( see fig . [
fig : diagram ] ) .
we drove the pzt with a small signal at 100 hz to monitor the calibration ( rad / v ) at both wavelengths , as any differential frequency noise between the two wavelengths would cause the 532 nm calibration to drift .
the calibration of the mach zehnder was confirmed by sweeping through multiple fringes .
the 1064 nm and 532 nm mach - zehnder length signals were read out with two pairs of balanced homodyne detectors .
the sum and difference signals were digitized at 256 hz and processed further offline . in the non - depleted pump approximation , with imperfect phase matching ,
the phase relation of the two fields in each arm before the recombining beamsplitter is : @xmath16 where @xmath17 and @xmath18 describe the relative phases of the fundamental and the second harmonic , respectively , @xmath19 is the length of the doubling crystal , and @xmath20 is the phase mismatch , the parameter normally used to describe the efficiency of shg . in theory ,
@xmath20 can be arbitrarily small ( limited in practice only by the ability to stabilize the temperature of the nonlinear crystal ) .
using superscripts to differentiate between the two arms , the difference in the second harmonic phase is thus @xmath21
the phase difference ( @xmath22 ) was suppressed by the servo , which had a unity gain frequency of @xmath23 10 hz .
typical values of @xmath24 and @xmath25 are shown in fig .
[ fig : phase ] .
( in red ) ( which is the in - loop signal of the mach - zehnder ) , @xmath26 ( in green ) ( which contains excess phase noise ) , and the subtraction residual ( in black ) .
the total rms frequency fluctuation in the measurement band is 3 mhz . on the right
we have a comparison of this work with previous bounds .
these bounds show up as frequency noise @xcite , timing error @xcite , and phase noise @xcite . ]
the frequency noise amplitude spectral density of the mach zehnder is shown in fig .
[ fig : nb ] with known noise sources . @xmath27 and @xmath28 are shown in red and green , respectively , where @xmath29 . above 10 hz
the phase noise is dominated by a forest of mechanical resonances on the optical table which show up strongly in both @xmath27 and @xmath28 .
we used the average transfer function over the measurement time , @xmath30 , to subtract the noise which is coherent between the two @xmath31s .
this subtracted level is shown as the black trace in fig .
[ fig : nb ] .
the black trace bounds any noise source which causes frequency noise between the fundamental and the harmonic and is uncorrelated between the two ovens , such as thermodynamic fluctuations in the crystals , or any temperature fluctuation not common to both ovens .
some common mode effects such as intensity dependent phase shifts and temperature fluctuations ( and thus phase matching fluctuations ) are suppressed by the experimental setup , so these technical noise sources will not be visible in these measurements .
however , a pessimistic estimate of the temperature noise coupling assuming no common mode rejection shown in fig .
[ fig : nb ] is below the measured excess frequency noise .
in addition , the intensity of the fundamental was only stabilized to a level of @xmath32 at 1 hz , where it can be lowered to the @xmath33 at 1 hz with current techniques , so unless the rejection of this effect was more than 46 db , we can safely ignore it .
it is highly improbable that the intensity to frequency noise coupling would be above the level shown in fig . [
fig : nb ] .
the excess noise below 10 hz was found to be correlated with air currents on the table , and would be reduced by moving the setup into a vacuum chamber .
the total rms excess frequency noise of the black trace is 3 mhz rms in the 10 mhz to 128 hz band .
in addition to the usual technical noise sources , it is worth considering whether there is a more fundamental limit to the relative phase between the fundamental and the harmonic .
a rough estimate of thermal noise from thermoelastic ( zener ) damping was obtained by directly applying @xcite the fluctuation - dissipation theorem .
we treat the crystal as an 0.8 mm radius cylinder , and follow the calculation done by heinert et al .
this yields a spectral density of @xmath20 taking into account both thermorefractive and thermoelastic fluctuations .
when expressed as frequency fluctuations , it is well approximated by @xmath34 above 10 hz .
below that , it must be flat or continue to decrease , or else the rms temperature integral would diverge . practically speaking , in our band of interest , the temperature fluctuations of the ovens
is at least 2 orders of magnitude greater than these fundamental thermodynamic temperature fluctuation . in the future ,
when researchers seek to make frequency comparisons at better than the @xmath35 level , these thermal noises will have to be calculated with more accuracy .
we have examined the results from a number of precision experiments which involve shg .
although these experiments are likely limited by other noise sources , these data can be used to place upper bounds on the possible noise generated by shg . in fig .
[ fig : nb ] we compare frequency noise @xcite , timing errors @xcite , and phase noise @xcite ; in fig .
[ fig : avar ] we compare allan deviations @xcite . while the comparison in fig .
[ fig : nb ] is straightforward , some caution should be taken interpreting fig .
[ fig : avar ] .
our allan deviation at the 0.1 s time scale is heavily influenced by the high frequency noise in the measurement ( 10 - 128 hz ) .
since we low pass the signal to acquire data at 256 hz , we reject noise which would make the allan deviation increase at all time scales .
it should also be emphasized that we only measure relative frequency fluctuations , and that there are some common mode noise sources which the experiment is insensitive to .
see @xcite for more information on allan deviations and phase noise .
in conclusion , we have demonstrated that the rms frequency fluctuations added from uncorrelated mechanisms between two shg crystals is less than 3@xmath0mhz at time scales over 10 ms .
the obvious correlated mechanisms ( temperature and intensity noise coupling ) are likely insignificant compared to this as discussed in section [ sec : res ] .
this is low enough to not limit the lock acquisition @xcite scheme for advanced ligo and other gravitational - wave detectors .
additionally , we have shown that there is no excess noise process at a level which is of interest to those doing precision atomic spectroscopy @xcite , using frequency combs to transfer optical harmonics @xcite and other tests of fundamental physics .
the authors would like to thank p. fritschel for suggesting this work as well as k. arai , j. harms , and k. mckenzie , for discussions leading to the understanding presented in this letter , as well as h. grote and m. evans for useful input on the interferometer stabilization .
we also gratefully acknowledge the national science foundation for support under grant phy-0757058 . | we report a bound on the relative frequency fluctuations in nonlinear second harmonic generation .
a 1064@xmath0 nm nd : yag laser is used to read out the phase of a mach - zehnder interferometer while ppktp , a nonlinear crystal , is placed in each arm to generate second harmonic light . by comparing the arm length difference of the mach zehnder as read out by the fundamental 1064 nm light , and its second harmonic at 532 nm
, we can bound the excess frequency noise introduced in the harmonic generation process .
we report an amplitude spectral density of frequency noise with total rms frequency deviation of @xmath1mhz and a minimum value of 20 @xmath2 over 250 seconds with a measurement bandwidth of 128 hz , corresponding to an allan deviation of @xmath3 at 20 seconds .
10 j. a. armstrong , n. bloembergen , j. ducuing , and p. s. pershan , `` interactions between light waves in a nonlinear dielectric , '' phys . rev . * 127 * , 19181939 ( 1962 ) .
j. stenger , h. schnatz , c. tamm , and h. r. telle , `` ultraprecise measurement of optical frequency ratios , '' phys .
rev .
lett . * 88 * , 073601 ( 2002 ) .
e. j. zang , j. p. cao , y. li , c. y. li , y. k. deng , and c. q. gao , `` realization of four - pass i@xmath4 absorption cell in 532-nm optical frequency standard , '' ieee transactions on instrumentation and measurement * 56 * , 673676 ( 2007 ) .
g. grosche , b. lipphardt , and h. schnatz , `` optical frequency synthesis and measurement using fibre - based femtosecond lasers , '' the european physical journal d - atomic , molecular , optical and plasma physics * 48 * , 2733 ( 2008 ) .
i. coddington , w. c. swann , l. lorini , j. c. bergquist , y. le coq , c. w. oates , q. quraishi , k. s. feder , j. w. nicholson , p. s. westbrook , s. a. diddams , and n. r. newbury , `` coherent optical link over hundreds of metres and hundreds of terahertz with subfemtosecond timing jitter , '' nature photonics * 1 * , 283287 ( 2007 ) .
t. rosenband , d. b. hume , p. o. schmidt , c. w. chou , a. brusch , l. lorini , w. h. oskay , r. e. drullinger , t. m. fortier , j. e. stalnaker , s. a. diddams , w. c. swann , n. r. newbury , w. m. itano , d. j. wineland , and j. c. bergquist , `` frequency ratio of al@xmath5 and hg@xmath5 single - ion optical clocks ; metrology at the 17th decimal place , '' science * 319 * , 18081812 ( 2008 ) .
t. m. fortier , n. ashby , j. c. bergquist , m. j. delaney , s. a. diddams , t. p. heavner , l. hollberg , w. m. itano , s. r. jefferts , k. kim , f. levi , l. lorini , w. h. oskay , t. e. parker , j. shirley , and j. e. stalnaker , `` precision atomic spectroscopy for improved limits on variation of the fine structure constant and local position invariance , '' phys .
rev .
lett . * 98 * , 070801 ( 2007 ) .
ligo scientific collaboration , `` ligo : the laser interferometer gravitational - wave observatory , '' reports on progress in physics * 72 * , 076901 ( 2009 ) . a. j. mullavey , b. j. j. slagmolen , j. miller , m. evans , p. fritschel , d. sigg , s. j. waldman , d. a. shaddock , and d. e. mcclelland , `` arm - length stabilisation for interferometric gravitational - wave detectors using frequency - doubled auxiliary lasers , '' optics express * 20 * , 8189 ( 2012 ) .
k. izumi , k. arai , b. barr , j. betzwieser , a. brooks , k. dahl , s. doravari , j. c. driggers , w. z. korth , h. miao , j. rollins , s. vass , d. yeaton - massey , r. x. adhikari , `` multi - wavelength cavity metrology , '' arxiv:1205.1257 ( 2012 ) f. khalili , s. danilishin , h. mller - ebhardt , h. miao , y. chen , and c. zhao , `` negative optical inertia for enhancing the sensitivity of future gravitational - wave detectors , '' phys .
rev .
d * 83 * , 062003 ( 2011 ) .
b. willke , n. uehara , e. k. gustafson , r. l. byer , p. j. king , s. u. seel , and j. r. l. savage , `` spatial and temporal filtering of a 10-w nd : yag laser with a fabry
perot ring - cavity premode cleaner , '' opt .
lett .
* 23 * , 17041706 ( 1998 )
. v. leonhardt and j. b. camp , `` space interferometry application of laser frequency stabilization with molecular iodine , '' appl . opt . *
45 * , 41424146 ( 2006 ) .
y. levin , `` fluctuation dissipation theorem for thermo - refractive noise , '' physics letters a * 372 * , 19411944 ( 2007 ) .
d. heinert , a. g. gurkovsky , r. nawrodt , s. p. vyatchanin , and k. yamamoto , `` thermorefractive noise of finite - sized cylindrical test masses , '' phys .
rev .
d * 84 * , 062001 ( 2011 ) .
l .- s .
ma , z. bi , a. bartels , l. robertsson , m. zucco , r. s. windeler , g. wilpers , c. oates , l. hollberg , and s. a. diddams , `` optical frequency synthesis and comparison with uncertainty at the 10@xmath6 level , '' science * 303 * , 18431845 ( 2004 ) .
l. robertsson , `` international comparison of @xmath7i@xmath4-stabilized frequency - doubled nd : yag lasers between the bipm , the nrlm and the bnm - inm , '' metrologia * 38 * ( 2001 ) .
j. e. stalnaker , s. a. diddams , t. m. fortier , k. kim , l. hollberg , j. c. bergquist , w. m. itano , m. j. delany , l. lorini , w. h. oskay , t. p. heavner , s. r. jefferts , f. levi , t. e. parker , and j. shirley , `` optical - to - microwave frequency comparison with fractional uncertainty of 10@xmath8 , '' applied physics b : lasers and optics * 89 * , 167176 ( 2007 ) .
rubiola , _ phase noise and frequency stability in oscillators _ ( cambridge university press , 2008 ) .
i. r. c. committee , `` characterization of frequency and phase noise , '' report ( 1986 ) . o. terra , g. grosche , k. predehl , r. holzwarth , t. legero , u. sterr , b. lipphardt , and h. schnatz , `` phase - coherent comparison of two optical frequency standards over 146 km using a telecommunication fiber link , '' applied physics b : lasers and optics * 97 * , 541551 ( 2009 ) .
10.1007/s00340 - 009 - 3653 - 2 . |
the strength of chiral interactions goes with derivatives of pion fields which allows one to organize a perturbative expansion in powers of pion momentum and mass the chiral perturbation theory ( @xmath3pt ) @xcite .
the small expansion parameter is @xmath4 , where @xmath5 is the momentum and @xmath6 gev stands for the scale of spontaneous chiral symmetry breaking .
based on this expansion one should be able to systematically compute the pion - mass dependence of static quantities , such as nucleon mass , magnetic moments , as well as the momentum dependence of scattering processes , such as pion - pion and pion - nucleon scattering .
this is an effective - field theory ( eft ) expansion , in this case a low - energy expansion of qcd .
one expects to obtain exactly the same answers as from qcd directly , provided the low - energy constants ( lecs ) the parameters of the effective lagrangian are known , either from experiment or by matching to qcd itself ( _ e.g. _ , by fitting to the lattice qcd results ) .
one of the principal ingredients of an eft expansion is _ power counting_. the power counting assigns an order to feynman graphs arising in loopwise expansion of the amplitudes , and thus defines which graphs need to be computed at a given order in the expansion . in a way it simply is a tool to estimate the size of different contributions without doing explicit calculations .
the main requirement on a power - counting scheme is that it should estimate the relative size of various contributions correctly . in @xmath3pt with pions and nucleons alone
@xcite , the power counting for a graph with @xmath7 loops , @xmath8 ( @xmath9 ) internal pion ( nucleon ) lines , and @xmath10 vertices from @xmath11th - order lagrangian , estimates its contribution to go as @xmath12 , with the power given by : n _ = 4 l - 2 n_- n_n + _ k k v_k .
note that in the manifestly lorentz - covariant formalism this power counting holds only in a specific renormalization scheme @xcite .
a cornerstone principle of effective field theories in general is _ naturalness _ , meaning that the dimensionless lecs must be of natural size , _
i.e. _ , of order of unity .
any significant fine - tuning of even a single lec leads , obviously , to a break - down of the eft expansion . therefore , even if an eft describes the experimental data , but at the expense of fine - tuned lecs , the eft expansion fails .
such an eft can still be useful for getting insights into the physics beyond the eft itself .
namely , by looking at the form of the fine - tuned operators , one might be able to deduce which contributions are missing .
for instance , it is known that the nlo @xmath3pt description of the pion - nucleon elastic scattering , near threshold , requires relatively large values for some of the lecs .
it is not difficult to see that the operators corresponding with those unnatural lecs can be matched to the `` integrated out '' @xmath13-resonance contributions .
the problem is that the @xmath13 is relatively light , its excitation energy , @xmath14 gev , is still quite small compared to @xmath15 gev .
integrating out the @xmath13-isobar degrees of freedom corresponds to an expansion in powers of @xmath16 , with @xmath17 , which certainly is not as good of an expansion as the one in the meson sector , in powers of @xmath4 .
the fine - tuning of the `` deltaless '' @xmath3pt seems to be lifted by the inclusion of an explicit @xmath13-isobar .
also , the limit of applicability of the eft expansion is then extended to momenta of order of the resonance excitation energy , @xmath18 .
such momenta can still be considered as soft , as long as @xmath19 can be treated as small .
the resulting @xmath3pt with pion , nucleon , and @xmath13-isobar degrees of freedom has two distinct light scales : @xmath20 and @xmath13 .
perhaps the most straightforward way to proceed is to organize a simultaneous expansion in two different small parameters : @xmath21 however , for power counting purposes it is certainly more convenient to have a single small parameter and thus a relation between @xmath22 and @xmath23 is usually imposed .
we stress that the relation is established only at the level of power counting and not in the actual calculations of graphs . at present
two such relations between @xmath22 and @xmath23 are used in the literature : * @xmath24 , the so - called `` small scale expansion '' @xcite , here refered to as the `` @xmath22-expansion '' ; + * @xmath25 , the `` @xmath23-expansion '' @xcite .
the table below ( ) summarizes the counting of momenta in the three expansions : deltaless ( @xmath26-@xmath3pt ) , @xmath22-expansion , and @xmath23-expansion .
.the counting of momenta in the three different @xmath3eft expansions . [ cols="^,^,^",options="header " , ] we conclude this discussion by giving the general formula for the power - counting index in the @xmath23-expansion . the power - counting index , @xmath27 , of a given graph simply tells us that the graph is of the size of @xmath28 . for a graph with @xmath7 loops , @xmath10 vertices of dimension @xmath11 , @xmath8 pion propagators , @xmath9 nucleon propagators , @xmath29 delta propagators , @xmath30 odr propagators and @xmath31 odi propagators ( such that @xmath32 ) the index is n = \ { cc 2 n _ - n _ , & p~m _ ; +
n _ - 3n_odr - n_odi , & p~ , . where @xmath33 , given by , is the index of the graph in @xmath3pt with no @xmath13 s . for further details on the @xmath23 counting
we refer to ref .
@xcite .
the @xmath34 form factors have been examined in both the @xmath22-expansion @xcite and the @xmath35-expansion @xcite . in the following we focus on the latter analysis .
the @xmath36 transition can be induced by a pion or a photon .
the correponding effective lagrangians are written as : ^(1)_n & = & n t^a ^ ( _ _ ) _ ^a + , + ^(2)_n & = & n t^3 ( _ _ ) f^ + , + ^(3)_n & = & n t^3 _ 5 f^+ , where @xmath37 , @xmath38 , @xmath39 stand respectively for the nucleon ( spinor , isodublet ) , @xmath0-isobar ( vector - spinor , isoquartet ) , pion ( pseudoscalar , isovector ) fields ; @xmath40 and @xmath41 are the electromagnetic field strength and its dual , @xmath42 are the isospin-1/2-to-3/2 transition @xmath43 ) matrices .
the coupling constants @xmath44 , @xmath45 , @xmath46 , and @xmath47 are the lecs describing the @xmath48 tansition at the tree level .
the @xmath49 transition is traditionally studied in the process of pion electroproduction on the nucleon .
let us consider this process to nlo in the @xmath23 expansion . since we are using the one - photon - exchange approximation , transition see refs .
the pion photoproduction can be viewed as the particular case of electroproduction at @xmath50 .
the pion electroproduction amplitude to nlo in the @xmath23-expansion , in the resonance region , is given by the graphs in , where the shaded blob in the 3rd graph denotes the nlo @xmath34 vertex , given by the graphs in .
the 1st graph in enters at the lo , which here is @xmath51 .
all the other graphs in are of nlo@xmath52 .
note that the @xmath13-resonance contribution at nlo is obtained by going to nlo in either the @xmath53 vertex ( 2nd graph ) or the @xmath34 vertex ( 3rd graph ) . accordingly ,
the @xmath13 self - energy in these graphs is included , respectively , to nlo .
the vector - meson diagram , ( 2 ) , contributes to nlo for @xmath54 .
one includes it effectively by giving the @xmath45-term a dipole @xmath55-dependence ( in analogy to how it is usually done for the nucleon isovector form factor ) : g_m .
the analogous effect for the @xmath46 and @xmath47 couplings begins at n@xmath56lo .
an important observation is that at @xmath50 only the imaginary part ( unitarity cut ) of the loop graphs in contributes to the nlo amplitude .
their real - part contributions , after the renormalization of the lecs , begin to contribute at n@xmath56lo , for @xmath57 . at present
we will consider only the nlo calculation where the @xmath58-loop contributions to the @xmath34-vertex are omitted since they do not give the imaginary contributions in the @xmath13-resonance region .
we emphasize that such loops might become important at this order for @xmath59 gev@xmath56 and should be included for the complete nlo result .
the present calculation is thus restricted to values @xmath60 gev@xmath56 . , @xmath61 , and @xmath62 multipoles for pion photoproduction as function of the invariant energy .
green dashed curves : @xmath0 contribution without the @xmath34-vertex loop corrections , [ i.e. , only the first three graphs in with ( 1 ) contribution are taken into account ] .
blue dotted curves : adding the born contributions , 2nd line in , to the dashed curves .
black solid curves : the nlo calculation , includes all graphs in as well as the loop corrections . the data point are from the said analysis ( fa04k ) @xcite ( red circles ) , and from the maid 2003 analysis @xcite ( blue squares ) . ,
title="fig : " ] , @xmath61 , and @xmath62 multipoles for pion photoproduction as function of the invariant energy .
green dashed curves : @xmath0 contribution without the @xmath34-vertex loop corrections , [ i.e. , only the first three graphs in with ( 1 ) contribution are taken into account ] .
blue dotted curves : adding the born contributions , 2nd line in , to the dashed curves .
black solid curves : the nlo calculation , includes all graphs in as well as the loop corrections . the data point are from the said analysis ( fa04k ) @xcite ( red circles ) , and from the maid 2003 analysis @xcite ( blue squares ) . ,
title="fig : " ] , @xmath61 , and @xmath62 multipoles for pion photoproduction as function of the invariant energy .
green dashed curves : @xmath0 contribution without the @xmath34-vertex loop corrections , [ i.e. , only the first three graphs in with ( 1 ) contribution are taken into account ] .
blue dotted curves : adding the born contributions , 2nd line in , to the dashed curves .
black solid curves : the nlo calculation , includes all graphs in as well as the loop corrections .
the data point are from the said analysis ( fa04k ) @xcite ( red circles ) , and from the maid 2003 analysis @xcite ( blue squares ) .
, title="fig : " ] in fig .
[ fig : gap_pin_m1mult ] shown are the resonant multipoles @xmath63 , @xmath61 , and @xmath62 as function of the invariant energy @xmath64 around the resonance position and @xmath50 . the @xmath63 and @xmath61 multipoles
are well established by the said @xcite and maid @xcite empirical partial - wave solutions , thus allowing one to fit two of the three @xmath34 lecs at this order as : @xmath65 , @xmath66 .
the third lec is adjusted to for a best description of the pion electroproduction data at low @xmath55 ( see ) , yielding @xmath67 .
the latter values translate into @xmath68 , @xmath69 , and @xmath70 for the jones scadron form - factors @xcite at @xmath50 .
as is seen from the figure , the nlo results ( solid curves ) give a good description of the energy dependence of the resonant multipoles in a window of 100 mev around the @xmath0-resonance position .
these values yield @xmath71 % and @xmath72 % .
the dashed curves in these figures show the contribution of the @xmath0-resonant diagram of the nlo loop corrections in . for the @xmath73 multipole
this is the lo and part of the nlo contributions .
for the @xmath74 and @xmath75 multipole the lo contribution is absent ( recall that @xmath46 and @xmath47 coupling are of one order higher than the @xmath45 coupling ) .
hence , the dashed curve represents a partial nlo contribution to @xmath74 and @xmath75 .
note that such a purely resonant contribution without the loop corrections satisfies unitarity in the sense of the fermi - watson theorem , which states that the phase of a pion electroproduction amplitude @xmath76 is given by the corresponding pion - nucleon phase - shift : @xmath77 . as a direct consequence of this theorem ,
the real - part of the resonant multipoles must vanish at the resonance position , where the phase - shift crosses @xmath78 degrees . upon adding the non - resonant born graphs ( 2nd line in ) to the dashed curves , one obtains the dotted curves .
the non - resonant contributions are purely real at this order and hence the imaginary part of the multipoles do not change .
while this is consistent with unitarity for the non - resonant multipoles ( recall that the non - resonant phase - shifts are zero at nlo ) , the fermi - watson theorem in the resonant channels is violated . in particular
, one sees that the real parts of the resonant multipoles now fail to cross zero at the resonance position .
the complete nlo calculation , shown by the solid curves in the figure includes in addition the @xmath79-loop corrections in , which restore unitarity .
the fermi - watson theorem is satisfied exactly in this calculation . in fig .
[ fig : epio_cross2 ] , the different virtual photon absorption cross sections around the resonance position are displayed at @xmath80 gev@xmath56 , where recent precision data are available .
we compare these data with the present calculations as well as with the results of sl , dmt , and duo models @xcite .
in the calculations , the low - energy constants @xmath45 and @xmath46 , were fixed from the resonant pion photoproduction multipoles . therefore , the only other low - energy constant from the chiral lagrangian entering the nlo calculation is @xmath47 .
the main sensitivity on @xmath47 enters in @xmath81 .
a best description of the @xmath82 data at low @xmath55 is obtained by choosing @xmath83 . from the figure one sees that the nlo calculation , within its accuracy , is consistent with the experimental data for these observables at low @xmath55 .
the dynamical models are in basic agreement with each other and the data for the transverse cross sections .
differences between the models do show up in the @xmath81 and @xmath84 cross sections which involve the longitudinal amplitude .
in particular for @xmath84 the differences reflect to a large extent how the non - resonant @xmath85 multipole is described in the models .
since the low - energy constants @xmath45 , @xmath46 , and @xmath47 have been fixed , our calculation can provide a prediction for the @xmath20 dependence of the @xmath86 transition form factors .
the study of the @xmath20-dependence is crucial to connect to lattice qcd results , which at present can only be obtained for larger pion masses ( typically @xmath87 mev ) . in fig .
[ fig : regmmpi ] we examine the @xmath20-dependence of the _ magnetic _ @xmath86-transition form factor @xmath88 , in the convention of jones and scadron @xcite . at the physical pion mass
, this form factor can be obtained from the imaginary part of the @xmath89 multipole at @xmath90 ( where the real part is zero by watson s theorem ) .
the value of @xmath88 at @xmath91 is determined by the low - energy constant @xmath45 .
the @xmath55-dependence then follows as a prediction of the nlo result , and fig .
[ fig : regmmpi ] shows that this prediction is consistent with the experimental value at @xmath92 gev@xmath56 and physical pion mass .
the @xmath20-dependence of @xmath88 is also completely fixed at nlo , no new parameters appear . in fig .
[ fig : regmmpi ] , the result for @xmath88 at @xmath92 gev@xmath56 is shown both when the @xmath20-dependence of the nucleon and @xmath0 masses is included ( solid line ) and when it is not ( dotted line ) .
accounting for the @xmath20-dependence in @xmath93 and @xmath94 , significantly affects @xmath88 .
the calculation , with the @xmath20 dependence of @xmath95 and @xmath94 included , is in a qualitatively good agreement with the lattice data shown in the figure .
the result also follows an approximately linear behavior in @xmath96 , although it falls about 10 - 15 % below the lattice data .
this is just within the uncertainty of the nlo results .
one should also keep in mind that the present lattice simulations are not done in full qcd , but are `` quenched '' , so discrepancies are not unexpected . in
, we show the @xmath20-dependence of the ratios @xmath97 and @xmath98 and compare them to lattice qcd calculations .
the recent state - of - the - art lattice calculations of @xmath97 and @xmath98 @xcite use a _ linear _ , in the quark mass ( @xmath99 ) , _ extrapolation _ to the physical point , thus assuming that the non - analytic @xmath100-dependencies are negligible . the thus obtained value for @xmath98 at the physical @xmath20 value displays a large discrepancy with the experimental result , as seen in .
our calculation , on the other hand , shows that the non - analytic dependencies are _ not _ negligible .
while at larger values of @xmath20 , where the @xmath0 is stable , the ratios display a smooth @xmath20 dependence , at @xmath101 there is an inflection point , and for @xmath102 the non - analytic effects are crucial , as was also observed for the @xmath13-resonance magnetic moment @xcite .
one also sees from that , unlike the result for @xmath88 , there is little difference between the calculations with the @xmath20-dependence of @xmath93 and @xmath94 accounted for , and our earlier calculation @xcite , where the ratios were evaluated neglecting the @xmath20-dependence of the masses .
this is easily understood , as the main effect due to the @xmath20-dependence of @xmath93 and @xmath94 arises due to a common factor in the evaluation of the @xmath86 form factors , which drops out of the ratios .
one can speculate that the `` quenching '' effects drop out , at least partially , from the ratios as well .
in we also show the @xmath20-dependence of the @xmath103 transition ratios , with the theoretical error band .
the @xmath20 dependence obtained here from clearly shows that the lattice results for @xmath98 may in fact be consistent with experiment .
we presented here an extension of the chiral perturbation theory framework to the energy domain of @xmath13(1232 ) . in this extension
the @xmath13-isobar appears as an explicit degree of freedom in the effective chiral lagrangian .
the other low - energy degrees of freedom appearing in this lagrangian are pions , nucleons , and @xmath104-mesons .
the power counting depends crucially on how the @xmath13-resonance excitation energy , @xmath105 , compares to the other scales in the problem . in the `` @xmath23-expansion '' scheme adopted here ,
one utilizes the scale hierarchy , @xmath106 .
in other words , we have an eft with two distinct light scales .
such a scale hierarchy is crucial for an adequate counting of the @xmath13-resonance contributions in both the low - energy and the resonance energy regions .
the @xmath23 power counting provides a justification for `` integrating out '' the resonance contribution at very low energies , as well as for resummation of certain resonant contributions in the resonance region .
we applied the @xmath23-expansion to the process of pion electroproduction .
this is a first study of this reaction in the @xmath107-resonance region .
our resulting next - to - leading order calculation was shown to satisfy the electromagnetic gauge and chiral symmetries , lorentz - covariance , analyticity , unitarity ( watson s theorem ) .
the free parameters entering at this order are the @xmath34 couplings @xmath45 , @xmath46 , @xmath47 characterizing the @xmath108 , @xmath109 , @xmath110 transitions , respectively . by comparing our nlo results with the standard multipole solutions ( maid and said ) for the photoproduction multipoles we have extracted @xmath45 = 2.97 and @xmath66 , corresponding to @xmath111 % .
the nlo result was also found to give a good description of the energy - dependence of most non - resonant @xmath112 , @xmath5 and @xmath113-wave photoproduction multipoles in a 100 mev window around the @xmath0-resonance position . from the pion electroproduction cross - section @xmath114
we have extracted @xmath115 , which yields @xmath116 % near @xmath92 gev@xmath56 . in overall ,
the nlo results are consistent with the experimental data of the recent high - precision measurements at mami and bates .
the framework plays a _ dual role _ in that it allows for an extraction of resonance parameters from observables _ and _ predicts their pion - mass dependence . in this way it may provide a crucial connection of present lattice qcd results ( obtained at larger than physical values of @xmath20 ) to the experiment .
we have shown here that the opening of the @xmath117 decay channel at @xmath118 induces a pronounced non - analytic behavior of the @xmath97 and @xmath98 ratios . while the linearly - extrapolated lattice qcd results for @xmath98 are in disagreement with experimental data , the prediction of the non - analytic dependencies suggests that these results are in fact consistent with experiment .
the presented results are systematically improvable .
we have indicated what are the next - next - to - leading order effects , however , at present we could only estimate the theoretical uncertainty of our calculations due to such effects .
we have defined and provided a corresponding error band on our nlo results .
an actual calculation of n@xmath56lo effects is a worthwhile topic for a future work .
we thank aron bernstein and costas papanicolas for invitation to this stimulating workshop and generous support .
this work is partially supported by doe grant no.de-fg02-04er41302 and contract de - ac05 - 06or23177 under which jefferson science associates operates the jefferson laboratory .
j. gasser , m. e. sainio and a. svarc , nucl .
b * 307 * , 779 ( 1988 ) .
j. gegelia and g. japaridze , phys .
d * 60 * , 114038 ( 1999 ) ; j. gegelia , g. japaridze and x. q. wang , j. phys .
g * 29 * , 2303 ( 2003 ) .
v. pascalutsa and m. vanderhaeghen , phys .
b * 636 * , 31 ( 2006 ) . g. c. gellas , t. r. hemmert , c. n. ktorides and g. i. poulis , phys . rev .
d * 60 * , 054022 ( 1999 ) .
t. a. gail and t. r. hemmert , arxiv : nucl - th/0512082 ; see also the contribution of t. gail to these proceedings .
v. pascalutsa and m. vanderhaeghen , phys .
232001 ( 2005 ) . v. pascalutsa and m. vanderhaeghen , phys .
d * 73 * , 034003 ( 2006 ) . v. pascalutsa , c. e. carlson and m. vanderhaeghen , phys . rev . lett . * 96 * , 012301 ( 2006 ) .
s. kondratyuk and p. g. blunden , nucl .
a * 778 * , 44 ( 2006 ) [ arxiv : nucl - th/0601063 ] .
v. pascalutsa and j. a. tjon , phys .
c * 61 * , 054003 ( 2000 ) ; _ ibid .
_ * 70 * , 035209 ( 2004 ) ; g. l. caia _
et al . _ ,
c * 70 * , 032201(r ) ( 2004 ) ; _ ibid .
_ * 72 * , 035203 ( 2005 ) . c. mertz _ et al . _ ,
lett . *
86 * , 2963 ( 2001 ) . c. kunz _ et al .
_ , phys .
b * 564 * , 21 ( 2003 ) .
n. f. sparveris _ et al .
_ [ oops collaboration ] , phys . rev .
lett . * 94 * , 022003 ( 2005 ) ; see also the contribution of nikos sparveris to these proceedings .
i. c. cloet , d. b. leinweber and a. w. thomas , phys .
b * 563 * , 157 ( 2003 ) ; r. d. young , d. b. leinweber and a. w. thomas , nucl .
suppl . * 129 * , 290 ( 2004 ) .
v. pascalutsa and m. vanderhaeghen , phys .
lett . * 94 * , 102003 ( 2005 ) . | we describe the pion electroproduction processes in the @xmath0(1232)-resonance region within the framework of chiral effective - field theory . by studying the observables of pion electroproduction in a next - to - leading order calculation
we are able to make predictions and draw conclusions on the properties of the @xmath1 electromagnetic form factors .
address = physics department , the college of william & mary , williamsburg , va 23187 , usa , altaddress = theory center , jefferson lab , 12000 jefferson ave , newport news , va 23606 , usa address = physics department , the college of william & mary , williamsburg , va 23187 , usa , altaddress = theory center , jefferson lab , 12000 jefferson ave , newport news , va 23606 , usa for the general overview and motivation of the electromagnetic @xmath2 transition
we refer the reader to the opening part of these proceedings and to our recent review @xcite . in this short contribution
we will focus on the chiral effective - field theoretic study of the @xmath2 transition . |
the present neutrino experimental data @xcite exhibit that the atmospheric neutrino deficit points toward a maximal mixing between the tau and muon neutrinos , however , the solar neutrino deficit favors a not - so - maximal mixing between the electron and muon neutrinos , @xmath11 , @xmath12 , where @xmath13 is the leptonic pmns mixing matrix .
in addition , although we do not have yet any firm evidence for the neutrino oscillation arisen from the 1st and 3rd generation flavor mixing , there is a bound on the mixing element @xmath14 from chooz reactor experiment , @xmath15 @xcite .
although neutrinos have gradually revealed their properties in various experiments since the historic super - kamiokande confirmation of neutrino oscillations @xcite , properties related to the leptonic cp violation are completely unknown yet .
cp violations in the leptonic sector are also obligatory , if the matter and antimatter asymmetry of the universe is made through leptogenesis scenario @xcite in the seesaw models @xcite .
recently , the so - called @xmath16-@xmath17 reflection symmetry @xcite has been imposed in light neutrino mass matrices so that the maximal atmospheric mixing @xmath18 as well as the maximal value for the dirac cp violating phase @xmath19 are predicted @xcite . under the @xmath16-@xmath17 reflection symmetry in the context of a seesaw model
, it is supposed that the right - handed neutrinos transform into the charge conjugates of themselves , @xmath20 and the light left - handed neutrinos into @xmath21 @xcite with the corresponding phases @xmath22 . the effective mass matrix derived from the seesaw mechanism is invariant under the @xmath1 reflection symmetry , too . in the light of the cp violation from the neutrino oscillations @xcite
, the @xmath1 reflection symmetry indicates that , for given values of mixing angles , the cp asymmetry @xmath23 is maximal . in this paper , we consider a seesaw model with the @xmath1 reflection symmetry .
in contrast with other models concerned with the @xmath1 reflection symmetry , we impose the exact symmetry at the gut scale and consider renormalization group ( rg ) evolution @xcite on the neutrino dirac yukawa couplings , heavy majorana neutrino masses and cp phases by running from the gut scale to a seesaw scale .
one of interesting issues concerned with neutrino models with the @xmath1 reflection symmetry is if baryogenesis can be successfully realized through leptogenesis or not .
in fact , high energy cosmological cp violation responsible for baryogenesis vanishes as long as the @xmath1 reflection symmetry is kept exact and flavor effects associated with the charged leptons are not taken into account @xcite .
however , it has been shown that flavor effects can lead to a successful leptogenesis for @xmath24 gev in the limit of the @xmath1 reflection symmetry as long as a partial lepton asymmetry associated with a lepton flavor does not vanish even if total lepton asymmetry is zero @xcite . according to our numerical estimates made in this paper , it is not true in some cases of heavy majorana neutrino mass spectrum with hierarchy structures like @xmath25 and @xmath26 . in such cases , as will be shown later , breaking of the @xmath1 reflection symmetry can lead to successful leptogenesis . instead of introducing _ ad - hoc _ soft symmetry breaking terms , in this paper , we consider a possibility that rg evolutions of neutrino parameters can beak the @xmath1 reflection symmetry and then examine if such breaking effects due to the rg evolutions can lead to successful baryogenesis . to do that
, we simply assume that two lighter heavy majorana neutrino masses are exactly degenerate at the gut scale .
if a mass splitting between two degenerate heavy majorana neutrinos due to the rg evolution comes out to be small , we expect that the lepton asymmetry can be resonantly enhanced @xcite . however , we show that even though the total lepton asymmetry can be resonantly enhanced , the magnitude of the total lepton asymmetry is not enough to achieve successful baryogenesis in the context of the standard model ( sm ) with the @xmath1 reflection symmetry broken by rg corrections
. in general , the absolute magnitudes of the lepton asymmetries for each lepton flavor can be larger than that of the total lepton asymmetry .
furthermore , since the interactions mediated by the tau and muon coupling are in thermal equilibrium at below the temperature @xmath27 gev , the processes which wash out lepton number are flavor dependent and thus the lepton asymmetries for each flavor should be treated separately with different wash - out factors .
it has been known that flavor effects can enlarge the area of parameter space where leptogenesis can work .
thus , it is meaningful to investigate if leptogenesis including flavor effects can be successfully realized in the case that leptogenesis without flavor effects does not work successfully . in this paper , we shall examine if the flavor effects on leptogenesis can help to achieve successful baryogenesis in the context of the sm with the @xmath1 reflection symmetry broken by rg corrections . in particular , we shall carefully discuss how the wash - out factors for each lepton flavor can be significant to achieve successful leptogenesis .
in the supersymmetric seesaw model ( ssm ) , the baryon asymmetry of our universe can be achieved leptogenesis via the decays of heavy majorana neutrinos .
one of interesting points is that the lepton asymmetry in the ssm can be enhanced in the case of large @xmath10 because it is proportional to the charged lepton yukawa couplings @xmath28 .
thus , even for the parameter space where leptogenesis in the sm does not work , supersymmetry can make leptogenesis working .
we also expect that since the rg evolutions in the ssm are different from those in the sm , the parameter space where leptogenesis can work must be different each other . in this paper
, we shall study how the results concerned with leptogenesis in the sm can change when we supersymmetrize it .
this paper is organized as follows : in sec .
ii , we present a seesaw model reflecting @xmath1 reflecting symmetry at the gut scale in the context of both sm and ssm . in sec .
iii , we discuss how we can obtain a rg - improved neutrino mass matrix . we show that the rg evolutions break the @xmath1 reflection symmetry and derive the deviations of low energy neutrino mixing angles and cp phase from their tree level values which can be parameterized in terms of a parameter concerned with the rg evolution .
the discussion for rg evolution from the gut scale to low scale and useful formulae are given in appendix . in sec .
iv , we show how successful leptogenesis can be radiatively induced in our scheme . here , we discuss how lepton flavor effects are important to achieve successful leptogenesis and carefully estimate the wash - out factors by taking flavor effects into account . numerical results and conclusion
are given in sec .
to begin with , let us consider the lagrangian of the lepton sector from which the seesaw mechanism works , @xmath29 where the family indices have been omitted and @xmath30 , @xmath31 stand for su(2 ) lepton doublet fields , charged lepton singlet fields and higgs scalar and singlet heavy majorana neutrino , respectively . in the above lagrangian , @xmath32 , @xmath33 and @xmath34 are the @xmath35 heavy majorana neutrino mass matrix , charged lepton and neutrino dirac yukawa matrices , respectively . after spontaneous electroweak symmetry breaking , the seesaw mechanism leads to the following effective light neutrino mass term , @xmath36 where @xmath37 is a vacuum expectation value of the higgs field with @xmath38 gev .
let us impose the @xmath1 reflection symmetry for the neutrino sectors at the gut scale , which is the combined operation of @xmath1 flavor exchange in pmns mixing matrix and cp transformation on the leptonic sector @xcite , in the basis where both the charged lepton mass and heavy majorana mass matrices are diagonal . under the @xmath1 reflection symmetry ,
the right - handed neutrinos transform into the charge conjugates of themselves , @xmath20 and thus the elements of @xmath32 should be real .
then , the neutrino dirac yukawa matrix and the heavy majorana neutrino mass matrix , at the gut scale , are taken to be @xmath39 where @xmath40 of the neutrino dirac yukawa matrix are all real .
we assume that the two heavy majorana neutrinos are exactly degenerate in mass , which is much smaller than that of the right - handed neutrino @xmath41 , _ i.e. _ @xmath42 .
it turns out , after our analysis , that in the light of seesaw mechanism with such a hierarchy of heavy majorana neutrinos the effects of the parameters @xmath43 and @xmath44 on low energy neutrino phenomenology and even leptogenesis are negligibly small .
thus , for our convenience , we take @xmath45 and @xmath46 at the gut scale . introducing several parameters defined by the ratios among the parameters appeared in eq .
( [ input1 ] ) , as follows , @xmath47 the effective light neutrino mass matrix generated through seesaw mechanism can be written as @xmath48 where @xmath49 we see that the effective neutrino mass matrix @xmath50 reflects the @xmath1 reflection symmetry @xcite . here
, we have not yet considered rg corrections to @xmath50 from the gut scale to a seesaw scale , so the form of @xmath50 given in eq .
( [ meff1 ] ) is regarded at tree level .
it is not difficult to see that the seesaw model based on eq .
( [ input1 ] ) leads to the normal hierarchical light neutrino mass spectrum because we take diagonal form of heavy majorana neutrino mass matrix @xcite .
thus , the rg effects from a seesaw scale to electroweak scale on @xmath50 as well as on the neutrino mixing matrix @xmath51 are expected to be very small .
starting from the @xmath1 reflection symmetric forms of @xmath2 and @xmath52 , we shall show that the @xmath1 reflection symmetry is broken due to the rg running effects between the gut and a seesaw scale and they can lead to successful leptogenesis without being in conflict with experimental low energy constraints .
concerned with cp violation , we notice from eq .
( [ meff1 ] ) that the cp phases @xmath53 coming from @xmath34 take part in low - energy cp violation . to see how the cp phases @xmath53 are correlated with low energy cp violation measurable through neutrino oscillations ,
let us consider cp violation parameter defined through jarlskog invariant @xcite @xmath54}{\delta m^{2}_{21}\delta m^{2}_{31}\delta m^{2}_{32 } } , \label{cp0 } \end{aligned}\ ] ] where @xmath55 and @xmath56 . from the matrix given in eq .
( [ meff1 ] ) it can be written as @xmath57\simeq m^{6}_{0}a\kappa\omega\sin\delta\phi_{12}+o(\eta^{-1 } ) , \end{aligned}\ ] ] where @xmath58 is the lengthy function composed of the parameters @xmath59 and @xmath60 .
we see from the above equation that @xmath61 is correlated with high energy cp parameter @xmath62 . as long as @xmath63 and @xmath64 ,
@xmath61 have non - vanishing value , which would be a signal of cp violation . the combination of dirac neutrino yukawa matrices , which is relevant for leptogenesis , is given by @xmath65 from this , we find that the hermitian quantity @xmath66 in the limit of the @xmath1 reflection symmetry leads to @xmath67=0 $ ] and thus vanishing lepton asymmetry which is undesirable for a successful leptogenesis . to generate non - vanishing lepton asymmetry , both the mass degeneracy of the 1st and 2nd heavy majorana neutrinos and the @xmath1 reflection symmetric texture of @xmath34 in eq .
( [ input1 ] ) should be broken . in the ssm ,
the leptonic superpotential is given by @xmath68 where the family indices have been omitted and @xmath69 stands for the chiral super - multiplets of the @xmath70 doublet lepton fields , @xmath71 are the higgs doublet fields with hypercharge @xmath72 , @xmath73 and @xmath74 are the super - multiplet of the @xmath70 singlet neutrino and charged lepton field , respectively . after spontaneous electroweak symmetry breaking , the seesaw mechanism leads to a following effective light neutrino mass term , @xmath75 where @xmath76 . for our purpose , we take the forms of @xmath34 and @xmath32 in the ssm to be the same forms given in eq .
( [ meff1 ] ) .
considering the rg effects in @xmath2 and @xmath77 , we obtain a rg improved effective neutrino mass matrix via the seesaw reconstruction at the decoupling scales of the heavy neutrino singlets .
the effective mass matrix at low energies can be found by running @xmath78 from the decoupling scales to the electroweak scale @xmath79 .
the procedures of the rg running and the useful formulae are given in appendix .
however , in the case of the normal hierarchical neutrino mass spectrum of light neutrinos , the rg running effects from the decoupling scales of heavy majorana neutrinos to the electroweak scale are turned out to be negligibly small . ignoring the rg running effects from seesaw scale to the electroweak scale in eqs .
( [ seesaw1],[seesaw11 ] ) , we can obtain the rg improved effective neutrino mass matrix at low energy presented as follows , @xmath80 where the parameter @xmath81 corresponding to the rg correction is given as @xmath82 we recast eq .
( [ cmatrix ] ) with the transformation @xmath83 and @xmath84 , as @xmath85 where @xmath86 and @xmath87 it is quite interesting to notice that the size of @xmath81 is very small , which breaks the @xmath1 reflection symmetry slightly in the mass matrix @xmath50 , however , this tiny breaking effects can play a crucial role in linking between low energy neutrino data and leptogenesis , as will be shown later .
this neutrino mass matrix is diagonalized by the pmns mixing matrix @xmath51 , @xmath88 $ ] , where @xmath89 indicates the mass eigenvalues of light majorana neutrinos . as a result of the rg effects ,
cp violation phase @xmath4 is shifted by a tiny amount , @xmath90
: @xmath91 here if the exact @xmath1 reflection symmetry is recovered , then @xmath92 and @xmath90 goes to zero .
atmospheric mixing angle @xmath3 is also deviated from @xmath93 due to the rg corrections , and the shift is approximately written by @xmath94 the magnitude of the unknown angle @xmath9 , which will complete our knowledge of neutrino mixing , can be written as @xmath95 note here that @xmath96 can be directly linked with leptogenesis through @xmath97 defined in eq .
( [ leptocomp ] ) .
solar neutrino mixing is also presented by @xmath98 the ratio between @xmath99 and @xmath100 can be approximated by the exact @xmath1 reflection parameter as @xmath101 which is independent of the rg correction . from eqs .
( [ deltacp1]-[sol1 ] ) , the jarlskog invariant @xmath61 can be approximated for @xmath102 by @xmath103 indicating that @xmath61 mainly depends on @xmath8 and @xmath9 , which is in turn proportional to @xmath104 defined in eq .
( [ cmatrix11 ] ) .
let us consider the cp asymmetry generated by the decays of the heavy majorana neutrinos @xmath105 ( i=1 , 2 ) .
in a basis where the right - handed majorana neutrino mass matrix is real and diagonal , the cp asymmetry generated through the interference between tree and one - loop diagrams for the decay of the heavy majorana neutrino @xmath105 is given , for each lepton flavor @xmath106 , by @xcite @xmath107}\\ & = & \frac{1}{8\pi(y_{\nu}y^{\dag}_{\nu})_{ii}}\sum_{j\neq i}{\rm im}\big\{(y_{\nu}y^{\dag}_{\nu})_{ij}(y_{\nu})_{i\alpha}(y_{\nu})^{\ast}_{j\alpha}\big\}g\big(\frac{m^{2}_{j}}{m^{2}_{i}}\big ) , \label{cpasym1 } \end{aligned}\ ] ] where the function @xmath108 is given by @xmath109~. \label{decayfunction } \end{aligned}\ ] ] here @xmath110 denotes a generation index and @xmath111 is the decay width of the @xmath110th - generation right - handed neutrino .
note here that the flavor effects generated due to the term @xmath112 disappears in the summation of @xmath113 for all flavors .
in order for @xmath114 to be non - vanishing , not only breaking of the degeneracy of right - handed neutrinos but also non - vanishing @xmath115 $ ] and/or @xmath116 $ ] are required at a seesaw scale @xmath117 .
note here that @xmath136 , which is justified from eq .
( [ tan2alpha ] ) , and the sign of plus and minus in eq .
( [ flalepto1 ] ) correspond to the case of the decay of @xmath137 and @xmath138 , respectively .
it is worthwhile to notice that since the angle @xmath139 given by eq .
( [ tan2alpha ] ) is limited by @xmath140 , @xmath141 ( order of 10 ) is always greater than @xmath142 ( order of 1 ) , as will be shown later .
this implies that @xmath143 is dominant over @xmath144 because of @xmath145 . since the total cp asymmetries @xmath146 are obtained by summing @xmath130 over the lepton flavors @xmath139 , with the help of eqs .
( [ leptocomp],[tan2alpha],[deltam],[rgparameter ] ) , we can express the lepton asymmetries as follows ) can also be obtained directly by using eq .
( [ imreh ] ) . ]
@xmath147 where the sign of minus and plus correspond to the case of the decay of @xmath137 and @xmath138 , respectively . as indicated in eq .
( [ flalepto ] ) , the cp asymmetries @xmath114 weakly depend on the heavy majorana neutrino scale @xmath117 . from eqs .
( [ flalepto1],[convlepto2 ] ) , we see that @xmath148 and @xmath149 . therefore , the cp asymmetry @xmath150 gets enhanced by @xmath151 due to flavor effects @xcite . below temperature @xmath152 gev
, it is known that muon and tau charged lepton yukawa interactions are much faster than the hubble expansion parameter rendering the @xmath16 and @xmath17 yukawa couplings in equilibrium .
then , the processes which wash out lepton number are flavor dependent and thus the lepton asymmetries for each flavor should be treated separately with different wash - out factors .
once the initial values of @xmath114 are fixed , the final result of @xmath153 or @xmath154 can be obtained by solving a set of flavor - dependent boltzmann equations including the decay , inverse decay , and scattering processes as well as the non - perturbative sphaleron interaction . in order to estimate the wash - out effects , we introduce the parameters @xmath155 which are the wash - out factors mainly due to the inverse decay of the majorana neutrino @xmath105 into the lepton flavor @xmath156 @xcite . the explicit form of @xmath155 is given by @xmath157 where @xmath158 ev , and @xmath159 and @xmath160 denote the partial decay rate of the process @xmath161 and the hubble parameter at temperature @xmath162 , respectively
. the decay rate of @xmath105 to the leptons with flavor @xmath163 is parameterized by @xmath164 and the trace @xmath165 coincides with the parameter @xmath166 .
from eq .
( [ k - factor2 ] ) and eq .
( [ seesaw1 ] ) , the wash - out parameters are given as @xmath167 the final baryon asymmetry @xmath168 is then given by @xcite @xmath169~. \label{baryon3 } \end{aligned}\ ] ] notice that each cp asymmetry for a single flavor given in eq .
( [ baryon3 ] ) is weighted differently by the corresponding wash - out parameter given by eq .
( [ k - factor2 ] ) , and appears with different weight in the final formula for the baryon asymmetry @xcite . from our numerical estimate
, we found that the wash - out factors @xmath170 except for @xmath171 are much greater than one indicating that corresponding lepton flavor asymmetries are strongly washed out , which is undesirable for successful leptogenesis . since @xmath172 can be much less than one and the lepton asymmetries given in eq .
( [ flalepto1 ] ) are the same order of magnitude , the lepton asymmetry for electron flavor in the weak wash - out regime leads to the dominant contribution to the lepton asymmetry , which can be enough to give rise to successful baryogenesis .
thus , @xmath168 is approximately given by @xmath173 and the magnitudes of @xmath174 and @xmath171 are approximately @xmath175 in the weak wash - out regime ( @xmath176 ) , the lepton asymmetry @xmath177 generated through the decay of @xmath138 is given by @xcite @xmath178 therefore , the resulting baryon - to - photon ratio @xmath179 can be simply given as @xmath180 as will be shown later , the wash - out factor @xmath181 is proportional to the low energy mixing angle @xmath9 and there is a connection between @xmath179 and low energy neutrino observables such as @xmath9 and @xmath61 .
on the other hand , the ratio of @xmath6 without lepton flavor effects compared to @xmath182 with lepton flavor effects @xcite is presented by @xmath183 thus , without taking lepton flavor effects into account , the prediction of @xmath6 is suppressed by @xmath184 , _
i.e. , _ @xmath185 orders of magnitude compared with @xmath182 , which is too small to give a successful leptogenesis .
it is worthwhile to notice that the radiatively induced lepton asymmetries @xmath192 are proportional to @xmath193 , and thus for large @xmath10 it can be highly enhanced and proportional to @xmath194 .
furthermore , it has an explicit dependence of the evolution scale @xmath129 .
these two points are already mentioned in ref .
taking lepton flavor effects into account , in this case the final baryon asymmetry is given as @xcite @xmath195~. \label{baryon3 m } \end{aligned}\ ] ] similar to the sm , below temperatures @xmath196 , muon and tau charged lepton yukawa interactions are much faster than the hubble expansion parameter rendering the @xmath16 and @xmath17 yukawa couplings in equilibrium . to obtain the final baryon asymmetry survived
, we should consider the wash - out factors .
similar to the sm , the magnitudes of @xmath197 except for @xmath198 are greater than one in our scenario .
so , we will consider two cases depending on the magnitude of @xmath198 : ( i)@xmath199 and @xmath200 , ( ii ) @xmath201 .
this case implies that the muon and tau lepton asymmetries are strongly washed out , whereas the electron asymmetry generated through the decay of @xmath120 is weakly washed out .
the form of @xmath177 in the weak wash - out regime in the ssm is the same as that in the sm given by eq .
( [ weak ] ) . in this case ,
similar to the sm , the contribution of @xmath177 to @xmath168 is dominant over the others .
thus , @xmath168 is approximately give by @xmath203 and the magnitudes of @xmath204 and @xmath171 are approximately @xmath205 the resulting baryon - to - photon ratio @xmath179 is simply given as @xmath206 just like the sm , one can see that there is a connection between baryon asymmetry and low energy neutrino observables .
the ratio of @xmath6 without lepton flavor effects to @xmath182 with lepton flavor effects are simply given by @xcite @xmath207 from this result , we see that without taking lepton flavor effects into account the prediction of @xmath6 is suppressed by @xmath208 , compared with @xmath182 . however , on the contrary to the flavor independent leptogenesis in the sm , it is possible to get a right amount of baryon asymmetry without flavor effects for the case of large @xmath209 . in the strong wash - out regime @xmath211 ,
the lepton asymmetry for the flavor @xmath163 generated through the decay of @xmath105 is given by @xcite @xmath212 in this case , given the initial thermal abundance of @xmath105 and the condition for @xmath155 , the baryon asymmetry including lepton flavor effects is approximately given by @xcite @xmath213 here , we note that each lepton asymmetry @xmath113 are almost the same order of magnitudes , even though @xmath214 for large @xmath10 and @xmath150 for small @xmath10 are slightly dominant within a few factors . according to our numerical estimate ,
the typical sizes of the suppression factors @xmath215 are @xmath216 which indicates that the flavor dependent wash - out factors equally contribute to leptogenesis . using eq .
( [ kstrong ] ) , the resulting baryon - to - photon ratio @xmath179 can be simply approximated as an order of magnitude @xmath217 which represents that at least @xmath218 is needed for the flavored leptogenesis to successfully work ( see , fig .
[ fig6 ] in sec .
v ) . as a consequence , in the strong wash - out case
, the resulting baryon asymmetry is too small to give a successful leptogenesis in the sm , whereas at least @xmath218 would be necessary to accommodate required baryon asymmetry in the ssm , and there exists also a connection between leptogenesis and low energy observable @xmath8 .
note that the strong wash - out case only allow @xmath9 to be @xmath219 which is from the condition @xmath220 .
in order to estimate the rg evolutions of neutrino dirac - yukawa matrix and heavy majorana neutrino masses from the gut scale to a seesaw scale , we numerically solve all the relevant rg equations presented in @xcite . in our numerical calculation of the rg evolutions
, we first fix the values of two masses of heavy majorana neutrinos with hierarchy @xmath221 to be @xmath222 , and @xmath223 as inputs can be as light as @xmath224 gev in our scenario . ]
. then we solve the rg equations by varying input values of all the parameter space @xmath225 given at the gut scale .
then , finally we select the parameter space allowed by low energy neutrino experimental data . at present
, we have five experimental data , which are taken as low energy constraints in our numerical analysis , given at @xmath226 by @xcite , @xmath227\leq2.8~,\quad\quad ~~7.1\leq\delta m^{2}_{\rm sol}[10^{-5}{\rm ev}^{2}]\leq8.3 , \label{exp bound } \end{aligned}\ ] ] where @xmath228 and @xmath229 . in addition
, the current measurement of baryon asymmetry of our universe @xmath230 we take in our numerical calculation is given by @xcite @xmath231 using the results of the rg evolutions , we estimate baryon asymmetry for the parameter space constrained from the low energy experimental data . for this case of wash - out factors , we found that baryogenesis could be successfully implemented in the sm and the ssm through electron flavor dominant leptogenesis .
in addition , we show that the successful leptogenesis could be linked with low energy observables ( @xmath9 and @xmath61 ) through the rg parameter @xmath133 given in eq .
( [ rgparameter ] ) as well as @xmath171 . in fig .
[ fig1 ] , we show the scatter plots for the model parameters in the sm ( upper panel ) and the ssm ( @xmath232 ) ( lower panel ) constrained by the experimental data given in eq .
( [ exp bound ] ) : the figures in the left - hand side exhibit how the parameters @xmath233 , @xmath234 , @xmath235 and @xmath236 are correlated with the angle @xmath139 defined in eq .
( [ tan2alpha ] ) , respectively and the figures in the right - hand side represent how the difference between the phases @xmath237 and @xmath238 in eq .
( [ input1 ] ) , @xmath239 , depends on the angle @xmath139 . in fig .
[ fig2 ] , we present how the wash - out factor @xmath198 ( fig . 2-a ) and cp violating observable @xmath61 ( fig .
2-b ) are correlated with the mixing angle @xmath9 .
the circles , triangles , stars and spots in fig . 2 correspond to the cases of the sm , and the ssm with @xmath240 , @xmath232 and @xmath241 , respectively . from fig .
2-(a ) , we see that in the case @xmath242 ( weak wash - out regime ) the mixing angle @xmath9 is limited . even in the case of the ssm with @xmath243 ,
the upper limit of @xmath9 is about 9 degree .
it is remarkable from fig .
2-(a ) that @xmath179 is expected to be enhanced through @xmath171 as @xmath9 gets lower . thus , this case with @xmath242 favors low values of @xmath9 .
the vertical dotted lines in fig .
2-(b ) indicate the values of @xmath9 corresponding to best - fit value of baryon asymmetry of our universe , @xmath244 , when @xmath245 and 60 , respectively , whereas the vertical solid line corresponds to the case of the sm . as can be seen in fig .
[ fig2]-(b ) , the prediction of @xmath61 for the ssm with a low @xmath10 can be measurable in the near future from long baseline neutrino oscillations . in fig .
[ fig3 ] , we present the predictions of ( a ) @xmath179 and ( b ) @xmath153 for the sm ( triangle ) , the ssm with @xmath240 ( daggers ) , 10 ( asters ) and 60 ( circles ) as a function of the mixing angle @xmath9 . in the case of flavor
dependent leptogenesis , the allowed values of @xmath182 prefers very low values of @xmath9 .
the horizontal dotted lines and solid line in fig .
[ fig3 ] ( also in figs .
[ fig4 ] and fig .
[ fig6 ] ) correspond to the current bounds on @xmath6 .
we see from fig .
[ fig3 ] that the prediction of @xmath179 in the ssm for a fixed @xmath9 is increased as @xmath10 increases . as explained with eq .
( [ ratiobaryonssm ] ) , we see from fig .
[ fig3 ] that @xmath6 is suppressed by factor @xmath185 compared with @xmath182 because of the different wash - out factors presented in eq .
( [ k - em ] ) .
we also see from fig .
[ fig3 ] that a successful baryogenesis can be achieved via flavor dependent leptogenesis in both the sm and the ssm , whereas a successful baryogenesis via flavor independent leptogenesis is possible only for the ssm with large @xmath246 . for this case
, we can obtain lower limit of @xmath10 from the current observation for @xmath153 based on eqs .
( [ cmatrix],[h],[deltam],[rgparameter ] ) as follows ; @xmath247\big)^{1/4}~. \end{aligned}\ ] ] imposing the best - fit value of observed baryon asymmetry , @xmath9 is determined to be @xmath248 for the sm and @xmath249 and @xmath250 for the ssm with @xmath245 and @xmath251 , respectively . for the case
@xmath252 , the dominant contribution to @xmath179 comes from the @xmath7-leptogenesis @xmath253 , as shown in eq .
( [ baryonasymssm ] ) . in the case of the ssm , the value of @xmath9 corresponding to @xmath254 becomes lower as @xmath10 gets higher because @xmath171 is proportional to @xmath9 . from eqs .
( [ flalepto1],[baryonasymssm],[etabexp ] ) , we can roughly get the value of @xmath10 corresponding to @xmath254 as follows : @xmath255 - 1\big)^{1/2}~. \label{tanbwk } \end{aligned}\ ] ] in fig .
[ fig4]-(a ) , we present how @xmath10 is correlated with the mixing angle @xmath9 , which reflects the above equation ( [ tanbwk ] ) .
[ fig4]-(b ) shows how the baryon asymmetry @xmath182 is sensitive to the cp violating observable @xmath61 given by eq .
( [ cp1 ] ) .
the triangles , daggers , asters and circles correspond to the cases of the sm and the ssm with @xmath256 , respectively . the value of @xmath182 corresponding to @xmath257 for the case of the sm leads to @xmath258 ( and @xmath259 ) , which can be measured in the future long - baseline neutrino oscillation experiments . from fig .
[ fig4]-(b ) and fig .
[ fig2]-(b ) , we see that the value of @xmath61 corresponding to @xmath257 for the case of the ssm becomes higher as @xmath10 gets smaller .
thus , if future experiments can measure @xmath61 with a sensitivity better than 0.01 , the predictions of @xmath182 in the ssm with low @xmath260 could be probed . in the case of strong wash - out for the decay of @xmath261 , @xmath262 ,
it turns out that baryogenesis in the sm can not be successfully realized .
however , as can be seen in eq .
( [ baryonasymst ] ) , flavor dependent leptogenesis in the ssm could successfully be implemented at least for @xmath218 as explained below . in fig .
[ fig5 ] , we show the scatter plots for the model parameters in the ssm with @xmath232 constrained by the experimental data given in eq .
( [ exp bound ] ) : fig .
5-(a ) exhibits how the parameters @xmath233 , @xmath234 , @xmath235 and @xmath263 are correlated with the angle @xmath139 defined in eq .
( [ tan2alpha ] ) , respectively .
5-(b ) represents how @xmath264 depends on the angle @xmath139 .
[ fig6 ] presents the predictions of baryon asymmetry for the allowed parameter regions presented in fig .
[ fig5 ] as a function of ( a ) the solar mixing angle @xmath8 and ( b ) the chooz mixing angle @xmath9 .
the results indicate that in the ssm flavor dependent leptogenesis can successfully work out for @xmath265 , whereas flavor independent leptogenesis does so only for the ssm with @xmath266 . from eqs .
( [ flalepto1],[etab],[etabexp ] ) , we can get the relation , @xmath267 where @xmath268 . the relation given by eq .
( [ tanbst ] ) indicates that there exits a correlation between @xmath10 and @xmath8 for a fixed value of @xmath6 .
in particular , taking @xmath269 , we plot the correlation between @xmath10 and @xmath8 in fig .
[ fig7]-(b ) . from our numerical estimate
, we found that in order for the prediction of @xmath182 in the case of strong wash - out to be consistent with the right amount of baryon asymmetry given by eq .
( [ etabexp ] ) the value of @xmath10 should be @xmath270 thus , future measurements for @xmath10 and @xmath8 could serve as an indirect test for the scenario of baryogenesis considered in this paper . fig .
[ fig7]-(b ) represents the correlation between @xmath61 and @xmath8 for a fixed value of @xmath271 .
the results show that the best fit value of @xmath6 and the current measurement of @xmath8 favors @xmath272 which can be measurable in the upcoming long baseline neutrino oscillation experiments .
as a summary , we have considered an exact @xmath0 reflection symmetry in neutrino sector realized at the gut scale in the context of the seesaw model .
the exact @xmath0 reflection symmetry has been imposed in the basis where both the charged lepton mass and heavy majorana neutrino mass matrices are real and diagonal .
we have assumed that the two lighter heavy majorana neutrinos are degenerate at the gut scale .
it has been shown that the rg evolution from the gut scale to the seesaw scale gives rise to breaking of the @xmath1 symmetry and a tiny splitting between two degenerate heavy majorana neutrino masses as well as small variations of the cp phases in @xmath2 , which are essential to achieve a successful leptogenesis .
such small rg effects lead to tiny deviations of @xmath3 from the maximal value and the cp phase @xmath4 from @xmath5 imposed at the gut scale due to @xmath1 reflection symmetry . in our scenario , the required amount of the baryon asymmetry @xmath6 could be generated via so - called resonant @xmath7-leptogenesis , in which the wash - out factor concerned with electron flavor plays a crucial role in reproducing a successful leptogenesis .
and we have found that the magnitude of the electron leptogenesis is enhanced by @xmath185 orders due to the flavor effects in the sm , and enhanced further by a factor of @xmath273 for the ssm .
a point deserved to notice is that cp violation responsible for the generation of baryon asymmetry of our universe comes from the breaking of both the two degenerate heavy majorana neutrinos and the cp phases @xmath53 in @xmath2 by rg evolutions , which corresponds to the tiny breakdown of @xmath1 reflection symmetry , and such small rg effects cause leptogenesis to directly be linked with the cp violation measurable through neutrino oscillation as well as neutrino mixing angles @xmath8 and @xmath9 .
we expect that in addition to the reactor and long baseline neutrino experiments for precise measurements of neutrino mixing angles and cp violation , the measurements for the supersymmetric parameter @xmath10 at future collider experiments would serve as an indirect test of our scenario of baryogenesis based on the @xmath1 reflection symmetry .
the radiative behavior of the heavy majorana neutrinos mass matrix @xmath280 is dictated by the following rg equations @xcite : @xmath281~,~~~\text{sm}\nonumber\\ \frac{d\textbf{m}_{r}}{dt } & = & 2[(\textbf{y}_\nu \textbf{y}_\nu^\dagger)\textbf{m}_r+\textbf{m}_r(\textbf{y}_\nu \textbf{y}_\nu^\dagger)^t]~,~~\text{ssm } \label{rg 1 } \end{aligned}\ ] ] where @xmath282 with an arbitrary renormalization scale @xmath283 .
the rg equations for the dirac - yukawa neutrino matrix can be written as @xmath284~,~~~\text{sm}\nonumber\\ \frac{d\textbf{y}_{\nu}}{dt } & = & \textbf{y}_{\nu}[(t-3g^{2}_2-\frac{3}{5}g^{2}_{1})+\textbf{y}^{\dag}_{l}\textbf{y}_{l}+3\textbf{y}^{\dag}_{\nu}\textbf{y}_{\nu}~,~~~~~~~\text{ssm } \label{rg 2 } \end{aligned}\ ] ] where @xmath285 , @xmath286 and @xmath287 are the yukawa matrices for up - type quarks and charged leptons and @xmath288 are the @xmath70 and @xmath289 gauge coupling constants .
let us first reformulate the rg equations ( [ rg 1 ] ) in the basis where @xmath32 is diagonal .
since @xmath32 is symmetric , it can be diagonalized by a unitary matrix @xmath290 , @xmath291 as the structure of the mass matrix @xmath32 changes with the evolution of the scale , the unitary matrix @xmath290 depends on the scale , too .
the rg evolution of the matrix @xmath292 can be written as @xmath293 where @xmath58 is an anti - hermitian matrix @xmath294 due to the unitary of @xmath290 . then , differentiating eq .
( [ rg 3 ] ) , we obtain @xmath295 v\}_{ij}~,~~\text{sm}\nonumber\\ \frac{dm_{i } \delta_{ij}}{dt } & = & a^t_{ij}m_j + m_i a_{ij } + 2\ { v^t[(\textbf{y}_\nu \textbf{y}_{\nu}^\dagger)\textbf{m}_r+\textbf{m}_r(\textbf{y}_\nu \textbf{y}_\nu ^\dagger)^t ] v\}_{ij}~,~\text{ssm}. \label{rg 4}\end{aligned}\ ] ] absorbing the unitary transformation into the dirac - yukawa coupling @xmath296 the real diagonal part of eq .
( [ rg 4 ] ) becomes @xmath297 the off diagonal part of eq .
( [ rg 4 ] ) leads to @xmath298 + i\frac{m_j - m_i}{m_j+m_i}{\rm im}[(y_\nu y_\nu^\dagger)_{ij}]~,~~~~~\text{sm}\nonumber\\ a_{ij}&= & 2\frac{m_j+m_i}{m_j - m_i}{\rm re}[(y_\nu y_\nu^\dagger)_{ij } ] + i2\frac{m_j - m_i}{m_j+m_i}{\rm im}[(y_\nu y_\nu^\dagger)_{ij}]~,~~\text{ssm}. \label{rg 6 } \end{aligned}\ ] ] the rg equations for @xmath2 in the basis where @xmath280 are diagonal is written as @xmath299+a^{t}y_{\nu}~,~~~\text{sm}\nonumber\\ \frac{dy_{\nu}}{dt } & = & y_{\nu}[(t-3g^{2}_2-\frac{3}{5}g^{2}_{1})+\textbf{y}^{\dag}_{l}\textbf{y}_{l}+3y^{\dag}_{\nu}y_{\nu})]+a^{t}y_{\nu}~,~~~~~~\text{ssm}. \label{rg 7 } \end{aligned}\ ] ] the rg equations for the quantity @xmath300 relevant for leptogenesis can be written as @xmath301 we see from eq .
( [ rg 6 ] ) that the real part of @xmath302 is singular when @xmath303 .
the singularity in @xmath304 $ ] can be eliminated with the help of an appropriate rotation between degenerate heavy majorana neutrino states . such a rotation does not change any physics and it is equivalent to absorb the rotation matrix @xmath305 into the dirac - yukawa neutrino matrix @xmath2 , @xmath306 where the matrix @xmath305 , in our case , particularly rotating the 1st and 2rd generations of heavy majorana neutrinos can be parameterized as @xmath307 then , the singularity in the real part of @xmath302 is indeed removed when the rotation angle @xmath139 is taken to be satisfied with the condition , @xmath308=0~,~~\text{for any pair}~i , j ~\text{corresponding to}~m_{i}=m_{j}~ , \label{singularity } \end{aligned}\ ] ] for @xmath309 , which leads to @xmath310 with @xmath311 , we construct a parameter @xmath312 as follows ; @xmath313 where @xmath314 and the components of @xmath312 are given by @xmath315 it is then obvious from eq .
( [ h1 ] ) that @xmath316=0 $ ] and thus the singularity in @xmath317 does not appear .
now , let us consider rg effects which may play an important role in successful leptogenesis .
first , we parameterize the mass splitting of the degenerate heavy majorana neutrinos in terms of a parameter @xmath318 defined by @xmath319 which is governed by the following rg equations derived from eq .
( [ rg 5 ] ) , @xmath320 \simeq \frac{4h_{12}}{\sin2\alpha}~,~~~~\text{sm}\nonumber\\ \frac{d\delta_n}{dt } & = & 4(1-\delta_{n})[\widetilde{h}_{11}-\widetilde{h}_{22 } ] \simeq \frac{8h_{12}}{\sin2\alpha}~,~~~~\text{ssm}. \label{degeneracy1 } \end{aligned}\ ] ] the solutions of the rg equations ( [ degeneracy1 ] ) are approximately given by @xmath321 where we used the parameters defined in eq .
( [ input2 ] ) .
next , we consider rg running of the dirac neutrino yukawa matrix from the gut scale to the seesaw scale , @xmath322 . since the rg evolution produces non - zero off - diagonal entries in @xmath77
, it has to be re - diagonalized by a unitary transformation , @xmath323 , which leads to the rotation , @xmath324 and @xmath325 .
neglecting the corrections proportional to charged-@xmath16 and -@xmath7 yukawa coupling , @xmath326 and @xmath327 , we can obtain the rg improved dirac neutrino yukawa matrix given for the sm by @xmath328 and for the ssm by @xmath329 where @xmath330 , the components of @xmath331 are @xmath332 and the parameter @xmath133 presenting rg corrections is given by @xmath333 note here that the third low of @xmath331 is not changed because @xmath334 does not depend on @xmath335 .
as will be shown later , the parameter @xmath133 is proportional to @xmath336 in the sm and @xmath337 in the ssm and it plays important roles in both leptogenesis and low - energy cp violation . at the seesaw scale , the combination of the yukawa - dirac coupling matrices and the mass matrix of right - handed heavy majorana neutrino , respectively , can be written as @xmath338 here the ( 1,2 ) and ( 2,1)-components of @xmath339 are the quantities radiatively generated by rg running .
considering the structure of @xmath312 in eq .
( [ h1 ] ) , up to non - zero leading contributions in the right side of eq .
( [ rg 8 ] ) , the rg equations of @xmath340 are given for the sm by @xmath341}{dt}&\simeq & -3y^{2}_{\tau}{\rm re}[(\widetilde{y}_{\nu1\tau}\widetilde{y}^\ast_{\nu2\tau})]+{\rm re}[a_{21}](\widetilde{h}_{22}-\widetilde{h}_{11})~,\nonumber\\ \frac{d { \rm im}[\widetilde{h}_{12}]}{dt}&\simeq & -3y^{2}_{\tau}{\rm im}[(\widetilde{y}_{\nu1\tau}\widetilde{y}^\ast_{\nu2\tau})]~ , \end{aligned}\ ] ] and for the ssm by @xmath341}{dt}&\simeq & 2y^{2}_{\tau}{\rm re}[(\widetilde{y}_{\nu1\tau}\widetilde{y}^\ast_{\nu2\tau})]+{\rm re}[a_{21}](\widetilde{h}_{22}-\widetilde{h}_{11})~,\nonumber\\ \frac{d { \rm im}[\widetilde{h}_{12}]}{dt}&\simeq & 2y^{2}_{\tau}{\rm im}[(\widetilde{y}_{\nu1\tau}\widetilde{y}^\ast_{\nu2\tau})]~. \end{aligned}\ ] ] using eqs .
( [ rg 6],[degeneracy1 ] ) , radiatively generated @xmath340 is given approximately in terms of the parameters in eqs .
( [ h1 ] ) and ( [ h1 ] ) , for the sm , by @xmath342 & = & { \rm re}[\widetilde{h}_{21}]\simeq -\frac{3}{2}y_\tau^2b_3
^ 2\big [ \frac{1}{2}(\kappa^2 -\omega^2)\sin2\alpha+\kappa\omega\cos 2\alpha\cos\delta\phi_{12}\big ] \cdot t~,\nonumber\\ { \rm im}[\widetilde{h}_{12 } ] & = & -{\rm im}[\widetilde{h}_{21}]\simeq 3y_\tau^2b_3 ^ 2\omega\kappa\sin\delta\phi_{12}\cdot t~ , \label{imreh ' } \end{aligned}\ ] ] and for the ssm by @xmath342 & = & { \rm re}[\widetilde{h}_{21}]\simeq y_\tau^2b_3 ^ 2\big [ \frac{1}{2}(\kappa^2 -\omega^2)\sin2\alpha+\kappa\omega\cos 2\alpha\cos\delta\phi_{12}\big ]
\cdot t~,\nonumber\\ { \rm im}[\widetilde{h}_{12 } ] & = & -{\rm im}[\widetilde{h}_{21}]\simeq -2y_\tau^2b_3 ^ 2\omega\kappa\sin\delta\phi_{12}\cdot t~. \label{imreh'm } \end{aligned}\ ] ] note here that radiatively generated quantity @xmath343 $ ] is proportional to @xmath344 which is fixed by the angle @xmath139 in eq .
( [ tan2alpha ] ) due to the @xmath1 reflection symmetry , and related with low energy observables @xmath9 and @xmath61 in eqs .
( [ theta13],[cp1 ] ) .
99 # 1#2#3phys .
b * # 1 * , ( # 3 ) # 2 # 1#2#3nucl .
* b#1 * , ( # 3 ) # 2 # 1#2#3phys .
* d#1 * , ( # 3 ) # 2 # 1#2#3phys .
* # 1 * , ( # 3 ) # 2 # 1#2#3mod .
* a#1 * , ( # 3 ) # 2 # 1#2#3phys
. rep . *
# 1 * , ( # 3 ) # 2 # 1#2#3science * # 1 * , ( # 3 ) # 2 # 1#2#3astrophys .
j. * # 1 * , ( # 3 ) # 2 # 1#2#3eur .
j. * c#1 * , ( # 3 ) # 2 # 1#2#3jhep * # 1 * , ( # 3 ) # 2 # 1#2#3j .
* g#1 * , ( # 3 ) # 2 # 1#2#3int . j. mod .
phys . * # 1 * , ( # 3 ) # 2 # 1#2#3prog .
* # 1 * , ( # 3 ) # 2 y. fukuda _ et al .
_ [ super - kamiokande collaboration ] , .
m. fukugita and t. yanagida , phys .
b * 174 * , ( 1986 ) 45 ; g. f. giudice _
et al . _ ,
b * 685 * ( 2004 ) 89 [ arxiv : hep - ph/0310123 ] ; w. buchmuller , p. di bari and m. plumacher , annals phys . * 315 * , ( 2005 ) 305 [ arxiv : hep - ph/0401240 ] ; a. pilaftsis and t. e. j. underwood , phys .
d * 72 * , ( 2005 ) 113001 [ arxiv : hep - ph/0506107 ] .
p. minkowski , phys .
b * 67 * , ( 1977 ) 421 ; m. gell - mann , p. ramond and r. slansky , _ proceedings of the supergravity stony brook workshop _
, new york 1979 , eds .
p. van nieuwenhuizen and d. freedman ; t. yanagida , _ proceedinds of the workshop on unified theories and baryon number in the universe _ ,
tsukuba , japan 1979 , ed.s a. sawada and a. sugamoto ; r. n. mohapatra , g. senjanovic , .
i. aizawa , t. kitabayashi and m. yasue , phys .
d * 72 * , ( 2005 )
055014 [ arxiv : hep - ph/0504172 ] ; i. aizawa , t. kitabayashi and m. yasue , nucl .
b * 728 * , ( 2005 ) 220 [ arxiv : hep - ph/0507332 ] ; z. z. xing , h. zhang and s. zhou , lett .
b * 641 * , ( 2006 ) 189 [ arxiv : hep - ph/0607091 ] .
a. de rujula , m. b. gavela and p. hernandez , nucl .
b * 547 * , ( 1999 ) 21 [ arxiv : hep - ph/9811390 ] ; a. cervera , a. donini , m. b. gavela , j. j. gomez cadenas , p. hernandez , o. mena and s. rigolin , nucl . phys .
b * 579 * , ( 2000 ) 17 [ erratum - ibid .
b * 593 * , ( 2001 ) 731 ] [ arxiv : hep - ph/0002108 ] ; c. albright _ et al .
_ , arxiv : hep - ex/0008064 ; k. dick , m. freund , m. lindner and a. romanino , nucl . phys .
b * 562 * , ( 1999 ) 29 [ arxiv : hep - ph/9903308 ] ; v. d. barger , s. geer , r. raja and k. whisnant , phys .
d * 63 * , ( 2001 ) 033002 [ arxiv : hep - ph/0007181 ] ; m. freund , p. huber and m. lindner , nucl .
b * 585 * , ( 2000 ) 105 [ arxiv : hep - ph/0004085 ] ; k. hagiwara , n. okamura and k. i. senda , arxiv : hep - ph/0607255 .
p. h. chankowski and z. pluciennik , phys .
b * 316 * , ( 1993 ) 312 [ arxiv : hep - ph/9306333 ] ; k. s. babu , c. n. leung and j. t. pantaleone , phys .
b * 319 * , ( 1993 ) 191 [ arxiv : hep - ph/9309223 ] ; m. tanimoto , phys .
b * 360 * , ( 1995 ) 41 [ arxiv : hep - ph/9508247 ] ; n. haba and n. okamura , eur .
j. c * 14 * , ( 2000 ) 347 [ arxiv : hep - ph/9906481 ] ; p. h. chankowski and s. pokorski , int .
j. mod .
a * 17 * , ( 2002 ) 575 [ arxiv : hep - ph/0110249 ] .
j. a. casas , j. r. espinosa , a. ibarra and i. navarro , nucl .
b * 573 * , ( 2000 ) 652 [ arxiv : hep - ph/9910420 ] ; j. a. casas , j. r. espinosa , a. ibarra and i. navarro , nucl .
b * 569 * , ( 2000 ) 82 [ arxiv : hep - ph/9905381 ] .
s. antusch , j. kersten , m. lindner and m. ratz , phys .
b * 538 * , ( 2002 ) 87 [ arxiv : hep - ph/0203233 ] .
a. pilaftsis , phys .
d * 56 * , ( 1997 ) 5431 [ arxiv : hep - ph/9707235 ] ; a. pilaftsis and t. e. j. underwood , nucl . phys .
b * 692 * , ( 2004 ) 303 [ arxiv : hep - ph/0309342 ] .
l. covi , e. roulet and f. vissani , phys . lett . *
b384 * , ( 1996 ) 169 ; a. pilaftsis , int .
j. mod .
a * 14 * , ( 1999 ) 1811 [ arxiv : hep - ph/9812256 ] .
t. endoh , t. morozumi , z. xiong , prog .
phys . * 111 * , ( 2004 ) 123 [ arxive : hep - ph/0308276 ] ; t. fujihara , s. kaneko , s. k. kang , d. kimura , t. morozumi , m. tanimoto , phys .
rev.d * 72 * , ( 2005 ) 016006 [ arxive : hep - ph/0505076 ] ; a. abada , s. davidson , a. ibarra , f. x. josse - michaux , m. losada and a. riotto , jhep * 0609 * , ( 2006 ) 010 [ arxiv : hep - ph/0605281 ] ; s. blanchet and p. di bari , jcap * 0703 * , ( 2007 ) 018 [ arxiv : hep - ph/0607330 ] ; s. antusch , s. f. king and a. riotto , jcap * 0611 * , ( 2006 ) 011 [ arxiv : hep - ph/0609038 ] ; s. pascoli , s. t. petcov and a. riotto , phys .
d * 75 * , ( 2007 ) 083511 [ arxiv : hep - ph/0609125 ] ; s. pascoli , s. t. petcov and a. riotto , nucl .
b * 774 * , ( 2007 ) 1 [ arxiv : hep - ph/0611338 ] ; g. c. branco , r. gonzalez felipe and f. r. joaquim , phys .
b * 645 * , ( 2007 ) 432 [ arxiv : hep - ph/0609297 ] ; g. c. branco , a. j. buras , s. jager , s. uhlig and a. weiler , jhep * 0709 * , ( 2007 ) 004 [ arxiv : hep - ph/0609067 ] .
y. h. ahn , s. k. kang , c. s. kim and j. lee , phys .
d * 77 * , ( 2008 ) 073009 [ arxiv : hep - ph/0711.1001 ] .
a. abada , s. davidson , f. x. josse - michaux , m. losada and a. riotto , jcap * 0604 * , ( 2006 ) 004 [ arxiv : hep - ph/0601083 ] ; s. antusch , s. f. king and a. riotto , jcap * 0611 * , ( 2006 )
011 [ arxiv : hep - ph/0609038 ] .
a. pilaftsis , phys .
lett . * 95 * , ( 2005 ) 081602 [ arxiv : hep - ph/0408103 ] ; a. pilaftsis and t. e. j. underwood , phys . rev .
d * 72 * , ( 2005 ) 113001 [ arxiv : hep - ph/0506107 ] .
wmap collaboration , d.n .
_ , astrophys
. j. suppl . *
148 * , ( 2003 ) 175 ; m. tegmark _
et al . _ ,
d * 69 * , ( 2004 ) 103501 ; c. l. bennett _ et al .
_ , astrophys .
j. suppl .
* 148 * , ( 2003 ) 1 [ arxiv : astro - ph/0302207 ] . | we consider an exact @xmath0 reflection symmetry in neutrino sector realized at the gut scale in the context of the seesaw model with and without supersymmetry . assuming the two lighter heavy majorana neutrinos are degenerate at the gut scale , it is shown that the renormalization group ( rg ) evolution from the gut scale to the seesaw scale gives rise to breaking of the @xmath1 symmetry and a tiny splitting between two degenerate heavy majorana neutrino masses as well as small variations of the cp phases in @xmath2 , which are essential to achieve a successful leptogenesis .
such small rg effects lead to tiny deviations of @xmath3 from the maximal value and the cp phase @xmath4 from @xmath5 imposed at the gut scale due to @xmath1 reflection symmetry . in our scenario , the required amount of the baryon asymmetry @xmath6 can be generated via so - called resonant @xmath7-leptogenesis , in which the wash - out factor concerned with electron flavor plays a crucial role in reproducing a successful leptogenesis .
we show that cp violation responsible for the generation of baryon asymmetry of our universe can be directly linked with cp violation measurable through neutrino oscillation as well as neutrino mixing angles @xmath8 and @xmath9 .
we expect that , in addition to the reactor and long baseline neutrino experiments , the measurements for the supersymmetric parameter @xmath10 at future collider experiments would serve as an indirect test of our scenario of baryogenesis based on the @xmath1 reflection symmetry . + pacs : 14.60.pq , 11.30.fs , 11.10.hi , 98.80.cq , 13.35.hb |
the one - dimensional ( 1d ) schrdinger equation including a symmetric potential structure produces single - particle wave functions of a definite parity , even or odd , with the ground state always corresponding to an even function without zeros .
however , a spatially symmetric hamiltonian of an interacting many - particle system can give rise to asymmetric states , which may be considered as a spontaneous - symmetry - breaking ( ssb ) effect . at the classical level
, the ssb occurs in optics , as a result of the interplay between the nonlinearity and waveguiding structures , when the strong nonlinearity partly suppresses the linear coupling between parallel guiding cores .
in particular , it was shown that a stable trapped mode may be asymmetric in a channel waveguide embedded in the self - focusing kerr medium @xcite .
the onset of a sharp symmetry - breaking instability in a double - hump two - component spatial optical soliton was demonstrated experimentally in a planar nonlinear waveguide @xcite . a natural setting in which ssb phenomena may arise in the context of nonlinear optics and bose - einstein condensation ( bec )
is provided by double - well potentials ( dwps ) . in the experiment ,
an effective optical dwp was created by a specially designed illumination pattern applied , in the ordinary polarization , to a photorefractive crystal ( the ssb was observed in a beam with extraordinary polarization , shone through this structure ) zhigang .
it was also proposed to realize similar effective potentials in coupled nonlinear microcavities @xcite , and in a structured core of an optical fiber @xcite .
a specific variety of the optical ssb was studied in a model of two parallel - coupled antiwaveguides with the self - focusing nonlinearity , which corresponds to an effective double - barrier _ _ _ _ potential , rather than dwp @xcite .
well - known dual - core optical fibers @xcite , which may serve as a basis for the power - controlled all - optical switching , if the kerr nonlinearity is taken into regard @xcite , may also be considered as dwp structures , with the difference that the tunneling between two potential wells is replaced by the linear coupling between the cores .
in addition to the ssb of continuous - wave states @xcite , the formation of asymmetric solitons in dual - core fibers was studied in detail theoretically @xcite .
similar analysis of the ssb for soliton modes was performed in models of dual - core fiber bragg gratings with the kerr nonlinearity @xcite , and coupled waveguides with the quadratic @xcite and cubic - quintic @xcite nonlinear terms , including a system of linearly coupled complex ginzburg - landau equations of the cubic - quintic type @xcite . in refs .
@xcite , @xcite and @xcite , the analysis of the ssb was extended to three - core linearly coupled triangular configurations for optical fibers , bragg gratings , and complex ginzburg - landau equations , respectively .
the concept of ssb also plays an important role in understanding experimental phenomena in bec , because , if interactions between atoms are strong enough , the ground state of the condensate may not follow the symmetry of the trapping potential @xcite . in particular , manifestations of ssb were observed in a quenched ferromagnetic state of a spinor ( three - component ) condensate @xcite . in the single - component bec ,
a natural setting for the realization of ssb may again be provided by dwp configurations .
an effectively one - dimensional dwp structure was realized experimentally in ref .
@xcite . loading
a condensate of @xmath0rb atoms with the repulsive interaction between them into this structure made it possible to observe josephson oscillations for a small number of atoms , and the macroscopic quantum self - trapping featuring an imbalance between populations of the two wells , for a larger number .
parallel to the experimental work , numerous theoretical studies of matter - wave dwp settings have been performed , for the cases of both repulsion and attraction between atoms .
these studies addressed problems such as finite - mode reductions finite - mode ( including two - component mixtures @xcite ) , obtaining analytical results for specific shapes of the potential @xcite , quantum effects @xcite , and some others .
recently investigated tunneling between vortex and antivortex states in bec trapped in a 2d anisotropic potential @xcite belongs to this category too .
theoretical analysis was also performed for 2d and 3d extensions of the dwp settings in bec , which add one or two extra dimensions to the model , either without an additional potential , or with a periodic optical - lattice ( ol ) potential acting in these directions .
these settings may be approximated , similar to the above - mentioned standard model of dual - core optical fibers , by a system of linearly coupled 1d @xcite or 2d @xcite equations . in a more accurate form , nearly-1d solitons can be found as solutions to the full 2d equation that includes the dwp ( the potential depends on the transverse coordinate , @xmath1 , allowing solitons to self - trap in the free longitudinal direction , @xmath2 ) @xcite .
the latter model is relevant to the case of the self - attractive nonlinearity . in the case of self - repulsion
, dual - core gap solitons have been be predicted in the setting with the ol potential applied along direction @xmath2 @xcite .
note that , in any setting , gap solitons can not realize the ground state of the respective system , but , nevertheless , they represent stable configurations , that have been created in the experiment using the condensate of @xmath0rb atoms with the repulsion between them @xcite .
a general principle , upheld by the analysis in various settings ( in nonlinear optics and bec alike ) , is that the ssb occurs through bifurcations of symmetric or antisymmetric states , in the models with the self - attraction and self - repulsion , respectively .
as mentioned above , models of the dwp / double - core type , combining cubic attractive and quintic repulsive nonlinearities , were studied too @xcite . in the latter case ,
the competition between the self - focusing and self - defocusing against the backdrop of the dwp structure gives rise to specific ssb bifurcation diagrams , in the form of non - convex closed loops @xcite , as well as to specific dynamical switching regimes @xcite . also predicted were manifestations of the ssb in a two - component bec mixture trapped in the dwp structure @xcite , for both cases of the self - attraction and self - repulsion .
all the extensive work on the ssb outlined above was performed in settings based on usual linear potentials of the double - well type .
the objective of the present work is to propose another physical framework , in which the ssb can be predicted in an effective _ nonlinear _ double - well potential ( ndwp ) , induced through a spatial modulation of the local nonlinearity coefficient . following the terminology commonly adopted in the solid - state theory @xcite
, this nonlinear ingredient of the physical model may also be a called a _
pseudopotential_. in bec settings , a pseudopotential structure may be readily induced through spatial modulation of the local value of the @xmath3-wave scattering length , @xmath4 , which determines the effective bec nonlinearity .
the modulation can be implemented , through the feshbach resonance , by means of a spatially inhomogeneous dc magnetic field @xcite , or by a resonant optical field , as predicted in ref .
@xcite and demonstrated experimentally in ref .
it was also proposed to control the feshbach resonance by dint of dc electric field fr - electric , which can be easily made inhomogeneous too .
the attractive and repulsive interactions between atoms correspond to @xmath5 and @xmath6 , respectively ; both signs , as well as sign - changing patterns , can be used to engineer effective nonlinear potentials .
so designed pseudopotential lattices have attracted much interest in studies of bec . in the 1d geometry , solitons , extended wave patterns , and various dynamical states supported by such structures were studied theoretically @xcite ( a random nonlinear lattice @xcite and pseudopotentials generated by a spatially monotonous ramp of the local scattering length @xcite were explored too ) .
recently , similar states were also considered in nonlinear optics , assuming a periodic modulation of the local kerr coefficient @xcite .
some ( but much fewer ) results were obtained too for 2d settings @xcite .
however , to the best of our knowledge , ssb phenomena in nonlinear pseudopotentials have not been studied yet . in this work ,
we focus on such effects in ndwp settings , which can be engineered by means of techniques mentioned above , using attractive interactions between atoms in bec , or the self - focusing nonlinearity in optics , as briefly described below . in section
ii , we formulate the model and give estimates of characteristic values of related physical parameters .
section iii reports full analytical solutions corresponding to symmetric , antisymmetric , and _ asymmetric _ states trapped by the ndwp , in the limit case with the modulation of the local nonlinearity coefficient is represented by a set of two dirac s delta - functions .
the relevance of the latter model is stressed , in particular , by the recently introduced @xcite bec model with a periodic nonlinear potential of the kronig - penney type , whose simplest version reduces to a periodic array of delta - functions ( ssb effects were not studied in ref .
@xcite ) . in section iv ,
we present numerical results for the general model , in which the delta - functions are replaced by a pair of gaussians of a finite width . in that case , the trapped states are found in a numerical form , and their stability is studied by means of direct simulations of slightly perturbed stationary states and also , independently , through computation of respective stability eigenvalues for small perturbations .
the result is that asymmetric states , which are unstable in the delta - function limit , can be readily stabilized in the general model .
in addition , antisymmetric states may be stabilized too , in two disjoint regions of the parameter state , giving rise to a _ bistability _ involving antisymmetric and asymmetric states .
a condition for the existence of symmetric states is that their norm must exceed a certain threshold value , in terms of their norm , an explanation to which is given .
the existence and stability of all states , including a line of the ssb bifurcation , are summarized in a single diagram , which is presented in section iv too .
results reported in this paper and perspectives for further work are summarized in section v.
the underlying 3d gross - pitaevskii equation for the mean - field wave function , @xmath7 , is taken in the ordinary form corresponding to the nearly-1d trap @xcite:@xmath8with @xmath9 the atomic mass and @xmath10 the frequency providing for the tight confinement of the condensate in the direction of @xmath11 .
as said above , we assume an axial modulation ( and negative sign ) of the scattering length , in the form of a pair of gaussians , each of width @xmath12 and amplitude @xmath13 , with centers set at points @xmath14 : @xmath15 .
\label{gauss}\]]in what follows below , we measure the axial coordinate in units of @xmath16 , and , accordingly , time in units of @xmath17 , i.e. , we define @xmath18further , following the usual approach to the derivation of the effective 1d equation , we take 3d field as a product of a slowly varying axial function , @xmath19 , and the ground - state wave function in the transverse plane , @xmath20with @xmath21 [ the scaling factor here is chosen so as to maintain normalization condition ( scaling ) , see below ] . the substitution of expression ( [ psi ] ) in eq .
( [ gpe ] ) and averaging in the transverse plane lead to the following 1d equation:@xmath22@xmath23 , \label{g}\]]where the modulated nonlinearity coefficient is subject to the following normalization condition : @xmath24profiles of modulation function ( [ g ] ) , which keeps the double - well structure for @xmath25 , are shown , for different values of @xmath26 , in fig .
[ fig1 ] . ) , normalized to its maximum value , are shown for different values of scaled width @xmath26 of the individual well.,width=336 ] equation ( [ psi2 ] ) conserves two dynamical invariants , _
viz_. , the norm and energy ( hamiltonian ) , @xmath27as follows from eqs .
( [ psi ] ) and ( [ a ] ) , the total number of atoms in the condensate , @xmath28 , is related to 1d norm @xmath29
as follows:@xmath30for physical parameters relevant to experiments with the condensate of @xmath31li atoms @xcite , i.e. , @xmath32 m , @xmath33 nm , and @xmath34 @xmath35 m , characteristic values of the number of atoms in various patterns reported below [ see , in particular , figs .
[ fig4](d ) and [ fig8 ] ] fall into the range of @xmath28 between @xmath36 and @xmath37 , which is quite sufficient for experimental manipulations and observation of the patterns . in the same range of physical parameters , @xmath38 in eq .
( [ psi2 ] ) , is estimated as being tantamount to @xmath39 ms , hence typical time scales for the instability development or intrinsic oscillations of breathers induced by the instabilities , that are reported below , are expected to be in the range of @xmath40 to @xmath41 s , which is realistic to the currently available experimental techniques @xcite . in terms of optical settings ,
a set of two narrow ( of width @xmath42 @xmath35 m ) parallel stripes with strong local nonlinearity can be built , in a planar waveguide , by means of available nanotechnological methods . in that case
, the power of the laser beam necessary for the self - trapping of transverse nonlinear patterns in the waveguide made of silica may be @xmath43 kw , @xcite while , using algaas , one may reduce the necessary power to the level of @xmath41 kw @xcite . in these settings , the characteristic evolution length of the spatial beam
can be made shorter than @xmath41 mm .
obviously , transitions between states of different types reported in this paper may be relevant to the design of power - controlled optical - switching schemes .
on the other hand , the description of the planar waveguide with the pair of embedded stripes may require a model more general than the one studied here , as it will plausibly combine the transverse modulation of the local nonlinearity with a similar linear potential ( which is briefly described at the end of the next section ) , as the material difference between the stripes and host medium ought to affect the linear index of refraction too .
stationary localized solutions to eq .
( [ psi2 ] ) are sought for as @xmath44 , where the chemical potential is negative , @xmath45 , and real function @xmath46 satisfies equation@xmath47 \phi ^{3}=0 .
\label{phi2}\ ] ] an analytically solvable version of the model corresponds to the limit of @xmath48 , with eq .
( [ psi2 ] ) going over into@xmath49 |\psi delta - function .
note that rescaled equation ( [ psi ] ) contains no free parameters ; however , the norm of the solution will play the role of an intrinsic parameter , see below . in the same limit , stationary equation ( [ phi2 ] ) takes the form of@xmath50
\phi ^{3}=0 .
\label{phi}\ ] ]
off points @xmath51 , eq .
( [ phi ] ) is linear .
a general solution to this equation , decaying at @xmath52 , can be written as@xmath53with constant amplitudes @xmath54 and @xmath55 .
the continuity of the wave function at @xmath51 imposes two relations on them , @xmath56 which allows one to eliminate @xmath13 and @xmath57 in favor of @xmath58 and @xmath59 , @xmath60further , the integration of eq .
( [ phi ] ) in infinitesimal vicinities of points @xmath51 yields expressions for jumps ( @xmath61 ) of the first derivative at these points , @xmath62 .
the substitution of solution ( linear ) in these relations leads to a system of cubic equations for the amplitudes,@xmath63after the substitution of expressions ( [ a0b0 ] ) into eqs .
( [ cubicb ] ) and ( [ cubica ] ) , we end up with two coupled cubic equations for @xmath58 and @xmath64:@xmath65@xmath66 solving eqs .
( [ a ] ) and ( [ b ] ) , we first find symmetric and antisymmetric solutions , @xmath67@xmath68the norm of these solutions , defined as per eq .
( [ n ] ) , is @xmath69 with @xmath70 and @xmath71 corresponding to the symmetric and antisymmetric states , respectively . in the limit of @xmath72 both expressions ( [ symm ] ) and ( [ antisymm ] ) yield @xmath73 . in this limit ,
norm ( [ + - ] ) of both solutions takes a common value , @xmath74on the other hand , in the limit of @xmath75 , the amplitude of the symmetric state vanishes , @xmath76 , while the amplitude of the antisymmetric state remains finite , @xmath77 , and its norm diverges , @xmath78 .
it can be easily checked that the decrease of @xmath79 and increase of @xmath80 with the variation of @xmath35 from @xmath81 to @xmath82 are not monotonous : @xmath83 attains a maximum , @xmath84 , at @xmath85 , and @xmath80 has a minimum , @xmath86 , at @xmath87 . actually , the finite minimum norm ( _ threshold _ ) , necessary for the existence of the symmetric states , which is @xmath88 in the present case , is a generic property , shared by the model with finite width @xmath26 of the nonlinear - potential wells , as shown in detail below in fig .
[ fig8 ] . on the other hand ,
the existence of the above - mentioned maximum value of the norm for the symmetric states , @xmath89 , is a specific feature of the model with @xmath90 ( the one based on the delta - functions ) : it is easy to see that , in the model with finite @xmath26 , the norm of any state grows @xmath91 at @xmath92 . however , this specific feature , which , as a matter of fact , indicates degeneracy of the model with @xmath90 , is less significant for the physical applications , as all symmetric states , for @xmath90 and finite @xmath26 alike , are unstable when the norm exceeds its value at the ssb bifurcation point [ further details can be seen in eq .
( [ nbif ] ) and fig .
[ fig8 ] below ] . a point of the ssb bifurcation , which gives rise to a pair of asymmetric solutions splitting off from the symmetric one , can be easily found .
indeed , the symmetry breaking means that symmetric solution ( [ symm ] ) acquires an infinitesimal _ antisymmetric _ addition , @xmath93 .
thus , infinitely close to the bifurcation point , the relevant solution is sought for as @xmath94 .
the substitution of this in eqs . ( a ) and ( [ b ] ) and linearization in infinitesimal @xmath95 lead to a simple equation that predicts the value of the chemical potential at the bifurcation point : @xmath96 , or@xmath97at this point
, the amplitude of symmetric solution ( [ symm ] ) is @xmath98 , and the value of norm ( [ n ] ) , with the upper sign , is @xmath99actually , value ( [ bif ] ) of the chemical potential at the bifurcation point ins approximately the same in the model with finite - width nonlinear - potential wells , up to @xmath100 , as seen from fig . [
fig4](d ) presented below .
the same analysis shows that antisymmetric solution ( [ antisymm ] ) never gives rise to an antisymmetry - breaking bifurcation .
indeed , for this solution the antisymmetry would be broken by an infinitesimal symmetric variation , @xmath101 , i.e. , infinitely close to the bifurcation point , the solution would be @xmath102 .
subsequent substitution in eqs .
( [ a ] ) and ( [ b ] ) and the linearization in @xmath103 yield an equation for @xmath35 that has no real solutions . equations ( [ a ] ) and ( [ b ] ) can be solved analytically for asymmetric states too:@xmath104note that full solution ( [ asymm ] ) predicts exactly the same bifurcation point as eq .
( [ bif ] ) , i.e. , @xmath105 [ at this point , the second radical in eq .
( [ asymm ] ) vanishes ] .
these solutions are characterized by the _ asymmetry ratio _ , which is defined as @xmath106typical examples of symmetric , asymmetric and antisymmetric states produced by he above analytical solutions are displayed in fig .
[ fig2 ] . .
the respective values of the norm are @xmath107 , @xmath108 , and @xmath109 .
the mutually symmetric lines ( blue and red ones , in the color version of the figure ) represent two asymmetric states that are mirror images of each other.,width=336 ] the analytical solution given by eqs .
( [ linear ] ) , ( [ a0b0 ] ) and ( asymm ) make it possible to plot the bifurcation diagrams in the planes of @xmath110 and @xmath111 , which are represented by curves pertaining to @xmath90 in figs .
[ fig4](a , b , c ) . to generate the diagrams , partial norms @xmath112 in expression ( [ theta ] ) for the asymmetric solutions were computed numerically [ analytical expressions for them are available , but they are very messy , cf .
( + - ) for the symmetric and antisymmetric states ] . a salient peculiarity of the ssb bifurcation for @xmath90 , evident in fig .
fig4(c ) , is its _ subcritical _ character , which means that the branches of asymmetric solutions emerge at the bifurcation point as unstable ones , and go in the _ backward _ direction .
a subcritical bifurcation also occurs in the above - mentioned model of the dual - core nonlinear fibers @xcite , but in that case the asymmetric branches quickly turn in the forward direction , getting stabilized at the turning point .
a remarkable feature of the present model with @xmath90 is that this does not happen , i.e. , the bifurcation in this model may be called a fully backward " one : the branches of the asymmetric solutions keep going backward up to the limit of @xmath113 , which corresponds to the asymmetric solutions with @xmath114 @xmath115 ( and @xmath116 , as shown in the following subsection ) .
indeed , @xmath117 follows from the fact that the amplitude appertaining to the lower sign between the radicals in eq .
( [ asymm ] ) vanishes in the limit of @xmath118 . in accordance with general properties of the subcritical ssb bifurcation @xcite
, the symmetric solution is expected to be stable below the bifurcation point [ at @xmath119 , see eq .
( [ nbif ] ) ] , and unstable above it .
the asymmetric branches emerging at @xmath120 are unstable as long as they go backward . in the present case ( @xmath90 ) , this means they are always unstable , as the respective branches in figs .
fig4(b , c ) never turn forward .
all these expectations are completely borne out by the stability analysis performed by means of both direct simulations and computation of stability eigenvalues , at finite but small values of @xmath26 ( technical details of the procedure are described in the next section ) . in particular , at @xmath121 the unstable symmetric state spontaneously transforms into a strongly asymmetric breather which features irregular oscillations , but remains robust as a whole ( quite similar to an example displayed below in fig .
[ newfig3 ] for @xmath122 ) . on the other hand ,
unstable asymmetric states transform themselves into breathers which maintain the original asymmetry of the unstable state , as shown in fig .
fig3 . and @xmath123.,width=384 ]
lastly , all antisymmetric states in the model with delta - functions are unstable too .
their instability is similar to that shown below in fig .
fig6(b ) for @xmath124 , transforming them into strongly asymmetric breathers .
as shown in the next section , both asymmetric and antisymmetric states may be stabilized in the general model , with finite @xmath26 . the existence of the asymmetric states , i.e.
, the presence of the ssb effect in the present model , can be easily explained by the consideration of the above - mentioned limit of @xmath118 . indeed , in this limit , the spatial scale of the solution , which is @xmath125 according to eq .
( [ linear ] ) , is much smaller than the separation between the two delta - functions , @xmath126 .
therefore , the full solution effectively splits into a superposition of those independently supported by each delta - function in isolation .
further , it is obvious that , for given large @xmath127 , eq .
( [ phi ] ) with an individual delta - function gives rise to two solutions : a trivial one , @xmath128 , and@xmath129where @xmath130 or @xmath131 ; note that the norm of solution ( [ xi ] ) is @xmath116 , for any @xmath35 .
the corresponding symmetric and antisymmetric states are built , respectively , as superpositions of solutions @xmath132 ( or , equivalently , @xmath133 ) centered at @xmath134 and @xmath135 , or @xmath133 centered at @xmath134 and @xmath132 centered at @xmath135 .
asymmetric solutions are represented , in the same limit , by a superposition of solution @xmath136 centered at @xmath134 and zero solution around @xmath135 , or vice versa .
of course , finding the bifurcation point requires one to perform the analysis of the model at finite @xmath35 , as done in the analytical form above for the case of the delta - functions , and will be done in a numerical form below for the general case of finite @xmath26 in eq .
( [ g ] ) .
it is relevant to compare the above exact results with those which can be easily obtained in the linear counterpart of the model , i.e. , the one with the dwp based on the set of two delta - functions ; as mentioned above , such a linear potential may be a plausible ingredient of a more general model , relevant to the description of ndwp settings in optics .
the stationary version of the linear equation reduces to@xmath137 \phi = 0 , \label{linear - delta}\]]with constant @xmath138 .
symmetric solutions , which must be continuous at @xmath139 , are sought for as @xmath140cf .
( [ linear ] ) .
the jump condition , @xmath141 , yields equation @xmath142 = 2\epsilon $ ] , which has a single solution for @xmath143and the respective jump condition takes the form of @xmath144 = 2\epsilon $ ] .
the latter equation has no solutions for @xmath145 , and exactly one solution for @xmath146 .
thus , symmetric solution ( [ symmetric ] ) is the single state in the linear dwp model at @xmath147 , while at @xmath146 the linear model supports precisely two states , symmetric and antisymmetric ones .
lastly , the combined model , which includes both the linear potential and its nonlinear counterpart ( that may be self - attractive , as above , or__self - repulsive _ _ too , in this case ) , is also solvable in the case when these features are based on the pair of delta - functions .
the combined model is described by the following stationary equation , cf .
( [ phi ] ) and ( [ linear - delta]):@xmath50 \left ( \epsilon \phi + \sigma \phi ^{3}\right ) = 0 , \label{combi}\]]where @xmath148 and @xmath149 corresponds to the nonlinear attraction and repulsion , respectively . in particular
, the bifurcation occurs only on the branch of the symmetric solutions in the case of @xmath148 , and only on the antisymmetric branch in the opposite case ( self - repulsion ) . in either case
, the value of the chemical potential ( @xmath45 ) at the symmetry- or antisymmetry - breaking bifurcation is determined by the following transcendental equation:@xmath150[for @xmath151 and @xmath148 , it reduces to eq .
( [ bif ] ) ] . with @xmath152 ,
( [ transc ] ) has exactly one solution for any @xmath153 and no solutions for @xmath154 ( negative @xmath155 corresponds to competition between the repulsive linear potential and attractive nonlinear pseudopotential ) . with @xmath156 , eq .
( [ transc ] ) has exactly one solution for any @xmath157 , and no solutions for @xmath158 .
analysis of the stability of states found in the combined model is beyond the scope of this work , and in the rest of the paper we consider the model without the linear potential .
to construct localized solutions to stationary equation ( [ phi2 ] ) of the symmetric , asymmetric and antisymmetric types , the newton iterative method was used , starting with the following inputs:@xmath159@xmath160with @xmath161 for symmetric solutions .
then , as mentioned above , numerical analysis of the stability of the stationary solutions was performed in two different ways : first , by means of direct simulations of the evolution of slightly perturbed solutions , and then through computation of ( in)stability eigenvalues for modes of small perturbations . in the former case
, the stability was tested by adding arbitrary perturbations to the initial state , at the level of @xmath162 of the amplitude ( in particular , care was taken to test effects of perturbations whose symmetry is different from that of the stationary state , such as antisymmetric perturbations added to symmetric states , and vice versa ) . for the computation of eigenvalues ,
perturbed solutions were looked for as @xmath163 \right\ } , \label{pert}\]]where @xmath46 is a stationary solution to eq .
( [ phi2 ] ) with chemical potential @xmath35 , while @xmath164 and @xmath165 are components of a perturbation mode with an infinitesimal amplitude @xmath166 , pertaining to instability growth rate @xmath167 .
the substitution of expression ( [ pert ] ) into eq .
( [ psi2 ] ) and linearization lead to the eigenvalue problem based on the following equations : @xmath168 the underlying solution , @xmath46 , is stable if all eigenvalues associated with it have @xmath169 .
equations ( [ eigen ] ) were solved numerically with the help of a finite - difference method .
conclusions concerning the stability of the patterns , drawn from direct simulations , always complied with results produced by the computation of eigenvalue .
the first significant change against the results reported above for the model with the delta - functions ( @xmath90 ) , which happens with the increase of @xmath170 , is quick stabilization of asymmetric states with larger values of the norm , while ones with smaller @xmath29 remain unstable , originally . at @xmath171 ,
the symmetric states are stable for all values of @xmath29 at which they exist .
another notable feature of the bifurcation diagrams at finite @xmath26 , demonstrated by figs . [ fig4](b ) and [ fig4](d ) , is that the norm at which the ssb bifurcation takes place , @xmath172 , first decreases with the growth of @xmath26 from small values up to @xmath173 , and then increases with the further growth of @xmath26 .
close to the their stabilization threshold ( in particular , at @xmath174 ) , asymmetric states with a smaller norm , which are still unstable , demonstrate a scenario of the instability development different from what was shown by their counterpart in fig .
[ fig3 ] in the case of very small @xmath26 .
namely , slow regular oscillations , observed in fig .
[ fig5 ] in this case , imply a _ dynamical re - symmetrization _ of the unstable asymmetric state .
indeed , densities @xmath175 , taken at points @xmath51 , perform identical periodic oscillations , with a phase shift of @xmath176 between them , as shown in [ fig5](b ) .
the stabilization of the asymmetric states at small finite values of @xmath26 is explained by the change in the character of the ssb bifurcation : at @xmath177 , there appear turning points on branches of asymmetric solutions in the bifurcation diagram , cf .
[ fig4](c ) .
past the turning point , the branch goes forward as a _
stable _ one .
in fact , fig .
[ fig4](c ) demonstrates a quick transformation , with the increase of @xmath26 , of the subcritical bifurcation into a supercritical one . when the bifurcation is supercritical , branches of the asymmetric solutions emerge as stable ones at the bifurcation point , and immediately go forward .
we do not display the quick transition from the sub- to supercritical bifurcation in full detail , as it actually happens at very small @xmath26 , in the range of @xmath178 .
the physical estimates given in section ii suggest that so small values of the scaled width of the nonlinear - potential wells correspond to physical widths @xmath179 @xmath35 m .
it seems doubtful that the feshbach - resonance technique would allow one to create a strong local inhomogeneity of the scattering length on such a small scale ( nevertheless , the exact analytical solutions obtained for @xmath90 , which provide clear clues for the understanding of the general model , are definitely relevant ) .
an additional problem impeding the full analysis of the case of very small @xmath26 is that , in this case , the accumulation of systematic numerical results requires very heavy simulations , as the stepsize of the spatial grid must be made much smaller than @xmath26 . above the bifurcation point ,
symmetric states found at finite @xmath26 demonstrate the familiar ssb instability , spontaneously transforming themselves into slightly nonstationary robust modes ( breathers ) , quite close in their shape to respective stable asymmetric solitons .
a typical example of this transformation in displayed in fig .
[ newfig3 ] . , @xmath180 and @xmath181.,width=384 ] as concerns antisymmetric solutions ,
both stable and unstable ones have been found at finite @xmath26 , as illustrated by figs .
[ fig6 ] and [ newfig2 ] .
panel ( b ) of the former figure shows that the density profile of unstable antisymmetric states evolves from the double - peak pattern into an asymmetric single - peak one , which features persistent intrinsic oscillations .
this outcome of the instability development complies with the fact that the instability of the antisymmetric states is oscillatory , being accounted for by a quartet of eigenvalues , as seen in fig .
[ newfig2](b ) .
in other words , the transition from stable to unstable antisymmetric states may be considered as the _ hamiltonian hopf bifurcation _
@xcite .
figure [ fig8 ] displays a combined diagram in the plane of the norm of the solution and width of the nonlinear potential wells , @xmath29 and @xmath26 , which summarizes the existence and stability results for the states of all the three types symmetric , asymmetric and antisymmetric ones .
the dashed - dotted line in the figure designates the symmetry - breaking bifurcation .
solely symmetric states exist below this line ( they are stable in that region ) , and stable asymmetric states exist above the line , where the symmetric ones are unstable .
solid curves in fig .
[ fig8 ] depict stability borders of antisymmetric solutions . .
the vertical dashed - dotted line corresponds to @xmath182 , beyond which the double - well structure in eq .
( [ g ] ) turns into a single - well one .
the lower dashed - dotted line depicts the symmetry - breaking bifurcation , with stable asymmetric solutions existing above it .
the chain of circles in the bottom of the parameter plane designates the minimum norm necessary for the existence of symmetric states , which are stable between the existence border and the bifurcation line .
stable asym - anti " marks regions of the bistability , where both asymmetric and antisymmetric states are stable.,width=336 ] for the reasons explained above , the region of very small values of @xmath26 , where the quick " stabilization of asymmetric states takes place , is not included . however , the region of the existence of the analytical symmetric and antisymmetric solutions in the model with delta - functions ( @xmath90 ) , and the respective bifurcation point , as given by eq .
( [ nbif ] ) , are shown by the bold vertical segment and square - marked dot on the axis of @xmath90 ( recall that the exact asymmetric solutions are unstable above the bifurcation point in the model with @xmath90 ) . because the modulation profile ( [ g ] ) does not feature the double - well structure for @xmath183 ( see fig .
[ fig1 ] ) , the ssb bifurcation tends to disappear as @xmath26 approaches @xmath184 .
actual results are included in fig .
[ fig8 ] for @xmath185 as the convergence of the numerical scheme becomes poor for values of @xmath26 still closer to @xmath184 .
the chain of circles in figure [ fig8 ] designates the threshold ( minimum norm ) , @xmath186 , necessary for the existence of symmetric states in the model . as mentioned above , in the case of @xmath90
the exact threshold is @xmath187 , and it is observed in fig .
[ fig8 ] that the threshold remains in the ballpark of this value at finite @xmath26 , which can be easily explained .
indeed , the minimum of @xmath29 is attained at @xmath188 , in which limit the spatial scale of the wave function , @xmath189 , is much larger than the size of the ndwp structure , @xmath190 .
thus , from the viewpoint of this weakly localized wave function , modulation pattern ( [ g ] ) looks like @xmath191 . in the corresponding approximation ,
the wave function takes the form of expression ( [ xi ] ) divided by @xmath184 , and the respective norm is , indeed , @xmath192 . for @xmath193 , the condensate confined to the trap of large length @xmath194 ( i.e. , in the thermodynamic limit ) will tend to form a quasi - uniform nearly linear state , with @xmath195 . as follows from eqs .
( [ n ] ) and ( [ g ] ) , the energy of the small - amplitude uniform state is@xmath196 in fact , this state realizes a minimum of the energy ( cf .
[ fig9 ] ) , i.e. , the system s ground state . nevertheless , a well - known fact is that dynamically stable localized states _ different _ from the ground state , such as the above - mentioned gap solitons in the repulsive condensate heidelberg , or their broader counterparts , in the form of the so - called gap - waves @xcite , can be created in the experiment .
another notable feature observed in fig .
[ fig8 ] is the _ bistability _ , i.e. , the coexistence of stable asymmetric and antisymmetric states above the stability border of the latter state .
in fact , the bistability always takes place when antisymmetric states are stable .
it is interesting too that the stability area for the antisymmetric states consists of two separate regions . finally , it is relevant to mention that , as well as in the analytically solvable model with the delta - functions ( @xmath48 ) , the antisymmetric states never undergo a bifurcation at finite @xmath26 .
we stress that the stability borders displayed in fig .
[ fig8 ] were identified by means of direct simulations and the computation of stability eigenvalues , both methods yielding identical results .
in particular , the sets of eigenvalues displayed in fig . [ newfig2 ] clearly confirm the presence of two disjoint stability areas for antisymmetric states . in the case of the bistability involving the asymmetric and antisymmetric states ,
it is interesting to compare their energies ( values of the hamiltonian ) . to this end ,
[ fig9 ] displays a typical example of the dependence of @xmath197 on norm @xmath29 .
the situation observed in this figure is also true in the general case : stable antisymmetric states realize smaller values of @xmath197 than the asymmetric counterparts coexisting with them .
as argued above , dynamically stable states can be created in the experiment even if their energy is higher than in some competing states . in particular , fig .
[ newfig3 ] demonstrates that an unstable symmetric state definitely self - traps into an asymmetric robust breather ( which is close to a stable stationary solution ) , despite the fact that a stable antisymmetric state exists at the same values of @xmath181 and @xmath122 , as seen from fig .
[ fig8 ] .
.,width=336 ] note that all curves in fig .
[ fig9 ] start from finite threshold values of @xmath29 corresponding , as said above , to the minimum norm ( @xmath186 ) necessary for the existence of the respective states .
in particular , for the branch of symmetric solutions , @xmath186 is close to @xmath192 , as argued above ( cf .
the existence border in the bottom of fig .
[ fig8 ] ) , while the asymmetric branch originates at the bifurcation point ( in agreement with the location of the respective dashed - dotted line in fig .
[ fig8 ] ) , at which the symmetric solution loses its stability .
the branch of antisymmetric solutions features a fold in fig .
[ fig9 ] ( in the region where these solutions are unstable ) , which is similar to the the above - mentioned fact that dependence @xmath198 in eq .
( [ + - ] ) for the unstable exact antisymmetric states has a minimum , @xmath199 at @xmath87 .
if replotted in terms of @xmath197 and @xmath29 , eq .
( [ + - ] ) features a similar fold , at @xmath200 .
in this work , we have introduced a model of the nonlinear double - well potential ( ndwp ) , alias a double - well pseudopotential , which can be created in bec , by means of the spatially inhomogeneous feshbach resonance , and also in nonlinear optics .
the model provides for a previoulsy unexplored setting in which effects of the spontaneous symmetry breaking ( ssb ) can be studied . in the limit
when each potential well is induced by the delta - function , full analytical solutions were obtained for symmetric , antisymmetric and asymmetric states .
the symmetric states are stable in that case up to the symmetry - breaking bifurcation point , but beyond the bifurcation both symmetric and emergent asymmetric states are unstable .
in particular , the asymmetric configurations transform themselves into breathers .
the instability of all the stationary asymmetric states in the model with the delta - functions is explained by fact that the respective ssb bifurcation is of a fully backward " type , with branches of the asymmetric solutions never turning forward .
all antisymmetric states are unstable too , in this limit form of the model .
the increase of the width of the potential wells readily stabilizes the asymmetric states , which concurs with the change of the character of the ssb bifurcation from sub- to supercritical .
close to the stabilization border , unstable asymmetric states develop slow intrinsic oscillations , featuring effective dynamical re - symmetrization .
antisymmetric states may also be stable in the ndwp structure with a finite width of the wells , which implies the bistability between asymmetric and antisymmetric states .
the symmetric states exist above a finite threshold , in terms of the norm ( number of atoms in the condensate ) , and they develop the usual ssb instability above the bifurcation point .
a simple explanation to the existence threshold was given , and an integrated diagram for the existence and stability of the trapped states of all the three types has been produced .
the analysis presented in this work suggests new experiments in the matter - wave and nonlinear - optical settings .
the analysis can also be developed in other directions .
in particular , it may be interesting to study a two - dimensional nonlinear - dwp configuration . in the 2d space , a triangular configuration with three nonlinear ( pseudo-)potential wells may be considered too .
the work of t.m . is supported , in a part , by a postdoctoral fellowship from the pikovsky - valazzi foundation , by the israel science foundation through the center - of - excellence grant no .
8006/03 , and by the thailand research fund under grant no .
mrg5080171 .
shcherbakov , sov .
17 , 724 ( 1987 ) ; s. trillo , s. wabnitz , e. m. wright , and g. i. stegeman , opt . lett .
* 13 * , 672 ( 1988 ) ; s. r. friberg , a. m. weiner , y. silberberg , b. g. sfez , and p. s. smith , _ ibid_. * 13 * , 904 ( 1988 ) .
abdullaev , r. m. abrarov , and s. a. darmanyan , opt . lett . * 14 * , 131 ( 1989 ) ; e. m. wright , g. i. stegeman and s. wabnitz , phys . rev . a * 40
* , 4455 ( 1989 ) ; c. par and m. florjanczyk , _ ibid_. * 41 * , 6287 ( 1990 ) ; n. akhmediev , a. ankiewicz , phys .
* 70 * ( 1993 ) 2395 ; p. l. chu , b. a. malomed , g. d. peng , j. opt .
b * 10 * , 1379 ( 1993 ) ; j. m. soto - crespo , n. akhmediev , phys .
e * 48 * , 4710 ( 1993 ) ; b. a. malomed , i. m. skinner , p. l. chu , g. d. peng , phys .
e * 53 * , 4084 ( 1996 ) .
s. raghavan , a. smerzi , s. fantoni , and s. r. shenoy , phys .
a * 59 * , 620 ( 1999 ) ; e. a. ostrovskaya , y. s. kivshar , m. lisak , b. hall , f. cattani and d. anderson , phys .
a * 61 * , 031601(r ) ( 2000 ) ; t. kapitula and p. g. kevrekidis , nonlinearity * 18 * , 2491 ( 2005 ) ; d. ananikian and t. bergeman , phys . rev .
a * 73 * , 013604 ( 2006 ) ; g. theocharis , p. g. kevrekidis , d. j. frantzeskakis , and p. schmelcher , phys .
e * 74 * , 056608 ( 2006 ) .
s. raghavan , a. smerzi , and v. m. kenkre , phys .
a * 60 * , r1787 ( 1999 ) ; a. smerzi and s. raghavan , phys .
a * 61 * , 063601 ( 2000 ) ; d. r. dounas - frazer , a. m. hermundstad , and l. d. carr , phys .
lett . * 99 * , 200402 ( 2007 ) .
s. inouye , m. r. andrews , j. stenger , h. j. miesner , d. m. stamper - kurn and w. ketterle , nature ( london ) * 392 * , 151 ( 1998 ) ; ph .
courteille , r. s. freeland , d. j. heinzen , f. a. van abeelen , and b. j. verhaar , phys .
81 * , 69 ( 1998 ) ; j. l. roberts , n. r. claussen , j. p. burke , c. h. greene , e. a. cornell , and c. e. wieman , _ ibid_. * * 8**1 , 5109 ( 1998 ) .
h. sakaguchi and b. a. malomed , phys . rev .
e * 72 * , 046610 ( 2005 ) ; g. fibich , y. sivan , and m. i. weinstein , physica d * 217 * , 31 ( 2006 ) ; j. garnier and f. k. abdullaev , phys . rev .
a * 74 * , 013604 ( 2006 ) ; d. l. machacek , e. a. foreman , q. e. hoq , p. g. kevrekidis , a. saxena , d. j. frantzeskakis , and a. r. bishop , phys .
e * 74 * , 036602 ( 2006 ) ; m. a. porter , p. g. kevrekidis , b. a. malomed , and d. j. frantzeskakis , physica d * 229 * , 104 ( 2007 ) ; j. belmonte - beitia , v. m. prez - garca , v. vekslerchik , and p. j. torres , phys .
* 98 * , 064102 ( 2007 ) ; f. abdullaev , a. abdumalikov , and r. galimzyanov , phys .
a * 367 * , 149 ( 2007 ) ; g. dong and b. hu , phys .
rev . a * 75
* , 013625 ( 2007 ) ; d. a. zezyulin , g. l. alfimov , v. v. konotop , and v. m. prez - garca , _
ibid_. a * 76 * , 013621 ( 2007 ) ; z. rapti , p. g. kevrekidis , v. v. konotop , and c. k. r. t. jones , j. phys .
a math . theor .
* 40 * , 14151 ( 2007 ) ; h. a. cruz , v. a. brazhnyi , and v. v. konotop , j. phys .
b at . mol .
. phys . * 41 * , 035304 ( 2008 ) ; l. c. qian , m. l. wall , s. l. zhang , z. w. zhou , and h. pu , phys .
a * 77 * , 013611 ( 2008 ) ; f. k. abdullaev , a. gammal , m. salerno , and l. tomio , _ ibid_. a * 77 * , 023615 ( 2008 ) . | we propose a model of a nonlinear double - well potential ( ndwp ) , alias a double - well pseudopotential , with the objective to study an alternative implementation of the spontaneous symmetry breaking ( ssb ) in bose - einstein condensates ( becs ) and optical media , under the action of a potential with two symmetric minima . in the limit case when the ndwp structure is induced by the local nonlinearity coefficient represented by a set of two delta - functions ,
a fully analytical solution is obtained for symmetric , antisymmetric and asymmetric states . in this solvable model ,
the ssb bifurcation has a fully subcritical character .
numerical analysis , based on both direct simulations and computation of stability eigenvalues , demonstrates that , while the symmetric states are stable up to the ssb bifurcation point , both symmetric and emerging asymmetric states , as well as all antisymmetric ones , are unstable in the model with the delta - functions . in the general model with a finite width of the nonlinear - potential wells ,
the asymmetric states quickly become stable , simultaneously with the switch of the ssb bifurcation from the subcritical to supercritical type .
antisymmetric solutions may also get stabilized in the ndwp structure of the general type , which gives rise to a bistability between them and asymmetric states .
the symmetric states require a finite norm for their existence , an explanation to which is given .
a full diagram for the existence and stability of the trapped states in the model is produced .
experimental observation of the predicted effects should be possible in bec formed by several hundred atoms . |
let @xmath1 be a ( germ of a ) normal complex surface singularity .
a _ smoothing _ of @xmath1 is a flat surjective map @xmath2 , where @xmath3 has an isolated 3-dimensional singularity and @xmath4 , such that @xmath5 is isomorphic to @xmath1 and @xmath6 is smooth for every @xmath7 .
assume that @xmath8 is embedded in @xmath9 .
then there exists an embedding of @xmath3 in @xmath10 such that the map @xmath11 is induced by the projection @xmath12 to the second factor .
the _ milnor fiber _ @xmath13 of a smoothing @xmath11 of @xmath8 is defined by the intersection of a fiber @xmath6 ( @xmath14 ) near the origin with a small ball about the origin , that is , @xmath15 ( @xmath16 ) . according to the general theory of milnor fibrations ( see looijenga @xcite ) ,
@xmath13 is a compact 4-manifold , with the _ link _ @xmath17 of the singularity @xmath8 as its boundary . in particular ,
the diffeomorphism type of @xmath13 depends only on the smoothing @xmath11 ; hence , the topological invariants of @xmath13 are invariants of the smoothing @xmath11 .
the 4-manifold @xmath13 has the homotopy type of a two - dimensional cw complex , thus we have @xmath18 for @xmath19 .
furthermore , by greuel
steenbrink ( * ? ? ?
* theorem 2 ) , the first betti number @xmath20 is zero .
therefore , an important invariant of @xmath13 ( hence , of the smoothing @xmath11 ) is @xmath21 , which is a finitely generated free abelian group .
the _ milnor number _ @xmath22 of the smoothing @xmath11 is given by the second betti number @xmath23 . if @xmath24 , that is , @xmath25 for @xmath26 , we say that @xmath13 is a _ rational homology disk _ ( @xmath27hdfor short ) .
correspondingly , a smoothing @xmath11 with @xmath24 is called a _ rational homology disk smoothing _
( ` @xmath27hd smoothing ' for short ) .
for example , any cyclic quotient singularity of type @xmath28 with two relatively prime integers @xmath29 admits a @xmath27hd smoothing .
indeed , according to looijenga
wahl ( * ? ? ?
* ( 5.10 ) ) and wahl ( * ? ? ?
* ( 5.9.1 ) ) , among cyclic quotient singularities these are the only ones having a @xmath27hd smoothing .
shepherd - barron @xcite made substantial use of the fact that the @xmath27hd smoothingof a singularity of type @xmath30 is a quotient of a smoothing of its index one cover ; they invented the term `` @xmath27-gorenstein smoothing '' .
singularities of type @xmath30 play an important role in the kollr shepherd - barron alexeev ( ksba ) compactification of moduli spaces of complex surfaces of general type .
for instance , y. lee j .
park constructed a singular surface with singularities of type @xmath30 . since the milnor fibers of these singularities are topologically very simple , it is easy to control ( topological ) invariants of the smoothing of the singular surface . hence , by smoothing the singular surface , they constructed examples of complex surfaces of general type with prescribed topological invariants .
in particular , they constructed a point lying on the boundary of the ksba compactification of a moduli space of complex surfaces of general type .
using similar ideas , many important examples of complex surfaces of general type have been constructed ; see , for example , @xcite .
these constructions were motivated by the _
rational blow - down _ construction of fintushel stern @xcite , and its generalization by j. park @xcite : in this smooth construction one substitutes the tubular neighbourhood of a configuration of surfaces in a 4-manifold intersecting each other according to the resolution graph of a singularity with a @xmath27hd smoothingof the same singularity .
these constructions played a crucial role in constructing exotic differentiable structures on many 4-manifolds , cf .
for example @xcite .
therefore it is an interesting problem to classify all normal surface singularities admitting a @xmath27hd smoothing .
such a singularity @xmath1 must be rational ; in particular the resolution dual graph is a tree and the vertices correspond to rational curves . besides the cyclic quotient ones
, further such examples were described by wahl @xcite , and a list of such singularities ( compiled by wahl ) was known to the experts , cf .
the remark on the bottom of page 505 of de jong van straten @xcite . using smooth topological ideas , in stipsicz szab
wahl @xcite strong necessary combinatorial conditions for the resolution graphs of singularities with a @xmath27hd smoothinghas been derived . besides the linear graphs ( that is , the resolution graphs of cyclic quotient singularities of type @xmath30 ) the potential graphs were classified into six classes @xmath31 and @xmath32 . in the first three classes
the resolution graphs are all star - shaped ( i.e. each admits a unique node ) , with the node having valency 3 , and all these graphs are taut in the sense of laufer @xcite .
( recall that a singularity is called _ taut _ if it is determined analytically by its resolution graph . )
the singularities corresponding to the graphs in @xmath33 and @xmath34 all admit @xmath27hd smoothing .
the further three types @xmath35 and @xmath36 are defined by the following construction .
let @xmath37 , @xmath38 , @xmath39 be the graphs given as follows .
[ bullet/.style = circle , draw = black!100,fill = black!100,thick , inner sep=0pt , minimum size=0.4em ] ( 00 ) at ( 0,0 ) [ label = below:@xmath40 ; ( 10 ) at ( 1,0 ) [ label = below:@xmath41 ; ( 11 ) at ( 1,1 ) [ label = above:@xmath40 ; ( 20 ) at ( 2,0 ) [ label = below:@xmath40 ; ( 01 ) at ( -0.5,1 ) @xmath37 : ; \(00 ) ( 10 ) ; ( 10 ) ( 11 ) ; ( 10 ) ( 20 ) ; [ bullet/.style = circle , draw = black!100,fill = black!100,thick , inner sep=0pt , minimum size=0.4em ] ( 00 ) at ( 0,0 ) [ label = below:@xmath42 ; ( 10 ) at ( 1,0 ) [ label = below:@xmath41 ; ( 11 ) at ( 1,1 ) [ label = above:@xmath42 ; ( 20 ) at ( 2,0 ) [ label = below:@xmath43 ; ( 01 ) at ( -0.5,1 ) @xmath38 : ; \(00 )
( 10 ) ; ( 10 ) ( 11 ) ; ( 10 ) ( 20 ) ; [ bullet/.style = circle , draw = black!100,fill = black!100,thick , inner sep=0pt , minimum size=0.4em ] ( 00 ) at ( 0,0 ) [ label = below:@xmath44 ; ( 10 ) at ( 1,0 ) [ label = below:@xmath41 ; ( 11 ) at ( 1,1 ) [ label = above:@xmath40 ; ( 20 ) at ( 2,0 ) [ label = below:@xmath43 ; ( 01 ) at ( -0.5,1 ) @xmath39 : ; \(00 ) ( 10 ) ; ( 10 ) ( 11 ) ; ( 10 ) ( 20 ) ; a _ non - minimal graph of type @xmath45 ( or @xmath46 or @xmath36 ) _ is a graph obtained as follows : starting with the graph @xmath37 ( respectively , @xmath38 or @xmath39 ) , apply the following two _ blowing up operations _ :
1 . blow up the ( @xmath47)-vertex + + [ bullet/.style = circle , draw = black!100,fill = black!100,thick , inner sep=0pt , minimum size=0.4em ] ( 00 ) at ( 0,0 ) [ ] ; ( 10 ) at ( 1,0 ) [ label = below:@xmath41 ; ( 20 ) at ( 2,0 ) [ ] ; + \(00 )
( 10 ) ; ( 10 ) ( 20 ) ; + \(250 ) at ( 2.5,0 ) ; ( 350 ) at ( 3.5,0 ) ; + \(250 ) ( 350 ) ; + \(40 ) at ( 4,0 ) [ ] ; ( 50 ) at ( 5,0 ) [ label = below:@xmath43 ; ( 51 ) at ( 5,1 ) [ label = right:@xmath41 ; ( 60 ) at ( 6,0 ) [ ] ; + \(40 ) ( 50 ) ; ( 50 )
( 51 ) ; ( 50 ) ( 60 ) ; 2 . or blow up any edge emanating from the ( @xmath47)-vertex + + [ bullet/.style = circle , draw = black!100,fill = black!100,thick , inner sep=0pt , minimum size=0.4em ] ( 00 ) at ( 0,0 ) [ ] ; ( 10 ) at ( 1,0 ) [ label = below:@xmath41 ; ( 20 ) at ( 2,0 ) [ label = below:@xmath48 ; ( 30 ) at ( 3,0 ) [ ] ; + \(00 ) ( 10 ) ; ( 10 ) ( 20 ) ; ( 20 ) ( 30 ) ; + \(350 ) at ( 3.5,0 ) ; ( 450 ) at ( 4.5,0 ) ; + \(350 ) ( 450 ) ; + \(50 ) at ( 5,0 ) [ ] ; ( 60 ) at ( 6,0 ) [ label = below:@xmath43 ; ( 70 ) at ( 7,0 ) [ label = below:@xmath41 ; ( 80 ) at ( 8,0 ) [ label = below:@xmath49 ; ( 90 ) at ( 9,0 ) [ ] ; + \(50 ) ( 60 ) ; ( 60 )
( 70 ) ; ( 70 ) ( 80 ) ; ( 80 ) ( 90 ) ; and repeat these procedures of blowing up ( either the new ( @xmath47)-vertex or an edge emanating from it ) finitely many times .
the result is a non - minimal graph @xmath50 .
a _ minimal graph @xmath51 of type @xmath45 ( or @xmath46 or @xmath36 ) _ corresponding to a non - minimal graph @xmath50 of type @xmath45 ( respectively , @xmath46 or @xmath36 ) is a graph obtained by 1 .
_ modifying _ the unique ( @xmath47)-decoration of a non - minimal graph @xmath50 of type @xmath45 ( respectively , @xmath46 or @xmath36 ) to ( @xmath52 ) ( respectively , ( @xmath53 ) or ( @xmath54 ) ) .
the classes @xmath35 and @xmath36 are the collections of minimal graphs of the respective types .
it is not hard to see that a graph in @xmath55 has at most one node of valency 4 ( corresponding to the node of @xmath56 or @xmath57 ) and all the others are of valency 3 . using methods of symplectic topology , in bhupal
stipsicz @xcite star - shaped graphs admitting a @xmath27hd smoothinghave been completely classified .
in particular , it has been shown that if a minimal good resolution graph @xmath58 is star - shaped and corresponds to a singularity with a @xmath27hd smoothing , then @xmath59 is one of the graphs given by figures [ fig : fig1 ] or [ fig : fig2 ] .
hd smoothing .
we assume that @xmath60 . ]
hd smoothing .
the quadruple @xmath61 is one of @xmath62 ; furthermore @xmath63 . ]
in fact , in @xcite wahl conjectured that the only complex surface singularities admitting a @xmath27hd smoothingare the formerly known examples , which are all weighted homogeneous ( hence , in particular , have resolution graphs with at most one node ) . in supporting this conjecture
, wahl showed that many graphs with exactly two nodes do not correspond to singularities with a @xmath27hd smoothing , cf .
* theorem 8.6 ) .
the aim of this paper is to prove wahl s conjecture : [ thm : maintheorem ] suppose that @xmath51 is a minimal negative definite graph with at least two nodes .
then there is no complex surface singularity with resolution graph @xmath51 which admits a @xmath27hd smoothing .
this result , with the aforementioned result of bhupal
stipsicz @xcite provides the following classification result : [ cor : classificationofgraphs ] suppose that @xmath64 is a minimal negative definite graph with the property that there is a singularity which admits a @xmath27hd smoothingand has @xmath64 as a resolution graph
. then @xmath64 is either the linear graph corresponding to one of the cyclic quotient singularities of type @xmath65 ( with @xmath66 relatively prime ) or @xmath64 is one of the graphs of figures [ fig : fig1 ] or [ fig : fig2 ] .
indeed , the above result leads to the complete classification of complex normal surface singularities with @xmath27hd smoothing . since the resolution graphs of cyclic quotient singularities and the graphs of figure [ fig : fig1 ] are all taut by laufer @xcite , for these cases the singularities themselves are determined by the resolution graph .
a graph of figure [ fig : fig2 ] does not determine a unique singularity the analytic type depends on a complex number , the cross ratio of the four intersection points on the rational curve corresponding to the node of valency @xmath67 with its four neighbours . according to a recent result of j. fowler (
* theorem 5(a ) ) , for any graph in figure [ fig : fig2 ] exactly one cross ratio determines a singularity admitting a @xmath27hdsmoothing .
this value of the cross ratio is also determined by fowler @xcite : it is anharmonic for @xmath68 , harmonic for @xmath69 , and @xmath70 for @xmath71
. therefore , as a combination of corollary [ cor : classificationofgraphs ] and the result of fowler @xcite we get the classification of singularities admitting a @xmath27hd smoothing : the set of complex normal surface singularities admitting a @xmath27hd smoothing is equal to the set of singularities we get by considering * the cyclic quotient singularities of type @xmath65 ( with @xmath66 relatively prime ) , * the weighted homogeneous singularities corresponding to the taut graphs of figure [ fig : fig1 ] , and * the weighted homogeneous singularities with resolution graphs of figure [ fig : fig2 ] , together with the cross ratios : anharmonic for @xmath68 , harmonic for @xmath69 , and @xmath70 for @xmath71 . it is known that a @xmath27hd smoothingcomponent of a cyclic quotient singularity of type @xmath65 ( with @xmath66 relatively prime ) has dimension one , and the @xmath27hd smoothingcan always be chosen to be a @xmath27-gorenstein smoothing . in @xcite wahl verified the same properties for any weighted homogeneous surface singularity admitting a @xmath27hd smoothing .
hence , combined with main theorem [ thm : maintheorem ] , we conclude that any @xmath27hd smoothingof a normal surface singularity is @xmath27-gorenstein occurring over a one - dimensional smoothing component .
one of the main ideas of the proof of theorem [ thm : maintheorem ] is an extension of the result of wahl in @xcite about graphs of two nodes .
let @xmath1 be a germ of a rational surface singularity , and let @xmath72 be the minimal good resolution of @xmath73 near @xmath74 with @xmath75 the exceptional set .
let @xmath76 be the decomposition of the exceptional divisor @xmath77 into irreducible components @xmath78 with @xmath79 .
an irreducible component of the base space of the semi - universal deformation of @xmath8 is called a _ smoothing component _ if a generic fiber over such a component is smooth .
every component of the base space of the semi - universal deformation of a rational surface singularity is a smoothing component , but their dimensions may vary . by wahl (
* theorem 8.1 ) a @xmath27hd smoothingcomponent ( i.e. a component containing a @xmath27hd smoothing , if any ) of @xmath8 has dimension @xmath80 where @xmath81 is the sheaf of logarithmic vector fields ( i.e. the dual of the sheaf @xmath82 of logarithmic differentials ) ; that is , it is the kernel of the natural surjection @xmath83 . in particular ,
if the above expression is nonpositive for a particular singularity , then it admits no @xmath27hd smoothing .
the proof of our main theorem [ thm : maintheorem ] will rest on the following technical result .
[ thm : nonexist ] suppose that @xmath8 is a rational surface singularity with resolution graph @xmath51 .
assume furthermore that @xmath51 is of type @xmath45 , @xmath46 , or @xmath36 and has at least two nodes .
then @xmath84 this result immediately implies the main result of the paper : suppose that @xmath51 is a minimal negative definite graph with at least two nodes .
suppose furthermore that the singularity @xmath8 has @xmath51 as the resolution graph , and @xmath1 admits a @xmath27hd smoothing . by stipsicz
wahl @xcite then @xmath51 is of type @xmath35 , or @xmath36 . by wahl (
* theorem 8.1 ) a @xmath27hd smoothingcomponent has dimension given by equation , which expression , by theorem [ thm : nonexist ] is nonpositive .
consequently the smoothing component does not exist , concluding the proof .
the difficulty in proving theorem [ thm : nonexist ] is that singularities with resolution graphs having at least two nodes are usually non - taut .
indeed , there may exist many analytically different singularities with the same resolution graph , and the dimension @xmath85 in formula depends on the analytic structure of the singularity @xmath8 .
in dealing with this difficulty , in section [ sec : plumbing ] we prove that there exists a ` natural ' singularity @xmath86 with minimal good resolution @xmath87 that has the same weighted resolution graph ( and the same cross ratio if any ) as @xmath8 such that @xmath88 that is , the singularity @xmath86 has maximal dimension @xmath89 among singularities having the same weighted resolution graph .
so we may call @xmath86 a ` maximal ' singularity . by controlling how the expression of formula changes under the construction of the graphs in @xmath35 and @xmath36
, we verify inequality for the maximal singularities , eventually providing the proof of theorem [ thm : nonexist ] .
the singularity with the maximal dimension property has been already introduced by laufer ( * ? ? ?
* theorem 3.9 ) using the plumbing construction . in the last paragraph of laufer @xcite , he observed that @xmath89 is _ usually _
maximal for the maximal singularity among singularities with the same resolution graph .
indeed , in ( * ? ? ?
* theorem 3.1 ) laufer proved the maximality property given by inequality for ( pseudo ) taut singularities , and used this fact to obtain a complete list of resolution graphs of such singularities @xcite . in theorem [
theorem : maximality ] we generalize laufer s observation for any rational surface singularity for which the resolution graph has nodes of valency 3 with at most one exception which is of valency 4 .
let us set up some more notation .
suppose that @xmath50 is a non - minimal graph of type @xmath35 , or @xmath36 , and define the _ augmented graph _
@xmath90 as the graph obtained from @xmath50 by blowing up once the ( @xmath47)-vertex and ( if needed ) by ( successively ) blowing up an edge emanating from the ( @xmath47)-vertex according to its type as below : 1 .
type @xmath45 + + [ bullet/.style = circle , draw = black!100,fill = black!100,thick , inner sep=0pt , minimum size=0.4em ] + \(a ) at ( -0.5,0 ) @xmath91 ; + \(01 ) at ( 0,0 ) ; ( 00 ) at ( 0,0 ) ; ( 0 - 1 ) at ( 1,-1 ) ; + \(10 ) at ( 1,0 ) [ label = above:@xmath41 ; + \(01 ) ( 10 ) ; ( 0 - 1 ) ( 10 ) ; + \(150 ) at ( 1.5,0 ) ; ( 30 ) at ( 3,0 ) ; + \(150 ) ( 30 ) ; + \(b ) at ( 3.5,0 ) @xmath92 ; + \(41 ) at ( 4,0 ) ; ( 40 ) at ( 4,0 ) ; ( 4 - 1 ) at ( 5,-1 ) ; + \(50 ) at ( 5,0 ) [ label = above:@xmath42 ; + \(60 ) at ( 6,0 ) [ label = above:@xmath41 ; ( 70 ) at ( 7,0 ) [ label = above:@xmath43 ; ( 80 ) at ( 8,0 ) [ label = above:@xmath43 ; + \(41 ) ( 50 ) ; ( 4 - 1 ) ( 50 ) ; + ( 50)(60)(70)(80 ) ; 2 .
type @xmath46 + + [ bullet/.style = circle , draw = black!100,fill = black!100,thick , inner sep=0pt , minimum size=0.4em ] + \(a ) at ( -0.5,0 ) @xmath91 ; + \(01 ) at ( 0,0 ) ; ( 00 ) at ( 0,0 ) ; ( 0 - 1 ) at ( 1,-1 ) ; + \(10 ) at ( 1,0 ) [ label = above:@xmath41 ; + \(01 ) ( 10 ) ; ( 0 - 1 )
( 10 ) ; + \(150 ) at ( 1.5,0 ) ; ( 30 ) at ( 3,0 ) ; + \(150 ) ( 30 ) ; + \(b ) at ( 3.5,0 ) @xmath92 ; + \(41 ) at ( 4,0 ) ; ( 40 ) at ( 4,0 ) ; ( 4 - 1 ) at ( 5,-1 ) ; + \(50 ) at ( 5,0 ) [ label = above:@xmath40 ; + \(60 ) at ( 6,0 ) [ label = above:@xmath41 ; ( 70 ) at ( 7,0 ) [ label = above:@xmath43 ; + \(41 ) ( 50 ) ; ( 4 - 1 ) ( 50 ) ; + ( 50)(60)(70 ) ; 3 .
type @xmath36 + + [ bullet/.style = circle , draw = black!100,fill = black!100,thick , inner sep=0pt , minimum size=0.4em ] + \(a ) at ( -0.5,0 ) @xmath91 ; + \(01 ) at ( 0,0 ) ; ( 00 ) at ( 0,0 ) ; ( 0 - 1 ) at ( 1,-1 ) ; + \(10 ) at ( 1,0 ) [ label = above:@xmath41 ; + \(01 ) ( 10 ) ; ( 0 - 1 )
( 10 ) ; + \(150 ) at ( 1.5,0 ) ; ( 30 ) at ( 3,0 ) ; + \(150 ) ( 30 ) ; + \(b ) at ( 3.5,0 ) @xmath92 ; + \(41 ) at ( 4,0 ) ; ( 40 ) at ( 4,0 ) ; ( 4 - 1 ) at ( 5,-1 ) ; + \(50 ) at ( 5,0 ) [ label = above:@xmath43 ; + \(60 ) at ( 6,0 ) [ label = above:@xmath41 ; + \(41 ) ( 50 ) ; ( 4 - 1 ) ( 50 ) ; + ( 50)(60 ) ; note that the minimal graph @xmath51 corresponding to @xmath50 can be obtained by deleting the redundant vertices and edges from @xmath90 ( three vertices and edges for type @xmath45 , two for type @xmath46 and one for type @xmath36 ) .
the paper is organized as follows . in section [ sec : plumbing ]
we prove that for a given resolution graph @xmath51 there is a natural singularity with @xmath51 as its resolution graph such that the @xmath93 appearing in the dimension formula of equation is maximal ( among singularities with the same resolution graph @xmath51 ) .
for this we review the construction of some specific surfaces ( called the _ plumbing surface _ ) . in section [ sec : cohomological - properties ] we verify some cohomological properties of these specific surfaces .
then , in section [ sec : fourth ] we provide formulae for the change of the dimension of equation under blow - ups and provide the proof of theorem [ thm : nonexist ] , which ultimately implies the main result of the paper . throughout this paper we work over the field of complex numbers .
the authors would like to thank j. wahl for his careful reading and valuable comments , and for pointing out an error in the proof of proposition [ proposition : independent ] of the first draft of this paper .
hp was supported by basic science research program through the national research foundation of korea ( nrf ) grant funded by the korean government ( 2011 - 0012111 ) .
ds was supported by basic science research program through the national research foundation of korea ( nrf ) grant funded by the korean government ( 2013r1a1a2010613 ) .
he thanks kias for warm hospitality when he was an associate member in kias . as was partially supported by otka nk81203 , by the _ lendlet program _ of the hungarian academy of sciences and by erc ldtbud .
the present work is part of the authors activities within cast , a research network program of the european science foundation .
in this section we prove that for a given negative definite weighted graph @xmath51 with certain properties , there is a normal surface singularity @xmath86 ( with minimal good resolution @xmath94 ) that has @xmath51 as its resolution graph and that @xmath95 is maximal among singularities having the same weighted resolution graph ( corollary [ corollary : maximality ] ) .
for this we recall the definitions of _ plumbing surfaces _ and _ plumbing curves _ associated to a weighted graph , and we investigate their properties .
( we refer to laufer ( * ? ? ?
* theorem 3.9 ) and schller @xcite for constructions of these schemes . )
let @xmath50 be a weighted graph which is a tree consisting of @xmath96-vertices @xmath78 ( @xmath97 ) with @xmath98 .
assume furthermore that the valencies of the nodes of @xmath50 are all equal to @xmath99 possibly except exactly one node with valency @xmath67 .
it is known that the analytic type of a singularity whose resolution graph has a node of valency @xmath67 depends on the cross ratio of the node of valency @xmath67 . throughout this paper a graph with a unique node of valency @xmath67 ( and all other nodes of valency @xmath99 ) is always assumed to be given with a complex number @xmath100 ( @xmath101 ) , called the _ cross ratio _ of the graph . for @xmath102 ,
let @xmath103 ( @xmath104 ) with coordinates @xmath105 .
we glue @xmath106 and @xmath107 via the isomorphism @xmath108 and obtain @xmath109 .
the ( @xmath110)-vertex @xmath78 is realized as the zero section @xmath111 of the @xmath112-bundle @xmath113 over @xmath114 with @xmath115 ( @xmath104 ) as fiber coordinates .
we first define a two - dimensional ( complex ) analytic space @xmath116 associated to @xmath50 , by gluing neighborhoods of the zero sections @xmath78 s of @xmath113 s together as explained below .
if @xmath117 for @xmath118 ( that is , if the two vertices @xmath78 and @xmath119 are connected by an edge in @xmath50 ) , we glue a neighborhood of @xmath120 and that of @xmath121 as follows : for a fixed @xmath122 , we place the ( at most four ) points @xmath123 at @xmath124 , @xmath125 , @xmath126 , @xmath127 ( @xmath128 ) , where @xmath129 is the cross ratio of the graph ( if given ) .
choose @xmath130 , @xmath131 , @xmath132 , @xmath133 as local base coordinates of @xmath113 near @xmath124 , @xmath125 , @xmath126 and @xmath127 , respectively .
near a point of @xmath134 we glue a neighborhood of @xmath120 and that of @xmath121 by interchanging the above chosen base coordinates and fiber coordinates for @xmath113 and @xmath135 . the _ plumbing surface @xmath136 associated to @xmath50 _ is a germ of the two - dimensional analytic space @xmath116 along the one - dimensional curves @xmath137 .
we will show that some relevant cohomological properties of plumbing surfaces are independent of the choice of the cross ratio @xmath129 .
so , by slight abuse of notation , we denote the plumbing surface associated to @xmath50 by @xmath136 for simplicity , instead of recording also @xmath129 in the notation .
let @xmath50 be a weighted graph and let @xmath138 be a graph obtained by blowing up a vertex or an edge of @xmath50 in a way that @xmath139 has the same number of valency 4 nodes as @xmath50 .
it is not hard to show that the plumbing surface @xmath140 associated to @xmath138 is equal to the surface @xmath141 blown up at the appropriate point on @xmath136 . for a non - minimal weighted graph @xmath50 of type @xmath45 , @xmath46 , or @xmath36 , a model for the plumbing surface @xmath136
can be obtained as follows : let @xmath142 be the negative section of the hirzebruch surface @xmath143 , i.e. @xmath142 is a section with @xmath144 .
choose three distinct fibers @xmath145 , @xmath146 , @xmath147 of @xmath148 intersecting @xmath142 at @xmath74 , @xmath149 , @xmath150 , respectively .
then a concrete model for the plumbing surface @xmath136 can be obtained by appropriately blowing up a small neighborhood of the negative section @xmath142 and three distinct fibers @xmath151 .
indeed , let @xmath152 be one of the weighted graphs @xmath37 , @xmath38 , or @xmath39 .
after the appropriate sequence of blow - ups we can identify a configuration of curves ( in the proper transform of the section @xmath142 and the three fibers @xmath151 ) which intersect each other according to @xmath152 in the resulting rational surface . the plumbing surface
@xmath153 is a germ of the resulting rational surface along the curves . by further blowing up the curves at appropriate points
, we can find a configuration of curves in the proper transform intersecting each other according to the given graph @xmath50 .
the germ of the surface along these curves then provides the plumbing surface @xmath136 .
let @xmath154 and let @xmath155 be the ideal sheaf of @xmath78 in @xmath136 .
we define the _ plumbing curve @xmath156 associated to @xmath50 and @xmath157 _ as a non - reduced one - dimensional scheme defined by the ideal sheaf @xmath158 , which is the same as the _ plumbing construction _ of laufer ( * ? ? ?
* theorem 3.9 ) . for brevity , in case of @xmath159 , we denote @xmath156 by @xmath160 .
here we briefly recall a more detailed construction of the plumbing curve @xmath156 given in schller @xcite and @xcite .
let @xmath161 , and let @xmath162 ( @xmath163 ) be the @xmath164 curves with @xmath165 .
we first define a @xmath149-dimensional scheme @xmath166 .
each @xmath166 consists of the following three affine open subschemes of @xmath156 .
( in what follows , if there is no node with valency @xmath67 , then one may remove the terms @xmath167 ( @xmath129 is the cross ratio ) in the formulae ) : if @xmath168 then @xmath169/\langle x_{i1}^{s_{i_1}}y_{i1}^{s_i}\rangle ) \setminus \{y_{i1}=1 , y_{i1}=c\ } \\ w_{i2 } & = \operatorname{spec}(\mathbb{c}[x_{i2 } , y_{i2}]/\langle y_{i2}^{s_i}\rangle ) \\ w_{i,12 } & = \operatorname{spec}\left(\frac{\mathbb{c}[x_{i1 } , y_{i1 } , x_{i2 } , y_{i2}]}{\langle x_{i1}x_{i2}-1 , y_{i1}-x_{i2}^{d_i}y_{i2 } , y_{i2}^{s_i}\rangle}\right ) \\ & \qquad \setminus \{y_{i1}=1 , y_{i1}=c\}.\end{aligned}\ ] ] if @xmath170 then @xmath169/\langle x_{i1}^{s_{i_1}}y_{i1}^{s_i}\rangle ) \setminus \{y_{i1}=1 , y_{i1}=c\ }
\\ w_{i2 } & = \operatorname{spec}(\mathbb{c}[x_{i2 } , y_{i2}]/\langle x_{i2}^{s_{i_2 } } y_{i2}^{s_i}\rangle ) \setminus \{y_{i2}=1 , y_{i2}=c\ } \\
w_{i,12 } & = \operatorname{spec}\left(\frac{\mathbb{c}[x_{i1 } , y_{i1 } , x_{i2 } , y_{i2}]}{\langle x_{i1}x_{i2}-1 , y_{i1}-x_{i2}^{d_i}y_{i2 } , y_{i2}^{s_i}\rangle}\right ) \\ & \qquad \setminus\{y_{i1}=1 , y_{i1}=c , y_{i2}=1 , y_{i2}=c\}.\end{aligned}\ ] ] if @xmath171 then @xmath169/\langle x_{i1}^{s_{i_1}}(x_{i1}-1)^{s_{i_3 } } y_{i1}^{s_i}\rangle ) \setminus \{y_{i1}=1 , y_{i1}=c\}\\ w_{i2 } & = \operatorname{spec}(\mathbb{c}[x_{i2 } , y_{i2}]/((x_{i1}-1)^{s_{i_3 } } x_{i2}^{s_{i_2 } } y_{i2}^{s_i } ) ) \setminus \{y_{i2}=1 , y_{i2}=c\}\\ w_{i,12 } & = \operatorname{spec}\left(\frac{\mathbb{c}[x_{i1 } , y_{i1 } , x_{i2 } , y_{i2}]}{\langle x_{i1}x_{i2}-1 , y_{i1}-x_{i2}^{d_i}y_{i2 } , ( x_{i1}-1)^{s_{i_3}}y_{i2}^{s_i}\rangle}\right ) \\ & \qquad \setminus \{y_{i1}=1 , y_{i1}=c , y_{i2}=1 , y_{i2}=c\}.\end{aligned}\ ] ] if @xmath172 then @xmath169/\langle x_{i1}^{s_{i_1}}(x_{i1}-1)^{s_{i_3}}(x_{i1}-c)^{s_{i_4}}y_{i1}^{s_i}\rangle ) \setminus \{y_{i1}=1\ } \\ w_{i2 } & = \operatorname{spec}(\mathbb{c}[x_{i2 } , y_{i2}]/\langle ( x_{i2}-1)^{s_{i_3 } } ( cx_{i2}-1)^{s_{i_4 } } x_{i2}^{s_{i_2 } } y_{i2}^{s_i}\rangle ) \setminus \{y_{i2}=1\ } \\ w_{i,12 } & = \operatorname{spec}\left(\frac{\mathbb{c}[x_{i1 } , y_{i1 } , x_{i2 } , y_{i2}]}{\langle x_{i1}x_{i2}-1 , y_{i1}-x_{i2}^{d_i}y_{i2 } , ( x_{i1}-1)^{s_{i_3}}(x_{i1}-c)^{s_{i_4}}y_{i2}^{s_i } \rangle}\right ) \\ & \qquad \setminus\{y_{i1}=1 , y_{i2}=1\}.\end{aligned}\ ] ] the plumbing curve @xmath156 is given by gluing @xmath166 and @xmath173 in case @xmath174 by interchanging the base coordinates and the fiber coordinates for @xmath166 and @xmath173 .
that is , if @xmath175 for @xmath176 , then we glue @xmath166 and @xmath173 by the relation @xmath177 with @xmath178 if @xmath179 with respect to @xmath166 , or @xmath180 if @xmath181 with respect to @xmath166 , or @xmath182 else , and analogously for @xmath183 .
let @xmath1 be a germ of a rational surface singularity .
let @xmath72 be the minimal good resolution of @xmath73 with @xmath75 the exceptional set .
let @xmath184 be the decomposition of the exceptional set @xmath77 into irreducible components .
then the @xmath78 s have only normal crossings and @xmath185 .
for @xmath154 let @xmath186 ( @xmath187 ) be an effective exceptional cycle supported on @xmath77 .
let @xmath51 be the weighted graph corresponding to @xmath77 .
in what follows we assume that the valencies of the vertices of @xmath51 are @xmath188 possibly except one node with valency @xmath67 ( as it is satisfied by graphs of type @xmath45 , @xmath46 , or @xmath36 ) , although the same method would give the results for more general graphs . furthermore ,
if there is a node of valency @xmath67 , say @xmath189 , then we assume that the cross ratio @xmath129 of the graph @xmath51 is given as that of the four intersection points in @xmath189 by its four neighbours .
[ proposition : z - via - gluing - w ] the scheme @xmath190 can be obtained by gluing the open subsets @xmath166 of the plumbing curve @xmath156 with @xmath191 by using various gluing maps . the proof is given in the proof of laufer ( * ? ? ?
* theorem 3.9 ) or in that of schller ( * ? ? ? * lemma 3.2 ) . here
we briefly recall
how to glue @xmath166 ( for details see schller ( * ? ? ?
* lemma 3.2 ) ) .
there are open neighborhoods of @xmath78 in @xmath192 isomorphic to @xmath166 for every @xmath78 .
for @xmath174 , letting @xmath193 , @xmath194 , @xmath195 , @xmath183 as before , we glue @xmath166 and @xmath173 by the relations @xmath196 for some @xmath197 and @xmath198 $ ] . [
proposition : deformation - z - c ] let @xmath186 ( @xmath187 ) be an effective exceptional cycle supported on @xmath77 . then there exist an integral affine scheme @xmath199 and a locally trivial flat surjective map
@xmath200 such that @xmath201 for some closed point @xmath202 and @xmath203 for some @xmath204 .
we briefly sketch the proof of schller ( * ? ? ? * proposition 3.14 ) for the convenience of the reader .
suppose that @xmath190 is defined by the relations in .
let @xmath205\ ] ] with @xmath206 running over all @xmath206 such that @xmath207 .
here we put @xmath208 and @xmath209 in @xmath210 because @xmath211 in the gluing map .
let @xmath212 .
then @xmath213 is defined as follows : @xmath214 and @xmath215 can be glued along @xmath216 via @xmath217 then it is not difficult to show that the second projection @xmath200 is flat , @xmath218 , and @xmath219 .
next we compare @xmath220 and @xmath221 : [ theorem : maximality ] let @xmath222 be the minimal good resolution of a rational surface singularity .
let @xmath223 be the decomposition of the exceptional set @xmath77 into irreducible components .
for @xmath224 let @xmath186 ( @xmath225 ) be an effective exceptional cycle supported on @xmath77 .
let @xmath51 be the weighted dual graph corresponding to @xmath77 ( given with the same cross ratio of the node of valency @xmath67 of @xmath77 , if any ) .
then we have @xmath226 by the mayer - vietoris sequence ( cf .
laufer ( * ? ? ? * ( 3.10 ) ) or schller ( * ? ? ?
* lemma 3.4 ) ) , we have @xmath227 where @xmath228 and @xmath229 are restriction maps . furthermore , in computing them , by laufer ( * ? ? ?
* ( 3.11 ) ) or schller ( * ? ? ?
* ( 4.16 ) ) it is enough to consider only elements of @xmath230 of the form @xmath231 we now consider the elements in @xmath232 and @xmath233 . at first , note that @xmath234 .
let @xmath235 as before .
depending on @xmath164 , the elements of @xmath234 are given as follows ( cf .
laufer @xcite and laufer ( * ? ? ?
* ( 4.4 ) ) ; or schller @xcite ) : for any @xmath164 , @xmath236 with @xmath237 , @xmath238 . for @xmath239
we have @xmath240 with @xmath241 , @xmath242 .
additionally , for @xmath168 , we have @xmath243 with @xmath242 .
for @xmath171 we have @xmath244 with @xmath245 , @xmath238 . finally for @xmath172
we have @xmath246 with @xmath247 , @xmath238 , where @xmath129 is the cross ratio . according to schller (
* corollary 3.9 ) , in order to compute @xmath220 and @xmath248 , we first construct matrices @xmath249 and @xmath250 in the following way : for every intersection point @xmath251 of @xmath252 and every element of equation we add one row to @xmath249 and @xmath250 , respectively . then for every @xmath166 and for every element of , , , , or , we add one column to @xmath249 and @xmath250 , respectively .
the entries of the matrices @xmath249 and @xmath250 are the coefficients of the element associated to the column as an expression in the element associated to the row .
note that the two matrices @xmath249 and @xmath250 have the same number of rows , say @xmath253 .
the entries of @xmath250 are complex numbers determined by @xmath51 ( and the cross ratio @xmath129 , if given ) and @xmath191 . on the other hand ,
the entries of @xmath249 are polynomials in @xmath210 of equation .
the difference between @xmath250 and @xmath249 is coming from the gluing data of equations and .
then it follows by that @xmath254 therefore , if @xmath255 is a nearby fiber of the deformation @xmath256 in proposition [ proposition : deformation - z - c ] , then we have @xmath257 because the @xmath258 is locally constant on the base space @xmath199 . therefore @xmath220 remains constant for the general fiber @xmath255 of the deformation @xmath256 .
the assertion then follows by upper semicontinuity .
[ lemma : w<-e>=tz ] for @xmath259 , we have @xmath260 this is a well - known fact ; here we give a proof for the convenience of the reader .
according to burns
wahl ( * ? ? ?
* subsection ( 1.6 ) ) , there is an exact sequence @xmath261 then we have the following commutative diagram : @xmath262 & \\ 0 \ar[r ] & \theta_v(-z(s ) ) \ar[r]^{\text{id } } & \theta_v(-z(s ) ) \ar[d ] \ar[r ] & 0 \ar[d ] \ar[r ] & 0 \\ 0 \ar[r ] & \theta_v(-\log{e } ) \ar[r ] & \theta_v \ar[r ] \ar[d ] & \bigoplus_{i=1}^{n } \sheaf{n_{e_i / v } } \ar[r ] \ar[d]^{\text{id } } & 0 \\ 0 \ar[r ] & \theta_{z(s ) } \ar[r ] & \theta_v \otimes \sheaf{o_{z(s ) } } \ar[r ] \ar[d ] & \bigoplus_{i=1}^{n } \sheaf{n_{e_i / v } } \ar[r ] & 0\\ & & 0 & & } \ ] ] by the snake lemma , we get an exact sequence @xmath263 since @xmath51 is negative definite , one may choose @xmath264 so that @xmath265 for all @xmath122 ( that is , @xmath266 is ample ) .
we have @xmath267 ( @xmath268 ) by kodaira vanishing , hence there is an isomorphism @xmath269 on the other hand , for any @xmath270 , the above isomorphism @xmath271 factors through @xmath272 note that the first map is surjective ; therefore it is an isomorphism .
hence we have @xmath273 the combination of theorem [ theorem : maximality ] and lemma [ lemma : w<-e>=tz ] immediately implies : [ corollary : maximality ] with the notation as in theorem [ theorem : maximality ] , @xmath274
let @xmath50 be a non - minimal graph of type @xmath45 , @xmath46 , or @xmath36 .
let @xmath51 be the corresponding minimal graph , and let @xmath90 be the augmented graph corresponding to @xmath50 .
the goal of this section is to compare @xmath275 to @xmath276 ( see theorem [ theorem : independent - alpha ] ) .
suppose that @xmath277 is the decomposition of the exceptional divisor @xmath278 in @xmath279 . since @xmath51 is negative definite , there is @xmath280 such that @xmath281 for all @xmath97 , that is , @xmath282 is ample in @xmath279 .
set @xmath283 .
[ lemma : easy - vanishing ] for @xmath284 , we have @xmath285 where @xmath286 means the cohomology with support on @xmath278 .
let @xmath287 . in the following , for simplicity
, we will denote @xmath288 , @xmath279 , @xmath289 , @xmath290 , @xmath291 , @xmath278 by @xmath292 , @xmath293 , @xmath294 , @xmath295 , @xmath296 , @xmath297 , respectively .
let @xmath298 be the map contracting @xmath297 to a point , say @xmath299 . since @xmath300 by kodaira vanishing , we have the following commutative diagram with exact rows : @xmath301 & \gamma(\overline{x } , \pi_{\ast}{\sheaf{f } } ) \ar@{=}[d ] \ar[r ] & \gamma(\overline{x}-p , \pi_{\ast}{\sheaf{f } } ) \ar[r ] \ar@{=}[d ] & h^1_p(\pi_{\ast}{\sheaf{f } } ) \\
0 \ar[r ] & \gamma(\overline{s } , \sheaf{f } ) \ar[r ] & \gamma(\overline{s}-\overline{z } , \sheaf{f } ) \ar[r ] & h^1_{\overline{z}}(\sheaf{f } ) \to 0 } \ ] ] by laufer ( * ? ? ?
* lemma 5.2 ) , @xmath302 is coherent . since @xmath303 is cohen - macaulay at @xmath299 ( being two - dimensional and normal ) , @xmath304 by schlessinger ( * ? ? ?
* lemma 1 ) ; hence @xmath305 . therefore @xmath306 ; thus , @xmath307 hence the assertion follows .
the above lemma may be proved by a general result , the _ easy vanishing theorem _ of wahl @xcite .
[ lemma : tzbar ] for @xmath284 , we have @xmath308 .
we use the same notations as in the proof of lemma [ lemma : easy - vanishing ] for simplicity . from the short exact sequence ,
we have @xmath309 since @xmath310 by kodaira vanishing , it is enough to show that @xmath311 is an isomorphism . by the definition of @xmath312 , we have the short exact sequence @xmath313 where @xmath314 .
so we have @xmath315 on the other hand , by equation , we have @xmath316 using the depth argument as in the proof of lemma [ lemma : easy - vanishing ] , we have @xmath317 hence the assertion follows .
[ proposition : independent ] for @xmath284 , the cohomology group @xmath318 depends only on the type ( @xmath45 , @xmath46 , or @xmath36 ) of the graph @xmath51 . with the same notations as in the proof of lemma [ lemma : easy - vanishing ] , we have the following exact sequence : @xmath319 where @xmath320 by lemma [ lemma : easy - vanishing ] and @xmath321 because @xmath322 is a locally free sheaf on the germ @xmath292 of an analytic space along a one - dimensional curve ( cf .
grauert ( * ? ? ?
* satz 1 , p. 355 ) ) .
we first prove that @xmath323 and @xmath324 in the above sequence depend only on the type of the graph @xmath51 . at first
, we will show that @xmath325 depends only on the type of the graph @xmath51 ; then , it is clear that @xmath323 depends only on the type of the graph @xmath51 .
this follows from the observation that @xmath326 where @xmath327 .
since @xmath292 is obtained by blowing up ( several times ) the plumbing surface corresponding to the graph @xmath37 , @xmath38 , or @xmath39 according to its type , the complement @xmath328 depends only on the type of the graph @xmath51 .
furthermore @xmath329 depends only on the type of the graph @xmath51 .
therefore @xmath325 depends only on the type of the graph @xmath51 as asserted .
next we prove that @xmath324 depends only on the type of the graph @xmath51 . in the following exact sequence @xmath330 & h^1(\overline{s } \setminus \overline{z } , \sheaf{f } )
\ar[r ] & h^2_{\overline{z}}(\sheaf{f } ) \ar[r ] & h^2(\overline{s } , \sheaf{f}),}\ ] ] we have @xmath331 by kodaira vanishing ; hence @xmath332 therefore it is enough to show that @xmath333 depends only on the type of the graph . for @xmath334 , which depends only on the type of the graph , we have @xmath335 since @xmath325 depends only on the type of the graph as we seen above
, so does @xmath336 .
hence @xmath324 depends only on the type of the graph @xmath51 . finally , the cohomology group @xmath337 is the kernel of the connecting homomorphism @xmath338 in the exact sequence equation , which is just a restriction map because @xmath339 by equation . since the two spaces @xmath325 and @xmath333 depend only on the type of the graph , so does @xmath340 .
therefore @xmath337 in the exact sequence of equation also depends only on the type of the graph @xmath51 as asserted .
the following result of flenner - zaidenberg shows how the cohomologies of logarithmic tangent sheaves change under blow - ups : [ proposition : flenner - zaidenberg ] let @xmath341 be a nonsingular surface , and let @xmath342 be a simple normal crossing divisor on @xmath341 .
let @xmath343 be the blow - up of @xmath341 at a point @xmath344 of @xmath342 .
let @xmath345 . 1 .
if @xmath344 is a smooth point of @xmath342 , then there is an exact sequence @xmath346 where the constant sheaf @xmath112 is supported on @xmath344 .
hence we have @xmath347 2 . if @xmath344 is on two components of @xmath342 , then @xmath348 for @xmath268 . after these preparations , we are ready to turn to the proof of the main result of this section . [
theorem : independent - alpha ] let @xmath50 be a non - minimal graph of type @xmath45 , @xmath46 , or @xmath36 .
let @xmath51 be the corresponding minimal graph , and let @xmath90 be the augmented graph corresponding to @xmath50 .
there is a constant @xmath349 which depends only on the type of @xmath50 ( not on the graph @xmath50 itself and the cross ratio , if any ) such that @xmath350 we then have @xmath351 as in equation , we have a short exact sequence @xmath352 by lemma [ lemma : w<-e>=tz ] , @xmath353 , and by lemma [ lemma : tzbar ] we have that @xmath308
. therefore we have an exact sequence @xmath354 where @xmath355 by grauert ( * ? ? ? * satz 1 , p. 355 ) ( as in the proof of proposition [ proposition : independent ] ) . by the above equation we have @xmath356 here
@xmath357 and @xmath358 are finite dimensional because @xmath292 is a germ of an analytic space along one - dimensional curves ( cf .
grauert ( * ? ? ?
* satz 1 , p. 355 ) ) . on the other hand , consider the short exact sequence @xmath359 where @xmath360 is the redundant divisor , i.e. @xmath361 and @xmath362 for @xmath363 .
since @xmath364 , we have @xmath365 where @xmath366 by grauert ( * ? ? ?
* satz 1 , p. 355 ) as before .
then @xmath367 from equations and we have @xmath368 set @xmath369 the redundant divisor @xmath360 depends only on the type of the graph . by proposition [ proposition :
independent ] , the quantity @xmath370 also depends only on the type of the graph @xmath51 .
therefore the constant @xmath349 depends only on the type of the graph @xmath51 , and so the first assertion of the proposition follows . for the second assertion ,
since @xmath292 is obtained from @xmath136 by blowing up once the ( @xmath47)-vertex of @xmath50 and ( if needed ) blowing up edges emanating from the ( @xmath47)-vertex , it follows by proposition [ proposition : flenner - zaidenberg ] that @xmath371 therefore the second inequality follows .
in this section we prove the main technical result , theorem [ thm : nonexist ] of the paper .
the proof will rest on the following two lemmas , where we treat the case of zero or one node of valency 4 . recall that a graph @xmath372 is called
_ @xmath373-shaped _ if it admits two nodes , both with valency 3 , and we say that a graph @xmath50 is _ key - shaped _ if it admits two nodes with valencies @xmath99 and @xmath67 , respectively .
recall that if a graph with a node @xmath77 of valency @xmath67 is a resolution graph of a rational surface singularity , then @xmath374 . 1 . [ bullet/.style = circle , draw = black!100,fill = black!100,thick , inner sep=0pt , minimum size=0.4em ] ( -105 ) at ( -1,0.5 ) @xmath376 ; + \(00 ) at ( 0,0 ) [ label = below:@xmath48 ; ( 10 ) at ( 1,0 ) [ label = below:@xmath377 ; ( 11 ) at ( 1,1 ) [ label = left:@xmath378 ; ( 40 ) at ( 4,0 ) [ label = below:@xmath379 ; ( 41 ) at ( 4,1 ) [ label = left:@xmath41 ; ( 42 ) at ( 4,2 ) ; ( 60 ) at ( 6,0 ) ; + \(00 ) ( 10 ) ; ( 10 ) ( 11 ) ; ( 10 ) ( 40 ) ; ( 40 ) ( 60 ) ; ( 40 ) ( 41 ) ; ( 41 ) ( 42 ) ; + for @xmath380 ; or 2 . [ bullet/.style = circle , draw = black!100,fill = black!100,thick , inner sep=0pt , minimum size=0.4em ] ( -105 ) at ( -1,0.5 ) @xmath376 ; + \(00 ) at ( 0,0 ) [ label = below:@xmath48 ; ( 10 ) at ( 1,0 ) [ label = below:@xmath377 ; ( 11 ) at ( 1,1 ) [ label = left:@xmath378 ; ( 40 ) at ( 4,0 ) [ label = below:@xmath43 ; ( 41 ) at ( 4,1 ) [ label = left:@xmath41 ; ( 60 ) at ( 6,0 ) ; + \(00 ) ( 10 ) ; ( 10 )
( 11 ) ; ( 10 ) ( 40 ) ; ( 40 ) ( 60 ) ; ( 40 ) ( 41 ) ; by the definition of the classes @xmath35 , and @xmath36 , one of the neighbours of the @xmath396-vertex is a @xmath397-vertex , while the other one is a @xmath398-vertex with @xmath380 .
we distinguish two cases according to whether the @xmath397-vertex is between the @xmath396-vertex and the node ( motivated by the diagram above , we say that the @xmath397-vertex is to the _ left _ of the @xmath396-vertex ) , or the @xmath397-vertex is on the other side of the @xmath396-vertex ( it is to the _ right _ of the @xmath396-vertex ) .
\(00 ) at ( 0,0 ) [ label = below:@xmath48 ; ( 10 ) at ( 1,0 ) [ label = below:@xmath399 , label=45:@xmath342 ] ; ( 11 ) at ( 1,1 ) [ label = left:@xmath378 ; ( 30 ) at ( 3,0 ) [ label = below:@xmath401 , label = above:@xmath77 ] ; ( 40 ) at ( 4,0 ) [ label = below:@xmath41 ; ( 50 ) at ( 5,0 ) [ label = below:@xmath54 , label = above:@xmath210 ] ; ( 60 ) at ( 6,0 ) ; if we blow down the ( @xmath47)-vertex of @xmath152 then the ( @xmath54)-vertex @xmath210 of @xmath152 becomes the ( @xmath47)-vertex of the new graph .
since a ( @xmath47)-vertex in a star - shaped graph of @xmath45 , @xmath46 , or @xmath36 with valencies @xmath188 is not a leaf , there must be a vertex attached to the right side of the ( @xmath54)-vertex @xmath210 of @xmath152 . in summary ,
the non - minimal graph @xmath375 is of the form : \(00 ) at ( 0,0 ) [ label = below:@xmath48 ; ( 10 ) at ( 1,0 ) [ label = below:@xmath399 , label=45:@xmath342 ] ; ( 11 ) at ( 1,1 ) [ label = left:@xmath378 ; ( 30 ) at ( 3,0 ) [ label = below:@xmath401 , label = above:@xmath77 ] ; ( 40 ) at ( 4,0 ) [ label = below:@xmath43 ; ( 41 ) at ( 4,1 ) [ label = left:@xmath41 ; ( 50 ) at ( 5,0 ) [ label = below:@xmath43 ; ( 60 ) at ( 6,0 ) ; ( 70 ) at ( 7,0 ) ; where @xmath380 . if @xmath404 , then the minimal graph @xmath386 is of type @xmath405 using the contraction @xmath406 .
if @xmath407 , the graph @xmath386 is of type @xmath405 for @xmath408 or @xmath409 for @xmath394 , concluding the argument in this subcase .
\(00 ) at ( 0,0 ) [ label = below:@xmath48 ; ( 10 ) at ( 1,0 ) [ label = below:@xmath399 , label=45:@xmath342 ] ; ( 11 ) at ( 1,1 ) [ label = left:@xmath378 ; ( 30 ) at ( 3,0 ) [ label = below:@xmath54 , label = above:@xmath210 ] ; ( 40 ) at ( 4,0 ) [ label = below:@xmath43 ; ( 41 ) at ( 4,1 ) [ label = left:@xmath41 ; ( 50 ) at ( 5,0 ) [ label = below:@xmath401 , label = above:@xmath77 ] ; ( 60 ) at ( 6,0 ) ; where @xmath380 . if @xmath410 , then the minimal graph @xmath386 is of type @xmath411 .
suppose that @xmath412 . if there is a ( @xmath399)-vertex with @xmath408 between @xmath342 and @xmath210 , then @xmath386 is of type @xmath413 for @xmath408 , and , of type @xmath414 for @xmath394 .
assume that there is no ( @xmath399)-vertex with @xmath408 between @xmath342 and @xmath210 .
if @xmath408 , then @xmath386 is of type @xmath413 using the contraction @xmath415 .
if @xmath394 , then @xmath77 is a leaf of the graph and @xmath386 is of type @xmath411 .
1 . [ bullet/.style = circle , draw = black!100,fill = black!100,thick , inner sep=0pt , minimum size=0.4em ] ( -105 ) at ( -1,0.5 ) @xmath376 ; + \(00 ) at ( 0,0 ) [ label = below:@xmath48 ; ( 10 ) at ( 1,0 ) [ label=315:@xmath377 ; ( 11 ) at ( 1,1 ) [ label = left:@xmath378 ; ( 1 - 1 ) at ( 1,-1 ) [ label = left:@xmath416 ; ( 40 ) at ( 4,0 ) [ label = below:@xmath379 ; ( 41 ) at ( 4,1 ) [ label = left:@xmath41 ; ( 42 ) at ( 4,2 ) ; ( 60 ) at ( 6,0 ) ; + \(00 ) ( 10 ) ; ( 10 ) ( 11 ) ; ( 10 ) ( 1 - 1 ) ; ( 10 ) ( 40 ) ; ( 40 ) ( 60 ) ; ( 40 ) ( 41 ) ; ( 41 ) ( 42 ) ; + for @xmath380 ; or 2 .
[ bullet/.style = circle , draw = black!100,fill = black!100,thick , inner sep=0pt , minimum size=0.4em ] ( -105 ) at ( -1,0.5 ) @xmath376 ; + \(00 ) at ( 0,0 ) [ label = below:@xmath48 ; ( 10 ) at ( 1,0 ) [ label=315:@xmath377 ; ( 11 ) at ( 1,1 ) [ label = left:@xmath378 ; ( 1 - 1 ) at ( 1,-1 ) [ label = left:@xmath416 ; ( 40 ) at ( 4,0 ) [ label = below:@xmath43 ; ( 41 ) at ( 4,1 ) [ label = left:@xmath41 ; ( 60 ) at ( 6,0 ) ; + \(00 ) ( 10 ) ; ( 10 ) ( 11 ) ; ( 10 ) ( 1 - 1 ) ; ( 10 ) ( 40 ) ; ( 40 ) ( 60 ) ; ( 40 ) ( 41 ) ; suppose that the analytic structure of the singularity @xmath422 ( obtained by contracting @xmath423 in @xmath424 ) is determined by the graph @xmath386 and the analytic structure on the reduced exceptional set @xmath423 .
by laufer ( * ? ?
* theorem 3.2 ) and laufer ( * ? ? ? * ( 4.1 ) ) , a necessary and sufficient condition for @xmath386 and @xmath423 to determine the singularity @xmath422 is that the restriction map @xmath425 is an isomorphism . since the analytic structure on @xmath423
is uniquely determined by the cross ratio of the 4 intersection points on the ( @xmath399)-curve with valency @xmath67 , we have @xmath426 hence it follows by lemma [ lemma : w<-e>=tz ] that @xmath427 therefore it is enough to show that the analytic structure of the singularity @xmath422 is determined by @xmath386 and @xmath423 .
for this we will use the list in laufer ( * ? ? ?
* theorem 4.1 ) of all dual graphs for singularities which are determined by the graph and the analytic structure on the reduced exceptional set .
the proof is similar to that of lemma [ lemma : h - taut ] .
\(00 ) at ( 0,0 ) [ label = below:@xmath48 ; ( 10 ) at ( 1,0 ) [ label=315:@xmath399 , label=45:@xmath342 ] ; ( 11 ) at ( 1,1 ) [ label = left:@xmath378 ; ( 1 - 1 ) at ( 1,-1 ) [ label = left:@xmath416 ; ( 30 ) at ( 3,0 ) [ label = below:@xmath401 , label = above:@xmath77 ] ; ( 40 ) at ( 4,0 ) [ label = below:@xmath43 ; ( 50 ) at ( 5,0 ) [ label = below:@xmath43 ; ( 41 ) at ( 4,1 ) [ label = left:@xmath41 ; ( 60 ) at ( 6,0 ) ; where @xmath380 .
notice that since @xmath434 ( for @xmath386 being a resolution graph of a rational surface singularity ) , the @xmath396-framed vertex in a star - shaped graph with valency 4 node and of type @xmath35 , or @xmath36 can not be a leaf . for @xmath429 ,
@xmath386 is of type @xmath435 possibly using the contraction @xmath415 . for @xmath431 ,
@xmath386 is of type @xmath436 possibly using the contraction @xmath406 .
for @xmath433 , if @xmath437 then @xmath386 is of type @xmath438 , or if @xmath439 then @xmath386 is of type @xmath440 .
\(00 ) at ( 0,0 ) [ label = below:@xmath48 ; ( 10 ) at ( 1,0 ) [ label=315:@xmath377 ; ( 11 ) at ( 1,1 ) [ label = left:@xmath378 ; ( 1 - 1 ) at ( 1,-1 ) [ label = left:@xmath416 ; ( 30 ) at ( 3,0 ) [ label = below:@xmath43 ; ( 40 ) at ( 4,0 ) [ label = below:@xmath43 ; ( 41 ) at ( 4,1 ) [ label = left:@xmath41 ; ( 50 ) at ( 5,0 ) [ label = below:@xmath379 ; ( 60 ) at ( 6,0 ) ; where @xmath380 . for @xmath429 , @xmath51 is of the form @xmath441 using the contraction @xmath415 . for @xmath431 , @xmath51 is of the form @xmath442 using the contraction @xmath415 . for @xmath433 , @xmath51 is of the form @xmath438 if there are only ( @xmath54)-vertices between two nodes , or @xmath51 is of the form @xmath443 otherwise .
[ lemma : gamma0 ] let @xmath375 be a non - minimal graph in lemma [ lemma : h - taut ] or lemma [ lemma : key - shaped ] , and let @xmath444 be the non - minimal graph obtained by blowing up once the ( @xmath47)-vertex of @xmath375 .
then @xmath445 where @xmath446 if @xmath375 is a graph in lemma [ lemma : h - taut ] , @xmath447 if @xmath375 is a graph in lemma [ lemma : key - shaped ] , and @xmath349 is given by theorem [ theorem : independent - alpha ] ( and it depends only on the type of @xmath375 ) .
it follows from theorem [ theorem : independent - alpha ] that @xmath448 since @xmath449 is obtained from @xmath450 by blowing up edges emanating from the ( @xmath47)-curve ( if needed ) , it follows by proposition [ proposition : flenner - zaidenberg ] that @xmath451 [ theorem : technical - main ] let @xmath375 be a non - minimal graph in lemma [ lemma : h - taut ] or lemma [ lemma : key - shaped ] and let @xmath444 be the non - minimal graph obtained by blowing up once the ( @xmath47)-vertex of @xmath375 .
let @xmath50 be a non - minimal graph obtained from @xmath444 by applying the blow - ups ( b-1 ) and ( b-2 ) ( described in section [ sec : intro ] ) finitely many times .
suppose that @xmath8 is a rational surface singularity with a resolution @xmath452 which admits the minimal graph @xmath51 as the resolution graph .
let @xmath453 be the decomposition of the exceptional divisor into irreducible components @xmath78 with @xmath79 .
then @xmath454 let @xmath457 be the number of blow - ups of the ( @xmath47)-vertices to obtain the non - minimal graph @xmath50 from @xmath444 .
since @xmath458 by lemma [ lemma : gamma0 ] , using proposition [ proposition : flenner - zaidenberg ] it follows that @xmath459 then theorem [ theorem : independent - alpha ] implies that @xmath460 note that for any minimal graph consisting of ( @xmath461)-vertices , the blow - up procedure ( b-1 ) with the modification ( m ) lowers the sum @xmath462 by @xmath149 , while the procedure ( b-2 ) with ( m ) leaves it unchanged ; cf .
the proof of wahl ( * ? ? ? * theorem 8.6 ) . since these sums for the starting graphs @xmath463 , @xmath464 , @xmath465 are equal to @xmath149 , this sum for @xmath466 is @xmath467 .
since we blow up the ( @xmath47)-vertices @xmath457 times to obtain @xmath51 from @xmath466 , we have @xmath468 from equation and equation it follows that @xmath469 concluding the proof .
suppose first that @xmath51 is of the form @xmath386 where @xmath375 is a non - minimal graph in lemma [ lemma : h - taut ] or lemma [ lemma : key - shaped ] .
if @xmath375 is a non - minimal graph in lemma [ lemma : h - taut ] , then we have @xmath470 by lemma [ lemma : h - taut ] and it is easy to see that @xmath471 . therefore , as in the proof of wahl ( * ? ? ?
* theorem 8.6 ) , we have @xmath472 on the other hand , if @xmath375 is a non - minimal graph in lemma [ lemma : key - shaped ] , then we have @xmath473 by lemma [ lemma : key - shaped ] .
but , since @xmath474 , we have @xmath475 suppose now that @xmath51 is not of the form @xmath386 of lemma [ lemma : h - taut ] or lemma [ lemma : key - shaped ] .
let @xmath444 be the non - minimal graph obtained by blowing up the ( @xmath47)-vertex of @xmath375 once ( as in theorem [ theorem : technical - main ] ) .
then @xmath51 is obtained by applying the blow - ups ( b-1 ) and ( b-2 ) to @xmath444 finitely many times , and finally the modification ( m ) .
the assertion of the theorem now follows from theorem [ theorem : technical - main ] .
h. laufer .
_ deformations of resolutions of two - dimensional singularities_. complex analysis , 1972 , vol .
i : geometry of singularities ( proc .
rice univ . ,
houston , tex . , 1972 ) .
rice univ .
studies * 59 * ( 1973 ) , no . 1 , 5396 . | we show that if the minimal good resolution graph of a normal surface singularity contains at least two nodes ( i.e. vertex with valency at least 3 ) then the singularity does not admit a smoothing with milnor fiber having rational homology equal to the rational homology of the 4-disk @xmath0 ( called a rational homology disk smoothing ) . combining with earlier results
, this theorem then provides a complete classification of resolution graphs of normal surface singularities with a rational homology disk smoothing , verifying a conjecture of j. wahl regarding such singularities . indeed , together with a recent result of j. fowler we get the complete list of normal surface singularities which admit rational homology disk smoothings . |
let @xmath0 be a convex body in @xmath4 .
a subset @xmath5 of @xmath0 is a _ face _ of @xmath0 if every open interval in @xmath0 that contains a point of @xmath6 is contained in @xmath5 .
extreme point _ is a 1-point face . if @xmath7 is any subset of @xmath0 , the _ _ face generated by _ _
@xmath7 is the minimal face of @xmath0 containing @xmath7 .
the set @xmath8 of all faces of @xmath0 ordered by inclusion is a lattice , where @xmath9 is the intersection of @xmath6 and @xmath10 , and @xmath11 is the face generated by @xmath12 .
the lattice @xmath13 is always _ algebraic _ ( the chains of faces are finite ) , _ atomic _ ( faces are generated by extreme points ) and _ complemented _ ( for every face @xmath5 there exists a face @xmath10 such that @xmath14 and @xmath15 ) .
we want to consider convex bodies for which @xmath16 is _ modular _
, i.e. @xmath17 whenever @xmath18 .
modularity is a ` weak distributivity ' property satisfied by the lattice of normal subgroups of a group and by the lattice of subspaces of a vector space . for algebraic , atomic lattices
, modularity is equivalent to the existence of a rank function such that @xmath19 for all @xmath5 and @xmath10 @xcite .
strictly convex bodies and simplices clearly have modular face lattices .
no other polytopes have this property @xcite , but there are beautiful examples of non - polytopal convex bodies in which every pair of extreme points is contained in a proper face and every pair of faces with more than one point meet . if @xmath20 and @xmath21 are lattices , their _ direct product _ is given by @xmath22 where @xmath23 if and only if @xmath24 and @xmath25 .
it follows that the direct product of two lattices is modular if and only if the factors are modular .
a lattice is called _ irreducible _ if it is not isomorphic to a direct product of two nontrivial lattices . if @xmath26 and @xmath27 are convex bodies , define @xmath28 as the convex hull of a copy of @xmath29 and a copy of @xmath30 placed in general position in the sense that their linear spans are disjoint and have no common directions .
so @xmath31 is well defined up to a linear transformation : it is the convex join of @xmath29 and @xmath30 of largest dimension .
for example @xmath32 is a pyramid with base @xmath0 .
let s say that a convex body @xmath0 is @xmath33-_decomposable _ if @xmath34 for two convex bodies @xmath29 and @xmath30 .
the natural correspondence ( up to linear transformation ) between convex bodies in @xmath4 and closed cones in @xmath35 gives an isomorphism of face lattices in which @xmath31 corresponds to the direct product of the cones , so the results of this paper apply to cones .
this project started with the undergraduate thesis of d. labardini - fragoso @xcite , who showed that in dimension less than 6 any cone with modular face lattice is strictly convex or is decomposable ( this was conjectured by barker in @xcite ) .
a convex body @xmath0 is @xmath33-decomposable if and only if its lattice of faces @xmath36 is reducible .
let @xmath37 .
observe that each point @xmath38 of @xmath31 with @xmath39 , lies in a unique segment joining a point @xmath40 of @xmath41 and a point @xmath42 of @xmath30 .
for , if a point lies in two segments @xmath43 and @xmath44 then the lines @xmath45 and @xmath46 are parallel or they intersect , contradicting the assumptions on the spans of @xmath29 and @xmath30 . moreover , if a point @xmath47moves along a straight line in @xmath31 then the corresponding points @xmath40 and @xmath42 move along straight lines in @xmath29 and @xmath30 : if @xmath38 and @xmath48 are points in @xmath0 and @xmath49 @xmath50 then @xmath51 which can be rewritten as a linear combination of a point in @xmath52 and a point in @xmath53 with coefficients adding up to 1 so @xmath54 and @xmath55 . now if @xmath56 is a face of @xmath57 then @xmath58 is a face of @xmath31 . for ,
if @xmath59 and @xmath60 with @xmath61 , then @xmath62 lies in @xmath52 and @xmath63 lies in @xmath64 so as @xmath65 is a face of @xmath57 , @xmath66 and @xmath67 lie in @xmath65 so @xmath38 and @xmath48 lie in @xmath58 .
conversely , if @xmath68 is a face of @xmath31 and @xmath69 then @xmath40 and @xmath42 lie in @xmath68 so @xmath70 .
it remains to show that @xmath71 is a face of @xmath57 .
if @xmath72 and @xmath60 with @xmath61 then as @xmath68 , and @xmath73 are faces of @xmath31 , @xmath38 and @xmath48 lie in @xmath68 and also in @xmath41 , so @xmath74 is a face of @xmath29 .
@xmath75 is a face of @xmath30 .
so @xmath76 if @xmath77 then @xmath78 and @xmath21 are isomorphic to sublattices of @xmath13 , so @xmath79 for two faces of @xmath0 with @xmath80 and @xmath81 . to show that @xmath37 we need to prove that @xmath82 and @xmath83 are disjoint and have no directions in common .
suppose that @xmath84 .
take @xmath85 then the line through @xmath86 and @xmath87 meets @xmath88 at two points @xmath89 and @xmath90 as @xmath91 lies in a proper subface @xmath92 of @xmath30 , the face generated by @xmath29 and @xmath91 lies in @xmath93 which is a proper subface of @xmath94 .
but the points @xmath95 , @xmath96 , @xmath91 , @xmath97 determine a plane quadrilateral whose side @xmath98 lies in the interior of @xmath57 so its diagonals intersect at an interior point @xmath99 of @xmath94 so the face generated by @xmath29 and @xmath91 ( which contains @xmath99 ) must be @xmath94 , a contradiction .
now suppose that @xmath82 and @xmath100 have a common direction @xmath101 .
take @xmath85 then the line through @xmath87 in the direction @xmath101 meets @xmath88 at two points @xmath89 and @xmath102 . as before @xmath95 , @xmath96 , @xmath91 , @xmath97
determine a plane quadrilateral whose diagonals intersect at an interior point @xmath99 of @xmath94 , but @xmath99 lies in the face generated by @xmath29 and @xmath91 which is a proper face of @xmath103 .
recall that a _ projective space _ consists of a set @xmath104 ( the points ) and a set @xmath105 ( the lines ) so that ( a ) each pair of points is contained in a unique line , ( b ) if @xmath106 are distinct points and the lines @xmath107 and @xmath108 intersect , then the lines @xmath109 and @xmath110 intersect ( c ) each line contains at least 3 points and there are at least 2 lines ( d ) every chain of subspaces ( also called _ flats _ ) has finite length .
the maximum length of a chain starting with a point is the _ projective dimension _ of the space .
the flats of a projective space form an algebraic , atomic , irreducible , modular lattice .
conversely , any lattice with these properties is the lattice of flats of a projective space , whose points are atoms and whose lines are joins of two atoms @xcite .
it is a classic result of hilbert @xcite that a projective space in which desargues theorem holds is isomorphic to the projective space @xmath111 determined by the linear subspaces of @xmath112 , for some division ring @xmath113 , and that @xmath111 and @xmath114 are isomorphic if and only if @xmath113 and @xmath115 are isomorphic and @xmath116 .
all projective spaces of dimension larger than @xmath117 are desarguesian , but there are many non - desarguesian projective planes
. examples of convex bodies whose face lattices determine the projective spaces @xmath118 , @xmath119 and @xmath120 , and the octonionic projective plane arise as sections of some classical cones .
let @xmath121,@xmath2,@xmath122 , let @xmath123 be the set of hermitian ( self - adjoint ) @xmath124 matrices with coefficients in @xmath125 , and let @xmath126 be the subset of positive - semidefinite matrices ( @xmath127 is positive - semidefinite if @xmath128 for all @xmath129 @xmath130 ) . then @xmath126is a real cone whose face lattice is isomorphic to the lattice of subspaces of @xmath130
. to see this , let @xmath131 , and let @xmath132 denote the face generated by @xmath133 . then @xmath134 if and only if @xmath135 . for , @xmath134 @xmath136 such that @xmath137 @xmath138 such that @xmath139 for all @xmath140 @xmath130 @xmath141 implies @xmath142 for all @xmath140 @xmath143 ( since for @xmath144 , @xmath142 if and only if @xmath145 ) .
therefore @xmath146 defines a bijection @xmath147 from the set of faces of @xmath148 to the set of linear subspaces of @xmath149 . to prove that @xmath147 is an isomorphism of lattices
observe that @xmath150 , and on the other hand , if @xmath151 is a non - empty face , then it is generated by a matrix @xmath0 with @xmath152 , so @xmath153 .
let @xmath154 be the set of hermitian @xmath155 matrices over @xmath156 .
then the subset @xmath157 of all sums of squares of elements in @xmath154 is a real cone whose face lattice determines an octonionic projective plane .
this can be shown using the nontrivial fact that each matrix in @xmath158 is diagonalizable by an automorphism of @xmath159 that leaves the trace invariant @xcite , so:(a ) a matrix @xmath127 in @xmath154 lies in @xmath157 if and only if it can be diagonalized to a matrix @xmath160 with non - negative entries , because if @xmath127 lies in @xmath157 then @xmath160 is a sum of squares of matrices in @xmath154 , which have non - negative diagonal entries .
( b ) all the idempotent matrices in @xmath154 lie in @xmath161 as they are squares @xmath162 .
the idempotent matrices of trace 1 correspond to the extreme rays of @xmath157 since they ca nt be written as non - negative combinations of other idempotent matrices.(c ) each face of @xmath157 is generated by an idempotent matrix , because in any cone all the positive linear combinations of the same set of vectors generate the same face , so a diagonal matrix with non - negative entries generates the same face as a matrix with only zeros and ones.(d ) any two idempotent matrices of trace 1 lie in a face generated by an idempotent matrix of trace 2 , because they can be put simultaneously in the form @xmath163 , $ ] and these lie in the face generated by @xmath164 $ ] .(e ) @xmath165 is an idempotent of trace 1 if and only if @xmath166 is an idempotent with trace 2 . if @xmath127 and @xmath133 are idempotents of trace 1 , then @xmath127 lies in the face generated by @xmath167 if and only if @xmath133 lies in the face generated by @xmath166 .
this duality and ( d ) show that any two faces generated by idempotent matrices of trace 2 meet in a face generated by an idempotent matrix of trace 1 .
if @xmath0 is a convex body whose face lattice is modular and irreducible and @xmath0 is not strictly convex , the set of extreme points of @xmath0 is a projective space with flats determined by the faces of @xmath0 .
we would like to know which projective spaces arise in this way , and what convex bodies give rise to them . by blaschke selection theorem @xcite ,
the space of all compact , convex subsets of a convex body in @xmath4 , with the hausdorff metric , is compact .
so the subspace formed by the compact convex subsets of the boundary is closed , but the subspace formed by the faces is not closed in general .
[ compact ] if @xmath0 is a convex body whose face lattice @xmath168 is modular , then the set @xmath169 of faces of rank @xmath170 , with the hausdorff metric , is compact for each @xmath170 .
let @xmath171 be a sequence of faces of rank @xmath170 .
then @xmath171 has a subsequence @xmath172 that is convergent in @xmath173 , and its limit is a compact convex set @xmath174 contained in @xmath175 , so @xmath174 generates a proper face @xmath5 of @xmath0 of some rank @xmath176 .
we claim that @xmath177 and @xmath178 . if the rank of @xmath5 was less than @xmath170 , there would be a face @xmath179 of @xmath0 of rank @xmath180 with @xmath181 .
as @xmath5 and @xmath179 are two disjoint compact sets in @xmath4 , there exists @xmath182 such that the @xmath183-neighborhoods of @xmath5 and @xmath179 in @xmath4 are disjoint . but as @xmath184 in the hausdorff metric , then for sufficiently large @xmath185 , @xmath172 is contained in the @xmath183-neighborhood of @xmath5 , therefore @xmath186 , but these 2 faces have ranks that add up to @xmath187 , so they should meet , a contradiction .
if the rank of @xmath5 is @xmath170 and @xmath188 , there is an extreme point @xmath189 , and there is a face @xmath190 of rank @xmath180 that meets @xmath5 only at @xmath38 , so @xmath191 and the previous argument gives a contradiction . to show that the rank of @xmath5 can not be larger than @xmath170 , proceed inductively on @xmath180 . as a limit of proper faces
is contained in a proper face , the claim holds if @xmath192 .
given a sequence @xmath171 of faces of rank @xmath170 , let @xmath193 be a face generated by the limit of a convergent subsequence @xmath172 . if @xmath38 is an extreme point of @xmath0 not in @xmath5 , then for sufficiently large @xmath194 , @xmath195 ( otherwise @xmath38 would be in @xmath5 ) .
let @xmath196 be a face of rank @xmath197 containing @xmath172 and @xmath38 .
now we can assume inductively that the limit of a convergent subsequence of @xmath196 generates a face of rank @xmath197 .
this face contains @xmath5 properly ( because it contains @xmath38 ) so @xmath198 .
now recall that a _ topological projective space _ is a projective space in which the sets of flats of each rank are given nontrivial topologies that make the join and meet operations @xmath199 and @xmath200 continuous , when restricted to pairs of flats of fixed ranks whose join or meet have a fixed rank @xcite .
if @xmath0 is a convex body whose face lattice is modular and irreducible then @xmath201 is strictly convex or the set of extreme points @xmath202 is a topological projective space which is compact and connected .
a natural topology for the set of flats is given by the hausdorff distance between the faces .
by lemma [ compact ] , @xmath203 is compact . as the lattice is irreducible and has more than 2 points , the 1-flats have more than 2 points ,
so ( as they are topological spheres ) they are connected .
now every pair of points in @xmath204 is contained in one of these spheres , so @xmath204 is connected ( one can actually show that each.@xmath205 is connected ) . it remains to show that @xmath199 and @xmath200 are continuous on the preimages of each @xmath169 .
suppose @xmath206 , @xmath207 where @xmath208 and @xmath209 are faces corresponding to @xmath170 flats .
we need to show that @xmath210 . by lemma [ compact ]
@xmath211 has convergent subsequences and the limit of a convergent subsequence @xmath212 is a face @xmath213 corresponding to an @xmath170 flat . as @xmath212
is contained in @xmath214 and @xmath215 , @xmath216 is contained in @xmath209 . but @xmath213 and @xmath209 are both faces corresponding to @xmath170 flats , so @xmath217 .
similarly , if @xmath206 , @xmath218 and @xmath219 , @xmath220 are faces corresponding to @xmath170 flats , the limit of each convergent subsequence of @xmath221 is a face @xmath222 corresponding to an @xmath170 flat . as @xmath223 contains @xmath224 and @xmath225
, @xmath222 contains @xmath220 , and as both faces correspond to @xmath170 flats they must be equal .
let @xmath0 be any convex body .
denote by @xmath226 the set of baricenters of faces of @xmath0 and let @xmath227 the function that assigns to each face its baricenter .
[ convergence ] ( a ) if @xmath8 is compact then @xmath102 is a homeomorphism .
\(b ) if @xmath228 is compact then a sequence of faces @xmath171 converges to a face @xmath5 if and only if @xmath229 converges to @xmath230 .
\(a ) the function that assigns to each compact convex set in @xmath4 its baricenter is continuous , so @xmath231 is a continuous bijective map from a compact hausdorff space to a metric space .
\(b ) the hausdorff distance between two compact convex sets is bounded above by the hausdorff distance between their sets of extreme points .
if @xmath171 converges to @xmath5 but @xmath229 does n't converge to @xmath232 then there is a subsequence @xmath233 that stays at distance at least @xmath234 from @xmath235 .
for each @xmath236 there is an extreme point @xmath237 whose distance from @xmath232 is larger than @xmath238 , or an extreme point @xmath239 whose distance from @xmath240 is larger than @xmath238 .
if there is a convergent subsequence @xmath241 then @xmath38 is at distance at least @xmath238 from @xmath232 , so @xmath38 ca nt be an extreme point of @xmath242 if there is a convergent subsequence @xmath243 , take @xmath244 with @xmath245 .
so the distance from @xmath247 to @xmath248 is at least @xmath249 , so @xmath247 is the center of a straight interval @xmath250 of length @xmath238 contained in @xmath251 .
a convergent subsequence of these intervals yields a straight interval centered at @xmath48 and contained in @xmath5 , so @xmath48 ca nt be an extreme point of @xmath0 , contradicting the compacity of @xmath228 .
[ extension ] if @xmath0 and @xmath68 are convex bodies with @xmath252 and @xmath253 compact , then any continuous `` face preserving '' map @xmath254 extends naturally to a continuous map @xmath255 .
@xmath256 determines a function @xmath257 .
@xmath258 is continuous because by lemma [ convergence ] , @xmath259 implies @xmath260 , then uniform continuity of @xmath256 on @xmath228 implies that @xmath261 so by definition @xmath262 and so
so @xmath256 can be extended to a continuous function @xmath264 as @xmath265 ( recall that @xmath266 ) .
now we can extend @xmath256 to the interiors of the faces of @xmath0 defining it linearly on rays , as follows .
for each point @xmath267 let @xmath268 be the unique face of @xmath0 containing @xmath269 in its interior and let @xmath270 be the baricenter of @xmath268 .
although @xmath268 and @xmath270 are not continuous functions of @xmath269 on all of @xmath0 , they are continuous on the union of the interiors of the faces corresponding to @xmath170-flats for each @xmath170 . if @xmath271 let @xmath272 be the projection of @xmath269 to @xmath273 from @xmath270 and let @xmath274 ( or @xmath275 if @xmath276 ) so @xmath277 .
define @xmath278 assume inductively that @xmath256 is continuous on the union of @xmath279 and the faces of @xmath0 of dimension less than @xmath280 ( this set is closed because the limit of faces of dimension less than @xmath280 has dimension less than @xmath280 ) . and let s show that for each sequence of points @xmath89 in the interiors of faces of dimension @xmath280 , @xmath281 implies @xmath282 .
we may assume that the @xmath89 are not baricenters , so @xmath283 is well defined .
@xmath284 then @xmath285 by the continuity of @xmath102 on faces .
if @xmath286 then @xmath287 and @xmath288 so @xmath289 . if @xmath290 then @xmath291 but @xmath283 may not converge , so consider a convergent subsequence @xmath292 : if @xmath293 @xmath294 then @xmath295 so @xmath296 \rightarrow \varphi ( b(a))+0=\varphi ( a)$ ] . if @xmath297 @xmath298 then @xmath299 ( by continuity of @xmath256 in the baricenters and faces of lower dimension ) and as @xmath300 lies between them , @xmath301 .
@xmath302 , then for any convergent subsequence @xmath303 with limit a face @xmath304 , @xmath269 lies in @xmath5 and so @xmath269 must lie in @xmath305 , so @xmath306 and @xmath307 , so @xmath308 ( by continuity of @xmath256 on the faces of lower dimension ) .
[ main ] let @xmath0 be a convex body whose face lattice defines a n - dimensional projective space . if @xmath309 , then @xmath0 has dimension @xmath310 or @xmath311 . if @xmath312 ( or the space is desarguesian ) there is a face - preserving homeomorphism from @xmath0 to a section of a cone of hermitian matrices over @xmath313 , @xmath2 , or @xmath314 .
first consider the case @xmath309 .
all the lines of a topological projective plane @xmath315 are homeomorphic because if @xmath316 is a line and @xmath38 is a point not in @xmath316 then the projection @xmath317 , @xmath318 is continuous and its restriction to each projective line not containing @xmath38 is one to one . if the projective lines are topological spheres then a famous result of adams @xcite , shows that their dimension must be @xmath319 or @xmath320 . to compute the dimension of @xmath0 take 3 faces of rank 1 , @xmath321 , @xmath322 and @xmath323 so that @xmath322 and @xmath323 meet at a point @xmath38 not in @xmath321 .
the projection @xmath324 extends to a continuous map @xmath325 whose restriction to each face is one to one ( see proof of lemma [ extension ] ) .
now @xmath326 is an open subset of @xmath175 and the function @xmath327 defined as @xmath328 is continuous and bijective , so @xmath329 has the same dimension as @xmath330 , which is @xmath331 , therefore @xmath0 has dimension @xmath332 .
note that the discrepancy between the dimensions of the union of the boundaries of the faces ( @xmath333 ) and the union of the faces ( @xmath331 ) arises because the boundaries of the faces overlap ( as the lines in a projective plane do ) but the interiors of the faces are disjoint .
when @xmath312 , there is a similar homeomorphism from an open subset of @xmath175 and a product @xmath334 where @xmath321 is a face of rank @xmath335 and @xmath336 are faces of rank 1 .
so @xmath337 and it follows by induction that @xmath338 .
now assume that the projective space determined by@xmath339 is desarguesian .
every topological desarguesian projective space is isomorphic to a projective space over a _
division ring @xmath127 ( defined on a line minus a point ) and the isomorphism is a homeomorphism hig . by a classic result of pontragin @xcite the only locally compact , connected division rings are @xmath340 @xmath2 and @xmath3 .
so the projective space determined by @xmath228 is isomorphic to @xmath341 or @xmath342 .
therefore @xmath228 is isomorphic to @xmath343 where @xmath68 is a section of a cone of hermitian matrices , and so by extension there is a face - preserving homeomorphism from @xmath0 to @xmath344 .
now let us consider closed ( but not necessarily compact ) convex sets in @xmath345 whose faces meet as the subspaces of an affine space .
an abstract affine plane consists of a set of points and a set of lines so that 1 ) there are at least 2 points and 2 lines 2 ) every pair of points is contained in a line and 3 ) given a line and a point not contained in it , there is a unique line containing the point and parallel to the line .
the axioms of an abstract affine space are not so simple , but it is enough to know that if @xmath104 is a projective space then the complement of a maximal flat of @xmath104 is an affine space , and any affine space @xmath127 can be embedded in a projective space in this fashion , by attaching to @xmath127 a point at infinity for each parallelism class of affine lines . observe that if a closed convex set @xmath0 in @xmath4 is non - compact , it contains a ray ( half of a euclidean line ) and if @xmath0 contains a ray then it contains all the parallel rays starting at points of @xmath0 ( we say that @xmath0 contains an infinite direction ) .
so if @xmath0 contains a line , @xmath0 is the product of that line and a closed convex set @xmath68 of lower dimension and the face lattice of @xmath0 and @xmath68 are isomorphic .
so from now on we will assume that @xmath0 does nt contain lines .
it is easy to see that the faces of a polytope can not determine an affine space ( the faces of rank @xmath346 would have dimension @xmath346 , two parallel faces of rank 1 generate a face of rank 2 with at least 4 vertices , but the sides of a polygon do nt define an affine plane ) .
let @xmath0 be a convex body in @xmath4 whose faces determine a projective space .
take a cone over @xmath0 and slice it with a hyperplane parallel to a support hyperplane containing a maximal face .
the result is a closed non - compact convex set @xmath68 in @xmath4 whose faces determine an affine space .
in particular , the cones of hermitian matrices have non - compact sections whose face lattice determines a real , complex or quaternionic affine space or an octonionic affine plane .
@xmath118 can be seen as the space of lines through the origin in @xmath347 or as the quotient of the unit sphere @xmath348 ( or the sphere at infinity of @xmath35 ) by the action of the antipodal map . identifying @xmath4 with a hyperplane of @xmath35 that does nt contain the origin gives an embedding of @xmath4 as a dense open subset of @xmath118 .
the remaining points of @xmath349 correspond to lines through the origin in @xmath35 that do nt meet the hyperplane , i.e. , parallelism classes of lines in @xmath350 ( or pairs of antipodal points in the sphere at infinity)@xmath351 define a set in @xmath118 to be _ convex _ if it is the image of a convex set in @xmath4 under one of these embeddings . as convex sets in @xmath118 correspond to convex cones based at the origin of @xmath347 , convexity in @xmath118
does nt depend on the particular embedding , and a convex set in @xmath118 has the usual properties of a convex set in @xmath4 .
now if @xmath0 is a closed convex set in @xmath4 that does nt contain lines , its closure @xmath352 is a convex set in @xmath118 .
the faces of @xmath352 are the closures of faces of @xmath0 and their intersections with the sphere at infinity modulo the antipodal map . in a closed convex set in @xmath4 that has semi - modular face lattice , two faces of rank 1
can share at most one direction , and it corresponds to a ray .
we are considering convex bodies without lines .
suppose that two rank 1 faces @xmath322 and @xmath323 have a common direction , i.e. , there are segments of parallel euclidean lines @xmath353 and @xmath354 lying in @xmath322 and @xmath323 .
we may assume that @xmath355 goes through an interior point of @xmath356 so @xmath357 meets @xmath358 in one or two extreme points .
if @xmath353 or @xmath359 has two extreme points then there is a convex quadrilateral with sides in @xmath353 and @xmath354 with 3 extreme points as vertices .
the interior of the quadrilateral lies in the interior of the rank 2 face generated by the 3 extreme points , but the intersection of its diagonals lies in the rank 1 face generated by 2 extreme points , a contradiction .
so @xmath355 meets @xmath360 in only one point and so @xmath171 contains a ray @xmath361 . if @xmath322 and @xmath323 have two common directions , there is a common direction which meets @xmath322 in 2 points , giving the same contradiction .
if the faces of a closed convex set @xmath0 in @xmath4 define an affine space , then each face representing a line contains a unique ray , and faces representing parallel lines contain parallel rays .
we are assuming again that @xmath0 does nt contain lines , so the points of the affine space correspond to the extreme points of @xmath0 .
each affine lines is represented by the boundary of a convex set of dimension at least 2 , which is connected , so set of extreme points in @xmath0 is connected .
let s first show that the faces representing affine lines can not be compact .
suppose that @xmath0 has a compact face @xmath5 .
let @xmath38 and @xmath48 be two extreme points in @xmath5 and let @xmath67 be a sequence of extreme points not in @xmath5 that converge to @xmath48 .
let @xmath171 be the face generated by @xmath38 and @xmath67 .
if @xmath171 is non - compact , it contains a ray @xmath362 through @xmath38 .
so @xmath171 contains the `` parallelogram '' determined by the interval @xmath363 and the ray @xmath364an infinite sequence of @xmath355 s would have a subsequence converging to a ray @xmath316 through @xmath48 , so the parallelogram determined by the interval @xmath50 and the ray @xmath316 would be contained in @xmath175 , so it would have to be contained in a face of @xmath0 , which would have to be @xmath5 because it contains @xmath38 and @xmath48 .
this contradicts the assumption that @xmath5 is compact and shows that if @xmath67 is sufficiently close to @xmath48 , the face generated by @xmath38 and @xmath67 is compact .
now take a face @xmath190 that does nt meet @xmath5 ( i.e. , @xmath5 and @xmath190 represent parallel affine lines ) . as @xmath5 is compact and @xmath190
is closed in @xmath4 , there is an @xmath238 neighborhood of @xmath5 that does nt intersect @xmath365 . by the previous argument
there is a point @xmath67 not in @xmath5 so that the face @xmath171 generated by @xmath38 and @xmath67 is contained in the @xmath183 neighborhood of @xmath5 .
so @xmath171 does nt meet @xmath190 , but @xmath5 was supposed to be the only face containing @xmath38 and disjoint from @xmath365 .
this proves that @xmath5 is non - compact , so it contains rays .
let s show thattwo faces representing parallel affine lines contain parallel rays .
let @xmath38 be an extreme point outside @xmath5 , so @xmath5 and @xmath38 generate a face @xmath366 representing an affine plane .
there are extreme points @xmath367 in @xmath5 so that the sequence of intervals @xmath368 converges to a ray @xmath369 contained in @xmath5 ( because @xmath5 is closed ) .
the sequence of intervals @xmath370 lie in @xmath371 and converge to a ray @xmath372 parallel to @xmath369 and containing @xmath38 so ( as @xmath366 is closed ) @xmath373 is contained in a face @xmath10 of @xmath371 representing an affine line . as two faces that contain parallel rays
can not meet at a single point , @xmath10 does nt meet @xmath5 so ( as @xmath5 and @xmath10 are contained in @xmath366 ) @xmath10 represents the affine line parallel to @xmath5 through @xmath38 . if @xmath5 has nonparallel rays , one can construct as before two nonparallel rays @xmath369 and @xmath374 in @xmath5 and faces @xmath10 and @xmath375 through @xmath38 and containing rays @xmath376 and @xmath377 .
the uniqueness of parallel affine lines implies that @xmath378 , so @xmath5 and @xmath10 have more than one common direction , contradicting the previous lemma . this shows that if the faces of a closed convex set @xmath0 define an affine space , @xmath0 is non - compact .
one can show that if the faces of a closed convex set @xmath0 ( containing no lines ) define a projective space , @xmath0 must be compact .
for this , one has to give a topology to the space of closed convex sets in @xmath4 that makes it locally compact , show that this makes @xmath228 into a locally compact projective space , and observe that these spaces are necessarily compact .
if @xmath0 is a closed convex set in @xmath4 whose faces determine an affine space , there is a projective transformation in @xmath118 taking @xmath352 to a compact convex set in @xmath4 whose faces determine a projective space . if the space is desarguesian , there is a face - preserving homeomorphism from @xmath0 to a non - compact slice of a cone of hermitian matrices .
we need to show that if the face lattice of @xmath0 determines an affine space , the face lattice of @xmath379 determines its projective completion .
the faces of @xmath0 representing affine lines are non - compact , and two of them share an infinite direction in @xmath4 if and only if they represent parallel affine lines .
the closure @xmath380 contains one point at infinity for each infinite direction in @xmath0 , so @xmath352 contains an extreme point at infinity for each class of faces of @xmath0 representing parallel lines .
this corresponds precisely with the definition of the projective completion of the affine space .
now the result for @xmath0 follows by applying theorem [ main ] to @xmath381 .
the face lattice of a convex body @xmath0 ( not a triangle ) determines a projective plane if every pair of extreme points is contained in a proper face and every pair of faces with more than one point meet . by theorem main this projective plane is compact and connected , so for some @xmath383 , all the faces of @xmath0 have dimension @xmath384 and @xmath0 has dimension @xmath385 .
@xmath390 we claim that if @xmath7 is not the span of a face then @xmath7 can not intersect @xmath386 in more than one point . for , if @xmath391 contains a line , then @xmath392 is 3 dimensional .
take an extreme point @xmath393 and let @xmath322 and @xmath323 be 2 faces containing @xmath38 , and meeting @xmath5 at points @xmath40 and @xmath42 not in @xmath7 . if @xmath394 and @xmath395 are points in @xmath396and @xmath397respectively , then @xmath38 , @xmath66 and @xmath398 are not aligned ( otherwise @xmath38 would be in the span of @xmath399 ) and so the span of @xmath38 , @xmath66 and @xmath400 , which is @xmath401 , is contained in @xmath402 @xmath403 . but @xmath402 @xmath403 is 4 dimensional , so it can not contain the 3 faces @xmath5 , @xmath322 and @xmath323 because if it did , it would contain each face that meets @xmath5 , @xmath322 and @xmath404 at @xmath405 different points , but every face is a limit of such faces , so it would contain all the faces of @xmath0 , but @xmath0 has dimension 5 .
this shows that @xmath7 intersects each @xmath386 at exactly one point , and so @xmath7 contains at most one extreme point of @xmath0 .
the function @xmath406 that maps each face @xmath171 to the point of intersection of @xmath386 with @xmath7 is continuous , and as the spans of faces meet only at extreme points , @xmath407 only fails to be injective on the faces containing the extreme point in @xmath7 ( if any ) .
but @xmath408 is a 2-dimensional closed surface in which the faces that contain an extreme point form a closed curve , and there are no continuous maps from a closed surface to the plane that fail to be injective only along a curve .
@xmath7 does nt contain extreme points of some face @xmath5 .
choose @xmath5 that minimizes the dimension of the subspace @xmath391 .
then for every @xmath365 in a neighborhood of @xmath5 , @xmath391 is a subspace of minimal dimension and with no extreme points . if @xmath409 is the orthogonal complement of @xmath391 in @xmath7 then @xmath409 intersects
@xmath386 in one point for all @xmath190 in a smaller neighborhood @xmath410 of @xmath411 then the function @xmath412 that maps @xmath190 to @xmath413 is continuous , and it is injective as the spans of faces only meet at extreme points .
but an injective map between manifolds can only exist when the domain has dimension no larger than the target so @xmath414 , so @xmath382 and we are in case 1 .
@xmath7 contains extreme points of each face @xmath411 as @xmath0 is convex , either @xmath415 or @xmath416 . in the first case @xmath417 is contained in a face @xmath322 of @xmath0 and
so either @xmath418 ( so @xmath419 ) or there is an extreme point @xmath38 of @xmath322 not contained in @xmath7 , but then a face @xmath323 that meets @xmath322 at @xmath38 does nt meet @xmath420 so @xmath7 does nt contain extreme points of @xmath323 .
let @xmath38 be an extreme point not in @xmath7 , and consider the set @xmath421 of faces of rank 1 containing @xmath38 .
if @xmath5 and @xmath190 are distinct faces in @xmath422 , @xmath423 and @xmath424 are disjoint .
choose @xmath5 so that @xmath423 has minimal dimension , then for all @xmath190 in some neighborhood @xmath410 of @xmath5 , @xmath424 has the same dimension and the map @xmath425
@xmath426 that sends @xmath190 to the baricenter of @xmath427 is continuous and injective . as @xmath428 is a map between manifolds , @xmath429 and so by domain invariance the image of @xmath428 is an open subset of @xmath430 .
this implies that for each @xmath431 , @xmath424 consists of one point ( if a face of a convex set has more than 1 point , its baricenter is arbitrarily close to points in the boundary that are not baricenters of other faces , namely , the points in the face ) and so , by hypothesis , @xmath432 is an extreme point of @xmath0 .
so part of the boundary of @xmath417 in @xmath7 is strictly convex , therefore the line segment joining two extreme points in it lies in @xmath433 , but that line segment lies in the face of @xmath0 containing the 2 extreme points , so it must lie in @xmath434 , a contradiction . by lemma
[ spans ] , the set @xmath436 of spans of faces of @xmath0 is the same as the set of @xmath384-dimensional subspaces of @xmath4 that intersect every element of @xmath436 .
the set of all @xmath384-dimensional affine subspaces of @xmath4 forms a real algebraic variety and the condition that the subspaces meet a fixed subspace is algebraic , so ( by the finite descending chain condition ) there is a finite family of spans @xmath437 such that @xmath438 if it intersects these @xmath439 s ( see @xcite ) .
now for @xmath440 , the subspace @xmath441 has dimension at least @xmath384 ( otherwise it would be contained in two subspaces of dimension @xmath384 that meet each @xmath439 , so they would both be in @xmath436 , but two spans can only meet in 1 point ) .
so @xmath441 lies in @xmath436 if and only if its dimension is @xmath384 , and this happens if and only if some determinants ( given by polynomials on @xmath442 ) vanish .
therefore the set @xmath443 is real algebraic , as is the set @xmath444 formed by the elements of @xmath445 that contain a fixed point @xmath38 .
if @xmath322 is the face in @xmath446 and @xmath38 is an extreme point of @xmath0 outside @xmath447 then @xmath448 consists of the intersections of @xmath446 with the elements of @xmath436 containing @xmath38 .
so @xmath448 is the one to one projection of the algebraic set @xmath444 to @xmath446 , so @xmath449 is at least semi - algebraic .
now assume @xmath382 so @xmath450 .
every projective plane has 7 points and 6 lines so that each line contains 3 points as in figure 1 , so @xmath0 has 7 extreme points and 6 faces intersecting in that way .
the 7 points are in general position in @xmath451 because as each face of @xmath0 is spanned by 3 points , the span of any 6 of those points contains the span of 3 faces , which is all of @xmath451
. therefore we may assume ( by applying a projective transformation ) that the 7 points are @xmath452 .
let @xmath439 be the plane spanned by the face @xmath171 .
a plane @xmath7 that intersects @xmath453 and @xmath454 has a parametrization @xmath455 with @xmath456 .
@xmath7 intersects @xmath457 and @xmath458 only if three systems of linear equations in @xmath459 represented by the following matrices have nontrivial solutions : this determinant factors as the product of a linear and a quadratic function of @xmath269 and @xmath102 ( with coefficients in @xmath99 and @xmath280 ) .
since the boundary of the face @xmath322 is formed by the intersections of @xmath446 with the planes that meet all @xmath439 s and go through a fixed point in the boundary of @xmath323 ( this corresponds to fixing @xmath99 and @xmath280 ) , the boundary of @xmath322 is contained in the union of a line and a conic .
as the boundary of @xmath322 is strictly convex , it must be the conic .
let @xmath0 and @xmath68 be two such bodies .
take extreme points @xmath464 and faces @xmath465 of @xmath0 as in figure 1 .
pick an extreme point @xmath466 in @xmath68 and two faces @xmath467 and @xmath468 of @xmath68 intersecting at @xmath469 .
let @xmath439 be the span of @xmath171 .
as the faces of @xmath0 and @xmath68 are conics , there are linear transformations from @xmath446 to @xmath470 taking @xmath322 to @xmath471 and from @xmath472 to @xmath473 taking @xmath323 to @xmath468 .
together , they define a linear transformation @xmath316 from @xmath474 to @xmath475 .
let @xmath476 for @xmath477 .
the faces @xmath478 are generated by unique pairs of @xmath66 s with @xmath479 .
let @xmath480 be the faces generated by the corresponding pairs of @xmath481s .
finally , let @xmath482 , let @xmath483 be the face generated by @xmath469 and @xmath484 and let @xmath485 .
the linear transformation @xmath316 can be extended to a projective transformation @xmath486 in @xmath487 that takes @xmath488 to @xmath484 and @xmath489 to @xmath490 .
as @xmath491 sends each @xmath66 to @xmath400 , it sends each @xmath439 to @xmath492 , so it sends each plane in @xmath451 intersecting every @xmath439 to a plane intersecting every @xmath493 .
since by construction @xmath486 takes those planes that meet @xmath448 and @xmath494 to planes that meet @xmath495 and @xmath496 , lemma [ spans ] implies that @xmath486 maps spans of faces of @xmath0 to spans of faces of @xmath68 and therefore it maps faces to faces . in dimensions _
_ _ _ @xmath320 and @xmath497 this is equivalent to ask if the projective planes are always desarguesian .
in dimension @xmath311 there might be enough space for non - equivalent non - desarguesian examples . | if a convex body @xmath0 has modular and irreducible face lattice ( and is not strictly convex ) , there is a face - preserving homeomorphism from @xmath0 to a section of a cone of hermitian matrices over @xmath1 , @xmath2 , or @xmath3 , or @xmath0 has dimension 8 , 14 or 26 . |
the average luminosities of quasars , and possibly many other properties , depend strongly on cosmological epoch .
shape , normalisation , and evolution of the quasar luminosity function ( qlf ) are among the most basic descriptions of the quasar population .
a good knowledge of the _ local _ ( @xmath6 ) qlf is essential for at least two purposes : first , to serve as zero - point for quasar evolution studies ; second , to provide a reference distribution law for unbiased subsamples constructed for statistical investigations ( e.g. , for spectral properties , host galaxies , etc ) .
unfortunately , current determinations of the qlf are limited mainly to @xmath7 , for the simple reason of lacking appropriate survey data .
most optical qso surveys discriminate against objects with extended morphology , causing severe incompleteness ( and bias ) at low redshifts . at the low - luminosity end of the qso
seyfert family , dedicated galaxy surveys have prompted investigations of the seyfert 1 luminosity function ( e.g. , cheng et al .
@xcite ; huchra & burg @xcite ) . however , these surveys are necessarily incomplete at _ high _ nuclear luminosities , where the active nuclei outshine the surrounding hosts , and the objects are not selected as galaxies anymore . in 1990
we started the ` hamburg / eso survey ' ( he s ) , a new wide - angle survey for bright qsos and seyferts , designed to avoid the above selection biases as far as possible .
a detailed description of the he s was given by wisotzki et al .
( @xcite ; hereafter paper i ) , and a first list of 160 newly discovered qsos was published by reimers et al .
( @xcite ; paper ii ) .
basically , qso candidates are selected by automated procedures from digitised objective - prism plates taken with the eso schmidt telescope , with subsequent slit spectroscopy of all candidates above the magnitude limit of typically @xmath817.5 , over an area of @xmath9deg@xmath0 at high galactic latitude . to prevent morphological bias against sources with resolved structure , such as seyferts or gravitational lenses , objects with non - stellar appearance are _
not _ excluded from the candidate lists . among other goals , we aim to construct a large , optically flux - limited sample of low - redshift ` type 1 agn ' , including qso / seyfert borderline objects . in this paper
we publish the results of an analysis of a first part of the he s , based on 33 schmidt fields , with an effective area of 611deg@xmath0 .
we present the first direct determination of the local quasar luminosity function from a single optical survey , and discuss some implications of our results . a comprehensive description of the analysis has been given by khler ( @xcite ) .
we adopt a cosmological model with @xmath10 , @xmath11 , and @xmath12 . lllll field & @xmath13 & @xmath14 & @xmath15 & @xmath16 + 501 & 17.38 & 5.5 & 0.39 & 16.38 ' '' '' + 503 & 17.25 & 4.8 & 0.34 & 17.75 + 505 & 16.95 & 7.4 & 0.53 & 17.05 + 506 & 17.66 & 7.5 & 0.54 & 17.17 + 507 & 17.11 & 7.2 & 0.51 & 17.10 + 509 & 16.95 & 5.3 & 0.38 & 17.04 + 568 & 16.95 & 5.8 & 0.41 & 18.20 + 570 & 16.78 & 4.2 & 0.30 & 18.92 + 578 & 16.64 & 6.2 & 0.44 & 17.59 + 637 & 16.90 & 5.5 & 0.39 & 17.94 + 638 & 16.95 & 7.5 & 0.53 & 16.14 + 639 & 17.47 & 5.1 & 0.36 & 19.35 + 640 & 17.50 & 4.9 & 0.35 & 16.23 + 643 & 17.41 & 4.4 & 0.31 & 21.47 + 644 & 17.31 & 3.9 & 0.28 & 18.46 + 645 & 17.42 & 3.8 & 0.27 & 20.72 + 646 & 17.23 & 4.0 & 0.28 & 17.68 + 647 & 17.46 & 5.2 & 0.37 & 20.02 + 648 & 17.15 & 6.2 & 0.44 & 17.95 + 649 & 17.05 & 7.3 & 0.52 & 16.68 + 650 & 17.15 & 7.4 & 0.53 & 11.86 + 708 & 17.21 & 4.2 & 0.30 & 20.12 + 709 & 17.46 & 5.7 & 0.41 & 20.07 + 710 & 17.62 & 4.8 & 0.34 & 22.91 + 711 & 17.11 & 3.5 & 0.25 & 19.74 + 715 & 17.35 & 4.1 & 0.29 & 21.52 + 716 & 16.78 & 3.2 & 0.23 & 18.71 + 717 & 17.14 & 2.5 & 0.17 & 17.64 + 718 & 17.35 & 3.4 & 0.24 & 20.39 + 719 & 17.25 & 2.5 & 0.18 & 20.20 + 720 & 17.30 & 3.7 & 0.26 & 19.85 + 721 & 17.14 & 3.6 & 0.25 & 19.39 + 722 & 16.86 & 5.8 & 0.41 & 18.95 + = llllll name & @xmath5 & @xmath17 & field & @xmath18 & @xmath19 + he 0952@xmath201552 & 0.108 & 16.56 & 637 & 16.17 & @xmath2022.87 ' '' '' + he 1006@xmath201211 & 0.693 & 16.37 & 709 & 15.96 & @xmath2027.03 + he 1007@xmath201405 & 0.583 & 16.05 & 637 & 15.66 & @xmath2026.94 + he 1012@xmath201637 & 0.433 & 16.22 & 638 & 15.69 & @xmath2026.25 + he 1015@xmath201618 & 0.247 & 15.91 : & 638 & 15.38 & @xmath2025.40 + he 1019@xmath201413 & 0.077 & 16.75 & 638 & 16.22 &
@xmath2022.09 + pks 1020@xmath20103 & 0.197 & 17.39 & 710 & 17.05 & @xmath2023.26 + he 1021@xmath200738 & 1.800 & 17.47 & 710 & 17.12 & @xmath2028.00 + pks 1022@xmath20102 & 2.000 & 17.58 & 710 & 17.24 & @xmath2028.14 + he 1025@xmath201915 & 0.323 & 16.91 & 568 & 16.50 & @xmath2024.84 + he 1029@xmath201401 & 0.086 & 14.07 & 638 & 13.54 & @xmath2025.01 + he 1031@xmath201457 & 0.652 & 17.39 & 639 & 17.03 & @xmath2025.83 + he 1041@xmath201447 & 1.569 & 17.26 : & 639 & 16.90 & @xmath2027.91 + he 1043@xmath201443 & 0.599 & 17.06 & 639 & 16.70 & @xmath2025.96 + he 1045@xmath202322 & 0.407 & 16.65 & 501 & 16.26 & @xmath2025.55 + pks 1048@xmath20090 & 0.345 & 16.51 & 711 & 16.27 & @xmath2025.21 + he 1104@xmath201805 & 2.319 & 16.30 & 570 & 16.00 & @xmath2029.62 + he 1109@xmath201255 & 0.596 & 17.19 & 640 & 16.84 & @xmath2025.80 + he 1110@xmath201910 & 0.111 & 16.77 & 570 & 16.47 & @xmath2022.63 + he 1115@xmath201735 & 0.217 & 16.25 & 570 & 15.95 & @xmath2024.56 + he 1120@xmath202713 & 0.389 & 16.94 & 503 & 16.60 & @xmath2025.12 + he 1159@xmath201338 & 0.506 & 16.68 & 643 & 16.36 & @xmath2025.91 + he 1200@xmath201234 & 0.553 & 17.22 & 643 & 16.91 & @xmath2025.56 + he 1201@xmath202409 & 0.137 & 16.74 & 505 & 16.21 & @xmath2023.33 + he 1211@xmath201322 & 1.125 & 16.15 & 644 & 15.87 & @xmath2028.25 + he 1223@xmath201543 & 1.735 & 17.14 & 644 & 16.86 & @xmath2028.17 + he 1228@xmath201637 & 0.102 & 16.78 & 644 & 16.50 & @xmath2022.43 + he 1233@xmath202313 & 0.238 & 17.06 & 506 & 16.52 & @xmath2024.18 + he 1237@xmath202252 & 0.096 & 17.29 & 506 & 16.75 & @xmath2022.05 + he 1239@xmath202426 & 0.082 & 16.73 & 507 & 16.22 & @xmath2022.25 + he 1254@xmath200934 & 0.139 & 15.72 & 718 & 15.48 & @xmath2024.11 + he 1255@xmath202231 & 0.492 & 17.02 & 507 & 16.51 & @xmath2025.71 + he 1258@xmath200823 & 1.153 & 16.44 & 718 & 16.20 & @xmath2027.97 + he 1258@xmath201627 & 1.709 & 17.22 & 646 & 16.94 & @xmath2028.06 + pks 1302@xmath20102 & 0.278 & 15.15 & 718 & 14.91 & @xmath2026.12 + he 1304@xmath201157 & 0.294 & 17.14 & 718 & 16.90 & @xmath2024.25 + he 1312@xmath201200 & 0.327 & 16.07 & 719 & 15.89 & @xmath2025.48 + he 1315@xmath201028 & 0.099 & 16.75 & 719 & 16.57 & @xmath2022.28 + q 1316@xmath200734 & 0.538 & 16.70 & 719 & 16.52 & @xmath2025.89 + he 1335@xmath200847 & 0.080 & 16.80 & 720 & 16.54 & @xmath2021.86 + he 1341@xmath201020 & 2.134 & 17.28 & 720 & 17.02 & @xmath2028.48 + he 1345@xmath200756 & 0.777 & 17.13 & 720 & 16.87 & @xmath2026.40 + he 1358@xmath201157 & 0.408 & 16.64 : & 721 & 16.39 & @xmath2025.43 + he 1403@xmath201137 & 0.589 & 16.86 & 721 & 16.61 & @xmath2026.01 + he 1405@xmath201545 & 0.194 & 16.51 & 649 & 15.99 & @xmath2024.29 + he 1405@xmath201722 & 0.661 & 15.85 & 649 & 15.33 & @xmath2027.55 +
pg 1416@xmath201256 & 0.129 & 16.88 & 650 & 16.35 & @xmath2023.07 + he 1419@xmath201048 & 0.265 & 16.77 & 722 & 16.36 & @xmath2024.58 +
the investigated area was defined by 33 eso schmidt fields located in the region @xmath21 and @xmath22 , thus all in the north galactic hemisphere .
a list of the survey fields and their properties is given in table [ tab : fields ] .
in each field , one direct plate from the eso ( b ) atlas , and one he s spectral ( objective prism ) plate were available .
the formal total area subtended by 33 eso plates is @xmath23 .
losses of usable area occurred because ( 1 ) plates overlapped in adjacent fields , especially if not precisely centred ; ( 2 ) occasionally direct and spectral plates were positionally mismatched , and only the common area could be used ; ( 3 ) most important , overlapping spectra ( and to a lesser degree , direct images ) rendered a certain percentage of spectra unprocessable .
these losses have been quantified and incorporated into ` effective areas ' @xmath16 for each field , given in table [ tab : fields ] . because of the field - to - field variations in limiting magnitude , the total effective survey area is a function of apparent magnitude , as plotted in fig .
[ fig : effarea ] .
the maximum area is 611deg@xmath0 for @xmath24 and decreases gradually to zero for sources fainter than @xmath25 . while many quasar surveys were carried out at very high galactic latitudes where foreground extinction corrections are small ( although systematically non - zero ) , this is not the case for our present set of fields .
we have used the data from stark et al .
( @xcite ) to obtain the neutral hydrogen column density in the centre of each field , and converted this into a @xmath17 band extinction using the formula @xmath26 where @xmath14 is given in units of @xmath27 ( cf .
spitzer @xcite ) ; table [ tab : fields ] lists @xmath14 and @xmath15 for each field .
the average extinction is 0.36mag .
the limiting brightness of qso candidates during the automated selection procedure was initially defined by a minimum s / n ratio of @xmath28 in the digital spectra .
we later found , as expected ( cf .
paper i ) , that close to the selection limit the object lists became incomplete .
several experiments showed that systematic incompleteness occurred only for the faintest spectra , and that the effect disappeared for a limiting s / n @xmath29 5 .
we translated this s / n cutoff into a magnitude limit on the direct plate , using the calibration of photographic magnitudes described below .
for the 33 fields , the values for @xmath13 range between 17.66 and 16.64 , depending largely on the quality of the spectral plates . altogether , 115 qsos and seyfert 1 galaxies were detected by the he s selection criteria in these fields ( primarily different measures of uv excess ; cf .
paper i ) that were either confirmed by our own follow - up spectroscopy ( paper ii ) , or previously catalogued as qso or seyfert 1 in the compilation of vron - cetty & vron ( @xcite ) .
note that in this paper we deliberately ignore the traditional subdivision seyferts vs. ` real ' qsos based on some arbitrary luminosity threshold .
a lower redshift limit of 0.07 was applied to avoid possible incompleteness due to host galaxy contamination ( but see the @xmath30 seyfert sample below ) .
several of the lowest redshift objects would nevertheless be called seyferts by many others .
there remained 48 qsos with @xmath31 located within the effective survey area and above the completeness limiting magnitude .
these form the flux - limited sample listed in table [ tab : qsos ] .
a comparison with the vron - cetty & vron ( @xcite ) catalogue revealed no additional qsos that should have been included in the flux - limited sample .
this test for ` completeness ' is , however , not very strong , as few other surveys have covered this region of the sky , and @xmath32% of the sample objects are new discoveries made by the he s .
llllllllll name & @xmath5 & @xmath17 & field & @xmath18 & @xmath33 & @xmath34 & @xmath35 & @xmath36 & @xmath37 + he 1043@xmath201346 & 0.0669 & 16.83 & 639 & 16.47 & 0.0692 & 0.902 & @xmath2021.55 & @xmath2021.06 & @xmath2020.44 ' '' '' + he 1248@xmath201357 & 0.0144 & 15.28 & 645 & 15.01 & 0.0152 & 0.855 & @xmath2019.67 & @xmath2019.39 & @xmath2018.06 + iras 1249@xmath2013 & 0.0129 & 14.91 & 645 & 14.64 & 0.0254 & 0.132 & @xmath2019.80 & @xmath2019.10 & @xmath2018.99 + r 12.01 & 0.0463 & 15.50 & 646 & 15.21 & 0.0700 & 0.289 & @xmath2022.00 & @xmath2020.90 & @xmath2021.52 + pg 1310@xmath201051 & 0.0337 & 15.64 & 719 & 15.46 & 0.0544 & 0.238 & @xmath2021.07 & @xmath2019.54 & @xmath2020.77 +
he 1330@xmath201013 & 0.0221 & 15.82 & 719 & 15.65 & 0.0293 & 0.428 & @xmath2019.96 & @xmath2018.96 & @xmath2019.41 + r 14.01 & 0.0408 & 14.68 & 648 & 14.24 & 0.0700 & 0.196 & @xmath2022.70 & @xmath2020.72 & @xmath2022.51 + obtaining consistent and unbiased photometry of low - redshift quasars with detectable host galaxies is not trivial . during the survey phase ,
normally only photographic data , in our case from the digitised direct @xmath17 plates , are available .
these plates usually reach much fainter magnitudes than objective - prism plates , but they also saturate earlier , especially if not the original plates but glass copies are used ( as was the case for several of our fields ) .
the conventionally employed isophotal or large aperture magnitudes , although well suited for point sources , tend to give much larger total luminosities for qsos with detectable hosts than desired for an investigation of the _ nuclear _ luminosity function .
additional complications can result from non - linear filtering effects inherent in the digitisation procedure .
for the present sample we have developed a two - step procedure to obtain unbiased and survey - consistent magnitudes .
the photographic plates were calibrated by ccd sequences obtained with the eso 90 cm telescope in 1993 and 1994 .
these sequences will be published in a separate paper .
the sample selection was based on simulated diaphragm photometry with an aperture of the size of the seeing disk . for point sources , an accurate calibration relation with very little intrinsic scatter , @xmath38 mag ,
could be derived from the ccd sequences ( which contained only stars ) .
these small - aperture magnitudes correlated also well with the corresponding s / n ratios of the digital spectra , important for an accurate definition of limiting magnitudes . in a second step
, we obtained ccd photometry for most of the sample qsos .
exposures in the @xmath17 band were available from the mandatory acquisition images at the eso 3.6 m and 2.2 m telescopes ( cf .
paper ii ) . after standard reduction and colour term corrections , the photometric zero points of these images
were shifted to the system of the corresponding photographic plate by using surrounding unsaturated field stars as references .
the resulting corrected ccd magnitudes , again measured in an aperture of diameter approximately equal to that of the seeing disk , should be much more accurate than the photographic data , at least for extended objects , while magnitudes of point sources should be more or less unchanged .
we tested this assertion by comparing photographic and zero - point corrected ccd photometry of the @xmath39 qsos ( which should be true point sources ) and found a mean difference of only 0.01mag .
for the further analysis we used the zero - point corrected @xmath17 band ccd magnitudes . for three objects no ccd data were available , and for those we substituted the photographic measurements ( cf .
table [ tab : qsos ] ) . to obtain further insight into the continuity of properties between high - luminosity qsos and low - luminosity seyfert nuclei
, we conducted a dedicated subsurvey for seyfert galaxies , targeting both type 1 and type 2 objects .
the criteria were different from those used in the main survey , adapted to discriminate between active and inactive galaxies rather than between stars and qsos .
more details about this subsurvey will be given in a later paper ( khler et al .
, in prep . ) where we shall also present our results concerning seyfert 2 galaxies , which we do not consider here .
we found 7 seyfert 1 galaxies with @xmath40 , listed in table [ tab : seyferts ] .
the redshifts were remeasured in new spectra obtained with the eso 1.52 m telescope , and corrected for heliocentric motion .
not all the fields from table [ tab : fields ] were involved in this subsurvey , and the effective area was only 477 deg@xmath0 ( but see sect . [ sec : sy1lf ] for the computation of space densities ) .
for these sources , proper photometry is at least as important as for qsos .
the photometric systems of many earlier investigations of the luminosity function of seyferts suffer from two severe drawbacks : ( i ) they were often based on heterogeneous collections of published ( aperture ) photometry , generally inconsistent with the magnitudes used for defining the survey flux limits .
( ii ) the contributions of host galaxies were either not corrected for at all , or some assumptions about universal intrinsic properties of hosts and nuclei had to be made , such as colours ( sandage @xcite ) or host luminosities ( cheng et al .
@xcite ) . on the assumption that seyfert 1 galaxies are the low - luminosity equivalent of quasars , the quantity of interest is the _ nuclear _ brightness , with the host galaxy contribution removed . although our small - aperture magnitudes were already much less affected by galaxy contamination than large - aperture or total brightness measurements , a major contribution could still be expected especially for the lowest nuclear luminosities .
we therefore estimated these corrections individually , based on @xmath17 and @xmath41 images of the seven sample objects taken with the eso 90 cm telescope .
we first constructed empirical two - dimensional point - spread functions from nearby stars .
these were then subtracted from the seyfert images , scaled by a factor chosen such that the residual galaxy surface brightness distribution did not decrease inwards . the resulting host galaxy brightness was finally integrated over the aperture and subtracted from the total aperture flux . as can be seen in table [ tab : seyferts ] , the corrections were not negligible , and expectedly
largest for the lowest redshift sources .
an important diagnostic tool for qso samples is the empirical relation between source fluxes and the number of sources per unit solid angle .
once the field properties and the sample photometry are complete , the cumulative surface density of qsos brighter than magnitude @xmath17 , @xmath42 , can be easily computed from summing over @xmath43 for all relevant sample objects . from our qso sample with @xmath44 ( 47 objects ) we have derived the @xmath42 relation shown as thick line in fig .
[ fig : surfdens ] . note that the abscissa values in this diagram are extinction - corrected magnitudes .
the bright end of the curve is dominated by the extremely luminous source he 1029@xmath201401 , but the main part can be well approximated by a simple power law of slope @xmath45 ( @xmath46 ) . if only the 33 sources with @xmath47 are considered , excluding thereby all qso / seyfert borderline cases ( thin line in fig .
[ fig : surfdens ] ) , the normalisation drops slightly but the slope is essentially unchanged with @xmath48 . for those readers who prefer differential surface densities ,
we have computed the figures in half magnitude bins between @xmath49 and @xmath50 ( table [ tab : surfdens ] ) .
the counts were multiplied by a factor of 2 to formally express number counts per unit magnitude intervals .
comparison relations from other surveys are numerous for fainter magnitudes , but rare for @xmath51 . the palomar bright quasar survey ( bqs ; schmidt & green @xcite , hereafter quoted as sg83 ) provides the only available optically selected qso sample with a well - defined flux limit subtending over a similar region in the hubble diagram .
sg83 listed differential surface densities computed from the full set of 114 qsos with @xmath44 found over an area of 10714 deg@xmath0 .
to make their numbers comparable to ours , we recomputed the bqs surface densities by removing all @xmath30 entries from their sample , and converted the result into a cumulative relation .
this relation is shown in fig .
[ fig : surfdens ] by the open squares .
a large discrepancy at all magnitudes is apparent . at @xmath52 ,
the surface density found in the he s is high by a factor of 3.6 compared to the bqs , certainly more than can be permitted by statistical fluctuations .
= llll @xmath17 & @xmath53 & @xmath54 & @xmath55 + 15.0 & 1 & 0.003 & 0.003 ' '' '' + 15.5 & 4 & 0.013 & 0.007 + 16.0 & 5 & 0.016 & 0.008 + 16.5 & 12 & 0.046 & 0.014 + 17.0 & 11 & 0.15 & 0.08 + a similar result was already reported by goldschmidt et al .
( @xcite ) from an analysis of the edinburgh survey , who found 5 qsos above the bqs flux limit in an area of 330deg@xmath0 ( non - overlapping with the present he s area ) while the bqs contained only one .
their estimate of @xmath56deg@xmath57 for @xmath58 is completely consistent with our corresponding value of @xmath59 .
furthermore , goldschmidt et al . found a zero - point offset of 0.28mag between the bqs photographic magnitudes and theirs , in the sense that the palomar measurements were systematically _ too bright _
, thus increasing rather than relaxing the discrepancy .
it should be noted that neither sg83 nor goldschmidt et al.corrected their magnitudes for galactic extinction , whereas we have done so .
however , our survey area is _ predominantly _ located in regions of rather high @xmath15 , while sg83 estimated that the extinction for most bqs quasars were ` not much more than 0.1mag ' ; this even might be , at least partly , compensated by the above mentioned zero - point offset . because of these uncertainties we have not attempted to correct the bqs results for extinction .
one can estimate globally that an average @xmath15(bqs ) of 0.61mag would be necessary to bring bqs and he s surface densities to a match a totally unrealistic value . at fainter magnitudes
the agreement of he s number counts with those of other surveys is excellent . as an example , we have plotted the values from the lbqs ( hewett et al .
@xcite ) into fig .
[ fig : surfdens ] , with the abscissa shifted by + 0.1mag to transform their @xmath60 photometric system approximately into @xmath17 magnitudes .
while the counts in the brightest bins are again lower than ours , this can be understood as a result of poor statistics in the lbqs for @xmath51 , possibly arising from saturation effects . around @xmath61 , he s and lbqs surface densities join smoothly , without any detectable significant offset .
similarly good agreement is reached with other surveys , cf . the recent compilation by cristiani et al .
we are thus confident that the photometric scale used in the present investigation is essentially unbiased and adequate for luminosity function work .
for all objects in table [ tab : qsos ] , we first estimated absolute magnitudes for the rest - frame @xmath17 band by computing their luminosity distances in an expanding friedmann universe with the formula of terrell ( @xcite ) , and applying @xmath62 corrections taken from cristiani & vio ( @xcite ) .
as low - redshift subsample we selected the 20 qsos with @xmath1 .
space densities were derived using the generalised @xmath63 estimator ( felten @xcite ; avni & bahcall @xcite ) , thus incorporating the full information about field - dependent flux limits , including fields with zero detections ( the ` coherent ' analysis in the terminology of avni & bahcall ) .
the resulting cumulative luminosity function is shown in fig .
[ fig : lqlf ] . locations of individual objects are indicated by the error bars .
we have chosen the cumulative representation @xmath64 as it allows one to avoid binning ( particularly problematic for smaller samples ) and to make the contribution of each object in the sample apparent .
the errors were estimated from poisson statistics in the sample , combined with the uncertainties in volume determination resulting from the photometric errors . restricting the he s luminosity function to @xmath65 ( 14 objects ) results in slightly larger error bars , but the space densities change only insignificantly , by less than 0.05dex , and , overplotted in fig .
[ fig : lqlf ] , could not even be distinguished from the @xmath1 relation .
we thus feel justified to neglect differential evolution within our redshift shell ; we shall further investigate and qualify this assumption below .
up to now , the bqs provides the only comparison sample in the region @xmath66 and @xmath1 , containing 45 objects that fulfill also the condition @xmath31 . surprisingly , there is yet to our knowledge no _ direct _ determination of the local qlf based on bqs objects alone .
we have therefore computed a qlf from this sample with the same methods used for our own data , taking magnitudes , survey limits , and areas from sg83 .
we show the results of these computations as a cumulative distribution in fig .
[ fig : lqlf ] .
( since all other literature data are available only in binned form , we have also binned the bqs lf in @xmath19 , to achieve a more uniform presentation . )
as mentioned above , galactic extinction is neglected for these data ; at any rate , the effect should be very small .
the space densities found from the he s data are much higher than those derived from the bqs , with the discrepancy increasing with luminosity .
for the most luminous sources , we find that there are almost an order of magnitude more low - redshift quasars per unit volume than discovered by the bqs .
possible reasons for this discrepancy are discussed in sect .
[ sec : discussion ] . to extend the local luminosity function towards fainter levels of nuclear activity ,
we have used the additional seyfert 1 sample with @xmath30 .
the absolute magnitudes were computed in the approximation of euclidean static geometry and neglecting @xmath62 corrections . to obtain reliable and unbiased estimates for the accessible volumes ,
the maximum redshift @xmath67 at which object @xmath68 would still have been included in the survey had to be determined for each object . because of the superposition of a host galaxy with constant intrinsic scale length and a point - like nucleus , the small - aperture magnitudes depended on redshift in a complicated way , and @xmath67 was therefore not only a function of @xmath17 and @xmath13 but also of intrinsic properties .
we used the available ccd images to simulate the mixing of agn / host contributions into the aperture as a function of redshift @xmath5 , resulting in an array of @xmath69 values , one for each field @xmath70 . for simplicity , table [ tab : seyferts ] supplies ` effective ' values @xmath33 that permit a direct conversion into space densities .
a @xmath71 test ( schmidt @xcite ) , implemented in the generalisation proposed by avni & bahcall ( @xcite ) , gives a mean value of @xmath72 ( individual values are also listed in table [ tab : seyferts ] ) .
compared with the expectation value of 0.5 for a complete , non - evolving , homogeneously distributed sample , there is no evidence for major incompleteness .
the resulting luminosity function is displayed in fig .
[ fig : sy1lf ] .
it can be seen that a host galaxy correction is important particularly for the intrinsically faintest objects in the sample : without correction , the luminosity function shows a steep upturn at @xmath73 ; for these objects , even the small - aperture magnitude is dominated by the host rather than by the agn .
although the sample is not large , the derived space densities are fully consistent with previous estimates ( cheng et al . @xcite ; huchra & burg @xcite ) .
however , none of these previous investigations have payed as much attention to an appropriate host galaxy correction , neither for source luminosities nor for the survey volume , as we have .
in particular , the analysis of the cfa sample by huchra & burg was based on uncorrected zwicky magnitudes ( essentially measuring the _ total _ galaxy brightness ) , making a straightforward comparison impossible . more similar to our approach
was the work of cheng et al .
( @xcite ) with markarian galaxies , but their sample , although much larger than ours , had to be manipulated with substantial incompleteness corrections , in addition to the afore mentioned problems of a heterogeneous photometric data base .
= llll @xmath19 & @xmath53 & @xmath74 & @xmath75 + -18.5 & 2 & @xmath76 & @xmath77 ' '' '' + -19.5 & 1 & @xmath78 & @xmath79 + -20.5 & 2 & @xmath80 & @xmath81 + -21.5 & 2 & @xmath82 & @xmath83 + -22.5 & 8 & @xmath84 & @xmath85 + -23.5 & 3 & @xmath86 & @xmath87 + -24.5 & 6 & @xmath88 & @xmath89 + -25.5 & 2 & @xmath90 & @xmath91 + -26.5 & 1 & @xmath92 & @xmath93 + the local universe is the only domain where it is possible to construct the entire qlf , including low - luminosity tail , directly from one single survey .
the local luminosity function of qsos and seyfert 1 nuclei , obtained by combining the two independently derived luminosity functions presented above , is valid between absolute magnitudes @xmath94 and @xmath95 , thus spanning three decades in luminosity . overlaying the cumulative relations of figs .
[ fig : lqlf ] and [ fig : sy1lf ] , the match in the common luminosity region is remarkably good ( cf . fig . [ fig : syqlf ] ) .
to arrive at a consistent system of _ nuclear _ luminosities , the magnitudes in the qso sample were corrected by subtracting a template host galaxy of @xmath96 . since the magnitudes had already been measured through a small aperture , this host subtraction was a minor correction for all objects in the qso sample , and the results are not sensitive to the assumed host luminosity .
the combined local qlf can be fitted well by a simple linear relation between @xmath19 and @xmath97 , corresponding to a single power - law for the _ differential _ qlf ( @xmath98 ) .
a slope parameter @xmath99 provides a statistically acceptable fit over the full range of 8 magnitudes . fitting only the qso sample ( @xmath31 )
does not alter the slope significantly ( @xmath100 ) ; fitting only the @xmath30 seyferts yields @xmath101 , again consistent with a constant slope of @xmath102 .
the slope of the bqs lf is much steeper : apart from flattening at both high and low luminosities , the data demand @xmath103 over the most of the range . from the combined dataset ,
we have also produced a binned differential luminosity function @xmath104 , tabulated in steps of one in absolute magnitude ( table [ tab : lqlf ] ) .
as was to be expected , this representation is quite sensitive to the actual choice of binning intervals , most obvious in the succession of a ` high ' and a ` low ' bin at @xmath105 and @xmath106 , respectively .
there is , however , full consistency of the results obtained ; in particular , the qlf slope parameter @xmath107 is reproduced within the expected uncertainties .
= the sample of table [ tab : qsos ] contains eight qsos with redshifts @xmath108 that we used to further constrain the potential error introduced by neglecting differential evolution . binning the full @xmath109 sample into two redshift regimes @xmath110 and @xmath47 , respectively
, there are 14 objects in each shell .
the luminosity functions are compared in fig .
[ fig : evolqlf ] : for @xmath111 , the local and the adjacent @xmath112 shell have essentially identical space densities , while for @xmath113 , the qlf in @xmath47 seems to be merely a continuation of the local qlf to higher luminosities , _ with the same slope and normalisation_. dividing the sample at @xmath114 instead of 0.2 does not affect the results . outside the local universe , a single survey can yield no more than a small segment of the qlf , and only the combination of many different surveys can provide similar coverage of source luminosities as in the local qlf .
the he s has the potential to provide new samples covering the brightest parts of the known qso population , and it will be a promising task for the future to combine our results with those of fainter surveys .
such an analysis is , however , not the scope of the present paper .
we only remark that even at redshifts higher than @xmath115 , the qlf pieces sampled by the he s do not show significant deviations from an extrapolation of the local relation .
a more quantitative discussion of this issue is deferred to a later paper .
our results imply that the palomar bright quasar survey is not only incomplete , but also heavily biased .
the general incompleteness may be estimated from the surface densities to be around a factor of 23 .
this effect has been noted by others before ( wampler & ponz @xcite ; goldschmidt et al .
@xcite ) , and we have at present no easy explanation for it . some mild adjustment of the bqs data to higher values of space densities and luminosities may be required because of neglected galactic extinction , but this can not account for the observed discrepany .
one possibility is that the rather large photometric errors in the bqs , reflected in the @xmath116 colours , scattered many qsos out of the uv excess domain .
there seems to be an additional deficit of luminous low - redshift qsos , by another factor of @xmath117 for @xmath118 .
this discrepancy is apparent already from simply counting the number of @xmath1 qsos more luminous than @xmath119 : there are 5 objects in the present he s sample , while the entire bqs contains only 15 such sources , in a more than 15 times larger area . as
this is the first detection of such an effect , one has to carefully consider systematic errors .
note , however , that the he s is the first optical qso survey since completion of the bqs that is sensitive to a similar range of redshifts and luminosities , and it is therefore not altogether inconceivable that substantial selection biases have remained unnoticed over a long time .
the good agreement between the he s @xmath47 or @xmath120 surface densities with other major surveys , e.g. , the lbqs ( hewett et al .
@xcite ) , the edinburgh survey ( goldschmidt et al .
@xcite ) , or the hbqs ( cristiani et al .
@xcite ) , makes it implausible that the he s photometric zeropoint could be seriously in error .
we also tested our method of zero - point corrected ccd photometry .
all relevant quantities were recomputed based on the standard approach of using only photographic magnitudes measured over large apertures , without any systematic change of results , except that the luminosities of low - redshift qsos with bright host galaxies were often _ over_estimated in the photographic data , creating a bias towards even higher qlf values .
the ccd approach thus seems justified and appropriate .
the reconstruction of the @xmath121 luminosity function for the bqs was based on the same procedures as applied to our own data .
we neglected differential evolution within the @xmath1 shell , but the _ observed _ discrepancy is independent of any assumed evolution law , and therefore most likely caused by real incompleteness in the bqs sample .
the exclusion of objects of non - stellar appearance in the bqs is an intuitively appealing candidate mechanism for this additional incompleteness .
it is obvious at first glance ( and has in fact never been disputed ) that the bqs sample is systematically lacking low - luminosity seyferts ; it has usually been assumed that this selection effect ceased to be valid for @xmath122 .
our results suggest that this is not the case . however , it is then somewhat surprising that the incompleteness should be smallest , almost disappearing , in the luminosity range around @xmath123 , as seems to follow from fig .
[ fig : lqlf ] .
in fact , this would then be the _ only _ region of the entire hubble diagram where the bqs counts were not significantly below those of the he s . increased incompleteness for the most luminous qsos seems a possible , but rather unlikely explanation . if , instead , the bqs magnitudes particularly for the slightly resolved @xmath124 objects were _ systematically too bright _ , overcompleteness would occur , scattering objects into the sample from below the flux limit . we know from our own experience with photometry of resolved objects on digitised photographic plates that such effects exist , especially if a large digitisation aperture is used ( this is why we decided to incorporate the ccd measurements ) .
the bqs was based on pds scans with an aperture of @xmath125 ( green & morrill @xcite ) , and magnitudes were basically isophotal in photographic density .
we _ suspect _ that these technical details provide a clue to the above contradiction , and that incompleteness due to the rejection of extended qso hosts could be partly counteracted by photometric bias and overcompleteness for only marginally resolved objects .
further analysis of this point is certainly required .
if our hypothesis is correct , two important conjectures can be made : ( 1 ) the slope of the qlf for @xmath126 is much flatter than previous investigators estimated , weakening the evidence for a ` break ' in the seyfert / qso transition regime ( e.g. , marshall @xcite ) the local qlf presented here showing essentially _ no _ break at all .
note that also in the seyfert lf of cheng et al .
( @xcite ) , the presence of such a break is not really demanded by the data .
( 2 ) the bqs sample , up to now the prime source for luminous radio - quiet low - redshift qsos , might suffer from severe selection biases with respect to the morphological properties of their galaxy hosts .
the local luminosity function derived in this paper is independent of samples collected at higher redshifts , and therefore well suited to test and constrain possible evolution laws .
existing estimates of the qlf at redshift @xmath127 ( e.g. , boyle et al .
@xcite ) show a marked break , or change of slope , at intermediate luminosities , that shifts towards lower values with decreasing @xmath5 .
we have argued above that the local equivalent of this break is probably an artefact .
the flat slope of the @xmath6 qlf derived in this paper , and the absence of clear features , demand , on the assumption that the high @xmath5 qlf determinations are correct , that both shape and slope of the qlf change strongly with redshift .
this clearly is in contradiction to the ` standard picture ' of pure luminosity evolution ( ple ) , where a double power - law luminosity function with invariant logarithmic shape is merely shifted along the luminosity axis .
however , it should be noted that virtually _ all _ evolution modelling since 1983 had to rely on the bqs data of sg83 , to cover an otherwise inaccessible part of the hubble diagram .
hence , in these regions , the models merely reproduce the incompleteness and possible biases of the bqs contribution .
the problem that qso surveys were notoriously incomplete at low redshifts was of course realised by most researchers , and it has become customary to apply a lower redshift cutoff of @xmath128 to all samples .
the construction of a @xmath129 qlf from such models has therefore always the character of an _ extrapolation _ outside the validated range , and should be treated with appropriate caution .
even so , a _ quantitative _ comparison between evolutionary models and actual measurements of the local seyfert 1/qlf is a powerful test which deserves higher attention than it has been paid to in the past , compared to the careful analyses of higher redshift data . in the construction of the @xmath1 local qlf
we have not corrected individual luminosities for statistical evolution .
while it is true that such a neglect could artificially flatten the luminosity function , evolution must be fairly strong for the effect to be relevant .
we show now that the no - evolution assumption was , in fact , the best approximation one could possibly make , over the range of luminosities covered by the he s . in the case of evolution ,
the lfs of adjacent , non - overlapping redshift shells should be different . for the simple case of ple with the rate suggested ,
e.g. , by boyle et al .
( @xcite ) , one would expect an abscissa offset of @xmath130mag between @xmath65 and @xmath112 . the necessity to invoke such an offset
is not apparent from fig .
[ fig : evolqlf ] . while it might be consistent with our data to allow for a somewhat steeper intrinsic slope of the qlf , combined with mild evolution , there is certainly no _ evidence _ for evolution in the data .
these conclusions are supported by results reported recently from other surveys .
hewett et al .
( @xcite ) , with a preliminary analysis of the lbqs , showed that the qlf flattens considerably at @xmath131 ; a similar conclusion was reached by miller et al .
( @xcite ) using the edinburgh survey .
although both surveys have the usual lower redshift limit excluding the local population ( @xmath120 for the edinburgh survey , @xmath47 for the lbqs ) , a prediction for the slope of the local qlf can be made by extrapolating the trend seen by miller et al . ;
the result is consistent with our measured value of @xmath4 .
a flatter qlf implies a higher fraction of luminous quasars among the total population .
our results indicate therefore , in agreement with hewett et al . and miller et al .
, that the evolution of the most luminous quasars must proceed considerably slower than suggested by the notion of pure luminosity evolution .
we have analysed a new , flux - limited and well - defined sample of qsos & seyfert 1 galaxies .
host - galaxy dependent selection and photometric biases , if not absent , are greatly reduced in comparison to other optically selected samples .
we find that the derived space density of luminous qsos in the local universe is much higher than previous surveys indicated , and that consequently these objects show a much slower cosmological evolution . with the present sample of 55 qsos and seyferts in 33 fields , only a small fraction of the he s area is covered .
we have already acquired , and partly processed , much additional plate material enlarging the survey by a significant factor . for example , we now have several fields in common with the bqs , and it shall soon be possible to pursue the question of incompleteness and biases in the bqs by a direct comparison of objects and photometry . to understand and model
the process of quasar evolution requires strong constraints from observations .
it seems now that we are farther away from a coherent empirical picture of the qso population and its evolution than thought a few years ago .
quasars of different luminosities evolve differently , and quasars do not necessarily evolve with the same rate at different cosmological epochs .
there are presumably more parameters needed to describe quasar evolution properly than in the simple model of ` pure luminosity evolution ' .
the ultimate aim of qso survey work is the deduction of intrinsic physical properties from statistical samples .
unfortunately the physical processes in active galaxy nuclei are poorly understood , and the relations between statistical and physical properties therefore far from unique .
the best study cases for nuclear activity in galaxies are nearby seyferts ; as shown above , the luminosity function of seyfert 1 nuclei can be smoothly continued into the classical quasar regime , without a significant change of slope or break .
while this by no means proves that seyfert 1 nuclei are simply scaled - down versions of quasars , it suggest a continuity of properties rather than distinct classes . with the completion of the he s
, a large and well - defined sample of highly luminous low - redshift qsos will become available , to study the relationship between qsos , their environments , and their evolution in detail .
avni y. , bahcall j.n .
, 1980 , apj 235 , 694 boyle b.j .
, jones l.r .
, shanks t. , 1991 , mnras 251 , 482 cheng f .- z . , danese l. , de zotti g. , franceschini a. , 1985 , mnras 212 , 857 cristiani s. , vio r. , 1990 , a&a 227 , 385 cristiani s. , la franca f. , andreani p. , et al . , 1995 , a&as 112 , 347 felten j.e . , 1976 , apj 207 , 700 goldschmidt p. , miller l. , la franca f. , cristiani s. , 1992 , mnras 256 , 65 green r.f . , morrill m.e . , 1978 ,
pasp 90 , 601 hewett p.c . ,
foltz c.b . ,
chaffee f.h . , 1993 , apj 406 , l43 hewett p.c .
, foltz c.b . ,
chaffee f.h . , 1995 , aj 109 , 1498 huchra j. , burg r. , 1992 , apj 393 , 90 khler t. , 1996 , doctoral thesis , universitt hamburg marshall h. , 1987 , aj 94 , 628 miller l. , goldschmidt p. , la franca f. , cristiani s. , 1993 , observational cosmology , eds .
g. chincarini et al . , asp conf .
51 , 614 reimers d. , khler t. , wisotzki l. , 1996 , a&as 115 , 235 ( paper ii ) sandage a. , 1973 , apj 180 , 687 schmidt m. , 1968 , apj 151 , 393 schmidt m. , green r.f . , 1983 , apj 269 , 352 ( sg83 ) spitzer l. , 1978 , physical processes in the interstellar medium , wiley&sons stark a.a . , gammie c.f . ,
wilson r.w . ,
et al . , 1992 , apjs 79 , 77 terrell j. , 1977 , am .
j. phys .
45 , 869 vron - cetty m .- p . , vron p. , 1993
, a catalogue of quasars and active galactic nuclei , 6th ed . ,
eso scientific report 13 wampler e.j .
, ponz d. , 1985 , apj 298 , 448 wisotzki l. , khler t. , groote d. , reimers d. , 1996 , a&as 115 , 227 ( paper i ) | we present the analysis of a new flux - limited sample of bright quasars and seyfert 1 galaxies in an effective area of 611deg@xmath0 , drawn from the hamburg / eso survey .
we confirm recent claims that bright quasars have a higher surface density than previously thought .
special care was taken to avoid morphological and photometric biases against low - redshift quasars , and about 50% of the sample objects are at @xmath1 , spanning a range of @xmath2 in absolute magnitudes . while our derived space densities for low - luminosity seyfert 1 nuclei are consistent with those found in the literature , we find that luminous qsos , @xmath3 , are much more numerous in the local universe than previous surveys indicated . the optical luminosity functions of seyfert 1 nuclei and qsos join smoothly , and if the host galaxy contributions are taken into account , a single power - law of slope @xmath4 describes the combined local luminosity function adequately , over the full range in absolute magnitude .
comparing our data with published results at higher redshifts , we can rule out pure luminosity evolution as an acceptable parametrisation ; the luminosity function of quasars changes shape and slope with @xmath5 , in the sense that the most luminous quasars show the weakest evolution . |
it is known that the superconformal field theory possesses two integrable structures @xcite . in the previous paper @xcite
we have considered the quantum super - kdv @xcite-@xcite hierarchy , which give rise to the infinite number of commuting integrals of motion ( i m ) , constructed via the generators of the superconformal algebra and therefore leading to one of these integrable structures .
but this set of i m is not invariant under the supersymmetry transformation .
+ here we consider the susy n=1 kdv hierarchy @xcite , @xcite which is a supersymmetric generalization of the kdv one , in the case of this model the supersymmetry generator is included in the commuting family of i m @xcite .
the outline of the paper is as follows . in the first part ( sec .
2 ) we consider the classical theory of susy n=1 kdv system , based on the twisted affine superalgebra @xmath0 .
we introduce the supersymmetric miura transformation and monodromy matrix associated with the corresponding l - operator
. then the auxiliary @xmath1-matrices are constructed , which satisfy the quadratic poisson bracket relation . as we will show in sec .
4 quantum counterparts of these matrices coincide with a vertex - operator - represented quantum r - matrix . the quantum version of the miura transformation i.e. the free field representation of the superconformal algebra is given in sec . 3 . in sec .
4 the quantum @xmath2 superalgebra @xcite is introduced .
then it is shown that the corresponding quantum r - matrix can be represented by two vertex operators , satisfying serre relations of lower borel subalgebra of @xmath2 and in the classical limit it coincides with @xmath1-matrix .
the vertex - operator - represented quantum r - matrix @xmath3 satisfies the so - called rtt - relation , which give us possibility to consider the model from a point of view of quantum inverse scattering method ( qism ) @xcite , @xcite .
this could be well applied ( sec.5 ) to the study integrable perturbed superconformal theories with supersymmetry unbroken , arising in the physics of 2d disordered systems , lattice models ( e.g. the tricritical ising model ) , and in the superstring physics ( e.g. supersymmetric d - branes ) @xcite-@xcite .
the susy n=1 kdv system can be constructed by means of the drinfeld - sokolov reduction applied to the @xmath4 twisted affine superalgebra @xcite .
the corresponding @xmath5-operator has the following form : @xmath6 where @xmath7 is a superderivative , the variable @xmath8 lies on a cylinder of circumference @xmath9 , @xmath10 is a grassmann variable , @xmath11 is a bosonic superfield ; @xmath12 , @xmath13 , @xmath14 , @xmath15 are the chevalley generators of c(2 ) with the following commutation relations : @xmath16=0,\quad [ h_1,{e_2}^{\pm}]=\pm{e_2}^{\pm } , \quad [ h_2,{e_1}^{\pm}]=\pm{e_1}^{\pm},\\ & & ad^{2}_{{e_1}^{\pm}}{e_2}^{\pm}=0 , \quad ad^{2}_{{e_2}^{\pm}}{e_1}^{\pm}=0 , \nonumber\\ & & [ h_{\alpha},{e_{\alpha}}^{\pm}]=0 \quad(\alpha = 1,2),\quad [ { e_{\beta}}^{\pm},{e_{\beta'}}^{\mp}]=\delta_{\beta , \beta'}h_{\beta}\quad ( \beta , \beta ' = 1,2),\nonumber\end{aligned}\ ] ] where the supercommutator [ , ] is defined as follows : @xmath17\equiv ad_{a } b\equiv ab- ( -1)^{p(a)p(b)}ba$ ] , where parity @xmath18 is equal to 1 for odd elements and is equal to 0 for even ones . in the particular case of c(2 ) @xmath19 are even and @xmath20 are odd .
the operator ( 1 ) can be considered as more general one , taken in the evaluation representation of @xmath4 : @xmath21 where @xmath22 , @xmath23 , @xmath24 are the chevalley generators of @xmath4 with such commutation relations : @xmath25=0,\quad [ h_{\alpha_0},e_{\pm\alpha_1}]= \mp e_{\pm\alpha_1},\quad [ h_{\alpha_1},e_{\pm\alpha_0}]= \mp e_{\pm\alpha_0},\\ & & [ h_{\alpha_i},e_{\pm\alpha_i}]=\pm e_{\pm\alpha_i},\quad [ e_{\pm\alpha_i } , e_{\mp\alpha_j}]=\delta_{i , j}h_{\alpha_i},\quad ( i , j=0,1 ) , \nonumber\\ & & ad^{3}_{e_{\pm\alpha_0 } } e_{\pm\alpha_1}=0 , \quad ad^{3}_{e_{\pm\alpha_1 } } e_{\pm\alpha_0}=0\nonumber\end{aligned}\ ] ] where @xmath26 , @xmath27 and @xmath28 , @xmath29 .
the poisson brackets for the field @xmath30 , obtained by means of the drinfeld - sokolov reduction are : @xmath31 and the following boundary conditions are imposed on the components of @xmath30 : @xmath32 , @xmath33 . the @xmath34-operator is written in the miura form , making a gauge transformation one can obtain a new superfield @xmath35 , where @xmath36 and @xmath37 generate the superconformal algebra under the poisson brackets : @xmath38 using one of the corresponding infinite family of i m , which are in involution under the poisson brackets ( these i m could be extracted from the monodromy matrix of @xmath39-operator , see below)@xcite : @xmath40 one can obtain an evolution equation ; for example , taking @xmath41 we get the susy n=1 kdv equation @xcite : @xmath42 and in components : @xmath43 , @xmath44 as we have noted in the introduction one can show that the i m are invariant under supersymmetry transformation generated by @xmath45 .
+ in order to construct the so - called monodromy matrix we introduce the @xmath39-operator , equivalent to the @xmath34 one : @xmath46 the equivalence can be easily established if one considers the linear problem associated with the @xmath34-operator : @xmath47 ( we consider this operator acting in some representation of @xmath4 and @xmath48 is the vector in this representation ) .
then , expressing @xmath49 in components : @xmath49=@xmath50 , we find : @xmath51 and @xmath52 .
+ the solution to the equation @xmath51 can be written in the following way : @xmath53 \big)\eta,\nonumber\end{aligned}\ ] ] where @xmath54 is a constant vector in the corresponding representation of @xmath4 .
therefore we can define the monodromy matrix in the following way : @xmath55 \big).\nonumber\end{aligned}\ ] ] introducing then ( as in @xcite ) the auxiliary @xmath1-operators : @xmath56 we find that in the evaluation representation ( when @xmath57 , the spectral parameter appears ) the following poisson bracket relation is satisfied @xcite : @xmath58,\end{aligned}\ ] ] where @xmath59 is trigonometric @xmath4 r - matrix @xcite . from this relation one
obtains that the supertraces of monodromy matrices @xmath60 commute under the poisson bracket : @xmath61 . expanding @xmath62 in @xmath57 in the evaluation representation corresponding to the defining 3-dimensional representation of @xmath63 @xmath64
we find : @xmath65 where @xmath66 .
so , one can obtain the i m from the supertrace of the monodromy matrix . using the poisson bracket relation for these supertraces with different values of spectral parameter
( see above ) we find that infinite family of i m is involutive , as it was mentioned earlier .
in this section we begin to build quantum counterparts of the introduced classical objects
. we will start from the quantum miura transformation , the free field representation of the superconformal algebra @xcite : @xmath67 where @xmath68=\frac{i}{2}\beta^2 , \quad [ a_n , a_m]=\frac{\beta^2}{2}n\delta_{n+m,0},\qquad \{\xi_n,\xi_m\}=\beta^2\delta_{n+m,0}.\nonumber\end{aligned}\ ] ] recall that there are two types of boundary conditions on @xmath69 : @xmath33 .
the sign `` + '' corresponds to the r sector , the case when @xmath69 is integer modded , the `` '' sign corresponds to the ns sector and @xmath69 is half - integer modded .
the variable @xmath70 in ( 13 ) is equal to zero in the r case and equal to 1 in the ns case .
+ one can expand @xmath71 and @xmath72 by modes in such a way : @xmath73 , @xmath74 , where @xmath75 and @xmath76 generate the superconformal algebra : @xmath77&=&(n - m)l_{n+m}+\frac{\hat{c}}{8}(n^3-n)\delta_{n ,- m } , \quad \lbrack l_n , g_m\rbrack=(\frac{n}{2}-m)g_{m+n}\nonumber\\ \lbrack g_n , g_m\rbrack&=&2l_{n+m}+\delta_{n ,- m}\frac{\hat{c}}{2}(n^2 - 1/4).\end{aligned}\ ] ] in the classical limit @xmath78 ( the same is @xmath79 ) the following substitution : @xmath80 , @xmath81 , @xmath82\to \frac{4\pi}{i\hat{c}}\{,\ } $ ] reduce the above algebra to the poisson bracket algebra of susy n=1 kdv theory .
+ let now @xmath83 be the highest weight module over the oscillator algebra of @xmath84 , @xmath85 with the highest weight vector ( ground state ) @xmath86 determined by the eigenvalue of @xmath87 and nilpotency condition of the action of the positive modes : @xmath88 where @xmath89 . in the case of the r sector
the highest weight becomes doubly degenerate due to the presence of zero mode @xmath90 .
so , there are two ground states @xmath91 and @xmath92 : @xmath93 . using the above free field representation of the superconformal algebra
one can obtain that for generic @xmath94 and @xmath18 , @xmath83 is isomorphic to the super - virasoro module with the highest weight vector @xmath86 : @xmath95 where @xmath96 in the ns sector and module with two highest weight vectors in the ramond case : @xmath97 the space @xmath83 , now considered as super - virasoro module , splits into the sum of finite - dimensional subspaces , determined by the value of @xmath98 : @xmath99 , @xmath100 .
the quantum versions of local integrals of motion should act invariantly on the subspaces @xmath101 .
thus , the diagonalization of i m reduces ( in a given subspace @xmath102 ) to the finite purely algebraic problem , which however rapidly become rather complicated for large @xmath103 .
it should be noted also that in the case of the ramond sector supersymmetry generator @xmath104 commute with i m , so i m act in @xmath91 and @xmath92 independently , without mixing of these two ground states ( unlike the super - kdv case @xcite ) .
the obtained rtt - relation ( 21 ) and corresponding quantum integrability condition give us possibility to consider the model possessing the associated i m from a point of view of qism @xcite , @xcite . in our case
the class of such systems is very wide .
it contains all the superconformal ( minimal ) models and some of their perturbations ( though in the unperturbed case scft usually provides easier methods ) .
the perturbations that do not break the integrability ( commuting with i m ) and therefore appropriate for the qism scheme are ( following the arguments of @xcite,@xcite ) @xmath197 and @xmath198 operators , corresponding to the vertex operators @xmath199 .
the topic of special interest is the scft with perturbation on the boundary related to the d - brane theory @xcite , @xcite .
+ in order to find the eigenvalues of the transfer matrices ( they are our main object of study , because their expansion in @xmath57 gives the quantum i m ) one can follow two routes . the first one , introduced in @xcite , is based on the so - called `` fusion relations '' @xcite for the transfer matrices .
in the case when @xmath200 is a root of unity ( @xmath201 and @xmath202 are rational ) , the system of fusion relations becomes finite and reduces to the thermodynamic bethe ansatz equations @xcite .
+ another approach is the baxter q - operator method , which could be applied for all values of @xmath201 ( therefore for all values of the central charge)@xcite .
the construction of the q - operator and fusion relations will be given in the paper under preparation @xcite .
we are grateful to f.a .
smirnov for useful discussions .
the work was supported by the dynasty foundation ( amz ) and rfbr grant 03 - 01 - 00593 ( ppk ) . | the supersymmetry invariant integrable structure of two - dimensional superconformal field theory is considered .
the classical limit of the corresponding infinite family of integrals of motion ( i m ) coincide with the family of i m of susy n=1 kdv hierarchy .
the quantum version of the monodromy matrix , generating quantum i m , associated with the susy n=1 kdv is constructed via vertex operator representation of the quantum r - matrix .
the possible applications to the perturbed superconformal models are discussed .
superconformal field theory , super - kdv , quantum superalgebras 11.25.hf ; 11.30.pb ; 02.20.uw ; 02.20.tw |
on 2014 january 2 , asteroid 2014 aa became the second example of an object discovered just prior to hitting the earth ( brown 2014 ; jenniskens 2014 ) . over five years before this event occurred
, similarly small 2008 tc@xmath0 had stricken our planet hours after being first spotted ( chodas et al .
2009 ; jenniskens et al . 2009 ; oszkiewicz et al .
the fireball caused by the entry of 2008 tc@xmath0 was observed by the meteosat 8 weather satellite ( borovika & charvt 2009 ) ; no images of the atmospheric entry of 2014 aa have emerged yet , although there is robust evidence for an impact over the atlantic ocean ( chesley et al .
2015 ; farnocchia et al .
2016 ) less than a day after this small asteroid was discovered .
the uncertainty associated with the pre - impact orbit of 2008 tc@xmath0 is relatively small because thousands of observations were made during the hours preceding the actual strike ( chodas et al .
2009 ; jenniskens et al . 2009 ; kozubal et al . 2011 ; oszkiewicz et al .
in sharp contrast , 2014 aa was only observed seven times before impact ( kowalski et al .
2014 ) and consequently its pre - impact orbit is somewhat uncertain ( chesley et al . 2014 , 2015 ;
farnocchia et al .
both objects hit the earth about 20 hours after they were first detected ( farnocchia et al .
asteroid 2008 tc@xmath0 completely broke up over northern sudan on 2008 october 7 ( jenniskens et al .
2009 ) ; asteroid 2014 aa probably met a similar end over the atlantic ocean .
meteorites were collected from 2008 tc@xmath0 ( jenniskens et al .
2009 ) ; any surviving meteorites from 2014 aa were likely lost to the sea .
both 2008 tc@xmath0 and 2014 aa had similar sizes of a few metres .
such small asteroids or meteoroids ( diameter @xmath15 10 m ) are probably fragments of larger objects , which may also be fragments themselves .
the study of the orbital dynamics of such fragments is a subject of considerable practical interest because small bodies dominate the risk of unanticipated earth impacts with just local effects ( brown et al .
asteroid fragmentation could be induced by collisional processes ( e.g. dorschner 1974 ; ryan 2000 ) but also be the combined result of thermal fatigue ( e.g. apek & vokrouhlick 2010 ) and rotational ( e.g. walsh et al .
2008 ) or tidal stresses ( e.g. richardson et al . 1998 ; tth et al .
the present - day rate of catastrophic disruption events of asteroids in the main belt has been recently studied by denneau et al .
these authors have found that the frequency of this phenomenon is much higher than previously thought , with rotational disruptions being the dominant source of fragments .
production of fragments can be understood within the context of active asteroids ( see e.g. jewitt 2012 ; jewitt et al .
2015 ; drahus et al . 2015 ; agarwal et al .
2016 ) . here
, we revisit the topic of the impact parameters of 2014 aa , then apply monte carlo and @xmath4-body techniques to obtain an independent determination of the pre - impact orbit of 2014 aa .
the computed orbital solution is used to investigate the existence of near - earth objects ( neos ) moving in similar orbits .
this paper is organized as follows . in sect .
2 , we review what is currently known about 2014 aa . the atmospheric entry of 2014 aa is revisited in sect .
3 ; an improved impact solution ( location and timing ) , that is based on a refined analysis of infrasound recordings , is presented .
a monte carlo technique is used in sect . 4 to estimate the most probable , in geometric terms , pre - impact orbit .
an @xmath4-body approach is described and applied in sect . 5 to obtain a more realistic orbital solution . in sect .
6 , we provide an extensive and detailed comparison with results obtained by other authors and show that all the solutions published so far are reasonably consistent .
based on the new solution , the recent past orbital evolution of 2014 aa is reconstructed in sect .
7 . a number of perhaps dynamically - related small bodies are discussed in sect . 8 .
section 9 summarizes our conclusions . | on 2008 october 7 , small asteroid 2008 tc@xmath0 turned itself into the parent body of the first meteor ever to be predicted before entering the earth s atmosphere . over five years later , the 2014 aa event became the second instance of such an occurrence .
the uncertainties associated with the pre - impact orbit of 2008 tc@xmath0 are relatively small because thousands of observations were made during the hours preceding the actual meteor airburst . in sharp contrast ,
2014 aa was only observed seven times before impact and consequently its trajectory is somewhat uncertain .
here , we present a recalculation of the impact parameters location and timing of this meteor based on infrasound recordings .
the new values (@xmath1 , @xmath2 , @xmath3 ) = ( -44 , + 11 , 2456659.618 jd utc) and their uncertainties together with monte carlo and @xmath4-body techniques , are applied to obtain an independent determination of the pre - impact orbit of 2014 aa : @xmath5 = 1.1623 au , @xmath6 = 0.2116 , @xmath7 = 1@xmath84156 , @xmath9 = 101@xmath86086 , and @xmath10 = 52@xmath83393 .
our orbital solution is used to investigate the possible presence of known near - earth objects ( neos ) moving in similar orbits . among the objects singled out by this search ,
the largest is 2013 ho@xmath11 with an absolute magnitude of 23.0 ( diameter 75169 m ) and a moid of 0.006 au . prior to impact
, 2014 aa was subjected to a web of overlapping secular resonances and it followed a path similar to those of 2011 gj@xmath0 , 2011 jv@xmath12 , 2012 dj@xmath13 , and 2013 nj@xmath14 .
neos in this transient group have their orbits controlled by close encounters with the earth
moon system at perihelion and mars at aphelion , perhaps constituting a dynamical family .
extensive comparison with other studies is also presented . |
we consider the incompressible navier
stokes equations @xmath0 in a bounded domain @xmath1 ( @xmath2 ) with a smooth boundary subject to homogeneous dirichlet boundary conditions @xmath3 on @xmath4 . in ( [ onetwo ] ) , @xmath5 is the velocity field , @xmath6 the pressure , and @xmath7 a given force field . for simplicity in the exposition we assume , as in @xcite , @xcite , @xcite , @xcite , @xcite , that the fluid density and viscosity have been normalized by an adequate change of scale in space and time .
let @xmath8 and @xmath9 be the semi - discrete ( in space ) mixed finite element ( mfe ) approximations to the velocity @xmath5 and pressure @xmath6 , respectively , solution of ( [ onetwo ] ) corresponding to a given initial condition @xmath10 we study the a posteriori error estimation of these approximations in the @xmath11 and @xmath12 norm for the velocity and in the @xmath13 norm for the pressure . to do this for a given time @xmath14 , we consider the solution ( @xmath15 , @xmath16 ) of the stokes problem @xmath17 we prove that @xmath15 and @xmath16 are approximations to @xmath5 and @xmath6 whose errors decay by a factor of @xmath18 faster than those of @xmath8 and @xmath9 ( @xmath19 being the mesh size ) . as a consequence ,
the quantities @xmath20 and @xmath21 , are asymptotically exact indicators of the errors @xmath22 and @xmath23 in the navier - stokes problem ( [ onetwo])([ic ] ) .
furthermore , the key observation in the present paper is that ( @xmath24 ) is also the mfe approximation to the solution @xmath25 of the stokes problem ( [ eq : stokes ] ) .
consequently , any available procedure to a posteriori estimate the errors in a stokes problem can be used to estimate the errors @xmath26 and @xmath21 which , as mentioned above , coincide asymptotically with the errors @xmath22 and @xmath23 in the evolutionary ns equations .
many references address the question of estimating the error in a stokes problem , see for example @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite and the references therein . in this paper
we prove that any efficient or asymptotically exact estimator of the error in the mfe approximation @xmath27 to the solution of the _ steady _ stokes problem ( [ eq : stokes ] ) is also an efficient or asymptotically exact estimator , respectively , of the error in the mfe approximation @xmath27 to the solution of the _ evolutionary _ navier - stokes equations ( [ onetwo])([ic ] ) . for the analysis in the present paper we do not assume to have more than second - order spatial derivatives bounded in @xmath28 up to initial time @xmath29 , since demanding further regularity requires the data to satisfy nonlocal compatibility conditions unlikely to be fulfilled in practical situations @xcite , @xcite .
the analysis of the errors @xmath30 and @xmath31 follows closely @xcite where mfe approximations to the stokes problem ( [ eq : stokes ] ) ( the so - called postprocessed approximations ) are considered with the aim of getting improved approximations to the solution of ( [ onetwo])([ic ] ) at any fixed time @xmath32 . in this paper
we will also refer to ( @xmath15 , @xmath16 ) as postprocessed approximations although they are of course not computable in practice and they are only considered for the analysis of a posteriori error estimators .
the postprocessed approximations to the navier - stokes equations were first developed for spectral methods in @xcite , @xcite , @xcite , @xcite and also developed for mfe methods for the navier - stokes equations in @xcite , @xcite , @xcite .
for the sake of completeness , in the present paper we also analyze the use of the computable postprocessed approximations of @xcite for a posteriori error estimation .
the use of this kind of postprocessing technique to get a posteriori error estimations has been studied in @xcite , @xcite and @xcite for nonlinear parabolic equations excluding the navier - stokes equations .
we refer also to @xcite where the so - called stokes reconstruction is used to a posteriori estimate the errors of the semi - discrete in space approximations to a linear time - dependent stokes problem .
we remark that the stokes reconstruction of @xcite is exactly the postprocessing approximation ( @xmath33 ) in the particular case of a linear model . in the second part of the paper we consider a posteriori error estimations for the fully discrete mfe approximations @xmath34 and @xmath35 , ( @xmath36 for @xmath37 ) obtained by integrating in time with either the backward euler method or the two - step backward differentiation formula ( bdf ) . for this purpose
, we define a stokes problem similar to ( [ eq : stokes ] ) but with the right - hand - side depending now on the fully discrete mfe approximation @xmath38 ( problem ( [ posth0n])([posth1n ] ) in section [ sec:4 ] below ) . we will call time - discrete postprocessed approximation to the solution @xmath39 of this new stokes problem . as before , @xmath40 is not computable in practice and it is only considered for the analysis of a posteriori error estimation .
observe that in the fully discrete case ( which is the case in actual computations ) the task of estimating the the error @xmath41 of the mfe approximation becomes more difficult due to the presence of time discretization errors @xmath42 , which are added to the spatial discretization errors @xmath43 .
however we show in section [ sec:4 ] that if temporal and spatial errors are not very different in size , the quantity @xmath44 correctly esimates the spatial error because the leading terms of the temporal errors in @xmath45 and @xmath38 get canceled out when subtracting @xmath44 , leaving only the spatial component of the error .
this is a very convenient property that allows to use independent procedures for the tasks of estimating the errors of the spatial and temporal discretizations .
we remark that the temporal error can be routinely controlled by resorting to well - known ordinary differential equations techniques .
analogous results were obtained in @xcite for fully discrete finite element approximations to evolutionary convection - reaction - diffusion equations using the backward euler method . as in the semidiscrete case , a key point in our results is again the fact that the fully discrete mfe approximation @xmath46 to the navier - stokes problem ( [ onetwo])([ic ] ) is also the mfe approximation to the solution @xmath47 of the stokes problem ( [ posth0n])([posth1n ] ) . as a consequence , we can use again any available error estimator for the stokes problem to estimate the spatial error of the fully discrete mfe approximations @xmath46 to the navier - stokes problem ( [ onetwo])([ic ] ) .
computable mixed finite element approximations to @xmath47 , the so - called fully discrete postprocessed approximations , were studied and analyzed in @xcite where we proved that the fully discrete postprocessed approximations maintain the increased spatial accuracy of the semi - discrete approximations . the analysis in the second part of the present paper borrows in part from @xmath48 .
also , we propose a computable error estimator based on the fully discrete postprocessed approximation of @xcite and show that it also has the excellent property of separating spatial and temporal errors .
the rest of the paper is as follows . in section 2
we introduce some preliminaries and notation . in section 3
we study the a posteriori error estimation of semi - discrete in space mfe approximations . in section 4
we study a posteriori error estimates for fully discrete approximations .
finally , some numerical experiments are shown in section 5 .
we will assume that @xmath49 is a bounded domain in @xmath50 , of class @xmath51 , for @xmath52 .
when dealing with linear elements ( @xmath53 below ) @xmath49 may also be a convex polygonal or polyhedral domain .
we consider the hilbert spaces @xmath54 endowed with the inner product of @xmath55 and @xmath56 , respectively . for @xmath57 integer and @xmath58 , we consider the standard spaces , @xmath59 , of functions with derivatives up to order @xmath60 in @xmath61 , and @xmath62 .
we will denote by @xmath63 the norm in @xmath64 , and @xmath65 will represent the norm of its dual space .
we consider also the quotient spaces @xmath66 with norm @xmath67 .
we recall the following sobolev s imbeddings @xcite : for @xmath68 , there exists a constant @xmath69 such that @xmath70 for @xmath71 , ( [ sob1 ] ) holds with @xmath72 .
the following inf - sup condition is satisfied ( see @xcite ) , there exists a constant @xmath73 such that @xmath74 where , here and in the sequel , @xmath75 denotes the standard inner product in @xmath76 or in @xmath28
. let @xmath77 be the @xmath78 projector onto @xmath79 .
we denote by @xmath80 the stokes operator on @xmath81 : @xmath82 applying leray s projector @xmath83 to ( [ onetwo ] ) , the equations can be written in the form @xmath84 where @xmath85 for @xmath5 , @xmath86 in @xmath87
. we shall use the trilinear form @xmath88 defined by @xmath89 it is straightforward to verify that @xmath90 enjoys skew - symmetry : @xmath91 let us observe that @xmath92 for @xmath93 @xmath94 .
let us consider for @xmath95 and @xmath96 the operators @xmath97 and @xmath98 , which are defined by means of the spectral properties of @xmath80 ( see , e.g. , @xcite , @xcite ) .
notice that @xmath80 is a positive self - adjoint operator with compact resolvent in @xmath79 .
an easy calculation shows that @xmath99 where , here and in what follows , @xmath100 when applied to an operator denotes the associated operator norm .
we shall assume that the solution @xmath5 of ( [ onetwo])-([ic ] ) satisfies @xmath101 for some constants @xmath102 and @xmath103 .
we shall also assume that there exists a constant @xmath104 such that @xmath105 finally , we shall assume that for some @xmath106 @xmath107 so that , according to theorems 2.4 and 2.5 in @xcite , there exist positive constants @xmath108 and @xmath109 such that the following bounds hold : @xmath110 where @xmath111 and @xmath112 for some @xmath113 .
observe that for @xmath114 , we can take @xmath115 and @xmath116 . for simplicity
, we will take these values of @xmath117 and @xmath118 .
let @xmath119 , @xmath120 be a family of partitions of suitable domains @xmath121 , where @xmath19 is the maximum diameter of the elements @xmath122 , and @xmath123 are the mappings of the reference simplex @xmath124 onto @xmath125 .
let @xmath126 , we consider the finite - element spaces @xmath127 where @xmath128 denotes the space of polynomials of degree at most @xmath129 on @xmath130 . as it is customary in the analysis of finite - element methods for the navier - stokes equations
( see e. g. , @xcite , @xcite , @xcite , @xcite , @xcite ) we restrict ourselves to quasiuniform and regular meshes @xmath131 , so that as a consequence of ( * ? ? ?
* theorem 3.2.6 ) , the following inverse inequality holds for each @xmath132 @xmath133 where @xmath134 , @xmath135 .
we shall denote by @xmath136 the so - called hood
taylor element @xcite , when @xmath137 , where @xmath138 and the so - called mini - element @xcite when @xmath53 , where @xmath139 , and @xmath140 . here
, @xmath141 is spanned by the bubble functions @xmath142 , @xmath143 , defined by @xmath144 , if @xmath145 and 0 elsewhere , where @xmath146 denote the barycentric coordinates of @xmath147 . for these elements
a uniform inf - sup condition is satisfied ( see @xcite ) , that is , there exists a constant @xmath73 independent of the mesh grid size @xmath19 such that @xmath148 we remark that our analysis can also be applied to other pairs of lbb - stable mixed finite elements ( see ( * ? ? ?
* remark 2.1 ) ) .
the approximate velocity belongs to the discrete divergence - free space @xmath149 which is not a subspace of @xmath150 .
we shall frequently write @xmath151 instead of @xmath152 whenever the value of @xmath153 plays no particular role . let @xmath154 be the discrete leray s projection defined by @xmath155 we will use the following well - known bounds @xmath156 we will denote by @xmath157 the discrete stokes operator defined by @xmath158 let @xmath159 be the solution of a stokes problem with right - hand side @xmath160 , we will denote by @xmath161 the so - called stokes projection ( see @xcite ) defined as the velocity component of solution of the following stokes problem : find @xmath162 such that @xmath163 the following bound holds for @xmath164 : @xmath165 the proof of ( [ stokespro ] ) for @xmath166 can be found in @xcite .
for the general case , @xmath121 must be such that the value of @xmath167 satisfies @xmath168 .
this can be achieved if , for example , @xmath169 is piecewise of class @xmath170 , and superparametric approximation at the boundary is used @xcite . under the same conditions ,
the bound for the pressure is @xcite @xmath171 where the constant @xmath172 depends on the constant @xmath173 in the inf - sup condition ( [ lbbh ] ) .
we will assume that the domain @xmath49 is of class @xmath51 , with @xmath174 so that standard bounds for the stokes problem @xcite , @xcite imply that @xmath175 for a domain @xmath49 of class @xmath176 we also have the bound ( see @xcite ) @xmath177 in what follows we will apply the above estimates to the particular case in which @xmath178 is the solution of the navier stokes problem ( [ onetwo])([ic ] ) . in that case
@xmath179 is the discrete velocity in problem ( [ stokesnew])([stokesnew2 ] ) with @xmath180 .
note that the temporal variable @xmath181 appears here merely as a parameter , and then , taking the time derivative , the error bound ( [ stokespro ] ) can also be applied to the time derivative of @xmath182 changing @xmath5 , @xmath6 by @xmath183 , @xmath184 .
since we are assuming that @xmath49 is of class @xmath51 and @xmath52 , from ( [ stokespro ] ) and standard bounds for the stokes problem @xcite , we deduce that @xmath185 we consider the semi - discrete finite - element approximation @xmath27 to @xmath178 , solution of ( [ onetwo])([ic ] ) .
that is , given @xmath186 , we compute @xmath187 and @xmath188 , @xmath189 $ ] , satisfying @xmath190 for @xmath191 , provided that ( [ stokespro])([stokespre ] ) hold for @xmath192 , and ( [ eq : u - inf])([eq : u - int ] ) hold for @xmath193 , then we have @xmath194 ( see , e.g. , @xcite ) , and also , @xmath195 where @xmath196 if @xmath197 and @xmath198 if @xmath199 .
see ( * ? ? ?
* proposition 3.2 ) .
let us consider the mfe approximation @xmath27 at any time @xmath200 $ ] to @xmath201 obtained by solving ( [ ten])([ten2 ] ) .
we consider the postprocessed approximation @xmath202 in @xmath203 which is the solution of the following stokes problem written in weak form @xmath204 for all @xmath205 and @xmath206 .
we remark that the mfe approximation @xmath207 to @xmath201 is also the mfe approximation to the solution @xmath202 of the stokes problem ( [ posth0])([posth1 ] ) . in theorems [ semi_disve ] and [ semi_dispre ] below
we prove that the postprocessed approximation @xmath202 is an improved approximation to the solution @xmath178 of the evolutionary navier - stokes equations ( [ onetwo])([ic ] ) at time @xmath208 .
although , as it is obvious , @xmath202 is not computable in practice , it is however a useful tool to provide a posteriori error estimates for the mfe approximation @xmath27 at any desired time @xmath14 . in theorem [ semi_disve ]
we obtain the error bounds for the velocity and in theorem [ semi_dispre ] the bounds for the pressure .
the improvement is achieved in the @xmath209 norm when using the mini - element ( @xmath53 ) and in both the @xmath28 and @xmath209 norms in the cases @xmath210 . in the sequel
we will use that for a forcing term satisfying ( [ tildem2 ] ) there exists a constant @xmath211 , depending only on @xmath104 , @xmath212 and @xmath213 , such that the following bound hold for @xmath214 : @xmath215 the following inequalities hold for all @xmath216 and @xmath217 , see ( * ? ? ? * ( 3.7 ) ) : @xmath218 the proof of theorem [ semi_disve ] requires some previous results which we now state and prove .
we will use the fact that @xmath219 for @xmath220 , from where it follows that @xmath221 where the constant @xmath222 is independent of @xmath19 .
let @xmath178 be the solution of ( [ onetwo])([ic ] ) and fix @xmath113 . then there exists a positive constant @xmath223 such that for @xmath224 satisfying the threshold condition @xmath225 the following inequalities hold for @xmath226 : @xmath227 due to the equivalence ( [ a-1equiv ] ) and and since @xmath228 for @xmath229 it is sufficient to prove @xmath230 for @xmath231 or @xmath232 .
we follow the proof ( * ? ? ?
* lemma 3.1 ) where a different threshold assumption is assumed .
we do this for @xmath231 , since the case @xmath232 is similar but yet simpler .
we write @xmath233 where @xmath234 .
we first observe that @xmath235 where , in the last inequality , we have used that thanks to sobolev s inequality ( [ sob1 ] ) we have @xmath236 .
similarly , @xmath237 the proof of the case @xmath238 in ( [ eq : aux_bat ] ) is finished if we show that for @xmath239 , both @xmath240 and @xmath241 are bounded in terms of @xmath103 and the value @xmath242 in the threshold assumption ( [ threshold ] ) . to do this
, we will use the inverse inequality ( [ inv ] ) and the fact that the stokes projection @xmath179 satisfies that @xmath243 for some constant @xmath244 ( see for example the proof of lemma 3.1 in @xcite ) .
we have @xmath245 where in the last inequality we have applied ( [ inv ] ) , and , similarly , @xmath246 where we also have used that @xmath247 for @xmath248 . now the threshold assumption ( [ threshold ] ) and ( [ stokespro ] ) show the boundedness of @xmath240 and @xmath241 .
finally , the proof of the case @xmath249 in ( [ eq : aux_bat ] ) is , with obvious changes , that of the equivalent result in ( * ? ? ?
* lemma 3.1 ) . in the sequel we consider the auxiliary function @xmath250\rightarrow v_h$ ] solution of @xmath251
according to ( * ? ? ?
* remark 4.2 ) we have @xmath252 for some constant @xmath253 .
the following lemma provides a superconvergence result .
[ le : sup1 ] let @xmath178 be the solution of ( [ onetwo])([ic ] ) .
then , there exists a positive constant @xmath222 such that the solution @xmath254 of ( [ atenopv ] ) and the galerkin approximation @xmath8 satisfy the following bound , @xmath255.\ ] ] since for @xmath256 we have @xmath257 where @xmath258 , it follows that @xmath259 applying ( [ 4.21 - 4.22 ] ) we have @xmath260 , so that taking into account that @xmath261 it follows that @xmath262 since applying ( * ? ? ?
* lemma 4.2 ) we obtain @xmath263 a generalized gronwall lemma @xcite , together with ( [ 4.21 - 4.22 ] ) allow us to conclude @xmath264 then by writing @xmath265 and applying ( [ eq : error - pi_h ] ) and ( [ eq : cota - z - low ] ) , the proof is finished if we check that the threshold condition ( [ threshold ] ) holds for @xmath266 and @xmath267 . in view of ( [ eq : cota - z - low ] ) , ( [ eq : error - pi_h ] ) and the inverse inequality ( [ inv ] ) we have indeed that @xmath268 , for @xmath226 . in the case of
@xmath8 the threshold condition holds due to ( [ eq : err_vel(t ) ] ) .
[ le : e_t ] let @xmath178 be the solution of ( [ onetwo])([ic ] ) . then there exists a positive constant @xmath222 such that @xmath269 , \label{e_t-1}\ ] ] where @xmath254 and @xmath270 are defined by ( [ atenopv ] ) and ( [ ten])-([ten2 ] ) respectively .
the difference @xmath271 satisfies that @xmath272 , so that multiplying by @xmath273 and taking norms , thanks to ( [ 4.21 - 4.22b ] ) , we have @xmath274 now we write @xmath275 so that ( [ eq : error - pi_h ] ) , ( [ eq : cota - z - low ] ) and ( [ supercuad ] ) , allow us to write , @xmath276 then , applying ( [ a-1equiv ] ) the proof is finished . [ le : z_t ] let @xmath178 be the solution of ( [ onetwo])([ic ] )
. then there exists a positive constant @xmath222 such that @xmath277 , \quad r=2,3,4,\ ] ] where @xmath278 when @xmath53 and @xmath279 otherwise .
the case @xmath210 is proved in ( * ? ? ? * lemma 5.1 ) . for the case
@xmath53 we write @xmath280 a simple duality argument and the fact that @xmath281 , easily show that @xmath282 the bound of the third term on the right - hand side of ( [ demini1 ] ) is given in lemma [ le : e_t ] , so that , thanks to the equivalence ( [ a-1equiv ] ) we are left with estimating @xmath283 we notice that @xmath284 where @xmath285 .
thus , @xmath286 recalling ( [ eq : vabien ] ) by means of the change of variables @xmath287 it is easy to show that @xmath288 where @xmath289 is the beta function ( see e. g. , @xcite ) .
thus , we have @xmath290 the first term on the right - hand side above is bounded by @xmath291 . for the second one we notice hat @xmath292 so that using ( [ eq : error - pi_h ] ) , ( [ stokespro ] ) and ( [ eq : cota - z - low ] ) it is bounded by @xmath293 .
[ semi_disve ] let @xmath178 be the solution of ( [ onetwo])-([ic ] ) .
then , there exists a positive constant @xmath222 such that the postprocessed velocity @xmath15 , defined in ( [ posth0])-([posth1 ] ) , satisfies the following bounds : \(i ) if @xmath53 then @xmath294 \(ii ) if @xmath210 then @xmath295 the proof follows the same steps as ( * ? ? ?
* theorem 5.2 ) . subtracting ( [ posth0 ] ) from ( [ onetwo ] ) ,
standard duality arguments show that @xmath296 to bound the second term on the right - hand side above we apply lemma [ le : z_t ] , whereas for the second we apply ( [ eq : aux_bat ] ) to get @xmath297 so that applying ( [ eq : err_vel(t ) ] ) the proof of ( [ utilde_lineal ] ) and the case @xmath249 of ( [ utilde_mas ] ) are finished .
we now get the error bounds in the @xmath11 norm .
it is easy to see that @xmath298 then , by applying @xmath299 to both sides of the above equations , we obtain @xmath300 as regards the nonlinear term , applying ( * ? ? ?
* lemma 4.1 ) we obtain @xmath301 to bound the second term on the right - hand side above we apply ( [ eq : err_vel(t ) ] ) , whereas the first one is bounded in the proof of ( * ? ? ?
* theorem 5.2 ) by @xmath302 finally , to bound @xmath303 we apply ( * ? ? ?
* lemma 5.1 ) to obtain @xmath304 which concludes the proof . in the following theorem
we obtain the error bounds for the pressure @xmath16 .
[ semi_dispre ] let @xmath178 be the solution of ( [ onetwo])-([ic ] ) .
then , there exists a positive constant @xmath222 such that the postprocessed pressure , @xmath16 , satisfies the following bounds : @xmath305 where @xmath278 if @xmath53 and @xmath279 if @xmath210 .
the proof follows the same steps as ( * ? ? ?
* theorem 5.3 ) . applying the inf - sup condition ( [ inf - sup ] )
it is easy to see that @xmath306 applying now ( [ utilde_lineal ] ) and ( [ utilde_mas ] ) to bound the first term and reasoning as in the proof of theorem [ semi_disve ] to bound the other two terms we conclude ( [ ptilde_mas ] ) .
[ remark31 ] as a consequence of theorems [ semi_disve ] and [ semi_dispre ] we obtain in the proof of theorem [ apos_semi1 ] that @xmath307 is an asymptotically exact estimator of the error @xmath308 while @xmath309 is an asymptotically exact estimator of the error @xmath310
. however , as we have already observed @xmath15 and @xmath16 are not computable in practice . in theorems
[ apos_semi1 ] , 4 and [ th_pos_esti ] we present different procedures to get computable error estimators . as we pointed out before the mfe approximations @xmath27 to the velocity and the pressure of the solution @xmath178 of the evolutionary navier - stokes equations ( [ onetwo])-([ic ] ) at any fixed time @xmath208 are also the approximations to the velocity and pressure of the steady stokes problem ( [ posth0])-([posth1 ] ) . in theorem [ apos_semi1 ]
we show that any a posteriori error estimator of the error in the steady stokes problem ( [ posth0])-([posth1 ] ) gives us an a posteriori indicator of the error in the approximations to the evolutionary navier - stokes equations . using the notation of @xcite we will denote in the sequel by @xmath311 , @xmath226 , any a posteriori error estimator of the error @xmath312 in the norm of @xmath313 in the approximation to the velocity in the steady stokes problem ( [ posth0])-([posth1 ] ) .
we will denote by @xmath314 any error estimator of the quantity @xmath315 .
[ apos_semi1 ] let @xmath178 be the solution of ( [ onetwo])-([ic ] ) and fix any positive time @xmath14 .
assume that the galerkin approximation @xmath27 satisfies , for @xmath19 small enough and @xmath316 , @xmath317 for some positive constant @xmath318 .
\(i ) if @xmath319 , @xmath226 , is an efficient error indicator of the error in the mfe approximation to the steady stokes problem ( [ posth0])-([posth1 ] ) .
that is , if there exist positive constants , @xmath320 and @xmath321 , that are independent of the mesh size @xmath19 , such that the following bound holds @xmath322 then @xmath319 , @xmath226 , it is also an efficient error indicator of the error in the mfe approximation to the evolutionary navier - stokes equations , i.e. there exist positive constants @xmath323 and @xmath324 that are independent of the mesh size @xmath19 such that the following bound holds for @xmath19 small enough @xmath325 \(ii ) if @xmath319 , @xmath226 is an asymptotically exact error estimator of the error in the steady stokes problem then it is also an asymptotically exact error estimator of the error in the evolutionary navier - stokes equations .
\(iii ) analogous results are obtained in the approximations to the pressure . in the case
@xmath53 , the results are valid only in the @xmath12 norm . for simplicity in the exposition we will concentrate on the cases @xmath210 in the approximations to the velocity , the proof for the approximations to the pressure and for the case @xmath53 being the same except for obvious changes .
let us first observe that @xmath326 dividing by @xmath327 , using ( [ saturacion ] ) and applying theorem [ semi_disve ] we obtain @xmath328 now , using ( [ efficient_steady ] ) we get @xmath329 taking @xmath19 small enough so that @xmath330 we get @xmath331 now , we use the decomposition @xmath332 reasoning as before we get @xmath333 since @xmath334 we finally reach @xmath335 from ( [ primer ] ) and ( [ segun ] ) we conclude ( [ efficient_evol ] ) with @xmath336 and @xmath337 .
let us now assume that @xmath338 is an asymptotically exact error estimator .
using again the decomposition ( [ eq : asterisco ] ) we have @xmath339 the last equality being a consequence of theorem ( [ semi_disve ] ) and the saturation hypothesis ( [ saturacion ] ) . as we pointed out before
, this limit implies that @xmath307 is an asymptotically exact estimator of the error @xmath308 .
then @xmath340 and @xmath338 is also an asymptotically exact estimator of the error in the approximation to the velocity of the evolutionary navier - stokes equations .
we remark that with hypothesis ( [ saturacion ] ) we are merely assuming that the term of order @xmath341 is really present in the asymptotic expansion of the galerkin error .
let us also notice that the constant @xmath342 in ( [ saturacion ] ) is , in general @xmath343 , so that the ratio @xmath344 in the proof of theorem [ apos_semi1 ] is , in general , @xmath345 .
the key point in theorem [ apos_semi1 ] comes from the observation that if we decompose @xmath346 the first term on the right hand side of ( [ decom_error ] ) , @xmath347 is in general smaller , by a factor of size @xmath348 , than the second one , @xmath20 ( theorem [ semi_disve ] ) .
then , to estimate the error @xmath22 we can safely omit the term @xmath30 in ( [ decom_error ] ) . comparing with the analysis of @xcite for a nonstationary linear stokes model problem the main difference is that the two terms in ( [ decom_error ] ) are taken into account . in theorem [ the_new ] below
we show that this kind of technique can also be applied to the nonlinear navier - stokes equations .
the advantage of this point of view is that hypothesis ( [ saturacion ] ) is not required for the proof of theorem [ the_new ] .
let us finally observe that @xmath349 are the mfe approximations to the solution @xmath350 of the stokes problem that we obtain deriving respect to the time variable the stokes problem ( [ posth0])-([posth1 ] ) .
then , we will denote by @xmath351 , @xmath352 , any a posteriori error estimator of the error @xmath353 in the norm of @xmath313 in the approximation to the velocity of the corresponding steady stokes problem .
the proof of the following theorem follows the steps of the proof of ( * ? ? ?
* theorem 1 ) .
[ the_new ] let @xmath178 be the solution of ( [ onetwo])-([ic ] ) and let @xmath27 be its mfe galerkin approximation .
then , the following a posteriori error bound holds for @xmath214 and a constant @xmath222 independent of @xmath19 .
@xmath354 let us denote by @xmath355 . from ( [ posth0])-([posth1 ] )
it follows that @xmath356 then @xmath357 satisfies the equation @xmath358 taking into account ( [ lee1 ] ) we get @xmath359 we first observe that for any @xmath360 we have @xmath361 .
then , taking into account ( [ eq : aux_bat ] ) we get @xmath362 let us observe that in order to apply ( [ eq : aux_bat ] ) we require @xmath8 to satisfy ( [ threshold ] ) , which holds due to ( [ eq : err_vel(t ) ] ) , and @xmath363 and @xmath364 to be bounded . using ( [ sob1 ] )
both norms are bounded in terms of @xmath365 . applying ( [ eq : may06-nueva ] ) we get @xmath366 finally , using that @xmath367 is uniformly bounded , see ( [ a_h_uh ] ) , and reasoning as in ( [ eq : adelanto0 ] ) to bound the second and forth terms above we conclude @xmath365 is uniformly bounded .
then , we arrive at @xmath368 and then @xmath369 a standard application of a generalized gronwall lemma @xcite gives @xmath370 now , using decomposition ( [ decom_error ] ) we conclude the proof .
we observe that using the same proof , a similar bound for the @xmath209 norm of the error can be obtained changing only @xmath371 by @xmath372 and @xmath373 by @xmath374 .
let us also remark that theorem [ the_new ] allows to a posteriori obtain upper error bounds for the error in the approximation to the nonlinear navier - stokes equations using only upper error bounds for some stokes problems depending only on the data and the computed approximation .
however , the estimation of the error at a time @xmath181 requires the estimation of the error of a family of stokes problems with right hand side depending on @xmath117 , for all @xmath375 $ ] .
we now propose a simple procedure to estimate the error which is based on computing a mfe approximation to the solution @xmath202 of ( [ posth0])-([posth1 ] ) on a mfe space with better approximation capabilities than @xmath376 in which the galerkin approximation @xmath27 is defined .
this procedure was applied to the @xmath6-version of the finite - element method for evolutionary convection - reaction - diffusion equations in @xcite .
the main idea here is to use a second approximation of different accuracy than that of the galerkin approximation of @xmath178 and whose computational cost hardly adds to that of the galerkin approximation itself .
let us fix any time @xmath377 $ ] and let approximate the solution @xmath378 of the stokes problem ( [ posth0])-([posth1 ] ) by solving the following discrete stokes problem : find @xmath379 satisfying @xmath380 where @xmath381 is either : 1 . the same - order mfe over a finer grid .
that is , for @xmath382 , we choose @xmath383 .
2 . a higher - order mfe over the same grid . in this case
we choose @xmath384 @xmath385 .
we now study the errors @xmath386 and @xmath387 .
[ teo4 ] let @xmath178 be the solution of ( [ onetwo])([ic ] ) and for @xmath316 , and let ( [ eq : u - inf])([eq : u - int ] ) hold with @xmath388 then , there exists a positive constant @xmath222 such that the postprocessed mfe approximation to @xmath5 , @xmath389 satisfies the following bounds for @xmath316 and @xmath390 $ ] : \(i ) if the postprocessing element is @xmath391 , then @xmath392 \(ii ) if the postprocessing element is @xmath393 , then @xmath394 for @xmath53 only the case @xmath249 in ( [ postcuad1 ] ) and ( [ postcuad2 ] ) holds . in ( [ postcuad1])([ulti2 ] ) , @xmath278 when @xmath53 and @xmath279 otherwise .
the cases @xmath210 have been proven in theorems 5.2 and 5.3 in @xcite . following the same arguments
, we now prove the results corresponding to @xmath53 and @xmath391 , the case @xmath395 being similar , yet easier .
we decompose the error @xmath396 , where @xmath397 is the solution of @xmath398 that is , @xmath399 is the stokes projection of @xmath5 onto @xmath400 .
since in view of ( [ stokespro])([stokespre ] ) we have @xmath401 we only have to estimate @xmath402 and @xmath403 .
to do this , we subtract ( [ posth0dis ] ) from ( [ july-24th11 ] ) , and take inner product with @xmath404 to get @xmath405 now applying lemma [ le : z_t ] , ( [ eq : aux_bat ] ) and ( [ eq : err_vel(t ) ] ) the proof of ( [ postcuad1 ] ) is finished . to prove ( [ ulti1 ] ) , again we subtract ( [ posth0dis ] ) from ( [ july-24th11 ] ) , rearrange terms and apply the inf - sup condition ( [ inf - sup ] ) to get @xmath406 and the proof is finished with the same arguments used to prove ( [ postcuad1 ] ) . to estimate the error in @xmath407 we propose to take the difference between the postprocessed and the galerkin approximations : @xmath408 in the following theorem we prove that this error estimator is efficient and asymptotically exact both in the @xmath28 and @xmath209 norms and it has the advantage of providing an improved approximation when added to the galerkin mfe approximation .
[ th_pos_esti ] let @xmath178 be the solution of ( [ onetwo])-([ic ] ) and fix any positive time @xmath14 .
assume that condition ( [ saturacion ] ) is satisfied .
then , there exist positive constants @xmath409 , @xmath410 , and @xmath320 , @xmath321 , @xmath323 and @xmath324 such that , for @xmath411 and @xmath412 , the error estimators @xmath413 @xmath414 satisfy the following bounds when @xmath415 and @xmath416 : @xmath417 furthermore , if @xmath415 , with @xmath418 , @xmath419 , or @xmath420 then @xmath421 for the mini element , the case @xmath238 in ( [ efi_posdis ] ) and ( [ asin_posdis ] ) must be excluded
. we will prove the estimates for the velocity in the case @xmath210 , since the estimates for the pressure and the case @xmath53 are obtained by similar arguments but with obvious changes .
let us observe that for @xmath226 @xmath422 on the other hand @xmath423 using ( [ saturacion ] ) we get @xmath424 taking @xmath425 and @xmath19 and @xmath426 sufficiently small , the bound ( [ efi_posdis ] ) is readily obtained .
the proof of ( [ asin_posdis ] ) follows straightforwardly from ( [ july-24th ] ) , since in the case when @xmath415 with @xmath418 , @xmath419 , the term @xmath427 when @xmath19 tends to zero , and in the case when @xmath420 the term containing the parameter @xmath428 is not present .
in practice , it is not possible to compute the mfe approximation exactly , and , instead , some time - stepping procedure must be used to approximate the solution of ( [ ten])-([ten2 ] ) .
hence , for some time levels @xmath429 , approximations @xmath430 and @xmath35 are obtained . in this section
we assume that the approximations are obtained with the backward euler method or the two - step bdf which we now describe . for simplicity , we consider only constant stepsizes , that is , for @xmath431 integer , we fix @xmath432 , and we denote @xmath433 , @xmath434 . for a sequence @xmath435 we denote @xmath436 given @xmath437 , a sequence @xmath438 of approximations to @xmath439 , @xmath440 , is obtained by means of the following recurrence relation : @xmath441 where @xmath442 in the case of the backward euler method and @xmath443 for the two - step bdf . in this last case , a second starting value @xmath444 is needed . here
, we will always assume that @xmath444 is obtained by one step of the backward euler method .
also , for both the backward euler and the two - step bdf , we assume that @xmath437 , which is usually the case in practical situations .
we now define the time - discrete postprocessed approximation . given an approximation @xmath445 to @xmath446 , the time - discrete postprocessed velocity and pressure @xmath47 are defined as the solution of the following stokes problem : @xmath447 for reasons already analyzed in @xcite and @xcite we define @xmath448 as an adequate approximation to the time derivative @xmath446 .
for the analysis of the errors @xmath449 and @xmath450 we follow @xcite , where the mfe approximations to the stokes problem ( [ posth0n])([posth1n ] ) are analyzed .
we start by decomposing the errors @xmath449 and @xmath450 as follows , @xmath451 where @xmath452 and @xmath453 are the temporal errors of the time - discrete postprocessed velocity and pressure @xmath454 .
the first terms on the right - hand sides of ( [ eq : decomtilde])([eq : decomtildep ] ) are the errors of the postprocessed approximation that were studied in the previous section .
let us denote by @xmath42 , the temporal error of the mfe approximation to the velocity , and by @xmath455 , the temporal error of the mfe approximation to the pressure . in the present section we bound @xmath456 and @xmath457 in terms of @xmath458 .
the error bounds in the following lemma are similar to those of ( * ? ? ?
* proposition 3.1 ) where error estimates for mfe approximations of the stokes problem ( [ posth0n])([posth1n ] ) are obtained .
[ lema_need]there exists a positive constant @xmath459 such that @xmath460 let us denote by @xmath461 where @xmath462 . subtracting ( [ posth0n])([posth1n ] ) from ( [ posth0])([posth1 ] ) we have that the temporal errors @xmath463 of the time - discrete postprocessed velocity and pressure are the solution of the following stokes problem @xmath464 on the other hand , subtracting ( [ tend])-([tend2 ] ) from ( [ ten])-([ten2 ] ) and taking into account that , thanks to definition ( [ dtu ] ) @xmath465
, we get that the temporal errors @xmath466 of the fully discrete mfe approximation satisfy @xmath467 and thus @xmath466 is the mfe approximation to the solution @xmath463 of ( [ otro_stokes])([otro_stokes2 ] ) . using
then ( [ stokespro+1 ] ) we get @xmath468 for the pressure we apply ( [ stokespre ] ) and ( [ prehr ] ) to obtain @xmath469 then , to conclude , it only remains to bound @xmath470 . from the definition of @xmath471
it is easy to see that @xmath472 so that @xmath473 now , by writing @xmath160 as @xmath474 and using ( [ eq : adelanto0])([eq : adelanto1 ] ) we get @xmath475 from which we finally conclude ( [ error_post_fully ] ) and ( [ error_post_fully_pre ] ) .
let us consider the quantities @xmath44 and @xmath476 as a posteriori indicators of the error in the fully discrete approximations to the velocity and pressure respectively .
then , we obtain the following result : [ th_esti_time_d ] let @xmath178 be the solution of ( [ onetwo])([ic ] ) and let ( [ tildem2 ] ) hold .
assume that the fully discrete mfe approximations @xmath438 , @xmath477 are obtained by the backward euler method or the two - step bdf ( [ tend])([tend2 ] ) , and let @xmath478 be the solution of ( [ posth0n])([posth1n ] ) .
then , for @xmath479 , @xmath480 where @xmath481 is the constant in ( [ pri_fully])([pri_fully_pres ] ) , @xmath482 for the backward euler method and @xmath483 for the two - step bdf . in ( * ? ? ?
* theorems 5.4 and 5.7 ) we prove that if ( [ tildem2 ] ) and the case @xmath484 in ( [ stokespro ] ) hold , the errors @xmath458 of these two time integration procedures satisfy for @xmath485 small enough that @xmath486 for a certain constants @xmath487 and @xmath321 , where @xmath482 for the backward euler method and @xmath483 for the two - step bdf .
since @xmath488 , and then @xmath489 , from ( [ eq : orden1 ] ) and ( [ error_post_fully])-([error_post_fully_pre ] ) we finally reach that for @xmath485 small enough @xmath490 where @xmath481 is a positive constant .
let us decompose the estimators as follows : @xmath491 which implies @xmath492 thus , in view of ( [ pri_fully])([pri_fully_pres ] ) we obtain ( [ decom_prin2 ] ) and ( [ decom_prin2p ] ) let us comment on the practical implications of this theorem .
observe that from ( [ decom_prin ] ) and ( [ decom_prinp ] ) the fully discrete estimators @xmath44 and @xmath493 can be both decomposed as the sum of two terms .
the first one is the semi - discrete a posteriori error estimator we have studied in the previous section ( see remark [ remark31 ] ) and which we showed it is an asymptotically exact estimator of the spatial error of @xmath38 and @xmath494 respectively . on the other hand , as shown in ( [ pri_fully])([pri_fully_pres ] ) , the size of the second term is in asymptotically smaller than the temporal error of @xmath38 and @xmath494 respectively .
we conclude that , as long as the spatial an temporal errors are not too unbalanced ( i.e. , they are not of very different sizes ) , the first term in ( [ decom_prin ] ) and ( [ decom_prinp ] ) is dominant and then the quantities @xmath495 and @xmath493 are a posteriori error estimators of the spatial error of the fully discrete approximations to the velocity and pressure respectively . the control of the temporal error can be then accomplished by standard and well - stablished techniques in the field of numerical integration of ordinary differential equations .
now , we remark that @xmath47 are obviously not computable .
however , we observe that the fully discrete approximation @xmath438 of the evolutionary navier - stokes equation is also the approximation to the stokes problem ( [ posth0n])-([posth1n ] ) whose solution is @xmath496 .
then , one can use any of the available error estimators for a steady stokes problem to estimate the quantities @xmath497 and @xmath498 , which , as we have already proved , are error indicators of the spatial errors of the fully discrete approximations to the velocity and pressure , respectively . to conclude , we show a procedure to get computable estimates of the error in the fully discrete approximations .
we define the fully discrete postprocessed approximation @xmath499 as the solution of the following stokes problem ( see @xcite ) : @xmath500 where @xmath381 is as in ( [ posth0dis])-([posth1dis ] ) .
let us denote by @xmath501 and @xmath502 the temporal errors of the fully discrete postprocessed approximation @xmath503 ( observe that the semi - discrete postprocessed approximation @xmath504 is defined in ( [ posth0dis])-([posth1dis ] ) ) . let us denote , as before , by @xmath458 the temporal error of the mfe approximation to the velocity .
then , we have the following bounds .
[ prop : err_post_fully ] there exists a positive constant @xmath505 such that for @xmath506 the following bounds hold @xmath507 the bound ( [ error_post_fully_vel ] ) is proved in ( * ? ? ?
* proposition 3.1 ) . to prove ( [ duda2pp ] )
we decompose @xmath508 the second term above is bounded in ( [ error_post_fully_pre ] ) of lemma [ lema_need ] . for the first we observe that @xmath509 is the mfe approximation in @xmath510 to the pressure @xmath511 in ( [ otro_stokes])-([otro_stokes2 ] ) so that the same reasoning used in the proof of lemma [ lema_need ] allow us to obtain @xmath512 using ( [ eq : orden1 ] ) as before , we get the analogous to ( [ pri_fully ] ) and ( [ pri_fully_pres ] ) , i.e. , for @xmath485 small enough the following bound holds @xmath513 where @xmath481 is a positive constant . similarly to ( [ decom_prin])([decom_prinp ] ) we write @xmath514 and @xmath515 , so that in view of ( [ pri_fully_dis])([pri_fully_dis_pre ] ) we have the following result .
[ th_esti_fully_d ] let @xmath178 be the solution of ( [ onetwo])([ic ] ) and let ( [ tildem2 ] ) hold .
assume that the fully discrete mfe approximations @xmath438 , @xmath477 are obtained by the backward euler method or the two - step bdf ( [ tend])([tend2 ] ) , and let @xmath516 be the solution of ( [ posth0ndis])([posth1ndis ] ) .
then , for @xmath479 , @xmath517 where @xmath481 is the constant in ( [ pri_fully_dis])([pri_fully_dis_pre ] ) , @xmath482 for the backward euler method and @xmath483 for the two - step bdf .
the practical implications of this result are similar to those of theorem [ th_esti_time_d ] , that is , the first term on the right - hand side of ( [ decom_prin2_dis ] ) is an error indicator of the spatial error ( see theorem [ th_pos_esti ] ) while the second one is asymptotically smaller than the temporal error . as a consequence ,
the quantity @xmath518 is a computable estimator of the spatial error of the fully discrete velocity @xmath38 whenever the temporal and spatial errors of @xmath38 are more or less of the same size .
as before , similar arguments apply for the pressure .
we remark that having balanced spatial and temporal errors in the fully discrete approximation is the more common case in practical computations since one usually looks for a final solution with small total error . as in the semi - discrete case ,
the advantage of these error estimators is that they produce enhanced ( in space ) approximations when they are added to the galerkin mfe approximations .
we consider the equations @xmath519 in the domain @xmath520\times[0,1]$ ] subject to homogeneous dirichlet boundary conditions . for the numerical experiments of this section we approximate the equations using the mini - element @xcite over a regular triangulation of @xmath49 induced by the set of nodes @xmath521 , @xmath522 , where @xmath523 is an integer . for the time integration
we use the two - step bdf method with fixed time step . for the first step
we apply the backward euler method . in the first numerical experiment we study the semi - discrete in space case . to this end in the numerical experiments we integrate in time with a time - step small enough in order to have negligible temporal errors .
we take the forcing term @xmath524 such that the solution of ( [ onetwo_nu ] ) with @xmath525 is @xmath526 we chose @xmath527 in the first numerical experiment . when using the mini - element it has been observed and reported in the literature
( see for instance @xcite , @xcite , @xcite @xcite , @xcite and @xcite ) that the linear part of the approximation to the velocity , @xmath528 , is a better approximation to the solution @xmath5 than @xmath8 itself .
the bubble part of the approximation is only introduced for stability reasons and does not improve the approximation to the velocity and pressure terms .
for this reason in the numerical experiments of this section we only consider the errors in the linear approximation to the velocity .
also , following @xcite , we postprocess only the linear approximation to the velocity , i.e. , we solve the stokes problem ( [ posth0dis])-([posth1dis ] ) with @xmath528 and @xmath529 on the right - hand - side instead of @xmath8 and @xmath530 .
the finite element space at the postprocessed step is the same mini - element defined over a refined mesh of size @xmath428 .
we show the galerkin errors and the a posteriori error estimates obtained at time @xmath531 by taking the difference between the postprocessed and the standard approximations to the velocity and the pressure . in figure
[ semi ] , we ( asterisks ) and @xmath12 ( circles ) for @xmath532 , @xmath533 , @xmath534 , @xmath535 and @xmath536 and @xmath537 , @xmath538 , @xmath539 , @xmath540 and @xmath541 respectively . on the left , error estimations for the first component of the velocity . on the right , error estimations for the pressure.,title="fig:",width=226 ] ( asterisks ) and @xmath12 ( circles ) for @xmath532 , @xmath533 , @xmath534 , @xmath535 and @xmath536 and @xmath537 , @xmath538 , @xmath539 , @xmath540 and @xmath541 respectively
. on the left , error estimations for the first component of the velocity . on the right , error estimations for the pressure.,title="fig:",width=226 ] have represented the errors in the first component of the velocity of the galerkin approximation in the @xmath11 and @xmath12 norms and the errors for the pressure in the @xmath11 norm using solid lines .
we have used dashed lines to represent the error estimations .
the results for the second component of the velocity are completely analogous and they are not reported here .
the @xmath11 errors of the pressure , on the right of figure [ semi ] , are approximately twice as those of the @xmath12 errors of the velocity , on the left of figure [ semi ] , in this example
. we can observe that with the procedure we propose in this paper we get very accurate estimations of the errors , specially in the @xmath12 norm of the velocity .
the difference between the behavior of the error estimations in the @xmath11 and @xmath12 norms of the velocity are due to the fact that for first order approximations the postprocessed procedure increases the rate of convergence of the standard method only in the @xmath12 norm for the velocity and the @xmath11 norm for the pressure . however , since the postprocessed method produces smaller errors than the galerkin method also in the @xmath11 norm it can also be used to estimate the errors in this norm , as it can be checked in the experiment . on the right of figure [ semi ] we can clearly observe the asymptotically exact behavior of the estimator in the @xmath11 errors in the pressure in agreement with ( [ asin_posdis ] ) of theorem [ th_pos_esti ] .
let us denote by @xmath542 the efficiency indexes for the first component of the velocity and for the pressure . in table
[ tabla1 ] we have represented the values of the @xmath11 and @xmath12 norms of the velocity index and the @xmath13 norm of the pressure index for the experiments in figure [ semi ] .
we deduce again from the values of the efficiency indexes that the a posteriori error estimates are very accurate , all the values are remarkably close to @xmath543 , which is the optimal value for the efficiency index .
more precisely , we can observe that the values of the efficiency index in the @xmath11 norm for the velocity in this experiment belong to the interval @xmath544 $ ] .
the values in the @xmath12 norm for the velocity lie on the interval @xmath545 $ ] and , finally , the values for the pressure are in the interval @xmath546 $ ] . to conclude , we show a numerical experiment to check the behavior of the estimators in the fully discrete case .
we choose the forcing term @xmath7 such that the solution of ( [ onetwo_nu ] ) is ( [ solu_fix ] ) with @xmath547 .
the value of @xmath525 and the final time @xmath531 are the same as before .
( asterisks ) and @xmath12 ( circles ) for @xmath548 . on the left : euler ; on the right : two - step bdf for @xmath549 to @xmath550.,title="fig:",width=226 ] ( asterisks ) and @xmath12 ( circles ) for @xmath548 . on the left : euler ; on the right : two - step bdf for @xmath549 to @xmath550.,title="fig:",width=226 ] in figure [ ulti ] , on the left , we have represented the errors obtained using the implicit euler method as a time integrator for different values of the fixed time step @xmath485 ranging from @xmath549 to @xmath550 halving each time the value of @xmath485 . for the spacial discretization
we use the mini - element with always the same value of @xmath548 .
we use solid lines for the errors in the galerkin method and dashed lines for the estimations , as before .
the @xmath11 norm errors are marked with asterisks while the @xmath12 norm errors are marked with circles .
we estimate the errors using the postprocessed method computed with the same mini - element over a refined mesh of size @xmath551 . we observe that the galerkin errors decrease as @xmath485 decreases until a value that corresponds to the spatial error of the approximation . on the contrary ,
the error estimations lie on an almost horizontal line , both for the velocity in the @xmath11 and @xmath12 norms and for the pressure .
this means , as we stated in section 4.2 , that the error estimations we propose are a measure of the spatial errors , even when the errors in the galerkin method are polluted by errors coming from the temporal discretization . in this experiment
the error estimations are very accurate for the spatial errors of the velocity in the @xmath12 norm and for the errors in the pressure .
as commented above , the fact that postprocessing linear elements does not increase the convergence rate in the @xmath11 norm is reflected in the precision of the error estimations in the @xmath11 norm .
on the right of figure [ ulti ] we have represented the errors obtained when we integrate in time with the two - step bdf and fixed time step .
the only remarkable difference is that , as we expected from the second order rate of convergence of the method in time , the temporal errors are smaller for the same values of the fixed time step @xmath485 .
again , the estimations lie on a horizontal line being essentially the same as in the experiment on the left .
, _ gamma function and related functions in handbook of mathematical functions with formulas , graphs , and mathematical tables _ , edited by milton abramowitz and irene a. stegun .
reprint of the 1972 edition .
dover publications , inc . , new york , 1992 .
j. g. heywood and r. rannacher , _ finite element approximation of the nonstationary navier
stokes problem_. i. _ regularity of solutions and second - order error estimates for spatial discretization _ ,
siam j. numer .
, 19 ( 1982 ) , pp . 275311 .
, _ finite element approximation of the nonstationary navier stokes problem .
_ iii : _ smoothing property and higher order error estimates for spatial discretization _ , siam j. numer .
, 25 ( 1988 ) , pp . 489512 . | a posteriori estimates for mixed finite element discretizations of the navier - stokes equations are derived .
we show that the task of estimating the error in the evolutionary navier - stokes equations can be reduced to the estimation of the error in a steady stokes problem . as a consequence ,
any available procedure to estimate the error in a stokes problem can be used to estimate the error in the nonlinear evolutionary problem . a practical procedure to estimate the error based on the so - called postprocessed approximation
is also considered .
both the semidiscrete ( in space ) and the fully discrete cases are analyzed .
some numerical experiments are provided . |
the first truly large variability survey of extragalactic sources at mid - infrared ( mid - ir , 3.68.0 ) wavelengths is the _ spitzer
_ deep , wide - field survey ( sdwfs ; @xcite ) of the noao deep wide - field survey ( @xcite ) botes field .
the survey spans the years 2004 - 2008 with four epochs , covers 8 deg@xmath10 , and contains variability statistics for nearly half a million sources ( @xcite ) .
while the majority ( @xmath11% ) of the extragalactic variable objects in the mid - ir are active galactic nuclei ( agns ) , there is room for serendipitous discoveries such as supernovae ( sne ) . in general ,
sdwfs was not expected to be interesting for sne searches .
first , the contrast between sne and their hosts is relatively poor in the mid - ir , essentially because of differences in the effective temperatures .
if we match the bolometric luminosities of an sn and a typical host galaxy spectral energy distribution ( sed ) , the sn contributes @xmath12 of the v - band flux but only @xmath13 of the @xmath14 m flux .
second , with only four epochs of data , we expected sdwfs would add little to our understanding sns even if one were detected . third , the expected rates are very low . typical type ia ( iip ) sn peak at @xmath15 } \sim m_k \sim -18.4 $ ] ( @xmath16 ) mag ( e.g. , @xcite ) , so sdwfs can detect them in a single epoch only for redshifts @xmath17 ( @xmath18 ) . given the sne ia rates from the sdss - ii supernova survey ( @xcite ) and sne type ii rate ( @xcite ) , we would then expect to detect only of order one type ia sn and of order seven type ii sne in the sdwfs survey even if we could safely search for them at the @xmath19 detection level for a single epoch ( @xmath20=19.1 $ ] mag ) . in practice , controlling the false positive rates for identifying variable sources means that the variability selection criteria must be significantly more conservative than a @xmath19 peak at the detection limit ( see @xcite ) , so we did not expect to detect any sne in sdwfs . there is a subclass of type iin sne that are far more luminous , reaching @xmath15 } \sim m_k \sim -20 $ ] mag ( e.g. , ) , which we could detect up to @xmath21 . the leading theory for these
sne uses collisions between two massive shells of material to efficiently radiate the kinetic energy of the sn in the optical ( @xcite , @xcite ) .
the fast moving sn ejecta collides with an outer , slower moving shell on scales of order @xmath22 au where the optical depth is high enough to produce a well - defined photosphere but low enough for the thermal energy from the shock to be radiated before adiabatic expansion converts it back into kinetic energy .
but these sne are rare , representing only @xmath23 of the local type ii sne rate ( @xcite ) , which more than balances the increased detection volume . here
we investigate the nature the brightest mid - ir transient in sdwfs . over a 6 month period , this @xmath24 source radiated @xmath25 ergs at a black body temperature of @xmath6 k. in section [ sec : data ] we present the available uv , optical and mid - ir data along with a keck spectrum of the host galaxy .
then , in section [ sec : results ] we consider a range of possible scenarios for producing this source and conclude that a simple variant of the models for the hyper - luminous type iin sne is the best explanation for this source .
the paper is summarized in section [ sec : summary ] . throughout this paper
we use a standard @xmath26cdm model with @xmath27 and @xmath28wright / cosmocalc.html ] .
all magnitudes are in the vega system .
sdwfs - mt-1 ( aka sn 2007va ; @xcite ) , where mt means mid - ir transient as a parallel notation to using ot for optical transient , is the most significantly variable source in the entire sdwfs field .
it corresponds to the sdwfs source sdwfs [email protected] .
the [ 3.6 ] and [ 4.5 ] light curves were strongly correlated ( @xmath30 ) and the variability amplitude was 76 standard deviations from the mean for its average magnitude . a detailed definition of these variability criteria is given in @xcite .
the sdwfs survey consists of four epochs , taken 3.5 years , 6 months , and 1 month apart .
the object is not detected in any of the four irac bands in the first epoch ( also known as the irac shallow survey , @xcite ) on 2004 january 10 - 14 ( @xmath31 ) .
the sdwfs single epoch @xmath32 detection limits of @xmath20=19.67 $ ] mag , @xmath33=18.73 $ ] mag , @xmath34=16.33 $ ] mag , and @xmath35=15.67 $ ] mag ( 4 arcsec apertures corrected to 24 arcsec ; vega mag ) , set our upper limits on the object brightness in the first epoch .
the observed part of the transient peaked in the second epoch taken 2007 august 8 - 13 ( @xmath36 ) at @xmath20=15.93 $ ] mag and @xmath33=15.61 $ ] mag ( figures [ fig : sdwfs ] and [ fig : lightcurve ] ) .
it was slightly fainter 6 months later in epoch 3 ( 2008 february 2 - 6 , @xmath37 ) , with @xmath20=16.09 $ ] mag and @xmath33=15.76 $ ] mag .
then , in a matter of a month , the object faded by almost 2 magnitudes to @xmath20=17.99 $ ] mag and @xmath33=17.72 $ ] mag in epoch 4 on 2008 march 6 - 10 ( @xmath38 ) .
we present all the available _ spitzer _ data in table [ tab : sdwfsdata ] .
irac ( @xcite ) simultaneously observes the @xmath20/[5.8]$ ] or @xmath33/[8.0]$ ] bands , and in sdwfs the @xmath33/[8.0]$ ] observations were taken 40 seconds after the @xmath20/[5.8]$ ] ones .
each epoch consists of three 30 second exposures in each band , taken no more than 2 days apart , and the images for the whole mosaics were taken within 5 days .
we confirmed that the transient was present in the individual artifact - corrected basic calibrated data ( cbcd ) frames corresponding to the position of the transient .
the mosaic image for each epoch is then a combination of three to four of these frames , and there is no indication of problems in the sdwfs reductions ( see @xcite ) .
we obtained a keck / lris ( @xcite ) spectrum of the transient s host galaxy on 2010 march 12 , long after the transient was gone .
the spectrum ( shown in figure [ fig : keck_spectrum ] ) shows clear emission lines of h@xmath39 , h@xmath40 , h@xmath41 , h@xmath42 , h@xmath43 , and h@xmath44 , [ oii ] at 3727 , [ o iii ] at 4363 , 4959 and 5007 , [ neiii ] at 3869 and 3968 .
the redshift of the host galaxy is @xmath45 , so the luminosity distance is 920 mpc , the distance modulus is 39.82 mag , and the angular scale is 3.15 kpc / arcsec . before measuring the line fluxes , we corrected the host galaxy spectrum for galactic extinction ( e(b@xmath46v@xmath47 mag ) using @xcite .
we do not detect the [ nii ] 6584 line , which implies log([nii]/h@xmath48 . combining it with log([oiii ] 5007 / h@xmath49
, we conclude that the host is an irregular hii galaxy ( see , e.g. , ) .
the lack of [ nii ] and [ nev ] 3426 lines and the clear detection of the 4363 line rule out the presence of an agn .
we followed the prescription of @xcite to derive the gas - phase oxygen abundance of the host .
we measured the fluxes of the [ oii ] 3727 , [ oiii ] 4363 , [ oiii ] 4959 , [ oiii ] 5007 and h@xmath40 4861 emission lines to find a metallicity of @xmath50 .
we assumed an electron density of @xmath51 @xmath52 , but the value does not change significantly with changes in the electron density ( e.g. , @xmath53 if we use @xmath54 @xmath52 instead ) .
next , we verified this metallicity measurement using the prescriptions of to get @xmath55 . as a cross check
, we also used the oxygen - to - hydrogen flux ratio @xmath56+[oiii]})/{\rm h}\beta=(f_{\lambda3727}+f_{\lambda4959}+f_{\lambda5007})/f_{\lambda4861}$ ] method of @xcite and the @xcite o3n2 method to derive oxygen abundances .
the clear detection of the 4363 line and log([nii]/h@xmath48 strongly suggests that we should use the `` lower '' metallicity branch in the @xmath57 method .
we use the relations from @xcite and with our measured @xmath58 , to get oxygen abundances of @xmath59 and @xmath60 , respectively .
the o3n2 method gives @xmath61 .
all these estimates are approximately 10% of the solar value ( assuming 8.8 from @xcite ) . in order to understand the origin of this mid - ir transient , we searched for any available uv and optical data for this area of the sky in the relevant time span .
the host galaxy was detected in the original noao deep wide - field survey ( ndwfs , @xcite ) with @xmath62 mag ( @xmath63 ergs / s ) , @xmath64 mag ( @xmath65 ergs / s ) , and @xmath66 mag ( @xmath67 ergs / s ) .
it was unresolved in @xmath68 arcsec ( fwhm ) seeing , which puts an upper limit on the size of the galaxy of 3.8 kpc . in archival galex ( @xcite )
data , the host has nuv@xmath69 mag and no detection in fuv . the galex source is formally 087 from the sdwfs / ndwfs source , but there are no other blue sources within 20 arcsec .
both the colors of the host and its size are consistent with a blue , low luminosity , low metallicity galaxy .
we used the template model of an irregular galaxy from @xcite to estimate the mid - ir magnitudes of the host .
fitting the galex nuv , and ndwfs @xmath70 , @xmath71 , and @xmath72-band magnitudes of the host galaxy , and assuming @xmath45 , the expected apparent ( absolute ) irac magnitudes are @xmath20 = 20.87~(-18.56)$ ] mag , @xmath33 = 20.69~(-18.57)$ ] mag , @xmath34 = 20.52~(-20.02)$ ] mag , and @xmath35 = 18.37~(-21.45)$ ] mag .
this is consistent with the failure to detect the host in the first sdwfs epoch .
we estimate rest frame host luminosities of @xmath73 mag and @xmath74 mag , making this an @xmath75 galaxy ( @xcite ) .
although the field was monitored by at least 5 optical variability surveys , the peak of the transient occurred when the field was `` behind '' the sun , so almost all the optical data corresponds to the periods before epoch 2 and after epoch 4 .
the botes field was observed seven times between @xmath76 and @xmath77 with the 8k@xmath788k mosaic ccd camera on the 2.4 meter mdm hiltner telescope .
the two most interesting observations took place 90 days before the second and 84 days after the fourth sdwfs epoch , bracketing the period of the transient . in these 180 second exposures
, we detected no flux from the transient ( or the host ) at a @xmath32 level of @xmath79 mag ( @xmath80 ergs / s ) and @xmath81 mag ( @xmath82 ergs / s ) , respectively . the catalina sky survey ( css , @xcite)/catalina real - time transient survey ( crts , @xcite ) provides light curves for the period prior to epoch 2 and overlapping epochs 3 and 4 .
each epoch of four 30 second exposures on the 0.7 meter catalina schmidt telescope reaches an r - band magnitude limit of @xmath83 mag .
these data provide the strongest limit on the optical - to - mid - ir flux ratios , where a stack of 12 images ( 3 epochs ) , gives a @xmath32 upper limit on the transient of @xmath84 mag , corresponding to @xmath85 ergs / s ( large dark green symbol to the right in figure [ fig : lightcurve ] ) , in the period between epochs 3 and 4 . the quest ( @xcite ) survey further limits the optical emission before epoch 2 and after epoch 4 .
there are 19 epochs each consisting of 1 to 4 , 60 - 100 sec exposures with the 1.2 m samuel oschin telescope in a wide red / ir filter that we calibrated to @xmath86-band .
the most interesting data were taken @xmath87 days prior epoch 2 , where we stacked six adjacent quest epochs ( @xmath88@xmath89 ) to obtain @xmath32 detection limit of @xmath90 mag ( @xmath91 ergs / s , large orange symbol in figure [ fig : lightcurve ] ) . similarly , by coadding the high cadence raptor ( rapid telescopes for optical response , @xcite ) observations of the field ,
each of which is a 30 sec exposure using the raptor - p array of four 200-mm canon telephoto lenses , we obtained 15 upper limits at various epochs between 2004 and 2008 , before and after the transient ( figure [ fig : lightcurve ] ) .
calibrating the unfiltered observations to an @xmath71-band equivalent using tycho 2 stars , we find no evidence for the transient , with @xmath92 mag ( @xmath32 ) .
only the northern station of the all sky automated survey ( asas ) data ( @xcite ) obtained data during the peak of the transient .
combining 7 high quality asas images during the transient ( @xmath93@xmath94 ) , we obtain a weak upper limit on the transient magnitude of @xmath95 mag ( @xmath32 ) , corresponding to @xmath96 ergs / s .
the steeply falling , roughly power law sed of the transient ( figure [ fig : transientsed ] ) is suggestive of thermal emission by a source cooler than normal stars .
if we fit the sed as dust emission using an emissivity of @xmath97 , we find good fits with @xmath98 as is typical of the mid - ir emission by warm dust ( e.g. ) .
the temperature of @xmath99 k varies little between epochs ( table [ tab : sdwfsdata ] ) and is below typical dust destruction temperatures of @xmath100 k ( e.g. , @xcite ) .
table [ tab : sdwfsdata ] summarizes these estimates of the temperatures and luminosities .
a zeroth order estimate ( ignoring the dust emissivity ) of the radius @xmath71 of the dust photosphere is @xmath101 we are assuming that the shell is optically thick and @xmath71 corresponds to the radius where the ( optical ) optical depth is of order unity .
if the dust is optically thin , as the case for typical dust echos , there is no characteristic radius or temperature since photons can be absorbed and emitted at any distance from the sn .
we discuss these issues in more detail below . for the second and third epochs of sdwfs , with total luminosity @xmath102
ergs / s and temperature @xmath99 k , we find the radius to be @xmath103 au ( 0.05 pc ) .
the total energy radiated in the ir is then @xmath104 ergs .
the weak optical limits mean that a comparable amount of energy might have been radiated in the optical .
we used the more sophisticated radiation transport model dusty ( @xcite ) to check this estimate . for incident black body spectra with temperatures 10,00020,000 k , shell geometries with constant density , and dust temperatures at the inner photosphere edge of 13002000 k , we could match the epoch 2 and 3 seds , including a weak optical limit , using shells dominated by graphite , a visual optical depth of @xmath1053 , and a radius @xmath106@xmath107 au .
pure silicate dusts were unable to fit the data well , producing the wrong spectral shape for reasonable @xmath108 and inner dust temperatures .
these results are consistent with our simple estimates and strongly suggest that there was an optical counterpart , albeit with significant extinction ( @xmath109 mag ) . before outlining our preferred explanation of the transient as a self - obscured supernova
, we briefly outline the hypotheses we rejected .
first , the transient can not be explained by a galactic star .
the arguments against are as follows : ( 1 ) based on the keck spectrum , we know that the source is spatially coincident with an irregular galaxy ; ( 2 ) few stars can brighten by at least 5 magnitudes ( by more than a hundred times ) in the mid - ir and then stay at this level for 6 months before fading ; and ( 3 ) most flaring stars do not flare in the mid - ir but in the uv .
there are exceptions to ( 2 ) and ( 3 ) , and these will generally be accreting systems where the disk can alter the spectral properties . however , the ndwfs field is a high latitude ( @xmath110 ) field with negligible galactic extinction and unlikely to contain young ( protostars ) or massive ( be stars ) with these characteristics , particularly if they must also have no optical counterpart in quiescence despite needing to be nearby sources because of the short path length out of the disk .
the second possibility is that we are really observing an agn .
first , we note that there are no signs of agn activity in the host spectrum .
the narrow lines in a galaxy are on relatively large scales , and should show line ratios averaged over decades ( or more ) of agn activity , so the lack of a spectrum during the transient does not invalidate using them to argue against an agn .
moreover , the mid - ir colors at the peak of the transient , @xmath20-[4.5]=0.3 $ ] mag and @xmath34-[8.0]=0.0 $ ] mag ( epoch 2 ) , are inconsistent with that of an agn ( e.g. , @xcite ) .
while the composite sed of an agn and its host can have these colors , we know from the host flux limits in the first epoch that the sed at the peak should be dominated by the agn , and a pure agn at @xmath111 should have mid - ir colors closer to @xmath20-[4.5 ] \simeq 0.8 $ ] mag and @xmath34-[8.0 ] \simeq 1.1 $ ] mag ( @xcite ) .
we also considered the possibility of a tidally disrupted star in the context of the super - eddington accretion models of @xcite . while these models have an inflated , relatively cool envelope around the black hole , the expected temperatures of @xmath112 m@xmath113 k for a black hole of mass @xmath114 are still too high . a black hole with low enough mass ( @xmath115 m@xmath116 ) to match the temperature ( @xmath117 k ) would be unable to produce the observed luminosity .
the eddington luminosity for such a low mass is only @xmath118 ergs / s .
the third possibility is that it is a @xmath41-ray burst ( grb ) .
we can rule it out as direct emission from a burst because the time scale ( 6 months ) is too long and the sed is wrong for direct beamed emission ( e.g. , @xcite ) .
it can not be indirect emission , where a dusty region absorbs and reradiates emission from the jet , because a relativistic jet expands past the scale of the photosphere ( @xmath119 au ) too quickly ( @xmath120 rest - frame days ) to produce the observed event duration .
finally , this can not be gravitational lensing of a background object by the host galaxy ( see @xcite for a review ) .
the time scales are such that it would have to be microlensing by the stars in the lens galaxy , but there are no known mid - ir sources that are sufficiently compact to show such a large , and apparently achromatic , degree of magnification .
moreover , almost any potential source would have to show a still stronger degree of optical magnification that would disagree even with our weak optical limits ( see e.g. , @xcite ) .
our working hypothesis is that sdwfs - mt-1 is a supernova analogous to sn 2006gy , an extremely luminous type iin sn ( @xcite ) .
sn 2006gy peaked at @xmath121 mag , corresponding to a luminosity of @xmath122 ergs / s , and it radiated of order @xmath123 ergs in total ( @xcite ) . in a normal type ii sn , a very small fraction ( @xmath124% ) of the available energy is radiated .
the initial energy from the shock is converted into kinetic energy by adiabatic expansion , leaving little to power the luminosity .
thus , a supernova ejecting @xmath125 at velocities of @xmath126 km / s contains enough kinetic energy , @xmath127 ergs to power the transient , but it is only radiated on time scales of order 10@xmath128 years as the sn remnant develops .
the scenario we present was proposed by @xcite for sn 2006gy with further elaborations in @xcite and @xcite .
very massive stars ( @xmath129 ) are known to undergo impulsive events ejecting significant fractions of the stellar envelope .
the most famous example is the eruption of @xmath8 carina in 1837 - 1857 ( see ) , but also includes some of the so - called `` supernova impostors '' ( e.g. @xcite ) .
the number and temporal distributions of these mass ejections relative to the final supernova is unknown , but there are several cases where one seems to have occurred shortly before the final supernova ( e.g. , @xcite ) .
pair instability sne can also produce a sequence of mass ejections just prior to the death of the star ( see @xcite ) .
let us suppose that the progenitor of sdwfs - mt-1 was such a massive star and had two such eruptions in its final phases , roughly 300 and 5 years before the final explosion , each of which ejected of order @xmath130 . for a typical ejection velocity of @xmath131 km / s ( e.g. @xcite )
, these shells would lie roughly 13000 and 200 au from the star when it exploded .
the role of the first shell is to convert the kinetic energy of the explosion into radiation following @xcite . roughly speaking
, the collision between the ejected material and the first shell can extract fraction @xmath132 of the kinetic energy and convert it to radiation .
the distance to the shell is tuned so that the shock crossing time is comparable to the thomson photon diffusion time , thereby allowing the radiation to escape without significant adiabatic losses .
this requires a total mass of approximately @xmath133 where @xmath134 au and @xmath135 km / s .
we can get up to @xmath136 either by making @xmath137 au or by allowing for adiabatic losses .
this then implies characteristic time , energy , luminosity and temperature scales of @xmath138 days , @xmath139 ergs , @xmath140 ergs / s , and @xmath141 k , respectively , where we have used the estimate of the total mass and equally divided it between the shell and the sn ejecta .
the basic energetics , but not the radiation transport , are largely confirmed by the simulations of @xcite .
this set our choice of incident temperature in the dusty models .
thus , a shell of ejected material at @xmath142 au can produce the necessary energetics and luminosities but can not match the observed time scales ( too short ) or temperatures ( too high ) .
the role of the second shell is to match the time and temperature scales . @xcite and
@xcite invoke a second @xmath143 shell located at @xmath144 pc in order to explain the relatively bright , long - lived ir emission observed for sn 2006gy .
their shell is , however , optically thin to optical light ( @xmath145 ) because the total infrared emission is only a small fraction of the total .
the optical depth of a shell is of order @xmath146 for @xmath147 in cm@xmath10/g , which leads to an optically thin shell with @xmath149 for an @xmath150 pc shell radius and @xmath151 cm@xmath10/g for a radiation temperature of order @xmath152 k ( e.g. , ) .
this opacity , @xmath153 cm@xmath10/g , corresponds to a dust - to - gas ratio of approximately 1% , which arguably may be high for a low metallicity galaxy .
note , however , that sn 2006gy is probably also associated with a low metallicity galaxy ( see 3.4 ) and the shell producing its dust echo contains enough dust for our scenario . to explain sdwfs - mt-1 we simply move the shell inward . at this point , however , we should improve on eqn .
( [ eq : radius ] ) with a simple model for an optically thick dust echo .
suppose that the dust lies in an optically thick spherical shell of radius @xmath71 and is exposed to luminosity @xmath154 for a period of time @xmath155 . because all the dust is exposed to the same luminosity , the dust temperature is independent of position or time while irradiated .
the observer , however , sees the luminosity pulse modulated by the light travel time @xmath156 across the shell . in this simple model ,
the light curves are trapezoids with linear rises and falls connected by a plateau created by the changing fraction of the shell illuminated by the outburst from the observer s perspective , but the dust temperature is constant .
models that can reproduce the drop in luminosity between epochs 3 and 4 have the pulse time @xmath157 , with the simplest model having the luminosity plateau run from epoch 2 to epoch 3 . the plateau duration is @xmath158 days while the rise and falls last @xmath155 .
the luminosity of the plateau is @xmath159 , which is smaller than @xmath160 by the ratio @xmath161 between the light crossing time and the pulse length ( if @xmath162 the plateau luminosity is @xmath154 ) .
the factor of @xmath163 drop in luminosity between epochs 3 and 4 is @xmath164 where @xmath165 days is the time between these epochs .
thus , we have @xmath166 days , and @xmath167 au or @xmath168 light days . with these time and distance
scales we come close to matching the epoch 2 and 3 luminosities given the observed temperatures , with an estimated luminosity of @xmath169 ergs / s .
the true luminosity is @xmath170 ergs / s , which is why the radius estimate is somewhat larger than the estimates of eqn .
( [ eq : radius ] ) .
there is some tension between the plateau luminosity and the drop in luminosity between the last two epochs , but the overall agreement is good given the crudeness of the model both for the dust emission and the luminosity profiles .
the mass required to produce an a shell with optical depth @xmath108 at this distance is @xmath171 , so there is little difficulty having an optically thick shell for masses of order @xmath172-@xmath173 even if the dust - to - gas ratio needs to be significantly reduced from @xmath174 because of the low metallicity of the galaxy .
while the dust temperature is close to that for destroying dust ( e.g. , @xcite ) , the situation is somewhat different from a normal supernova because the ( thomson ) optically thick inner shell protects the outer from any initial high luminosity bursts of hard radiation . for the ir - radiation , @xmath175 cm@xmath10/g , and
the shell is optically thin to the mid - ir emission for reasonable shell masses , so the dust emission streams freely to the observer .
given the estimated pulse duration , @xmath166 days , and luminosity , @xmath176 ergs / s , we can estimate that the inner shell is at @xmath177 au , the velocity must be @xmath178 km / s and the mass involved must be @xmath179 , where the rapid decline pushes these values to small radii / masses and high velocities in order to keep @xmath155 short . given the crudeness of the model , it is remarkable how well all the observational features can be matched using relatively sensible physical parameters .
the nature of the host galaxy indirectly supports this hypothesis because both grbs and most hyper - luminous type iin sne are also associated with low luminosity , low metallicity dwarf galaxies .
figure [ fig : stanek ] , similar to figure 1 in @xcite , shows the position of the host galaxy for sdwfs - mt-1 in the metallicity - luminosity plane in relation to @xmath180 star - forming sdss dr4 galaxies ( @xcite ) .
we also include the hosts of six local ( @xmath181 ) grbs using updated estimates from @xcite and including the host of grb 100316d ( @xcite ) . while sdwfs - mt-1 was not a grb ,
its occurrence in a very low metallicity galaxy is striking , since such galaxies produce only a small fraction of massive stars in the local universe ( e.g. , see figure 2 in @xcite ) . in figure
[ fig : stanek ] , we ignore small ( @xmath182% ) effects on metallicity measurements arising from the method being used to derive it ( see @xcite ) .
in fact , the hosts of most other hyper - luminous sne also tend to be very low luminosity ( @xmath183 ) , suggestive of their low metallicity ( unfortunately , oxygen abundances have not been measured for most of these hosts ) .
sn 2005ap ( @xcite ) at @xmath184 peaked at unfiltered magnitude of @xmath185 mag and is associated with a @xmath186 ( @xmath187 ) mag dwarf galaxy .
sn 2006tf ( @xcite ) at @xmath188 exploded in a host of low luminosity @xmath189 ( @xmath190 ) mag .
the over - luminous @xmath191 mag type ic sn 2007bi exploded in a faint , @xmath192 mag dwarf galaxy , with @xmath193 ( ) .
sn 2008es at @xmath194 ( @xcite ) reached @xmath195 mag and exploded in a host which was not detected down to a faint luminosity of @xmath196 ( @xmath197 ) mag .
sn 2008fz at @xmath198 peaked at @xmath195 mag , but no host has been detected down to an upper limit of @xmath199 ( @xmath200 ) mag ( @xcite ) .
sn 2006gy ( @xcite ) is one of two exceptions here , as it exploded in a fairly luminous s0 galaxy , ngc 1260 with @xmath201 mag .
another example is sn 2003ma ( @xcite ) that released @xmath202 ergs in a @xmath203 mag , @xmath204 galaxy .
the general message is clear : grbs and luminous sne prefer low metallicity , low luminosity galaxies , so does sdwfs - mt-1 .
we systematically searched for transients similar to sdwfs - mt-1 and sne in general in the sdwfs data .
we considered highly correlated @xmath20 $ ] and @xmath33 $ ] light curves ( @xmath205 ) with flag @xmath206 ( photometry unaffected by the wings of bright stars ) , variability significances @xmath207 } > 1 $ ] and @xmath208}>1 $ ] and colors @xmath20-[4.5]<0.4 $ ] to eliminate agn ( see @xcite ) .
we selected as candidates objects that at their peak were brighter than @xmath20<18.67 $ ] mag , corresponding to 1 mag brighter than the @xmath32 detection limit , and which changed in brightness by at least one magnitude between either epochs 1 and 2 or epochs 2 and 3/4 .
the latter two epochs are close enough together that an sn might not show a significant brightness change between them .
we found six candidate objects , one of which is sdwfs - mt-1 .
the next four of them are artifacts due to psf structures ( long spikes ) from bright stars .
the remaining object , sdwfs j142557.64 + 330732.6 ( hereafter sdwfs - mt-2 ) , peaks in epoch 2 close to our selection limit at @xmath20=18.52\pm0.14 $ ] mag and @xmath33=18.71\pm0.33 $ ] mag .
it lies @xmath209 south - west of a potential host galaxy with _ galex _
@xmath210 mag , ndwfs @xmath211 mag , @xmath212 mag , and @xmath213 mag that is not detected in sdwfs . in the deep images from ndwfs
, the host galaxy appears to be an extended , possibly edge - on ( length @xmath214 arcsec in @xmath68 arcsec seeing ) disk , with a redder core and bluer outer regions .
the transient lies in the bluer outer regions , although its position is only known to @xmath215 due to its faintness .
we estimate a photometric redshift of @xmath216 for the host using the template models of @xcite , corresponding to a comoving distance of @xmath217 gpc and a linear scale of @xmath218 kpc / arcsec .
this translates into the linear diameter of the host of @xmath219 kpc .
the best fit model for the sed of the host is a mixture of an agn and a starburst template .
the mid - ir color of the transient , @xmath20-[4.5]=0.2\pm0.5 $ ] mag , is ambiguous on the question of an sn or agn origin for the transient , although the offset position would favor an sn .
given the redshift estimate , however , the implied luminosity is too high even for the brightest normal sn for @xmath15 } \sim m_k \sim -20 $ ] mag ( ) the peak magnitude of @xmath20 \sim 24 $ ] mag would be far below our detection limits .
a second event like swfs - mt-1 would have to be even more extreme .
given the paucity of the data , interpreting this source as a false positive seems the most probable explanation .
for events with six month durations like sdwfs - mt-1 our survey time coverage is approximately @xmath220 years .
sdwfs - mt-1 peaked in epoch two at fluxes of @xmath221 mjy in [ 3.6 ] and @xmath222 mjy in [ 4.5 ] at a luminosity distance of @xmath223 mpc .
the peak flux limit for our systematic search is @xmath224jy ( @xmath19 in [ 4.5 ] ) which means we could have detected sdwfs - mt-1 at a luminosity distance of @xmath225 mpc ( @xmath226 ) or about @xmath227 ignoring @xmath228-corrections , which will initially aid detection because the sed peaks bluewards of the irac bands .
thus , we could detect the source in a comoving volume of approximately @xmath229 mpc@xmath128 . combining these factors
, we estimate a rate for sdwfs - mt-1-like transients of @xmath230 mpc@xmath231 yr@xmath232 where we have not corrected the time scales for redshifting effects given the otherwise large uncertainties from the statistics of one event .
this is approximately @xmath233 of the type ii sne rate ( @xcite ) , @xmath234 of the hyper - luminous type ii sne ( @xcite ) , and @xmath7 of the grb rate ( @xcite ) .
in this paper , we presented the discovery of a luminous mid - ir transient , sdwfs - mt-1 , that is unusual in several ways .
its most striking features are the high luminosity , @xmath235 ergs / s , the low apparent temperature , @xmath236 k , the long , 6 month duration of the event , and the resulting high amount of radiated energy , @xmath237 ergs . a model that explains the transient is a self - obscured supernova , that has two concentric ( patchy ) shells , each with the mass of @xmath7 m@xmath116 , ejected prior to the sn explosion . the inner shell at @xmath238 au protects the dust in the outer shell from initial luminosity spikes and converts the high kinetic energy of the sn shock wave into thermalized radiation with @xmath239 k. the dust in the more distant ( 0.05 pc ) second shell absorbs the radiation of the first shell , and reradiates it at @xmath240 k , while simultaneously smoothing out and stretching the light curve .
this scenario matches the observed energy , luminosity , time and temperature scales of the transient by slightly modifying the scenarios used for other hyper - luminous type iin sne ( e.g. , @xcite ) .
the host galaxy of the transient is also typical of these sne , and our estimate of the rate is consistent with such events being @xmath7% of hyper - luminous type ii sne . the discovery of the progenitors of sn 2008s and ngc300-ot-1 ( @xcite ) as self - obscured ( @xmath241 k , @xmath242 ) stars already indicated that mid - ir observations are necessary to understanding the fates of massive stars .
if our interpretation of sdwfs - mt-1 is correct , then this may extend to the actual transients as well as to their progenitors , although it is quite likely that sdwfs - mt-1 was also a significant optical transient .
ir dust echos from sne are relatively common , ( e.g. , @xcite ) , but in these cases the optical depth of the dust is low and the ir emission is only a fraction of what is observed in the optical . a few recent sne with significant echo luminosities are sn 2007od ( @xcite ) , sn 2006gy ( @xcite ) , sn 2004et ( @xcite ) , and sn 2002hh ( @xcite ) .
obviously there is a strong selection effect against optically selecting sne where the optical depth of the surviving dust is near unity , so the true abundance of such sources can only be determined in near / mid - ir searches .
there have been some searches made in the near - ir ( e.g. , ) , principally focused on the problem of unassociated foreground absorption rather than dust associated with the progenitor star .
two important steps are to try and identify massive stars with such dense shells of ejected material , similar to the search for self - obscured stars in @xcite and @xcite , and to better characterize the mid - ir emission of other hyper - luminous type iin sne .
if the inner shell invoked to convert the kinetic energy into radiation is dusty , the progenitor will appear as a rapidly cooling near / mid - ir source as the shell expands .
such shells will have temperatures of order the 400 k observed for sn 2008s or ngc300-ot-1 but should show the time variability known to be absent for the sn 2008s progenitor ( see @xcite ) . for the dust to survive the sn explosion and produce a powerful mid - ir transient
, it must be considerably more distant and colder than the @xmath243 k dust photosphere of the progenitors of sn 2008s or ngc300-ot-1 , with temperatures of order @xmath244 k and seds peaking at 20 - 30@xmath9 m that will make them invisible to warm _ spitzer_. in fact , the progenitors may well resemble a moderately colder version of @xmath8 carina , which radiates most of its energy in the mid - ir and has an sed peaking at @xmath245 m ( @xcite ) .
the temperature of such a distant shell will evolve slowly because the time scale for the expansion of the shell is now quite long . warm _
spitzer _ will remain a valuable tool for identifying warm , inner shells , but a careful search for cold outer shells will need to wait for the _ james webb space telescope _
( _ jwst _ ) .
the _ wide - field infrared survey explorer _ ( _ wise _ , @xcite ) , launched in december 2009 , is currently carrying out an all - sky survey at similar wavelengths to _ spitzer _ , as a natural by - product of its design , find many transients similar to sdwfs - mt-1 . with a @xmath19 sensitivity of @xmath221 mjy at 3.4 and @xmath246 mjy at 4.7 for a region with 8 exposures ( @xcite ) , _ wise _
has far less sensitivity than the sdwfs survey . crudely speaking the survey volume compared to sdwfs scales as the ratio of the @xmath247 of the two surveys , so @xmath248 and the greater solid angle wins over the reduced sensitivity . if _
wise _ completes two full years of observations it will also have 4 epochs , 1 epoch every 6 months , and so should detect a significant number of these mid - ir transients ( @xmath249 ) independent of their origin . for sources closer to the poles of the _ wise _ scan pattern , _ wise
_ will be able to produce detailed light curves . on longer time scales ,
_ jwst _ will be able to identify and monitor such luminous transients at almost any redshift .
we thank the anonymous referee , whose comments helped us to improve the manuscript .
this work is based on observations made with the _
spitzer space telescope _ , which is operated by the jet propulsion laboratory ( jpl ) , california institute of the technology ( caltech ) under contract with the national aeronautics and space administration ( nasa ) .
support for this work was provided by nasa through award numbers 1310744 ( c.s.k . and s.k .
) , 1314516 ( m.l.n.a . ) issued by jpl / caltech .
, t.a.t . , and s.k .
are also supported by national science foundation ( nsf ) grant ast-0908816 . j.l.p .
acknowledges support from nasa through hubble fellowship grant hf-51261.01-a awarded by stsci , which is operated by aura , inc . for nasa , under contract nas 5 - 2655
this work made use of images and/or data products provided by ndwfs ( jannuzi and dey 1999 ; jannuzi et al .
2005 ; dey et al .
the ndwfs and the research of a.d and b.t.j .
are supported by the national optical astronomy observatory ( noao ) .
noao is operated by aura , inc .
, under a cooperative agreement with nsf .
the crts survey is supported by nsf under grants ast-0407448 and ast-0909182 .
the css survey is funded by nasa under grant no .
nng05gf22 g issued through the science mission directorate near - earth objects observations program .
arcmin@xmath10 , centered on the transient .
the rows correspond to the [ 3.6 ] , [ 4.5 ] , [ 5.8 ] , and [ 8.0 ] bands ( downward from top ) , and the columns show epochs 1 through 4 ( left to right ) , taken in 2004 , 2007 and twice in 2008 , respectively .
the object is marked with the cross hairs .
the source is clearly absent in epoch 1 and then present in all subsequent epochs except for the epoch 4 [ 8.0 ] data.,width=604 ] days [ 3.6 ] , @xmath250 day [ 4.5 ] , @xmath251 day [ 5.8 ] and , @xmath252 days [ 8.0 ] . the right @xmath253-axis shows the absolute magnitudes without @xmath228-corrections .
a normal type ii sn would peak at @xmath254 mag at 3.6 .
the gray area marks the region when the mid - ir transient was not visible to airmass @xmath255 .
the large dark green symbols mark the crts upper limits found by stacking the three nearest epochs ( small symbols ) into one image .
the raptor upper limits ( purple symbols ) are stacks of 2162 images .
the quest upper limits ( orange ) are derived based on 14 stacked images per epoch .
the large symbol for quest combines the four nearest epochs .
bottom panel : points from the upper panel converted to luminosity as @xmath256.,width=604 ] ) to the four irac bands ( filled circles ) and find a best fitting temperature of @xmath257 k ( dotted line ) .
we also show a template sed for the host galaxy modeled as an irregular ( dashed line ; @xcite ) and the combination of the modified black body and irregular galaxy seds ( thick solid line ) .
overplotted are the v , r and irac filter transmission curves , `` normalized '' to @xmath258 ergs / s .
the mdm ( quest ) limit is derived for an observation taken 90 ( @xmath87 ) days prior to the second epoch of sdwfs .
the crts limit corresponds to the stack of 3 epochs ( @xmath259 sec images ) taken between sdwfs epochs 3 and 4 .
the open squares mark the ndwfs @xmath260 , @xmath71 , and @xmath72-band measurements for the host galaxy ( from left to right).,width=604 ] bins ) are drawn for 1 , 10 , 50 , 100 and 200 objects per bin , counting from the outer contour .
the figure is analogous to figure 1 from @xcite .
sdwfs - mt-1 is shown as the filled pentagon ( lower left ) .
a handful of local grbs with associated sne ( and two sne ) with known hosts metallicities and luminosities are shown with open circles ( small pentagons ) .
the arrows show @xmath261 for five host galaxies of high luminosity sne with unknown oxygen abundances .
they are , from left to right , sn 2005ap , sn 2006tf , sn 2008fz , sn 2008es , and sn 2006gy ( see section [ sec : results : host ] ) .
the lmc / smc oxygen abundances are adopted following @xcite and milky way s ( solar ) are from @xcite , while the absolute magnitudes are from @xcite.,width=529 ] l|cccc @xmath20 $ ] ( mag ) & @xmath262 & @xmath263 & @xmath264 & @xmath265 + @xmath33 $ ] ( mag ) & @xmath266 & @xmath267 & @xmath268 & @xmath269 + @xmath34 $ ] ( mag ) & @xmath270 & @xmath271 & @xmath272 & @xmath273 + @xmath35 $ ] ( mag ) & @xmath274 & @xmath275 & @xmath276 & @xmath274 + @xmath277)$ ] ( @xmath278 ergs / s ) & @xmath279 & @xmath280 & @xmath281 & @xmath282 + @xmath283)$ ] ( @xmath278 ergs / s ) & @xmath284 & @xmath285 & @xmath286 & @xmath287 + @xmath288)$ ] ( @xmath278 ergs / s ) & @xmath289 & @xmath290 & @xmath291 & @xmath292 + @xmath293)$ ] ( @xmath278 ergs / s ) & @xmath294 & @xmath295 & @xmath296 & @xmath294 + + + & + & + + black body temp .
@xmath297 & & @xmath298 & @xmath299 & @xmath300 + `` total '' @xmath301 ( @xmath302 ergs / s ) & & @xmath303 & @xmath304 & @xmath305 + radius @xmath71 ( au)@xmath306 & & @xmath307 & @xmath308 & @xmath309 + | we report the discovery of a six - month - long mid - infrared transient , sdwfs - mt-1 ( aka sn 2007va ) , in the _ spitzer _ deep , wide - field survey of the noao deep wide - field survey botes field . the transient , located in a @xmath0 low luminosity ( @xmath1}\simeq-18.6 $ ] mag , @xmath2 ) metal - poor ( @xmath3 ) irregular galaxy , peaked at a mid - infrared absolute magnitude of @xmath1}\simeq-24.2 $ ] in the @xmath4 _
spitzer_/irac band and emitted a total energy of at least @xmath5 ergs .
the optical emission was likely fainter than the mid - infrared , although our constraints on the optical emission are poor because the transient peaked when the source was `` behind '' the sun .
the _ spitzer _
data are consistent with emission by a modified black body with a temperature of @xmath6 k. we rule out a number of scenarios for the origin of the transient such as a galactic star , agn activity , grb , tidal disruption of a star by a black hole and gravitational lensing .
the most plausible scenario is a supernova exploding inside a massive , optically thick circumstellar medium , composed of multiple shells of previously ejected material .
if the proposed scenario is correct , then a significant fraction ( @xmath7% ) of the most luminous supernova may be self - enshrouded by dust not only before but also after the supernova occurs .
the spectral energy distribution of the progenitor of such a supernova would be a slightly cooler version of @xmath8 carina , peaking at 2030@xmath9 m . |
it is known that lepton - hadron collisions have been playing a crucial role in exploration of deep inside of matter . today , thera
( tesla on hera ) and linac*lhc can be considered as realistic candidates for future lepton - hadron and photon hadron colliders .
we discuss the main parameters and physics search potential of the lepton - hadron colliders and draw the attention of the high energy and nuclear physics communities to future @xmath1 and fel @xmath2 collider facilities .
it is known that the tesla collider will be a powerful tool for exploraiton of the multi - hundred gev scale @xcite .
taking into account the possible polarized proton and nucleus options for hera will provide a number of additional opportunities to investigate lepton - hadron and photon hadron interactions at tev scale .
recently , the work on tesla tdr has been finished , and tesla@xmath3hera based @xmath4 and @xmath2 colliders are included into tesla project .
main parameters of tesla@xmath5hera based @xmath6 collider are given in @xcite .
it is seen that one has @xmath7 @xmath8 with @xmath9 gev and @xmath10 @xmath11 . also two additional versions (
@xmath12 @xmath13 with @xmath14 @xmath8 and @xmath15 @xmath16 @xmath13 with @xmath17 @xmath8 ) have been mentioned in tesla tdr . in principle , tesla@xmath3 hera based ep collider will extend the hera kinematics region by an order in both q@xmath18 and @xmath19 and , therefore , the parton saturation regime can be achieved .
main parameters and physics search potential of thera based @xmath20 * * * * collider are given in ciftci95 , brinkmann97 .
main limitations for @xmath21 option comes from fast emittance growth of nucleus beam due to intra - beam scattering . in our opinion @xmath2 option
is the most promising option of tesla@xmath3hera complex , because it will give unique opportunity to investigate small @xmath22 region in nuclear medium .
colliding of tesla fel beam with nucleus bunches from hera may give a unique possibility to investigate old nuclear phenomena in rather unusual conditions .
the main idea is very simple @xcite : ultra - relativistic ions will see laser photons with energy @xmath23 as a beam of photons with energy 2@xmath24 , where @xmath25 is the lorentz factor of the ion beam .
the region 0.1@xmath2610 mev , which is matter of interest for nuclear spectroscopy , corresponds to 0.1@xmath2710 kev lasers , which coincide with the energy region of tesla fel .
the center - of - mass energies which will be achieved at different options of this machine @xcite are an order larger than those at hera are and @xmath283 times larger than the energy region of tesla@xmath3hera .
center - of - mass energy and luminosity for this option are @xmath29 tev and @xmath30 @xmath31s@xmath32 this machine , which will extend both the @xmath33-range and @xmath34-range by more than two order of magnitude comparing to those explored by hera , has a strong potential for both standard model and new physics research . using @xmath20 * * * * option * * * * of * *
* * this collider thousands di - jets with @xmath35 gev and hundreds thousands single w bosons will be produced , hundred millions of @xmath36- and @xmath37- pairs will give opportunity to explore the region of extremely small x@xmath38 etc .
details on main parameters and physics search potential of linac*lhc based @xmath39 , @xmath2 and fel @xmath2 coliders can be found in brinkmann97,yavas00 .
the clic , an electron - positron collider with @xmath40 @xmath11 and @xmath41 @xmath42 is considered as one of the future options for post - lhc era at cern .
the work on clic*lhc based @xmath43 and fel @xmath2 options is under progress .
it seems that neither hera nor lhc@xmath3lep will be the end points for lepton - hadron colliders .
we see that tev scale linac - ring type @xmath6 machines will give an opportunity to go far in this direction ( see table 1 ) .
in addition , more knowledge on the subject can be found in @xcite . | main parameters of future lepton - hadron colliders are estimated .
namely , thera and linac*lhc based ep , @xmath0p , ea , @xmath0a and fel@xmath0a colliders are considered .
the physics search potential of these machines are considered .
= 14.5pt |
this work was partially supported by infn , murst and by the european commission program mrtn - ct-2004 - 005104 .
99 k. gdel , `` an example of a new type of cosmological solutions of einstein s field equations of gravitation , '' rev . mod .
* 21 * ( 1949 ) 447 .
r. c. myers and m. j. perry , `` black holes in higher dimensional spacetimes , '' annals phys . *
172 * , 304 ( 1986 ) .
g. w. gibbons and c. a. r. herdeiro , `` supersymmetric rotating black holes and causality violation , '' class .
* 16 * ( 1999 ) 3619 [ arxiv : hep - th/9906098 ] .
m. m. caldarelli , d. klemm and w. a. sabra , `` causality violation and naked time machines in ads@xmath0 , '' jhep * 0105 * ( 2001 ) 014 [ arxiv : hep - th/0103133 ] . j. r. i. gott , `` closed timelike curves produced by pairs of moving cosmic strings : exact solutions , '' phys .
* 66 * ( 1991 ) 1126 .
s. m. carroll , e. farhi , a. h. guth and k. d. olum , `` energy momentum restrictions on the creation of gott time machines , '' phys .
d * 50 * ( 1994 ) 6190 [ arxiv : gr - qc/9404065 ] .
a. ori , `` a class of time - machine solutions with a compact vacuum core , '' phys .
* 95 * ( 2005 ) 021101 .
s. w. hawking , `` the chronology protection conjecture , '' phys .
d * 46 * ( 1992 ) 603 .
b. s. kay , m. j. radzikowski and r. m. wald , `` quantum field theory on spacetimes with a compactly generated cauchy horizon , '' commun .
* 183 * ( 1997 ) 533 [ arxiv : gr - qc/9603012 ] .
m. visser , `` the quantum physics of chronology protection , '' arxiv : gr - qc/0204022 .
j. m. maldacena , `` the large n limit of superconformal field theories and supergravity , '' adv .
* 2 * ( 1998 ) 231 [ int .
j. theor .
* 38 * ( 1999 ) 1113 ] [ arxiv : hep - th/9711200 ] .
d. berenstein , `` a toy model for the ads / cft correspondence , '' jhep * 0407 * , 018 ( 2004 ) [ arxiv : hep - th/0403110 ]
. m. m. caldarelli and p. j. silva , `` giant gravitons in ads / cft .
i : matrix model and back reaction , '' jhep * 0408 * , 029 ( 2004 ) [ arxiv : hep - th/0406096 ] .
h. lin , o. lunin and j. maldacena , `` bubbling ads space and 1/2 bps geometries , '' jhep * 0410 * ( 2004 ) 025 [ arxiv : hep - th/0409174 ] .
m. m. caldarelli , d. klemm and p. j. silva , `` chronology protection in anti - de sitter '' class .
* 22 * ( 2005 ) 3461 .
[ arxiv : hep - th/0411203 ] .
m. cveti _ et al .
_ , `` embedding ads black holes in ten and eleven dimensions , '' nucl .
b * 558 * ( 1999 ) 96 [ arxiv : hep - th/9903214 ] .
k. behrndt , a. h. chamseddine and w. a. sabra , `` bps black holes in n = 2 five dimensional ads supergravity , '' phys .
b * 442 * , 97 ( 1998 ) [ arxiv : hep - th/9807187 ] .
k. behrndt , m. cveti and w. a. sabra , `` non - extreme black holes of five dimensional n = 2 ads supergravity , '' nucl .
b * 553 * , 317 ( 1999 ) [ arxiv : hep - th/9810227 ] .
r. c. myers and o. tafjord , `` superstars and giant gravitons , '' jhep * 0111 * , 009 ( 2001 ) [ arxiv : hep - th/0109127 ] .
g. milanesi and m. oloughlin , `` singularities and closed time - like curves in type iib 1/2 bps geometries , '' jhep * 0509 * , 008 ( 2005 ) [ arxiv : hep - th/0507056 ] . j. c. breckenridge , r. c. myers , a. w. peet and c. vafa , `` d - branes and spinning black holes , '' phys . lett .
b * 391 * ( 1997 ) 93 [ arxiv : hep - th/9602065 ] . c. a. r. herdeiro , `` special properties of five dimensional bps rotating black holes , '' nucl .
b * 582 * ( 2000 ) 363 [ arxiv : hep - th/0003063 ] . | we review the issue of chronology protection and show how string theory can solve it in the half bps sector of ads / cft . according to the llm prescription , half bps excitations of ads@xmath0 @xmath1 @xmath2 geometries in type iib string theory
can be mapped into free fermion configurations .
we show that unitarity of the theory describing these fermions is intimately related to the protection of the chronology in the dual geometries .
the principle of causality stands at the foundation of all our understanding of physics .
it can not be relaxed without falling into paradoxes which , in the final analysis , consist either in the attempt to change the past conditions at the origin of the event or in the so - called bootstrap paradox , occurring when the effect is the cause of its own existence .
the concept of causality is so important that it has been hardcoded into newtonian physics and special relativity .
but as soon one takes into account the gravitational force , space and time become dynamical .
their evolution is driven by local equations and the equivalence principle ensures that the causality requirements of special relativity are locally respected .
however , the gravitational field equations do not put any global constraint on the topology of time , leaving the possibility for the existence of closed timelike curves ( ctcs ) , seeds of the violation of causality .
the global nature of ctcs is what makes them so elusive and difficult to treat .
essentially , ctcs can arise by two mechanisms , rotation ( or dragging of the inertial frames ) and effectively superluminal propagation .
the most well - known example of non topologically trivial ctcs induced by rotation is given by the gdel universe @xcite , a homogeneous and simply connected manifold described by the metric @xmath3 here , @xmath4 is a periodic angular coordinate , whose orbits become timelike for large @xmath5 due to the frame - dragging term in the metric and form therefore ctcs . the fact that these ctcs exist at all times is a reason one can invoke to discard this solution .
another example is the kerr black hole that has harmless ctcs induced by its rotation , but they are hidden by the cauchy horizon , which is unstable .
the situation is worse in higher dimensions , where naked ctcs are a generic feature of kerr black holes @xcite .
it should be noted that supersymmetry does not help to rule out chronological pathologies @xcite , nor does the lifting to higher dimensional supergravities @xcite .
the other mechanism to create ctcs is to construct geometries allowing for effectively superluminal propagation of signals .
for example , gott s time machine @xcite consists in two parallel cosmic strings in relative motion .
the effect of the cosmic strings on the spacetime is to create conical defects , which can be cleverly exploited to build ctcs using the properties of lorentz transformations ( see figure [ fig - gott ] ) .
again , this solution has problems , for example its spacelike total energy momentum , which make it unphysical @xcite .
however , ori has recently been able to show the existence of a family of solutions representing time machines with a compact vacuum core @xcite , in which ctcs are created in a compact region evolved from a causally well - behaved past .
these solutions satisfy all energy conditions , are regular , asymptotically flat and topologically trivial , proving that general relativity alone can not prevent the formation of ctcs . in 1992 , hawking proposed the chronology protection conjecture @xcite , which states that the laws of physics do not allow the appearance of ctcs . in particular , he argued that the strong backreaction due to quantum fluctuations should destroy the chronology horizon .
however , in 1996 , kay , radzikowski and wald showed that there are points on the chronology horizon where the semiclassical einstein s equations fail to hold @xcite , and therefore the ultimate resolution of the chronology protection issue must involve the full theory of quantum gravity ( see matt visser s review @xcite ) .
string theory is believed to yield a consistent theory of quantum gravity , therefore we shall take a more fundamental point of view , and consider the ( super-)gravity theories merely as its effective low energy limits . then , even if the effective theory can not distinguish between causally well - behaved and ill spacetimes , one can hope that string theory provides a mechanism to rule out the pathological solutions or eliminate the ctcs . to be more specific
, we will concentrate on the half bps sector of ads / cft @xcite , where we have a strong control on the physics .
using the ads / cft correspondence , we have a non - perturbative formulation of type iib superstring theory in terms of @xmath6 sym in four dimensions @xcite . in the half bps subsector
, the sym theory reduces to a gauged quantum mechanical system of @xmath7 harmonic oscillators , with the harmonic potential generated by the conformal coupling to the @xmath8 background .
half of these oscillators consist in left - moving excitations , the other half in right - moving ones .
the bps condition translates then in having just left - moving excitations on - shell .
moreover , the model is gauged , and only gauge - invariant states are physical .
one can use the available gauge invariance to diagonalize the matrix of left - movers , thus leaving only @xmath9 physical degrees of freedom . performing carefully this change of variables ,
one introduces a jacobian in the path integral , and effectively changes the statistics of the degrees of freedom which become fermionic . in conclusion ,
the half bps sector is described by a system of @xmath9 free fermions in a harmonic well .
its ground state corresponds to the @xmath9 fermions distibuted on the @xmath9 lowest energy levels of the harmonic oscillator , while the excitations correspond to excited fermions with energy higher than the fermi energy , or holes inside the fermi sea . using the ads / cft correspondence
, one can check that the former correspond to giant gravitons growing on the @xmath10 factor of the metric , and the latter to giant gravitons growing on the ads@xmath0 factor @xcite ( see figure [ fig - harmonic ] ) .
+ to obtain the ( semi-)classical configurations of the system , we trade the energy eigenstate basis for coherent states . then , each fermion occupies an area @xmath11 in the phase space of the harmonic oscillator , and the classical configurations are given by quantum hall droplets in this space , i.e. by the description of a region of area @xmath12 where the fermions are localized .
, with @xmath13 the planck length , and the radius @xmath14 of ads@xmath0 is related to them by @xmath15 . ] the ground state of the system corresponds to a disk centered in the origin , of area @xmath16 .
small deformation of the surface of the disk correspond to closed string excitations , or perturbative gravitons , propagating in the bulk .
finally , small holes in the disk correspond to stacks of @xmath10 giant gravitons and droplets of fermions high above the fermi sea to ads@xmath0 giant gravitons ( see figure [ fig - phasespace ] ) . and holes in the fermi sea to giant gravitons in @xmath10 . ]
thanks to the work by lin , lunin and maldacena @xcite , we have the complete supergravity solution dual to these classical cft states .
they define a function @xmath17 which takes the value @xmath18 inside the hall droplets , and @xmath19 outside .
the gravitational dual to this configuration is then given by the metric @xmath20 with @xmath21 the function @xmath22 and the shift vector @xmath23 are determined by fermion configurations @xmath24 , which act as boundary value , @xmath25 ^ 2}\,,\ ] ] @xmath26 ^ 2}\,.\ ] ] we have therefore a very nice realization of holography , in which the dual field theory data is mapped on the @xmath27 two - plane of the gravity solution .
it is useful to describe the fermi liquid using the variable @xmath28 , which takes the value one inside the hall droplets , and zero outside .
then , it is tempting to go beyond the classical limit , and interpret the function @xmath29 as the probability density of the fermions . from the point of view of the sym theory , as long as @xmath30 $ ] , one has a well - defined semi - classical state of the field theory yielding this probability density . due to the noncommutativity of the phase space variables , the variation of @xmath31 must be slow on the scale of @xmath32 .
a @xmath33 would correspond to a violation of pauli s exclusion principle , while negative density clearly has no meaning .
however , on the gravitional side of the correspondence , _ any _ real function @xmath29 gives a one half bps solution of the supergravity equations of motion , in general with some singularities .
as observed in @xcite , whenever the above physical conditions on @xmath29 are violated , the dual spacetime has closed timelike curves .
this suggests that _ the requirement of unitarity of the dual cft implements the chronology protection for superstring theory on asymptotically ads@xmath34spacetimes_. for simplicity , we shall illustrate this behaviour in the case of the superstar solution , which is the simplest family of half bps solutions of type iib supergravity on ads@xmath34 , given by the metric @xmath35 with @xmath36 , @xmath37 and @xmath38 .
the @xmath39 factor of the metric in these coordinates is given by @xmath40 .
this solution is obtained by oxidation to ten dimensions @xcite of the extremal limit of the black hole solution with one r - charge of the @xmath41 model @xcite .
again , there is a five - form field strength turned on , which can be found in @xcite .
for @xmath42 we recover standard ads@xmath34 .
the @xmath43 case exhibits a naked singularity at the origin @xmath44 of ads , where a condensate of giant gravitons growing in the five - sphere sits , and acts as a source for the supergravity fields @xcite .
this view has been confirmed in @xcite , where the corresponding fermion distribution in the dual cft was interpreted as a dilute gas of holes in the fermi sea .
in fact , using the coordinate transformation @xmath45 one recovers the llm metric ( [ llm ] ) with polar coordinates @xmath46 on the @xmath47 plane . with some algebra , one
can then compute the corresponding function @xmath48 , that reads @xmath49\,,\ ] ] where @xmath50 .
its @xmath27 limit yields the fermion distribution @xmath51 llm showed that the vacuum ads@xmath34 is represented by a fermi droplet of density @xmath52 and radius @xmath53 .
its total area @xmath54 consists therefore exactly of @xmath9 fermi cells of area @xmath55 . by turning on the r - charge @xmath56
, the probability density spreads to a disk of radius @xmath57 , but with lower density , in such a way that the correct number of fermions @xmath58 is recovered .
this fact supports the interpretation of @xmath29 as a density distribution of fermions .
the fermion system represents a uniform gas of holes delocalized in the fermi sea , and since such holes correspond to giant gravitons growing on the five - sphere @xcite , the superstar can be thought of as the backreaction on spacetime produced by a condensate of giant gravitons @xcite .
let us suppose now that @xmath59 .
corresponds to putting sources away from the @xmath60-plane , and is therefore not covered by the llm construction . ]
the metric ( [ superstar ] ) still represents a genuine solution of the supergravity theory .
this spacetime is defined for @xmath61 , and has a naked singularity at @xmath62 .
for @xmath63 it is perfectly well - behaved , but one checks that in - between , for @xmath64 , the coefficient @xmath65 of the metric becomes negative , and the orbits of @xmath66 are ctcs ( see figure [ fig - superstar ] ) . . ]
note that the metric remains lorentzian everywhere . as in the superstar case
, we can cast the metric in the llm form to read off the source which produces this spacetime .
the result is again ( [ superstar - density ] ) , but this time , since @xmath67 , the fermi sea has been squeezed to a smaller radius , resulting in a density @xmath33 .
this implies the presence of more than one fermion per fermi cell , in plain violation of pauli s exclusion principle .
the anti - superstar has no dual cft state .
taking the dual cft description equivalent to type iib superstring theory as the fundamental one , we see that the pauli exclusion principle automatically rules out the solutions of the form ( [ superstar ] ) with causal anomalies .
this mechanism works in fact in much more generality , and the chronology protection conjecture has been proven for general fermionic distributions with circular symmetry . using continuity arguments , it was shown in @xcite that supergravity solutions resulting from non - physical distributions @xmath29 ( i.e. which do not define a good probability density ) have indeed ctcs .
subsequently , milanesi and oloughlin proved the converse by a detailed analysis of the geometries , i.e. that solutions generated by well - behaved fermion densities are free of ctcs @xcite .
to summarize , in the half bps sector a spacetime suffers from ctcs if and only if the boundary condition @xmath68 for the supergravity solution violates somewhere the bound @xmath69 .
if we interpret @xmath29 as the probability density of fermions in the phase space , this bound corresponds to a unitarity bound of the dual field theory ; @xmath33 would violate pauli s exclusion principle , while negative @xmath29 would imply negative norm states .
therefore , configurations violating this bound do not correspond to any state in the hilbert space of the sym theory , which gives a fundamental description of superstring theory on ads@xmath34 , and the corresponding causality - violating backgrounds are forbidden .
this mechanism implements in a simple form the chronology protection conjecture for a subsector of type iib superstring theory in asymptotic ads@xmath34 backgrounds , and suggests a deep link between causality in gravitational theories and unitarity of the dual field theory .
such a link also emerged in the analysis of the bmpv black hole @xcite , for which a similar protection mechanism acts .
this black hole has ctcs , which are hidden by the event horizon in the underrotating regime , but develops naked ctcs if one increases the angular momentum and enters the overrotating regime . using the microscopic description in terms of a two - dimensional conformal field theory , provided by string theory
, herdeiro showed that the causality bound corresponds to a unitarity bound in the cft @xcite , and therefore that the overrotating spacetimes are not genuine solutions of string theory .
this analysis has then been generalized by boni and silva , who extended the protection mechanism to general ads@xmath70 bubbling solutions @xcite .
moreover , this mechanism could apply to non - supersymmetric configurations , and it was proposed in @xcite to hold for non - extremal kerr - ads@xmath0 black holes of type iib supergravity .
such solutions have naked ctcs for equal angular momenta and sufficiently negative mass . on the other hand ,
unitary representations of the superconformal algebra put a lower bound on the mass .
it would be interesting to check if the bounds agree in this non - supersymmetric case .
finally , it would be interesting to understand how this chronology protection mechanism can be realized in other sectors of ads / cft . |
classical electrodynamics is well known to be a linear theory leading to the superposition principle . at the quantum level
the basic qed lagrangian remains quadratic in the electromagnetic fields , so that the theory still appears to be linear .
however , quantum fluctuations in the qed vacuum induce nonlinear effects that lead to a breakdown of the superposition principle @xcite . in particular
, these qed fluctuations make the vacuum appear as an electrically and magnetically polarizable medium .
the size of these corrections in nonlinear qed ( nlqed ) is very tiny , so that experiments with ultra - high intensity lasers have been proposed to search for these effects , e.g. @xmath11 pair production from the vacuum @xcite-@xcite , vacuum birefringence @xcite-@xcite , light diffraction by a strong standing electromagnetic wave @xcite , and nonlinear compton scattering @xcite .
a different proposal , involving quasistatic external electromagnetic fields interacting with given electric or magnetic sources , has been made recently @xcite-@xcite . in @xcite
general expressions were obtained for the induced electric and magnetic fields in such circumstances , and applied to the case of an electrically charged sphere in the presence of an external , quasistatic magnetic field . as a result of qed nonlinearity there appears an induced magnetic dipole moment , as well as corrections to the coulomb field of the sphere . in spite of this being a dramatic effect , experimental detection appears very challenging .
the complementary case of a purely magnetic dipole moment placed in an external , quasistatic electric field @xmath12 was considered in @xcite .
the result is an induced electric dipole moment @xmath13 , plus corrections to the magnetic field produced by the magnetic dipole .
it was then suggested that the neutron could be used as a probe in the presence of large electric fields of order @xmath14 v / m , such as present in certain crystals . a distinctive feature of this induced electric dipole moment , which should help in its detection , is its peculiar dependence on the angle between @xmath13 and @xmath12 , or equivalently the angle between @xmath13 and the neutron spin .
+ in this paper we follow up on the experimental observability of such an induced electric dipole moment of the neutron . on the theoretical side
we complete the analysis of @xcite by computing the interaction hamiltonian of the neutron immersed in a large external quasistatic electric field @xmath12 , and an external , quasistatic , magnetic field @xmath15 of ordinary strength . given the nonlinearity of the problem one needs to check that ( a ) the magnetic interaction energy is of the usual form , ( nonlinear magnetic corrections due to @xmath15 are negligible ) , and ( b ) that the induced electric dipole does interact with the electric field @xmath12 that generates it .
the latter interaction energy is expected to have the standard functional form @xmath16 , albeit with an a - priori unknown coefficient which we determine .
next , we study the quantum behaviour of @xmath13 by means of the heisenberg equation of motion .
this is important for experiments based on potential changes in the larmor frequency of the neutron spin around an external magnetic field due to the presence of @xmath17 .
we find no effect here , thus ruling out experiments of this type to detect an induced electric dipole moment of the neutron .
finally , we discuss in some detail a different approach based on neutron - nucleus scattering and conclude that this experiment offers an excellent opportunity to observe such an effect . this is due to the peculiar angular dependence of @xmath13 .
we find that for sufficiently large momentum transfers , a scattering asymmetry is induced with such particular characteristics that it would be easy to distinguish from other standard effects .
the experimental discovery of such an asymmetry would be the first ever signal of a nonlinear effect in electrodynamics due to quantum fluctuations in the qed vacuum .
an appropriate framework to discuss nonlinear effects induced by quantum fluctuations in the qed vacuum is that of the euler - heisenberg lagrangian @xcite .
this is obtained from the weak field asymptotic expansion of the qed effective action at one loop order leading to@xmath18where the omitted terms are of higher order in the expansion parameter @xmath19 . in si units
@xmath20with @xmath21 the electromagnetic fine structure constant , @xmath22 and @xmath23 are the mass and charge of the electron , respectively , and @xmath24 the speed of light .
the invariants @xmath25 and @xmath26 are defined as @xmath27 @xmath28with @xmath29 and @xmath30 .
the so - called critical field @xmath31 , which plays the role of a reference field strength for the onset of nonlinearity , is given by @xmath32this estimate is obtained by computing the electric field needed to produce an electron - positron pair in a spatial length of one compton wavelength . for fields stronger than @xmath31 the weak - field asymptotic expansion leading to eq .
( [ l - eh ] ) breaks down . in @xcite
general expressions were obtained for electric and magnetic fields induced by nonlinearity , to leading order in @xmath19 , in the presence of external quasistatic weak fields ( smaller than @xmath31 ) and arbitrary sources .
these induced fields are @xmath33 @xmath34where @xmath35 and @xmath36 are the maxwell ( classical ) fields produced by the arbitrary sources . notice that these fields vanish as @xmath37 ( @xmath38 )
. in the case of a current density uniformly distributed on the surface of a sphere of radius @xmath39 , or equivalently , for a uniformly magnetized sphere of the same radius , the maxwell , magnetic dipole type field , is given by @xmath40 where @xmath41 is identified with the magnetic dipole moment of the source , and @xmath42 is a unit vector in the radial direction . since the central expressions , eqs.([e - ind ] ) and ( [ b - ind ] ) , were derived assuming @xmath43 the following constraint follows @xmath44 for instance , if @xmath45 , as for the neutron , then it follows that @xmath46 .
if this magnetic source is placed in an external , constant electric field @xmath47 it has been shown @xcite that there is an induced electric field of the dipole type @xmath48 + \mathcal{o}\left ( \left\vert \mathbf{x}\right\vert ^{-6}\right ) .
\label{e - induced}\ ] ] where @xmath49 is the angle between the external electric field , lying in the @xmath50-@xmath51 plane , and the magnetic dipole moment pointing along the @xmath51 axis , i.e. @xmath52 , and @xmath53 .
the induced electric dipole moment @xmath54 is given by @xmath55 \ , .
\label{p - ind}\]]this induced electric field is of the electric dipole type in its radial @xmath56 dependence , but it has a manifestly peculiar angular dependence .
for instance , along the @xmath51 axis , and unlike a standard electric dipole field , it has a non - zero component along @xmath57 that depends on the azimuthal angle @xmath58 .
it also has a non - zero component along the direction of @xmath59 , as may be appreciated by writing the induced electric field in spherical coordinates ( @xmath60 , @xmath61 , @xmath58 ) , i.e. @xmath62 \,{\mathbf{e}_{r}}\right .
\nonumber \\ & & + \left . \left
[ 13\cos \theta \cos \phi \sin \psi + 36\sin \theta \cos \psi \right ] { \mathbf{e}_{\theta } } -13\,\sin \phi \sin \psi \;{\mathbf{e}_{\phi } } \frac { } { } \right\ } \;.\end{aligned}\]]in addition to the induced electric field eq.([e - induced ] ) , there is an induced magnetic field ( a correction to the field produced by the magnetic dipole source ) , which can be derived from a vector potential , i.e. @xmath63 , where after a lengthy calculation one finds @xmath64 \left ( \mathbf{e}_{r}\times \mathbf{e}_{0}\right ) \right .
\label{a - calli } \\ & & \left .
+ 7\left ( \mathbf{e}_{0}\cdot \mathbf{e}_{r}\right ) \left ( \mathbf{m}\times \mathbf{e}_{0}\right ) \frac { } { } \right\ } \left [ 1+\mathcal{o}\left ( \frac{\mu _ { 0}\left\vert \mathbf{m}\right\vert ^{2}}{a^{6}\epsilon _ { 0}\left\vert \mathbf{e}_{0}\right\vert ^{2}}\right ) \right ] + \mathcal{o}\left ( \left\vert \mathbf{x}\right\vert ^{-4}\right ) .
\nonumber\end{aligned}\]]notice that while @xmath65 grows linearly with @xmath66 , @xmath67 depends quadratically on @xmath66 .
+ we proceed to discuss the interaction energy of the magnetic dipole source and its induced electric dipole with the external constant field @xmath47 , and with an external uniform magnetic field @xmath68 weak enough not to induce nonlinear effects , i.e. @xmath69 . given the nonlinearity of the problem it is important to verify that the magnetic interaction hamiltonian has the expected form @xmath70 , given the strength of @xmath68 .
in addition , the electric interaction energy of the induced electric dipole and the external field @xmath47 is a - priori unknown .
this need not be exactly of the form @xmath71as one would obtain for a linearly polarizable particle immersed in an external electric field , e.g. for a polarizable neutron on account of its quark substructure .
in fact , the electric interaction hamiltonian due to nonlinearity lacks the factor @xmath72 as shown next .
the canonical energy - momentum tensor is defined as @xmath73where the total lagrangian density is @xmath74 , with @xmath75 and @xmath76 given in eq.([l - eh ] ) . this equation can be rewritten as @xmath77 .\end{aligned}\]]since the last term on the right hand side above is the total divergence of an anti - symmetric tensor , employing the equations of motion @xmath78one can define another energy - momentum tensor as @xmath79 \nonumber \\ & = & \left ( \frac{\partial \mathcal{l}}{\partial \mathcal{f}}f^{\mu \alpha } + \frac{\partial \mathcal{l}}{\partial \mathcal{g}}\widetilde{f}^{\mu \alpha } \right ) f_{\alpha \nu } -\mathcal{l}\ , \delta _ { { \mbox } { } \nu } ^{\mu } + \left ( j_{\alpha } a^{\alpha } \right ) \delta _ { { \mbox } { } \nu } ^{\mu } -j^{\mu } a_{\nu } .\end{aligned}\]]notice that this tensor is symmetric and gauge invariant except for the last two terms .
the total energy density of the system is defined as the component @xmath80 , @xmath81 in general , for a given configuration of the fields the interaction hamiltonian is defined as the difference of the total hamiltonian with and without the sources . in a quantum theory it is defined as the difference of the total hamiltonian evaluated at the fields in the interaction picture , with and without the external sources . then , the interaction hamiltonian @xmath82 , i.e. the volume integral of the interaction hamiltonian density @xmath83 is @xmath84where @xmath85 , with @xmath86 given in eq.([a - calli ] ) , and @xmath87 is the vector potential associated with @xmath68 , i.e. @xmath88 , and @xmath89 .
the current @xmath90 corresponding to the magnetized sphere producing the field , eq.([b - d ] ) , is @xmath91 . in eq.([h - int ] ) the self energy of the magnetized sphere , independent of the external field , has been omitted . after performing the integration in eq.([h - int ] )
one finds @xmath92which has the correct magnetic interaction term as in the linear theory .
the electric interaction energy is of the expected form , but it involves a coefficient different from the case of linear qed as a result of nonlinearity . in the absence of the external magnetic field @xmath68 , and using eq.([p - ind ] ) , the interaction hamiltonian becomes @xmath93notice the dependence of @xmath82 on @xmath94 .
it should be pointed out that in an experimental situation one would typically be interested in a point magnetic dipole .
this source would produce very strong fields in its proximity so that the limit @xmath95 would obviously not be allowed .
instead , we assume that even in such a case the large distance solution for the fields is well described by the first order approximation to the effective lagrangian @xmath76 in eq.([l - eh ] ) .
we also assume that this solution is robust against short distance modifications of the source as long as its symmetry is preserved . in this sense
the parameter @xmath39 is to be considered as a measure of our ignorance about the higher order corrections to this effective lagrangian , something necessary when dealing with strong fields .
the specific value of @xmath39 will be discussed later in section 4 .
the fact that @xmath82 depends on the orientation of @xmath41 with respect to @xmath47 through the angle @xmath96 can be used as a distinctive feature in the design of an experimental asymmetry as described below in section 5 .
+ we consider next the quantum behaviour of @xmath97 using the heisenberg equation of motion . to this end
we consider a particle with magnetic dipole moment @xmath98 related to the spin through the standard relation @xmath99 , where @xmath100 is the gyromagnetic ratio . assuming that the dynamics of this particle is described by the hamiltonian eq.([pot - energy ] ) , and given that @xmath101 , @xmath82 to first order in @xmath19 contains only quadratic terms in the spin , whose components are the dynamical variables of the problem .
the effective hamiltonian involving these dynamical variables must be symmetrized in order to ensure hermiticity . hence
, the quadratic terms in the spin entering the heisenberg equation of motion lead to the commutator @xmath102 = i\epsilon _ { jkl}\left\ { s_{i},s_{l}\right\ } + i\epsilon _ { ikl}\left\ { s_{j},s_{l}\right\ } .\]]for a spin @xmath72 particle , such as the neutron , we have @xmath103 , and @xmath104 . in this case ,
@xmath102 = i\epsilon _ { jkl}\frac{1}{2}\delta _ { il}+i\epsilon _ { ikl}\frac{1}{2}\delta _ { jl}=\frac{i}{2}\left ( \epsilon _ { jki}+\epsilon _ { ikj}\right ) = 0.\]]therefore , @xmath105 so that if one is interested in the time evolution of a spin @xmath72 particle , and eq.([pot - energy ] ) describes its effective hamiltonian , we find no contribution from this leading order nonlinear correction .
in other words , the precession of the spin is not affected .
this is not the case , though , for spin - one particles .
this unfortunate feature rules out experiments to detect the induced electric dipole moment of the neutron based on larmor frequency changes . a different approach involving neutron scattering off nuclei is discussed next .
scattering of slow neutrons by a free atom can be described by a scattering amplitude in the born - approximation , which in the center of mass system is given by ( see e.g. @xcite ) @xmath106where @xmath107 is the reduced mass @xmath108with @xmath109 the neutron mass , and @xmath110 the mass of the atom . the three - momentum transfer @xmath111 is @xmath112 , with @xmath113 and @xmath114 the neutron wave vectors before and after scattering , respectively , and @xmath115 * * * * is the neutron spin in units of @xmath116 .
the magnitude of @xmath117 will be denoted as @xmath118 in the sequel .
the total hamiltonian @xmath82 involves all known interactions between the neutron and the atom , to which we add now the new interaction due to nlqed given in eq.([hamiltonian ] ) .
correspondingly , the total scattering amplitude can be written as @xmath119where the various contributions are as follows .
the term @xmath120 is due to the hadronic interaction of the neutron with the nucleus , and is usually the dominant term .
the amplitude @xmath121 corresponds to the interaction of the neutron magnetic moment with the atomic magnetic field ( for atoms with unpaired electrons ) .
this term can be of a similar size as @xmath120 .
the next three terms arise from various electromagnetic interactions , i.e. @xmath122 is due to scattering of the neutron charge radius by the electric charges in the atom , @xmath123 arises from the electric polarizability of the neutron due to its quark substructure , and @xmath124 corresponds to the spin - orbit interaction of the neutron in the electric field of the nucleus .
the term @xmath125 is a weak interaction , parity - violating amplitude which we list separately from @xmath120 as it has a different dependence on neutron spin .
finally , @xmath126 is the new component due to the induced electric dipole moment of the neutron , which we wish to isolate experimentally .
+ the scattering amplitude , eq.([scattering amp ] ) , enters the differential cross section for elastic neutron scattering by a single atom in the ground state , @xmath127which includes an ensemble average over nuclear and electronic spin degrees of freedom ( if present ) , and the neutron spin .
the incident neutrons are characterized by a polarization defined as @xmath128 . in the absence of nuclear and electronic polarization of the atom ,
the case of interest here , the kinematic scattering variables are @xmath111 and @xmath129 .
experimentally , one determines neutron scattering cross sections using a sample containing a macroscopic number of atoms . considering a single atomic species , the ensemble average in eq.(sigma )
still has to account for the isotopic composition and the different states of total spin of a neutron scattering off a nucleus with non - zero spin . for slow neutrons with wavelengths much larger than the nuclear radius @xmath130
the hadronic amplitude @xmath131 is practically independent of @xmath111 .
this is in the absence of nuclear resonances for thermal and epithermal neutrons , i.e. for the energy range of interest here .
scattering thus proceeds in an s - wave and is isotropic in the center of mass system .
one defines a neutron scattering length operator as @xmath132for a nucleus with spin @xmath133 one has @xmath134where @xmath135 and @xmath136 are the eigenvalues of @xmath137 for the two states of total spin @xmath138 ( see e.g. @xcite ) . for a sample with all nuclear species unpolarized , scattering by the @xmath139th isotope enters with statistical weight factors @xmath140 and @xmath141 . the leading term in the cross section is then given by @xmath142 , \]]where the bar indicates the averaging over isotopes and spin states , and @xmath143 stands for the relative abundance of the @xmath139th isotope . in next - to - leading order the cross section contains interference terms between small amplitudes like @xmath126 and a usually dominant coherent nuclear scattering length @xmath144 , which for unpolarized nuclei is given by @xmath145 .\]]most scattering lengths @xmath144 are found to be positive with typical values of a few @xmath146 .
neutron optical measurements determine a coherent _ bound scattering length _
@xmath147 , related to the corresponding scattering length @xmath148 of a free atom through @xmath149this relation includes the contributions @xmath150 due to all amplitudes ( @xmath151 @xmath152 ) appearing in eq.([scattering amp ] ) .
lacking sufficiently accurate theoretical predictions for the nuclear part , however , one can not even extract from @xmath147 the sum of all non - hadronic components , which normally contribute less than @xmath153 .
instead , one needs to perform measurements for @xmath154 .
values for @xmath147 and the total neutron scattering cross section of an atom fixed in space , @xmath155 , can be found e.g. in @xcite . for later use
we quote the values for lead with natural isotopic abundances @xmath156 for low neutron energies one also has to take into account interference effects of the neutron waves scattered from different atoms , as e.g. in condensed - matter studies .
classical examples are bragg scattering by single crystals and measurements of phonon dispersion relations . however , for sufficiently large momentum transfer @xmath157 as considered here , interatomic interferences and eventual nuclear spin correlations between different atoms can be neglected .
we thus consider the cross section in the center of mass system as given by @xmath158 \right\rangle .
\label{sigma - diff}\]]where the dots stand for the remaining contributions of squared amplitudes from eq.([scattering amp ] ) , and the sum is over @xmath159 , mag , pol , so , pv and ind
. the formulas to transform this expression to the laboratory reference frame can be found in ref .
as long as the nucleus is free to recoil , the total cross section does not change and changes in the angular distribution of the scattered neutrons appear only at order @xmath160
. since we are not interested in angular distributions and will only consider atoms much heavier than the neutron , we may use the scattering cross section as given above in eq.([sigma - diff ] ) .
we now turn to the calculation of the new amplitude @xmath126 due to the nlqed induced electric dipole moment @xmath0 given in eq.([p - ind ] ) .
the magnetic moment of the neutron , @xmath41 , can be written as @xmath161where @xmath162 and @xmath163 are the pauli matrices . from eq .
( [ a - constraint ] ) there follows the lower bound @xmath164according to eq.([pot - energy ] ) , @xmath0 interacts with the atomic electrostatic field , which for simplicity we consider as given by a pointlike nucleus with electric charge @xmath165 , @xmath166as discussed later , one can neglect electric field shielding due to the atomic electrons . using eqs.([field nucleus ] ) , ( [ hamiltonian ] ) and ( [ f ] ) one obtains @xmath167 , \label{f - qed}\]]where @xmath168 is the angle between @xmath115 and @xmath111 .
the two integrals @xmath169 and @xmath170 can be easily calculated analytically in polar coordinates with @xmath111 along the polar axis .
they are given by @xmath171and @xmath172the radial integration extends over all space , excluding a sphere of radius @xmath173 around the nucleus . for a heavy nucleus like lead , electric fields as strong as @xmath174 exist close to the nuclear surface
this exceeds by far the critical field , eq.([e - crit ] ) , beyond which higher order terms in the one - loop effective lagrangian in eq.([l - eh ] ) become important @xcite and thus can not be neglected .
therefore , @xmath173 has to be much larger than the nuclear radius @xmath175 , and we choose it here as the distance from the nucleus to where the critical field is reached , i.e. @xmath176for lead isotopes with @xmath177 one has @xmath178 .
the integrals eqs.(41 ) and ( 42 ) can be solved analytically with the result @xmath179 qr\right\ } \bigskip \mathstrut \medskip \label{i-1-sol } \\[0.4 cm ] & \simeq & \frac{4\pi } { r}\left [ 1-\frac{\pi } { 4}qr+\frac{1}{6}\left ( qr\right ) ^{2}- ... \right ] , \nonumber\end{aligned}\]]where @xmath180 is the sine integral , and @xmath181 \cos \left ( qr\right ) \right .
\nonumber \\ & & -\frac{1}{\left ( qr\right ) ^{3}}\left [ 2 - 3\left ( qr\right ) ^{2}+\left ( 6-\left ( qr\right ) ^{2}\right ) \cos \left ( 2\beta \right ) \right ] \sin \left ( qr\right ) \nonumber \\ & & \left .
-qr\left ( 3+\cos \left ( 2\beta \right ) \right ) \left ( \frac{\pi } { 2}-{\mbox}{si}\left ( qr\right ) \right ) \right\ } \medskip \label{i-2-sol } \\[0.4 cm ] & \simeq & \frac{4\pi } { 3r}\left [ 1-\frac{3\pi } { 32}\left ( 3+\cos \left ( 2\beta \right ) \right ) qr+\frac{1}{10}\left ( 2+\cos \left ( 2\beta \right ) \right ) \left ( qr\right ) ^{2}- ... \right ] .
\nonumber\end{aligned}\]]using eqs.([i-1-sol ] ) and ( [ i-2-sol ] ) in eq.([f - qed ] ) one finds the final expression for the scattering amplitude due to the nlqed induced electric dipole moment@xmath182 \left [ qr\cos \left ( qr\right ) -\sin \left ( qr\right ) \right]}{2\left ( qr\right ) ^{3}}\right . \label{f - final } \\ & & \left . + \frac{1}{4}\left [ 43 + 49\cos \left ( 2\beta \right ) \right ] \left [ \cos \left ( qr\right ) + \frac{\sin \left ( qr\right ) } { qr}+\left ( { \mbox}{si}\left ( qr\right ) -\frac{\pi } { 2}\right ) qr\right ] \right\ } \medskip \nonumber \\[0.4 cm ] & \simeq & \frac{\zeta m\mu _ { 0}\mu _ { \mathrm{n}}^{2}z^{2}e^{2}}{24\pi ^{3}\hbar ^{2}\epsilon _ { 0}^{3}a^{3}r}\left\ { 1-\frac{3\pi } { 320}\left [ 43 + 49\cos \left ( 2\beta \right ) \right ] qr\right .
\nonumber \\ & & \left .
+ \frac{1}{100}\left [ 33 + 49\cos \left ( 2\beta \right ) \right ] \left ( qr\right ) ^{2}- ... \right\ } . \nonumber\end{aligned}\]]the scattering amplitude @xmath126 exhibits a welcome peculiar dependence on the angle @xmath183 between the neutron spin @xmath115 and the three - momentum transfer @xmath111 .
this dependence introduces an asymmetry which for suitable experimental conditions is essentially free from _ background _
contributions due to other well known effects .
the largest effect is obtained by evaluating @xmath126 at @xmath184 and at @xmath185 .
this feature will then play a crucial role in the experimental detection of the nlqed induced electric dipole moment of the neutron , as discussed in the following section .
we define the scattering asymmetry as @xmath186where the differential cross section , eq.([sigma - diff ] ) , is evaluated for two neutron polarization states @xmath187 and @xmath188 , parallel and perpendicular to the scattering vector @xmath111 , respectively . the interference term between the coherent nuclear amplitude and the amplitude of interest leads to @xmath189where @xmath190 and @xmath191 , @xmath192 , and @xmath193 is the total scattering cross section of the free atom .
we argue below that in a well - designed experiment possible influences of interference terms other than between @xmath144 and @xmath126 are negligible .
hence , @xmath194 , so that the asymmetry defined in eq.(def - a ) should allow for a detection and determination of the new amplitude . using the values for natural lead from eq.([values - pb ] ) in eq.([a - new ] ) one obtains @xmath195from eqs.([f - final ] ) and ( [ a - lead ] ) it follows that @xmath196 , where @xmath197 is a function of the dimensionless parameter @xmath198 . the maximum of @xmath199 occurs for ( see fig .
1 ) @xmath200using the value of @xmath39 given in eq.([a ] ) the asymmetry becomes @xmath201a result which appears experimentally accessible .
neglecting nuclear recoil , valid in good approximation for neutron scattering off a heavy target , one may use the relation @xmath202where @xmath203 is the angle between @xmath113 and @xmath204 , and @xmath205 , with @xmath206 the neutron kinetic energy in @xmath207 . in a backscattering geometry , i.e. for @xmath208 , the maximum asymmetry , eq.([a - epi ] ) , is obtained with epithermal neutrons of energy @xmath209 .
, eq.(a - lead ) , normalized to the neutron polarization @xmath210 as a function of the dimensionless variable @xmath198.[figure1 ] ] the result for the asymmetry depends explicitly on the radii @xmath173 and @xmath39 of spheres centered on the nucleus and on the neutron , respectively .
these parameters play the role of separating the short - distance physics close to the electromagnetic sources from the long - distance effects involving fields weak enough for the euler - heisenberg approximation to be valid . in
this regard it is important to point out that electric and magnetic fields induced by quantum fluctuations in the qed vacuum involve various multipolarities @xcite .
for instance , in the case of a neutron in an external , quasistatic electric field @xmath211 , the induced electric field eq.([e - induced ] ) has a dipole type term of order @xmath212 , as well as a higher multipole of order @xmath213 , while the induced magnetic field involves terms of order @xmath212 , @xmath214 , and @xmath215 , with @xmath216 .
these higher order multipoles can be safely neglected in the interaction energy as long as the weak - field approximation remains valid .
however , it is not clear what happens at distances closer to the nucleus or to the neutron .
this problem is similar to that of the separation into far and near field regions around a localized charge / current distribution in classical electrodynamics , where the lowest - order multipole provides the long - distance solution .
although one can not compute the interaction energy due to qed vacuum effects stemming from the regions @xmath217 and @xmath218 , their contribution to the asymmetry would presumably have a different dependence on @xmath115 and @xmath111 . for large @xmath219 it might smear out the oscillations appearing in fig . [ figure1 ] . for small @xmath220
, corresponding to small spatial resolution in probing the qed vacuum , the asymmetry should not be affected much by the short - distance physics .
this statement is underlined by the fact that at leading order in @xmath198 the @xmath168-dependent term in @xmath126 , eq.([f - final ] ) , does not depend on @xmath173 . for @xmath221 , one thus obtain a prediction for the asymmetry which should be robust against variations in the choice of @xmath173 , i.e. @xmath222for a neutron energy of @xmath223 in a backscattering geometry one expects @xmath224 from a practical point of view the polarization of epithermal neutrons with more than @xmath225 requires a spin filter of polarized protons .
this is technically demanding if one wishes to polarize a beam with a diameter of several cm .
the measured energy - dependent neutron polarization cross section of a polarized - proton spin filter for the energy range of interest may be found in ref . @xcite . on the other hand ,
neutrons with energies up to @xmath226 are available from a hot moderator in a reactor neutron source with much higher intensity than epithermal neutrons .
they can be polarized using magnetic monochromator crystals , or by a spin filter of polarized @xmath227he gas @xcite which has a polarization cross section proportional to @xmath228 .
for the fluxes available at the institut laue - langevin ( ill ) in grenoble an asymmetry as in eq.([a - hot ] ) appears experimentally accessible within a few days of beam time .
+ to conclude this section we stress that the neutron scattering asymmetry due to nonlinear qed has two characteristic properties which should make it rather easy to detect and distinguish from other effects .
first , the asymmetry attains its maximum value for perpendicular orientations of the neutron polarization and vanishes for opposite orientations .
this is in contrast to most ordinary asymmetries which become maximal for opposite orientations .
second , @xmath229 exhibits a characteristic @xmath219 dependence with a broad maximum around the value of @xmath219 given in eq.([qr ] ) .
these features are discussed in more detail in the sequel .
in this section we study the contributions to the asymmetry @xmath230 , eq.([def - a ] ) , from the various ordinary scattering amplitudes defined in eq.([scattering amp ] ) .
the neutron spin - dependent amplitudes can be written as @xmath231 \label{f - spin - dep}\]]where @xmath232 is spin independent , and @xmath233 is a vector not correlated with the neutron spin . in the case of the weak amplitude @xmath234 the vector @xmath111
must be replaced by @xmath113 .
for instance , for the nuclear amplitude in eq.([a - spin ] ) , @xmath233 is independent of @xmath111 and given by the nuclear spin @xmath235 .
it can be shown in general that the terms proportional to @xmath236 in the differential cross section , eq.([sigma - diff ] ) , are all independent of the neutron polarization and therefore can not generate an asymmetry . in principle
these terms influence the size of @xmath230 through the total scattering cross section @xmath237 , eq.([a - new ] ) .
for scattering angles @xmath238 the pure spin - orbit cross section , quadratic in the amplitude @xmath239 , might become large enough to have an impact on @xmath237 .
however , for sufficiently large @xmath203 , and in the absence of nuclear and electronic polarization , corrections to @xmath237 due to squared - amplitude terms can be safely neglected .
the interference terms between the nuclear and the other scattering amplitudes in eq.([sigma - diff ] ) may however affect the asymmetry @xmath199 through their dependence on the neutron polarization .
this requires careful consideration , and we start with the amplitude @xmath239 for spin - orbit scattering .
it originates in the interaction of the neutron magnetic moment with the magnetic field present in the neutron rest frame due to its motion through the atomic electric fields .
its expression is ( see e.g. @xcite ) @xmath240 \left ( \mathbf{s}\cdot \mathbf{n}\right ) , \label{f - so}\]]where @xmath241 $ ] is the fourier transform of the electric charge density of the atom .
this term involves the nuclear charge @xmath242 and the atomic form factor @xmath243 normalized to @xmath244 .
this form factor is measured e.g. in x - ray scattering off atoms , and is a real function of the momentum .
the unit vector @xmath245 points along @xmath246 , so that the amplitude can contribute only if the neutron polarization has a component out of the scattering plane .
the asymmetry due to the spin - orbit interaction is given by @xmath247 \left ( \mathbf{p}_{\vert } \cdot \mathbf{n}-\mathbf{p}_{\perp } \cdot \mathbf{n}\right ) .
\label{a - so}\]]the imaginary part of @xmath144 above is due to nuclear absorption and can be calculated using the optical theorem .
lead nuclei absorb neutrons only weakly so that @xmath248 for neutron energies of interest here . for a scattering angle
@xmath249 a neutron kinetic energy of @xmath250 would be required to observe the maximum asymmetry according to eq.([a - epi ] ) . in this case
the atomic form factor @xmath251 , and a maximum asymmetry @xmath252 should be observed for @xmath253 perpendicular to the scattering plane , i.e. @xmath254 ( while @xmath255 by definition ) . as a result , for @xmath256 one
has @xmath257 , while for @xmath258 together with the conservative value @xmath259 , the asymmetry becomes @xmath260 .
these values are already smaller than the corresponding values of @xmath261 , but with suitable experimental settings they can be reduced even further .
notice that the condition @xmath262 required for @xmath263 can be realized for different orientations of @xmath253 , while also fulfilling @xmath264 .
hence , choosing a backscattering geometry , for which @xmath265 together with @xmath266 , and using realistic assumptions about the experimental definition of the directions of @xmath267 and @xmath129 , one can easily suppress @xmath252 by a factor @xmath268 .
hence , the impact of @xmath252 can be kept well under the @xmath269 level .
+ next , the amplitude @xmath270 of the neutron - nuclear interaction in eq.([a - spin ] ) , when squared , gives rise to interference between its spin - dependent and spin - independent parts .
after ensemble averaging this becomes proportional to @xmath210 and to the nuclear polarization @xmath271 . in thermal equilibrium ,
@xmath272 for nuclei with magnetic moment @xmath273 in a magnetic field @xmath274 ( @xmath275 is the boltzmann constant ) .
if the target is at room temperature , and given that no magnetic field is needed at the position of the sample , the @xmath210-dependent cross section is orders of magnitude too small to have an impact on the asymmetry .
+ we consider next the amplitude @xmath276 , which is due to the interaction of the neutron magnetic moment with the magnetic field produced by unpaired atomic electrons of paramagnetic contaminants in the sample .
the operator structure of this amplitude is given by @xmath277 , \label{f - mag}\]]where @xmath278 is the fourier transform of the total ( spin and orbital ) magnetization of the atom , and @xmath279 .
the interference term with @xmath144 thus involves the neutron polarization and the sample - averaged magnetization .
it would only influence the asymmetry if ( 1 ) the ratio @xmath280 is sufficiently high to result in a sizable magnetization , ( 2 ) paramagnetic centers are sufficiently abundant , ( 3 ) the two neutron polarization states in the asymmetry have different projections perpendicular to the scattering plane ( which is a consequence of the term in brackets in eq.([f - mag ] ) ) , and ( 4 ) measurements are performed for sufficiently small @xmath219 , where the magnetic form factor still has a sizable value . regarding the latter , even for the smaller value of @xmath219 envisaged in eq.([a - hot ] ) , the magnetic form factor leads to a strong suppression of @xmath276 .
hence , with the conditions 1 , 2 and 3 under experimental control one can safely disregard magnetism as a source of an asymmetry .
+ finally , the parity - violating amplitude @xmath234 due to the hadronic weak interaction may lead to a different type of asymmetry which has indeed been observed in neutron transmission experiments .
effects depend on the neutron helicity , hence @xmath281 in eq.(f - spin - dep ) , with a complex coefficient to describe both parity violating spin rotation and transmission asymmetry .
the amplitude is normally so small that it requires special efforts to detect it .
for thermal neutrons , transmission asymmetries for longitudinally polarized neutrons forte/1980 have typical sizes of a few times @xmath282 . however , for neutron energies in the vicinity of p - wave resonances of complex nuclei a strong enhancement due to the weak nuclear interactions may appear .
a prominent example is the transmission asymmetry of @xmath283 found at the p - wave resonance of @xmath284 in @xmath285la @xcite .
however , no effect sufficiently strong to affect the asymmetry @xmath261 is known for lead in the relevant energy range . in addition , an experimental test can easily be performed .
in fact , taking the neutron polarization @xmath129 parallel and anti - parallel to @xmath111 , i.e. @xmath286 and @xmath287 , it follows from eqs.([f - final ] ) and ( [ a - new ] ) that @xmath288 for these two polarization orientations .
in contrast , for @xmath234 one has @xmath289 , which could be measured separately and corrected for if the need arises .
the electric amplitudes @xmath290 and @xmath291 do not generate any known scattering asymmetry .
however , owing to their characteristic @xmath219-dependences a comparison with the nlqed amplitude @xmath126 is needed .
like @xmath126 , the amplitude @xmath290 due to the electric polarizability of the neutron , @xmath292 , is induced by the nuclear electric field , eq.([field nucleus ] ) . in si units
@xmath292 is defined by @xmath293 , so that its dimension is @xmath294 = \mathrm{m}^{3}$ ] .
the calculation of @xmath290 follows from eq.([f ] ) with the interaction energy given by eq.([h - el - pol ] ) .
this leads to the integral eq.([i-1 ] ) with the result given in eq.(i-1-sol ) .
however , the lower limit of the radial integration is now different from that for @xmath126 .
in fact , this lower limit can now be extended down to the nuclear radius @xmath175 , since for @xmath295 the neutron probes only the long - range electric forces . for @xmath296
the electric interaction is small in comparison with the nuclear force , so that in early calculations @xcite-@xcite it has simply been included in the nuclear amplitude . in si units and for the electric field given in eq.([field nucleus ] ) , the dependence of @xmath290 on @xmath219 is given by @xmath297the term linear in @xmath219 is characteristic of the @xmath298 dependence of the hamiltonian , and it also enters the interference term in the cross section .
this feature has been exploited in the past to measure @xmath292 .
conflicting results from experiments performed during more than three decades show that a proper assessment of all systematic errors has been difficult ( see e.g. the table in ref .
the most recent result @xcite , derived from energy - dependent neutron transmission through a @xmath299pb target , and reporting the smallest uncertainty is @xmath300calculations using quark bag models @xcite agree with this result .
an early estimate of breit and rustgi @xcite using data on pion photoproduction already indicated that @xmath301 .
these authors also analyzed other effects which might mimic a signal from the neutron electric polarizability . from an estimate of vacuum polarization effects close to the nucleus , and using the uehling potential @xcite
, they concluded that this contribution to neutron scattering can be safely neglected .
turning now to the question of shielding of the nuclear charge by the atomic electrons , one notices that this might quench the amplitude @xmath302 .
since @xmath290 and @xmath126 both depend quadratically on the electric field @xmath47 , one can draw parallels with the analysis of @xmath290 .
we recall that in our calculation of @xmath126 one needs to exclude a spherical region of radius @xmath173 around the nucleus inside which the weak - field expansion of the euler - heisenberg lagrangian breaks down .
this procedure was followed ( for different reasons ) in the early calculations of @xmath290 @xcite-@xcite , @xcite where the nuclear region with radius @xmath175 was excluded . a more recent analysis which does not rely on a simple model for the nuclear charge distribution
gives @xcite @xmath303where @xmath304 is the root mean square charge radius of the nucleus .
this shows that slow neutrons are indeed insensitive to details at this length scale .
notice that this result is nearly identical to the leading term in eq.([f - qcd ] ) after replacing @xmath175 by @xmath304 . in the derivation of eq.([f - pol - q - to-0 ] ) the following intermediate result was obtained in @xcite @xmath305 , \label{f with f}\ ] ] where @xmath306 and @xmath307 are the charge form factors of the nucleus and of the electron distribution in the atom , respectively . with @xmath308 , shielding of the nuclear charge can be neglected in @xmath290 . even in the limit @xmath309
the neutron feels the full unscreened nuclear charge as far as the electric polarizability is concerned . for the nlqed
induced electric dipole moment it follows that , with @xmath178 for lead , the corresponding correction term of order @xmath310 is much larger than the one of order @xmath311 .
however , @xmath312 is still small enough so that one can neglect shielding of the electric field in the region around the nucleus .
we now compare the two amplitudes @xmath290 and @xmath126 in the limit @xmath309 , i.e. their respective contributions to the neutron scattering length .
setting @xmath313 and using eq.([alpha - exp ] ) one can estimate the leading order term in eq.([f - qcd ] ) as @xmath314from eq.([f - final ] ) the corresponding leading term in @xmath126 is @xmath315 where @xmath316 \right ) ^{3}}\,.\ ] ] with the value of @xmath39 from eq.([a - constraint ] ) one obtains @xmath317 which is larger than @xmath292 , eq.([alpha - exp ] ) .
however , the impact of this polarizability on the amplitude @xmath126 for @xmath318 is suppressed with respect to @xmath290 due to @xmath319 . in fact
, for lead one obtains @xmath320 this result , though , is not even an order of magnitude smaller than @xmath321 , eq.([f - pol - q0 ] ) .
it would thus contribute about @xmath322 to the neutron optical potential of solid lead , which is quite substantial given the high precision of some neutron optical methods .
the contribution of the nlqed electric dipole moment to the total cross section is @xmath323 , where @xmath324 in eq.([sigma - diff ] ) has been used .
numerically this becomes @xmath325 for unpolarized neutrons , eq.([f_ind - av ] ) , as a function of the dimensionless variable @xmath198.[figure2 ] ] we discuss next the amplitude @xmath291 which describes the interaction of the electric charges of the atom with the internal charge distribution of the neutron as characterized by its mean squared charge radius .
the amplitude for a bound nucleus is given by @xmath326 $ ] , with the atomic form factor @xmath243 given in eqs.([f - so ] ) and ( [ f with f ] ) , and the neutron - electron scattering length is @xmath327 as determined from measurements of the total cross sections of lead and bismuth at different neutron energies @xcite-@xcite . for lead and sufficiently large @xmath219 , @xmath328 , which leads to a contribution to the total scattering cross section sears/1986 @xmath329 .
it is also interesting to notice that since @xmath330 this amplitude does not contribute to the neutron scattering length .
the atomic form factor changes significantly at small @xmath219 where interference of neutron waves from different atoms can not be neglected .
hence , the macroscopic state of the sample enters crucially in the analysis of scattering data .
in contrast , in the case of @xmath126 where larger effects show up at much higher values of @xmath219 , interatomic interferences do not play any significant role .
+ to conclude this section we discuss the contribution of @xmath126 as a potential background in measurements of the amplitudes @xmath331 and @xmath291 performed with unpolarized neutrons . for the polarization averaged @xmath126 one obtains @xmath332 qr\right\ } \label{f_ind - av}\end{aligned}\]]to analyze the impact on @xmath291 one may approximate relativistic hartree - fock results for @xmath243 by the simple function @xmath333 ^{-1/2}$ ] @xcite . with @xmath334 the momentum
@xmath335 provides a scale at which significant changes in @xmath291 take place . for lead , @xmath336 , and @xmath337 in contrast , the change of @xmath338 is much smaller due to its milder @xmath219-dependence and its smaller magnitude ( see fig . [ figure2 ] ) , @xmath339since the precision of the best measurements of @xmath340 is at the level of a few percent , potential background due to @xmath338 is negligible .
a similar argument leads to the same conclusion for the determination of @xmath292 from @xmath290 .
many , if not most proposals to detect nonlinear effects due to quantum fluctuations in the qed vacuum rely on experiments involving lasers of ultra - high intensities @xcite-@xcite .
these intensities , though , are at least two orders of magnitude below current values .
an alternative approach has been discussed in this paper , based on the theoretical prediction of an induced electric dipole moment of the neutron , @xmath13 , in an external quasistatic electric field @xcite .
the peculiar features of this dipole moment , particularly its dependence on the angle between @xmath13 and the neutron spin , suggests the definition of an asymmetry which could be detected in the scattering of polarized neutrons from heavy nuclei .
we have introduced this asymmetry and discussed all possible sources of background asymmetries .
we have also compared the new nlqed amplitude with ordinary electric scattering amplitudes , particularly the one due to the polarization of the neutron in an electric field due to its quark substructure .
the conclusion from this detailed analysis is that the asymmetry due to nlqed should be observable using epithermal neutrons , and even using thermalized neutrons from a hot moderator
. this would be the first ever experimental confirmation of nonlinearity in electrodynamics due to qed vacuum fluctuations .
the numerical predictions for the asymmetry made in this paper were calculated using definite values for the parameters @xmath173 and @xmath39 .
these were derived from the condition that the electric and magnetic fields should be below their critical values , beyond which the weak - field expansion of the effective lagrangian breaks down .
while the value of the asymmetry @xmath341 for small @xmath219 does not depend on @xmath173 , it does depend on @xmath39 as seen from eq.([hamiltonian ] ) hence , the numerical results given here should be correct up to a numerical factor of order one .
this work was supported in part by fondecyt 1095217 ( chile ) , proyecto anillos act119 ( chile ) , by conicet ( pip 01787 ) ( argentina ) , anpcyt ( pict 00909)(argentina ) , and unlp ( proy .
11/x492 ) ( argentina ) , nrf ( south africa ) , and national institute for theoretical physics ( south africa ) .
99 w. heisenberg and h. euler , z. phys . * 98 * , 714 ( 1936 ) , english translation in w. korolevski and h. kleinert , arxiv : physics/0605038 ; v. weisskopf , k. dan .
selsk . mat .
* 14 * , 6 ( 1936 ) , english translation in early quantum electrodynamics : a source book , a.i .
miller ed . , ( cambridge university press , 1994 ) .
j. schwinger , phys . rev . * 82 * , 664 ( 1951 ) .
a. di piazza , e. loetstedt , a.i .
milstein , and c.h .
keitel , arxiv:0906.0726 .
t. heinzl , b. liesfeld , k.u .
amthor , h. schwoerer , r. sauerbrey , and a. wipf , opt . commun . * 267 * , 318 ( 2006 ) . s.l .
adler , j. phys .
a : math . theor . * 40 * , f143 ( 2007 ) .
a. di piazza , k.z .
hatsagortysan , and c.h .
keitel , phys .
lett . * 97 * , 083603 ( 2008 ) .
z. fried , a. baker , and d. korff , phys .
151 * , 1040 ( 1966 ) .
dominguez , h. falomir , m. ipinza , m. loewe , and j.c .
rojas , mod .
lett . a * 24 * , 1857 ( 2009 ) .
dominguez , h. falomir , m. ipinza , s. kohler , m. loewe , and j.c .
rojas , phys .
d * 80 * , 033008 ( 2009 ) .
turchin , slow neutrons , israel program for scientific translations ltd .
h. rauch and w. waschkowski , neutron scattering lengths , in : neutron data booklet , eds . a .- j .
dianoux and g. lander , institut laue langevin ( 2002 ) .
dunne , in _ from fields to strings _ , vol . 1 , 445 , m. shifman , a. vainshtein , and j. wheater eds .
( world scientific , singapore , 2005 ) .
lushchikov , yu.v .
taran , and f.l .
shapiro , sov .
. phys . * 10 * , 669 ( 1970 ) .
w. heil , k. andersen , d. hofmann , h. humblot , j. kulda , e. lelievre - berna , o. schrpf , f. tasset , physica b * 241 - 243 * , 56 ( 1998 ) .
squires , introduction to the theory of thermal neutron scattering , cambridge university press ( 1978 ) .
m. forte , b.r .
heckel , n.f .
ramsey , k. green , g.l .
greene , j. byrne , j.m .
pendlebury , phys .
45 * , 564 ( 1980 ) .
alfimenkov , s.b .
borzakov , vo van thuan , et al .
a * 398 * , 93 ( 1983 ) .
l. koester , w. nistler , and w. waschkowski , phys . rev
. lett . * 36 * , 1021 ( 1976 ) .
s. kopecky , j.a .
harvey , n.w .
hill , m. krenn , m.pernicka , p. riehs , and s. steiner , phys .
c * 56 * , 2229 ( 1997 ) .
barashenkov , i.p .
stakhanov , and yu.a .
aleksandrov , j. exp .
phys . * 32 * , 154 ( 1957 ) .
thaler , phys . rev . * 114 * , 827 ( 1959 ) .
j. schmiedmayer , h. rauch , and p. riehs , nucl .
instr . meth .
a * 284 * , 137 ( 1989 ) .
j. schmiedmayer , p. riehs , j.a .
harvey , and n.w .
hill , phys .
lett * 66 * , 1015 ( 1991 ) .
v. bernard , b. hiller , and w. weise , phys .
b * 205 * , 16 ( 1988 ) , and references therein .
g. breit and m.l .
rustgi , phys . rev . * 114 * , 830 ( 1959 ) .
uehling , phys .
* 48 * , 55 ( 1935 ) .
sears , phys .
rep . * 141 * , 281 ( 1986 ) . | it has been shown recently that a neutron placed in an external quasistatic electric field develops an induced electric dipole moment @xmath0 due to quantum fluctuations in the qed vacuum . a feasible experiment which could detect such an effect
is proposed and described here .
it is shown that the peculiar angular dependence of @xmath0 on the orientation of the neutron spin leads to a characteristic asymmetry in polarized neutron scattering by heavy nuclei .
this asymmetry can be of the order of @xmath1 for neutrons with epithermal energies .
for thermalized neutrons from a hot moderator one still expects experimentally accessible values of the order of @xmath2 .
the contribution of the induced effect to the neutron scattering length is expected to be only one order of magnitude smaller than that due to the neutron polarizability from its quark substructure .
the experimental observation of this scattering asymmetry would be the first ever signal of nonlinearity in electrodynamics due to quantum fluctuations in the qed vacuum .
+ pacs numbers : 12.20.ds , 11.10.-z , 11.10.lm , 29.27.hj , 14.20.dh * o. zimmer*@xmath3 , * c. a. dominguez*@xmath4 , * h. falomir*@xmath5 , * m. loewe*@xmath6 @xmath3 institut laue langevin , 38042 grenoble , france @xmath7centre for theoretical & mathematical physics , university of cape town , rondebosch 7700 , south africa @xmath8 department of physics , stellenbosch university , matieland 7600 , south africa @xmath9 national institute of theoretical physics , private bag x1 , matieland 7602 , south africa @xmath10 instituto de fisica la plata , consejo nacional de ciencia y tecnica , facultad de ciencias exactas , universidad nacional de la plata , argentina @xmath6facultad de fsica , pontificia universidad catlica de chile , casilla 306 , santiago 22 , chile |
the overall energy budget of the universe is dictated by dark matter and dark energy with a minor contamination from baryonic matter , where dark energy is supposed to dominate the cosmic landscape causing the acceleration of universe . since 1998 , we have witnessed astrophysical observations , still we are struggling to find the suitable candidates for dark energy from fundamental physics .
many cosmologists believe that the simplest candidate for the dark energy is the cosmological constant @xmath2 since it fits well with observational data . during the cosmological evolution ,
the @xmath3-term has the density density and pressure @xmath4 .
however , one has the reason to dislike @xmath3-term because it always suffers from the `` fine - tunning '' problems and `` cosmic coincidence '' puzzles @xcite on the theoretical ground .
+ in this regard we note that the universe is mostly observed to be flat and isotropic and supports the predictions of @xmath3cdm model .
however , observations of high resolution cmb radiation data from wilkinson microwave anisotropy probe ( wmap ) showing some large angle anomalies of the large scale structure of the universe with an asymmetric expansion @xcite .
plank data also show a slight redshift of the primordial power spectrum of curvature perturbation from exact scale invariant @xcite .
these observations obviously hint towards the presence of some anisotropy energy source in the universe with anisotropic pressures .
the issue of global anisotropy can be settled if anisotropy can be incorporated to be frw models as a sort of small perturbation . in order to address the issue of smallness in the angular power spectrum ,
some anisotropic models have been proposed in recent times @xcite .
these models have a similarity to the bianchi morphology @xcite .
spatially homogeneous bianchi i model is more general than the frw model and have anisotropic spatial sections .
the bianchi type models provide an opportunity to consider asymmetric expansion along different spatial sections . very recently
a study have been carried out @xcite in bianchi i space time with dominance of dark energy .
some other recent works on anisotropic dark energy are also available in the literature @xcite@xmath5@xcite .
+ lie group of transformations has been extensively applied to linear and nonlinear differential equations in the area of theoretical physics such as : general relativity , particle physics and cosmology @xcite .
the method of lie symmetry group is one of the most useful tools for finding exact solutions for the einstein field equations described by a system of nlpdes .
recently we have developed a formalism to solve non linear einstein s field equations in general relativity @xcite . in our earlier work @xcite , we have proposed the invariant solution of dark energy ( de ) model in cylindrically symmetric space - time while in this paper , we confine ourselves to investigate the similarity solution of anisotropic de model in bianchi - i space - time which is entirely different from @xcite . in the astrophysical community ,
the inhomogeneous cosmological models have gained interest due to exact perturbation of frw model and are more often employed to study cosmological phenomenon .
that is why , here , we consider inhomogeneous bianchi - i space - time .
the paper is organized as follows : in section 2 , the basic formalism for anisotropic de has been discussed for an anisotropic and inhomogeneous bianchi i space - time .
similar formalism has been already developed in our earlier work @xcite . in section 3 ,
lie group analysis method is developed for bianchi i space - time .
section 4 and 5 deal with the optimal system and similarity solutions of the models .
the physical viability of the discussed dark energy model is tested with graphical analysis of cosmological parameters in section 6 .
finally the conclusions are summarized in section 7 .
the bianchi type - i space - time is given by @xmath6 where the metric potentials @xmath7 , @xmath8 and @xmath9 are functions of @xmath10 and @xmath11 .
einstien s field equations in the case of a mixture of perfect fluid and anisotropic dark energy are given by @xmath12 with @xmath13=diag[-1 , \omega^{(pf ) } , \omega^{(pf ) } , \omega^{(pf)}]\rho^{(pf)}\ ] ] and @xmath14 = diag[-1 , \omega_x^{(de ) } , \omega_y^{(de ) } , \omega_z^{(de)}]\rho^{(de)}\ ] ] where @xmath15 are the metric tensor with @xmath16 ; @xmath17 is the flow vector ; @xmath18 is the ricci tensor ; @xmath19 is the ricci scalar ; @xmath20 , @xmath21 and @xmath22 are , respectively the pressure and energy density of the perfect fluid and dark energy components ; @xmath23 is the eos parameter of perfect fluid with @xmath24 ; @xmath25 , @xmath26 and @xmath27 are the deviation - free eos parameters of dark energy , respectively , on the @xmath10 , @xmath28 and @xmath29 axis .
+ in co - moving coordinate system , the field equation ( [ efe ] ) , for the inhomogeneous space - time ( [ spacetime ] ) , read as @xmath30 @xmath31-\frac{\ddot{a}}{a}-\frac{\dot{a}\dot{c}}{ac}-\frac{\ddot{c}}{c},\ ] ] @xmath32-\frac{\ddot{a}}{a}-\frac{\dot{a}\dot{b}}{ab}-\frac{\ddot{b}}{b},\ ] ] @xmath33-\frac{\dot{a}\dot{b}}{ab}-\frac{\dot{a}\dot{c}}{ac}-\frac{\dot{b}\dot{c}}{bc},\ ] ] @xmath34.\ ] ] here @xmath35 , @xmath36 and so on .
+ the velocity field @xmath17 is ir - rotational . the scalar expansion @xmath37 , shear scalar @xmath38 and volume scalar @xmath39 are respectively have the form @xcite : @xmath40 @xmath41 @xmath42 where @xmath43 is the determinant of the metric ( [ spacetime ] ) .
the shear tensor is @xmath44 and the non - vanishing components of the @xmath45 are @xmath46 the einstein s field equations ( [ efe1])-([efe5 ] ) constitute a system of five highly nlpdes with six unknowns variables , @xmath7 , @xmath8 , @xmath9 , @xmath20 , @xmath21 and @xmath22 .
therefore , one physically reasonable conditions amongst these parameters are required to obtain explicit solutions of the field equations .
let us assume that the metric potential function @xmath7 is a function of the time only , i.e. , @xmath47 .
if we substitute the metric function @xmath47 in the einstein field equations , the equations ( [ efe1])-([efe5 ] ) transform to the nlpdes of the coefficients @xmath8 and @xmath9 only , as the following new form : @xmath48\\ \\ \,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\
, + \left(\dfrac{\omega_x^{(de)}-\omega_z^{(de)}}{\omega_z^{(de)}-\omega_y^{(de)}}\right)\left[\dfrac{\ddot{b}}{b}-\dfrac{\dot{a}\dot{c}}{ac}+\dfrac{c''}{a^2c}\right ] = 0 , \end{array } % \right.\ ] ] @xmath49 where @xmath50 @xmath51 @xmath52
we consider a one - parameter lie group of transformations @xmath53 with a small parameter @xmath54 , where @xmath55 , @xmath56 , @xmath57 and @xmath58 .
the coefficients @xmath59 , @xmath60 , @xmath61 and @xmath62 are the functions of @xmath10 , @xmath11 , @xmath8 and @xmath9 .
the system ( [ u210 - 1])-([u210 - 2 ] ) is invariant under the transformations given in eq .
( [ u31 ] ) and the corresponding infinitesimal generator of lie groups ( symmetries ) @xmath63 must satisfy the invariance conditions : @xmath64 where @xmath65 are the system ( [ u210 - 1])-([u210 - 2 ] ) under study and @xmath66 is the second prolongation of the symmetries @xmath67 .
the detail of lie point symmetry has been already given in @xcite .
finally the characteristic equations are given by @xmath68
the general lie point symmetries ( [ u32 ] ) becomes @xmath69 where , the non - linear einstein field equations ( [ u210 - 1])-([u210 - 2 ] ) admits the 6-dimensional lie algebra spanned by the independent symmetries shown below : @xmath70 the forms of the symmetries @xmath71 , @xmath72 suggest their significations : @xmath73 , @xmath74 generate the symmetry of space translation , @xmath75 , @xmath74 , @xmath76 , @xmath77 are associated with the scaling transformations .
when the lie algebra of these symmetries is computed , the only non - vanishing relations are : @xmath78\,=-x_2,\,\,\,\,\,\,\,\,\,\,[x_3,x_4]=-x_4.\ ] ] following yadav and ali @xcite , we acquire an optimal system of one - dimensional subalgebras to be those spanned by : @xmath79
if we considered the symmetries @xmath80 or @xmath81 or @xmath82 or @xmath83 , then @xmath84 , we shall analyze the similarity solutions associated with the optimal systems of symmetries @xmath85 , @xmath86 , @xmath87 and @xmath88 only as the following : + * solution ( i ) : * the symmetries @xmath85 has the characteristic equations : @xmath89 then the similarity variable and the similarity transformations takes the form : @xmath90 where @xmath91 , @xmath92 and @xmath93 are an arbitrary constants . in this case
, we have @xmath94 where @xmath95 , @xmath96 and @xmath97 are an arbitrary constants . substituting the transformations ( [ u42 - 1 ] ) in the field eqs .
( [ u210 - 1])-([u210 - 2 ] ) lead to the following system of ordinary differential equations : @xmath98 \,=\,\dfrac{(a-1)(b+c)}{a\,\xi } , \end{array}\ ] ] @xmath99 -a\,d^2\,\xi\bigg[\big(\alpha_2+a\,(r - q)\,\xi\dfrac{\phi'}{\phi}\big)\dfrac{\psi'}{\psi}+a\,r\,\xi\,\dfrac{\psi''}{\psi}\bigg ] \\ \,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\ , + d^2\bigg[(a-1)(q - r)+a\,\xi\,\big(\alpha_3\,\dfrac{\phi'}{\phi}+a\,q\,\xi\,\dfrac{\phi''}{\phi}\big)\bigg ] -a^2\,\xi^{-\frac{2}{a}}\bigg[\alpha_4+a\,\xi\,\big(\alpha_5\,\dfrac{\phi'}{\phi}+a\,r\,\xi\,\dfrac{\phi''}{\phi}\big)\bigg]=0 , \end{array}\ ] ] where @xmath100 the equations ( [ u43 - 1 ] ) and ( [ u43 - 2 ] ) are non - linear ordinary differential equations
. one can not solve these equation in general .
however , in a special cases , one can find a solution .
we propose the following conditions : @xmath101 where @xmath102 , @xmath103 , @xmath104 , @xmath105 and @xmath105 are an arbitrary constants . substitute ( [ u43 - 4 ] ) in ( [ u43 - 1 ] ) , we have the following equation : @xmath106\\ \,\,\,\,\,\,\,\,\,\ , + \beta_3\,\left[4\,\beta_2\,(1+\beta_2)-2\,\beta_4\,\big(b+c-2+b\,\beta_2-\beta_2\big)-\beta_4 ^ 2\,(b+3\,c-6)\right]\,\xi^{\beta_4}=0 . \end{array}\ ] ] the coefficients of @xmath107 and the absolute value must be equal zero . solving the two resulting conditions with respect to @xmath108 and @xmath109 , we have : @xmath110 therefore the ( [ u43 - 2 ] becomes @xmath111 where @xmath112,\\ \\
\gamma_1=d^2\,\beta_3\,\beta_4 ^ 2\,(2\,\beta_2+\beta_4)\,(2 + 2\,\beta_2+\beta_4)^2\big [ r\,(2 + 2\,\beta_2+\beta_4)-q\,(2 + 3\,\beta_4)\big]\\ \,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,+4\,\beta_5\,(2 + 3\,\beta_4)\,\big ( r\,(2 + 2\,\beta_2+\beta_4)\,\big[4\,\beta_2\,(1+\beta_4)+(2+\beta_4)\,(4 + 5\,\beta_4)\big]\\ \,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\ , -q\,(2+\beta_4)\,\big[(2+\beta_4)\,(4 + 7\,\beta_4)+4\,\beta_2\,(1 + 2\,\beta_4)\big]\big),\\ \\ \gamma_2=4\,\beta_3\,(2+\beta_4)\,\big ( r\,(2 + 2\,\beta_2+\beta_4)\big[4\,\beta_2+(2+\beta_4)\,(4 + 3\,\beta_4)\big]\\ \,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\ , -q\,(2 + 3\,\beta_4)\big[4\,\beta_2\,(1+\beta_4)+(2+\beta_4)\,(4 + 5\,\beta_4)\big ] \big ) .
\end{array } \right.\ ] ] solving the three equations @xmath113 , @xmath114 and @xmath115 with respect to @xmath116 , @xmath103 and @xmath117 , we can get the following solutions : @xmath118 here , we consider only one of the above solutions . + by using the solution ( 3 ) in ( [ u43 - 11 ] ) , ( [ u43 - 8 ] ) , ( [ u43 - 4 ] ) , ( [ u42 - 2 ] ) and ( [ u42 - 1 ] ) , we obtain the solution of the einstein field equations as the following : @xmath119 where @xmath120 , @xmath121 and @xmath105 are an arbitrary constants , while @xmath122 is an arbitrary function of @xmath11 . + it is observed from equations ( [ uu1 ] ) , the line element ( [ spacetime ] ) can be written in the following form : @xmath123 * remark : * in the above solution , we can replace @xmath11 by @xmath124 and @xmath10 by @xmath125 without loss of generality , where @xmath126 and @xmath127 are an some arbitrary constants .
+ * solution ( ii ) : * the symmetries @xmath86 has the characteristic equations : @xmath128 then the similarity variable and the similarity transformations takes the form : @xmath129,\,\,\,\,\,\,b(x , t)=\psi(\xi)\,\exp\big[b\,t\big],\,\,\,\,\,\,c(x , t)=\phi(\xi)\,\exp\big[c\,t\big ] , \end{array}\ ] ] where @xmath130 , @xmath92 and @xmath93 are an arbitrary constants . in this case
, we have @xmath131,\,\,\,\,\,\,\ , \omega_y^{(de)}(t)=\omega_x^{(de)}(t)+q\,\exp\big[-2\,a\,t\big],\,\,\,\,\,\,\ , \omega_z^{(de)}(t)=\omega_x^{(de)}(t)+r\,\exp\big[-2\,a\,t\big ] , \end{array } % \right.\ ] ] where @xmath132 , @xmath133 and @xmath134 are an arbitrary constants . substituting the transformations ( [ u52 - 1 ] ) in the field eqs .
( [ u210 - 1])-([u210 - 2 ] ) , we can get the following system of ordinary differential equations : @xmath135 @xmath136 + \big(a^2\,d^2\,r\,\xi^2-q\big)\dfrac{\psi''}{\psi}-\big(a^2\,d^2\,q\,\xi^2-r\big)\dfrac{\phi''}{\phi}=0 , \end{array}\ ] ] where @xmath137n\,(1+q)-1\big],\\ \alpha_2=a\,d^2\,\big[(a+b+2\,c)\,q+(a - b)\,r\big],\\ \alpha_3=a\,d^2\,\big[(a - c)\,q+(a+2\,b+c)\,r\big ] .
\end{array } \right.\ ] ] * solution ( iii ) : * the symmetries @xmath87 has the characteristic equations : @xmath138 then the similarity variable and the similarity transformations takes the form : @xmath139,\,\,\,\,\,\,b(x , t)=\psi(\xi)\,\exp\big[b\,x\big],\,\,\,\,\,\,c(x , t)=\phi(\xi)\,\exp\big[c\,x\big ] , \end{array}\ ] ] where @xmath140 , @xmath92 and @xmath93 are an arbitrary constants . in this case
, we have @xmath141 where @xmath132 , @xmath133 and @xmath134 are an arbitrary constants . substituting the transformations ( [ u52 - 1 ] ) in the field eqs .
( [ u210 - 1])-([u210 - 2 ] ) , we obtain the following system of ordinary differential equations : @xmath142 @xmath143 where @xmath144+a\,c\,(q - r),\\ \alpha_2=(a - d)\,(a+d)\,(q - r),\\ \alpha_3=r\,d^2-q\,a^2,\\ \alpha_4=(c - b)\,\big[b\,q+c\,r\big],\\ \alpha_5=a\,b\,q - r\,d^2+a\,r\,(a - b+2\,c),\\ \alpha_6=a^2\,r - d^2\,q . \end{array } \right.\ ] ] equation ( [ u63 - 1 ] ) can be written in the following form : @xmath145 , \end{array}\ ] ] substituting @xmath146 into ( [ u63 - 2 ] ) , we obtain the following equation : @xmath147=\dfrac{\alpha_6}{\xi}+\dfrac{\alpha_7\,\psi'}{\psi } + \dfrac{\alpha_8\,\phi'}{\phi } , \end{array}\ ] ] where @xmath148 it is easy to solve the above equation when @xmath149 or @xmath150 . when @xmath149 , the equation ( [ u64 - 2 ] ) takes the following form : @xmath151\,\dfrac{\psi'}{\psi } + \big[b\,(q - r)+c(q+r)\big]\,\dfrac{\phi'}{\phi}. \end{array}\ ] ] the solution of the above equation is @xmath152 where @xmath102 is an arbitrary constant of integration .
+ substituting the above equation into ( [ u63 - 1 ] ) , we get the following equation @xmath153=0 , \end{array}\ ] ] where @xmath154-(b^2+c^2)\,\big[b\,(q - r)^2-c\,(q^2 + 3\,r^2)\big],\\ \\ \gamma_3=b^2\,r\,(r - q)+c^2\,q\,(q+r)+b\,c\,(q^2 - 2\,q\,r - r^2),\\ \\ \gamma_4=c^2\,r\,(r - q)+b^2\,q\,(q+r)+b\,c\,(r^2 + 2\,q\,r - q^2 ) . \end{array } \right.\ ] ] the general solution of the equation ( [ u67 - 1 ] ) is read as @xmath155 where @xmath103 and @xmath104 are an arbitrary constants of integration and @xmath156}{(b^2+c^2)\,(q^2+r^2)}. \end{array } \right.\ ] ] substituting ( [ u68 - 1 ] ) in ( [ u66 ] ) , we have @xmath157 where @xmath158 is the new arbitrary constant and @xmath159}{(b^2+c^2)\,(q^2+r^2)}. \end{array } \right.\ ] ] without loss of generality , we can take @xmath160 . now using the solution ( [ u68 - 1 ] ) and ( [ u69 - 1 ] ) in ( [ u43 - 11 ] ) , ( [ u43 - 8 ] ) , ( [ u43 - 4 ] ) , ( [ u42 - 2 ] ) and ( [ u42 - 1 ] ) , we can find the solution of the einstein field equations as the following : @xmath161^{\frac{b\,(b+c)}{b^2+c^2}},\\ \\ c(x , t)=\beta_2\,t\,e^{(a+c)\,x}\,\left[\beta_3+\big(t\,e^{a\,x}\big)^{\frac{b^2 + 2\,a\,c+c^2}{2\,a\,b}}\right]^{\frac{b\,(b - c)}{b^2+c^2}},\\ \end{array } \right.\ ] ] where @xmath162 , @xmath108 @xmath109 , @xmath103 , @xmath104 , @xmath105 and @xmath121 are an arbitrary constants , while @xmath122 is an arbitrary function of @xmath11 .
it is observed from equations ( [ uu2 ] ) , the line element ( [ spacetime ] ) can be written in the following form : @xmath163^{\frac{2\,b\,(b+c)}{b^2+c^2}}\,dy^2\\ \\ \,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\ , + \beta_2 ^ 2\,t^2\,e^{2\,(a+c)\,x}\,\left[\beta_3+\big(t\,e^{a\,x}\big)^{\frac{b^2 + 2\,a\,c+c^2}{2\,a\,b}}\right]^{\frac{2\,b\,(b - c)}{b^2+c^2}}\,dz^2-dt^2 . \end{array}\ ] ] * solution ( iv ) : * the symmetries @xmath88 has the characteristic equations : @xmath164 then the similarity variable and the similarity transformations takes the form : @xmath165,\,\,\,\,\,\,c(x , t)=\phi(\xi)\,\exp\big[c\,t\big ] , \end{array}\ ] ] where @xmath166 , @xmath167 and @xmath168 are an arbitrary constants . in this case
, we have @xmath169 where @xmath132 , @xmath133 and @xmath134 are an arbitrary constants . + substituting the transformations ( [ u52 - 1 ] ) in the field eqs .
( [ u210 - 1])-([u210 - 2 ] ) , we can get the following system of ordinary differential equations : @xmath170 @xmath171+d^2\big[2\,b\,r+c\,(r - q)\big]\,\dfrac{\psi'}{\psi}+d^2\big[b\,(r - q)-2\,c\,r\big]\,\dfrac{\phi'}{\phi}\\ \,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\ , + ( a^2-d^2)\,(q - r)\,\dfrac{\psi'\phi'}{\psi\phi}+(d^2\,r - a^2\,q)\dfrac{\psi''}{\psi}+(a^2\,r - d^2\,q)\dfrac{\phi''}{\phi}=0 . \end{array}\ ] ]
* for the model ( [ s1 ] ) : * + the expressions of @xmath20 , @xmath21 and @xmath22 for the model ( [ s1 ] ) , are given by : @xmath172\,x^2\,t^{-2}\\ \,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\ , -q\,d_0 ^ 4\,(2+\beta_4)\,(4 + 3\,\beta_4)\,\big[16 + 56\,\beta_4 + 60\,\beta_4 ^ 2 + 19\,\beta_4 ^ 3\big]\,x^2\,t^{-2+\beta_4}\\ \,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\ , + 4\,(1+\beta_4)\,(8 + 12\,\beta_4 + 3\,\beta_4 ^ 2)\big[2\,(1+\beta_4)\,(2+\beta_4)+d_0 ^ 2\,(4 + 10\,\beta_4 + 5\,\beta_4 ^ 2)\,x^2\,t^{\beta_4}\big]\omega_x^{(pf)}(t ) \bigg ] , \end{array } % \right.\ ] ] @xmath173\\ \,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\ , -4\,q\,(1+\beta_4)\,(2+\beta_4)\,(8 + 24\,\beta_4 + 26\,\beta_4 ^ 2 + 9\,\beta_4 ^ 3)\,t^{-2-\beta_4}\\ \,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\ , -q\,d_0 ^ 2\,\big[320 + 1344\,\beta_4 + 1968\,\beta_4 ^ 2 + 1160\,\beta_4 ^ 3 + 194\,\beta_4 ^ 4 - 27\,\beta_4 ^
5\big]\,x^2\,t^{-2}\\ \,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\ , + q\,d_0 ^ 4\,(2+\beta_4)\,(4 + 3\,\beta_4)\,\big[16 + 56\,\beta_4 + 60\,\beta_4 ^ 2 + 19\,\beta_4 ^ 3\big]\,x^2\,t^{-2+\beta_4 } \bigg ] , \end{array } % \right.\ ] ] @xmath174}{f(x , t ) } , \end{array } % \right.\ ] ] where @xmath175\big(d_0 ^ 2\,x^4\,t^{\beta_4}-x^2\big).\ ] ] the volume element is @xmath176 the expansion scalar is given by : @xmath177 the non - vanishing components of the shear tensor , @xmath45 , are : @xmath178 } , \end{array } % \right.\ ] ] @xmath179 } , \end{array } % \right.\ ] ] @xmath180}. \end{array } % \right.\ ] ] the shear scalar is : @xmath181 ^ 2}. \end{array } % \right.\ ] ] the deceleration parameter is given by @xcite @xmath182}{t^4\,\big[d_0 ^ 2\,x^2\,t^{\beta_4}-1\big]^4}. \end{array } % \right.\ ] ] [ cols= " > , < " , ] * for the model ( [ s2 ] ) : * + the expressions of @xmath20 , @xmath21 and @xmath22 for the model ( [ s2 ] ) , are given by : @xmath183\big(t\,e^{a\,x}\big)^{\frac{b^2 + 2\,a\,c+c^2}{2\,a\,b } } -\beta_3\,(b^2+c^2)\,\big[b^2\,q+(b^2 + 2\,a\,c+c^2)\,t^2\,\omega_x^{(de)}(t)\big]}{g(x , t ) } , \end{array } % \right.\ ] ] @xmath184\,\big(t\,e^{a\,x}\big)^{\frac{b^2 + 2\,a\,c+c^2}{2\,a\,b}}\\ \\ \,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\ , \,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\
, + \beta_3\,(b^2+c^2)\,\big[q\,\big[c^2 + 2\,a\,(b+c)\big]-(b^2 + 2\,a\,c+c^2)\,t^2\big]\bigg ) , \end{array } % \right.\ ] ] @xmath185 where @xmath186 $ ] .
+ the volume element is @xmath187^{\frac{2\,b^2}{b^2+c^2}}.\ ] ] the expansion scalar is given by : @xmath188}\bigg ) , \end{array } % \right.\ ] ] the non - vanishing components of the shear tensor , @xmath45 , are : @xmath189 @xmath190\,\big(t\,e^{a\,x}\big)^{\frac{b^2 + 2\,a\,c+c^2}{2\,a\,b } } } { 6\,\beta_3\,(b-2\,a)\,(b^2+c^2)-12\,a\,(b^2+b\,c+c^2)^2\,\big(t\,e^{a\,x}\big)^{\frac{b^2 + 2\,a\,c+c^2}{2\,a\,b } } } , \end{array } % \right.\ ] ] @xmath191\,\big(t\,e^{a\,x}\big)^{\frac{b^2 + 2\,a\,c+c^2}{2\,a\,b } } } { 6\,\beta_3\,(2\,a - b)\,(b^2+c^2)+12\,a\,(b^2+b\,c+c^2)^2\,\big(t\,e^{a\,x}\big)^{\frac{b^2 + 2\,a\,c+c^2}{2\,a\,b}}}. \end{array } % \right.\ ] ] the shear scalar is : @xmath192\,\big(t\,e^{a\,x}\big)^{\frac{b^2 + 2\,a\,c+c^2}{2\,a\,b}}\\ \\ \,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\ , + ( b - c)^2\,\big[6\,a\,(b+c)\,(b^2+c^2)+3\,(b^2+c^2)^2 + 4\,a\,(b^2+b\,c + c^2)\big]\,\big(t\,e^{a\,x}\big)^{\frac{b^2 + 2\,a\,c+c^2}{a\,b}}\bigg ) , \end{array } % \right.\ ] ] where @xmath193 ^ 2 .
\end{array } % \right.\ ] ] the deceleration parameter is given by : @xmath194\,\big(t\,e^{a\,x}\big)^{\frac{b^2 + 2\,a\,c+c^2}{2\,a\,b}}\\ \\ \,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\
, + 4\,a^2(b - c)^2\,(b^2+b\,c+c^2)\,\big(t\,e^{a\,x}\big)^{\frac{b^2 + 2\,a\,c+c^2}{a\,b}}\bigg)\\ \,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\ , \times\,\bigg(b\,\beta_3\,(b^2 + 2\,a\,c+c^2)+2\,a\,(b^2+b\,c+c^2)^2\,\big(t\,e^{a\,x}\big)^{\frac{b^2 + 2\,a\,c+c^2}{2\,a\,b}}\bigg)^2 , \end{array } % \right.\ ] ] where @xmath195 ^ 4 .
\end{array } % \right.\ ] ]
in this paper , we have investigated some models of accelerating universe with minimal interaction between normal matter and anisotropic de in anisotropic and inhomogeneous bianchi i space - time .
generally , the models behave like an expanding , shearing and non - rotating universe in which flow vector is geodetic . on the basis of optimal systems of symmetries @xmath85 and @xmath86 , we obtain two models ( [ s1 ] ) and ( [ s2 ] ) respectively . the main features of the work are as follows : * the models are based on similarity solution of field equation and we have obtained the new class of exact solution . * we have , in general discussed several physical features and geometrical properties of the models . however , as a special case , most notable aspect of the solutions have been studied that are non - singular in nature .
all figures depicts interesting features of the present cosmological models in terms of dp and other physical parameters . * in the derived models , the matter energy density and dark energy density remains positive .
therefore , the wec and nec are satisfied , which in turn imply that the derived models are physically realistic .
* as @xmath196 , @xmath197 but @xmath198 , hence the volume increases with passage of time while de density decreases . *
the derived models seem to describes the dynamics of universe from big bang to present epoch and de dominates the universe at present time which may be attributed to the current accelerated expansion of universe . *
hypothetical de is the most reasonable way of explaining why the universe is expanding at an ever increasing rate
. de plays a massive part in shaping our reality , however , it is to be noted that no body seems certain of what the dang stuff actually is .
future space mission hope to solve this mystery and shake up our current understanding of the universe . to our knowledge , this work is the first study of minimally interacting anisotropic de with normal matter in bianchi - i inhomogeneous space - time and its general form .
it is important to note that @xmath10 = constant removes the inhomogeneity from the derived models and they simply represent model of universe based on power law cosmology .
numerous cosmological models with power law expansion exist in the literature .
thus the solutions , presented in this paper generalize the solution obtained by numerous authors in spatially homogeneous and anisotropic bianchi i space - time , in particular yadav and saha @xcite .
+ * acknowledgment : * this paper is funded by the deanship of scientific research ( dsr ) , king abdulaziz university , jeddah , under grant no . (
130682d1435 ) .
the authors , therefore , acknowledge with thank dsr technical and financial support . | we investigate some new similarity inhomogeneous solutions of anisotropic dark energy and perfect fluid in bianchi type - i space - time .
three different equation of state parameters along the spatial directions are introduced to quantify the deviation of pressure from isotropy .
we consider the case when the dark energy is minimally coupled to the perfect fluid as well as direct interaction with it .
the lie symmetry generators that leave the equation invariant are identified and we generate an optimal system of one - dimensional subalgebras . each element of the optimal system is used to reduce the partial differential equation to an ordinary differential equation which is further analyzed .
we solve the einstein field equations , described by a system of non - linear partial differential equations ( nlpdes ) , by using the lie point symmetry analysis method .
the geometrical and kinematical features of the models and the behavior of the anisotropy of dark energy , are examined in detail .
, anil kumar yadav@xmath0 and abdulah k alzahrani@xmath1 + + e - mail : [email protected] + + + _ pacs : _
98.80.jk , 98.80.-k . _ keywords
_ : optimal system , similarity solutions , dark energy , general relativity . |
in @xcite s. negami proved that if a graph @xmath2 has a finite , unbranched , regular , planar cover then @xmath2 itself embeds in the projective plane @xmath3 ( for definitions see section [ sec : prelims ] ) .
we call a graph that embeds in @xmath3 _ projective_. note that all planar graphs are projective . in the same paper ,
negami conjectured that this holds in general : [ cnj : negami ] if a graph @xmath2 s has a finite unbranched planar cover then @xmath2 embeds in the projective plane @xmath3 .
negami s result was extended by s. kitakubo to branched regular covers in @xcite ( with the exception of section [ sec : necessity ] , any cover considered in this paper may be branched ) .
we note that `` branched cover '' is not a standard term for graphs ; for a precise definition and discussion see definitions [ dfn : branched - cover ] and remark [ rmk : branched - cover ] .
the definition used by kitakubo is what we call _ weak cover _ ; it is immediate from the definitions that every branched cover is a weak cover . for regular covers the converse holds as well : every regular weak cover is a branched cover .
we do not have a clear idea about negami s conjecture for weak covers : is negami s conjecture true for weak covers ? given a finite cover @xmath0 and an embedding of @xmath1 into @xmath6 we define two properties called property @xmath4 and property @xmath5 .
these properties depend on the covering map @xmath7 and the embedding of @xmath1 into @xmath6 .
we prove negami s conjecture for covers @xmath0 fulfilling properties @xmath4 and @xmath5 .
we show that regular covers fulfill properties @xmath4 and @xmath5 ( perhaps after re - embedding @xmath1 in @xmath6 ) .
the converse does not hold : in example [ ex : irregular ] we show an irregular cover fulfilling properties @xmath4 and @xmath5 , showing that the work here is more general than @xcite and @xcite .
we now give a more detailed description of our results and the structure of this paper .
in section [ sec : prelims ] we give the necessary definition and background . in section [ sec : sufficiency ] we define properties @xmath4 and @xmath5 ( definitions [ dfn : pv ] and [ dfn : pe ] ) and show ( theorem [ thm : sufficiency ] and corollary [ cor : negami - cnj ] ) that if @xmath0 is a finite planar cover fulfilling properties @xmath4 and @xmath5 then @xmath2 is projective , thus proving negami s conjecture in that case .
in fact , we show a little more : the map @xmath7 can be extended to a map @xmath8 , for some surface @xmath9 , and this map is a branched cover .
it is then easy to see ( lemma [ pro : f - covered - bysorp ] ) that either @xmath10 or @xmath11 . in section [ sec :
regular ] we show that if @xmath12 is a _ regular _ finite , planar cover then it fulfills properties @xmath4 and @xmath5
. in @xcite negami mentions the following strategy for proving conjecture [ cnj : negami ] : given a finite cover @xmath0 it is a well - known application of group theory that there exists finite cover @xmath13 , so that the composition of the covers ( _ i.e. _ , @xmath14 ) is a finite regular cover ( we remark that the degree of @xmath15 is usually quite high )
. however , negami continues , even if we assume that @xmath1 is planar , it does _ not _ follow that @xmath16 is planar as well . in section [ sec : necessity ]
we prove theorem [ thm : necessity ] : let @xmath12 be a finite _ unbranched _ cover and let @xmath17 be a cover with @xmath18 . if the cover @xmath19 fulfills properties @xmath4 and @xmath5 , then either @xmath1 is planar and @xmath0 fulfills properties @xmath4 and @xmath5 , or @xmath1 embeds in @xmath3 ; in that case , lifting @xmath1 to the universal cover of @xmath3 ( _ i.e. _ , to the double cover @xmath20 ) we obtain a finite planar cover of @xmath2 fulfilling properties @xmath4 and @xmath5(this is sharp
see example [ ex : have - to - lift ] ) .
thus , if we wish to pass to a cover of @xmath1 in oder to prove that @xmath2 is planar , we need not look for further than double covers . in section [ sec :
examples ] we give examples .
the first is example [ ex : irregular ] , which is an irregular cover fulfilling properties @xmath4 and @xmath5 .
next ( example [ ex : k4 ] ) we give an example of two distinct double covers of the planar graph @xmath21 , both fulfilling properties @xmath4 and @xmath5 .
we use both covers to embed @xmath21 .
the first gives an embedding into @xmath6 and the second into @xmath3 , demonstrating that the embedding depends on the cover in a non - trivial way .
next , we exemplify theorem [ thm : necessity ] by showing ( example [ ex : have - to - lift ] ) a planar cover @xmath22 that does not fulfill properties @xmath4 and @xmath5 for any planar embedding of @xmath23
. however , @xmath23 embeds in @xmath3 and the lift of that embedding to @xmath6 does fulfill properties @xmath4 and @xmath5 . finally , we show a more disturbing phenomenon : in conjecture [ ex : no properties ] we show a planar finite cover @xmath22 and we _ conjecture _ that is has no finite cover @xmath24 so the composition @xmath25 fulfills properties @xmath4 and @xmath5 .
what happens here is that the cover @xmath23 is simply the _
wrong cover _ _ , and this can not be fixed by passing to a higher cover .
this , perhaps , explains the difficulty in proving negami s conjecture .
_ _ _ so is negami s conjecture true ? at the current time , the answer is not known . however , the work of many people ( including d. archdeacon , m. fellows , p. hlinn , negami , r. thomas ) accumulated in proving that negami s conjecture is equivalent to the statement : _ the graph @xmath26 has no finite planar cover .
_ this seems to us very strong evidence supporting conjecture [ cnj : negami ] .
assuming conjecture [ cnj : negami ] for a moment , in light of the results and examples in this paper we ask : suppose we are given a finite planar cover @xmath0 not fulfilling properties @xmath4 and @xmath5 for any embedding of @xmath1 ( or a cover of @xmath1 ) into @xmath6 ( _ i.e. _ , a `` wrong cover '' ) , how can we correct that cover ?
we end section [ sec : examples ] with two ways to `` fix '' the cover in conjecture [ ex : no properties ] .
the first is a cut - and - paste procedure that produces a cover fulfilling properties @xmath4 and @xmath5 ; this changes the graph @xmath1 .
the second does not change @xmath1 but embeds it into some non - orientable surface ( not unlike theorem [ thm : necessity ] ) .
this gives an embedding of @xmath2 into _ some _ surface , not necessarily @xmath3 , with control over the euler characteristic of that surface .
for a graph @xmath2 , we denote the vertices of @xmath2 by @xmath27 and the edges of @xmath2 by @xmath28 .
naturally , any map between graphs is assumed to map vertices to vertices and edges to edges .
we follow standard definitions and terminology used in topology : @xmath29 means _ closed _ regular ( or normal ) neighborhood , @xmath30 is read boundary , @xmath31 is read closure , and @xmath32 is read interior . a homeomorphism is a continuous bijection with a continuous inverse .
@xmath33 stands for euler characteristic .
all surface are assumed to be connected . as it is easy to reduce negami s conjecture to graphs @xmath2 with no cycles of length one or two ( that is
, graphs @xmath2 in which every edge in @xmath28 connects distinct vertices and distinct edges have at most one vertex in common ) .
it is also easy to reduce negami s conjecture to connected graphs @xmath2 and @xmath1 .
therefore , throughout this paper we assume : [ assumptions ] the graphs @xmath2 and @xmath1 are connected and @xmath2 has no cycles of length 2 or less .
we define branched covers in the two relevant situations , graphs and surfaces . in these definitions disks
are modeled on @xmath34 , and a _ half disk _ is a set homeomorphic to ( and modeled on ) @xmath35 ( here , @xmath36 is the imaginary part of @xmath37 ) .
the boundary of a half disk are the points corresponding to @xmath38 . the symbol @xmath39 is used for disjoint unions .
[ dfn : branched - cover ] in ( 1)(3 ) below let @xmath1 and @xmath2 be finite graphs and @xmath0 a map . in ( 4)(5 )
below let @xmath40 and @xmath41 be compact surfaces and @xmath42 a proper map ( that is , @xmath43 ) . 1 . @xmath0 is called a _
unbranched cover _ if @xmath7 is onto and for any @xmath44 @xmath7 maps the neighbors of @xmath45 bijectively onto the neighbors of @xmath46 .
@xmath0 is called a _ weak cover _ if @xmath7 is onto and for any @xmath44 @xmath7 maps the neighbors of @xmath45 onto the neighbors of @xmath46 .
@xmath0 is called a _ branched cover _
if @xmath7 is onto and for any @xmath44 there is a positive integer @xmath47 so that every neighbor of @xmath46 has exactly @xmath48 preimages that are neighbors of @xmath45 , _
i.e. _ , the restriction of @xmath7 to the neighbors of @xmath45 as a map to the neighbors of @xmath46 is onto and @xmath48-to-1 .
+ we call @xmath48 the _ local degree _ of @xmath7 at @xmath45 . if @xmath49 then @xmath45 is called a
_ singular point_. the set of all singular points is called the _ singular set _ and the image of the singular set is called the
_ branched set_. 4 .
map @xmath42 is called a _ branched cover _ if the following holds : 1 .
every point @xmath50 has a neighborhood @xmath51 so that @xmath52 is half a disk ( with @xmath53 corresponding to 0 ) and @xmath54 is a disjoint collection of half disks @xmath55 ( for some @xmath56 ) with @xmath57 , so that for every @xmath58 , @xmath59 is a homeomorphism .
2 . every point @xmath60 has a neighborhood @xmath51 ( with @xmath53 corresponding to 0 ) so that @xmath52 is a disk and @xmath54 is a disjoint collection of disks @xmath55 ( for some @xmath56 ) so that for every @xmath58 , @xmath59 is modeled on @xmath61 for some non - zero integer @xmath48 . +
if @xmath62 is modeled on @xmath61 then @xmath48 is called the _ local degree _ at the center of @xmath63 ( of course , for distinct values of @xmath58 we may have distinct local degrees ) . a point with local degree
greater than one ( in absolute value ) is called a _ singular point _ , the union of the singular points is called the _ singular set _ , and the image of the singular set
is called the _ branched set_. note that the branch set is finite and contained in @xmath64 .
thus we see that every unbranched cover of graphs is a branched cover of graphs with all local degrees one , and conversely a branched cover with all local degrees one is an unbranched cover .
( equivalently , unbranched covers are covers with empty branch set ) . [ rmk : branched - cover]the definition of branched cover used in @xcite is the definition of weak cover given above .
however , kitakubo only considered regular covers .
it is left as an exercise to the reader to show that ( under assumptions [ assumptions ] ) a weak regular cover is in fact a branched cover .
we do not know if negami s conjecture holds for weak covers , and it will probably be a nice project to the reader to find and classify the counterexamples .
weak covers will not appear in this paper again .
we identify @xmath6 with the riemann sphere .
then any rational function @xmath7 gives a branched cover @xmath65 .
conversely given a branched cover @xmath65 we can multiply @xmath7 by @xmath66 ( for an appropriately chosen polynomial @xmath53 ) so that no point in @xmath67 is sent to @xmath68 .
it then follows from the riemann uniformization theorem that ( perhaps after conjugation ) @xmath69 is a polynomial .
hence , after conjugating if necessary , @xmath7 is a rational function .
we conclude this section with a few well - known lemmas about branched covers ; some of the proofs are sketched for the convenience of the reader . for the first lemma ,
the reader may consult , for example , @xcite for the definition of cover from the circle @xmath70 to itself .
[ lem : cover - along - boundary ] let @xmath42 be a branched cover between surfaces with non - empty boundary . then the restriction @xmath71 an unbranched cover from @xmath72 .
[ notation : universal - cover]the _ universal cover _ of @xmath3 is the map @xmath73 given by identifying antipodal points .
it is an unbranched double cover .
one of the basic facts about the universal cover of @xmath3 is : [ lem : covers - of - p2-factor ] let @xmath74 be a cover . then @xmath7 factors through @xmath75 , that is , there exists a cover @xmath76 so that @xmath77 .
let @xmath78 be the branch set .
then @xmath79 is an unbranched cover .
since @xmath80 is non - orientable , it has an _ orientation double cover _ , . ]
that is , an unbranched double cover @xmath81 from some orientable surface @xmath9 .
a basic property of the orientation double cover is that any cover from an orientable surface factors through it .
is orientable if and only if the corresponding subgroup contains only orientation preserving loops . any such subgroup is contained in the subgroup of all orientation preserving loops , and hence any such cover factors through the orientation double cover . ]
applying this in our setting , we see there a cover @xmath82 so that @xmath83 .
the surfaces @xmath84 , @xmath9 and @xmath85 are not compact .
we compactify them by adding one point to each end . denoting the compactification of @xmath40 by @xmath41
, we get a cover @xmath86 so that @xmath7 factors through @xmath87 .
all that remains to show is that @xmath87 is in fact the universal cover @xmath75 .
we can easily see that @xmath87 is a double cover from a closed orientable surface @xmath41 to @xmath3 .
since curves parallel to the punctures of @xmath84 are orientation preserving they lift to the double cover ; hence this cover is not branched .
euler characteristic is multiplicative under unbranched cover , and so @xmath88 . as @xmath6 is the only connected surface with euler characteristic 2
, we get that @xmath89 . by uniqueness of the universal cover , @xmath90 ( perhaps after conjugation ) .
the proof of the lemma below is an easy exercise in euler characteristic : [ pro : f - covered - bysorp ] if @xmath6 branch covers @xmath9 then either @xmath10 or @xmath11 .
the following lemma tells us when @xmath9 is @xmath3 ; ( 3 ) is particularly convenient since it requires only looking at one point : [ lem : when - is - f - on - orientable ] let @xmath40 , @xmath41 be surfaces and suppose @xmath40 is orientable .
let @xmath91 be a cover .
for @xmath92 , let @xmath93 be a open normal neighborhood of @xmath53 endowed with and orientation ( since @xmath93 is a disk this is always possible ) .
for @xmath94 with @xmath95 let @xmath96 be a disk so that @xmath97 is modeled on @xmath98 ( for some @xmath48 ) .
note that the orientation on @xmath93 induces an orientation on @xmath96 .
then the following conditions are equivalent : 1 .
@xmath41 is non - orientable .
2 . for any @xmath92
, there exist @xmath99 with @xmath100 so that the orientations induced orientations on @xmath101 and @xmath102 define opposite orientation on @xmath40 .
3 . for some @xmath92 there exist @xmath99 with @xmath100 so that the orientations induced orientations on @xmath101 and @xmath102 define opposite orientation on @xmath40 .
@xmath103 : let @xmath104 be an arbitrary point , and let @xmath105 be an orientation reversing loop ( which exists by assumption ) based at @xmath53 . since @xmath40 admits no orientation reversing loops , the lift of @xmath106 is an arc connecting two points ( say @xmath107 and @xmath108 ) that project to @xmath53 .
it is now easy to verify that the @xmath101 and @xmath102 induce opposite orientations on @xmath40 ( in fact , this is equivalent to @xmath106 being orientation reversing ) .
@xmath109 : trivial .
@xmath110 : given @xmath111 as in the statement let @xmath112 be any arc connecting @xmath107 and @xmath108 . then the image of @xmath113 on @xmath41 is an orientation reversing loop , and hence @xmath41 is non - orientable .
let @xmath0 be a finite planar cover .
the embedding of @xmath1 into @xmath6 induces a cyclic order around each vertex of @xmath1 . our first condition ,
property @xmath4 , is a consistency condition requiring that this order induces a cyclic order around each vertex of @xmath27 .
fix @xmath114 and @xmath115 in the preimage of @xmath116 .
there are two obstructions to inducing a cyclic order around @xmath116 .
the first obstruction is local : if @xmath45 is a singular vertex ( say with local degree @xmath48 ) , then every neighbor of @xmath116 has @xmath48 preimages around @xmath45 ; the ordering of the different preimages may contain inconsistencies .
for example , if @xmath116 has three neighbors ( say @xmath117 , @xmath118 , and @xmath119 ) and @xmath120 then there are six lifts of @xmath117 , @xmath118 and @xmath119 adjacent to @xmath45 , say @xmath121 ( with @xmath122 , @xmath123 ) .
these vertices may be cyclically ordered as @xmath124 .
it is now not possible to induce a cyclic order around @xmath116 .
the second obstruction is global : it is possible that the order around each preimage induces an order around @xmath116 but different preimages induce distinct orders .
we now define property @xmath4 : [ dfn : pv ] let @xmath0 be a finite planar cover .
we say that @xmath0 fulfills property @xmath4 if for any @xmath125 the following two conditions are satisfied : 1 . for any @xmath126
, the cyclic order of the neighbors of @xmath45 ( induced by the embedding into @xmath6 ) induces an order around @xmath116 . that is to say , denoting the neighbors of @xmath116 by @xmath127 , after reordering the indices of if necessary , projecting the neighbors of @xmath45 to @xmath27 in order we get @xmath128 @xmath129 @xmath130 .
( this condition is vacuous is @xmath45 is not singular ) .
the order obtained is independent of choice of preimage .
_ example : _ suppose that @xmath125 has four neighbors .
figure [ fig : no - pv ] shows the two possibilities that a cover does not fulfill property @xmath4 .
the vertices @xmath131 in that figure project to @xmath116 ; their neighbors are labeled using the labels @xmath132 , according to which vertex of @xmath2 they project to . using the same labeling scheme , in figure [ fig : yes - pv ]
we show a little part of a cover that fulfills property @xmath4 .
, height=113 ] , height=113 ] let @xmath0 be a finite planar cover fulfilling property @xmath4
. given @xmath133 and @xmath126 , moving around a the neighbors of @xmath45 counterclockwise allows us to distinguish a particular cyclic order from its reverse order on the neighbors of @xmath116 .
we call this the _ counterclockwise cyclic order .
_ we assign @xmath45 an arbitrary sign ( plus or minus ) .. ] if @xmath134 is another preimage of @xmath116 we assign @xmath134 the same sign as @xmath45 if the counterclockwise cyclic order around @xmath134 projects to the same order around @xmath116 , and the opposite sign otherwise . a sign assignment as above
is called _
valid_. we refer the reader to section [ sec : examples ] for examples of valid sign assignments . like property @xmath4 ,
property @xmath5 is a question of consistency : [ dfn : pe ] let @xmath0 be a finite planar cover fulfilling property @xmath4 with valid signs on @xmath135 .
we say that @xmath0 fulfills property @xmath5 if for any edge @xmath136 and any two edges @xmath137 that project to @xmath138 , we have that @xmath139 connects vertices of the same sign if and only if @xmath140 does .
we now state and prove theorem [ thm : sufficiency ] . in that theorem
, we consider a graph @xmath1 embedded in @xmath6 ( @xmath2 embedded in some surface @xmath9 , resp . ) .
the closure of the components of @xmath141 ( @xmath142 , resp . ) are called _ faces _ of @xmath6 ( @xmath9 , resp . ) .
[ thm : sufficiency ] let @xmath1 be a connected planar graph , and let @xmath143 be a finite cover
. then @xmath0 fulfills properties @xmath4 and @xmath5 if and only if there exist a surface @xmath9 containing @xmath2 and a map @xmath144 with the following properties : 1 .
@xmath145 extends @xmath7 , that is , for every point @xmath146 , @xmath147 .
2 . @xmath145 is a branched cover , and hence ( by lemma [ pro : f - covered - bysorp ] ) @xmath148 or @xmath10 .
3 . the intersection of the branch points of @xmath145 with @xmath2 is contained in @xmath27 .
more specifically , it is exactly the branch set of @xmath7 .
the intersection of the singular points of @xmath145 with @xmath1 is contained in @xmath149 .
more specifically , it is exactly the set of singular points of @xmath7 and with the same local degrees .
the faces of @xmath6 ( @xmath9 , resp . )
are all disks , and every such face contains at most one singular point ( branch point , resp . ) . 6 .
[ order ] for each @xmath125 , the cyclic order induced on the neighbors of @xmath116 by @xmath7 and the cyclic order given by the embedding of @xmath2 into @xmath9 are the same .
negami s conjecture for finite planar covers fulfilling properties @xmath4 and @xmath5(corollary [ cor : negami - cnj ] below ) follows easily from points ( 1 ) and ( 2 ) above .
however , since negami s conjecture does not require ( 3)(6 ) , corollary [ cor : negami - cnj ] is _ not _ and `` if and only if '' statement .
[ cor : negami - cnj]let @xmath1 be a connected graph , @xmath150 .
let @xmath0 be a cover fulfilling properties @xmath4 and @xmath5 .
then @xmath2 is projective .
suppose @xmath0 is a finite planar cover fulfilling properties @xmath4 and @xmath5 .
we construct the surface @xmath9 and the map @xmath144 : * step one : vertices . *
we first construct @xmath9 near each vertex of @xmath2 and extend the map @xmath145 from a neighborhood of @xmath149 to this neighborhood of @xmath27 . for
each @xmath44 let @xmath151 be a regular neighborhood of @xmath45 . by the normal neighborhood theorem we may assume each @xmath151 is a closed ( _ sic .
_ ) disk and these disks are disjoint ; moreover , the edges of @xmath1 intersect @xmath151 in radial arcs , as in figure [ fig : d - v ] . ,
height=113 ] [ fig : d - v ] for @xmath125 we pick a small disk @xmath152 and embed a neighborhood of @xmath116 in @xmath152 as follows : @xmath116 is the center of @xmath152 , and @xmath28 intersects @xmath152 in radial arcs , each arc corresponding to the tip of an edge that is incident to @xmath116 .
this can be done in several different ways resulting in distinct cyclic orders around @xmath116 .
we embed @xmath116 in @xmath152 so that the cyclic order around @xmath116 agrees with the order induced by the neighbors of the preimages of @xmath116 ; this is well - defined since the cover fulfills property @xmath4 .
we assign @xmath149 a valid sign convention .
endowing @xmath152 with an orientation gives a specific counterclockwise cyclic order around @xmath116 .
we orient @xmath152 so that the counterclockwise cyclic order around @xmath152 agrees with the counterclockwise cyclic order along induced by the neighbors of a vertex of @xmath153 with plus sign and is opposite the counterclockwise cyclic order induced by vertices of negative sign . since the sign assignment is valid , this is well defined . the surface constructed so far , @xmath154 , is denoted @xmath155 .
next we extend the map @xmath7 to @xmath156 by mapping @xmath151 to @xmath157 by a map modeled @xmath61 ( where @xmath48 is the local degree at @xmath45 ) .
we check each property listed in theorem [ thm : sufficiency ] , in the same order : 1 .
holds by construction : for any point @xmath158 , @xmath159 .
2 . again by construction , @xmath160 is a branched cover .
3 . holds by construction .
4 . holds by construction .
5 . to be verified later .
. holds by construction .
since this is a local property and we will not modify @xmath145 near @xmath149 any more , we will not need to check this property again .
[ rmk : orientation ] @xmath161 is orientation preserving if and only if the sign of @xmath116 is plus .
* step two : edges . * next , we extend the construction of @xmath155 to a neighborhood of @xmath28 and extend the range of @xmath145 from a neighborhood of @xmath162 .
the neighborhood of an edge @xmath136 ( say @xmath163 , for some @xmath164 ) is a closed ( _ sic . _ ) band that connects @xmath152 to @xmath165 .
we first glue the band to @xmath152 and extend the orientation of @xmath152 along the band .
we then glue the band to @xmath165 as follows : choose an edge @xmath166 .
if @xmath167 connects vertices with the same sign , the orientation of the band agrees with the orientation of @xmath165 and if @xmath167 connects vertices with the opposite signs , the orientation of the band disagrees with the orientation of @xmath165 .
are oriented , we can place them on a coffee table all facing up .
for the edge @xmath168 , the band @xmath169 connects @xmath165 to @xmath152 .
if @xmath167 connects vertices of the same sign , this band is untwisted ( i.e. , lies flat on the table ) and if @xmath167 connects vertices of the opposite sign , the band is twisted . ]
since @xmath7 fulfills property @xmath5 this construction is independent of choice .
we denote the part of @xmath9 constructed so far @xmath170 .
again , we check each property listed in theorem [ thm : sufficiency ] , ignoring the properties we are done with . 1 .
holds by construction for any point of @xmath1 .
since this is a local property and we will not modify @xmath145 on @xmath171 any more , we will not need to check this property again .
2 . the map @xmath172 is a branched cover by construction .
( 4 ) since we did not introduce any new branch points or singular points , ( 3 ) and ( 4 ) still hold . we will not change @xmath145 near @xmath1 so do not need to check these properties again , 4 .
note that so far , all singular ( resp .
branch ) points are vertices of @xmath1 ( resp .
@xmath2 ) .
( 5 ) will be verified later .
* step three : closing @xmath170 .
* we have constructed a branched cover @xmath173 , both compact surfaces with non - empty boundary .
let @xmath174 denote a boundary component of @xmath175 .
by lemma [ lem : cover - along - boundary ] @xmath176 maps @xmath174 to a boundary component of @xmath170 , say @xmath106 , and the map @xmath177 is a cover .
since @xmath174 and @xmath106 are both circle , @xmath178 is modeled on the restriction of @xmath61 to the unit circle ( for some @xmath179 , called the _ winding number _ of @xmath178 ) .
since @xmath1 is connected and @xmath180 is a neighborhood of @xmath1 , @xmath181 is connected as well ; hence the components of @xmath182 are all disks .
let @xmath183 be the closed disk bound by @xmath174 disjoint from @xmath1 , that is , the closure of the component of @xmath184 adjacent to @xmath174 .
we attach a disk ( say @xmath185 ) to @xmath106 and extend the map @xmath145 by mapping @xmath183 to @xmath185 by coning . and @xmath186 on the unit disk in @xmath187 .
by _ coning _ we mean that @xmath145 is modeled on @xmath98 , extending @xmath188 . ]
if the absolute value of the winding number @xmath178 is more than one we introduce exactly one singular point on @xmath183 and one branch point on @xmath185 , otherwise no new singular or branch point is introduced . continuing this way
, we cap off every component of the boundary of @xmath170 , finally constructing a closed surface @xmath9 containing @xmath2 and a branched cover @xmath8 . with the exception of ( 2 ) and ( 5 ) , we verified that the cover @xmath144 fulfills all the conditions of theorem [ thm : sufficiency ] .
we note that @xmath189 is a branched cover , and by construction , @xmath190 is a branched cover as well ; this establishes ( 2 ) .
it is straightforward to see that the construction gives ( 5 ) as well .
conversely , given @xmath0 and a map @xmath145 as in the statement of theorem [ thm : sufficiency ] , the embedding of @xmath2 into @xmath9 induces a cyclic order around the vertices of @xmath2 . lifting these orders to @xmath1 we get the order around each vertex of @xmath1 ; this order coincides with the order given by the embedding of @xmath1 into @xmath6 .
it is now easy to see that properties @xmath4 and @xmath5 follow .
this completes the proof of theorem [ thm : sufficiency ] .
we can know if @xmath10 or @xmath11 by looking at labels only : [ pro : p2iff - both - signs ] the following are equivalent : 1 .
the surface @xmath9 constructed in theorem [ thm : sufficiency ] is homeomorphic to @xmath3 .
2 . there exists a vertex @xmath125 with preimages of opposite signs .
every vertex @xmath125 has preimages of opposite signs .
this is immediate from remark [ rmk : orientation ] and lemma [ lem : when - is - f - on - orientable ] .
in this section we show that if @xmath12 is a finite planar regular cover then it fulfills properties @xmath4 and @xmath5 .
some of the material in this section is from @xcite and @xcite .
[ pro : regular - is - nice ] let @xmath0 be a regular planar cover . then ( perhaps after re - embedding @xmath1 ) @xmath0 fulfills properties @xmath4 and @xmath5 .
recall that a cover @xmath0 is called _ regular _ if there is a group @xmath191 acting on @xmath1 so that for any @xmath192 with @xmath193 there exist @xmath194 with @xmath195 , and for any @xmath196 with @xmath197 there exist @xmath194 with @xmath198 .
[ lem : translations - extend ] let @xmath0 be a finite planar regular cover with group @xmath191 .
then ( perhaps after re - embedding @xmath1 ) the action of @xmath191 can be extended to an action on @xmath6 .
the proof is an induction on the number of vertices of @xmath1 ; when @xmath1 is not 3-connected we reduce this number .
hence the base case of the induction is : * base case : * @xmath1 is 3-connected .
this is a well - known theorem of whitney @xcite .
* inductive step : * we assume @xmath1 is not 3-connected . the proofs for 2-connected and 1-connected graphs are similar , and we omit the easier case of 1-connected graphs .
let @xmath199 be a cut pair for @xmath1 and suppose removing @xmath200 from @xmath1 we obtain the graphs @xmath201 and @xmath202 ; we assume further that @xmath203 were chosen so that @xmath201 is minimal with respect to inclusion
. is 2-connected , but @xmath202 need not be connected . ] note that @xmath204 and @xmath205 . for any @xmath194
, @xmath206 is a cut pair .
therefore minimality of @xmath201 implies that @xmath207 ( @xmath208 ) .
we use the notation @xmath209 ( called the _ extension _ of @xmath201 ) for the subgraph spun by @xmath210 and similarly @xmath211 is the graph spun by @xmath212 . since @xmath1 is 2-connected there is a pass ( say @xmath213 ) in @xmath214 ( equivalently , in @xmath211 ) connecting @xmath215 to @xmath216 . by choosing the shortest pass , we can guarantee that @xmath213 is embedded . therefore , the graph obtained by adding an edge ( say @xmath167 ) that connects @xmath215 to @xmath216 to @xmath209 is planar .
we denote this graph by @xmath217 .
let @xmath218 be the ( possibly trivial ) subgroup of @xmath191 that leaves @xmath209 invariant ( equivalently , leaves @xmath201 invariant ) .
then for any @xmath219 , @xmath220 , and hence @xmath221 . therefore the action of @xmath218 can be extended to an action on @xmath222 . since @xmath223 , latexmath:[$|v(\mbox{ext}(\widetilde{g}_0 )
\cup \ { \tilde{e } \})| < vertices ) and we may apply the inductive hypothesis to get a re - embedding of @xmath222 into @xmath6 so that the action of @xmath218 extends to @xmath6 . removing @xmath167 , we obtain a re - embedding of @xmath209 , denoted @xmath225 .
[ rmk : hat - is - same ] it is not hard to argue that @xmath226 is 3-connected .
hence by whitney @xcite ( the base case of the induction ) @xmath225 is the original embedding of @xmath209 .
since @xmath1 is 2-connected there is a pass ( say @xmath227 ) in @xmath209 connecting @xmath215 to @xmath216 . by choosing the shortest pass , we can guarantee that @xmath227 is embedded .
similar to the process above , we replace @xmath201 with an edge ( say @xmath228 ) connecting @xmath215 and @xmath216 , obtaining the planar graph @xmath229 . of course
, we can not expect @xmath191 to act on this graph . to that end , we repeat this operation on the image of @xmath201 under @xmath191 and obtain the planar graph : @xmath230denote this graph by @xmath231 , and note that by construction @xmath191 acts on @xmath231 . since @xmath232 , @xmath233 and we may apply the inductive hypothesis to get a re - embedding of @xmath231 into @xmath6 so that the action of @xmath191 extends to @xmath6 . for every @xmath194 , we replace @xmath228 by @xmath234 .
it is now easy to see that we obtain an embedding of @xmath1 into @xmath6 and the action of @xmath191 on this graph extends to @xmath6 , as desired .
this completes the proof of lemma [ lem : translations - extend ] . after extending the action of @xmath191 to @xmath6
, we denote the group elements by @xmath235 .
[ lem : extends - implies - pev ] let @xmath0 be a finite planar regular cover with group @xmath191 . if any @xmath194 can be extended to a homeomorphism @xmath235 then the cover fulfills properties @xmath4 and @xmath5 .
first , let @xmath45 be a singular vertex ( say @xmath45 projects to @xmath116 ) .
we need to show that the cyclic order of the neighbors of @xmath45 induces an order on the neighbors @xmath116 .
let @xmath236 be consecutive vertices in the counterclockwise cyclic order around @xmath45 and let @xmath237 be a neighbor of @xmath45 so that @xmath238 then the vertex that follows @xmath237 in the counterclockwise cyclic order around @xmath45 projects to the same vertex as @xmath239 .
proof of claim 1 : since @xmath240 and @xmath237 project to the same vertex , there exists @xmath194 so that @xmath241 .
denote the edge @xmath242 by @xmath243 and the image of @xmath244 under @xmath106 by @xmath245 ( that is , @xmath246 ) . if @xmath247 then @xmath248 projects to a cycle of length 1 or 2 , contradicting our assumption ( recall assumptions [ assumptions ] ) .
we conclude that @xmath106 fixes @xmath45 . by assumption
, there exists @xmath249 a homeomorphism that extends @xmath106 ; thus @xmath250 , @xmath251 , and @xmath252 is a vertex that projects to the same vertex as @xmath239 , and follows @xmath237 in the counterclockwise cyclic order around @xmath45 .
this proves claim 1 .
it follows immediately from claim 1 that the counterclockwise cyclic order around @xmath45 indices a cyclic order around @xmath116 .
next , let @xmath253 be distinct vertices that project to the same vertex @xmath125 .
then there exists @xmath254 so that @xmath255 , and by assumption there exists @xmath235 extending @xmath106 .
it is easy to see that @xmath256 induces an order preserving bijection between the neighbors of @xmath45 and the neighbors of @xmath134 .
therefore the neighbors of @xmath45 and the neighbors of @xmath134 induce the same order on the neighbors of @xmath116 ; this establishes property @xmath4 .
since @xmath0 fulfills property @xmath4 , we can assign a valid sign assignment for @xmath149 ( as described in section [ sec : sufficiency ] ) .
let @xmath137 be edges that project to the same edge @xmath136 ; say @xmath257 and @xmath258 .
than there exists @xmath254 so that @xmath259 , equivalently @xmath260 . if @xmath106 is preserves the orientation of @xmath6 , the sign of @xmath240 is the same as the sign of @xmath261 and the sign of @xmath239 is the same as the sign of @xmath262 ; if @xmath106 reverses the orientation of @xmath6 the sign of @xmath240 is the opposite the sign of @xmath261 and the sign of @xmath239 is opposite the sign of @xmath262 . in both cases , @xmath139 connects vertices of the same sign if and only if @xmath140 does .
this establishes property @xmath5 and completes the proof of lemma [ lem : extends - implies - pev ] .
clearly , proposition [ pro : regular - is - nice ] follows from lemmas [ lem : translations - extend ] and [ lem : extends - implies - pev ] .
in this section , we try to better our situation by passing to higher covers .
let @xmath0 be a finite cover .
when does there exist a finite planar cover @xmath263 so that the composition @xmath264 fulfills properties @xmath4 and @xmath5 ?
assume such cover exists . to address this question ,
we wish to apply theorem [ thm : sufficiency ] to the cover @xmath263 .
however , it does not follow that @xmath263 fulfills properties @xmath4 and @xmath5 . to guarantee that
, we need to add the assumption that the cover @xmath0 is not branched .
recall from notation [ notation : universal - cover ] that the map @xmath265 given by identifying antipodal points is an unbranched , double cover called the _ universal cover _ of @xmath3 .
therefore , given a graph @xmath266 , @xmath267 is a graph that double covers @xmath1 and naturally embeds in @xmath6 [ thm : necessity ] let @xmath0 be an unbranched finite ( not necessarily planar ) cover and @xmath268 be a finite planar covers .
suppose that the composition @xmath269 fulfills properties @xmath4 and @xmath5 .
then one of the following holds : 1 .
@xmath1 is planar and for some embedding of @xmath1 into @xmath6 , @xmath0 fulfills properties @xmath4 and @xmath5 .
2 . @xmath1 embeds in @xmath3 so that the cover @xmath270 fulfills properties @xmath4 and @xmath5 .
we begin by showing : [ pro : tilde - f - has - pev ] let @xmath0 be an unbranched finite planar cover and @xmath263 be a finite planar covers .
suppose that the composition @xmath271 fulfills properties @xmath4 and @xmath5
. then @xmath263 fulfills properties @xmath4 and @xmath5 .
let @xmath272 be two vertices that project to the same vertex under @xmath273 ( say @xmath45 ) and denote @xmath46 by @xmath116 .
by assumption , @xmath274 fulfills property @xmath4 .
therefore , the cyclic order on the neighbors @xmath275 induces a cyclic order on the neighbors of @xmath116 . since @xmath7 is unbranched
, the restriction of @xmath7 to the neighbors of @xmath45 is a bijection to the neighbors of @xmath116 .
hence , the cyclic order around @xmath276 ( which is a subset of @xmath277 ) induces a cyclic order on the neighbors of @xmath45 .
this establishes property @xmath4 . by construction
the bijection induced by @xmath7 between the neighbors of @xmath45 and the neighbors of @xmath46 is order preserving . before establishing property @xmath5
we must assign valid signs to the vertices of @xmath16 . since @xmath278 fulfills property @xmath5 , some signs
have been assigned already , and these signs are valid for @xmath279 .
we show that the same signs are valid for @xmath273 .
let @xmath280 be vertices so that @xmath281
. then @xmath282 and therefore @xmath283 and @xmath284 have the same ( resp .
opposite ) sign if and only if the counterclockwise cyclic order around them is the same ( resp .
opposite ) under @xmath285 .
the order preserving bijection induced on the neighbors of @xmath45 by @xmath7 shows that the sign choice is valid for @xmath273 as well .
let @xmath286 be two edges that project to the same edge ( say @xmath167 ) under @xmath273
. by assumption @xmath278 fulfills property @xmath5 .
since @xmath287 and @xmath288 project to the same edge under @xmath285 , @xmath287 connects vertices of the same sign if and only if @xmath288 does .
hence @xmath263 fulfills property @xmath5 .
this completes the proof of proposition [ pro : tilde - f - has - pev ] . by proposition
[ pro : tilde - f - has - pev ] we may apply theorem [ thm : sufficiency ] to @xmath289 and get an embedding of @xmath1 into @xmath9 , where @xmath290 or @xmath11 .
* case one : @xmath10 : * this case corresponds to case ( 1 ) of theorem [ thm : necessity ] .
we need to show that @xmath0 fulfills properties @xmath4 and @xmath5 . by theorem [ thm : sufficiency ]
( [ order ] ) the embedding of @xmath1 into @xmath6 and the map @xmath273 induce the same order around the vertices of @xmath1 .
similarly , applying theorem [ thm : sufficiency ] to @xmath285 we obtain an embedding of @xmath2 into some surface @xmath291 , and the embedding of @xmath2 into @xmath291 and the map @xmath292 induce the same order around the vertices of @xmath2 . to verify property @xmath4 , fix @xmath44 and denote @xmath46 by @xmath116 . by assumption , the cover @xmath7 is not branched .
therefore @xmath7 induces a bijection between the neighbors of @xmath45 and the neighbors of @xmath116 . since the orders around @xmath45 and around @xmath116 are both induced from the order around their preimages in @xmath16 , this bijection is order preserving
. therefore , @xmath7 fulfills property @xmath4 .
next we need a valid sign assignment on @xmath149 .
fix @xmath293 . by assumption , @xmath10 and therefore by proposition [ pro : p2iff - both - signs ] all the preimages of @xmath45 have the same sign .
we assign @xmath45 that sign .
we need to verify that the assignment is valid .
fix @xmath294 that project to the same vertex , say @xmath116 .
let @xmath152 be a small disk neighborhood of @xmath116 oriented so that the restriction of @xmath285 to a component of @xmath295 is orientation preserving if and only if the sign at the corresponding preimage of @xmath116 is plus ( this is possible since the sign assignment on @xmath296 is valid for @xmath279 ) .
we choose an orientation for @xmath10 so that @xmath273 is orientation preserving .
it is easy to see that given @xmath45 with @xmath297 , @xmath298 is orientation preserving if and only if only if the sign at @xmath45 is plus ; it follows that the sign assignment is valid .
finally we verify property @xmath5 . given @xmath136 and @xmath299 that project to @xmath138 , the signs at the endpoints of @xmath167 are the same as the signs at the endpoints of of any edge that projects to @xmath167 under @xmath273 , and similarly for @xmath167 .
property @xmath5 for @xmath7 follows from property @xmath5for @xmath273 ( that was established is proposition [ pro : tilde - f - has - pev ] ) .
* case two : @xmath11 : * this case corresponds to case ( 2 ) of theorem [ thm : necessity ] .
let @xmath300 be the lift of @xmath1 to @xmath3 , _
i.e. _ , @xmath301 .
( recall the definition of the universal cover @xmath73 in notation [ notation : universal - cover ] . ) although not essential to the proof , we show that @xmath300 is connected .
every face of @xmath6 cut open along @xmath300 is an unbranched cover of a face of @xmath1 , which by theorem [ thm : sufficiency ] ( 5 ) ( applied to @xmath273 ) is a disk .
therefore the faces of @xmath300 are disks as well and @xmath300 is connected . by lemma [ lem : covers - of - p2-factor ] @xmath273 factors through @xmath75 ; that is , there exist a cover @xmath302 so that @xmath303
. then @xmath304 .
therefore the orders induced on the neighbors of every vertex of @xmath305 by the embedding into @xmath6 and by @xmath306 are the same , and we conclude that @xmath307 induces that same order on the neighbors of every vertex in @xmath27 as @xmath285 ; in particular , @xmath308 induces _ some _ order on the neighbors of every vertex in @xmath27 and therefore fulfills properties @xmath4 and @xmath5 .
this completes the proof of theorem [ thm : necessity ] .
our first example is very simple .
it shows a cover of a graph with one vertex and two edges ( the bouquet of two circles ) .
the cover is given in figure [ fig : irregular ] and is an irregular triple cover that fulfills properties @xmath4 and @xmath5 .
( we note that any double cover is regular . ) in that figure , we use the following labels @xmath309 and @xmath310 , where the three edges labeled @xmath309 project to the same edge and the three edges labeled @xmath310 project to the other edge . the the bouquet of two circles is not shown in the figure .
[ ex : irregular ] let @xmath2 be the bouquet of two circles and @xmath1 the triple cover given in figure [ fig : irregular ] ; the projection @xmath7 is indicated by labels and arrows . then @xmath0 is an irregular cover fulfilling properties @xmath4 and @xmath5 . of course
, the bouquet of two circles can not be regarded as an `` interesting '' graph .
the reader can soup this example up by replacing the edges of the bouquet of two circles by planar graphs .
however , the resulting cover is not 3-connected .
we ask : all the remaining examples in this section are covers of @xmath21 .
the labels used are as follows : the vertices of @xmath21 are labeled @xmath309 , @xmath310 , @xmath311 , and @xmath48 and vertices of the covers are labeled by the vertex they project to ; edges are not labeled .
[ ex : k4 ] figures [ fig : k4 - 1 ] and [ fig : k4 - 2 ] give two regular , unbranched , planar double covers of @xmath21 .
the first cover yields an embedding of @xmath21 into @xmath6 while the second cover yields an embedding of @xmath21 into @xmath3 .
we see by inspection that both covers fulfill properties @xmath4 and @xmath5 .
it follows from proposition [ pro : p2iff - both - signs ] the first cover gives an embedding of @xmath21 into @xmath6 while the second embeds @xmath21 in @xmath3 .
alternatively , we can see that the second cover yields an embedding into @xmath3 by observing that it has two cycles of length four that bound faces .
these cycles project to two cycles of length two or a single cycle of length four in @xmath21 ( in fact , they project to a single cycle of length four ) . by theorem [ thm : necessity ] ( 5 )
this cycle bounds a face .
but the unique embedding of @xmath21 into @xmath6 has only triangular faces .
the next example , example [ ex : have - to - lift ] , is the triple cover @xmath22 given in figure [ fig : have - to - lift1 ] .
we observe that with the given embedding the cover does not fulfill property @xmath5 .
this cover is not 3-connected so the embedding into @xmath6 is not unique .
however , the only other embedding is given by `` flipping '' the disk contained in the dashed circle .
it is easy to see that it too does not fulfill property @xmath5 .
we seek a planar cover @xmath312 so that the composition @xmath313 fulfills properties @xmath4 and @xmath5 .
although covers of this type may have arbitrarily high degree , by theorem [ thm : necessity ] if such cover exists , then a double cover with this property exists as well .
moreover , theorem [ thm : necessity ] tells us how to find this cover : let @xmath73 be the universal cover ; embed @xmath23 in @xmath3 and lift to @xmath6 ; this is the cover we need .
[ ex : have - to - lift ] the cover given in figure [ fig : have - to - lift1 ] does not fulfill properties @xmath4 and @xmath5 for any embedding of @xmath1 into @xmath6 .
however , the embedding of @xmath23 into @xmath3 given in figure [ fig : have - to - lift2 ] ( where we view @xmath3 as a disk union a mbius band ) has a double cover ( given in figure [ fig : have - to - lift4 ] ) that embeds in @xmath6 and fulfills properties @xmath4 and @xmath5 .
figure [ fig : have - to - lift4 ] was constructed by taking two copies of the disk in figure [ fig : have - to - lift2 ] and an annulus that double covers the mbius band in figure [ fig : have - to - lift2 ] ; see figure [ fig : have - to - lift3 ]
. pasting the disks to the annulus in figure [ fig : have - to - lift3 ] gives figure [ fig : have - to - lift4 ] .
our final example is far more disturbing in nature .
we give a cover @xmath22 that does not fulfill properties @xmath4 and @xmath5 ( quite similar to example [ ex : have - to - lift ] ) and _ conjecture _ that for any finite planar cover @xmath314 , the composition @xmath313 does not fulfilling properties @xmath4 and @xmath5 . [ ex : no properties ] consider the cover @xmath23 given in figure [ fig : no - properties ] .
let @xmath314 be a finite planar cover .
then the composition @xmath315 does not fulfill properties @xmath4 and @xmath5 .
we end this paper by describing two ways of dealing with the cover @xmath23 given in figure [ fig : no - properties ] .
first , note that if the four disks enclosed in dashed circles in figure [ fig : no - properties ] are replaced by the disks given in figure [ fig : fixme ] ( and some of the `` mushrooms '' are reflected ) , we obtain the cover shown in figure [ fig : fixed ] that does fulfill properties @xmath4 and @xmath5 .
the work in this paper suggests that topological `` cut - and - paste '' techniques such as this could be very useful .
next , recall figures [ fig : have - to - lift1 ] and [ fig : have - to - lift2 ] where we replaced a disk with a mbius band . similarly , we can replace the four disks in figure [ fig : no - properties ] with four mbius bands , obtaining an embedding of @xmath23 into a non - orientable surface @xmath316 ; since replacing a disk with a mbius band lowers the euler characteristic by 1 , @xmath317 .
just like @xmath318 , @xmath316 has an oriented double cover ( say @xmath319 ) .
the euler characteristic is multiplicative under unbranched covers , and we conclude that @xmath319 is the surface of genus 3 .
( in general , if we replace @xmath320 disks with mbius bands the oriented double cover will have genus @xmath321 . )
the reader can verify that the proof of theorem [ thm : sufficiency ] is valid , and gives a surface @xmath9 , an embedding @xmath322 , and a cover @xmath323 that extends @xmath7 .
as in lemma [ pro : f - covered - bysorp ] , @xmath324 .
( in general , @xmath325 . )
this does not give an embedding of @xmath2 into @xmath3 , but does give an embedding of @xmath2 into a surface with some control over its euler characteristic . | given a finite cover @xmath0 and an embedding of @xmath1 in the plane , negami conjectures that @xmath2 embeds in @xmath3 .
negami proved this conjecture for regular covers . in this paper
we define two properties ( called properties @xmath4 and @xmath5 ) , depending on the cover @xmath1 and its embedding into @xmath6 , and generalize negami s result by showing : ( 1 ) if properties @xmath4 and @xmath5are fulfilled then @xmath2 embeds in @xmath3 .
( 2 ) regular covers always fulfill properties @xmath4 and @xmath5 .
we give an example of an irregular cover fulfilling properties @xmath4 and @xmath5 .
covers not fulfilling properties @xmath4 and @xmath5 are discussed as well . |
Subsets and Splits
No saved queries yet
Save your SQL queries to embed, download, and access them later. Queries will appear here once saved.